Merge the type-specific data with the payload data into one four-word chunk
as it seems pointless to keep them separate.
Use user_key_payload() for accessing the payloads of overloaded
user-defined keys.
Signed-off-by: David Howells
cc: linux-c...@vger.kernel.org
cc: ecryp...@vger.kernel.org
cc
Milosz Tanski wrote:
> FSCache already tracks statistics about how long writes and reads take
> (at least if you enable that option). With those stats in hand we
> should be able generate a default timeout value that works well and
> avoid a tunable.
If stat generation is going to be on all the
Milosz Tanski wrote:
> That's the same thing exact fix I started testing on Saturday. I found that
> there already is a wait_event_timeout (even without your recent changes). The
> thing I'm not quite sure is what timeout it should use?
That's probably something to make an external tuning knob f
Milosz Tanski wrote:
> The patch that you sent me in private last week solved this issue and
> I haven't experienced it. I haven't seen the patch go into your
> linus-fs.git tree, so I don't know if you submitted the fix upstream
> to the linux kernel. So I'm following up to see if you plan are
>
Milosz Tanski wrote:
> Stores : ops=10257 run=67477 pgs=57220 rxd=62216 olm=14
I think this line probably shows the problem. The olm=14 indicates that 14
pages were found over the store limit set on the object. Look in
fscache_write_op() for:
if (page->index > op->store_limit) {
Milosz Tanski wrote:
> In my test cluster I started seeing an issue when the fscache get
> stuck waiting on pending writes when doing page invalidate. The
> problem happens because it looks like the page never leaves the
> cookie->store page tree.
>
> I've been reading the different code paths i
o control whether to get an fscache cookie in
nfs_open() (2013-09-27 18:40:25 +0100)
(from the branch description for fscache local branch)
clone of "master"
----------
se no one else
can get at the cookie before this returns.
Signed-off-by: David Howells cache =
fscache_acquire_cookie(afs_cache_netfs.primary_index,
&afs_cell_cache_index_def,
-
's also possible to have a two-way race on this if the file is opened
O_TRUNC|O_RDONLY instead.
Reported-by: Mark Moseley
Signed-off-by: David Howells
---
fs/nfs/dir.c |2 -
fs/nfs/fscache.c | 192 ++--
fs/nfs/fscache.h |
that we don't have to deal with undoing it if anything
earlier fails. Possibly this should be moved into fscache_submit_op() which
could look at FSCACHE_OP_UNUSE_COOKIE.
Signed-off-by: David Howells
---
fs/fscache/cookie.c |7 +--
fs/fscache
s probably safe.
David
---
David Howells (3):
FS-Cache: Add use/unuse/wake cookie wrappers
FS-Cache: Provide the ability to enable/disable cookies
NFS: Use i_writecount to control whether to get an fscache cookie in
nfs_open()
Documentation/filesystems/caching/netfs-api.
->netfs_data pointers have been cleared - and a NULL pointer exception will
ensue, usually just after a:
CacheFiles: Error: Unexpected object collision
error is reported.
Signed-off-by: David Howells
---
fs/cachefiles/namei.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(
appropriately typed variable rather than to ret.
Signed-off-by: Josh Boyer
Signed-off-by: David Howells
cc: Hongyi Jia
cc: Milosz Tanski
---
fs/cachefiles/xattr.c | 29 +++--
1 file changed, 15 insertions(+), 14 deletions(-)
diff --git a/fs/cachefiles/xattr.c
it to an appropriately typed variable rather than to ret.
Signed-off-by: Josh Boyer
Signed-off-by: David Howells
cc: Hongyi Jia
cc: Milosz Tanski
---
fs/cachefiles/xattr.c | 29 +++--
1 file changed, 15 insertions(+), 14 deletions(-)
diff --git a/fs/cachefiles/xattr.c
/0x200 [ceph]
> [] ? __fsnotify_inode_delete+0xe/0x10
> [] destroy_inode+0x3c/0x70
> [] evict+0x119/0x1b0
>
> Signed-off-by: Milosz Tanski
Acked-by: David Howells
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to maj
Milosz Tanski wrote:
> I think that change does the trick. I had it running on the same
> machine for 5 hours and had the kernel forcefully drop some of the
> inodes in the cache (via drop caches) without a crash. I'll send a
> proper patch email after you take a look and make sure I did the righ
Milosz Tanski wrote:
> - (1 << FSCACHE_OP_WAITING);
> + (1 << FSCACHE_OP_WAITING) |
> + (1 << FSCACHE_OP_UNUSE_COOKIE);
Yeah... That'll do it. We could just decrement n_active directly after
calling into the backend - after all, we cannot reduce n_acti
Milosz Tanski wrote:
> After running this for a day on some loaded machines I ran into what
> looks like an old issue with the new code. I remember you saw an issue
> that manifested it self in a similar way a while back.
>
> [13837253.462779] FS-Cache: Assertion failed
> [13837253.462782] 3 ==
Implement the FS-Cache interface to check the consistency of a cache object in
CacheFiles.
Original-author: Hongyi Jia
Signed-off-by: David Howells
cc: Hongyi Jia
cc: Milosz Tanski
---
fs/cachefiles/interface.c | 26 ++
fs/cachefiles/internal.h |1 +
fs
Cache takes care of pinning the
object and the cookie in memory and managing this call with respect to the
object state.
Original-author: Hongyi Jia
Signed-off-by: David Howells
cc: Hongyi Jia
cc: Milosz Tanski
---
Documentation/filesystems/caching/backend-api.txt |9 +++
Document
I've pushed the non-Ceph bits of the bad-page fix to here:
http://git.kernel.org/cgit/linux/kernel/git/dhowells/linux-fs.git/log/?h=fscache
and added a CIFS fix.
David
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majord...@vger.ke
Milosz Tanski wrote:
> Is it as simple as stick a mutex at the top of the
> __fscache_check_consistency function before we try to find the object?
You can lock a mutex in a function, but where are you actually going to place
the mutex struct? And what other code is going to take it? To do this
Milosz Tanski wrote:
> If the cache is withdrawn and we're starting anew I would consider
> that to okay. I would consider an empty page cache for a cookie to be
> consistent since there's nothing stale that I can read. Unless there's
> another synchronization issue that I'm missing in fscache.
Hongyi Jia wrote:
> +bool __fscache_check_consistency(struct fscache_cookie *cookie)
> +{
> + struct fscache_object *object;
> +
> + if (cookie->def->type != FSCACHE_COOKIE_TYPE_DATAFILE)
> + return false;
> +
> + if (hlist_empty(&cookie->backing_objects))
> +
Sage Weil wrote:
> David, are the fscache patches here ready for the next merge window? Do
> you have a preference for whose tree they go through?
There's only one problem - patch 1 needs to come _after_ patch 2 to avoid
breaking git bisect. Plus these patches 2 and 4 extend the fscache API
w
Milosz Tanski wrote:
> Am I on the right path here?
I think you are! :-)
Sorry for not getting back to you earlier, but August has been somewhat
compressed due to PTO.
David
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majord...@vger.ker
Milosz Tanski wrote:
> I would like to find out if you could add a new API for re-validating
> a cookie. My use case is still FSCache support for Ceph.
>
> Ceph has a notion of capabilities (like CACHE cap) that can be
> granted, revoked or given up. For example the code there to give up
> the c
Milosz Tanski wrote:
> It looks like both the cifs and NFS code do not bother with any
> locking around cifs_fscache_set_inode_cookie. Is there no concern over
> multiple open() calls racing to create the cookie in those
> filesystems?
Yeah... That's probably wrong. AFS obviates the need for s
Milosz Tanski wrote:
> > Looking at your index structure, ceph has per-fsid indices under the top
> > "ceph" index and then per-inode indices under those? Are fsids universally
> > unique - or just for a given server/cell/whatever?
>
> It's my understanding that's a guuid assigned to the cluste
Okay, my analysis of the patch:
Looking at your index structure, ceph has per-fsid indices under the top
"ceph" index and then per-inode indices under those? Are fsids universally
unique - or just for a given server/cell/whatever?
> +#ifdef CONFIG_CEPH_FSCACHE
> + if (PageFsCache(page))
> +
Milosz Tanski wrote:
> You can also get the 'wip-ceph-fscache' branch from my gitrepo:
There's only one patch from you there. Shouldn't there be at least two as you
posted?
David
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majord...@vger
Milosz Tanski wrote:
> I just looked and saw that it's been pulled a couple hours ago.
>
> Can I also trouble you into looking at my patches for Ceph for the
> FSCache? We're using it (in production starting today actually); we're
> not able to find any bugs with the current iteration. But it's
Milosz Tanski wrote:
> Now that 3.10 is out I hope that David Howell's fscache improvement
> branch will go into the next kernel.
It has just been pulled.
David
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majord...@vger.kernel.org
More ma
33 matches
Mail list logo
