minix/servers/mfs/main.c
David van Moolenbroek b4d909d415 Split block/character protocols and libdriver
This patch separates the character and block driver communication
protocols. The old character protocol remains the same, but a new
block protocol is introduced. The libdriver library is replaced by
two new libraries: libchardriver and libblockdriver. Their exposed
API, and drivers that use them, have been updated accordingly.
Together, libbdev and libblockdriver now completely abstract away
the message format used by the block protocol. As the memory driver
is both a character and a block device driver, it now implements its
own message loop.

The most important semantic change made to the block protocol is that
it is no longer possible to return both partial results and an error
for a single transfer. This simplifies the interaction between the
caller and the driver, as the I/O vector no longer needs to be copied
back. Also, drivers are now no longer supposed to decide based on the
layout of the I/O vector when a transfer should be cut short. Put
simply, transfers are now supposed to either succeed completely, or
result in an error.

After this patch, the state of the various pieces is as follows:
- block protocol: stable
- libbdev API: stable for synchronous communication
- libblockdriver API: needs slight revision (the drvlib/partition API
  in particular; the threading API will also change shortly)
- character protocol: needs cleanup
- libchardriver API: needs cleanup accordingly
- driver restarts: largely unsupported until endpoint changes are
  reintroduced

As a side effect, this patch eliminates several bugs, hacks, and gcc
-Wall and -W warnings all over the place. It probably introduces a
few new ones, too.

Update warning: this patch changes the protocol between MFS and disk
drivers, so in order to use old/new images, the MFS from the ramdisk
must be used to mount all file systems.
2011-11-23 14:06:37 +01:00

207 lines
5.6 KiB
C

#include "fs.h"
#include <assert.h>
#include <minix/callnr.h>
#include <signal.h>
#include <stdlib.h>
#include <unistd.h>
#include <minix/dmap.h>
#include <minix/endpoint.h>
#include <minix/vfsif.h>
#include "buf.h"
#include "inode.h"
/* Declare some local functions. */
FORWARD _PROTOTYPE(void get_work, (message *m_in) );
FORWARD _PROTOTYPE( void reply, (endpoint_t who, message *m_out) );
/* SEF functions and variables. */
FORWARD _PROTOTYPE( void sef_local_startup, (void) );
FORWARD _PROTOTYPE( int sef_cb_init_fresh, (int type, sef_init_info_t *info) );
FORWARD _PROTOTYPE( void sef_cb_signal_handler, (int signo) );
/*===========================================================================*
* main *
*===========================================================================*/
PUBLIC int main(int argc, char *argv[])
{
/* This is the main routine of this service. The main loop consists of
* three major activities: getting new work, processing the work, and
* sending the reply. The loop never terminates, unless a panic occurs.
*/
int error = OK, ind, transid;
/* SEF local startup. */
env_setargs(argc, argv);
sef_local_startup();
while(!unmountdone || !exitsignaled) {
endpoint_t src;
/* Wait for request message. */
get_work(&fs_m_in);
transid = TRNS_GET_ID(fs_m_in.m_type);
fs_m_in.m_type = TRNS_DEL_ID(fs_m_in.m_type);
if (fs_m_in.m_type == 0) {
assert(!IS_VFS_FS_TRANSID(transid));
fs_m_in.m_type = transid; /* Backwards compat. */
transid = 0;
} else
assert(IS_VFS_FS_TRANSID(transid));
src = fs_m_in.m_source;
caller_uid = INVAL_UID; /* To trap errors */
caller_gid = INVAL_GID;
req_nr = fs_m_in.m_type;
if (req_nr < VFS_BASE) {
fs_m_in.m_type += VFS_BASE;
req_nr = fs_m_in.m_type;
}
ind = req_nr - VFS_BASE;
if (ind < 0 || ind >= NREQS) {
printf("MFS: bad request %d from %d\n", req_nr, src);
printf("ind = %d\n", ind);
error = EINVAL;
} else {
error = (*fs_call_vec[ind])();
/*cch_check();*/
}
fs_m_out.m_type = error;
if (IS_VFS_FS_TRANSID(transid)) {
/* If a transaction ID was set, reset it */
fs_m_out.m_type = TRNS_ADD_ID(fs_m_out.m_type, transid);
}
reply(src, &fs_m_out);
}
return(OK);
}
/*===========================================================================*
* sef_local_startup *
*===========================================================================*/
PRIVATE void sef_local_startup()
{
/* Register init callbacks. */
sef_setcb_init_fresh(sef_cb_init_fresh);
sef_setcb_init_restart(sef_cb_init_fail);
/* No live update support for now. */
/* Register signal callbacks. */
sef_setcb_signal_handler(sef_cb_signal_handler);
/* Let SEF perform startup. */
sef_startup();
}
/*===========================================================================*
* sef_cb_init_fresh *
*===========================================================================*/
PRIVATE int sef_cb_init_fresh(int UNUSED(type), sef_init_info_t *UNUSED(info))
{
/* Initialize the Minix file server. */
int i;
may_use_vmcache = 1;
/* Init inode table */
for (i = 0; i < NR_INODES; ++i) {
inode[i].i_count = 0;
cch[i] = 0;
}
init_inode_cache();
SELF_E = getprocnr();
buf_pool(DEFAULT_NR_BUFS);
fs_block_size = _MIN_BLOCK_SIZE;
return(OK);
}
/*===========================================================================*
* sef_cb_signal_handler *
*===========================================================================*/
PRIVATE void sef_cb_signal_handler(int signo)
{
/* Only check for termination signal, ignore anything else. */
if (signo != SIGTERM) return;
exitsignaled = 1;
(void) fs_sync();
/* If unmounting has already been performed, exit immediately.
* We might not get another message.
*/
if (unmountdone) exit(0);
}
/*===========================================================================*
* get_work *
*===========================================================================*/
PRIVATE void get_work(m_in)
message *m_in; /* pointer to message */
{
int r, srcok = 0;
endpoint_t src;
do {
if ((r = sef_receive(ANY, m_in)) != OK) /* wait for message */
panic("sef_receive failed: %d", r);
src = m_in->m_source;
if(src == VFS_PROC_NR) {
if(unmountdone)
printf("MFS: unmounted: unexpected message from FS\n");
else
srcok = 1; /* Normal FS request. */
} else
printf("MFS: unexpected source %d\n", src);
} while(!srcok);
assert((src == VFS_PROC_NR && !unmountdone));
}
/*===========================================================================*
* reply *
*===========================================================================*/
PRIVATE void reply(
endpoint_t who,
message *m_out /* report result */
)
{
if (OK != send(who, m_out)) /* send the message */
printf("MFS(%d) was unable to send reply\n", SELF_E);
}
#if 0
/*===========================================================================*
* cch_check *
*===========================================================================*/
PRIVATE void cch_check(void)
{
int i;
for (i = 0; i < NR_INODES; ++i) {
if (inode[i].i_count != cch[i] && req_nr != REQ_GETNODE &&
req_nr != REQ_PUTNODE && req_nr != REQ_READSUPER &&
req_nr != REQ_MOUNTPOINT && req_nr != REQ_UNMOUNT &&
req_nr != REQ_SYNC && req_nr != REQ_LOOKUP) {
printf("MFS(%d) inode(%lu) cc: %d req_nr: %d\n", SELF_E,
inode[i].i_num, inode[i].i_count - cch[i], req_nr);
}
cch[i] = inode[i].i_count;
}
}
#endif