minix/servers/vfs/stadir.c
David van Moolenbroek 723e51327f VFS: worker thread model overhaul
The main purpose of this patch is to fix handling of unpause calls
from PM while another call is ongoing. The solution to this problem
sparked a full revision of the threading model, consisting of a large
number of related changes:

- all active worker threads are now always associated with a process,
  and every process has at most one active thread working for it;
- the process lock is always held by a process's worker thread;
- a process can now have both normal work and postponed PM work
  associated to it;
- timer expiry and non-postponed PM work is done from the main thread;
- filp garbage collection is done from a thread associated with VFS;
- reboot calls from PM are now done from a thread associated with PM;
- the DS events handler is protected from starting multiple threads;
- support for a system worker thread has been removed;
- the deadlock recovery thread has been replaced by a parameter to the
  worker_start() function; the number of worker threads has
  consequently been increased by one;
- saving and restoring of global but per-thread variables is now
  centralized in worker_suspend() and worker_resume(); err_code is now
  saved and restored in all cases;
- the concept of jobs has been removed, and job_m_in now points to a
  message stored in the worker thread structure instead;
- the PM lock has been removed;
- the separate exec lock has been replaced by a lock on the VM
  process, which was already being locked for exec calls anyway;
- PM_UNPAUSE is now processed as a postponed PM request, from a thread
  associated with the target process;
- the FP_DROP_WORK flag has been removed, since it is no longer more
  than just an optimization and only applied to processes operating on
  a pipe when getting killed;
- assignment to "fp" now takes place only when obtaining new work in
  the main thread or a worker thread, when resuming execution of a
  thread, and in the special case of exiting processes during reboot;
- there are no longer special cases where the yield() call is used to
  force a thread to run.

Change-Id: I7a97b9b95c2450454a9b5318dfa0e6150d4e6858
2014-02-18 11:25:03 +01:00

462 lines
13 KiB
C

/* This file contains the code for performing four system calls relating to
* status and directories.
*
* The entry points into this file are
* do_chdir: perform the CHDIR system call
* do_chroot: perform the CHROOT system call
* do_lstat: perform the LSTAT system call
* do_stat: perform the STAT system call
* do_fstat: perform the FSTAT system call
* do_statvfs: perform the STATVFS1 system call
* do_fstatvfs: perform the FSTATVFS1 system call
* do_getvfsstat: perform the GETVFSSTAT system call
*/
#include "fs.h"
#include <sys/stat.h>
#include <minix/com.h>
#include <minix/u64.h>
#include <string.h>
#include "file.h"
#include "path.h"
#include "param.h"
#include <minix/vfsif.h>
#include <minix/callnr.h>
#include "vnode.h"
#include "vmnt.h"
static int change_into(struct vnode **iip, struct vnode *vp);
/*===========================================================================*
* do_fchdir *
*===========================================================================*/
int do_fchdir(message *UNUSED(m_out))
{
/* Change directory on already-opened fd. */
struct filp *rfilp;
int r, rfd;
rfd = job_m_in.fd;
/* Is the file descriptor valid? */
if ((rfilp = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
r = change_into(&fp->fp_wd, rfilp->filp_vno);
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* do_chdir *
*===========================================================================*/
int do_chdir(message *UNUSED(m_out))
{
/* Perform the chdir(name) system call.
* syscall might provide 'name' embedded in the message.
*/
int r;
struct vnode *vp;
struct vmnt *vmp;
char fullpath[PATH_MAX];
struct lookup resolve;
vir_bytes vname;
size_t vname_length;
vname = (vir_bytes) job_m_in.name;
vname_length = (size_t) job_m_in.name_length;
if (copy_name(vname_length, fullpath) != OK) {
/* Direct copy failed, try fetching from user space */
if (fetch_name(vname, vname_length, fullpath) != OK)
return(err_code);
}
/* Try to open the directory */
lookup_init(&resolve, fullpath, PATH_NOFLAGS, &vmp, &vp);
resolve.l_vmnt_lock = VMNT_READ;
resolve.l_vnode_lock = VNODE_READ;
if ((vp = eat_path(&resolve, fp)) == NULL) return(err_code);
r = change_into(&fp->fp_wd, vp);
unlock_vnode(vp);
unlock_vmnt(vmp);
put_vnode(vp);
return(r);
}
/*===========================================================================*
* do_chroot *
*===========================================================================*/
int do_chroot(message *UNUSED(m_out))
{
/* Perform the chroot(name) system call.
* syscall might provide 'name' embedded in the message.
*/
int r;
struct vnode *vp;
struct vmnt *vmp;
char fullpath[PATH_MAX];
struct lookup resolve;
vir_bytes vname;
size_t vname_length;
vname = (vir_bytes) job_m_in.name;
vname_length = (size_t) job_m_in.name_length;
if (!super_user) return(EPERM); /* only su may chroot() */
if (copy_name(vname_length, fullpath) != OK) {
/* Direct copy failed, try fetching from user space */
if (fetch_name(vname, vname_length, fullpath) != OK)
return(err_code);
}
/* Try to open the directory */
lookup_init(&resolve, fullpath, PATH_NOFLAGS, &vmp, &vp);
resolve.l_vmnt_lock = VMNT_READ;
resolve.l_vnode_lock = VNODE_READ;
if ((vp = eat_path(&resolve, fp)) == NULL) return(err_code);
r = change_into(&fp->fp_rd, vp);
unlock_vnode(vp);
unlock_vmnt(vmp);
put_vnode(vp);
return(r);
}
/*===========================================================================*
* change_into *
*===========================================================================*/
static int change_into(struct vnode **result, struct vnode *vp)
{
int r;
if (*result == vp) return(OK); /* Nothing to do */
/* It must be a directory and also be searchable */
if (!S_ISDIR(vp->v_mode))
r = ENOTDIR;
else
r = forbidden(fp, vp, X_BIT); /* Check if dir is searchable*/
if (r != OK) return(r);
/* Everything is OK. Make the change. */
put_vnode(*result); /* release the old directory */
dup_vnode(vp);
*result = vp; /* acquire the new one */
return(OK);
}
/*===========================================================================*
* do_stat *
*===========================================================================*/
int do_stat(message *UNUSED(m_out))
{
/* Perform the stat(name, buf) system call. */
int r;
struct vnode *vp;
struct vmnt *vmp;
char fullpath[PATH_MAX];
struct lookup resolve;
vir_bytes vname1, statbuf;
size_t vname1_length;
vname1 = (vir_bytes) job_m_in.name1;
vname1_length = (size_t) job_m_in.name1_length;
statbuf = (vir_bytes) job_m_in.m1_p2;
lookup_init(&resolve, fullpath, PATH_NOFLAGS, &vmp, &vp);
resolve.l_vmnt_lock = VMNT_READ;
resolve.l_vnode_lock = VNODE_READ;
if (fetch_name(vname1, vname1_length, fullpath) != OK) return(err_code);
if ((vp = eat_path(&resolve, fp)) == NULL) return(err_code);
r = req_stat(vp->v_fs_e, vp->v_inode_nr, who_e, statbuf);
unlock_vnode(vp);
unlock_vmnt(vmp);
put_vnode(vp);
return r;
}
/*===========================================================================*
* do_fstat *
*===========================================================================*/
int do_fstat(message *UNUSED(m_out))
{
/* Perform the fstat(fd, buf) system call. */
register struct filp *rfilp;
int r, rfd;
vir_bytes statbuf;
statbuf = (vir_bytes) job_m_in.buffer;
rfd = job_m_in.fd;
/* Is the file descriptor valid? */
if ((rfilp = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
r = req_stat(rfilp->filp_vno->v_fs_e, rfilp->filp_vno->v_inode_nr,
who_e, statbuf);
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* update_statvfs *
*===========================================================================*/
int update_statvfs(struct vmnt *vmp, struct statvfs *buf)
{
/* Get statistics from a file system, and cache part of the results. */
int r;
if ((r = req_statvfs(vmp->m_fs_e, buf)) != OK)
return r;
vmp->m_stats.f_flag = buf->f_flag;
vmp->m_stats.f_bsize = buf->f_bsize;
vmp->m_stats.f_frsize = buf->f_frsize;
vmp->m_stats.f_iosize = buf->f_iosize;
vmp->m_stats.f_blocks = buf->f_blocks;
vmp->m_stats.f_bfree = buf->f_bfree;
vmp->m_stats.f_bavail = buf->f_bavail;
vmp->m_stats.f_bresvd = buf->f_bresvd;
vmp->m_stats.f_files = buf->f_files;
vmp->m_stats.f_ffree = buf->f_ffree;
vmp->m_stats.f_favail = buf->f_favail;
vmp->m_stats.f_fresvd = buf->f_fresvd;
vmp->m_stats.f_syncreads = buf->f_syncreads;
vmp->m_stats.f_syncwrites = buf->f_syncwrites;
vmp->m_stats.f_asyncreads = buf->f_asyncreads;
vmp->m_stats.f_asyncwrites = buf->f_asyncwrites;
vmp->m_stats.f_namemax = buf->f_namemax;
return OK;
}
/*===========================================================================*
* fill_statvfs *
*===========================================================================*/
static int fill_statvfs(struct vmnt *vmp, endpoint_t endpt, vir_bytes buf_addr,
int flags)
{
/* Fill a statvfs structure in a userspace process. First let the target file
* server fill in most fields, or use the cached copy if ST_NOWAIT is given.
* Then fill in some remaining fields with local information. Finally, copy
* the result to user space.
*/
struct statvfs buf;
if (!(flags & ST_NOWAIT)) {
/* Get fresh statistics from the file system. */
if (update_statvfs(vmp, &buf) != OK)
return EIO;
} else {
/* Use the cached statistics. */
memset(&buf, 0, sizeof(buf));
buf.f_flag = vmp->m_stats.f_flag;
buf.f_bsize = vmp->m_stats.f_bsize;
buf.f_frsize = vmp->m_stats.f_frsize;
buf.f_iosize = vmp->m_stats.f_iosize;
buf.f_blocks = vmp->m_stats.f_blocks;
buf.f_bfree = vmp->m_stats.f_bfree;
buf.f_bavail = vmp->m_stats.f_bavail;
buf.f_bresvd = vmp->m_stats.f_bresvd;
buf.f_files = vmp->m_stats.f_files;
buf.f_ffree = vmp->m_stats.f_ffree;
buf.f_favail = vmp->m_stats.f_favail;
buf.f_fresvd = vmp->m_stats.f_fresvd;
buf.f_syncreads = vmp->m_stats.f_syncreads;
buf.f_syncwrites = vmp->m_stats.f_syncwrites;
buf.f_asyncreads = vmp->m_stats.f_asyncreads;
buf.f_asyncwrites = vmp->m_stats.f_asyncwrites;
buf.f_namemax = vmp->m_stats.f_namemax;
}
if (vmp->m_flags & VMNT_READONLY)
buf.f_flag |= ST_RDONLY;
buf.f_fsid = vmp->m_dev;
buf.f_fsidx.__fsid_val[0] = 0;
buf.f_fsidx.__fsid_val[1] = vmp->m_dev;
strlcpy(buf.f_fstypename, vmp->m_fstype, sizeof(buf.f_fstypename));
strlcpy(buf.f_mntonname, vmp->m_mount_path, sizeof(buf.f_mntonname));
strlcpy(buf.f_mntfromname, vmp->m_mount_dev, sizeof(buf.f_mntfromname));
return sys_datacopy(SELF, (vir_bytes) &buf, endpt, buf_addr, sizeof(buf));
}
/*===========================================================================*
* do_statvfs *
*===========================================================================*/
int do_statvfs(message *UNUSED(m_out))
{
/* Perform the statvfs1(name, buf, flags) system call. */
int r, flags;
struct vnode *vp;
struct vmnt *vmp;
char fullpath[PATH_MAX];
struct lookup resolve;
vir_bytes vname1, statbuf;
size_t vname1_length;
vname1 = (vir_bytes) job_m_in.VFS_STATVFS1_NAME;
vname1_length = (size_t) job_m_in.VFS_STATVFS1_LEN;
statbuf = (vir_bytes) job_m_in.VFS_STATVFS1_BUF;
flags = job_m_in.VFS_STATVFS1_FLAGS;
lookup_init(&resolve, fullpath, PATH_NOFLAGS, &vmp, &vp);
resolve.l_vmnt_lock = VMNT_READ;
resolve.l_vnode_lock = VNODE_READ;
if (fetch_name(vname1, vname1_length, fullpath) != OK) return(err_code);
if ((vp = eat_path(&resolve, fp)) == NULL) return(err_code);
r = fill_statvfs(vp->v_vmnt, who_e, statbuf, flags);
unlock_vnode(vp);
unlock_vmnt(vmp);
put_vnode(vp);
return r;
}
/*===========================================================================*
* do_fstatvfs *
*===========================================================================*/
int do_fstatvfs(message *UNUSED(m_out))
{
/* Perform the fstatvfs1(fd, buf, flags) system call. */
register struct filp *rfilp;
int r, rfd, flags;
vir_bytes statbuf;
rfd = job_m_in.VFS_FSTATVFS1_FD;
statbuf = (vir_bytes) job_m_in.VFS_FSTATVFS1_BUF;
flags = job_m_in.VFS_FSTATVFS1_FLAGS;
/* Is the file descriptor valid? */
if ((rfilp = get_filp(rfd, VNODE_READ)) == NULL) return(err_code);
r = fill_statvfs(rfilp->filp_vno->v_vmnt, who_e, statbuf, flags);
unlock_filp(rfilp);
return(r);
}
/*===========================================================================*
* do_getvfsstat *
*===========================================================================*/
int do_getvfsstat(message *UNUSED(m_out))
{
/* Perform the getvfsstat(buf, bufsize, flags) system call. */
struct vmnt *vmp;
vir_bytes buf;
size_t bufsize;
int r, flags, count, do_lock;
buf = (vir_bytes) job_m_in.VFS_GETVFSSTAT_BUF;
bufsize = job_m_in.VFS_GETVFSSTAT_SIZE;
flags = job_m_in.VFS_GETVFSSTAT_FLAGS;
count = 0;
if (buf != 0) {
/* We only need to lock target file systems if we are going to query
* them. This will only happen if ST_NOWAIT is not given. If we do
* not lock, we rely on the VMNT_CANSTAT flag to protect us from
* concurrent (un)mount operations. Note that procfs relies on
* ST_NOWAIT calls being lock free, as it is a file system itself.
*/
do_lock = !(flags & ST_NOWAIT);
for (vmp = &vmnt[0]; vmp < &vmnt[NR_MNTS]; vmp++) {
/* If there is no more space, return the count so far. */
if (bufsize < sizeof(struct statvfs))
break;
/* Lock the file system before checking any fields. */
if (do_lock && (r = lock_vmnt(vmp, VMNT_READ)) != OK)
return r;
/* Obtain information for this file system, if it is in use and
* can be reported. File systems that are being (un)mounted
* are skipped, as is PFS. The fill call will block only if
* ST_NOWAIT was not given.
*/
if (vmp->m_dev != NO_DEV && (vmp->m_flags & VMNT_CANSTAT)) {
if ((r = fill_statvfs(vmp, who_e, buf, flags)) != OK) {
if (do_lock)
unlock_vmnt(vmp);
return r;
}
count++;
buf += sizeof(struct statvfs);
bufsize -= sizeof(struct statvfs);
}
if (do_lock)
unlock_vmnt(vmp);
}
} else {
/* Just report a file system count. No need to lock, as above. */
for (vmp = &vmnt[0]; vmp < &vmnt[NR_MNTS]; vmp++) {
if (vmp->m_dev != NO_DEV && (vmp->m_flags & VMNT_CANSTAT))
count++;
}
}
return count;
}
/*===========================================================================*
* do_lstat *
*===========================================================================*/
int do_lstat(message *UNUSED(m_out))
{
/* Perform the lstat(name, buf) system call. */
struct vnode *vp;
struct vmnt *vmp;
int r;
char fullpath[PATH_MAX];
struct lookup resolve;
vir_bytes vname1, statbuf;
size_t vname1_length;
vname1 = (vir_bytes) job_m_in.name1;
vname1_length = (size_t) job_m_in.name1_length;
statbuf = (vir_bytes) job_m_in.name2;
lookup_init(&resolve, fullpath, PATH_RET_SYMLINK, &vmp, &vp);
resolve.l_vmnt_lock = VMNT_READ;
resolve.l_vnode_lock = VNODE_READ;
if (fetch_name(vname1, vname1_length, fullpath) != OK) return(err_code);
if ((vp = eat_path(&resolve, fp)) == NULL) return(err_code);
r = req_stat(vp->v_fs_e, vp->v_inode_nr, who_e, statbuf);
unlock_vnode(vp);
unlock_vmnt(vmp);
put_vnode(vp);
return(r);
}