glusterfs.git/libglusterfs/src, branch v3.1.2qa1

inode: fix crash in logging during cyclic loop detection

2010-12-07T06:18:55+00:00

Signed-off-by: Anand V. Avati 
Signed-off-by: Anand V. Avati 

BUG: 2196 ()
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2196

iatt: fix bug in ia_makedev

2010-11-24T14:39:55+00:00

the wrong binary operator was resulting in lookup/stat showing wrong
major/minor numbers for device files

Signed-off-by: Anand V. Avati 
Signed-off-by: Anand V. Avati 

BUG: 2145 ([3.1.1qa9] mknod does not work properly)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2145

inode: catch loop formation during inode_link() and fail linking

2010-11-24T08:33:41+00:00

- explores all parent branches
- performs loop formation check only if the operation is resulting in relinking
  of an inode already existing in the table

Signed-off-by: Anand V. Avati 
Signed-off-by: Anand V. Avati 
Signed-off-by: Anand V. Avati 

BUG: 971 (dynamic volume management)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=971

nfs: opendir/closedir for every readdir

2010-11-15T12:03:50+00:00

Revert "nfs3: Unref & unbind dir fd with inode lock on EOF"

This reverts commit 4e6fb304ce41acbaf7c9ba67c06bf443e65082e8.

The above commit (which unbinds fds at EOF) does not fix the original
bug (1619) because a readdir from a second app could have already
started before the readdir_cbk of the first app's readdir reaches
NFS code. Hence the race still exists.

Performing extra unrefs when EOF is received is not a reliable way
of detecting that a client has performed a closedir (and to close
the fd ourselves). Neither is interpreting a 0 cookies a new opendir.
Clients can always use telldir/seekdir and hit EOFs twice.

Due to the way NFS3 protocol is designed, it is just not possible
for the server to reliably detect opendirs/closedirs performed by
the client and map the corresponding readdirs to the same dir fd
on the server side.

The only reliable way of fixing this is to perform opendir/closedir
at the cost of performance. Any optimization towards keeping dir fds
open attempting to map them with application's opendir/closedir will
either result in fd leaks or extra fd unrefs.

Signed-off-by: Anand V. Avati 
Signed-off-by: Anand V. Avati 

BUG: 2061 (NFS server crashes in readdir_fstat_cbk due to extra fd unref)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2061

Solaris: vasprint fix for %llu crash

2010-11-13T15:02:27+00:00

vasprint now inturn call gf_vasprintf, which calls vsnprintf.
vsnprintf allocates buffer of required size, and hence prevents
the sigsegv seen in vasprintf w.r.t %llu

Signed-off-by: shishir gowda 
Signed-off-by: Anand V. Avati 

BUG: 1058 (vasprintf dumps core when %llu is involved)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=1058

libglusterfs: Memory accounting changes

2010-11-09T11:07:03+00:00

This patch disables memory accounting by default and provides better
alignemnt for faster fetch.

Signed-off-by: Vijay Bellur 
Signed-off-by: Anand V. Avati 

BUG: 2068 (performance enhancements)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2068

rpc-transport/rdma: Fix 2KB as threshold size for msgs that can be transferred inline.

2010-11-08T04:15:08+00:00

- Any fop/mop that can result a reply whose size is greater than this threshold
    allocates and submits a buffer along with the request to receive reply.

Signed-off-by: Raghavendra G 
Signed-off-by: Anand V. Avati 

BUG: 513 (Introduce 0 copy rdma)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=513

Solaris build fixes: Patch for 64bit build

2010-10-28T06:37:32+00:00

This will fix build failures when -m64 option is used to build
gluster

Signed-off-by: Super-User 
Signed-off-by: Anand V. Avati 

BUG: 1990 (Gluster mainline build on solaris fails with errors)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=1990

libglusterfs: fix buffer overflow in gf_string2int()

2010-10-27T11:13:30+00:00

Signed-off-by: Csaba Henk 
Signed-off-by: Anand V. Avati 

BUG: 2022 (gf_string2int() overflows target pointer)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2022

fuse: retire the faked-FLUSH-upon-RELEASE hack

2010-10-27T06:56:28+00:00

Now the code has lived up to the glorious state of not relying
on getting the FLUSH whenever a file is released. So we don't need
to forge one in release for the cases when the kernel doesn't send
it.

Undo commits:
- 155ffe5c
- c50bc710
- b8779318 (partly, just release related parts)

Signed-off-by: Csaba Henk 
Signed-off-by: Anand V. Avati 

BUG: 223 (flush not sent)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=223