Skip to content

Commit

Permalink
fs/9p: remove writeback fid and fix per-file modes
Browse files Browse the repository at this point in the history 
This patch removes the creating of an additional writeback_fid
for opened files.  The patch addresses problems when files
were opened write-only or getattr on files with dirty caches.

This patch also incorporates information about cache behavior
in the fid for every file.  This allows us to reflect cache
behavior from mount flags, open mode, and information from
the server to inform readahead and writeback behavior.

This includes adding support for a 9p semantic that qid.version==0
is used to mark a file as non-cachable which is important for
synthetic files.  This may have a side-effect of not supporting
caching on certain legacy file servers that do not properly set
qid.version.  There is also now a mount flag which can disable
the qid.version behavior.

Signed-off-by: Eric Van Hensbergen <ericvh@kernel.org>
  • Loading branch information
@ericvh
ericvh committed Mar 27, 2023
1 parent 6deffc8 commit 1543b4c
Show file tree
Hide file tree
Showing 8 changed files with 135 additions and 186 deletions.
48 changes: 19 additions & 29 deletions fs/9p/fid.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -41,14 +41,24 @@ void v9fs_fid_add(struct dentry *dentry, struct p9_fid **pfid)
*pfid = NULL;
}

static bool v9fs_is_writeable(int mode)
{
if (mode & (P9_OWRITE|P9_ORDWR))
return true;
else
return false;
}

/**
* v9fs_fid_find_inode - search for an open fid off of the inode list
* @inode: return a fid pointing to a specific inode
* @want_writeable: only consider fids which are writeable
* @uid: return a fid belonging to the specified user
* @any: ignore uid as a selection criteria
*
*/

static struct p9_fid *v9fs_fid_find_inode(struct inode *inode, kuid_t uid)
struct p9_fid *v9fs_fid_find_inode(struct inode *inode, bool want_writeable,
kuid_t uid, bool any)
{
struct hlist_head *h;
struct p9_fid *fid, *ret = NULL;
Expand All @@ -58,7 +68,12 @@ static struct p9_fid *v9fs_fid_find_inode(struct inode *inode, kuid_t uid)
spin_lock(&inode->i_lock);
h = (struct hlist_head *)&inode->i_private;
hlist_for_each_entry(fid, h, ilist) {
if (uid_eq(fid->uid, uid)) {
if (any || uid_eq(fid->uid, uid)) {
if (want_writeable && !v9fs_is_writeable(fid->mode)) {
p9_debug(P9_DEBUG_VFS, " mode: %x not writeable?\n",
fid->mode);
continue;
}
p9_fid_get(fid);
ret = fid;
break;
Expand Down Expand Up @@ -118,7 +133,7 @@ static struct p9_fid *v9fs_fid_find(struct dentry *dentry, kuid_t uid, int any)
spin_unlock(&dentry->d_lock);
} else {
if (dentry->d_inode)
ret = v9fs_fid_find_inode(dentry->d_inode, uid);
ret = v9fs_fid_find_inode(dentry->d_inode, false, uid, any);
}

return ret;
Expand Down Expand Up @@ -299,28 +314,3 @@ struct p9_fid *v9fs_fid_lookup(struct dentry *dentry)
return v9fs_fid_lookup_with_uid(dentry, uid, any);
}

struct p9_fid *v9fs_writeback_fid(struct dentry *dentry)
{
int err;
struct p9_fid *fid, *ofid;

ofid = v9fs_fid_lookup_with_uid(dentry, GLOBAL_ROOT_UID, 0);
fid = clone_fid(ofid);
if (IS_ERR(fid))
goto error_out;
p9_fid_put(ofid);
/*
* writeback fid will only be used to write back the
* dirty pages. We always request for the open fid in read-write
* mode so that a partial page write which result in page
* read can work.
*/
err = p9_client_open(fid, O_RDWR);
if (err < 0) {
p9_fid_put(fid);
fid = ERR_PTR(err);
goto error_out;
}
error_out:
return fid;
}
33 changes: 32 additions & 1 deletion fs/9p/fid.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,14 +7,16 @@
#ifndef FS_9P_FID_H
#define FS_9P_FID_H
#include <linux/list.h>
#include "v9fs.h"

struct p9_fid *v9fs_fid_find_inode(struct inode *inode, bool want_writeable,
kuid_t uid, bool any);
struct p9_fid *v9fs_fid_lookup(struct dentry *dentry);
static inline struct p9_fid *v9fs_parent_fid(struct dentry *dentry)
{
return v9fs_fid_lookup(dentry->d_parent);
}
void v9fs_fid_add(struct dentry *dentry, struct p9_fid **fid);
struct p9_fid *v9fs_writeback_fid(struct dentry *dentry);
void v9fs_open_fid_add(struct inode *inode, struct p9_fid **fid);
static inline struct p9_fid *clone_fid(struct p9_fid *fid)
{
Expand All @@ -32,4 +34,33 @@ static inline struct p9_fid *v9fs_fid_clone(struct dentry *dentry)
p9_fid_put(fid);
return nfid;
}
/**
* v9fs_fid_addmodes - add cache flags to fid mode (for client use only)
* @fid: fid to augment
* @s_flags: session info mount flags
* @s_cache: session info cache flags
* @f_flags: unix open flags
*
* make sure mode reflects flags of underlying mounts
* also qid.version == 0 reflects a synthetic or legacy file system
* NOTE: these are set after open so only reflect 9p client not
* underlying file system on server.
*/
static inline void v9fs_fid_add_modes(struct p9_fid *fid, int s_flags,
int s_cache, unsigned int f_flags)
{
if (fid->qid.type != P9_QTFILE)
return;

if ((!s_cache) ||
((fid->qid.version == 0) && !(s_flags & V9FS_IGNORE_QV)) ||
(s_flags & V9FS_DIRECT_IO) || (f_flags & O_DIRECT)) {
fid->mode |= P9L_DIRECT; /* no read or write cache */
} else if ((s_cache < CACHE_WRITEBACK) ||
(f_flags & O_DSYNC) | (s_flags & V9FS_SYNC)) {
fid->mode |= P9L_NOWRITECACHE;
} else if (s_cache == CACHE_LOOSE) {
fid->mode |= P9L_LOOSE; /* noncoherent cache */
}
}
#endif
1 change: 0 additions & 1 deletion fs/9p/v9fs.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -118,7 +118,6 @@ struct v9fs_inode {
struct netfs_inode netfs; /* Netfslib context and vfs inode */
struct p9_qid qid;
unsigned int cache_validity;
struct p9_fid *writeback_fid;
struct mutex v_mutex;
};

Expand Down
22 changes: 12 additions & 10 deletions fs/9p/vfs_addr.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -57,20 +57,15 @@ static void v9fs_issue_read(struct netfs_io_subrequest *subreq)
*/
static int v9fs_init_request(struct netfs_io_request *rreq, struct file *file)
{
struct inode *inode = file_inode(file);
struct v9fs_inode *v9inode = V9FS_I(inode);
struct p9_fid *fid = file->private_data;

BUG_ON(!fid);

/* we might need to read from a fid that was opened write-only
* for read-modify-write of page cache, use the writeback fid
* for that */
if (rreq->origin == NETFS_READ_FOR_WRITE &&
(fid->mode & O_ACCMODE) == O_WRONLY) {
fid = v9inode->writeback_fid;
BUG_ON(!fid);
}
WARN_ON(rreq->origin == NETFS_READ_FOR_WRITE &&
!(fid->mode & P9_ORDWR));

p9_fid_get(fid);
rreq->netfs_priv = fid;
Expand Down Expand Up @@ -164,6 +159,7 @@ static int v9fs_vfs_write_folio_locked(struct folio *folio)
loff_t i_size = i_size_read(inode);
struct iov_iter from;
size_t len = folio_size(folio);
struct p9_fid *writeback_fid;
int err;

if (start >= i_size)
Expand All @@ -173,13 +169,17 @@ static int v9fs_vfs_write_folio_locked(struct folio *folio)

iov_iter_xarray(&from, ITER_SOURCE, &folio_mapping(folio)->i_pages, start, len);

/* We should have writeback_fid always set */
BUG_ON(!v9inode->writeback_fid);
writeback_fid = v9fs_fid_find_inode(inode, true, INVALID_UID, true);
if (!writeback_fid) {
WARN_ONCE(1, "folio expected an open fid inode->i_private=%p\n",
inode->i_private);
return -EINVAL;
}

folio_wait_fscache(folio);
folio_start_writeback(folio);

p9_client_write(v9inode->writeback_fid, start, &from, &err);
p9_client_write(writeback_fid, start, &from, &err);

if (err == 0 &&
fscache_cookie_enabled(cookie) &&
Expand All @@ -192,6 +192,8 @@ static int v9fs_vfs_write_folio_locked(struct folio *folio)
}

folio_end_writeback(folio);
p9_fid_put(writeback_fid);

return err;
}

Expand Down
91 changes: 31 additions & 60 deletions fs/9p/vfs_file.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,7 @@ int v9fs_file_open(struct inode *inode, struct file *file)
int err;
struct v9fs_inode *v9inode;
struct v9fs_session_info *v9ses;
struct p9_fid *fid, *writeback_fid;
struct p9_fid *fid;
int omode;

p9_debug(P9_DEBUG_VFS, "inode: %p file: %p\n", inode, file);
Expand All @@ -60,7 +60,19 @@ int v9fs_file_open(struct inode *inode, struct file *file)
if (IS_ERR(fid))
return PTR_ERR(fid);

err = p9_client_open(fid, omode);
if ((v9ses->cache >= CACHE_WRITEBACK) && (omode & P9_OWRITE)) {
int writeback_omode = (omode & ~P9_OWRITE) | P9_ORDWR;

p9_debug(P9_DEBUG_CACHE, "write-only file with writeback enabled, try opening O_RDWR\n");
err = p9_client_open(fid, writeback_omode);
if (err < 0) {
p9_debug(P9_DEBUG_CACHE, "could not open O_RDWR, disabling caches\n");
err = p9_client_open(fid, omode);
fid->mode |= P9L_DIRECT;
}
} else {
err = p9_client_open(fid, omode);
}
if (err < 0) {
p9_fid_put(fid);
return err;
Expand All @@ -72,36 +84,14 @@ int v9fs_file_open(struct inode *inode, struct file *file)
file->private_data = fid;
}

mutex_lock(&v9inode->v_mutex);
if ((v9ses->cache >= CACHE_WRITEBACK) && !v9inode->writeback_fid &&
((file->f_flags & O_ACCMODE) != O_RDONLY)) {
/*
* clone a fid and add it to writeback_fid
* we do it during open time instead of
* page dirty time via write_begin/page_mkwrite
* because we want write after unlink usecase
* to work.
*/
writeback_fid = v9fs_writeback_fid(file_dentry(file));
if (IS_ERR(writeback_fid)) {
err = PTR_ERR(writeback_fid);
mutex_unlock(&v9inode->v_mutex);
goto out_error;
}
v9inode->writeback_fid = (void *) writeback_fid;
}
mutex_unlock(&v9inode->v_mutex);
#ifdef CONFIG_9P_FSCACHE
if (v9ses->cache == CACHE_FSCACHE)
fscache_use_cookie(v9fs_inode_cookie(v9inode),
file->f_mode & FMODE_WRITE);
#endif
v9fs_fid_add_modes(fid, v9ses->flags, v9ses->cache, file->f_flags);
v9fs_open_fid_add(inode, &fid);
return 0;
out_error:
p9_fid_put(file->private_data);
file->private_data = NULL;
return err;
}

/**
Expand Down Expand Up @@ -367,14 +357,14 @@ v9fs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
{
struct p9_fid *fid = iocb->ki_filp->private_data;
int ret, err = 0;
struct inode *inode = file_inode(iocb->ki_filp);
struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode);

p9_debug(P9_DEBUG_VFS, "count %zu offset %lld\n",
iov_iter_count(to), iocb->ki_pos);
p9_debug(P9_DEBUG_VFS, "fid %d count %zu offset %lld\n",
fid->fid, iov_iter_count(to), iocb->ki_pos);

if (v9ses->cache > CACHE_MMAP)
if (!(fid->mode & P9L_DIRECT)) {
p9_debug(P9_DEBUG_VFS, "(cached)\n");
return generic_file_read_iter(iocb, to);
}

if (iocb->ki_filp->f_flags & O_NONBLOCK)
ret = p9_client_read_once(fid, iocb->ki_pos, to, &err);
Expand All @@ -397,14 +387,17 @@ static ssize_t
v9fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
{
struct file *file = iocb->ki_filp;
struct p9_fid *fid = file->private_data;
ssize_t retval;
loff_t origin;
int err = 0;
struct inode *inode = file_inode(iocb->ki_filp);
struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode);

if (v9ses->cache >= CACHE_WRITEBACK)
p9_debug(P9_DEBUG_VFS, "fid %d\n", fid->fid);

if (!(fid->mode & (P9L_DIRECT | P9L_NOWRITECACHE))) {
p9_debug(P9_DEBUG_CACHE, "(cached)\n");
return generic_file_write_iter(iocb, from);
}

retval = generic_write_checks(iocb, from);
if (retval <= 0)
Expand Down Expand Up @@ -488,36 +481,18 @@ v9fs_file_mmap(struct file *filp, struct vm_area_struct *vma)
{
int retval;
struct inode *inode = file_inode(filp);
struct v9fs_inode *v9inode = V9FS_I(inode);
struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode);
struct p9_fid *fid;

p9_debug(P9_DEBUG_MMAP, "filp :%p\n", filp);

if (v9ses->cache < CACHE_MMAP) {
p9_debug(P9_DEBUG_CACHE, "(no mmap mode)");
if (vma->vm_flags & VM_MAYSHARE)
return -ENODEV;
invalidate_inode_pages2(filp->f_mapping);
return generic_file_readonly_mmap(filp, vma);
}

mutex_lock(&v9inode->v_mutex);
if (!v9inode->writeback_fid &&
(vma->vm_flags & VM_SHARED) &&
(vma->vm_flags & VM_WRITE)) {
/*
* clone a fid and add it to writeback_fid
* we do it during mmap instead of
* page dirty time via write_begin/page_mkwrite
* because we want write after unlink usecase
* to work.
*/
fid = v9fs_writeback_fid(file_dentry(filp));
if (IS_ERR(fid)) {
retval = PTR_ERR(fid);
mutex_unlock(&v9inode->v_mutex);
return retval;
}
v9inode->writeback_fid = (void *) fid;
}
mutex_unlock(&v9inode->v_mutex);

retval = generic_file_mmap(filp, vma);
if (!retval)
vma->vm_ops = &v9fs_mmap_file_vm_ops;
Expand All @@ -528,7 +503,6 @@ v9fs_file_mmap(struct file *filp, struct vm_area_struct *vma)
static vm_fault_t
v9fs_vm_page_mkwrite(struct vm_fault *vmf)
{
struct v9fs_inode *v9inode;
struct folio *folio = page_folio(vmf->page);
struct file *filp = vmf->vma->vm_file;
struct inode *inode = file_inode(filp);
Expand All @@ -537,8 +511,6 @@ v9fs_vm_page_mkwrite(struct vm_fault *vmf)
p9_debug(P9_DEBUG_VFS, "folio %p fid %lx\n",
folio, (unsigned long)filp->private_data);

v9inode = V9FS_I(inode);

/* Wait for the page to be written to the cache before we allow it to
* be modified. We then assume the entire page will need writing back.
*/
Expand All @@ -551,7 +523,6 @@ v9fs_vm_page_mkwrite(struct vm_fault *vmf)
/* Update file times before taking page lock */
file_update_time(filp);

BUG_ON(!v9inode->writeback_fid);
if (folio_lock_killable(folio) < 0)
return VM_FAULT_RETRY;
if (folio_mapping(folio) != inode->i_mapping)
Expand Down
Loading

0 comments on commit 1543b4c

Please sign in to comment.