kernel-fxtec-pro1x/fs/cifs/file.c

2368 lines
63 KiB
C
Raw Normal View History

/*
* fs/cifs/file.c
*
* vfs operations that deal with files
*
* Copyright (C) International Business Machines Corp., 2002,2010
* Author(s): Steve French (sfrench@us.ibm.com)
* Jeremy Allison (jra@samba.org)
*
* This library is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as published
* by the Free Software Foundation; either version 2.1 of the License, or
* (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
* the GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public License
* along with this library; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
#include <linux/fs.h>
#include <linux/backing-dev.h>
#include <linux/stat.h>
#include <linux/fcntl.h>
#include <linux/pagemap.h>
#include <linux/pagevec.h>
#include <linux/writeback.h>
#include <linux/task_io_accounting_ops.h>
#include <linux/delay.h>
#include <linux/mount.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 02:04:11 -06:00
#include <linux/slab.h>
#include <asm/div64.h>
#include "cifsfs.h"
#include "cifspdu.h"
#include "cifsglob.h"
#include "cifsproto.h"
#include "cifs_unicode.h"
#include "cifs_debug.h"
#include "cifs_fs_sb.h"
#include "fscache.h"
static inline int cifs_convert_flags(unsigned int flags)
{
if ((flags & O_ACCMODE) == O_RDONLY)
return GENERIC_READ;
else if ((flags & O_ACCMODE) == O_WRONLY)
return GENERIC_WRITE;
else if ((flags & O_ACCMODE) == O_RDWR) {
/* GENERIC_ALL is too much permission to request
can cause unnecessary access denied on create */
/* return GENERIC_ALL; */
return (GENERIC_READ | GENERIC_WRITE);
}
return (READ_CONTROL | FILE_WRITE_ATTRIBUTES | FILE_READ_ATTRIBUTES |
FILE_WRITE_EA | FILE_APPEND_DATA | FILE_WRITE_DATA |
FILE_READ_DATA);
}
static u32 cifs_posix_convert_flags(unsigned int flags)
{
u32 posix_flags = 0;
if ((flags & O_ACCMODE) == O_RDONLY)
posix_flags = SMB_O_RDONLY;
else if ((flags & O_ACCMODE) == O_WRONLY)
posix_flags = SMB_O_WRONLY;
else if ((flags & O_ACCMODE) == O_RDWR)
posix_flags = SMB_O_RDWR;
if (flags & O_CREAT)
posix_flags |= SMB_O_CREAT;
if (flags & O_EXCL)
posix_flags |= SMB_O_EXCL;
if (flags & O_TRUNC)
posix_flags |= SMB_O_TRUNC;
/* be safe and imply O_SYNC for O_DSYNC */
vfs: Implement proper O_SYNC semantics While Linux provided an O_SYNC flag basically since day 1, it took until Linux 2.4.0-test12pre2 to actually get it implemented for filesystems, since that day we had generic_osync_around with only minor changes and the great "For now, when the user asks for O_SYNC, we'll actually give O_DSYNC" comment. This patch intends to actually give us real O_SYNC semantics in addition to the O_DSYNC semantics. After Jan's O_SYNC patches which are required before this patch it's actually surprisingly simple, we just need to figure out when to set the datasync flag to vfs_fsync_range and when not. This patch renames the existing O_SYNC flag to O_DSYNC while keeping it's numerical value to keep binary compatibility, and adds a new real O_SYNC flag. To guarantee backwards compatiblity it is defined as expanding to both the O_DSYNC and the new additional binary flag (__O_SYNC) to make sure we are backwards-compatible when compiled against the new headers. This also means that all places that don't care about the differences can just check O_DSYNC and get the right behaviour for O_SYNC, too - only places that actuall care need to check __O_SYNC in addition. Drivers and network filesystems have been updated in a fail safe way to always do the full sync magic if O_DSYNC is set. The few places setting O_SYNC for lower layers are kept that way for now to stay failsafe. We enforce that O_DSYNC is set when __O_SYNC is set early in the open path to make sure we always get these sane options. Note that parisc really screwed up their headers as they already define a O_DSYNC that has always been a no-op. We try to repair it by using it for the new O_DSYNC and redefinining O_SYNC to send both the traditional O_SYNC numerical value _and_ the O_DSYNC one. Cc: Richard Henderson <rth@twiddle.net> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Grant Grundler <grundler@parisc-linux.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Ingo Molnar <mingo@elte.hu> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andreas Dilger <adilger@sun.com> Acked-by: Trond Myklebust <Trond.Myklebust@netapp.com> Acked-by: Kyle McMartin <kyle@mcmartin.ca> Acked-by: Ulrich Drepper <drepper@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jan Kara <jack@suse.cz>
2009-10-27 04:05:28 -06:00
if (flags & O_DSYNC)
posix_flags |= SMB_O_SYNC;
if (flags & O_DIRECTORY)
posix_flags |= SMB_O_DIRECTORY;
if (flags & O_NOFOLLOW)
posix_flags |= SMB_O_NOFOLLOW;
if (flags & O_DIRECT)
posix_flags |= SMB_O_DIRECT;
return posix_flags;
}
static inline int cifs_get_disposition(unsigned int flags)
{
if ((flags & (O_CREAT | O_EXCL)) == (O_CREAT | O_EXCL))
return FILE_CREATE;
else if ((flags & (O_CREAT | O_TRUNC)) == (O_CREAT | O_TRUNC))
return FILE_OVERWRITE_IF;
else if ((flags & O_CREAT) == O_CREAT)
return FILE_OPEN_IF;
else if ((flags & O_TRUNC) == O_TRUNC)
return FILE_OVERWRITE;
else
return FILE_OPEN;
}
static inline int cifs_open_inode_helper(struct inode *inode,
struct cifsTconInfo *pTcon, __u32 oplock, FILE_ALL_INFO *buf,
char *full_path, int xid)
{
struct cifsInodeInfo *pCifsInode = CIFS_I(inode);
struct timespec temp;
int rc;
if (pCifsInode->clientCanCacheRead) {
/* we have the inode open somewhere else
no need to discard cache data */
goto client_can_cache;
}
/* BB need same check in cifs_create too? */
/* if not oplocked, invalidate inode pages if mtime or file
size changed */
temp = cifs_NTtimeToUnix(buf->LastWriteTime);
if (timespec_equal(&inode->i_mtime, &temp) &&
(inode->i_size ==
(loff_t)le64_to_cpu(buf->EndOfFile))) {
cFYI(1, "inode unchanged on server");
} else {
if (inode->i_mapping) {
/* BB no need to lock inode until after invalidate
since namei code should already have it locked? */
rc = filemap_write_and_wait(inode->i_mapping);
[CIFS] Fix potential data corruption when writing out cached dirty pages Fix RedHat bug 329431 The idea here is separate "conscious" from "unconscious" flushes. Conscious flushes are those due to a fsync() or close(). Unconscious ones are flushes that occur as a side effect of some other operation or due to memory pressure. Currently, when an error occurs during an unconscious flush (ENOSPC or EIO), we toss out the page and don't preserve that error to report to the user when a conscious flush occurs. If after the unconscious flush, there are no more dirty pages for the inode, the conscious flush will simply return success even though there were previous errors when writing out pages. This can lead to data corruption. The easiest way to reproduce this is to mount up a CIFS share that's very close to being full or where the user is very close to quota. mv a file to the share that's slightly larger than the quota allows. The writes will all succeed (since they go to pagecache). The mv will do a setattr to set the new file's attributes. This calls filemap_write_and_wait, which will return an error since all of the pages can't be written out. Then later, when the flush and release ops occur, there are no more dirty pages in pagecache for the file and those operations return 0. mv then assumes that the file was written out correctly and deletes the original. CIFS already has a write_behind_rc variable where it stores the results from earlier flushes, but that value is only reported in cifs_close. Since the VFS ignores the return value from the release operation, this isn't helpful. We should be reporting this error during the flush operation. This patch does the following: 1) changes cifs_fsync to use filemap_write_and_wait and cifs_flush and also sync to check its return code. If it returns successful, they then check the value of write_behind_rc to see if an earlier flush had reported any errors. If so, they return that error and clear write_behind_rc. 2) sets write_behind_rc in a few other places where pages are written out as a side effect of other operations and the code waits on them. 3) changes cifs_setattr to only call filemap_write_and_wait for ATTR_SIZE changes. 4) makes cifs_writepages accurately distinguish between EIO and ENOSPC errors when writing out pages. Some simple testing indicates that the patch works as expected and that it fixes the reproduceable known problem. Acked-by: Dave Kleikamp <shaggy@austin.rr.com> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2007-11-20 16:19:03 -07:00
if (rc != 0)
pCifsInode->write_behind_rc = rc;
}
cFYI(1, "invalidating remote inode since open detected it "
"changed");
invalidate_remote_inode(inode);
}
client_can_cache:
if (pTcon->unix_ext)
rc = cifs_get_inode_info_unix(&inode, full_path, inode->i_sb,
xid);
else
rc = cifs_get_inode_info(&inode, full_path, buf, inode->i_sb,
xid, NULL);
if ((oplock & 0xF) == OPLOCK_EXCLUSIVE) {
pCifsInode->clientCanCacheAll = true;
pCifsInode->clientCanCacheRead = true;
cFYI(1, "Exclusive Oplock granted on inode %p", inode);
} else if ((oplock & 0xF) == OPLOCK_READ)
pCifsInode->clientCanCacheRead = true;
return rc;
}
int cifs_posix_open(char *full_path, struct inode **pinode,
struct super_block *sb, int mode, unsigned int f_flags,
__u32 *poplock, __u16 *pnetfid, int xid)
{
int rc;
FILE_UNIX_BASIC_INFO *presp_data;
__u32 posix_flags = 0;
struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
struct cifs_fattr fattr;
struct tcon_link *tlink;
struct cifsTconInfo *tcon;
cFYI(1, "posix open %s", full_path);
presp_data = kzalloc(sizeof(FILE_UNIX_BASIC_INFO), GFP_KERNEL);
if (presp_data == NULL)
return -ENOMEM;
tlink = cifs_sb_tlink(cifs_sb);
if (IS_ERR(tlink)) {
rc = PTR_ERR(tlink);
goto posix_open_ret;
}
tcon = tlink_tcon(tlink);
mode &= ~current_umask();
posix_flags = cifs_posix_convert_flags(f_flags);
rc = CIFSPOSIXCreate(xid, tcon, posix_flags, mode, pnetfid, presp_data,
poplock, full_path, cifs_sb->local_nls,
cifs_sb->mnt_cifs_flags &
CIFS_MOUNT_MAP_SPECIAL_CHR);
cifs_put_tlink(tlink);
if (rc)
goto posix_open_ret;
if (presp_data->Type == cpu_to_le32(-1))
goto posix_open_ret; /* open ok, caller does qpathinfo */
if (!pinode)
goto posix_open_ret; /* caller does not need info */
cifs_unix_basic_to_fattr(&fattr, presp_data, cifs_sb);
/* get new inode and set it up */
if (*pinode == NULL) {
cifs_fill_uniqueid(sb, &fattr);
*pinode = cifs_iget(sb, &fattr);
if (!*pinode) {
rc = -ENOMEM;
goto posix_open_ret;
}
} else {
cifs_fattr_to_inode(*pinode, &fattr);
}
posix_open_ret:
kfree(presp_data);
return rc;
}
struct cifsFileInfo *
cifs_new_fileinfo(__u16 fileHandle, struct file *file,
struct tcon_link *tlink, __u32 oplock)
{
struct dentry *dentry = file->f_path.dentry;
struct inode *inode = dentry->d_inode;
struct cifsInodeInfo *pCifsInode = CIFS_I(inode);
struct cifsFileInfo *pCifsFile;
pCifsFile = kzalloc(sizeof(struct cifsFileInfo), GFP_KERNEL);
if (pCifsFile == NULL)
return pCifsFile;
pCifsFile->netfid = fileHandle;
pCifsFile->pid = current->tgid;
pCifsFile->uid = current_fsuid();
pCifsFile->dentry = dget(dentry);
pCifsFile->f_flags = file->f_flags;
pCifsFile->invalidHandle = false;
pCifsFile->tlink = cifs_get_tlink(tlink);
mutex_init(&pCifsFile->fh_mutex);
mutex_init(&pCifsFile->lock_mutex);
INIT_LIST_HEAD(&pCifsFile->llist);
atomic_set(&pCifsFile->count, 1);
INIT_WORK(&pCifsFile->oplock_break, cifs_oplock_break);
spin_lock(&cifs_file_list_lock);
list_add(&pCifsFile->tlist, &(tlink_tcon(tlink)->openFileList));
/* if readable file instance put first in list*/
if (file->f_mode & FMODE_READ)
list_add(&pCifsFile->flist, &pCifsInode->openFileList);
else
list_add_tail(&pCifsFile->flist, &pCifsInode->openFileList);
spin_unlock(&cifs_file_list_lock);
if ((oplock & 0xF) == OPLOCK_EXCLUSIVE) {
pCifsInode->clientCanCacheAll = true;
pCifsInode->clientCanCacheRead = true;
cFYI(1, "Exclusive Oplock inode %p", inode);
} else if ((oplock & 0xF) == OPLOCK_READ)
pCifsInode->clientCanCacheRead = true;
file->private_data = pCifsFile;
return pCifsFile;
}
/*
* Release a reference on the file private data. This may involve closing
* the filehandle out on the server.
*/
void cifsFileInfo_put(struct cifsFileInfo *cifs_file)
{
struct cifsTconInfo *tcon = tlink_tcon(cifs_file->tlink);
struct cifsInodeInfo *cifsi = CIFS_I(cifs_file->dentry->d_inode);
struct cifsLockInfo *li, *tmp;
spin_lock(&cifs_file_list_lock);
if (!atomic_dec_and_test(&cifs_file->count)) {
spin_unlock(&cifs_file_list_lock);
return;
}
/* remove it from the lists */
list_del(&cifs_file->flist);
list_del(&cifs_file->tlist);
if (list_empty(&cifsi->openFileList)) {
cFYI(1, "closing last open instance for inode %p",
cifs_file->dentry->d_inode);
cifsi->clientCanCacheRead = false;
cifsi->clientCanCacheAll = false;
}
spin_unlock(&cifs_file_list_lock);
if (!tcon->need_reconnect && !cifs_file->invalidHandle) {
int xid, rc;
xid = GetXid();
rc = CIFSSMBClose(xid, tcon, cifs_file->netfid);
FreeXid(xid);
}
/* Delete any outstanding lock records. We'll lose them when the file
* is closed anyway.
*/
mutex_lock(&cifs_file->lock_mutex);
list_for_each_entry_safe(li, tmp, &cifs_file->llist, llist) {
list_del(&li->llist);
kfree(li);
}
mutex_unlock(&cifs_file->lock_mutex);
cifs_put_tlink(cifs_file->tlink);
dput(cifs_file->dentry);
kfree(cifs_file);
}
int cifs_open(struct inode *inode, struct file *file)
{
int rc = -EACCES;
int xid;
__u32 oplock;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *tcon;
struct tcon_link *tlink;
struct cifsFileInfo *pCifsFile = NULL;
struct cifsInodeInfo *pCifsInode;
char *full_path = NULL;
int desiredAccess;
int disposition;
__u16 netfid;
FILE_ALL_INFO *buf = NULL;
xid = GetXid();
cifs_sb = CIFS_SB(inode->i_sb);
tlink = cifs_sb_tlink(cifs_sb);
if (IS_ERR(tlink)) {
FreeXid(xid);
return PTR_ERR(tlink);
}
tcon = tlink_tcon(tlink);
pCifsInode = CIFS_I(file->f_path.dentry->d_inode);
full_path = build_path_from_dentry(file->f_path.dentry);
if (full_path == NULL) {
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
rc = -ENOMEM;
goto out;
}
cFYI(1, "inode = 0x%p file flags are 0x%x for %s",
inode, file->f_flags, full_path);
if (oplockEnabled)
oplock = REQ_OPLOCK;
else
oplock = 0;
if (!tcon->broken_posix_open && tcon->unix_ext &&
(tcon->ses->capabilities & CAP_UNIX) &&
(CIFS_UNIX_POSIX_PATH_OPS_CAP &
le64_to_cpu(tcon->fsUnixInfo.Capability))) {
/* can not refresh inode info since size could be stale */
rc = cifs_posix_open(full_path, &inode, inode->i_sb,
cifs_sb->mnt_file_mode /* ignored */,
file->f_flags, &oplock, &netfid, xid);
if (rc == 0) {
cFYI(1, "posix open succeeded");
pCifsFile = cifs_new_fileinfo(netfid, file, tlink,
oplock);
if (pCifsFile == NULL) {
CIFSSMBClose(xid, tcon, netfid);
rc = -ENOMEM;
}
cifs_fscache_set_inode_cookie(inode, file);
goto out;
} else if ((rc == -EINVAL) || (rc == -EOPNOTSUPP)) {
if (tcon->ses->serverNOS)
cERROR(1, "server %s of type %s returned"
" unexpected error on SMB posix open"
", disabling posix open support."
" Check if server update available.",
tcon->ses->serverName,
tcon->ses->serverNOS);
tcon->broken_posix_open = true;
} else if ((rc != -EIO) && (rc != -EREMOTE) &&
(rc != -EOPNOTSUPP)) /* path not found or net err */
goto out;
/* else fallthrough to retry open the old way on network i/o
or DFS errors */
}
desiredAccess = cifs_convert_flags(file->f_flags);
/*********************************************************************
* open flag mapping table:
*
* POSIX Flag CIFS Disposition
* ---------- ----------------
* O_CREAT FILE_OPEN_IF
* O_CREAT | O_EXCL FILE_CREATE
* O_CREAT | O_TRUNC FILE_OVERWRITE_IF
* O_TRUNC FILE_OVERWRITE
* none of the above FILE_OPEN
*
* Note that there is not a direct match between disposition
* FILE_SUPERSEDE (ie create whether or not file exists although
* O_CREAT | O_TRUNC is similar but truncates the existing
* file rather than creating a new file as FILE_SUPERSEDE does
* (which uses the attributes / metadata passed in on open call)
*?
*? O_SYNC is a reasonable match to CIFS writethrough flag
*? and the read write flags match reasonably. O_LARGEFILE
*? is irrelevant because largefile support is always used
*? by this client. Flags O_APPEND, O_DIRECT, O_DIRECTORY,
* O_FASYNC, O_NOFOLLOW, O_NONBLOCK need further investigation
*********************************************************************/
disposition = cifs_get_disposition(file->f_flags);
/* BB pass O_SYNC flag through on file attributes .. BB */
/* Also refresh inode by passing in file_info buf returned by SMBOpen
and calling get_inode_info with returned buf (at least helps
non-Unix server case) */
/* BB we can not do this if this is the second open of a file
and the first handle has writebehind data, we might be
able to simply do a filemap_fdatawrite/filemap_fdatawait first */
buf = kmalloc(sizeof(FILE_ALL_INFO), GFP_KERNEL);
if (!buf) {
rc = -ENOMEM;
goto out;
}
if (tcon->ses->capabilities & CAP_NT_SMBS)
rc = CIFSSMBOpen(xid, tcon, full_path, disposition,
desiredAccess, CREATE_NOT_DIR, &netfid, &oplock, buf,
cifs_sb->local_nls, cifs_sb->mnt_cifs_flags
& CIFS_MOUNT_MAP_SPECIAL_CHR);
else
rc = -EIO; /* no NT SMB support fall into legacy open below */
if (rc == -EIO) {
/* Old server, try legacy style OpenX */
rc = SMBLegacyOpen(xid, tcon, full_path, disposition,
desiredAccess, CREATE_NOT_DIR, &netfid, &oplock, buf,
cifs_sb->local_nls, cifs_sb->mnt_cifs_flags
& CIFS_MOUNT_MAP_SPECIAL_CHR);
}
if (rc) {
cFYI(1, "cifs_open returned 0x%x", rc);
goto out;
}
rc = cifs_open_inode_helper(inode, tcon, oplock, buf, full_path, xid);
if (rc != 0)
goto out;
pCifsFile = cifs_new_fileinfo(netfid, file, tlink, oplock);
if (pCifsFile == NULL) {
rc = -ENOMEM;
goto out;
}
cifs_fscache_set_inode_cookie(inode, file);
if (oplock & CIFS_CREATE_ACTION) {
/* time to set mode which we can not set earlier due to
problems creating new read-only files */
if (tcon->unix_ext) {
struct cifs_unix_set_info_args args = {
.mode = inode->i_mode,
.uid = NO_CHANGE_64,
.gid = NO_CHANGE_64,
.ctime = NO_CHANGE_64,
.atime = NO_CHANGE_64,
.mtime = NO_CHANGE_64,
.device = 0,
};
CIFSSMBUnixSetPathInfo(xid, tcon, full_path, &args,
cifs_sb->local_nls,
cifs_sb->mnt_cifs_flags &
CIFS_MOUNT_MAP_SPECIAL_CHR);
}
}
out:
kfree(buf);
kfree(full_path);
FreeXid(xid);
cifs_put_tlink(tlink);
return rc;
}
/* Try to reacquire byte range locks that were released when session */
/* to server was lost */
static int cifs_relock_file(struct cifsFileInfo *cifsFile)
{
int rc = 0;
/* BB list all locks open on this file and relock */
return rc;
}
static int cifs_reopen_file(struct cifsFileInfo *pCifsFile, bool can_flush)
{
int rc = -EACCES;
int xid;
__u32 oplock;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *tcon;
struct cifsInodeInfo *pCifsInode;
struct inode *inode;
char *full_path = NULL;
int desiredAccess;
int disposition = FILE_OPEN;
__u16 netfid;
xid = GetXid();
mutex_lock(&pCifsFile->fh_mutex);
if (!pCifsFile->invalidHandle) {
mutex_unlock(&pCifsFile->fh_mutex);
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
rc = 0;
FreeXid(xid);
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
return rc;
}
inode = pCifsFile->dentry->d_inode;
cifs_sb = CIFS_SB(inode->i_sb);
tcon = tlink_tcon(pCifsFile->tlink);
/* can not grab rename sem here because various ops, including
those that already have the rename sem can end up causing writepage
to get called and if the server was down that means we end up here,
and we can never tell if the caller already has the rename_sem */
full_path = build_path_from_dentry(pCifsFile->dentry);
if (full_path == NULL) {
rc = -ENOMEM;
mutex_unlock(&pCifsFile->fh_mutex);
FreeXid(xid);
return rc;
}
cFYI(1, "inode = 0x%p file flags 0x%x for %s",
inode, pCifsFile->f_flags, full_path);
if (oplockEnabled)
oplock = REQ_OPLOCK;
else
oplock = 0;
if (tcon->unix_ext && (tcon->ses->capabilities & CAP_UNIX) &&
(CIFS_UNIX_POSIX_PATH_OPS_CAP &
le64_to_cpu(tcon->fsUnixInfo.Capability))) {
/*
* O_CREAT, O_EXCL and O_TRUNC already had their effect on the
* original open. Must mask them off for a reopen.
*/
unsigned int oflags = pCifsFile->f_flags &
~(O_CREAT | O_EXCL | O_TRUNC);
rc = cifs_posix_open(full_path, NULL, inode->i_sb,
cifs_sb->mnt_file_mode /* ignored */,
oflags, &oplock, &netfid, xid);
if (rc == 0) {
cFYI(1, "posix reopen succeeded");
goto reopen_success;
}
/* fallthrough to retry open the old way on errors, especially
in the reconnect path it is important to retry hard */
}
desiredAccess = cifs_convert_flags(pCifsFile->f_flags);
/* Can not refresh inode by passing in file_info buf to be returned
by SMBOpen and then calling get_inode_info with returned buf
since file might have write behind data that needs to be flushed
and server version of file size can be stale. If we knew for sure
that inode was not dirty locally we could do this */
rc = CIFSSMBOpen(xid, tcon, full_path, disposition, desiredAccess,
CREATE_NOT_DIR, &netfid, &oplock, NULL,
cifs_sb->local_nls, cifs_sb->mnt_cifs_flags &
CIFS_MOUNT_MAP_SPECIAL_CHR);
if (rc) {
mutex_unlock(&pCifsFile->fh_mutex);
cFYI(1, "cifs_open returned 0x%x", rc);
cFYI(1, "oplock: %d", oplock);
goto reopen_error_exit;
}
reopen_success:
pCifsFile->netfid = netfid;
pCifsFile->invalidHandle = false;
mutex_unlock(&pCifsFile->fh_mutex);
pCifsInode = CIFS_I(inode);
if (can_flush) {
rc = filemap_write_and_wait(inode->i_mapping);
if (rc != 0)
CIFS_I(inode)->write_behind_rc = rc;
pCifsInode->clientCanCacheAll = false;
pCifsInode->clientCanCacheRead = false;
if (tcon->unix_ext)
rc = cifs_get_inode_info_unix(&inode,
full_path, inode->i_sb, xid);
else
rc = cifs_get_inode_info(&inode,
full_path, NULL, inode->i_sb,
xid, NULL);
} /* else we are writing out data to server already
and could deadlock if we tried to flush data, and
since we do not know if we have data that would
invalidate the current end of file on the server
we can not go to the server to get the new inod
info */
if ((oplock & 0xF) == OPLOCK_EXCLUSIVE) {
pCifsInode->clientCanCacheAll = true;
pCifsInode->clientCanCacheRead = true;
cFYI(1, "Exclusive Oplock granted on inode %p",
pCifsFile->dentry->d_inode);
} else if ((oplock & 0xF) == OPLOCK_READ) {
pCifsInode->clientCanCacheRead = true;
pCifsInode->clientCanCacheAll = false;
} else {
pCifsInode->clientCanCacheRead = false;
pCifsInode->clientCanCacheAll = false;
}
cifs_relock_file(pCifsFile);
reopen_error_exit:
kfree(full_path);
FreeXid(xid);
return rc;
}
int cifs_close(struct inode *inode, struct file *file)
{
cifsFileInfo_put(file->private_data);
file->private_data = NULL;
/* return code from the ->release op is always ignored */
return 0;
}
int cifs_closedir(struct inode *inode, struct file *file)
{
int rc = 0;
int xid;
struct cifsFileInfo *pCFileStruct = file->private_data;
char *ptmp;
cFYI(1, "Closedir inode = 0x%p", inode);
xid = GetXid();
if (pCFileStruct) {
struct cifsTconInfo *pTcon = tlink_tcon(pCFileStruct->tlink);
cFYI(1, "Freeing private data in close dir");
spin_lock(&cifs_file_list_lock);
if (!pCFileStruct->srch_inf.endOfSearch &&
!pCFileStruct->invalidHandle) {
pCFileStruct->invalidHandle = true;
spin_unlock(&cifs_file_list_lock);
rc = CIFSFindClose(xid, pTcon, pCFileStruct->netfid);
cFYI(1, "Closing uncompleted readdir with rc %d",
rc);
/* not much we can do if it fails anyway, ignore rc */
rc = 0;
} else
spin_unlock(&cifs_file_list_lock);
ptmp = pCFileStruct->srch_inf.ntwrk_buf_start;
if (ptmp) {
cFYI(1, "closedir free smb buf in srch struct");
pCFileStruct->srch_inf.ntwrk_buf_start = NULL;
if (pCFileStruct->srch_inf.smallBuf)
cifs_small_buf_release(ptmp);
else
cifs_buf_release(ptmp);
}
cifs_put_tlink(pCFileStruct->tlink);
kfree(file->private_data);
file->private_data = NULL;
}
/* BB can we lock the filestruct while this is going on? */
FreeXid(xid);
return rc;
}
static int store_file_lock(struct cifsFileInfo *fid, __u64 len,
__u64 offset, __u8 lockType)
{
struct cifsLockInfo *li =
kmalloc(sizeof(struct cifsLockInfo), GFP_KERNEL);
if (li == NULL)
return -ENOMEM;
li->offset = offset;
li->length = len;
li->type = lockType;
mutex_lock(&fid->lock_mutex);
list_add(&li->llist, &fid->llist);
mutex_unlock(&fid->lock_mutex);
return 0;
}
int cifs_lock(struct file *file, int cmd, struct file_lock *pfLock)
{
int rc, xid;
__u32 numLock = 0;
__u32 numUnlock = 0;
__u64 length;
bool wait_flag = false;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *tcon;
__u16 netfid;
__u8 lockType = LOCKING_ANDX_LARGE_FILES;
bool posix_locking = 0;
length = 1 + pfLock->fl_end - pfLock->fl_start;
rc = -EACCES;
xid = GetXid();
cFYI(1, "Lock parm: 0x%x flockflags: "
"0x%x flocktype: 0x%x start: %lld end: %lld",
cmd, pfLock->fl_flags, pfLock->fl_type, pfLock->fl_start,
pfLock->fl_end);
if (pfLock->fl_flags & FL_POSIX)
cFYI(1, "Posix");
if (pfLock->fl_flags & FL_FLOCK)
cFYI(1, "Flock");
if (pfLock->fl_flags & FL_SLEEP) {
cFYI(1, "Blocking lock");
wait_flag = true;
}
if (pfLock->fl_flags & FL_ACCESS)
cFYI(1, "Process suspended by mandatory locking - "
"not implemented yet");
if (pfLock->fl_flags & FL_LEASE)
cFYI(1, "Lease on file - not implemented yet");
if (pfLock->fl_flags &
(~(FL_POSIX | FL_FLOCK | FL_SLEEP | FL_ACCESS | FL_LEASE)))
cFYI(1, "Unknown lock flags 0x%x", pfLock->fl_flags);
if (pfLock->fl_type == F_WRLCK) {
cFYI(1, "F_WRLCK ");
numLock = 1;
} else if (pfLock->fl_type == F_UNLCK) {
cFYI(1, "F_UNLCK");
numUnlock = 1;
/* Check if unlock includes more than
one lock range */
} else if (pfLock->fl_type == F_RDLCK) {
cFYI(1, "F_RDLCK");
lockType |= LOCKING_ANDX_SHARED_LOCK;
numLock = 1;
} else if (pfLock->fl_type == F_EXLCK) {
cFYI(1, "F_EXLCK");
numLock = 1;
} else if (pfLock->fl_type == F_SHLCK) {
cFYI(1, "F_SHLCK");
lockType |= LOCKING_ANDX_SHARED_LOCK;
numLock = 1;
} else
cFYI(1, "Unknown type of lock");
cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
tcon = tlink_tcon(((struct cifsFileInfo *)file->private_data)->tlink);
if (file->private_data == NULL) {
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
rc = -EBADF;
FreeXid(xid);
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
return rc;
}
netfid = ((struct cifsFileInfo *)file->private_data)->netfid;
if ((tcon->ses->capabilities & CAP_UNIX) &&
(CIFS_UNIX_FCNTL_CAP & le64_to_cpu(tcon->fsUnixInfo.Capability)) &&
((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) == 0))
posix_locking = 1;
/* BB add code here to normalize offset and length to
account for negative length which we can not accept over the
wire */
if (IS_GETLK(cmd)) {
if (posix_locking) {
int posix_lock_type;
if (lockType & LOCKING_ANDX_SHARED_LOCK)
posix_lock_type = CIFS_RDLCK;
else
posix_lock_type = CIFS_WRLCK;
rc = CIFSSMBPosixLock(xid, tcon, netfid, 1 /* get */,
length, pfLock,
posix_lock_type, wait_flag);
FreeXid(xid);
return rc;
}
/* BB we could chain these into one lock request BB */
rc = CIFSSMBLock(xid, tcon, netfid, length, pfLock->fl_start,
0, 1, lockType, 0 /* wait flag */ );
if (rc == 0) {
rc = CIFSSMBLock(xid, tcon, netfid, length,
pfLock->fl_start, 1 /* numUnlock */ ,
0 /* numLock */ , lockType,
0 /* wait flag */ );
pfLock->fl_type = F_UNLCK;
if (rc != 0)
cERROR(1, "Error unlocking previously locked "
"range %d during test of lock", rc);
rc = 0;
} else {
/* if rc == ERR_SHARING_VIOLATION ? */
rc = 0;
if (lockType & LOCKING_ANDX_SHARED_LOCK) {
pfLock->fl_type = F_WRLCK;
} else {
rc = CIFSSMBLock(xid, tcon, netfid, length,
pfLock->fl_start, 0, 1,
lockType | LOCKING_ANDX_SHARED_LOCK,
0 /* wait flag */);
if (rc == 0) {
rc = CIFSSMBLock(xid, tcon, netfid,
length, pfLock->fl_start, 1, 0,
lockType |
LOCKING_ANDX_SHARED_LOCK,
0 /* wait flag */);
pfLock->fl_type = F_RDLCK;
if (rc != 0)
cERROR(1, "Error unlocking "
"previously locked range %d "
"during test of lock", rc);
rc = 0;
} else {
pfLock->fl_type = F_WRLCK;
rc = 0;
}
}
}
FreeXid(xid);
return rc;
}
if (!numLock && !numUnlock) {
/* if no lock or unlock then nothing
to do since we do not know what it is */
FreeXid(xid);
return -EOPNOTSUPP;
}
if (posix_locking) {
int posix_lock_type;
if (lockType & LOCKING_ANDX_SHARED_LOCK)
posix_lock_type = CIFS_RDLCK;
else
posix_lock_type = CIFS_WRLCK;
if (numUnlock == 1)
posix_lock_type = CIFS_UNLCK;
rc = CIFSSMBPosixLock(xid, tcon, netfid, 0 /* set */,
length, pfLock,
posix_lock_type, wait_flag);
} else {
struct cifsFileInfo *fid = file->private_data;
if (numLock) {
rc = CIFSSMBLock(xid, tcon, netfid, length,
pfLock->fl_start,
0, numLock, lockType, wait_flag);
if (rc == 0) {
/* For Windows locks we must store them. */
rc = store_file_lock(fid, length,
pfLock->fl_start, lockType);
}
} else if (numUnlock) {
/* For each stored lock that this unlock overlaps
completely, unlock it. */
int stored_rc = 0;
struct cifsLockInfo *li, *tmp;
rc = 0;
mutex_lock(&fid->lock_mutex);
list_for_each_entry_safe(li, tmp, &fid->llist, llist) {
if (pfLock->fl_start <= li->offset &&
(pfLock->fl_start + length) >=
(li->offset + li->length)) {
stored_rc = CIFSSMBLock(xid, tcon,
netfid,
li->length, li->offset,
1, 0, li->type, false);
if (stored_rc)
rc = stored_rc;
else {
list_del(&li->llist);
kfree(li);
}
}
}
mutex_unlock(&fid->lock_mutex);
}
}
[PATCH] Fix oops in fs/locks.c on close of file with pending locks The recent change to locks_remove_flock code in fs/locks.c changes how byte range locks are removed from closing files, which shows up a bug in cifs. The assumption in the cifs code was that the close call sent to the server would remove any pending locks on the server on this file, but that is no longer safe as the fs/locks.c code on the client wants unlock of 0 to PATH_MAX to remove all locks (at least from this client, it is not possible AFAIK to remove all locks from other clients made to the server copy of the file). Note that cifs locks are different from posix locks - and it is not possible to map posix locks perfectly on the wire yet, due to restrictions of the cifs network protocol, even to Samba without adding a new request type to the network protocol (which we plan to do for Samba 3.0.21 within a few months), but the local client will have the correct, posix view, of the lock in most cases. The correct fix for cifs for this would involve a bigger change than I would like to do this late in the 2.6.13-rc cycle - and would involve cifs keeping track of all unmerged (uncoalesced) byte range locks for each remote inode and scanning that list to remove locks that intersect or fall wholly within the range - locks that intersect may have to be reaquired with the smaller, remaining range. Signed-off-by: Steve French <sfrench@us.ibm.com> Signed-off-by: Dave Kleikamp <shaggy@austin.ibm.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-08-26 13:42:59 -06:00
if (pfLock->fl_flags & FL_POSIX)
posix_lock_file_wait(file, pfLock);
FreeXid(xid);
return rc;
}
/*
* Set the timeout on write requests past EOF. For some servers (Windows)
* these calls can be very long.
*
* If we're writing >10M past the EOF we give a 180s timeout. Anything less
* than that gets a 45s timeout. Writes not past EOF get 15s timeouts.
* The 10M cutoff is totally arbitrary. A better scheme for this would be
* welcome if someone wants to suggest one.
*
* We may be able to do a better job with this if there were some way to
* declare that a file should be sparse.
*/
static int
cifs_write_timeout(struct cifsInodeInfo *cifsi, loff_t offset)
{
if (offset <= cifsi->server_eof)
return CIFS_STD_OP;
else if (offset > (cifsi->server_eof + (10 * 1024 * 1024)))
return CIFS_VLONG_OP;
else
return CIFS_LONG_OP;
}
/* update the file size (if needed) after a write */
static void
cifs_update_eof(struct cifsInodeInfo *cifsi, loff_t offset,
unsigned int bytes_written)
{
loff_t end_of_write = offset + bytes_written;
if (end_of_write > cifsi->server_eof)
cifsi->server_eof = end_of_write;
}
ssize_t cifs_user_write(struct file *file, const char __user *write_data,
size_t write_size, loff_t *poffset)
{
int rc = 0;
unsigned int bytes_written = 0;
unsigned int total_written;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *pTcon;
int xid, long_op;
struct cifsFileInfo *open_file;
struct cifsInodeInfo *cifsi = CIFS_I(file->f_path.dentry->d_inode);
cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
/* cFYI(1, " write %d bytes to offset %lld of %s", write_size,
*poffset, file->f_path.dentry->d_name.name); */
if (file->private_data == NULL)
return -EBADF;
open_file = file->private_data;
pTcon = tlink_tcon(open_file->tlink);
2008-08-28 05:54:59 -06:00
rc = generic_write_checks(file, poffset, &write_size, 0);
if (rc)
return rc;
xid = GetXid();
long_op = cifs_write_timeout(cifsi, *poffset);
for (total_written = 0; write_size > total_written;
total_written += bytes_written) {
rc = -EAGAIN;
while (rc == -EAGAIN) {
if (file->private_data == NULL) {
/* file has been closed on us */
FreeXid(xid);
/* if we have gotten here we have written some data
and blocked, and the file has been freed on us while
we blocked so return what we managed to write */
return total_written;
}
if (open_file->invalidHandle) {
/* we could deadlock if we called
filemap_fdatawait from here so tell
reopen_file not to flush data to server
now */
rc = cifs_reopen_file(open_file, false);
if (rc != 0)
break;
}
rc = CIFSSMBWrite(xid, pTcon,
open_file->netfid,
min_t(const int, cifs_sb->wsize,
write_size - total_written),
*poffset, &bytes_written,
NULL, write_data + total_written, long_op);
}
if (rc || (bytes_written == 0)) {
if (total_written)
break;
else {
FreeXid(xid);
return rc;
}
} else {
cifs_update_eof(cifsi, *poffset, bytes_written);
*poffset += bytes_written;
}
long_op = CIFS_STD_OP; /* subsequent writes fast -
15 seconds is plenty */
}
cifs_stats_bytes_written(pTcon, total_written);
/* since the write may have blocked check these pointers again */
if ((file->f_path.dentry) && (file->f_path.dentry->d_inode)) {
struct inode *inode = file->f_path.dentry->d_inode;
/* Do not update local mtime - server will set its actual value on write
* inode->i_ctime = inode->i_mtime =
* current_fs_time(inode->i_sb);*/
if (total_written > 0) {
spin_lock(&inode->i_lock);
if (*poffset > file->f_path.dentry->d_inode->i_size)
i_size_write(file->f_path.dentry->d_inode,
*poffset);
spin_unlock(&inode->i_lock);
}
mark_inode_dirty_sync(file->f_path.dentry->d_inode);
}
FreeXid(xid);
return total_written;
}
static ssize_t cifs_write(struct cifsFileInfo *open_file,
const char *write_data, size_t write_size,
loff_t *poffset)
{
int rc = 0;
unsigned int bytes_written = 0;
unsigned int total_written;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *pTcon;
int xid, long_op;
struct dentry *dentry = open_file->dentry;
struct cifsInodeInfo *cifsi = CIFS_I(dentry->d_inode);
cifs_sb = CIFS_SB(dentry->d_sb);
cFYI(1, "write %zd bytes to offset %lld of %s", write_size,
*poffset, dentry->d_name.name);
pTcon = tlink_tcon(open_file->tlink);
xid = GetXid();
long_op = cifs_write_timeout(cifsi, *poffset);
for (total_written = 0; write_size > total_written;
total_written += bytes_written) {
rc = -EAGAIN;
while (rc == -EAGAIN) {
if (open_file->invalidHandle) {
/* we could deadlock if we called
filemap_fdatawait from here so tell
reopen_file not to flush data to
server now */
rc = cifs_reopen_file(open_file, false);
if (rc != 0)
break;
}
if (experimEnabled || (pTcon->ses->server &&
((pTcon->ses->server->secMode &
(SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED))
== 0))) {
struct kvec iov[2];
unsigned int len;
len = min((size_t)cifs_sb->wsize,
write_size - total_written);
/* iov[0] is reserved for smb header */
iov[1].iov_base = (char *)write_data +
total_written;
iov[1].iov_len = len;
rc = CIFSSMBWrite2(xid, pTcon,
open_file->netfid, len,
*poffset, &bytes_written,
iov, 1, long_op);
} else
rc = CIFSSMBWrite(xid, pTcon,
open_file->netfid,
min_t(const int, cifs_sb->wsize,
write_size - total_written),
*poffset, &bytes_written,
write_data + total_written,
NULL, long_op);
}
if (rc || (bytes_written == 0)) {
if (total_written)
break;
else {
FreeXid(xid);
return rc;
}
} else {
cifs_update_eof(cifsi, *poffset, bytes_written);
*poffset += bytes_written;
}
long_op = CIFS_STD_OP; /* subsequent writes fast -
15 seconds is plenty */
}
cifs_stats_bytes_written(pTcon, total_written);
if (total_written > 0) {
spin_lock(&dentry->d_inode->i_lock);
if (*poffset > dentry->d_inode->i_size)
i_size_write(dentry->d_inode, *poffset);
spin_unlock(&dentry->d_inode->i_lock);
}
mark_inode_dirty_sync(dentry->d_inode);
FreeXid(xid);
return total_written;
}
#ifdef CONFIG_CIFS_EXPERIMENTAL
struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode,
bool fsuid_only)
{
struct cifsFileInfo *open_file = NULL;
struct cifs_sb_info *cifs_sb = CIFS_SB(cifs_inode->vfs_inode.i_sb);
/* only filter by fsuid on multiuser mounts */
if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER))
fsuid_only = false;
spin_lock(&cifs_file_list_lock);
/* we could simply get the first_list_entry since write-only entries
are always at the end of the list but since the first entry might
have a close pending, we go through the whole list */
list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
if (fsuid_only && open_file->uid != current_fsuid())
continue;
if (OPEN_FMODE(open_file->f_flags) & FMODE_READ) {
if (!open_file->invalidHandle) {
/* found a good file */
/* lock it so it will not be closed on us */
cifsFileInfo_get(open_file);
spin_unlock(&cifs_file_list_lock);
return open_file;
} /* else might as well continue, and look for
another, or simply have the caller reopen it
again rather than trying to fix this handle */
} else /* write only file */
break; /* write only files are last so must be done */
}
spin_unlock(&cifs_file_list_lock);
return NULL;
}
#endif
struct cifsFileInfo *find_writable_file(struct cifsInodeInfo *cifs_inode,
bool fsuid_only)
{
struct cifsFileInfo *open_file;
struct cifs_sb_info *cifs_sb = CIFS_SB(cifs_inode->vfs_inode.i_sb);
bool any_available = false;
int rc;
/* Having a null inode here (because mapping->host was set to zero by
the VFS or MM) should not happen but we had reports of on oops (due to
it being zero) during stress testcases so we need to check for it */
if (cifs_inode == NULL) {
cERROR(1, "Null inode passed to cifs_writeable_file");
dump_stack();
return NULL;
}
/* only filter by fsuid on multiuser mounts */
if (!(cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER))
fsuid_only = false;
spin_lock(&cifs_file_list_lock);
refind_writable:
list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
if (!any_available && open_file->pid != current->tgid)
continue;
if (fsuid_only && open_file->uid != current_fsuid())
continue;
if (OPEN_FMODE(open_file->f_flags) & FMODE_WRITE) {
cifsFileInfo_get(open_file);
if (!open_file->invalidHandle) {
/* found a good writable file */
spin_unlock(&cifs_file_list_lock);
return open_file;
}
spin_unlock(&cifs_file_list_lock);
/* Had to unlock since following call can block */
rc = cifs_reopen_file(open_file, false);
if (!rc)
return open_file;
/* if it fails, try another handle if possible */
cFYI(1, "wp failed on reopen file");
cifsFileInfo_put(open_file);
spin_lock(&cifs_file_list_lock);
/* else we simply continue to the next entry. Thus
we do not loop on reopen errors. If we
can not reopen the file, for example if we
reconnected to a server with another client
racing to delete or lock the file we would not
make progress if we restarted before the beginning
of the loop here. */
}
}
/* couldn't find useable FH with same pid, try any available */
if (!any_available) {
any_available = true;
goto refind_writable;
}
spin_unlock(&cifs_file_list_lock);
return NULL;
}
static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to)
{
struct address_space *mapping = page->mapping;
loff_t offset = (loff_t)page->index << PAGE_CACHE_SHIFT;
char *write_data;
int rc = -EFAULT;
int bytes_written = 0;
struct cifs_sb_info *cifs_sb;
struct inode *inode;
struct cifsFileInfo *open_file;
if (!mapping || !mapping->host)
return -EFAULT;
inode = page->mapping->host;
cifs_sb = CIFS_SB(inode->i_sb);
offset += (loff_t)from;
write_data = kmap(page);
write_data += from;
if ((to > PAGE_CACHE_SIZE) || (from > to)) {
kunmap(page);
return -EIO;
}
/* racing with truncate? */
if (offset > mapping->host->i_size) {
kunmap(page);
return 0; /* don't care */
}
/* check to make sure that we are not extending the file */
if (mapping->host->i_size - offset < (loff_t)to)
to = (unsigned)(mapping->host->i_size - offset);
open_file = find_writable_file(CIFS_I(mapping->host), false);
if (open_file) {
bytes_written = cifs_write(open_file, write_data,
to - from, &offset);
cifsFileInfo_put(open_file);
/* Does mm or vfs already set times? */
inode->i_atime = inode->i_mtime = current_fs_time(inode->i_sb);
if ((bytes_written > 0) && (offset))
rc = 0;
else if (bytes_written < 0)
rc = bytes_written;
} else {
cFYI(1, "No writeable filehandles for inode");
rc = -EIO;
}
kunmap(page);
return rc;
}
static int cifs_writepages(struct address_space *mapping,
struct writeback_control *wbc)
{
struct backing_dev_info *bdi = mapping->backing_dev_info;
unsigned int bytes_to_write;
unsigned int bytes_written;
struct cifs_sb_info *cifs_sb;
int done = 0;
[PATCH] writeback: fix range handling When a writeback_control's `start' and `end' fields are used to indicate a one-byte-range starting at file offset zero, the required values of .start=0,.end=0 mean that the ->writepages() implementation has no way of telling that it is being asked to perform a range request. Because we're currently overloading (start == 0 && end == 0) to mean "this is not a write-a-range request". To make all this sane, the patch changes range of writeback_control. So caller does: If it is calling ->writepages() to write pages, it sets range (range_start/end or range_cyclic) always. And if range_cyclic is true, ->writepages() thinks the range is cyclic, otherwise it just uses range_start and range_end. This patch does, - Add LLONG_MAX, LLONG_MIN, ULLONG_MAX to include/linux/kernel.h -1 is usually ok for range_end (type is long long). But, if someone did, range_end += val; range_end is "val - 1" u64val = range_end >> bits; u64val is "~(0ULL)" or something, they are wrong. So, this adds LLONG_MAX to avoid nasty things, and uses LLONG_MAX for range_end. - All callers of ->writepages() sets range_start/end or range_cyclic. - Fix updates of ->writeback_index. It seems already bit strange. If it starts at 0 and ended by check of nr_to_write, this last index may reduce chance to scan end of file. So, this updates ->writeback_index only if range_cyclic is true or whole-file is scanned. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Nathan Scott <nathans@sgi.com> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Steven French <sfrench@us.ibm.com> Cc: "Vladimir V. Saveliev" <vs@namesys.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 03:03:26 -06:00
pgoff_t end;
pgoff_t index;
int range_whole = 0;
struct kvec *iov;
int len;
int n_iov = 0;
pgoff_t next;
int nr_pages;
__u64 offset = 0;
struct cifsFileInfo *open_file;
struct cifsTconInfo *tcon;
struct cifsInodeInfo *cifsi = CIFS_I(mapping->host);
struct page *page;
struct pagevec pvec;
int rc = 0;
int scanned = 0;
int xid, long_op;
/*
* BB: Is this meaningful for a non-block-device file system?
* If it is, we should test it again after we do I/O
*/
if (wbc->nonblocking && bdi_write_congested(bdi)) {
wbc->encountered_congestion = 1;
return 0;
}
cifs_sb = CIFS_SB(mapping->host->i_sb);
/*
* If wsize is smaller that the page cache size, default to writing
* one page at a time via cifs_writepage
*/
if (cifs_sb->wsize < PAGE_CACHE_SIZE)
return generic_writepages(mapping, wbc);
iov = kmalloc(32 * sizeof(struct kvec), GFP_KERNEL);
if (iov == NULL)
return generic_writepages(mapping, wbc);
/*
* if there's no open file, then this is likely to fail too,
* but it'll at least handle the return. Maybe it should be
* a BUG() instead?
*/
open_file = find_writable_file(CIFS_I(mapping->host), false);
if (!open_file) {
kfree(iov);
return generic_writepages(mapping, wbc);
}
tcon = tlink_tcon(open_file->tlink);
if (!experimEnabled && tcon->ses->server->secMode &
(SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) {
cifsFileInfo_put(open_file);
return generic_writepages(mapping, wbc);
}
cifsFileInfo_put(open_file);
xid = GetXid();
pagevec_init(&pvec, 0);
[PATCH] writeback: fix range handling When a writeback_control's `start' and `end' fields are used to indicate a one-byte-range starting at file offset zero, the required values of .start=0,.end=0 mean that the ->writepages() implementation has no way of telling that it is being asked to perform a range request. Because we're currently overloading (start == 0 && end == 0) to mean "this is not a write-a-range request". To make all this sane, the patch changes range of writeback_control. So caller does: If it is calling ->writepages() to write pages, it sets range (range_start/end or range_cyclic) always. And if range_cyclic is true, ->writepages() thinks the range is cyclic, otherwise it just uses range_start and range_end. This patch does, - Add LLONG_MAX, LLONG_MIN, ULLONG_MAX to include/linux/kernel.h -1 is usually ok for range_end (type is long long). But, if someone did, range_end += val; range_end is "val - 1" u64val = range_end >> bits; u64val is "~(0ULL)" or something, they are wrong. So, this adds LLONG_MAX to avoid nasty things, and uses LLONG_MAX for range_end. - All callers of ->writepages() sets range_start/end or range_cyclic. - Fix updates of ->writeback_index. It seems already bit strange. If it starts at 0 and ended by check of nr_to_write, this last index may reduce chance to scan end of file. So, this updates ->writeback_index only if range_cyclic is true or whole-file is scanned. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Nathan Scott <nathans@sgi.com> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Steven French <sfrench@us.ibm.com> Cc: "Vladimir V. Saveliev" <vs@namesys.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 03:03:26 -06:00
if (wbc->range_cyclic) {
index = mapping->writeback_index; /* Start from prev offset */
[PATCH] writeback: fix range handling When a writeback_control's `start' and `end' fields are used to indicate a one-byte-range starting at file offset zero, the required values of .start=0,.end=0 mean that the ->writepages() implementation has no way of telling that it is being asked to perform a range request. Because we're currently overloading (start == 0 && end == 0) to mean "this is not a write-a-range request". To make all this sane, the patch changes range of writeback_control. So caller does: If it is calling ->writepages() to write pages, it sets range (range_start/end or range_cyclic) always. And if range_cyclic is true, ->writepages() thinks the range is cyclic, otherwise it just uses range_start and range_end. This patch does, - Add LLONG_MAX, LLONG_MIN, ULLONG_MAX to include/linux/kernel.h -1 is usually ok for range_end (type is long long). But, if someone did, range_end += val; range_end is "val - 1" u64val = range_end >> bits; u64val is "~(0ULL)" or something, they are wrong. So, this adds LLONG_MAX to avoid nasty things, and uses LLONG_MAX for range_end. - All callers of ->writepages() sets range_start/end or range_cyclic. - Fix updates of ->writeback_index. It seems already bit strange. If it starts at 0 and ended by check of nr_to_write, this last index may reduce chance to scan end of file. So, this updates ->writeback_index only if range_cyclic is true or whole-file is scanned. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Nathan Scott <nathans@sgi.com> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Steven French <sfrench@us.ibm.com> Cc: "Vladimir V. Saveliev" <vs@namesys.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 03:03:26 -06:00
end = -1;
} else {
index = wbc->range_start >> PAGE_CACHE_SHIFT;
end = wbc->range_end >> PAGE_CACHE_SHIFT;
if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
range_whole = 1;
scanned = 1;
}
retry:
while (!done && (index <= end) &&
(nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
PAGECACHE_TAG_DIRTY,
min(end - index, (pgoff_t)PAGEVEC_SIZE - 1) + 1))) {
int first;
unsigned int i;
first = -1;
next = 0;
n_iov = 0;
bytes_to_write = 0;
for (i = 0; i < nr_pages; i++) {
page = pvec.pages[i];
/*
* At this point we hold neither mapping->tree_lock nor
* lock on the page itself: the page may be truncated or
* invalidated (changing page->mapping to NULL), or even
* swizzled back from swapper_space to tmpfs file
* mapping
*/
if (first < 0)
lock_page(page);
else if (!trylock_page(page))
break;
if (unlikely(page->mapping != mapping)) {
unlock_page(page);
break;
}
[PATCH] writeback: fix range handling When a writeback_control's `start' and `end' fields are used to indicate a one-byte-range starting at file offset zero, the required values of .start=0,.end=0 mean that the ->writepages() implementation has no way of telling that it is being asked to perform a range request. Because we're currently overloading (start == 0 && end == 0) to mean "this is not a write-a-range request". To make all this sane, the patch changes range of writeback_control. So caller does: If it is calling ->writepages() to write pages, it sets range (range_start/end or range_cyclic) always. And if range_cyclic is true, ->writepages() thinks the range is cyclic, otherwise it just uses range_start and range_end. This patch does, - Add LLONG_MAX, LLONG_MIN, ULLONG_MAX to include/linux/kernel.h -1 is usually ok for range_end (type is long long). But, if someone did, range_end += val; range_end is "val - 1" u64val = range_end >> bits; u64val is "~(0ULL)" or something, they are wrong. So, this adds LLONG_MAX to avoid nasty things, and uses LLONG_MAX for range_end. - All callers of ->writepages() sets range_start/end or range_cyclic. - Fix updates of ->writeback_index. It seems already bit strange. If it starts at 0 and ended by check of nr_to_write, this last index may reduce chance to scan end of file. So, this updates ->writeback_index only if range_cyclic is true or whole-file is scanned. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Nathan Scott <nathans@sgi.com> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Steven French <sfrench@us.ibm.com> Cc: "Vladimir V. Saveliev" <vs@namesys.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 03:03:26 -06:00
if (!wbc->range_cyclic && page->index > end) {
done = 1;
unlock_page(page);
break;
}
if (next && (page->index != next)) {
/* Not next consecutive page */
unlock_page(page);
break;
}
if (wbc->sync_mode != WB_SYNC_NONE)
wait_on_page_writeback(page);
if (PageWriteback(page) ||
!clear_page_dirty_for_io(page)) {
unlock_page(page);
break;
}
/*
* This actually clears the dirty bit in the radix tree.
* See cifs_writepage() for more commentary.
*/
set_page_writeback(page);
if (page_offset(page) >= mapping->host->i_size) {
done = 1;
unlock_page(page);
end_page_writeback(page);
break;
}
/*
* BB can we get rid of this? pages are held by pvec
*/
page_cache_get(page);
len = min(mapping->host->i_size - page_offset(page),
(loff_t)PAGE_CACHE_SIZE);
/* reserve iov[0] for the smb header */
n_iov++;
iov[n_iov].iov_base = kmap(page);
iov[n_iov].iov_len = len;
bytes_to_write += len;
if (first < 0) {
first = i;
offset = page_offset(page);
}
next = page->index + 1;
if (bytes_to_write + PAGE_CACHE_SIZE > cifs_sb->wsize)
break;
}
if (n_iov) {
open_file = find_writable_file(CIFS_I(mapping->host),
false);
if (!open_file) {
cERROR(1, "No writable handles for inode");
rc = -EBADF;
} else {
long_op = cifs_write_timeout(cifsi, offset);
rc = CIFSSMBWrite2(xid, tcon, open_file->netfid,
bytes_to_write, offset,
&bytes_written, iov, n_iov,
long_op);
cifsFileInfo_put(open_file);
cifs_update_eof(cifsi, offset, bytes_written);
}
if (rc || bytes_written < bytes_to_write) {
cERROR(1, "Write2 ret %d, wrote %d",
rc, bytes_written);
/* BB what if continued retry is
requested via mount flags? */
if (rc == -ENOSPC)
set_bit(AS_ENOSPC, &mapping->flags);
else
set_bit(AS_EIO, &mapping->flags);
} else {
cifs_stats_bytes_written(tcon, bytes_written);
}
for (i = 0; i < n_iov; i++) {
page = pvec.pages[first + i];
/* Should we also set page error on
success rc but too little data written? */
/* BB investigate retry logic on temporary
server crash cases and how recovery works
when page marked as error */
if (rc)
SetPageError(page);
kunmap(page);
unlock_page(page);
end_page_writeback(page);
page_cache_release(page);
}
if ((wbc->nr_to_write -= n_iov) <= 0)
done = 1;
index = next;
} else
/* Need to re-find the pages we skipped */
index = pvec.pages[0]->index + 1;
pagevec_release(&pvec);
}
if (!scanned && !done) {
/*
* We hit the last page and there is more work to be done: wrap
* back to the start of the file
*/
scanned = 1;
index = 0;
goto retry;
}
[PATCH] writeback: fix range handling When a writeback_control's `start' and `end' fields are used to indicate a one-byte-range starting at file offset zero, the required values of .start=0,.end=0 mean that the ->writepages() implementation has no way of telling that it is being asked to perform a range request. Because we're currently overloading (start == 0 && end == 0) to mean "this is not a write-a-range request". To make all this sane, the patch changes range of writeback_control. So caller does: If it is calling ->writepages() to write pages, it sets range (range_start/end or range_cyclic) always. And if range_cyclic is true, ->writepages() thinks the range is cyclic, otherwise it just uses range_start and range_end. This patch does, - Add LLONG_MAX, LLONG_MIN, ULLONG_MAX to include/linux/kernel.h -1 is usually ok for range_end (type is long long). But, if someone did, range_end += val; range_end is "val - 1" u64val = range_end >> bits; u64val is "~(0ULL)" or something, they are wrong. So, this adds LLONG_MAX to avoid nasty things, and uses LLONG_MAX for range_end. - All callers of ->writepages() sets range_start/end or range_cyclic. - Fix updates of ->writeback_index. It seems already bit strange. If it starts at 0 and ended by check of nr_to_write, this last index may reduce chance to scan end of file. So, this updates ->writeback_index only if range_cyclic is true or whole-file is scanned. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Cc: Nathan Scott <nathans@sgi.com> Cc: Anton Altaparmakov <aia21@cantab.net> Cc: Steven French <sfrench@us.ibm.com> Cc: "Vladimir V. Saveliev" <vs@namesys.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 03:03:26 -06:00
if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
mapping->writeback_index = index;
FreeXid(xid);
kfree(iov);
return rc;
}
static int cifs_writepage(struct page *page, struct writeback_control *wbc)
{
int rc = -EFAULT;
int xid;
xid = GetXid();
/* BB add check for wbc flags */
page_cache_get(page);
if (!PageUptodate(page))
cFYI(1, "ppw - page not up to date");
/*
* Set the "writeback" flag, and clear "dirty" in the radix tree.
*
* A writepage() implementation always needs to do either this,
* or re-dirty the page with "redirty_page_for_writepage()" in
* the case of a failure.
*
* Just unlocking the page will cause the radix tree tag-bits
* to fail to update with the state of the page correctly.
*/
set_page_writeback(page);
rc = cifs_partialpagewrite(page, 0, PAGE_CACHE_SIZE);
SetPageUptodate(page); /* BB add check for error and Clearuptodate? */
unlock_page(page);
end_page_writeback(page);
page_cache_release(page);
FreeXid(xid);
return rc;
}
static int cifs_write_end(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata)
{
int rc;
struct inode *inode = mapping->host;
cFYI(1, "write_end for page %p from pos %lld with %d bytes",
page, pos, copied);
if (PageChecked(page)) {
if (copied == len)
SetPageUptodate(page);
ClearPageChecked(page);
} else if (!PageUptodate(page) && copied == PAGE_CACHE_SIZE)
SetPageUptodate(page);
if (!PageUptodate(page)) {
char *page_data;
unsigned offset = pos & (PAGE_CACHE_SIZE - 1);
int xid;
xid = GetXid();
/* this is probably better than directly calling
partialpage_write since in this function the file handle is
known which we might as well leverage */
/* BB check if anything else missing out of ppw
such as updating last write time */
page_data = kmap(page);
rc = cifs_write(file->private_data, page_data + offset,
copied, &pos);
/* if (rc < 0) should we set writebehind rc? */
kunmap(page);
FreeXid(xid);
} else {
rc = copied;
pos += copied;
set_page_dirty(page);
}
if (rc > 0) {
spin_lock(&inode->i_lock);
if (pos > inode->i_size)
i_size_write(inode, pos);
spin_unlock(&inode->i_lock);
}
unlock_page(page);
page_cache_release(page);
return rc;
}
int cifs_fsync(struct file *file, int datasync)
{
int xid;
int rc = 0;
struct cifsTconInfo *tcon;
struct cifsFileInfo *smbfile = file->private_data;
struct inode *inode = file->f_path.dentry->d_inode;
xid = GetXid();
cFYI(1, "Sync file - name: %s datasync: 0x%x",
file->f_path.dentry->d_name.name, datasync);
[CIFS] Fix potential data corruption when writing out cached dirty pages Fix RedHat bug 329431 The idea here is separate "conscious" from "unconscious" flushes. Conscious flushes are those due to a fsync() or close(). Unconscious ones are flushes that occur as a side effect of some other operation or due to memory pressure. Currently, when an error occurs during an unconscious flush (ENOSPC or EIO), we toss out the page and don't preserve that error to report to the user when a conscious flush occurs. If after the unconscious flush, there are no more dirty pages for the inode, the conscious flush will simply return success even though there were previous errors when writing out pages. This can lead to data corruption. The easiest way to reproduce this is to mount up a CIFS share that's very close to being full or where the user is very close to quota. mv a file to the share that's slightly larger than the quota allows. The writes will all succeed (since they go to pagecache). The mv will do a setattr to set the new file's attributes. This calls filemap_write_and_wait, which will return an error since all of the pages can't be written out. Then later, when the flush and release ops occur, there are no more dirty pages in pagecache for the file and those operations return 0. mv then assumes that the file was written out correctly and deletes the original. CIFS already has a write_behind_rc variable where it stores the results from earlier flushes, but that value is only reported in cifs_close. Since the VFS ignores the return value from the release operation, this isn't helpful. We should be reporting this error during the flush operation. This patch does the following: 1) changes cifs_fsync to use filemap_write_and_wait and cifs_flush and also sync to check its return code. If it returns successful, they then check the value of write_behind_rc to see if an earlier flush had reported any errors. If so, they return that error and clear write_behind_rc. 2) sets write_behind_rc in a few other places where pages are written out as a side effect of other operations and the code waits on them. 3) changes cifs_setattr to only call filemap_write_and_wait for ATTR_SIZE changes. 4) makes cifs_writepages accurately distinguish between EIO and ENOSPC errors when writing out pages. Some simple testing indicates that the patch works as expected and that it fixes the reproduceable known problem. Acked-by: Dave Kleikamp <shaggy@austin.rr.com> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2007-11-20 16:19:03 -07:00
rc = filemap_write_and_wait(inode->i_mapping);
if (rc == 0) {
rc = CIFS_I(inode)->write_behind_rc;
CIFS_I(inode)->write_behind_rc = 0;
tcon = tlink_tcon(smbfile->tlink);
if (!rc && tcon && smbfile &&
!(CIFS_SB(inode->i_sb)->mnt_cifs_flags & CIFS_MOUNT_NOSSYNC))
rc = CIFSSMBFlush(xid, tcon, smbfile->netfid);
[CIFS] Fix potential data corruption when writing out cached dirty pages Fix RedHat bug 329431 The idea here is separate "conscious" from "unconscious" flushes. Conscious flushes are those due to a fsync() or close(). Unconscious ones are flushes that occur as a side effect of some other operation or due to memory pressure. Currently, when an error occurs during an unconscious flush (ENOSPC or EIO), we toss out the page and don't preserve that error to report to the user when a conscious flush occurs. If after the unconscious flush, there are no more dirty pages for the inode, the conscious flush will simply return success even though there were previous errors when writing out pages. This can lead to data corruption. The easiest way to reproduce this is to mount up a CIFS share that's very close to being full or where the user is very close to quota. mv a file to the share that's slightly larger than the quota allows. The writes will all succeed (since they go to pagecache). The mv will do a setattr to set the new file's attributes. This calls filemap_write_and_wait, which will return an error since all of the pages can't be written out. Then later, when the flush and release ops occur, there are no more dirty pages in pagecache for the file and those operations return 0. mv then assumes that the file was written out correctly and deletes the original. CIFS already has a write_behind_rc variable where it stores the results from earlier flushes, but that value is only reported in cifs_close. Since the VFS ignores the return value from the release operation, this isn't helpful. We should be reporting this error during the flush operation. This patch does the following: 1) changes cifs_fsync to use filemap_write_and_wait and cifs_flush and also sync to check its return code. If it returns successful, they then check the value of write_behind_rc to see if an earlier flush had reported any errors. If so, they return that error and clear write_behind_rc. 2) sets write_behind_rc in a few other places where pages are written out as a side effect of other operations and the code waits on them. 3) changes cifs_setattr to only call filemap_write_and_wait for ATTR_SIZE changes. 4) makes cifs_writepages accurately distinguish between EIO and ENOSPC errors when writing out pages. Some simple testing indicates that the patch works as expected and that it fixes the reproduceable known problem. Acked-by: Dave Kleikamp <shaggy@austin.rr.com> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2007-11-20 16:19:03 -07:00
}
FreeXid(xid);
return rc;
}
/* static void cifs_sync_page(struct page *page)
{
struct address_space *mapping;
struct inode *inode;
unsigned long index = page->index;
unsigned int rpages = 0;
int rc = 0;
cFYI(1, "sync page %p", page);
mapping = page->mapping;
if (!mapping)
return 0;
inode = mapping->host;
if (!inode)
return; */
/* fill in rpages then
result = cifs_pagein_inode(inode, index, rpages); */ /* BB finish */
/* cFYI(1, "rpages is %d for sync page of Index %ld", rpages, index);
#if 0
if (rc < 0)
return rc;
return 0;
#endif
} */
/*
* As file closes, flush all cached write data for this inode checking
* for write behind errors.
*/
int cifs_flush(struct file *file, fl_owner_t id)
{
struct inode *inode = file->f_path.dentry->d_inode;
int rc = 0;
/* Rather than do the steps manually:
lock the inode for writing
loop through pages looking for write behind data (dirty pages)
coalesce into contiguous 16K (or smaller) chunks to write to server
send to server (prefer in parallel)
deal with writebehind errors
unlock inode for writing
filemapfdatawrite appears easier for the time being */
rc = filemap_fdatawrite(inode->i_mapping);
[CIFS] Fix potential data corruption when writing out cached dirty pages Fix RedHat bug 329431 The idea here is separate "conscious" from "unconscious" flushes. Conscious flushes are those due to a fsync() or close(). Unconscious ones are flushes that occur as a side effect of some other operation or due to memory pressure. Currently, when an error occurs during an unconscious flush (ENOSPC or EIO), we toss out the page and don't preserve that error to report to the user when a conscious flush occurs. If after the unconscious flush, there are no more dirty pages for the inode, the conscious flush will simply return success even though there were previous errors when writing out pages. This can lead to data corruption. The easiest way to reproduce this is to mount up a CIFS share that's very close to being full or where the user is very close to quota. mv a file to the share that's slightly larger than the quota allows. The writes will all succeed (since they go to pagecache). The mv will do a setattr to set the new file's attributes. This calls filemap_write_and_wait, which will return an error since all of the pages can't be written out. Then later, when the flush and release ops occur, there are no more dirty pages in pagecache for the file and those operations return 0. mv then assumes that the file was written out correctly and deletes the original. CIFS already has a write_behind_rc variable where it stores the results from earlier flushes, but that value is only reported in cifs_close. Since the VFS ignores the return value from the release operation, this isn't helpful. We should be reporting this error during the flush operation. This patch does the following: 1) changes cifs_fsync to use filemap_write_and_wait and cifs_flush and also sync to check its return code. If it returns successful, they then check the value of write_behind_rc to see if an earlier flush had reported any errors. If so, they return that error and clear write_behind_rc. 2) sets write_behind_rc in a few other places where pages are written out as a side effect of other operations and the code waits on them. 3) changes cifs_setattr to only call filemap_write_and_wait for ATTR_SIZE changes. 4) makes cifs_writepages accurately distinguish between EIO and ENOSPC errors when writing out pages. Some simple testing indicates that the patch works as expected and that it fixes the reproduceable known problem. Acked-by: Dave Kleikamp <shaggy@austin.rr.com> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2007-11-20 16:19:03 -07:00
/* reset wb rc if we were able to write out dirty pages */
if (!rc) {
rc = CIFS_I(inode)->write_behind_rc;
CIFS_I(inode)->write_behind_rc = 0;
[CIFS] Fix potential data corruption when writing out cached dirty pages Fix RedHat bug 329431 The idea here is separate "conscious" from "unconscious" flushes. Conscious flushes are those due to a fsync() or close(). Unconscious ones are flushes that occur as a side effect of some other operation or due to memory pressure. Currently, when an error occurs during an unconscious flush (ENOSPC or EIO), we toss out the page and don't preserve that error to report to the user when a conscious flush occurs. If after the unconscious flush, there are no more dirty pages for the inode, the conscious flush will simply return success even though there were previous errors when writing out pages. This can lead to data corruption. The easiest way to reproduce this is to mount up a CIFS share that's very close to being full or where the user is very close to quota. mv a file to the share that's slightly larger than the quota allows. The writes will all succeed (since they go to pagecache). The mv will do a setattr to set the new file's attributes. This calls filemap_write_and_wait, which will return an error since all of the pages can't be written out. Then later, when the flush and release ops occur, there are no more dirty pages in pagecache for the file and those operations return 0. mv then assumes that the file was written out correctly and deletes the original. CIFS already has a write_behind_rc variable where it stores the results from earlier flushes, but that value is only reported in cifs_close. Since the VFS ignores the return value from the release operation, this isn't helpful. We should be reporting this error during the flush operation. This patch does the following: 1) changes cifs_fsync to use filemap_write_and_wait and cifs_flush and also sync to check its return code. If it returns successful, they then check the value of write_behind_rc to see if an earlier flush had reported any errors. If so, they return that error and clear write_behind_rc. 2) sets write_behind_rc in a few other places where pages are written out as a side effect of other operations and the code waits on them. 3) changes cifs_setattr to only call filemap_write_and_wait for ATTR_SIZE changes. 4) makes cifs_writepages accurately distinguish between EIO and ENOSPC errors when writing out pages. Some simple testing indicates that the patch works as expected and that it fixes the reproduceable known problem. Acked-by: Dave Kleikamp <shaggy@austin.rr.com> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
2007-11-20 16:19:03 -07:00
}
cFYI(1, "Flush inode %p file %p rc %d", inode, file, rc);
return rc;
}
ssize_t cifs_user_read(struct file *file, char __user *read_data,
size_t read_size, loff_t *poffset)
{
int rc = -EACCES;
unsigned int bytes_read = 0;
unsigned int total_read = 0;
unsigned int current_read_size;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *pTcon;
int xid;
struct cifsFileInfo *open_file;
char *smb_read_data;
char __user *current_offset;
struct smb_com_read_rsp *pSMBr;
xid = GetXid();
cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
if (file->private_data == NULL) {
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
rc = -EBADF;
FreeXid(xid);
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
return rc;
}
open_file = file->private_data;
pTcon = tlink_tcon(open_file->tlink);
if ((file->f_flags & O_ACCMODE) == O_WRONLY)
cFYI(1, "attempting read on write only file instance");
for (total_read = 0, current_offset = read_data;
read_size > total_read;
total_read += bytes_read, current_offset += bytes_read) {
current_read_size = min_t(const int, read_size - total_read,
cifs_sb->rsize);
rc = -EAGAIN;
smb_read_data = NULL;
while (rc == -EAGAIN) {
int buf_type = CIFS_NO_BUFFER;
if (open_file->invalidHandle) {
rc = cifs_reopen_file(open_file, true);
if (rc != 0)
break;
}
rc = CIFSSMBRead(xid, pTcon,
open_file->netfid,
current_read_size, *poffset,
&bytes_read, &smb_read_data,
&buf_type);
pSMBr = (struct smb_com_read_rsp *)smb_read_data;
if (smb_read_data) {
if (copy_to_user(current_offset,
smb_read_data +
4 /* RFC1001 length field */ +
le16_to_cpu(pSMBr->DataOffset),
bytes_read))
rc = -EFAULT;
if (buf_type == CIFS_SMALL_BUFFER)
cifs_small_buf_release(smb_read_data);
else if (buf_type == CIFS_LARGE_BUFFER)
cifs_buf_release(smb_read_data);
smb_read_data = NULL;
}
}
if (rc || (bytes_read == 0)) {
if (total_read) {
break;
} else {
FreeXid(xid);
return rc;
}
} else {
cifs_stats_bytes_read(pTcon, bytes_read);
*poffset += bytes_read;
}
}
FreeXid(xid);
return total_read;
}
static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size,
loff_t *poffset)
{
int rc = -EACCES;
unsigned int bytes_read = 0;
unsigned int total_read;
unsigned int current_read_size;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *pTcon;
int xid;
char *current_offset;
struct cifsFileInfo *open_file;
int buf_type = CIFS_NO_BUFFER;
xid = GetXid();
cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
if (file->private_data == NULL) {
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
rc = -EBADF;
FreeXid(xid);
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
return rc;
}
open_file = file->private_data;
pTcon = tlink_tcon(open_file->tlink);
if ((file->f_flags & O_ACCMODE) == O_WRONLY)
cFYI(1, "attempting read on write only file instance");
for (total_read = 0, current_offset = read_data;
read_size > total_read;
total_read += bytes_read, current_offset += bytes_read) {
current_read_size = min_t(const int, read_size - total_read,
cifs_sb->rsize);
/* For windows me and 9x we do not want to request more
than it negotiated since it will refuse the read then */
if ((pTcon->ses) &&
!(pTcon->ses->capabilities & CAP_LARGE_FILES)) {
current_read_size = min_t(const int, current_read_size,
pTcon->ses->server->maxBuf - 128);
}
rc = -EAGAIN;
while (rc == -EAGAIN) {
if (open_file->invalidHandle) {
rc = cifs_reopen_file(open_file, true);
if (rc != 0)
break;
}
rc = CIFSSMBRead(xid, pTcon,
open_file->netfid,
current_read_size, *poffset,
&bytes_read, &current_offset,
&buf_type);
}
if (rc || (bytes_read == 0)) {
if (total_read) {
break;
} else {
FreeXid(xid);
return rc;
}
} else {
cifs_stats_bytes_read(pTcon, total_read);
*poffset += bytes_read;
}
}
FreeXid(xid);
return total_read;
}
int cifs_file_mmap(struct file *file, struct vm_area_struct *vma)
{
int rc, xid;
xid = GetXid();
rc = cifs_revalidate_file(file);
if (rc) {
cFYI(1, "Validation prior to mmap failed, error=%d", rc);
FreeXid(xid);
return rc;
}
rc = generic_file_mmap(file, vma);
FreeXid(xid);
return rc;
}
static void cifs_copy_cache_pages(struct address_space *mapping,
struct list_head *pages, int bytes_read, char *data)
{
struct page *page;
char *target;
while (bytes_read > 0) {
if (list_empty(pages))
break;
page = list_entry(pages->prev, struct page, lru);
list_del(&page->lru);
if (add_to_page_cache_lru(page, mapping, page->index,
GFP_KERNEL)) {
page_cache_release(page);
cFYI(1, "Add page cache failed");
data += PAGE_CACHE_SIZE;
bytes_read -= PAGE_CACHE_SIZE;
continue;
}
page_cache_release(page);
target = kmap_atomic(page, KM_USER0);
if (PAGE_CACHE_SIZE > bytes_read) {
memcpy(target, data, bytes_read);
/* zero the tail end of this partial page */
memset(target + bytes_read, 0,
PAGE_CACHE_SIZE - bytes_read);
bytes_read = 0;
} else {
memcpy(target, data, PAGE_CACHE_SIZE);
bytes_read -= PAGE_CACHE_SIZE;
}
kunmap_atomic(target, KM_USER0);
flush_dcache_page(page);
SetPageUptodate(page);
unlock_page(page);
data += PAGE_CACHE_SIZE;
/* add page to FS-Cache */
cifs_readpage_to_fscache(mapping->host, page);
}
return;
}
static int cifs_readpages(struct file *file, struct address_space *mapping,
struct list_head *page_list, unsigned num_pages)
{
int rc = -EACCES;
int xid;
loff_t offset;
struct page *page;
struct cifs_sb_info *cifs_sb;
struct cifsTconInfo *pTcon;
unsigned int bytes_read = 0;
unsigned int read_size, i;
char *smb_read_data = NULL;
struct smb_com_read_rsp *pSMBr;
struct cifsFileInfo *open_file;
int buf_type = CIFS_NO_BUFFER;
xid = GetXid();
if (file->private_data == NULL) {
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
rc = -EBADF;
FreeXid(xid);
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
return rc;
}
open_file = file->private_data;
cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
pTcon = tlink_tcon(open_file->tlink);
/*
* Reads as many pages as possible from fscache. Returns -ENOBUFS
* immediately if the cookie is negative
*/
rc = cifs_readpages_from_fscache(mapping->host, mapping, page_list,
&num_pages);
if (rc == 0)
goto read_complete;
cFYI(DBG2, "rpages: num pages %d", num_pages);
for (i = 0; i < num_pages; ) {
unsigned contig_pages;
struct page *tmp_page;
unsigned long expected_index;
if (list_empty(page_list))
break;
page = list_entry(page_list->prev, struct page, lru);
offset = (loff_t)page->index << PAGE_CACHE_SHIFT;
/* count adjacent pages that we will read into */
contig_pages = 0;
expected_index =
list_entry(page_list->prev, struct page, lru)->index;
list_for_each_entry_reverse(tmp_page, page_list, lru) {
if (tmp_page->index == expected_index) {
contig_pages++;
expected_index++;
} else
break;
}
if (contig_pages + i > num_pages)
contig_pages = num_pages - i;
/* for reads over a certain size could initiate async
read ahead */
read_size = contig_pages * PAGE_CACHE_SIZE;
/* Read size needs to be in multiples of one page */
read_size = min_t(const unsigned int, read_size,
cifs_sb->rsize & PAGE_CACHE_MASK);
cFYI(DBG2, "rpages: read size 0x%x contiguous pages %d",
read_size, contig_pages);
rc = -EAGAIN;
while (rc == -EAGAIN) {
if (open_file->invalidHandle) {
rc = cifs_reopen_file(open_file, true);
if (rc != 0)
break;
}
rc = CIFSSMBRead(xid, pTcon,
open_file->netfid,
read_size, offset,
&bytes_read, &smb_read_data,
&buf_type);
/* BB more RC checks ? */
if (rc == -EAGAIN) {
if (smb_read_data) {
if (buf_type == CIFS_SMALL_BUFFER)
cifs_small_buf_release(smb_read_data);
else if (buf_type == CIFS_LARGE_BUFFER)
cifs_buf_release(smb_read_data);
smb_read_data = NULL;
}
}
}
if ((rc < 0) || (smb_read_data == NULL)) {
cFYI(1, "Read error in readpages: %d", rc);
break;
} else if (bytes_read > 0) {
task_io_account_read(bytes_read);
pSMBr = (struct smb_com_read_rsp *)smb_read_data;
cifs_copy_cache_pages(mapping, page_list, bytes_read,
smb_read_data + 4 /* RFC1001 hdr */ +
le16_to_cpu(pSMBr->DataOffset));
i += bytes_read >> PAGE_CACHE_SHIFT;
cifs_stats_bytes_read(pTcon, bytes_read);
if ((bytes_read & PAGE_CACHE_MASK) != bytes_read) {
i++; /* account for partial page */
/* server copy of file can have smaller size
than client */
/* BB do we need to verify this common case ?
this case is ok - if we are at server EOF
we will hit it on next read */
/* break; */
}
} else {
cFYI(1, "No bytes read (%d) at offset %lld . "
"Cleaning remaining pages from readahead list",
bytes_read, offset);
/* BB turn off caching and do new lookup on
file size at server? */
break;
}
if (smb_read_data) {
if (buf_type == CIFS_SMALL_BUFFER)
cifs_small_buf_release(smb_read_data);
else if (buf_type == CIFS_LARGE_BUFFER)
cifs_buf_release(smb_read_data);
smb_read_data = NULL;
}
bytes_read = 0;
}
/* need to free smb_read_data buf before exit */
if (smb_read_data) {
if (buf_type == CIFS_SMALL_BUFFER)
cifs_small_buf_release(smb_read_data);
else if (buf_type == CIFS_LARGE_BUFFER)
cifs_buf_release(smb_read_data);
smb_read_data = NULL;
}
read_complete:
FreeXid(xid);
return rc;
}
static int cifs_readpage_worker(struct file *file, struct page *page,
loff_t *poffset)
{
char *read_data;
int rc;
/* Is the page cached? */
rc = cifs_readpage_from_fscache(file->f_path.dentry->d_inode, page);
if (rc == 0)
goto read_complete;
page_cache_get(page);
read_data = kmap(page);
/* for reads over a certain size could initiate async read ahead */
rc = cifs_read(file, read_data, PAGE_CACHE_SIZE, poffset);
if (rc < 0)
goto io_error;
else
cFYI(1, "Bytes read %d", rc);
file->f_path.dentry->d_inode->i_atime =
current_fs_time(file->f_path.dentry->d_inode->i_sb);
if (PAGE_CACHE_SIZE > rc)
memset(read_data + rc, 0, PAGE_CACHE_SIZE - rc);
flush_dcache_page(page);
SetPageUptodate(page);
/* send this page to the cache */
cifs_readpage_to_fscache(file->f_path.dentry->d_inode, page);
rc = 0;
io_error:
kunmap(page);
page_cache_release(page);
read_complete:
return rc;
}
static int cifs_readpage(struct file *file, struct page *page)
{
loff_t offset = (loff_t)page->index << PAGE_CACHE_SHIFT;
int rc = -EACCES;
int xid;
xid = GetXid();
if (file->private_data == NULL) {
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
rc = -EBADF;
FreeXid(xid);
cifs: Fix incorrect return code being printed in cFYI messages FreeXid() along with freeing Xid does add a cifsFYI debug message that prints rc (return code) as well. In some code paths where we set/return error code after calling FreeXid(), incorrect error code is being printed when cifsFYI is enabled. This could be misleading in few cases. For eg. In cifs_open() if cifs_fill_filedata() returns a valid pointer to cifsFileInfo, FreeXid() prints rc=-13 whereas 0 is actually being returned. Fix this by setting rc before calling FreeXid(). Basically convert FreeXid(xid); rc = -ERR; return -ERR; => FreeXid(xid); return rc; [Note that Christoph would like to replace the GetXid/FreeXid calls, which are primarily used for debugging. This seems like a good longer term goal, but although there is an alternative tracing facility, there are no examples yet available that I know of that we can use (yet) to convert this cifs function entry/exit logging, and for creating an identifier that we can use to correlate all dmesg log entries for a particular vfs operation (ie identify all log entries for a particular vfs request to cifs: e.g. a particular close or read or write or byte range lock call ... and just using the thread id is harder). Eventually when a replacement for this is available (e.g. when NFS switches over and various samples to look at in other file systems) we can remove the GetXid/FreeXid macro but in the meantime multiple people use this run time configurable logging all the time for debugging, and Suresh's patch fixes a problem which made it harder to notice some low memory problems in the log so it is worthwhile to fix this problem until a better logging approach is able to be used] Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de> Signed-off-by: Steve French <sfrench@us.ibm.com>
2009-06-25 06:42:34 -06:00
return rc;
}
cFYI(1, "readpage %p at offset %d 0x%x\n",
page, (int)offset, (int)offset);
rc = cifs_readpage_worker(file, page, &offset);
unlock_page(page);
FreeXid(xid);
return rc;
}
static int is_inode_writable(struct cifsInodeInfo *cifs_inode)
{
struct cifsFileInfo *open_file;
spin_lock(&cifs_file_list_lock);
list_for_each_entry(open_file, &cifs_inode->openFileList, flist) {
if (OPEN_FMODE(open_file->f_flags) & FMODE_WRITE) {
spin_unlock(&cifs_file_list_lock);
return 1;
}
}
spin_unlock(&cifs_file_list_lock);
return 0;
}
/* We do not want to update the file size from server for inodes
open for write - to avoid races with writepage extending
the file - in the future we could consider allowing
refreshing the inode only on increases in the file size
but this is tricky to do without racing with writebehind
page caching in the current Linux kernel design */
bool is_size_safe_to_change(struct cifsInodeInfo *cifsInode, __u64 end_of_file)
{
if (!cifsInode)
return true;
if (is_inode_writable(cifsInode)) {
/* This inode is open for write at least once */
struct cifs_sb_info *cifs_sb;
cifs_sb = CIFS_SB(cifsInode->vfs_inode.i_sb);
if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DIRECT_IO) {
/* since no page cache to corrupt on directio
we can change size safely */
return true;
}
if (i_size_read(&cifsInode->vfs_inode) < end_of_file)
return true;
return false;
} else
return true;
}
static int cifs_write_begin(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **pagep, void **fsdata)
{
pgoff_t index = pos >> PAGE_CACHE_SHIFT;
loff_t offset = pos & (PAGE_CACHE_SIZE - 1);
loff_t page_start = pos & PAGE_MASK;
loff_t i_size;
struct page *page;
int rc = 0;
cFYI(1, "write_begin from %lld len %d", (long long)pos, len);
fs: symlink write_begin allocation context fix With the write_begin/write_end aops, page_symlink was broken because it could no longer pass a GFP_NOFS type mask into the point where the allocations happened. They are done in write_begin, which would always assume that the filesystem can be entered from reclaim. This bug could cause filesystem deadlocks. The funny thing with having a gfp_t mask there is that it doesn't really allow the caller to arbitrarily tinker with the context in which it can be called. It couldn't ever be GFP_ATOMIC, for example, because it needs to take the page lock. The only thing any callers care about is __GFP_FS anyway, so turn that into a single flag. Add a new flag for write_begin, AOP_FLAG_NOFS. Filesystems can now act on this flag in their write_begin function. Change __grab_cache_page to accept a nofs argument as well, to honour that flag (while we're there, change the name to grab_cache_page_write_begin which is more instructive and does away with random leading underscores). This is really a more flexible way to go in the end anyway -- if a filesystem happens to want any extra allocations aside from the pagecache ones in ints write_begin function, it may now use GFP_KERNEL (rather than GFP_NOFS) for common case allocations (eg. ocfs2_alloc_write_ctxt, for a random example). [kosaki.motohiro@jp.fujitsu.com: fix ubifs] [kosaki.motohiro@jp.fujitsu.com: fix fuse] Signed-off-by: Nick Piggin <npiggin@suse.de> Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: <stable@kernel.org> [2.6.28.x] Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> [ Cleaned up the calling convention: just pass in the AOP flags untouched to the grab_cache_page_write_begin() function. That just simplifies everybody, and may even allow future expansion of the logic. - Linus ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-04 13:00:53 -07:00
page = grab_cache_page_write_begin(mapping, index, flags);
if (!page) {
rc = -ENOMEM;
goto out;
}
if (PageUptodate(page))
goto out;
/*
* If we write a full page it will be up to date, no need to read from
* the server. If the write is short, we'll end up doing a sync write
* instead.
*/
if (len == PAGE_CACHE_SIZE)
goto out;
/*
* optimize away the read when we have an oplock, and we're not
* expecting to use any of the data we'd be reading in. That
* is, when the page lies beyond the EOF, or straddles the EOF
* and the write will cover all of the existing data.
*/
if (CIFS_I(mapping->host)->clientCanCacheRead) {
i_size = i_size_read(mapping->host);
if (page_start >= i_size ||
(offset == 0 && (pos + len) >= i_size)) {
zero_user_segments(page, 0, offset,
offset + len,
PAGE_CACHE_SIZE);
/*
* PageChecked means that the parts of the page
* to which we're not writing are considered up
* to date. Once the data is copied to the
* page, it can be set uptodate.
*/
SetPageChecked(page);
goto out;
}
}
if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
/*
* might as well read a page, it is fast enough. If we get
* an error, we don't need to return it. cifs_write_end will
* do a sync write instead since PG_uptodate isn't set.
*/
cifs_readpage_worker(file, page, &page_start);
} else {
/* we could try using another file handle if there is one -
but how would we lock it to prevent close of that handle
racing with this read? In any case
this will be written out by write_end so is fine */
}
out:
*pagep = page;
return rc;
}
static int cifs_release_page(struct page *page, gfp_t gfp)
{
if (PagePrivate(page))
return 0;
return cifs_fscache_release_page(page, gfp);
}
static void cifs_invalidate_page(struct page *page, unsigned long offset)
{
struct cifsInodeInfo *cifsi = CIFS_I(page->mapping->host);
if (offset == 0)
cifs_fscache_invalidate_page(page, &cifsi->vfs_inode);
}
void cifs_oplock_break(struct work_struct *work)
{
struct cifsFileInfo *cfile = container_of(work, struct cifsFileInfo,
oplock_break);
struct inode *inode = cfile->dentry->d_inode;
struct cifsInodeInfo *cinode = CIFS_I(inode);
int rc, waitrc = 0;
if (inode && S_ISREG(inode->i_mode)) {
if (cinode->clientCanCacheRead)
break_lease(inode, O_RDONLY);
else
break_lease(inode, O_WRONLY);
rc = filemap_fdatawrite(inode->i_mapping);
if (cinode->clientCanCacheRead == 0) {
waitrc = filemap_fdatawait(inode->i_mapping);
invalidate_remote_inode(inode);
}
if (!rc)
rc = waitrc;
if (rc)
cinode->write_behind_rc = rc;
cFYI(1, "Oplock flush inode %p rc %d", inode, rc);
}
/*
* releasing stale oplock after recent reconnect of smb session using
* a now incorrect file handle is not a data integrity issue but do
* not bother sending an oplock release if session to server still is
* disconnected since oplock already released by the server
*/
if (!cfile->oplock_break_cancelled) {
rc = CIFSSMBLock(0, tlink_tcon(cfile->tlink), cfile->netfid, 0,
0, 0, 0, LOCKING_ANDX_OPLOCK_RELEASE, false);
cFYI(1, "Oplock release rc = %d", rc);
}
/*
* We might have kicked in before is_valid_oplock_break()
* finished grabbing reference for us. Make sure it's done by
* waiting for GlobalSMSSeslock.
*/
spin_lock(&cifs_file_list_lock);
spin_unlock(&cifs_file_list_lock);
cifs_oplock_break_put(cfile);
}
void cifs_oplock_break_get(struct cifsFileInfo *cfile)
{
cifs_sb_active(cfile->dentry->d_sb);
cifsFileInfo_get(cfile);
}
void cifs_oplock_break_put(struct cifsFileInfo *cfile)
{
cifsFileInfo_put(cfile);
cifs_sb_deactive(cfile->dentry->d_sb);
}
const struct address_space_operations cifs_addr_ops = {
.readpage = cifs_readpage,
.readpages = cifs_readpages,
.writepage = cifs_writepage,
.writepages = cifs_writepages,
.write_begin = cifs_write_begin,
.write_end = cifs_write_end,
.set_page_dirty = __set_page_dirty_nobuffers,
.releasepage = cifs_release_page,
.invalidatepage = cifs_invalidate_page,
/* .sync_page = cifs_sync_page, */
/* .direct_IO = */
};
/*
* cifs_readpages requires the server to support a buffer large enough to
* contain the header plus one complete page of data. Otherwise, we need
* to leave cifs_readpages out of the address space operations.
*/
const struct address_space_operations cifs_addr_ops_smallbuf = {
.readpage = cifs_readpage,
.writepage = cifs_writepage,
.writepages = cifs_writepages,
.write_begin = cifs_write_begin,
.write_end = cifs_write_end,
.set_page_dirty = __set_page_dirty_nobuffers,
.releasepage = cifs_release_page,
.invalidatepage = cifs_invalidate_page,
/* .sync_page = cifs_sync_page, */
/* .direct_IO = */
};