InfiniBand updates
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 <[email protected]> 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.
commit 354ba39cf96e439149541acf3c6c7c0df0a3ef25
Author: John Kingman <[email protected]> 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).
commit 63aaf647529e8a56bdf31fd8f2979d4371c6a332
Author: Roland Dreier <[email protected]> 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.
commit 2e9f7cb7869059e55cd91f5e23c6380f3763db56
Author: Roland Dreier <[email protected]> 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.
commit fbed8eee70cf7e11fbf231afafc0ccb313acc62e
Author: Hal Rosenstock <[email protected]> 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.
commit 1325cc79163058739b70bed9860fccbecac6236b
Author: Hal Rosenstock <[email protected]> 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)
commit 17781cd6186cb3472ff34b2d9a15e647bd311e8b
Author: James Lentini <[email protected]> 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.
commit b5dcbf47e10e568273213a4410daa27c11cdba3a
Author: Hal Rosenstock <[email protected]> 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).
commit 30a7e8ef13b2ff0db7b15af9afdd12b93783f01e
Author: Michael S. Tsirkin <[email protected]> Date: Wed Sep 7 09:45:00 2005 -0700
[PATCH] IB: Initialize qp->wait
Add missing call to init_waitqueue_head().
commit c9fe2b3287498b80781284306064104ef9c8a31a
Author: Roland Dreier <[email protected]> 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.
commit 0b2b35f68140ceeb1b78ef85680198e63ebc8649
Author: Sean Hefty <[email protected]> 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.
commit 1d6801f9dd3ebb054ae685153a01b1a4ec817f46
Author: Michael S. Tsirkin <[email protected]> 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.
commit 06c56e44f3e32a859420ecac97996cc6f12827bb
Author: Michael S. Tsirkin <[email protected]> Date: Thu Sep 1 09:19:02 2005 -0700
[PATCH] IPoIB: fix memory leak
Fix IPoIB memory leak on device removal.
Quoted 5 lines · Roland Dreier
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?
Quoted 10 lines · Chris Wedgwood
Quoted 5 lines · Roland Dreier
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/inifinibandfor 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.
Quoted 8 lines
Quoted 5 lines
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/inifinibandfor 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.