Commit Graph

65 Commits

Author SHA1 Message Date
Michael S. Tsirkin
23f3bc0f2c IB/mthca: Fix posting lists of 256 receive requests for Tavor
If we post a list of length 256 exactly, nreq in doorbell gets set to
256 which is wrong: it should be encoded by 0.  This is because we
only zero it out on the next WR, which may not be there.  The solution
is to ring the doorbell after posting a WQE, not before posting the
next one.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-05-18 11:37:03 -07:00
Roland Dreier
a3285aa4ee IB/mthca: Fix race in reference counting
Fix races in in destroying various objects.  If a destroy routine
waits for an object to become free by doing

	wait_event(&obj->wait, !atomic_read(&obj->refcount));
	/* now clean up and destroy the object */

and another place drops a reference to the object by doing

	if (atomic_dec_and_test(&obj->refcount))
		wake_up(&obj->wait);

then this is susceptible to a race where the wait_event() and final
freeing of the object occur between the atomic_dec_and_test() and the
wake_up().  And this is a use-after-free, since wake_up() will be
called on part of the already-freed object.

Fix this in mthca by replacing the atomic_t refcounts with plain old
integers protected by a spinlock.  This makes it possible to do the
decrement of the reference count and the wake_up() so that it appears
as a single atomic operation to the code waiting on the wait queue.

While touching this code, also simplify mthca_cq_clean(): the CQ being
cleaned cannot go away, because it still has a QP attached to it.  So
there's no reason to be paranoid and look up the CQ by number; it's
perfectly safe to use the pointer that the callers already have.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-05-09 10:50:29 -07:00
Jack Morgenstein
bf6a9e31cf IB: simplify static rate encoding
Push translation of static rate to HCA format into low-level drivers,
where it belongs.  For static rate encoding, use encoding of rate
field from IB standard PathRecord, with addition of value 0, for
backwards compatibility with current usage.  The changes are:

 - Add enum ib_rate to midlayer includes.
 - Get rid of static rate translation in IPoIB; just use static rate
   directly from Path and MulticastGroup records.
 - Update mthca driver to translate absolute static rate into the
   format used by hardware.  This also fixes mthca's static rate
   handling for HCAs that are capable of 4X DDR.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-04-10 09:43:47 -07:00
Roland Dreier
e1f7868c80 IB/mthca: Fix section mismatch problems
Quite a few cleanup functions in mthca were marked as __devexit.
However, they could also be called from error paths during
initialization, so they cannot be marked that way.  Just delete all of
the incorrect annotations.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-29 09:36:46 -08:00
Roland Dreier
192daa18dd IB/mthca: Fix modify QP error path
If the call to mthca_MODIFY_QP() failed, then mthca_modify_qp() would
still do some things it shouldn't, such as store away attributes for
special QPs.  Fix this, and simplify the code, by simply jumping to
the exit path if mthca_MODIFY_QP() fails.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-24 15:47:30 -08:00
Roland Dreier
b0b3a8e193 IB/mthca: Fix indentation
Fix some whitespace damage (indenting with spaces) that snuck in.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-24 15:47:29 -08:00
Jack Morgenstein
b3f64967fa IB/mthca: Fix uninitialized variable in mthca_alloc_qp()
mthca_alloc_sqp() by mthca_set_qp_size() need to set qp->transport
before calling mthca_set_qp_size(), since the value is used there.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-24 15:47:29 -08:00
Dotan Barak
0ef61db837 IB/mthca: Check that sgid_index and path_mtu are valid in modify_qp
Add a check that the modify QP parameters sgid_index and path_mtu are
valid, since they might come from userspace.

Signed-off-by: Dotan Barak <dotanb@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-24 15:47:27 -08:00
Dotan Barak
67e7377661 IB/mthca: Check alternate P_Key index when setting alternate path
Check that the alternate P_Key index is in range when setting the
alternate path for a QP.  Also make a cosmetic touch up to the debug
message printed when the main P_Key index is out of range.

Signed-off-by: Dotan Barak <dotanb@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-20 10:08:18 -08:00
Dotan Barak
7667abd152 IB/mthca: Add support for send work request fence flag
Add support for IB_SEND_FENCE flag in post_send methods.

Signed-off-by: Dotan Barak <dotanb@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-20 10:08:18 -08:00
Eli Cohen
8ebe5077e3 IB/mthca: Support for query QP and SRQ
Implement the query_qp and query_srq methods in mthca.

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-20 10:08:15 -08:00
Roland Dreier
d844183d9c IB/mthca: Convert to use ib_modify_qp_is_ok()
Use ib_modify_qp_is_ok() in mthca, and delete the big table of
attributes for queue pair state transitions.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-20 10:08:13 -08:00
Roland Dreier
3fa1fa3e80 IB/mthca: Generate SQ drained events when requested
Add low-level driver support to ib_mthca so that consumers can request
a "send queue drained" event be generated when a transiton to the SQD
state completes.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-20 10:08:12 -08:00
Roland Dreier
2fa5e2ebbe IB/mthca: Whitespace cleanups
Remove trailing whitespace and fix indentation that with spaces
instead of tabs.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-20 10:08:09 -08:00
Roland Dreier
d9b98b0f11 IB/mthca: Make functions that never fail return void
The function mthca_free_err_wqe() can never fail, so get rid of its
return value.  That means handle_error_cqe() doesn't have to check
what mthca_free_err_wqe() returns, which means it can't fail either
and doesn't have to return anything either.  All this results in
simpler source code and a slight object code improvement:

add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-10 (-10)
function                                     old     new   delta
mthca_free_err_wqe                            83      81      -2
mthca_poll_cq                               1758    1750      -8

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-03-20 10:08:07 -08:00
Michael S. Tsirkin
9eacee2ac6 IB/mthca: Initialize grh_present before using it
build_mlx_header() was using sqp->ud_header.grh_present before it was
initialized by mthca_read_ah().  Furthermore, header->grh_present is
set by ib_ud_header_init, so there's no need to set it again in
mthca_read_ah().

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-01-12 15:55:41 -08:00
Dotan Barak
4de144bf72 IB/mthca: Add support for automatic path migration (APM)
Add code to modify QP operation to handle setting alternate paths for
connected QPs.

Signed-off-by: Dotan Barak <dotanb@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-01-06 13:23:58 -08:00
Jack Morgenstein
0d3b525fff IB/mthca: fix for RTR-to-RTS transition in modify QP
PKEY_INDEX is not a legal parameter in the RTR->RTS transition.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-01-06 13:03:43 -08:00
Jack Morgenstein
0364ffc3e8 IB/mthca: fix for SQEr-to-RTS transition in modify QP
Fixes to SQEr->RTS transition in modify_qp:
1. The flag IB_QP_ACCESS_FLAGS is optional for UC qps
2. The SQEr state is not supported for RC qps

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-01-06 13:01:27 -08:00
Jack Morgenstein
5b3bc7a681 IB/mthca: max_inline_data handling tweaks
Fix a case where copying max_inline_data from a successful create_qp
capabilities output to create_qp input could cause EINVAL error:

mthca_set_qp_size must check max_inline_data directly against
max_desc_sz; checking qp->sq.max_gs is wrong since max_inline_data
depends on the qp type and does not involve max_sg.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-01-06 12:57:30 -08:00
Jack Morgenstein
38d1e79347 IB/mthca: check port validity in modify_qp
Modify_qp should check that the physical port number provided
is a legal value.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-01-05 16:13:46 -08:00
Jack Morgenstein
c4342d8a4d IB/mthca: Fix corner cases in max_rd_atomic value handling in modify QP
sae and sre bits should only be set when setting sra_max.  Further, in
the old code, if the caller specifies max_rd_atomic = 0, the sre and
sae bits are still set, with the result that the QP ends up with
max_rd_atomic = 1 in effect.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-12-15 19:59:01 -08:00
Jack Morgenstein
d1646f86a2 IB/mthca: Fix IB_QP_ACCESS_FLAGS handling.
This patch corrects some corner cases in managing the RAE/RRE bits in
the mthca qp context.  These bits need to be zero if the user requests
max_dest_rd_atomic of zero.  The bits need to be restored to the value
implied by the qp access flags attribute in a previous (or the
current) modify-qp command if the dest_rd_atomic variable is changed
to non-zero.

In the current implementation, the following scenario will not work:
RESET-to-INIT 	set QP access flags to all disabled (zeroes)
INIT-to-RTR     set max_dest_rd_atomic=10, AND
		set qp_access_flags = IB_ACCESS_REMOTE_READ | IB_ACCESS_REMOTE_ATOMIC

The current code will incorrectly take the access-flags value set in
the RESET-to-INIT transition.

We can simplify, and correct, this IB_QP_ACCESS_FLAGS handling: it is
always safe to set qp access flags in the firmware command if either
of IB_QP_MAX_DEST_RD_ATOMIC or IB_QP_ACCESS_FLAGS is set, so let's
just set it to the correct value, always.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-12-15 14:36:24 -08:00
Jack Morgenstein
44b5b03033 IB/mthca: don't change driver's copy of attributes if modify QP fails
Only change the driver's copy of the QP attributes in modify QP after
checking the modify QP command completed successfully.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-12-09 16:40:14 -08:00
Jack Morgenstein
6aa2e4e806 IB/mthca: correct log2 calculation
Fix thinko in rd_atomic calculation: ffs(x) - 1 does not find the next
power of 2 -- it should be fls(x - 1).

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-12-09 16:38:04 -08:00
Jack Morgenstein
94361cf74a IB/mthca: check RDMA limits
Add limit checking on rd_atomic and dest_rd_atomic attributes:
especially for max_dest_rd_atomic, a value that is larger than HCA
capability can cause RDB overflow and corruption of another QP.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-12-09 16:32:21 -08:00
Michael S. Tsirkin
e0ae9ecf46 IB/mthca: fix posting of send lists of length >= 255 on mem-free HCAs
On mem-free HCAs, when posting a long list of send requests, a
doorbell must be rung every 255 requests.  Add code to handle this.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-29 11:33:46 -08:00
Michael S. Tsirkin
187a25863f IB/mthca: reset QP's last pointers when transitioning to reset state
last pointer is not updated when QP is modified to reset state.  This
causes data corruption if WQEs are already posted on the queue.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-28 11:19:43 -08:00
Michael S. Tsirkin
48fd0d1fdd IB/mthca: Safer max_send_sge/max_recv_sge calculation
Calculation of QP capabilities still isn't exactly right in mthca:
max_send_sge/max_recv_sge fields returned in create_qp can exceed the
handware supported limits.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-18 14:11:17 -08:00
Roland Dreier
cbc5b2bb9e [IB] mthca: don't disable RDMA writes if no responder resources
Responder resources are only required to handle RDMA reads and atomic
operations, not RDMA writes.  So the driver should allow RDMA writes
even if responder resources are set to 0.  This is especially
important for the UC transport -- with the old code, it was impossible
to enable RDMA writes for UC QPs.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-15 00:24:23 -08:00
Michael S. Tsirkin
ae57e24a40 [IB] mthca: fix posting long lists of receive work requests
In Tavor mode, when posting a long list of receive work requests, a
doorbell must be rung every 256 requests.  Add code to do this when
required.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-10 10:22:51 -08:00
Michael S. Tsirkin
62abb8416f [IB] mthca: fix posting of atomic operations
The size of work requests for atomic operations was computed
incorrectly in mthca: all sizeofs need to be divided by 16.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-10 10:22:50 -08:00
Jack Morgenstein
77369ed31d [IB] uverbs: have kernel return QP capabilities
Move the computation of QP capabilities (max scatter/gather entries,
max inline data, etc) into the kernel, and have the uverbs module
return the values as part of the create QP response.  This keeps
precise knowledge of device limits in the low-level kernel driver.

This requires an ABI bump, so while we're making changes, get rid of
the max_sge parameter for the modify SRQ command -- it's not used and
shouldn't be there.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-10 10:22:50 -08:00
Linus Torvalds
ba77df570c Merge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband 2005-11-04 16:31:54 -08:00
Jack Morgenstein
d09e327641 [IB] mthca: check P_Key index in modify QP
Make sure that the P_Key index passed into mthca_modify_qp() is
within the device's P_Key table.

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-11-03 14:58:33 -08:00
Tim Schmielau
4e57b68178 [PATCH] fix missing includes
I recently picked up my older work to remove unnecessary #includes of
sched.h, starting from a patch by Dave Jones to not include sched.h
from module.h. This reduces the number of indirect includes of sched.h
by ~300. Another ~400 pointless direct includes can be removed after
this disentangling (patch to follow later).
However, quite a few indirect includes need to be fixed up for this.

In order to feed the patches through -mm with as little disturbance as
possible, I've split out the fixes I accumulated up to now (complete for
i386 and x86_64, more archs to follow later) and post them before the real
patch.  This way this large part of the patch is kept simple with only
adding #includes, and all hunks are independent of each other.  So if any
hunk rejects or gets in the way of other patches, just drop it.  My scripts
will pick it up again in the next round.

Signed-off-by: Tim Schmielau <tim@physik3.uni-rostock.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-30 17:37:32 -08:00
Roland Dreier
547e309073 [IB] mthca: correct modify QP attribute masks for UC
The UC transport does not support RDMA reads or atomic operations, so
we shouldn't require or even allow the consumer to set attributes
relating to these operations for UC QPs.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-10-25 10:57:32 -07:00
Jack Morgenstein
efaae8f71f [IB] mthca: Better limit checking and reporting
Check the sizes of CQs, QPs and SRQs when creating objects, and fail
instead of creating too-big queues.  Also return real limits instead
of just plausible-sounding values from mthca_query_device().

Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-10-17 15:20:29 -07:00
Michael S. Tsirkin
71eea47d85 [PATCH] IB/mthca: Fix device removal memory leak
Clean up QP table array on device removal.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-09-20 10:54:48 -07:00
Roland Dreier
d6cff021e2 [PATCH] IB/mthca: fix posting of first work request
Fix posting first WQE for mem-free HCAs: we need to link to previous
WQE even in that case.  While we're at it, simplify code for
Tavor-mode HCAs.  We don't really need the conditional test there
either; we can similarly always link to the previous WQE.

Based on Michael S. Tsirkin's analogous fix for userspace libmthca.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-09-18 22:02:37 -07:00
Roland Dreier
bb4a7f0da7 [PATCH] IB/mthca: assign ACK timeout field correctly
The hardware reads the ACK timeout field from the most significant 5
bits of struct mthca_qp_path's ackto field, not the least significant
bits.  This fix has the driver put the timeout in the right place.
Without this, we get a timeout that is 2^8 times too small.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-09-18 22:02:37 -07:00
Michael S. Tsirkin
30a7e8ef13 [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>
2005-09-07 09:48:53 -07:00
Roland Dreier
c9fe2b3287 [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>
2005-09-07 09:48:52 -07:00
Roland Dreier
a4d61e8480 [PATCH] IB: move include files to include/rdma
Move the InfiniBand headers from drivers/infiniband/include to include/rdma.
This allows InfiniBand-using code to live elsewhere, and lets us remove the
ugly EXTRA_CFLAGS include path from the InfiniBand Makefiles.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-08-26 20:37:38 -07:00
Roland Dreier
ec34a922d2 [PATCH] IB/mthca: Add SRQ implementation
Add mthca support for shared receive queues (SRQs),
including userspace SRQs.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-08-26 20:37:37 -07:00
Roland Dreier
c04bc3d1f4 [PATCH] IB/mthca: Move WQE structures into their own header
Move the definitions of the WQE structures from mthca_qp.c into
mthca_wqe.h, so that we'll be able to share them when we add the
SRQ code in mthca_srq.c.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-08-26 20:37:37 -07:00
Roland Dreier
288bdeb4bc [PATCH] IB/mthca: Simplify handling of completions with error
Mem-free HCAs never generate error CQEs that complete multiple WQEs,
so just skip the call to mthca_free_err_wqe() for them rather than
having logic to handle the mem-free case in mthca_free_err_wqe().

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-08-26 20:37:37 -07:00
Roland Dreier
87b816706b [PATCH] IB/mthca: Factor out common queue alloc code
Clean up the allocation of memory for queues by factoring out the
common code into mthca_buf_alloc() and mthca_buf_free().  Now CQs and
QPs share the same queue allocation code, which we'll also use for SRQs.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-08-26 20:37:37 -07:00
Roland Dreier
da6561c285 [PATCH] IB/mthca: Use correct port width capability value
When we call the INIT_IB firmware command to bring up a port, use
the actual port width capability returned by the QUERY_DEV_LIM
command instead of always trying to enable both 1X and 4X.  This
fixes breakage seen when the firmware is build to allow 4X only.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-08-26 20:37:36 -07:00
Sean Hefty
97f52eb438 [PATCH] IB: sparse endianness cleanup
Fix sparse warnings.  Use __be* where appropriate.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2005-08-26 20:37:35 -07:00