]> git.openfabrics.org - ~shefty/librdmacm.git/log
~shefty/librdmacm.git
11 years agopop
Sean Hefty [Tue, 9 Oct 2012 16:54:40 +0000 (09:54 -0700)]
pop

11 years agopop
Sean Hefty [Tue, 9 Oct 2012 16:54:40 +0000 (09:54 -0700)]
pop

11 years agoimport
Sean Hefty [Tue, 9 Oct 2012 16:54:25 +0000 (09:54 -0700)]
import

11 years agoimport
Sean Hefty [Tue, 9 Oct 2012 16:54:25 +0000 (09:54 -0700)]
import

11 years ago[5/5,librdmacm] rping: added checks to the return values functions
Dotan Barak [Tue, 9 Oct 2012 12:27:52 +0000 (12:27 +0000)]
[5/5,librdmacm] rping: added checks to the return values functions

This will make rping to exit with return value other than zero in case of an
error.

Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
11 years ago[4/5,librdmacm] rstream: added missing return is accept() failed
Dotan Barak [Tue, 9 Oct 2012 12:27:51 +0000 (12:27 +0000)]
[4/5,librdmacm] rstream: added missing return is accept() failed

Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
11 years ago[3/5,librdmacm] rstream: initialize return value in server_connect()
Dotan Barak [Tue, 9 Oct 2012 12:27:50 +0000 (12:27 +0000)]
[3/5,librdmacm] rstream: initialize return value in server_connect()

If use_async == 0 and rs_accept() passes (i.e. non negative value), then
the return value from the function was uninitialized.

Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
11 years ago[2/5,librdmacm] rsocket: added missing break
Dotan Barak [Tue, 9 Oct 2012 12:27:49 +0000 (12:27 +0000)]
[2/5,librdmacm] rsocket: added missing break

Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
11 years ago[1/5,librdmacm] rsocket: add missing va_end() after calling va_end()
Dotan Barak [Tue, 9 Oct 2012 12:27:48 +0000 (12:27 +0000)]
[1/5,librdmacm] rsocket: add missing va_end() after calling va_end()

Not doing so, may lead to resource leak.

Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
11 years agopop
Sean Hefty [Tue, 9 Oct 2012 16:52:20 +0000 (09:52 -0700)]
pop

11 years agopop
Sean Hefty [Tue, 9 Oct 2012 16:52:20 +0000 (09:52 -0700)]
pop

11 years agorefresh
Sean Hefty [Mon, 8 Oct 2012 17:47:37 +0000 (10:47 -0700)]
refresh

11 years agorefresh
Sean Hefty [Mon, 8 Oct 2012 17:47:37 +0000 (10:47 -0700)]
refresh

11 years agorefresh (create temporary patch)
Sean Hefty [Mon, 8 Oct 2012 17:47:12 +0000 (10:47 -0700)]
refresh (create temporary patch)

11 years agorefresh (create temporary patch)
Sean Hefty [Mon, 8 Oct 2012 17:47:12 +0000 (10:47 -0700)]
refresh (create temporary patch)

11 years agoRefresh of acm-port
Sean Hefty [Mon, 8 Oct 2012 17:47:12 +0000 (10:47 -0700)]
Refresh of acm-port

11 years agorefresh
Sean Hefty [Mon, 8 Oct 2012 17:37:43 +0000 (10:37 -0700)]
refresh

11 years agorefresh
Sean Hefty [Mon, 8 Oct 2012 17:37:43 +0000 (10:37 -0700)]
refresh

11 years agorefresh (create temporary patch)
Sean Hefty [Mon, 8 Oct 2012 17:37:43 +0000 (10:37 -0700)]
refresh (create temporary patch)

11 years agorefresh (create temporary patch)
Sean Hefty [Mon, 8 Oct 2012 17:37:43 +0000 (10:37 -0700)]
refresh (create temporary patch)

11 years agoRefresh of acm-port
Sean Hefty [Mon, 8 Oct 2012 17:37:43 +0000 (10:37 -0700)]
Refresh of acm-port

11 years agolibrdmacm: Disable ACM support if ibacm.port is not found
Sean Hefty [Mon, 8 Oct 2012 17:33:21 +0000 (10:33 -0700)]
librdmacm: Disable ACM support if ibacm.port is not found

The librdmacm will try to connect port 6125 if ibacm.port is
not found.  The problem is that some other service or application
could be using that port and respond with garbage.  Rather
than falling back to a hard coded port number, if ibacm.port
is not found, simply disable ACM support.

This has the effect of removing support for older versions
of ibacm, unless the port file is created manually.

Patch created based on feedback from Doug Ledford and Florian
Weimer from RedHat.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agolibrdmacm: Disable ACM support if ibacm.port is not found
Sean Hefty [Mon, 8 Oct 2012 17:33:21 +0000 (10:33 -0700)]
librdmacm: Disable ACM support if ibacm.port is not found

The librdmacm will try to connect port 6125 if ibacm.port is
not found.  The problem is that some other service or application
could be using that port and respond with garbage.  Rather
than falling back to a hard coded port number, if ibacm.port
is not found, simply disable ACM support.

This has the effect of removing support for older versions
of ibacm, unless the port file is created manually.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agonew
Sean Hefty [Mon, 8 Oct 2012 17:33:21 +0000 (10:33 -0700)]
new

11 years agonew
Sean Hefty [Mon, 8 Oct 2012 17:33:21 +0000 (10:33 -0700)]
new

11 years agolibrdmacm: Disable ACM support if ibacm.port is not found
Sean Hefty [Mon, 8 Oct 2012 17:33:21 +0000 (10:33 -0700)]
librdmacm: Disable ACM support if ibacm.port is not found

The librdmacm will try to connect port 6125 if ibacm.port is
not found.  The problem is that some other service or application
could be using that port and respond with garbage.  Rather
than falling back to a hard coded port number, if ibacm.port
is not found, simply disable ACM support.

This has the effect of removing support for older versions
of ibacm, unless the port file is created manually.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Mon, 8 Oct 2012 17:30:13 +0000 (10:30 -0700)]
pop

11 years agopop
Sean Hefty [Mon, 8 Oct 2012 17:30:13 +0000 (10:30 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:57:36 +0000 (12:57 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:57:36 +0000 (12:57 -0700)]
pop

11 years agocommit
Sean Hefty [Thu, 4 Oct 2012 19:57:33 +0000 (12:57 -0700)]
commit

11 years agocommit
Sean Hefty [Thu, 4 Oct 2012 19:57:33 +0000 (12:57 -0700)]
commit

11 years agocommit
Sean Hefty [Thu, 4 Oct 2012 19:57:32 +0000 (12:57 -0700)]
commit

11 years agocommit
Sean Hefty [Thu, 4 Oct 2012 19:57:32 +0000 (12:57 -0700)]
commit

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:56:26 +0000 (12:56 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:56:26 +0000 (12:56 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:51:05 +0000 (12:51 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:51:05 +0000 (12:51 -0700)]
pop

11 years agocmatose: Allow user to specify address format
Sean Hefty [Wed, 29 Aug 2012 22:02:54 +0000 (15:02 -0700)]
cmatose: Allow user to specify address format

Provide an option for the user to indicate the type of
addresses used as input.  Support hostname, IPv4, IPv6,
and GIDs.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:51:04 +0000 (12:51 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:51:04 +0000 (12:51 -0700)]
pop

11 years agolibrdmacm: Set connection information for AF_IB
Sean Hefty [Fri, 21 Sep 2012 22:38:15 +0000 (15:38 -0700)]
librdmacm: Set connection information for AF_IB

If an rdma_cm id is associated with AF_IB, then all private
data is exposed to the user.  Format the private data
according to the rdma_cm service annex.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agoucmatose: Remove connect parameter passed into rdma_accept
Sean Hefty [Thu, 4 Oct 2012 19:01:50 +0000 (12:01 -0700)]
ucmatose: Remove connect parameter passed into rdma_accept

Pass in NULL for conn_param into rdma_accept to indicate
that the passive side will use the values specified by the
active side.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:51:01 +0000 (12:51 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:51:01 +0000 (12:51 -0700)]
pop

11 years agoucmatose: Fix number of connections to disconnect
Sean Hefty [Thu, 4 Oct 2012 18:49:59 +0000 (11:49 -0700)]
ucmatose: Fix number of connections to disconnect

When ucmatose aborts because of issues trying to connect
to the server, it moves to disconnecting all connections.
However, not all connections may have been established.
The result is that ucmatose will hang in disconnect_events.
Fix this by setting the number of times that we need to
disconnect to the number of times that we successfully
connect.

This problem is based on a report by Doug Ledford
<dledford@redhat.com>

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:50:57 +0000 (12:50 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:50:57 +0000 (12:50 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:50:54 +0000 (12:50 -0700)]
pop

11 years agopop
Sean Hefty [Thu, 4 Oct 2012 19:50:54 +0000 (12:50 -0700)]
pop

11 years agorefresh
Sean Hefty [Thu, 4 Oct 2012 19:02:07 +0000 (12:02 -0700)]
refresh

11 years agorefresh
Sean Hefty [Thu, 4 Oct 2012 19:02:07 +0000 (12:02 -0700)]
refresh

11 years agorefresh (create temporary patch)
Sean Hefty [Thu, 4 Oct 2012 19:02:07 +0000 (12:02 -0700)]
refresh (create temporary patch)

11 years agorefresh (create temporary patch)
Sean Hefty [Thu, 4 Oct 2012 19:02:07 +0000 (12:02 -0700)]
refresh (create temporary patch)

11 years agoRefresh of ucmatose-accept
Sean Hefty [Thu, 4 Oct 2012 19:02:06 +0000 (12:02 -0700)]
Refresh of ucmatose-accept

11 years agoucmatose: Remove connect parameter passed into rdma_accept
Sean Hefty [Thu, 4 Oct 2012 19:01:50 +0000 (12:01 -0700)]
ucmatose: Remove connect parameter passed into rdma_accept

Pass in NULL for conn_param into rdma_accept to indicate
that the passive side will use the values specified by the
active side.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agonew
Sean Hefty [Thu, 4 Oct 2012 19:01:50 +0000 (12:01 -0700)]
new

11 years agonew
Sean Hefty [Thu, 4 Oct 2012 19:01:50 +0000 (12:01 -0700)]
new

11 years agoucmatose: Remove connect parameter passed into rdma_accept
Sean Hefty [Thu, 4 Oct 2012 19:01:50 +0000 (12:01 -0700)]
ucmatose: Remove connect parameter passed into rdma_accept

Pass in NULL for conn_param into rdma_accept to indicate
that the passive side will use the values specified by the
active side.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agorefresh
Sean Hefty [Thu, 4 Oct 2012 18:59:12 +0000 (11:59 -0700)]
refresh

11 years agorefresh
Sean Hefty [Thu, 4 Oct 2012 18:59:12 +0000 (11:59 -0700)]
refresh

11 years agorefresh (create temporary patch)
Sean Hefty [Thu, 4 Oct 2012 18:59:12 +0000 (11:59 -0700)]
refresh (create temporary patch)

11 years agorefresh (create temporary patch)
Sean Hefty [Thu, 4 Oct 2012 18:59:12 +0000 (11:59 -0700)]
refresh (create temporary patch)

11 years agoRefresh of ucmatose
Sean Hefty [Thu, 4 Oct 2012 18:59:11 +0000 (11:59 -0700)]
Refresh of ucmatose

11 years agorefresh
Sean Hefty [Thu, 4 Oct 2012 18:55:29 +0000 (11:55 -0700)]
refresh

11 years agorefresh
Sean Hefty [Thu, 4 Oct 2012 18:55:29 +0000 (11:55 -0700)]
refresh

11 years agorefresh (create temporary patch)
Sean Hefty [Thu, 4 Oct 2012 18:55:29 +0000 (11:55 -0700)]
refresh (create temporary patch)

11 years agorefresh (create temporary patch)
Sean Hefty [Thu, 4 Oct 2012 18:55:29 +0000 (11:55 -0700)]
refresh (create temporary patch)

11 years agoRefresh of ucmatose
Sean Hefty [Thu, 4 Oct 2012 18:55:29 +0000 (11:55 -0700)]
Refresh of ucmatose

11 years agoucmatose: Fix number of connections to disconnect
Sean Hefty [Thu, 4 Oct 2012 18:49:59 +0000 (11:49 -0700)]
ucmatose: Fix number of connections to disconnect

When ucmatose aborts because of issues trying to connect
to the server, it moves to disconnecting all connections.
However, not all connections may have been established.
The result is that ucmatose will hang in disconnect_events.
Fix this by setting the number of times that we need to
disconnect to the number of times that we successfully
connect.

This problem is based on a report by Doug Ledford
<dledford@redhat.com>

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agonew
Sean Hefty [Thu, 4 Oct 2012 18:49:59 +0000 (11:49 -0700)]
new

11 years agonew
Sean Hefty [Thu, 4 Oct 2012 18:49:59 +0000 (11:49 -0700)]
new

11 years agoucmatose: Fix number of connections to disconnect
Sean Hefty [Thu, 4 Oct 2012 18:49:59 +0000 (11:49 -0700)]
ucmatose: Fix number of connections to disconnect

When ucmatose aborts because of issues trying to connect
to the server, it moves to disconnecting all connections.
However, not all connections may have been established.
The result is that ucmatose will hang in disconnect_events.
Fix this by setting the number of times that we need to
disconnect to the number of times that we successfully
connect.

This problem is based on a report by Doug Ledford
<dledford@redhat.com>

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agocmatose: Allow user to specify address format
Sean Hefty [Wed, 29 Aug 2012 22:02:54 +0000 (15:02 -0700)]
cmatose: Allow user to specify address format

Provide an option for the user to indicate the type of
addresses used as input.  Support hostname, IPv4, IPv6,
and GIDs.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:38:45 +0000 (15:38 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:38:45 +0000 (15:38 -0700)]
pop

11 years agolibrdmacm: Set connection information for AF_IB
Sean Hefty [Fri, 21 Sep 2012 22:38:15 +0000 (15:38 -0700)]
librdmacm: Set connection information for AF_IB

If an rdma_cm id is associated with AF_IB, then all private
data is exposed to the user.  Format the private data
according to the rdma_cm service annex.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agocommit
Sean Hefty [Wed, 3 Oct 2012 22:38:41 +0000 (15:38 -0700)]
commit

11 years agocommit
Sean Hefty [Wed, 3 Oct 2012 22:38:41 +0000 (15:38 -0700)]
commit

11 years agocommit
Sean Hefty [Wed, 3 Oct 2012 22:38:40 +0000 (15:38 -0700)]
commit

11 years agocommit
Sean Hefty [Wed, 3 Oct 2012 22:38:40 +0000 (15:38 -0700)]
commit

11 years agocommit
Sean Hefty [Wed, 3 Oct 2012 22:38:39 +0000 (15:38 -0700)]
commit

11 years agocommit
Sean Hefty [Wed, 3 Oct 2012 22:38:39 +0000 (15:38 -0700)]
commit

11 years agorping: Reduce retry_count to fit in 3-bits
Sean Hefty [Wed, 3 Oct 2012 22:05:20 +0000 (15:05 -0700)]
rping: Reduce retry_count to fit in 3-bits

retry_count is a 3 bit value on IB, reduce it from
10 to 7.

A value of 10 prevents rping from working over the Intel
IB HCA.  Problem reported by Doug Ledford <dledford@redhat.com>

The retry_count is also not set when calling rdma_accept.
Rather than passing different values into rdma_accept than
what was specified by the remote side, use the values given
in the connection request.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:49 +0000 (15:21 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:49 +0000 (15:21 -0700)]
pop

11 years agolibrdmacm: Place container_of inside #ifdef
Sean Hefty [Sat, 22 Sep 2012 00:16:09 +0000 (17:16 -0700)]
librdmacm: Place container_of inside #ifdef

verbs.h defines container_of.  Only define it if not defined

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:46 +0000 (15:21 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:46 +0000 (15:21 -0700)]
pop

11 years agoaddrinfo: Remove debug printf calls
Sean Hefty [Wed, 3 Oct 2012 22:18:29 +0000 (15:18 -0700)]
addrinfo: Remove debug printf calls

These never should have made it into the commit.  :P

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:40 +0000 (15:21 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:40 +0000 (15:21 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:36 +0000 (15:21 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:21:36 +0000 (15:21 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:19:27 +0000 (15:19 -0700)]
pop

11 years agopop
Sean Hefty [Wed, 3 Oct 2012 22:19:27 +0000 (15:19 -0700)]
pop

11 years agolibrdmacm: Place container_of inside #ifdef
Sean Hefty [Sat, 22 Sep 2012 00:16:09 +0000 (17:16 -0700)]
librdmacm: Place container_of inside #ifdef

verbs.h defines container_of.  Only define it if not defined

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
11 years agorefresh
Sean Hefty [Wed, 3 Oct 2012 22:18:57 +0000 (15:18 -0700)]
refresh

11 years agorefresh
Sean Hefty [Wed, 3 Oct 2012 22:18:57 +0000 (15:18 -0700)]
refresh

11 years agorefresh (create temporary patch)
Sean Hefty [Wed, 3 Oct 2012 22:18:57 +0000 (15:18 -0700)]
refresh (create temporary patch)