sanchayanmaity/minix
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity
minix/servers/vfs/select.c

1065 lines
32 KiB
C
Raw Normal View History

small compile fixes; initial select() stubs; release.sh script doesn't need /tmp any more since 16MB root device; increase to 3.0.5 to make new CD with working FXP driver. (not tagged 3.0.5 yet as at driver bios-copy workaround hasn't been done.)
2005-06-06 13:40:32 +02:00
/* Implement entry point to select system call.
*
* The entry points into this file are
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
* do_select: perform the SELECT system call
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
* select_callback: notify select system of possible fd operation
endpoint-aware conversion of servers. 'who', indicating caller number in pm and fs and some other servers, has been removed in favour of 'who_e' (endpoint) and 'who_p' (proc nr.). In both PM and FS, isokendpt() convert endpoints to process slot numbers, returning OK if it was a valid and consistent endpoint number. okendpt() does the same but panic()s if it doesn't succeed. (In PM, this is pm_isok..) pm and fs keep their own records of process endpoints in their proc tables, which are needed to make kernel calls about those processes. message field names have changed. fs drivers are endpoints. fs now doesn't try to get out of driver deadlock, as the protocol isn't supposed to let that happen any more. (A warning is printed if ELOCKED is detected though.) fproc[].fp_task (indicating which driver the process is suspended on) became an int. PM and FS now get endpoint numbers of initial boot processes from the kernel. These happen to be the same as the old proc numbers, to let user processes reach them with the old numbers, but FS and PM don't know that. All new processes after INIT, even after the generation number wraps around, get endpoint numbers with generation 1 and higher, so the first instances of the boot processes are the only processes ever to have endpoint numbers in the old proc number range. More return code checks of sys_* functions have been added. IS has become endpoint-aware. Ditched the 'text' and 'data' fields in the kernel dump (which show locations, not sizes, so aren't terribly useful) in favour of the endpoint number. Proc number is still visible. Some other dumps (e.g. dmap, rs) show endpoint numbers now too which got the formatting changed. PM reading segments using rw_seg() has changed - it uses other fields in the message now instead of encoding the segment and process number and fd in the fd field. For that it uses _read_pm() and _write_pm() which to _taskcall()s directly in pm/misc.c. PM now sys_exit()s itself on panic(), instead of sys_abort(). RS also talks in endpoints instead of process numbers.
2006-03-03 11:20:58 +01:00
* select_unsuspend_by_endpt: cancel a blocking select on exiting driver
small compile fixes; initial select() stubs; release.sh script doesn't need /tmp any more since 16MB root device; increase to 3.0.5 to make new CD with working FXP driver. (not tagged 3.0.5 yet as at driver bios-copy workaround hasn't been done.)
2005-06-06 13:40:32 +02:00
*/
#include "fs.h"
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
#include <sys/time.h>
#include <sys/select.h>
#include <minix/com.h>
First cut at 64-bit file offsets in block devices for mkfs/fsck.
2006-11-27 15:21:43 +01:00
#include <minix/u64.h>
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
#include <string.h>
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
#include <assert.h>
#include "file.h"
#include "fproc.h"
#include "dmap.h"
#include "vnode.h"
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/* max. number of simultaneously pending select() calls */
#define MAXSELECTS 25
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
#define FROM_PROC 0
#define TO_PROC 1
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
PRIVATE struct selectentry {
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
struct fproc *requestor; /* slot is free iff this is NULL */
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
endpoint_t req_endpt;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
fd_set readfds, writefds, errorfds;
fd_set ready_readfds, ready_writefds, ready_errorfds;
fd_set *vir_readfds, *vir_writefds, *vir_errorfds;
struct filp *filps[OPEN_MAX];
int type[OPEN_MAX];
int nfds, nreadyfds;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int error;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
char block;
clock_t expiry;
timer_t timer; /* if expiry > 0 */
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
} selecttab[MAXSELECTS];
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
FORWARD _PROTOTYPE(int copy_fdsets, (struct selectentry *se, int nfds,
int direction) );
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
FORWARD _PROTOTYPE(int do_select_request, (struct selectentry *se, int fd,
int *ops) );
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
FORWARD _PROTOTYPE(void filp_status, (struct filp *fp, int status) );
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
FORWARD _PROTOTYPE(int is_deferred, (struct selectentry *se) );
FORWARD _PROTOTYPE(void restart_proc, (struct selectentry *se) );
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
FORWARD _PROTOTYPE(void ops2tab, (int ops, int fd, struct selectentry *e));
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
FORWARD _PROTOTYPE(int is_regular_file, (struct filp *f) );
FORWARD _PROTOTYPE(int is_pipe, (struct filp *f) );
FORWARD _PROTOTYPE(int is_supported_major, (struct filp *f) );
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
FORWARD _PROTOTYPE(void select_lock_filp, (struct filp *f, int ops) );
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
FORWARD _PROTOTYPE(int select_request_async, (struct filp *f, int *ops,
int block) );
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
FORWARD _PROTOTYPE(int select_request_file, (struct filp *f, int *ops,
int block) );
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
FORWARD _PROTOTYPE(int select_request_major, (struct filp *f, int *ops,
int block) );
FORWARD _PROTOTYPE(int select_request_pipe, (struct filp *f, int *ops,
int block) );
FORWARD _PROTOTYPE(int select_request_sync, (struct filp *f, int *ops,
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
int block) );
FORWARD _PROTOTYPE(void select_cancel_all, (struct selectentry *e) );
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
FORWARD _PROTOTYPE(void select_cancel_filp, (struct filp *f) );
FORWARD _PROTOTYPE(void select_return, (struct selectentry *) );
FORWARD _PROTOTYPE(void select_restart_filps, (void) );
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
FORWARD _PROTOTYPE(int tab2ops, (int fd, struct selectentry *e) );
FORWARD _PROTOTYPE(void wipe_select, (struct selectentry *s) );
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
PRIVATE struct fdtype {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
int (*select_request)(struct filp *, int *ops, int block);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int (*type_match)(struct filp *f);
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
} fdtypes[] = {
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
{ select_request_major, is_supported_major },
{ select_request_file, is_regular_file },
{ select_request_pipe, is_pipe },
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
};
Fixed some select bugs related to pipes. Removed SELFD_* because they were bogus and unused.
2006-06-07 16:41:47 +02:00
#define SEL_FDS (sizeof(fdtypes) / sizeof(fdtypes[0]))
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE int select_majors[] = { /* List of majors that support selecting on */
TTY_MAJOR,
INET_MAJOR,
UDS_MAJOR,
LOG_MAJOR,
};
#define SEL_MAJORS (sizeof(select_majors) / sizeof(select_majors[0]))
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/*===========================================================================*
* do_select *
*===========================================================================*/
PUBLIC int do_select(void)
{
/* Implement the select(nfds, readfds, writefds, errorfds, timeout) system
* call. First we copy the arguments and verify their sanity. Then we check
* whether there are file descriptors that satisfy the select call right of the
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
* bat. If so, or if there are no ready file descriptors but the process
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
* requested to return immediately, we return the result. Otherwise we set a
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
* timeout and wait for either the file descriptors to become ready or the
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
* timer to go off. If no timeout value was provided, we wait indefinitely. */
Miscellaneous code cleanup.
2010-03-22 21:43:06 +01:00
int r, nfds, do_timeout = 0, fd, s;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
struct timeval timeout;
struct selectentry *se;
nfds = m_in.SEL_NFDS;
/* Sane amount of file descriptors? */
if (nfds < 0 || nfds > OPEN_MAX) return(EINVAL);
/* Find a slot to store this select request */
for (s = 0; s < MAXSELECTS; s++)
if (selecttab[s].requestor == NULL) /* Unused slot */
break;
if (s >= MAXSELECTS) return(ENOSPC);
se = &selecttab[s];
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
wipe_select(se); /* Clear results of previous usage */
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
se->req_endpt = who_e;
se->vir_readfds = (fd_set *) m_in.SEL_READFDS;
se->vir_writefds = (fd_set *) m_in.SEL_WRITEFDS;
se->vir_errorfds = (fd_set *) m_in.SEL_ERRORFDS;
/* Copy fdsets from the process */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if ((r = copy_fdsets(se, nfds, FROM_PROC)) != OK) return(r);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/* Did the process set a timeout value? If so, retrieve it. */
if (m_in.SEL_TIMEOUT != NULL) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
do_timeout = 1;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
r = sys_vircopy(who_e, D, (vir_bytes) m_in.SEL_TIMEOUT, SELF, D,
(vir_bytes) &timeout, sizeof(timeout));
if (r != OK) return(r);
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* No nonsense in the timeval */
if (do_timeout && (timeout.tv_sec < 0 || timeout.tv_usec < 0))
return(EINVAL);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/* If there is no timeout, we block forever. Otherwise, we block up to the
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
* specified time interval.
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (!do_timeout) /* No timeout value set */
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
se->block = 1;
else if (do_timeout && (timeout.tv_sec > 0 || timeout.tv_usec > 0))
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
se->block = 1;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
else /* timeout set as (0,0) - this effects a poll */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
se->block = 0;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
se->expiry = 0; /* no timer set (yet) */
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Verify that file descriptors are okay to select on */
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
for (fd = 0; fd < nfds; fd++) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
struct filp *f;
int type, ops;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Because the select() interface implicitly includes file descriptors
* you might not want to select on, we have to figure out whether we're
* interested in them. Typically, these file descriptors include fd's
* inherited from the parent proc and file descriptors that have been
* close()d, but had a lower fd than one in the current set.
*/
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (!(ops = tab2ops(fd, se)))
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
continue; /* No operations set; nothing to do for this fd */
/* Get filp belonging to this fd */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
f = se->filps[fd] = get_filp(fd, VNODE_READ);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (f == NULL) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (err_code == EBADF)
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
r = err_code;
else /* File descriptor is 'ready' to return EIO */
r = EINTR;
return(r);
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Check file types. According to POSIX 2008:
* "The pselect() and select() functions shall support regular files,
* terminal and pseudo-terminal devices, FIFOs, pipes, and sockets. The
* behavior of pselect() and select() on file descriptors that refer to
* other types of file is unspecified."
*
* In our case, terminal and pseudo-terminal devices are handled by the
* TTY major and sockets by either INET major (socket type AF_INET) or
* PFS major (socket type AF_UNIX). PFS acts as an FS when it handles
* pipes and as a driver when it handles sockets. Additionally, we
* support select on the LOG major to handle kernel logging, which is
* beyond the POSIX spec. */
se->type[fd] = -1;
for (type = 0; type < SEL_FDS; type++) {
if (fdtypes[type].type_match(f)) {
se->type[fd] = type;
se->nfds = fd+1;
se->filps[fd]->filp_selectors++;
break;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
}
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
unlock_filp(f);
if (se->type[fd] == -1) /* Type not found */
Fix select-related bugs: missing cancellations led to potentially forgetting notifies, especially in the case of async drivers
2010-10-08 14:50:52 +02:00
return(EBADF);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Check all file descriptors in the set whether one is 'ready' now */
for (fd = 0; fd < nfds; fd++) {
fix clang warnings -R in kernel/ and servers/
2011-06-09 16:09:13 +02:00
int ops, r;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
struct filp *f;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Again, check for involuntarily selected fd's */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (!(ops = tab2ops(fd, se)))
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
continue; /* No operations set; nothing to do for this fd */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
/* Test filp for select operations if not already done so. e.g.,
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* processes sharing a filp and both doing a select on that filp. */
f = se->filps[fd];
if ((f->filp_select_ops & ops) != ops) {
fix clang warnings -R in kernel/ and servers/
2011-06-09 16:09:13 +02:00
int wantops;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
wantops = (f->filp_select_ops |= ops);
r = do_select_request(se, fd, &wantops);
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
if (r != OK) {
if (r == SUSPEND) continue;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
else break; /* Error or bogus return code; abort */
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/* The select request above might have turned on/off some
* operations because they were 'ready' or not meaningful.
* Either way, we might have a result and we need to store them
* in the select table entry. */
if (wantops & ops) ops2tab(wantops, fd, se);
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if ((se->nreadyfds > 0 || !se->block) && !is_deferred(se)) {
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/* fd's were found that were ready to go right away, and/or
* we were instructed not to block at all. Must return
* immediately.
*/
r = copy_fdsets(se, se->nfds, TO_PROC);
select_cancel_all(se);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (r != OK)
return(r);
else if (se->error != OK)
return(se->error);
return(se->nreadyfds);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/* Convert timeval to ticks and set the timer. If it fails, undo
* all, return error.
*/
if (do_timeout) {
int ticks;
/* Open Group:
* "If the requested timeout interval requires a finer
* granularity than the implementation supports, the
* actual timeout interval shall be rounded up to the next
* supported value."
*/
#define USECPERSEC 1000000
while(timeout.tv_usec >= USECPERSEC) {
/* this is to avoid overflow with *system_hz below */
timeout.tv_usec -= USECPERSEC;
timeout.tv_sec++;
}
ticks = timeout.tv_sec * system_hz +
(timeout.tv_usec * system_hz + USECPERSEC-1) / USECPERSEC;
se->expiry = ticks;
move timers code to libsys
2010-07-09 14:58:18 +02:00
set_timer(&se->timer, ticks, select_timeout_check, s);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* If we're blocking, the table entry is now valid */
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
se->requestor = fp;
/* process now blocked */
suspend(FP_BLOCKED_ON_SELECT);
return(SUSPEND);
}
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* is_deferred *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE int is_deferred(struct selectentry *se)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Find out whether this select has pending initial replies */
int fd;
struct filp *f;
for (fd = 0; fd < se->nfds; fd++) {
if ((f = se->filps[fd]) == NULL) continue;
if (f->filp_select_flags & (FSF_UPDATE|FSF_BUSY)) return(TRUE);
}
return(FALSE);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* is_regular_file *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE int is_regular_file(struct filp *f)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
return(f && f->filp_vno && (f->filp_vno->v_mode & I_TYPE) == I_REGULAR);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/*===========================================================================*
* is_pipe *
*===========================================================================*/
PRIVATE int is_pipe(struct filp *f)
{
/* Recognize either anonymous pipe or named pipe (FIFO) */
return(f && f->filp_vno && (f->filp_vno->v_mode & I_TYPE) == I_NAMED_PIPE);
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* is_supported_major *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE int is_supported_major(struct filp *f)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* See if this filp is a handle on a device on which we support select() */
int m;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (!(f && f->filp_vno)) return(FALSE);
if ((f->filp_vno->v_mode & I_TYPE) != I_CHAR_SPECIAL) return(FALSE);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
for (m = 0; m < SEL_MAJORS; m++)
if (major(f->filp_vno->v_sdev) == select_majors[m])
return(TRUE);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
return(FALSE);
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* select_request_async *
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE int select_request_async(struct filp *f, int *ops, int block)
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int r, rops, major;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
struct dmap *dp;
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
rops = *ops;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (!block && (f->filp_select_flags & FSF_BLOCKED)) {
/* This filp is blocked waiting for a reply, but we don't want to
* block ourselves. Unless we're awaiting the initial reply, these
* operations won't be ready */
if (!(f->filp_select_flags & FSF_BUSY)) {
if ((rops & SEL_RD) && (f->filp_select_flags & FSF_RD_BLOCK))
rops &= ~SEL_RD;
if ((rops & SEL_WR) && (f->filp_select_flags & FSF_WR_BLOCK))
rops &= ~SEL_WR;
if ((rops & SEL_ERR) && (f->filp_select_flags & FSF_ERR_BLOCK))
rops &= ~SEL_ERR;
if (!(rops & (SEL_RD|SEL_WR|SEL_ERR))) {
/* Nothing left to do */
*ops = 0;
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(OK);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
}
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
f->filp_select_flags |= FSF_UPDATE;
if (block) {
rops |= SEL_NOTIFY;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (rops & SEL_RD) f->filp_select_flags |= FSF_RD_BLOCK;
if (rops & SEL_WR) f->filp_select_flags |= FSF_WR_BLOCK;
if (rops & SEL_ERR) f->filp_select_flags |= FSF_ERR_BLOCK;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
if (f->filp_select_flags & FSF_BUSY)
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(SUSPEND);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
major = major(f->filp_vno->v_sdev);
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
if (major < 0 || major >= NR_DEVICES) return(ENXIO);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
dp = &dmap[major];
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
if (dp->dmap_sel_filp)
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(SUSPEND);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
f->filp_select_flags &= ~FSF_UPDATE;
r = dev_io(VFS_DEV_SELECT, f->filp_vno->v_sdev, rops, NULL,
cvu64(0), 0, 0, FALSE);
if (r < 0 && r != SUSPEND)
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(r);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (r != SUSPEND)
panic() cleanup. this change - makes panic() variadic, doing full printf() formatting - no more NO_NUM, and no more separate printf() statements needed to print extra info (or something in hex) before panicing - unifies panic() - same panic() name and usage for everyone - vm, kernel and rest have different names/syntax currently in order to implement their own luxuries, but no longer - throws out the 1st argument, to make source less noisy. the panic() in syslib retrieves the server name from the kernel so it should be clear enough who is panicing; e.g. panic("sigaction failed: %d", errno); looks like: at_wini(73130): panic: sigaction failed: 0 syslib:panic.c: stacktrace: 0x74dc 0x2025 0x100a - throws out report() - printf() is more convenient and powerful - harmonizes/fixes the use of panic() - there were a few places that used printf-style formatting (didn't work) and newlines (messes up the formatting) in panic() - throws out a few per-server panic() functions - cleans up a tie-in of tty with panic() merging printf() and panic() statements to be done incrementally.
2010-03-05 16:05:11 +01:00
panic("select_request_asynch: expected SUSPEND got: %d", r);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
dp->dmap_sel_filp = f;
f->filp_select_flags |= FSF_BUSY;
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(SUSPEND);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* select_request_file *
*===========================================================================*/
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
PRIVATE int select_request_file(struct filp *UNUSED(f), int *UNUSED(ops),
int UNUSED(block))
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
{
/* Files are always ready, so output *ops is input *ops */
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(OK);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
/*===========================================================================*
* select_request_major *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE int select_request_major(struct filp *f, int *ops, int block)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int major, r;
major = major(f->filp_vno->v_sdev);
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
if (major < 0 || major >= NR_DEVICES) return(ENXIO);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (dmap[major].dmap_style == STYLE_DEVA ||
dmap[major].dmap_style == STYLE_CLONE_A)
r = select_request_async(f, ops, block);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
else
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r = select_request_sync(f, ops, block);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
return(r);
}
/*===========================================================================*
* select_request_sync *
*===========================================================================*/
PRIVATE int select_request_sync(struct filp *f, int *ops, int block)
{
int rops;
rops = *ops;
if (block) rops |= SEL_NOTIFY;
*ops = dev_io(VFS_DEV_SELECT, f->filp_vno->v_sdev, rops, NULL,
cvu64(0), 0, 0, FALSE);
if (*ops < 0)
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(*ops);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(OK);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/*===========================================================================*
* select_request_pipe *
*===========================================================================*/
PRIVATE int select_request_pipe(struct filp *f, int *ops, int block)
{
int orig_ops, r = 0, err;
orig_ops = *ops;
if ((*ops & (SEL_RD|SEL_ERR))) {
err = pipe_check(f->filp_vno, READING, 0, 1, f->filp_pos, 1);
if (err != SUSPEND)
r |= SEL_RD;
if (err < 0 && err != SUSPEND)
r |= SEL_ERR;
if (err == SUSPEND && !(f->filp_mode & R_BIT)) {
/* A "meaningless" read select, therefore ready
* for reading and no error set. */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r |= SEL_RD;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
r &= ~SEL_ERR;
}
}
if ((*ops & (SEL_WR|SEL_ERR))) {
err = pipe_check(f->filp_vno, WRITING, 0, 1, f->filp_pos, 1);
if (err != SUSPEND)
r |= SEL_WR;
if (err < 0 && err != SUSPEND)
r |= SEL_ERR;
if (err == SUSPEND && !(f->filp_mode & W_BIT)) {
/* A "meaningless" write select, therefore ready
for writing and no error set. */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r |= SEL_WR;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
r &= ~SEL_ERR;
}
}
/* Some options we collected might not be requested. */
*ops = r & orig_ops;
if (!*ops && block)
f->filp_pipe_select_ops |= orig_ops;
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
return(OK);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Als select.c fixes.
2005-09-22 23:17:22 +02:00
/*===========================================================================*
* tab2ops *
*===========================================================================*/
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
PRIVATE int tab2ops(int fd, struct selectentry *e)
{
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
int ops = 0;
if (FD_ISSET(fd, &e->readfds)) ops |= SEL_RD;
if (FD_ISSET(fd, &e->writefds)) ops |= SEL_WR;
if (FD_ISSET(fd, &e->errorfds)) ops |= SEL_ERR;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
return(ops);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Als select.c fixes.
2005-09-22 23:17:22 +02:00
/*===========================================================================*
* ops2tab *
*===========================================================================*/
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
PRIVATE void ops2tab(int ops, int fd, struct selectentry *e)
{
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
if ((ops & SEL_RD) && e->vir_readfds && FD_ISSET(fd, &e->readfds) &&
!FD_ISSET(fd, &e->ready_readfds)) {
FD_SET(fd, &e->ready_readfds);
e->nreadyfds++;
}
if ((ops & SEL_WR) && e->vir_writefds && FD_ISSET(fd, &e->writefds) &&
!FD_ISSET(fd, &e->ready_writefds)) {
FD_SET(fd, &e->ready_writefds);
e->nreadyfds++;
}
if ((ops & SEL_ERR) && e->vir_errorfds && FD_ISSET(fd, &e->errorfds) &&
!FD_ISSET(fd, &e->ready_errorfds)) {
FD_SET(fd, &e->ready_errorfds);
e->nreadyfds++;
}
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Als select.c fixes.
2005-09-22 23:17:22 +02:00
/*===========================================================================*
* copy_fdsets *
*===========================================================================*/
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
PRIVATE int copy_fdsets(struct selectentry *se, int nfds, int direction)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int r;
size_t fd_setsize;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
endpoint_t src_e, dst_e;
fd_set *src_fds, *dst_fds;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (nfds < 0 || nfds > OPEN_MAX)
panic() cleanup. this change - makes panic() variadic, doing full printf() formatting - no more NO_NUM, and no more separate printf() statements needed to print extra info (or something in hex) before panicing - unifies panic() - same panic() name and usage for everyone - vm, kernel and rest have different names/syntax currently in order to implement their own luxuries, but no longer - throws out the 1st argument, to make source less noisy. the panic() in syslib retrieves the server name from the kernel so it should be clear enough who is panicing; e.g. panic("sigaction failed: %d", errno); looks like: at_wini(73130): panic: sigaction failed: 0 syslib:panic.c: stacktrace: 0x74dc 0x2025 0x100a - throws out report() - printf() is more convenient and powerful - harmonizes/fixes the use of panic() - there were a few places that used printf-style formatting (didn't work) and newlines (messes up the formatting) in panic() - throws out a few per-server panic() functions - cleans up a tie-in of tty with panic() merging printf() and panic() statements to be done incrementally.
2010-03-05 16:05:11 +01:00
panic("select copy_fdsets: nfds wrong: %d", nfds);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/* Only copy back as many bits as the user expects. */
Build NetBSD libc library in world in ELF mode. 3 sets of libraries are built now: . ack: all libraries that ack can compile (/usr/lib/i386/) . clang+elf: all libraries with minix headers (/usr/lib/) . clang+elf: all libraries with netbsd headers (/usr/netbsd/) Once everything can be compiled with netbsd libraries and headers, the /usr/netbsd hierarchy will be obsolete and its libraries compiled with netbsd headers will be installed in /usr/lib, and its headers in /usr/include. (i.e. minix libc and current minix headers set will be gone.) To use the NetBSD libc system (libraries + headers) before it is the default libc, see: http://wiki.minix3.org/en/DevelopersGuide/UsingNetBSDCode This wiki page also documents the maintenance of the patch files of minix-specific changes to imported NetBSD code. Changes in this commit: . libsys: Add NBSD compilation and create a safe NBSD-based libc. . Port rest of libraries (except libddekit) to new header system. . Enable compilation of libddekit with new headers. . Enable kernel compilation with new headers. . Enable drivers compilation with new headers. . Port legacy commands to new headers and libc. . Port servers to new headers. . Add <sys/sigcontext.h> in compat library. . Remove dependency file in tree. . Enable compilation of common/lib/libc/atomic in libsys . Do not generate RCSID strings in libc. . Temporarily disable zoneinfo as they are incompatible with NetBSD format . obj-nbsd for .gitignore . Procfs: use only integer arithmetic. (Antoine Leca) . Increase ramdisk size to create NBSD-based images. . Remove INCSYMLINKS handling hack. . Add nbsd_include/sys/exec_elf.h . Enable ELF compilation with NBSD libc. . Add 'make nbsdsrc' in tools to download reference NetBSD sources. . Automate minix-port.patch creation. . Avoid using fstavfs() as it is *extremely* slow and unneeded. . Set err() as PRIVATE to avoid name clash with libc. . [NBSD] servers/vm: remove compilation warnings. . u32 is not a long in NBSD headers. . UPDATING info on netbsd hierarchy . commands fixes for netbsd libc
2011-04-27 15:00:52 +02:00
#ifdef __NBSD_LIBC
fd_setsize = (size_t) (howmany(nfds, __NFDBITS) * sizeof(__fd_mask));
#else
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
fd_setsize = (size_t) (_FDSETWORDS(nfds) * _FDSETBITSPERWORD/8);
Build NetBSD libc library in world in ELF mode. 3 sets of libraries are built now: . ack: all libraries that ack can compile (/usr/lib/i386/) . clang+elf: all libraries with minix headers (/usr/lib/) . clang+elf: all libraries with netbsd headers (/usr/netbsd/) Once everything can be compiled with netbsd libraries and headers, the /usr/netbsd hierarchy will be obsolete and its libraries compiled with netbsd headers will be installed in /usr/lib, and its headers in /usr/include. (i.e. minix libc and current minix headers set will be gone.) To use the NetBSD libc system (libraries + headers) before it is the default libc, see: http://wiki.minix3.org/en/DevelopersGuide/UsingNetBSDCode This wiki page also documents the maintenance of the patch files of minix-specific changes to imported NetBSD code. Changes in this commit: . libsys: Add NBSD compilation and create a safe NBSD-based libc. . Port rest of libraries (except libddekit) to new header system. . Enable compilation of libddekit with new headers. . Enable kernel compilation with new headers. . Enable drivers compilation with new headers. . Port legacy commands to new headers and libc. . Port servers to new headers. . Add <sys/sigcontext.h> in compat library. . Remove dependency file in tree. . Enable compilation of common/lib/libc/atomic in libsys . Do not generate RCSID strings in libc. . Temporarily disable zoneinfo as they are incompatible with NetBSD format . obj-nbsd for .gitignore . Procfs: use only integer arithmetic. (Antoine Leca) . Increase ramdisk size to create NBSD-based images. . Remove INCSYMLINKS handling hack. . Add nbsd_include/sys/exec_elf.h . Enable ELF compilation with NBSD libc. . Add 'make nbsdsrc' in tools to download reference NetBSD sources. . Automate minix-port.patch creation. . Avoid using fstavfs() as it is *extremely* slow and unneeded. . Set err() as PRIVATE to avoid name clash with libc. . [NBSD] servers/vm: remove compilation warnings. . u32 is not a long in NBSD headers. . UPDATING info on netbsd hierarchy . commands fixes for netbsd libc
2011-04-27 15:00:52 +02:00
#endif
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/* Set source and destination endpoints */
src_e = (direction == FROM_PROC) ? se->req_endpt : SELF;
dst_e = (direction == FROM_PROC) ? SELF : se->req_endpt;
/* read set */
src_fds = (direction == FROM_PROC) ? se->vir_readfds : &se->ready_readfds;
dst_fds = (direction == FROM_PROC) ? &se->readfds : se->vir_readfds;
if (se->vir_readfds) {
r = sys_vircopy(src_e, D, (vir_bytes) src_fds, dst_e, D,
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
(vir_bytes) dst_fds, fd_setsize);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
if (r != OK) return(r);
}
/* write set */
src_fds = (direction == FROM_PROC) ? se->vir_writefds : &se->ready_writefds;
dst_fds = (direction == FROM_PROC) ? &se->writefds : se->vir_writefds;
if (se->vir_writefds) {
r = sys_vircopy(src_e, D, (vir_bytes) src_fds, dst_e, D,
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
(vir_bytes) dst_fds, fd_setsize);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
if (r != OK) return(r);
}
/* error set */
src_fds = (direction == FROM_PROC) ? se->vir_errorfds : &se->ready_errorfds;
dst_fds = (direction == FROM_PROC) ? &se->errorfds : se->vir_errorfds;
if (se->vir_errorfds) {
r = sys_vircopy(src_e, D, (vir_bytes) src_fds, dst_e, D,
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
(vir_bytes) dst_fds, fd_setsize);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
if (r != OK) return(r);
}
return(OK);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
small compile fixes; initial select() stubs; release.sh script doesn't need /tmp any more since 16MB root device; increase to 3.0.5 to make new CD with working FXP driver. (not tagged 3.0.5 yet as at driver bios-copy workaround hasn't been done.)
2005-06-06 13:40:32 +02:00
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Als select.c fixes.
2005-09-22 23:17:22 +02:00
/*===========================================================================*
* select_cancel_all *
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE void select_cancel_all(struct selectentry *se)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Cancel select. Decrease select usage and cancel timer */
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int fd;
struct filp *f;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
for (fd = 0; fd < se->nfds; fd++) {
if ((f = se->filps[fd]) == NULL) continue;
se->filps[fd] = NULL;
select_cancel_filp(f);
}
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (se->expiry > 0) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
cancel_timer(&se->timer);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
se->expiry = 0;
}
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
se->requestor = NULL;
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* select_cancel_filp *
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE void select_cancel_filp(struct filp *f)
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Reduce number of select users of this filp */
assert(f);
assert(f->filp_selectors >= 0);
if (f->filp_selectors == 0) return;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (f->filp_count == 0) return;
select_lock_filp(f, f->filp_select_ops);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
f->filp_selectors--;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (f->filp_selectors == 0) {
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* No one selecting on this filp anymore, forget about select state */
f->filp_select_ops = 0;
f->filp_select_flags = 0;
f->filp_pipe_select_ops = 0;
}
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
unlock_filp(f);
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
}
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* select_return *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PRIVATE void select_return(struct selectentry *se)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int r, r1;
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
assert(!is_deferred(se)); /* Not done yet, first wait for async reply */
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
select_cancel_all(se);
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r1 = copy_fdsets(se, se->nfds, TO_PROC);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (r1 != OK)
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r = r1;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
else if (se->error != OK)
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r = se->error;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
else
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r = se->nreadyfds;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
revive(se->req_endpt, r);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* select_callback *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PUBLIC void select_callback(struct filp *f, int status)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
filp_status(f, status);
small compile fixes; initial select() stubs; release.sh script doesn't need /tmp any more since 16MB root device; increase to 3.0.5 to make new CD with working FXP driver. (not tagged 3.0.5 yet as at driver bios-copy workaround hasn't been done.)
2005-06-06 13:40:32 +02:00
}
FS understanding revives and selects as alerts (new notifies)
2005-07-27 15:08:52 +02:00
. added cd-probe logic; tries to find the right root filesystem if the boot monitor parameter is set. it also sets a boot monitor parameter (by talking to the PM) so userland knows it. . ramdisk copy code made slightly smarter . various select() prettifications (work in progress)
2005-07-22 20:28:32 +02:00
/*===========================================================================*
* init_select *
*===========================================================================*/
PUBLIC void init_select(void)
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int s;
. added cd-probe logic; tries to find the right root filesystem if the boot monitor parameter is set. it also sets a boot monitor parameter (by talking to the PM) so userland knows it. . ramdisk copy code made slightly smarter . various select() prettifications (work in progress)
2005-07-22 20:28:32 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
for (s = 0; s < MAXSELECTS; s++)
init_timer(&selecttab[s].timer);
. added cd-probe logic; tries to find the right root filesystem if the boot monitor parameter is set. it also sets a boot monitor parameter (by talking to the PM) so userland knows it. . ramdisk copy code made slightly smarter . various select() prettifications (work in progress)
2005-07-22 20:28:32 +02:00
}
small compile fixes; initial select() stubs; release.sh script doesn't need /tmp any more since 16MB root device; increase to 3.0.5 to make new CD with working FXP driver. (not tagged 3.0.5 yet as at driver bios-copy workaround hasn't been done.)
2005-06-06 13:40:32 +02:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Andy's formatting changes.
2005-09-11 18:45:46 +02:00
* select_forget *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PUBLIC void select_forget(endpoint_t proc_e)
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
{
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
/* Something has happened (e.g. signal delivered that interrupts select()).
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* Totally forget about the select(). */
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int slot;
struct selectentry *se;
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
for (slot = 0; slot < MAXSELECTS; slot++) {
se = &selecttab[slot];
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (se->requestor != NULL && se->req_endpt == proc_e)
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
break;
}
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (slot >= MAXSELECTS) return; /* Entry not found */
se->error = EINTR;
if (is_deferred(se)) return; /* Still awaiting initial reply */
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
select_cancel_all(se);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
/*===========================================================================*
Andy's formatting changes.
2005-09-11 18:45:46 +02:00
* select_timeout_check *
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
*===========================================================================*/
PUBLIC void select_timeout_check(timer_t *timer)
{
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
int s;
struct selectentry *se;
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
s = tmr_arg(timer)->ta_int;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (s < 0 || s >= MAXSELECTS) return; /* Entry does not exist */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
se = &selecttab[s];
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (se->requestor == NULL) return;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
fp = se->requestor;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (se->expiry <= 0) return; /* Strange, did we even ask for a timeout? */
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
se->expiry = 0;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (is_deferred(se)) return; /* Wait for initial replies to DEV_SELECT */
select_return(se);
Started select() implementation. Added interface to select() for pipes (also named pipes), and select() stubs for regular files. Added timer library in FS that select() is the first customer of. This is unfinished, but committed anyway to get a new release out to Al and testers.
2005-06-17 15:41:12 +02:00
}
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
/*===========================================================================*
endpoint-aware conversion of servers. 'who', indicating caller number in pm and fs and some other servers, has been removed in favour of 'who_e' (endpoint) and 'who_p' (proc nr.). In both PM and FS, isokendpt() convert endpoints to process slot numbers, returning OK if it was a valid and consistent endpoint number. okendpt() does the same but panic()s if it doesn't succeed. (In PM, this is pm_isok..) pm and fs keep their own records of process endpoints in their proc tables, which are needed to make kernel calls about those processes. message field names have changed. fs drivers are endpoints. fs now doesn't try to get out of driver deadlock, as the protocol isn't supposed to let that happen any more. (A warning is printed if ELOCKED is detected though.) fproc[].fp_task (indicating which driver the process is suspended on) became an int. PM and FS now get endpoint numbers of initial boot processes from the kernel. These happen to be the same as the old proc numbers, to let user processes reach them with the old numbers, but FS and PM don't know that. All new processes after INIT, even after the generation number wraps around, get endpoint numbers with generation 1 and higher, so the first instances of the boot processes are the only processes ever to have endpoint numbers in the old proc number range. More return code checks of sys_* functions have been added. IS has become endpoint-aware. Ditched the 'text' and 'data' fields in the kernel dump (which show locations, not sizes, so aren't terribly useful) in favour of the endpoint number. Proc number is still visible. Some other dumps (e.g. dmap, rs) show endpoint numbers now too which got the formatting changed. PM reading segments using rw_seg() has changed - it uses other fields in the message now instead of encoding the segment and process number and fd in the fd field. For that it uses _read_pm() and _write_pm() which to _taskcall()s directly in pm/misc.c. PM now sys_exit()s itself on panic(), instead of sys_abort(). RS also talks in endpoints instead of process numbers.
2006-03-03 11:20:58 +01:00
* select_unsuspend_by_endpt *
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
*===========================================================================*/
Removed dependency of vfs on NR_TASKS macro - all macros in consts.h that depend on NR_TASKS replaced by a FP_BLOCKED_ON_* - fp_suspended removed and replaced by fp_blocked_on. Testing whether a process is supended is qeual to testing whether fp_blocked_on is FP_BLOCKED_ON_NONE or not - fp_task is valid only if fp_blocked_on == FP_BLOCKED_ON_OTHER - no need of special values that do not colide with valid and special endpoints since they are not used as endpoints anymore - suspend only takes FP_BLOCKED_ON_* values not endpoints anymore - suspend(task) replaced by wait_for(task) which sets fp_task so we remember who are we waiting for and suspend sets fp_blocked_on to FP_BLOCKED_ON_OTHER to signal that we are waiting for some other process - some functions should take endpoint_t instead of int, fixed
2009-09-22 23:48:26 +02:00
PUBLIC void select_unsuspend_by_endpt(endpoint_t proc_e)
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Revive blocked processes when a driver has disappeared */
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int fd, s, major;
struct selectentry *se;
struct filp *f;
for (s = 0; s < MAXSELECTS; s++) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
int wakehim = 0;
se = &selecttab[s];
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (se->requestor == NULL) continue;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (se->requestor->fp_endpoint == proc_e) {
assert(se->requestor->fp_flags & FP_EXITING);
select_cancel_all(se);
continue;
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
for (fd = 0; fd < se->nfds; fd++) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if ((f = se->filps[fd]) == NULL || f->filp_vno == NULL)
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
continue;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
major = major(f->filp_vno->v_sdev);
if (dmap_driver_match(proc_e, major)) {
se->filps[fd] = NULL;
se->error = EINTR;
select_cancel_filp(f);
wakehim = 1;
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (wakehim && !is_deferred(se))
select_return(se);
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
}
Two 'dynamic driver' features in FS: . When drivers disappear that have pending select()s, wake up those user processes with EAGAIN so that they can retry their select() and won't hang forever on it. . When drivers re-appear and are mapped into the dmap, run through the list of mounted filesystems and re-dev_open() every one (for partition tables and such). This can't happen before the driver has exec()ced itself, so processes that have fork()ed but not exec()ced yet are marked as DMAP_BABY in the dmap table if they are dmapped before they are execced. If that happens, the above procedure happens after the exec(). If the exec() happens before the dmapping, it (the dev_open()ing) happens right away.
2005-10-20 21:39:32 +02:00
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
/*===========================================================================*
* select_reply1 *
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PUBLIC void select_reply1(driver_e, minor, status)
endpoint_t driver_e;
int minor;
int status;
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Handle reply to DEV_SELECT request */
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int major;
dev_t dev;
struct filp *f;
struct dmap *dp;
struct vnode *vp;
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Figure out which device is replying */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if ((dp = get_dmap(driver_e)) == NULL) return;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
major = dp-dmap;
dev = makedev(major, minor);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Get filp belonging to character special file */
if ((f = dp->dmap_sel_filp) == NULL) {
printf("VFS (%s:%d): major %d was not expecting a DEV_SELECT reply\n",
__FILE__, __LINE__, major);
return;
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Don't panic if owner has vanished before reply
2011-07-15 16:11:34 +02:00
/* Is the filp still in use and busy waiting for a reply? The owner might
* have vanished before the driver was able to reply. */
if (f->filp_count >= 1 && (f->filp_select_flags & FSF_BUSY)) {
/* Find vnode and check we got a reply from the device we expected */
vp = f->filp_vno;
assert(vp != NULL);
assert((vp->v_mode & I_TYPE) == I_CHAR_SPECIAL); /* Must be char. special */
if (vp->v_sdev != dev) {
printf("VFS (%s:%d): expected reply from dev %d not %d\n",
Cleanup servers to make GCC/Clang a little happier
2011-09-08 15:57:03 +02:00
__FILE__, __LINE__, vp->v_sdev, dev);
Don't panic if owner has vanished before reply
2011-07-15 16:11:34 +02:00
return;
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* No longer waiting for a reply from this device */
dp->dmap_sel_filp = NULL;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
/* Process select result only if requestor is still around. That is, the
* corresponding filp is still in use.
*/
if (f->filp_count >= 1) {
select_lock_filp(f, f->filp_select_ops);
f->filp_select_flags &= ~FSF_BUSY;
/* The select call is done now, except when
* - another process started a select on the same filp with possibly a
* different set of operations.
* - a process does a select on the same filp but using different file
* descriptors.
* - the select has a timeout. Upon receiving this reply the operations
* might not be ready yet, so we want to wait for that to ultimately
* happen.
* Therefore we need to keep remembering what the operations are.
*/
if (!(f->filp_select_flags & (FSF_UPDATE|FSF_BLOCKED)))
f->filp_select_ops = 0; /* done selecting */
else if (!(f->filp_select_flags & FSF_UPDATE))
/* there may be operations pending */
f->filp_select_ops &= ~status;
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
/* Record new filp status */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (!(status == 0 && (f->filp_select_flags & FSF_BLOCKED))) {
if (status > 0) { /* operations ready */
if (status & SEL_RD)
f->filp_select_flags &= ~FSF_RD_BLOCK;
if (status & SEL_WR)
f->filp_select_flags &= ~FSF_WR_BLOCK;
if (status & SEL_ERR)
f->filp_select_flags &= ~FSF_ERR_BLOCK;
} else if (status < 0) { /* error */
/* Always unblock upon error */
f->filp_select_flags &= ~FSF_BLOCKED;
}
}
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
unlock_filp(f);
filp_status(f, status); /* Tell filp owners about the results */
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
select_restart_filps();
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
/*===========================================================================*
* select_reply2 *
*===========================================================================*/
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
PUBLIC void select_reply2(driver_e, minor, status)
endpoint_t driver_e;
int minor;
int status;
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Handle secondary reply to DEV_SELECT request. A secondary reply occurs when
* the select request is 'blocking' until an operation becomes ready. */
int major, slot, fd;
dev_t dev;
struct filp *f;
struct dmap *dp;
struct vnode *vp;
struct selectentry *se;
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (status == 0) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
printf("VFS (%s:%d): weird status (%d) to report\n",
__FILE__, __LINE__, status);
return;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
/* Figure out which device is replying */
if ((dp = get_dmap(driver_e)) == NULL) {
printf("VFS (%s:%d): endpoint %d is not a known driver endpoint\n",
__FILE__, __LINE__, driver_e);
return;
}
major = dp-dmap;
dev = makedev(major, minor);
/* Find all file descriptors selecting for this device */
for (slot = 0; slot < MAXSELECTS; slot++) {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
se = &selecttab[slot];
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (se->requestor == NULL) continue; /* empty slot */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
for (fd = 0; fd < se->nfds; fd++) {
if ((f = se->filps[fd]) == NULL) continue;
if ((vp = f->filp_vno) == NULL) continue;
if ((vp->v_mode & I_TYPE) != I_CHAR_SPECIAL) continue;
if (vp->v_sdev != dev) continue;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
select_lock_filp(f, f->filp_select_ops);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (status > 0) { /* Operations ready */
/* Clear the replied bits from the request
* mask unless FSF_UPDATE is set.
*/
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (!(f->filp_select_flags & FSF_UPDATE))
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
f->filp_select_ops &= ~status;
if (status & SEL_RD)
f->filp_select_flags &= ~FSF_RD_BLOCK;
if (status & SEL_WR)
f->filp_select_flags &= ~FSF_WR_BLOCK;
if (status & SEL_ERR)
f->filp_select_flags &= ~FSF_ERR_BLOCK;
ops2tab(status, fd, se);
} else {
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
f->filp_select_flags &= ~FSF_BLOCKED;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
ops2tab(SEL_RD|SEL_WR|SEL_ERR, fd, se);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
unlock_filp(f);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (se->nreadyfds > 0) restart_proc(se);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
select_restart_filps();
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/*===========================================================================*
* select_restart_filps *
*===========================================================================*/
PRIVATE void select_restart_filps()
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
int fd, slot;
struct filp *f;
struct vnode *vp;
struct selectentry *se;
/* Locate filps that can be restarted */
for (slot = 0; slot < MAXSELECTS; slot++) {
se = &selecttab[slot];
if (se->requestor == NULL) continue; /* empty slot */
/* Only 'deferred' processes are eligible to restart */
if (!is_deferred(se)) continue;
/* Find filps that are not waiting for a reply, but have an updated
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
* status (i.e., another select on the same filp with possibly a
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
* different set of operations is to be done), and thus requires the
* select request to be sent again).
*/
for (fd = 0; fd < se->nfds; fd++) {
fix clang warnings -R in kernel/ and servers/
2011-06-09 16:09:13 +02:00
int r, wantops, ops;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if ((f = se->filps[fd]) == NULL) continue;
if (f->filp_select_flags & FSF_BUSY) /* Still waiting for */
continue; /* initial reply */
if (!(f->filp_select_flags & FSF_UPDATE)) /* Must be in */
continue; /* 'update' state */
wantops = ops = f->filp_select_ops;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
vp = f->filp_vno;
assert((vp->v_mode & I_TYPE) == I_CHAR_SPECIAL);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
r = do_select_request(se, fd, &wantops);
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
if (r != OK) {
if (r == SUSPEND) continue;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
else break; /* Error or bogus return code; abort */
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
if (wantops & ops) ops2tab(wantops, fd, se);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/*===========================================================================*
* do_select_request *
*===========================================================================*/
PRIVATE int do_select_request(se, fd, ops)
struct selectentry *se;
int fd;
int *ops;
{
/* Perform actual select request for file descriptor fd */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
int r, type;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
struct filp *f;
type = se->type[fd];
f = se->filps[fd];
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
select_lock_filp(f, *ops);
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
r = fdtypes[type].select_request(f, ops, se->block);
VFS: various select fixes - Fix locking bug when unable to send DEV_SELECT request. Upon failure VFS tried to cancel the select operation, but this failed due to trying to lock a filp that was already locked to send the request in the first place. Do_select_request now handles locking of filps itself instead of relying on the caller to do it. This fixes a crash when killing INET. - Fix failure to revive a process after a non-blocking select operation yielded no ready select operations when replying DEV_SEL_REPL1. - Improve readability by using OK, SUSPEND, and standard error values as results instead of having separate macros in select. - Don't print not having a driver for a major device; after killing a driver select will trigger this printf.
2012-02-17 14:36:57 +01:00
unlock_filp(f);
if (r != OK && r != SUSPEND) {
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
se->error = EINTR;
se->block = 0; /* Stop blocking to return asap */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if (!is_deferred(se)) select_cancel_all(se);
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
return(r);
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/*===========================================================================*
* filp_status *
*===========================================================================*/
PRIVATE void filp_status(f, status)
struct filp *f;
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
int status;
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Tell processes that need to know about the status of this filp */
int fd, slot;
struct selectentry *se;
for (slot = 0; slot < MAXSELECTS; slot++) {
se = &selecttab[slot];
if (se->requestor == NULL) continue; /* empty slot */
for (fd = 0; fd < se->nfds; fd++) {
if (se->filps[fd] != f) continue;
if (status < 0)
ops2tab(SEL_RD|SEL_WR|SEL_ERR, fd, se);
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
else
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
ops2tab(status, fd, se);
restart_proc(se);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
}
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/*===========================================================================*
* restart_proc *
*===========================================================================*/
PRIVATE void restart_proc(se)
struct selectentry *se;
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
{
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
/* Tell process about select results (if any) unless there are still results
* pending. */
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
if ((se->nreadyfds > 0 || !se->block) && !is_deferred(se))
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
select_return(se);
Asynchronous select implementation.
2008-02-22 16:46:59 +01:00
}
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
/*===========================================================================*
* wipe_select *
*===========================================================================*/
PRIVATE void wipe_select(struct selectentry *se)
{
se->nfds = 0;
se->nreadyfds = 0;
Clean up and fix multiple bugs in select: - Remove redundant code. - Always wait for the initial reply from an asynchronous select request, even if the select has been satisfied on another file descriptor or was canceled due to a serious error. - Restart asynchronous selects if upon reply from the driver turns out that there are deferred operations (and do not forget we're still interested in the results of the deferred operations). - Do not hang a non-blocking select when another blocking select on the same filp is still blocking. - Split blocking operations in read, write, and exceptions (i.e., blocking on read does not imply the write will block as well). - Some loops would iterate over OPEN_MAX file descriptors instead of the "highest" file descriptor. - Use proper internal error return values. - A secondary reply from a synchronous driver is essentially the same as from an asynchronous driver (the only difference being how the answer is received). Merge. - Return proper error code after a driver failure. - Auto-detect whether a driver is synchronous or asynchronous. - Remove some code duplication. - Clean up code (coding style, add missing comments, put all select related code together).
2011-04-13 15:25:34 +02:00
se->error = OK;
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
se->block = 0;
memset(se->filps, 0, sizeof(se->filps));
- Introduce support for sticky bit. - Revise VFS-FS protocol and update VFS/MFS/ISOFS accordingly. - Clean up MFS by removing old, dead code (backwards compatibility is broken by the new VFS-FS protocol, anyway) and rewrite other parts. Also, make sure all functions have proper banners and prototypes. - VFS should always provide a (syntactically) valid path to the FS; no need for the FS to do sanity checks when leaving/entering mount points. - Fix several bugs in MFS: - Several path lookup bugs in MFS. - A link can be too big for the path buffer. - A mountpoint can become inaccessible when the creation of a new inode fails, because the inode already exists and is a mountpoint. - Introduce support for supplemental groups. - Add test 46 to test supplemental group functionality (and removed obsolete suppl. tests from test 2). - Clean up VFS (not everything is done yet). - ISOFS now opens device read-only. This makes the -r flag in the mount command unnecessary (but will still report to be mounted read-write). - Introduce PipeFS. PipeFS is a new FS that handles all anonymous and named pipes. However, named pipes still reside on the (M)FS, as they are part of the file system on disk. To make this work VFS now has a concept of 'mapped' inodes, which causes read, write, truncate and stat requests to be redirected to the mapped FS, and all other requests to the original FS.
2009-12-20 21:27:14 +01:00
FD_ZERO(&se->readfds);
FD_ZERO(&se->writefds);
FD_ZERO(&se->errorfds);
FD_ZERO(&se->ready_readfds);
FD_ZERO(&se->ready_writefds);
FD_ZERO(&se->ready_errorfds);
}
VFS: replace VFS with AVFS
2012-02-13 16:28:04 +01:00
/*===========================================================================*
* select_lock_filp *
*===========================================================================*/
PRIVATE void select_lock_filp(struct filp *f, int ops)
{
/* Lock a filp and vnode based on which operations are requested */
tll_access_t locktype;;
locktype = VNODE_READ; /* By default */
if (ops & (SEL_WR|SEL_ERR))
/* Selecting for error or writing requires exclusive access */
locktype = VNODE_WRITE;
lock_filp(f, locktype);
}
Reference in a new issue Copy permalink