* [git pull] InfiniBand updates
@ 2005-09-10 23:02 Roland Dreier
2005-09-11 3:03 ` Chris Wedgwood
0 siblings, 1 reply; 5+ messages in thread
From: Roland Dreier @ 2005-09-10 23:02 UTC (permalink / raw)
To: torvalds; +Cc: linux-kernel, openib-general
Linus, please pull from
master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband.git
This tree is also available from kernel.org mirrors at:
rsync://rsync.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git
This will update the following files:
drivers/infiniband/Kconfig | 25 +-
drivers/infiniband/core/Makefile | 5
drivers/infiniband/core/cm.c | 5
drivers/infiniband/core/mad_rmpp.c | 4
drivers/infiniband/core/sa_query.c | 30 ---
drivers/infiniband/core/ucm.c | 289 +++++++++++++++++++-----------
drivers/infiniband/core/ucm.h | 11 -
drivers/infiniband/core/uverbs.h | 26 +-
drivers/infiniband/core/uverbs_cmd.c | 155 +++++++++++-----
drivers/infiniband/core/uverbs_main.c | 98 ++++++----
drivers/infiniband/hw/mthca/mthca_qp.c | 45 +++-
drivers/infiniband/ulp/ipoib/ipoib_main.c | 2
include/rdma/ib_cm.h | 1
include/rdma/ib_mad.h | 21 ++
include/rdma/ib_sa.h | 31 +++
include/rdma/ib_user_cm.h | 72 +++++++
include/rdma/ib_user_verbs.h | 21 ++
17 files changed, 590 insertions(+), 251 deletions(-)
through the following changes:
commit 1b205c2d2464bfecbba80227e74b412596dc5521
Author: Roland Dreier <roland@eddore.topspincom.com>
Date: Fri Sep 9 20:52:00 2005 -0700
[PATCH] IB: fix CM use-after-free
If the CM REQ handling function gets to error2, then it frees
cm_id_priv->timewait_info. But the next line goes through
ib_destroy_cm_id() -> ib_send_cm_rej() -> cm_reset_to_idle(),
which ends up calling cm_cleanup_timewait(), which dereferences the
pointer we just freed. Make sure we clear cm_id_priv->timewait_info
after freeing it, so that doesn't happen.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 354ba39cf96e439149541acf3c6c7c0df0a3ef25
Author: John Kingman <kingman@storagegear.com>
Date: Fri Sep 9 18:23:32 2005 -0700
[PATCH] IB CM: support CM redir
Changes to CM to support CM and port redirection (REJ reason 24).
Signed-off-by: John Kingman <kingman <at> storagegear.com>
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 63aaf647529e8a56bdf31fd8f2979d4371c6a332
Author: Roland Dreier <roland@eddore.topspincom.com>
Date: Fri Sep 9 15:55:08 2005 -0700
Make sure that userspace does not retrieve stale asynchronous or
completion events after destroying a CQ, QP or SRQ. We do this by
sweeping the event lists before returning from a destroy calls, and
then return the number of events already reported before the destroy
call. This allows userspace wait until it has processed all events
for an object returned from the kernel before it frees its context for
the object.
The ABI of the destroy CQ, destroy QP and destroy SRQ commands has to
change to return the event count, so bump the ABI version from 1 to 2.
The userspace libibverbs library has already been updated to handle
both the old and new ABI versions.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 2e9f7cb7869059e55cd91f5e23c6380f3763db56
Author: Roland Dreier <roland@eddore.topspincom.com>
Date: Fri Sep 9 15:45:57 2005 -0700
[PATCH] IB: Add struct for ClassPortInfo
Add structure definition for ClassPortInfo format. This is
needed for (at least) handling CM redirects.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit fbed8eee70cf7e11fbf231afafc0ccb313acc62e
Author: Hal Rosenstock <halr@voltaire.com>
Date: Fri Sep 9 15:24:04 2005 -0700
[PATCH] IB: Move SA attributes to ib_sa.h
SA: Move SA attributes to ib_sa.h so are accessible to more than
sa_query.c. Also, remove deprecated attributes and add one missing one.
Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 1325cc79163058739b70bed9860fccbecac6236b
Author: Hal Rosenstock <halr@voltaire.com>
Date: Fri Sep 9 13:45:51 2005 -0700
[PATCH] IB: Define more SA methods
ib_sa.h: Define more SA methods (initially for madeye decode)
Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 17781cd6186cb3472ff34b2d9a15e647bd311e8b
Author: James Lentini <jlentini@netapp.com>
Date: Wed Sep 7 12:43:08 2005 -0700
[PATCH] IB: clean up user access config options
Add a new config option INFINIBAND_USER_MAD to control whether we
build ib_umad. Change INFINIBAND_USER_VERBS to INFINIBAND_USER_ACCESS,
and have it control ib_ucm and ib_uat as well as ib_uverbs.
Signed-off-by: James Lentini <jlentini@netapp.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit b5dcbf47e10e568273213a4410daa27c11cdba3a
Author: Hal Rosenstock <halr@voltaire.com>
Date: Wed Sep 7 11:03:41 2005 -0700
[PATCH] IB: RMPP fixes
- Fix payload length of middle RMPP sent segments. Middle payload
lengths should be 0 on the send side.
(This is perhaps a compliance and should not be an interop issue as
middle payload lengths are supposed to be ignored on receive).
- Fix length in first segment of multipacket sends
(This is a compliance issue but does not affect at least OpenIB to
OpenIB RMPP transfers).
Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 30a7e8ef13b2ff0db7b15af9afdd12b93783f01e
Author: Michael S. Tsirkin <mst@mellanox.co.il>
Date: Wed Sep 7 09:45:00 2005 -0700
[PATCH] IB: Initialize qp->wait
Add missing call to init_waitqueue_head().
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit c9fe2b3287498b80781284306064104ef9c8a31a
Author: Roland Dreier <roland@eddore.topspincom.com>
Date: Wed Sep 7 09:43:23 2005 -0700
[PATCH] IB: really reset QPs
When we modify a QP to the RESET state, completely clean up the QP
so that it is really and truly reset.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 0b2b35f68140ceeb1b78ef85680198e63ebc8649
Author: Sean Hefty <sean.hefty@intel.com>
Date: Thu Sep 1 09:28:03 2005 -0700
[PATCH] IB: Add user-supplied context to userspace CM ABI
- Add user specified context to all uCM events. Users will not retrieve
any events associated with the context after destroying the corresponding
cm_id.
- Provide the ib_cm_init_qp_attr() call to userspace clients of the CM.
This call may be used to set QP attributes properly before modifying the QP.
- Fixes some error handling synchonization and cleanup issues.
- Performs some minor code cleanup.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 1d6801f9dd3ebb054ae685153a01b1a4ec817f46
Author: Michael S. Tsirkin <mst@mellanox.co.il>
Date: Thu Sep 1 09:19:44 2005 -0700
[PATCH] IB/sa_query: avoid unnecessary list scan
Using ib_get_client_data in SA event handler performs a list scan.
It's better to use container_of to get the sa device directly.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
commit 06c56e44f3e32a859420ecac97996cc6f12827bb
Author: Michael S. Tsirkin <mst@mellanox.co.il>
Date: Thu Sep 1 09:19:02 2005 -0700
[PATCH] IPoIB: fix memory leak
Fix IPoIB memory leak on device removal.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [git pull] InfiniBand updates
2005-09-10 23:02 [git pull] InfiniBand updates Roland Dreier
@ 2005-09-11 3:03 ` Chris Wedgwood
2005-09-11 9:00 ` Christoph Hellwig
2005-09-11 15:15 ` Roland Dreier
0 siblings, 2 replies; 5+ messages in thread
From: Chris Wedgwood @ 2005-09-11 3:03 UTC (permalink / raw)
To: Roland Dreier; +Cc: torvalds, linux-kernel, openib-general
On Sat, Sep 10, 2005 at 04:02:12PM -0700, Roland Dreier wrote:
> include/rdma/ib_cm.h | 1
> include/rdma/ib_mad.h | 21 ++
> include/rdma/ib_sa.h | 31 +++
> include/rdma/ib_user_cm.h | 72 +++++++
> include/rdma/ib_user_verbs.h | 21 ++
Do these really need to be here? if we really must merge RDMA can we
not hide these headers in drivers/inifiniband for now?
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [git pull] InfiniBand updates
2005-09-11 3:03 ` Chris Wedgwood
@ 2005-09-11 9:00 ` Christoph Hellwig
2005-09-11 15:15 ` Roland Dreier
1 sibling, 0 replies; 5+ messages in thread
From: Christoph Hellwig @ 2005-09-11 9:00 UTC (permalink / raw)
To: Chris Wedgwood; +Cc: Roland Dreier, torvalds, linux-kernel, openib-general
On Sat, Sep 10, 2005 at 08:03:45PM -0700, Chris Wedgwood wrote:
> On Sat, Sep 10, 2005 at 04:02:12PM -0700, Roland Dreier wrote:
>
> > include/rdma/ib_cm.h | 1
> > include/rdma/ib_mad.h | 21 ++
> > include/rdma/ib_sa.h | 31 +++
> > include/rdma/ib_user_cm.h | 72 +++++++
> > include/rdma/ib_user_verbs.h | 21 ++
>
> Do these really need to be here? if we really must merge RDMA can we
> not hide these headers in drivers/inifiniband for now?
No. They've been there before, but it's just wrong. This stuff is
kernel-wide interfaces and having them under drivers/ was wrong to start
with.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [git pull] InfiniBand updates
2005-09-11 3:03 ` Chris Wedgwood
2005-09-11 9:00 ` Christoph Hellwig
@ 2005-09-11 15:15 ` Roland Dreier
1 sibling, 0 replies; 5+ messages in thread
From: Roland Dreier @ 2005-09-11 15:15 UTC (permalink / raw)
To: Chris Wedgwood; +Cc: torvalds, linux-kernel, openib-general
> > include/rdma/ib_cm.h | 1
> > include/rdma/ib_mad.h | 21 ++
> > include/rdma/ib_sa.h | 31 +++
> > include/rdma/ib_user_cm.h | 72 +++++++
> > include/rdma/ib_user_verbs.h | 21 ++
> Do these really need to be here? if we really must merge RDMA can we
> not hide these headers in drivers/inifiniband for now?
The includes were moved from drivers/infiniband a few weeks ago for
various good reasons.
I really wish you had replied to the initial RFC
(http://lkml.org/lkml/2005/8/4/191) or the merge where the headers
were actually moved (http://lkml.org/lkml/2005/8/29/105). I don't
think there's much point in moving the files back now.
- R.
^ permalink raw reply [flat|nested] 5+ messages in thread
* [git pull] InfiniBand updates
@ 2006-01-08 22:25 Roland Dreier
0 siblings, 0 replies; 5+ messages in thread
From: Roland Dreier @ 2006-01-08 22:25 UTC (permalink / raw)
To: torvalds; +Cc: linux-kernel, openib-general
Linus, please pull from
master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband.git for-linus
This tree is also available from kernel.org mirrors at:
git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git for-linus
The pull will get the following changes:
Dotan Barak:
IB/mthca: Add support for automatic path migration (APM)
Jack Morgenstein:
IB/mthca: fix QP size limits for mem-free HCAs
IB/umad: fix memory leaks
IB/mthca: fix memory user DB table leak
IB/mthca: check RDMA limits
IB/mthca: correct log2 calculation
IB/mthca: don't change driver's copy of attributes if modify QP fails
IB/mthca: Fix SRQ cleanup during QP destroy
IB/mthca: Fix IB_QP_ACCESS_FLAGS handling.
IB/mthca: Fix corner cases in max_rd_atomic value handling in modify QP
IB/mthca: fix WQE size calculation in create-srq
IB/mthca: check return value in mthca_dev_lim call
IB/mthca: check port validity in modify_qp
IB/mthca: max_inline_data handling tweaks
IB/mthca: fix for SQEr-to-RTS transition in modify QP
IB/mthca: fix for RTR-to-RTS transition in modify QP
IB/mthca: multiple fixes for multicast group handling
IB/uverbs: Fix reference counting on error paths
IB/uverbs: Release event file reference on ib_uverbs_create_cq() error
Michael S. Tsirkin:
IB/mthca: Fix thinko in mthca_table_find()
IB/mthca: create_eq with size not a power of 2
IB/mthca: Fill in vendor_err field in completion with error
Ralph Campbell:
IB/uverbs: set ah_flags when creating address handle
IB: Set GIDs correctly in ib_create_ah_from_wc()
Sean Hefty:
IB/cm: correct reported reject code
IB/cm: avoid reusing local ID
drivers/infiniband/core/cm.c | 16 +-
drivers/infiniband/core/user_mad.c | 4
drivers/infiniband/core/uverbs_cmd.c | 21 ++
drivers/infiniband/core/verbs.c | 4
drivers/infiniband/hw/mthca/mthca_cmd.c | 12 +
drivers/infiniband/hw/mthca/mthca_cq.c | 23 ++
drivers/infiniband/hw/mthca/mthca_eq.c | 4
drivers/infiniband/hw/mthca/mthca_main.c | 4
drivers/infiniband/hw/mthca/mthca_mcg.c | 54 ++++--
drivers/infiniband/hw/mthca/mthca_memfree.c | 4
drivers/infiniband/hw/mthca/mthca_qp.c | 265 +++++++++++++++------------
drivers/infiniband/hw/mthca/mthca_srq.c | 2
12 files changed, 250 insertions(+), 163 deletions(-)
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2006-01-08 22:25 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-09-10 23:02 [git pull] InfiniBand updates Roland Dreier
2005-09-11 3:03 ` Chris Wedgwood
2005-09-11 9:00 ` Christoph Hellwig
2005-09-11 15:15 ` Roland Dreier
-- strict thread matches above, loose matches on Subject: below --
2006-01-08 22:25 Roland Dreier
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox