glusterfs.git, branch v3.1.3solaris

Fixing argp build failures seen on solaris

2011-03-23T12:03:18+00:00

Fix courtesy of pavan (tcp@gluster.com).

Signed-off-by: shishir gowda 
Signed-off-by: Vijay Bellur 

BUG: 2563 ()
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2563

nfs3: Flush file I/O call states on open failure

2011-03-22T13:04:38+00:00

Signed-off-by: Shehjar Tikoo 
Signed-off-by: Vijay Bellur 

BUG: 2566 (NFS read hangs when arequal-checksum script is run)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2566

CLI : Fix memory free for key_fixed.

2011-03-22T13:04:34+00:00

Signed-off-by: Gaurav 
Signed-off-by: Vijay Bellur 

BUG: 2554 (Invalid memory access in glusterd.)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2554

Solaris: redefine O_DIRECTORY flag to the correct value

2011-03-22T13:04:30+00:00

O_DIRECTORY is not supported on non-linux systems. By defining it
as 0, the check in qr_open always fails with error ENOTDIR.

We need to document the limiation of not supporting O_DIRECTORY flag
when solaris is used as a backend.

Signed-off-by: shishir gowda 
Signed-off-by: Vijay Bellur 

BUG: 2566 (NFS read hangs when arequal-checksum script is run)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2566

Change NFS_PATH_MAX value to 4096

2011-03-19T03:05:52+00:00

This is tmp fix for crash seen on solaris gnfs server is received path is
longer than PATH_MAX. On solaris default PATH_MAX is 1024, and linux default
PATH_MAX is 4096.

Signed-off-by: shishir gowda 
Signed-off-by: Vijay Bellur 

BUG: 2476 ()
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2476

Disable geo-replication build

2011-03-14T12:46:27+00:00

Signed-off-by: Vijay Bellur 
Signed-off-by: Vijay Bellur 

BUG: 1570 (geosync related changes)
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=1570

rpc-transport/rdma: handle cases where peer is NULL in send_completion.

2011-03-12T13:10:39+00:00

- There can be valid cases where peer is NULL - For eg., race b/w pollin
    received on handshake socket and as a result transport getting disconnected
    and handling of a send/recv completion on the same qp for some previous
    read/write - in send/recv_completion_process. Hence we should be checking
    for a valid peer pointer before accessing it.

Signed-off-by: Raghavendra G 
Signed-off-by: Vijay Bellur 

BUG: 2514 ()
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2514

nfs-rpc: Fix order of freeing request and connection objects

2011-03-12T13:10:35+00:00

..to prevent segfaults due to double frees and dereferencing of freed
up memory.

Signed-off-by: Shehjar Tikoo 
Signed-off-by: Vijay Bellur 

BUG: 2504 ()
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=2504

posix: Set op_errno to ENOMEM on failed iobuf_get

2011-03-12T05:58:26+00:00

This prevents a situation where op_ret for the read fop is
-1 but the op_errno is still 0.

Signed-off-by: Shehjar Tikoo 
Signed-off-by: Vijay Bellur 

BUG: 1977 ()
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=1977

posix: Handle offset greater than file size of EOF determination

2011-03-12T05:58:21+00:00

Signed-off-by: Shehjar Tikoo 
Signed-off-by: Vijay Bellur 

BUG: 1977 ()
URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=1977