Arlin Davis [Mon, 11 Feb 2008 22:50:33 +0000 (14:50 -0800)]
The OFA dapl provider is checking for incorrect return code
from rdma_bind_addr and rdma_listen calls. This causes an error
to be returned back to the consumer instead of correctly
incrementing the seed port and re-calling the OFA provider
until a valid port is issued. The seed value (1000) is also
incorrect and should start a non-privledged port (1024) to
avoid EPERM errors when seeding the starting port value.
Signed-off by: Arlin Davis <ardavis@ichips.intel.com>
Arlin Davis [Thu, 24 Jan 2008 00:40:48 +0000 (16:40 -0800)]
dapltest does not include definitions for inet_ntoa.
At load time the symbol was resolved but with the
default definition of int, instead of char*, it caused
segfault. Add correct include files in dapl_mdep_user.h
for linux.
Signed-off by: Arlin Davis <ardavis@ichips.intel.com>
Arlin Davis [Wed, 23 Jan 2008 22:25:21 +0000 (14:25 -0800)]
Fix for uDAPL v2 using extended operation. After extension completion,
the DTO cookie must be checked for type before deallocating to handle
inbound immediate data in receive. The sample dtestx client will fail
when running loopback if the rdma immediate is received from remote
before the rdma immediate request completion fires.
Bug causes following dtestx client error:
dat_ib_post_rdma_write_immed
returned DAT_INSUFFICIENT_RESOURCES : DAT_RESOURCE_MEMORY
Signed-off by: Arlin Davis <ardavis@ichips.intel.com>
Arlin Davis [Fri, 4 Jan 2008 16:32:39 +0000 (08:32 -0800)]
Windows specific -
IBAL support in evd_create
Build IB extensions by default
Common code -
check return status, evd_free, evd_wait
add dapl_event_str function
definitions for dat_os_library_error, dat_os_ungetc
Signed-off by: Stan Smith <stan.smith@intel.com>
Signed-off by: Arlin Davis <ardavis@ichips.intel.com>
Arlin Davis [Fri, 4 Jan 2008 16:31:27 +0000 (08:31 -0800)]
Windows specific -
Add dapl_ep fields ibal_cm_handle, recv_disc, sent_disc for IBAL provider
Support for direct object on CQ
INIT and FINI changes
setup dat.conf default path, fix sr parsing
Common code -
Add Stan Smith as contributor
O/S independent dat_os_library_error()
Signed-off by: Stan Smith <stan.smith@intel.com>
Signed-off by: Arlin Davis <ardavis@ichips.intel.com>
Arlin Davis [Tue, 11 Dec 2007 22:44:24 +0000 (14:44 -0800)]
2/2 uDAPL changes to sync common code base with WinOF
- add DAT_API to specify calling conventions (windows=__stdcall, linux= )
- cleanup platform specific definitions for windows
- c++ support
- add handle check macros DAT_IA_HANDLE_TO_UL and DAT_UL_TO_IA_HANDLE
Arlin Davis [Tue, 11 Dec 2007 22:43:05 +0000 (14:43 -0800)]
1/2 uDAT changes to sync common code base with WinOF
- add DAT_API to specify calling conventions (windows=__stdcall, linux= )
- cleanup platform specific definitions for windows
- c++ support
- add handle check macros DAT_IA_HANDLE_TO_UL and DAT_UL_TO_IA_HANDLE
Signed-off by: Arlin Davis <ardavis@ichips.intel.com>
Signed-off by: Stan Smith <stan.smith@intel.com>
Arlin Davis [Tue, 16 Oct 2007 21:23:17 +0000 (14:23 -0700)]
Fix dtest to build properly with extensions.
Modify dat.h dat_event to include event_extension_data[8].
Extend struct dat_event outside of extension build
switch to enable non-extended applications to work
with extended libraries. Otherwise, there is a potential
for the event callee to write back too much event data
and exceed callers non-extended event buffer.
Signed-off by: Arlin Davis <ardavis@ichips.intel.com>
Arlin Davis [Tue, 16 Oct 2007 21:10:52 +0000 (14:10 -0700)]
Use inet_ntoa instead of open coding it. IP addresses were being
reversed on PowerPC.
On PowerPC the timebase ticks at a different frequency to the CPU.
dapltest currently assumes a 1:1 relationship, and gives bogus values
when scaling timebase to real time.
To fix this, look at the timebase field in /proc/cpuinfo instead. To
keep things consistent with x86, scale that value to MHz.
Arlin Davis [Tue, 16 Oct 2007 21:09:56 +0000 (14:09 -0700)]
Use inet_ntoa instead of open coding it. IP addresses were being
reversed on PowerPC.
On PowerPC the timebase ticks at a different frequency to the CPU.
dapltest currently assumes a 1:1 relationship, and gives bogus values
when scaling timebase to real time.
To fix this, look at the timebase field in /proc/cpuinfo instead. To
keep things consistent with x86, scale that value to MHz.
Arlin Davis [Thu, 31 May 2007 19:43:28 +0000 (12:43 -0700)]
Add provider specific interface to uDAPL for extensions support.
Fix memory leak with extensions, missing cookie deallocation when processing extended DTO events
Remove unnecessary check on dats_set_ia_handle in dat_ia_openv
Clean up specfile and some extra LF's
Arlin Davis [Mon, 7 May 2007 19:49:18 +0000 (12:49 -0700)]
Fix issues with IB DTO extensions
- debug print_cqe updated for latest IBV definitions
- dapli_evd_cq_to_event modified to handle both post and receive extensions
- dtestx example tested with rdma_write_imm, fetch_add, and cmp_swap
- modify cookie DTO types to support extensions properly
Arlin Davis [Tue, 7 Nov 2006 20:22:05 +0000 (20:22 +0000)]
r10074: Added support for new ib verbs client register event. No extra processing required at the uDAPL
level.
Fix some issues supporting create qp without recv cq handle or recv qp resources. IB verbs assume a
recv_cq handle and uDAPL dapl_ep_create assumes there is always recv_sge resources specified.
Fix some timeout and long disconnect delay issues discovered during scale-out testing. Added support
to retry rdma_cm address and route resolution with configuration options. Provide a disconnect call
when receiving the disconnect request to guarantee a disconnect reply and event on the remote side.
The rdma_disconnect was not being called from dat_ep_disconnect() as a result of the state changing
to DISCONNECTED in the event callback.
Here are the new options (environment variables) with the default setting:
Steve Wise [Tue, 12 Sep 2006 18:15:39 +0000 (18:15 +0000)]
r9442: Update obsolete CLK_TCK to CLOCKS_PER_SEC Signed-off-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: James Lentini <jlentini@netapp.com>