/distfiles/ | wc -l
47470
/usr/portage contains a lot of directories, /usr/portage/distfiles
contains lots of files.
A bisect in fs/cifs gives:
$ git bisect bad
c4d3396b261473ded6f370edd1e79ba34e089d7e is the first bad commit
commit c4d3396b261473ded6f370edd1e79ba34e089d7e
Author: Jeff Layton
On Tue, 2 Aug 2011 06:44:55 -0400
Jeff Layton jlay...@redhat.com wrote:
On Tue, 2 Aug 2011 02:30:35 +0200
Jan Seiffert kaffeemons...@googlemail.com wrote:
Please CC, as not subscibed.
Looks like something broke CIFS.
Kernel 3.0.0 works, master after the CIFS merge does not ATM
that in the not too distant future, but it requires
some relatively major work to do it right.
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo
) {
mutex_unlock(server-srv_mutex);
+ atomic_dec(server-inFlight);
+ wake_up(server-request_q);
return -ENOMEM;
}
Good catch -- looks correct:
Reviewed-by: Jeff Layton jlay...@redhat.com
--
To unsubscribe from this list: send the line
...we may need to come up with a
scheme more like NFS does. It will alter the tree when it discovers
that a root dir that it previously was working on is actually a subdir
of another mount.
Thoughts?
--
Jeff Layton jlay...@redhat.com
--
To unsubscribe from this list: send the line unsubscribe
://bugzilla.redhat.com/show_bug.cgi?id=727927
Reported-by: TR Bentley h...@trarbentley.net
Reported-by: Iain Arnell iarn...@gmail.com
Cc: Al Viro v...@zeniv.linux.org.uk
Cc: sta...@kernel.org
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsfs.c |4
1 files changed, 4 insertions(+), 0
Cc: Pavel Shilovsky piastr...@gmail.com
Cc: sta...@kernel.org
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/inode.c | 14 ++
1 files changed, 2 insertions(+), 12 deletions(-)
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 9b018c8..a7b2dcd 100644
--- a/fs/cifs
during debugging.
Cc: sta...@kernel.org
Cc: Al Viro v...@zeniv.linux.org.uk
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/dir.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index ae576fb..72d448b 100644
--- a/fs/cifs/dir.c
run I hit it 7
times within 1 second. I think this started with commit dc137bf553
which is pretty recent. Before that, the read_seqretry check wasn't
present.
--
Jeff Layton jlay...@redhat.com
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message
On Tue, 9 Aug 2011 20:49:57 +0400
Pavel Shilovsky piastr...@gmail.com wrote:
2011/8/4 Pavel Shilovsky piastr...@gmail.com:
2011/8/4 Jeff Layton jlay...@redhat.com:
Hi Pavel,
We have a bit of a problem...
https://bugzilla.redhat.com/show_bug.cgi?id=727834
The shared superblock
again:
80975d21aae2136ccae1ce914a1602dc1d8b0795
f9e8c45002cacad536b338dfa9e910e341a49c31
Cheers,
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http
that might be helpful is turning on page poisoning and
redoing this test, that might make it crash sooner and point out the
source of the corruption.
Even better would be a bisect to track down the cause...
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe
(volume_info-iocharset);
kfree(volume_info-prepath);
Looks correct, nice catch.
Reviewed-by: Jeff Layton jlay...@redhat.com
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org
(volume_info-iocharset);
kfree(volume_info-prepath);
Also, this patch should obviously go to stable too.
--
Jeff Layton jlay...@redhat.com
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http
.
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, 18 Aug 2011 12:25:50 -0500
Steve French smfre...@gmail.com wrote:
On Thu, Aug 18, 2011 at 12:16 PM, Justin Piszcz jpis...@lucidpixels.com
wrote:
On Thu, 18 Aug 2011, Jeff Layton wrote:
On Thu, 18 Aug 2011 09:15:36 -0400 (EDT)
Justin Piszcz jpis...@lucidpixels.com wrote
, but that just recently changed...
Reads are done using relatively small buffers and then copied to
pagecache. Part of what I'm working on will be to allow for much larger
reads directly into the pagecache. That should also help performance
significantly.
--
Jeff Layton jlay...@samba.org
...as that's more efficient when we know that the lengths are equal.
Reported-by: David Howells dhowe...@redhat.com
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifssmb.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
...it's more efficient since we know the length.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsencrypt.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
index e76bfeb..f11dfb6 100644
--- a/fs/cifs
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsfs.c |3 ---
1 files changed, 0 insertions(+), 3 deletions(-)
diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index f93eb94..eac8e087 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -530,7 +530,6 @@ cifs_get_root(struct
interchangeably in the code.
This makes little sense as these two values are not related at all.
Separate them and make sure the code uses the right values in the right
places.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifssmb.c | 33 ++---
fs/cifs/connect.c | 11
once these changes are in place.
Comments and suggestions welcome...
Jeff Layton (14):
cifs: trivial: remove obsolete comment
cifs: make smb_msg local to read_from_socket
cifs: check for unresponsive server every time we call kernel_recvmsg
cifs: simplify read_from_socket
cifs: turn
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c |6 --
1 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 0ffba09..55efd1d 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -694,12 +694,6 @@ incomplete_rcv
If msg_controllen is 0, then the socket layer should never touch these
fields. Thus, there's no need to continually reset them. Also, there's
no need to keep this field on the stack for the demultiplex thread, just
make it a local variable in read_from_socket.
Signed-off-by: Jeff Layton jlay
If the server stops sending data while in the middle of sending a
response then we still want to reconnect it if it doesn't come back.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 33 ++---
1 files changed, 22 insertions(+), 11 deletions
the same way.
Also, make it return an int (since we'll never receive enough to worry
about the sign bit anyway), and simply make it return the amount of bytes
read or a negative error code.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 77
to use a copy of the kvec array
and refresh that copy on each pass through the loop.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 72 +++-
1 files changed, 65 insertions(+), 7 deletions(-)
diff --git a/fs/cifs/connect.c b/fs
a new receive phase and normalize the cFYI messages.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 63 +---
1 files changed, 30 insertions(+), 33 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 430e563
Have the demultiplex thread receive just enough to get to the MID, and
then find it before receiving the rest. Later, we'll use this to swap
in a preallocated receive buffer for some calls.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 31
.
Finally, change checkSMB to return a real error code (-EIO) when
it finds an error. That will help simplify some coming changes in the
callers.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c |2 +-
fs/cifs/misc.c| 51 +--
2
-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 113 +++--
1 files changed, 66 insertions(+), 47 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 28007fd..bfd9622 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs
as
well and just pass by value. If a mid is found, then we can just mark
it as malformed. If one isn't found, then the value of length won't
change anyway.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 35 ---
1 files changed, 16 insertions
We have several functions that need to access these pointers. Currently
that's done with a lot of double pointer passing. Instead, move them
into the TCP_Server_Info and simplify the handling.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsglob.h |4 ++
fs/cifs/connect.c
In order to receive directly into a preallocated buffer, we need to ID
the mid earlier, before the bulk of the response is read. Call the mid
finding routine as soon as we're able to read the mid.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 24
Move the entire 3rd phase of the receive codepath into a separate
function in preparation for the addition of a pluggable receive
function.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 101 +++--
1 files changed, 59
the mid_q_entry has been identified
in order to receive the rest of the SMB. If the mid can't be identified
or the receive pointer is unset, then the standard 3rd phase receive
function will be called.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsglob.h | 23
));
_FreeXid(xid);
kfree(full_path);
What exactly does this fix? The only possible thing I could see it is
if sb-s_root was a negative dentry, and I'm not sure that would ever
happen. Would it?
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux
this by making the name_len an unsigned value in CIFSFindNext.
Cc: sta...@kernel.org
Reported-by: Darren Lavender d...@hppine99.gbr.hp.com
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifssmb.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/fs/cifs/cifssmb.c b
On Tue, 23 Aug 2011 17:55:25 +0530
Suresh Jayaraman sjayara...@suse.de wrote:
On 08/23/2011 04:51 PM, Jeff Layton wrote:
The name_len variable in CIFSFindNext is a signed int that gets set to
the resume_name_len in the cifs_search_info. The resume_name_len however
is unsigned and for some
these perms */
+};
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
, but this certainly doesn't preclude
that, and seems useful.
One minor nit: you should add the manpages before the autoconf changes
to make sure that bisectability is preserved.
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message
It currently says no is the default, but it should be yes.
Reported-by: Elias Pipping pipp...@lavabit.com
Signed-off-by: Jeff Layton jlay...@samba.org
---
configure.ac |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/configure.ac b/configure.ac
index 780abe7..6a9ad75
with the appropriate flag. If all the backup users were in
a certan group, for instance, then you could use that info to decide
whether to set the flag in the open calls.
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message
, but if the necessary stuff wasn't
present then it would be autodisabled. That's how cifs.upcall works
too. If default=auto is more illustrative of this situation then I'll
respin the patch and fix the cifs.upcall help text too.
Quoting Jeff Layton jlay...@samba.org:
It currently says no is the default
On Sun, 21 Aug 2011 12:22:51 -0400
Jeff Layton jlay...@samba.org wrote:
On Sun, 21 Aug 2011 19:30:15 +0400
Pavel Shilovsky piastr...@gmail.com wrote:
move it to the beginning of the loop.
Signed-off-by: Pavel Shilovsky piastr...@gmail.com
---
fs/cifs/cifsfs.c | 10
;
}
+
error_exit:
kfree(full_path);
cifs_put_tlink(tlink);
--
Jeff Layton jlay...@redhat.com
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
to review
the kernel pieces however. If Steve plans to review and commit those,
then I'll plan to pull the userspace patch into the next cifs-utils
release.
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord
autoconf tests pull in extra includes, but I don't have my
autoconf book handy.
Shirish, can you look into this, fix it and resubmit this patch?
Simo tells me there's an open samba bug to fix this...
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux
On Thu, 25 Aug 2011 13:50:06 -0500
Shirish Pargaonkar shirishpargaon...@gmail.com wrote:
On Thu, Aug 25, 2011 at 1:40 PM, Shirish Pargaonkar
shirishpargaon...@gmail.com wrote:
On Thu, Aug 25, 2011 at 1:34 PM, Jeff Layton jlay...@samba.org wrote:
On Tue, 23 Aug 2011 14:04:24 -0500
Some versions of wbclient.h have function declarations with bool type
args, but they don't include wbclient.h themselves. Make sure that
we can deal with that by telling the autoconf test to include stdbool.h
explicitly.
Signed-off-by: Jeff Layton jlay...@samba.org
---
configure.ac | 59
of dcache trickery, but I tend
to think we probably ought to follow NFS' example here. It has sort of
an opportunistic mechanism for filling in the dcache for a particular
superblock. This is detailed in the comments on commit 54ceac45.
David, any thoughts here?
--
Jeff Layton jlay...@redhat.com
On Thu, 25 Aug 2011 15:55:14 -0500
Shirish Pargaonkar shirishpargaon...@gmail.com wrote:
On Thu, Aug 25, 2011 at 3:31 PM, Jeff Layton jlay...@samba.org wrote:
Some versions of wbclient.h have function declarations with bool type
Jeff, did not understand this verbiage, did you mean winbind
this infrastructure to allow for a larger rsize.
Jeff Layton (16):
cifs: clean up checkSMB
cifs: consolidate signature generating code
cifs: trivial: remove obsolete comment
cifs: make smb_msg local to read_from_socket
cifs: check for unresponsive server every time we call kernel_recvmsg
.
Finally, change checkSMB to return a real error code (-EIO) when
it finds an error. That will help simplify some coming changes in the
callers.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/misc.c | 51 +--
1 files changed, 25 insertions
We have two versions of signature generating code. A vectorized and
non-vectorized version. Eliminate a large chunk of cut-and-paste
code by turning the non-vectorized version into a wrapper around the
vectorized one.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsencrypt.c | 103
If msg_controllen is 0, then the socket layer should never touch these
fields. Thus, there's no need to continually reset them. Also, there's
no need to keep this field on the stack for the demultiplex thread, just
make it a local variable in read_from_socket.
Signed-off-by: Jeff Layton jlay
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c |6 --
1 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 0ffba09..55efd1d 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -694,12 +694,6 @@ incomplete_rcv
If the server stops sending data while in the middle of sending a
response then we still want to reconnect it if it doesn't come back.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 33 ++---
1 files changed, 22 insertions(+), 11 deletions
the same way.
Also, make it return an int (since we'll never receive enough to worry
about the sign bit anyway), and simply make it return the amount of bytes
read or a negative error code.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 77
to use a copy of the kvec array
and refresh that copy on each pass through the loop.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 72 +++-
1 files changed, 65 insertions(+), 7 deletions(-)
diff --git a/fs/cifs/connect.c b/fs
Having to continually allocate a new kvec array is expensive. Allocate
one that's big enough, and only reallocate it as needed.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsglob.h |2 ++
fs/cifs/connect.c | 26 --
2 files changed, 22 insertions
a new receive phase and normalize the cFYI messages.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 63 +---
1 files changed, 30 insertions(+), 33 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index f7bd9ba
Have the demultiplex thread receive just enough to get to the MID, and
then find it before receiving the rest. Later, we'll use this to swap
in a preallocated receive buffer for some calls.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 31
as
well and just pass by value. If a mid is found, then we can just mark
it as malformed. If one isn't found, then the value of length won't
change anyway.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 35 ---
1 files changed, 16 insertions
We have several functions that need to access these pointers. Currently
that's done with a lot of double pointer passing. Instead, move them
into the TCP_Server_Info and simplify the handling.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsglob.h |4 ++
fs/cifs/connect.c
In order to receive directly into a preallocated buffer, we need to ID
the mid earlier, before the bulk of the response is read. Call the mid
finding routine as soon as we're able to read the mid.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 24
.
The rest is handled by the async readv infrastructure.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/file.c | 260 +++
1 files changed, 90 insertions(+), 170 deletions(-)
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index fd57165..2f263bd
and wsize are less than PAGE_CACHE_SIZE. Although it has never been
enforced at mount time, cifs.ko has never handled those cases properly.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 101 +++--
fs/cifs/file.c| 12 +-
2
in the response, and the page is beyond the
server-eof, then simply zero it out, set it uptodate and leave it
in place. This ensures that the VFS won't keep trying to issue reads
for these regions.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifssmb.c | 15 +++
1 files changed, 15
embedded in struct page. Once
the read completes, the pages are added to the LRU list, set Uptodate
and unlocked.
While this may seem like an odd way to handle the pagecache, it does
prevent the VFS from attempting to issue duplicate reads on the same
pages.
Signed-off-by: Jeff Layton jlay
On Sat, 27 Aug 2011 00:11:57 +0400
Pavel Shilovsky pias...@etersoft.ru wrote:
2011/8/26 Jeff Layton jlay...@redhat.com:
On Fri, 26 Aug 2011 10:52:21 +0400
Pavel Shilovsky pias...@etersoft.ru wrote:
Reorganize code and make it send qpath info request only for a full
path (//server/share
On Sat, 27 Aug 2011 14:41:36 +0400
Pavel Shilovsky pias...@etersoft.ru wrote:
2011/8/27 Jeff Layton jlay...@poochiereds.net:
On Fri, 26 Aug 2011 22:03:03 +0400
Pavel Shilovsky pias...@etersoft.ru wrote:
Both these options are started with rw - that's why the first one
isn't switched
.
Thanks,
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
If msg_controllen is 0, then the socket layer should never touch these
fields. Thus, there's no need to continually reset them. Also, there's
no need to keep this field on the stack for the demultiplex thread, just
make it a local variable in read_from_socket.
Signed-off-by: Jeff Layton jlay
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c |6 --
1 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index b5f0d90..5667964 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -694,12 +694,6 @@ incomplete_rcv
.
Finally, change checkSMB to return a real error code (-EIO) when
it finds an error. That will help simplify some coming changes in the
callers.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/misc.c | 51 +--
1 files changed, 25 insertions
in this code. I
feel that's essential as bugs in the receive path can be very subtle and
hard to detect. Being able to bisect to find bugs is a generally a good
thing. Hence, I highly recommend committing these in order.
Jeff Layton (16):
cifs: clean up checkSMB
cifs: consolidate signature generating
If the server stops sending data while in the middle of sending a
response then we still want to reconnect it if it doesn't come back.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 33 ++---
1 files changed, 22 insertions(+), 11 deletions
the same way.
Also, make it return an int (since we'll never receive enough to worry
about the sign bit anyway), and simply make it return the amount of bytes
read or a negative error code.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 77
to use a copy of the kvec array
and refresh that copy on each pass through the loop.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 72 +++-
1 files changed, 65 insertions(+), 7 deletions(-)
diff --git a/fs/cifs/connect.c b/fs
Having to continually allocate a new kvec array is expensive. Allocate
one that's big enough, and only reallocate it as needed. Also, we need
to use GFP_NOFS here to ensure that this doesn't deadlock.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsglob.h |2 ++
fs/cifs
Have the demultiplex thread receive just enough to get to the MID, and
then find it before receiving the rest. Later, we'll use this to swap
in a preallocated receive buffer for some calls.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 31
-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 113 +++--
1 files changed, 66 insertions(+), 47 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index bc16873..6b4fcb3 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs
as
well so just pass by value. If a mid is found, then we can just mark
it as malformed. If one isn't found, then the value of length won't
change anyway.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 35 ---
1 files changed, 16 insertions
We have several functions that need to access these pointers. Currently
that's done with a lot of double pointer passing. Instead, move them
into the TCP_Server_Info and simplify the handling.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsglob.h |4 ++
fs/cifs/connect.c
Move the entire 3rd phase of the receive codepath into a separate
function in preparation for the addition of a pluggable receive
function.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 101 +++--
1 files changed, 59
the mid_q_entry has been identified
in order to receive the rest of the SMB. If the mid can't be identified
or the receive pointer is unset, then the standard 3rd phase receive
function will be called.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifsglob.h | 23
or on higher-latency connections.
Steve, I'd like to see both of these sets go into 3.2 if at all
possible. Both of these sets should be bisectable, so committing them in
order is highly recommended.
Jeff Layton (4):
cifs: fix protocol definition for READ_RSP
cifs: add cifs_async_readv
cifs: convert
There is no pad, and it simplifies the code to remove the Data field.
None of the existing code relies on these fields, or on the READ_RSP
being a particular length.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/cifspdu.h |4 +---
1 files changed, 1 insertions(+), 3 deletions
that to receive the rest of the
frame. cifs_readv_receive will then marshal up a kvec array with
kmapped pages from the pagecache, which eliminates one copy of the
data. Once the data is received, the pages are added to the pagecache,
set uptodate, and unlocked.
Signed-off-by: Jeff Layton jlay...@redhat.com
Now that we have code in place to do asynchronous reads, convert
cifs_readpages to use it. The new cifs_readpages walks the page_list
that gets passed in, sets up cifs_readdata to handle the reads.
The rest is handled by the async readv infrastructure.
Signed-off-by: Jeff Layton jlay
handled those cases properly.
Signed-off-by: Jeff Layton jlay...@redhat.com
---
fs/cifs/connect.c | 113
fs/cifs/file.c| 12 +-
2 files changed, 79 insertions(+), 46 deletions(-)
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
-UNCip);
+ kfree(volume_info-UNC);
kfree(volume_info-domainname);
kfree(volume_info-iocharset);
kfree(volume_info-prepath);
Seems reasonable for silencing the checker...
Reviewed-by: Jeff Layton jlay...@redhat.com
--
To unsubscribe from this list: send the line
unsubscribe linux-cifs in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Latest set of 8 patches committed.
--
Jeff Layton jlay...@samba.org
--
To unsubscribe from this list: send the line unsubscribe linux-cifs in
the body
On Tue, 23 Aug 2011 09:03:23 -0400
Jeff Layton jlay...@samba.org wrote:
It currently says no is the default, but it should be yes.
Reported-by: Elias Pipping pipp...@lavabit.com
Signed-off-by: Jeff Layton jlay...@samba.org
---
configure.ac |2 +-
1 files changed, 1 insertions(+), 1
On Thu, 25 Aug 2011 16:31:03 -0400
Jeff Layton jlay...@samba.org wrote:
Some versions of wbclient.h have function declarations with bool type
args, but they don't include wbclient.h themselves. Make sure that
we can deal with that by telling the autoconf test to include stdbool.h
explicitly
;
if (strncmp(token, ro, 2) == 0)
return OPT_RO;
- if (strncmp(token, rw, 2) == 0)
+ if (strncmp(token, rw, 2) == 0 strlen(token) == 2)
return OPT_RW;
if (strncmp(token, remount, 7) == 0)
return OPT_REMOUNT;
Committed.
--
Jeff
On Mon, 29 Aug 2011 16:24:33 -0500
Shirish Pargaonkar shirishpargaon...@gmail.com wrote:
On Tue, Aug 23, 2011 at 8:15 AM, Jeff Layton jlay...@samba.org wrote:
On Mon, 22 Aug 2011 08:33:49 -0500
Shirish Pargaonkar shirishpargaon...@gmail.com wrote:
On Fri, Aug 12, 2011 at 11:33 AM
On Mon, 29 Aug 2011 20:11:13 -0500
Steve French smfre...@gmail.com wrote:
On Mon, Aug 29, 2011 at 7:29 PM, Jeff Layton jlay...@samba.org wrote:
On Mon, 29 Aug 2011 16:24:33 -0500
Shirish Pargaonkar shirishpargaon...@gmail.com wrote:
On Tue, Aug 23, 2011 at 8:15 AM, Jeff Layton jlay
801 - 900 of 2365 matches
Mail list logo