commit e163bcb66b04bed7ab946a6f2cecc390d3940fb5 Author: Mitko Haralanov Date: Thu Jan 20 15:03:05 2011 -0500 IB/qib: Hold link for Tx SERDES settings Hold the IB link at DISABLED until we get the correct TX settings on mezz boards. Signed-off-by: Mike Marciniszyn commit dc76474c013e8d4d1e222e8c2cee3a2908424668 Author: Dan Carpenter Date: Thu Jan 20 10:48:07 2011 +0200 IB/uverbs: Handle large number of entries in poll CQ In ib_uverbs_poll_cq() code there is a potential integer overflow if userspace passes in a large cmd.ne. The calls to kmalloc() would allocate smaller buffers than intended, leading to memory corruption. There iss also an information leak if resp wasn't all used. Unprivileged userspace may call this function, although only if an RDMA device that uses this function is present. Fix this by copying CQ entries one at a time, which avoids the allocation entirely, and also by moving this copying into a function that makes sure to initialize all memory copied to userspace. Special thanks to Jason Gunthorpe for his help and advice. Cc: Signed-off-by: Dan Carpenter [ Monkey around with things a bit to avoid bad code generation by gcc when designated initializers are used. - Roland ] Signed-off-by: Roland Dreier commit e9dda1545111aaee070827baefae53a59105c106 Author: miroslaw.walukiewicz@intel.com Date: Wed Jan 19 12:56:26 2011 +0100 RDMA/core: backports for 2.6.35 and 2.6.36 kernels Added missing backports Signed-off-by: Mirek Walukiewicz commit e4b79f1ce8a28b300aa6df4def4a247752fbcdb2 Author: Vladimir Sokolovsky Date: Wed Jan 19 12:32:48 2011 +0200 mlx4,mthca: Change version to 1.5.3 Signed-off-by: Vladimir Sokolovsky commit 54bd87d97c9f878029231fb647897ac6b883ae1f Author: Vladimir Sokolovsky Date: Wed Jan 19 10:07:33 2011 +0200 openibd: unload cxgb3i on restart Signed-off-by: Vladimir Sokolovsky commit ba6fb97979aed79a54556de9a73f0d2527389f1a Author: Vladimir Sokolovsky Date: Tue Jan 18 13:03:36 2011 +0200 Fix support for kernels > 2.6.30 Replace dma_sync_single() use with proper functions dma_sync_single() is deprecated now, and the use in mthca is wrong: there should be a dma_sync_single_for_cpu() before touching the memory from the CPU, and a dma_sync_single_for_device() afterwards. Fix this, prompted by a kick in the pants from a patch from FUJITA Tomonori . Signed-off-by: Roland Dreier Signed-off-by: Vladimir Sokolovsky commit e75f1be772a7c48b0467a3cb98bb9731814be650 Author: Amir Vadai Date: Tue Jan 18 19:03:32 2011 +0200 sdp: do not reenter sdp_abort_rx_srcavail SrcAvailCancel could be handled from within sdp_abort_rx_srcavail. This shouldn't happen. To prevent it, ssk->rx_sa is set to NULL at the begining. Signed-off-by: Amir Vadai commit 925cbfc97f8395abc8ac04f463e676561dca8d16 Author: Amir Vadai Date: Sun Jan 16 18:32:38 2011 +0200 sdp: Abort rx SrcAvail when out of credits SrcAvail sink side could run out of credits when having bcopy sends to the other direction. Because of that no RdmaRdCompl could be sent nor SendSM. This could hang the sender side forever (No SendSM on SrcAvailCancel). Signed-off-by: Amir Vadai commit ea27c49e4fb320ab84645d57f375a05a8224109b Author: Vladimir Sokolovsky Date: Sun Jan 16 15:41:37 2011 +0200 cxgb3: Fix backport for RHEL5.6 Signed-off-by: Vladimir Sokolovsky commit c330eedc7149e70f94ca4a30e0be0657c97a2944 Author: Jack Morgenstein Date: Sun Jan 16 15:02:51 2011 +0200 mlx4: enable raw eth QPs to work with inline and blueflame. When using blueflame, an assigned qp number cannot contain "1" bits in positions 6 and 7 (which become the CS and SV bits in the second dword of the CTL segment when using blueflame). To solve this, we add a "skip_mask" dword to resource range allocation. The acquired resource is skipped if its number "and'ed" with the skip_mask is non_zero. For RAW_ETH qp's, we use a skip_mask of 0xC0, corresponding to bits 6 and 7 of the QP number. Signed-off-by: Jack Morgenstein commit f29ad409551c11d77656d00279dc1b49571b4add Author: Vladimir Sokolovsky Date: Thu Jan 13 16:24:36 2011 +0200 Add RHEL5.6 support Signed-off-by: Vladimir Sokolovsky commit 906bfc2bd29adce4fefa8600a505c9f8b2b04a8f Author: Vladimir Sokolovsky Date: Wed Jan 12 11:48:48 2011 +0200 tuning/sysctl: Rename ib_ipoib_sysctl -> sysctl_perf_tuning Tuning parameters are valid for all ULPs/network drivers. So, remove IPoIB dependency. Signed-off-by: Vladimir Sokolovsky commit 5a8f89849629b994934821a9f36c88e3bf6b29f1 Author: Vladimir Sokolovsky Date: Tue Jan 11 17:12:42 2011 +0200 openibd: add option to disable kernel parameters tuning Add RUN_SYSCTL parameter to the openib.conf. If RUN_SYSCTL=no the /sbin/ib_ipoib_sysctl will not be executed. Signed-off-by: Vladimir Sokolovsky commit 97f08a7e422516791900b060aa47493fce7d454c Author: Vladimir Sokolovsky Date: Tue Jan 11 17:00:58 2011 +0200 openibd: replace ifconfig by ip ifconfig should not be used to get/set network device properties. It was replaced by ip. BUG: http://bugs.openfabrics.org/bugzilla/show_bug.cgi?id=2209 Signed-off-by: Vladimir Sokolovsky commit 0ac3c8b53ed10101a1d7cef914a31dda100fe464 Merge: 449b58e 02e0103 Author: Vladimir Sokolovsky Date: Tue Jan 11 16:54:08 2011 +0200 Merge remote branch 'amirv/ofed_kernel_1_5' into ofed_kernel_1_5 commit 02e010336e34b3485a393bce5f462b029986cbe5 Author: Amir Vadai Date: Tue Jan 11 15:16:51 2011 +0200 sdp: Fixed BUG2207 - EINVAL when connect after IPv6 bind Connect to IPv4 over IPv6 address need rdma id to be created with IPv4 address. If bound before with IPv6 adderss, need to destroy and recreate the id. Also, when connecting after bind, keep the same source port number. Signed-off-by: Amir Vadai commit 449b58e2ef0be13b317ecaca7adc402c8ce213c5 Author: Vladimir Sokolovsky Date: Tue Jan 11 10:35:57 2011 +0200 Fix show_abi_version on 2.6.34 Signed-off-by: Vladimir Sokolovsky commit 38ffe0066bc8327256b8f73242906491fb7b2db3 Author: Vladimir Sokolovsky Date: Sun Jan 9 14:48:34 2011 +0200 sdp: Fix 2.6.36 backport patch Signed-off-by: Vladimir Sokolovsky commit f77da9f97403312c69373df8b4eabc1657751fad Author: Vladimir Sokolovsky Date: Sun Jan 9 14:28:54 2011 +0200 Add support for 2.6.18.8-xen kernel Signed-off-by: Vladimir Sokolovsky commit ce9d42bebe9729655a874a985bc6cb7b850232d9 Author: Amir Vadai Date: Thu Jan 6 15:49:55 2011 +0200 sdp: check address family before connecting Signed-off-by: Amir Vadai commit 9805faa6d787f379ad0c67528cfb636937f3829e Author: Amir Vadai Date: Thu Jan 6 14:38:34 2011 +0200 sdp: Do not ignore scope_id in IPv6 Signed-off-by: Amir Vadai commit e640d343c50c14b9eb3aab39dd407cb1d5292c26 Merge: 0148f04 b9cc54d Author: Vladimir Sokolovsky Date: Sun Jan 9 09:25:44 2011 +0200 Merge branch 'ofed_kernel_1_5' of git://git.openfabrics.org/~mmarciniszyn/linux-2.6.to_ofed into ofed_kernel_1_5 commit b9cc54dc33cb21b3b32037131aa53ccdd37e6dfa Author: Mike Marciniszyn Date: Thu Jan 6 16:33:25 2011 -0500 IB/qib: fix refcount leak in lkey/rkey validation The mr optimization introduced a reference count leak on an exception test. The lock/refcount manipulation is moved down and the problematic exception test now calls bail to insure that the lock is released. Additional fixes as suggested by Ralph Campbell : - reduce lock scope of dma regions - use explicit values on returns vs. automatic ret value Signed-off-by: Mike Marciniszyn commit 0148f048919082023c1dd94f00149462a6e04ed6 Merge: d74a5ae 708221f Author: Vladimir Sokolovsky Date: Wed Jan 5 15:54:13 2011 +0200 Merge remote branch 'amirv/ofed_kernel_1_5' into ofed_kernel_1_5 commit 708221f03c7f110350691746c6d4f4ae28611fb2 Author: Amir Vadai Date: Wed Jan 5 13:26:32 2011 +0200 sdp: make backport sk_inet()->xxx simpler Signed-off-by: Amir Vadai commit d74a5ae3cd0471770b5d5293b16cd92d6b058dda Author: Vladimir Sokolovsky Date: Wed Jan 5 14:17:07 2011 +0200 ppc64/backport: Add workaround for compilation on PPC64 with 2.6.32 Add an empty asm/iomap.h under kernel_addons for 2.6.32* kernels BUG: http://bugs.openfabrics.org/bugzilla/show_bug.cgi?id=2203 Signed-off-by: Vladimir Sokolovsky commit 890e7721d4c2c19a4c7a9b6d23a6502182ec476b Author: Amir Vadai Date: Wed Jan 5 11:33:21 2011 +0200 sdp: Fix get getsockname/getpeername in IPv6 Signed-off-by: Amir Vadai commit 8effad4bafd8a98e49e44d64859951bbc536e031 Merge: a61000b 9fdd445 Author: Vladimir Sokolovsky Date: Wed Jan 5 10:58:39 2011 +0200 Merge branch 'ofed_kernel_1_5' of git://git.openfabrics.org/~mmarciniszyn/linux-2.6.to_ofed into ofed_kernel_1_5 commit 9fdd445e04aac7cbe3fa3780e385bce4fa9d5db1 Author: Mike Marciniszyn Date: Tue Jan 4 15:19:07 2011 -0500 ofa_kernel.spec: Account for QLogic cards for counting ipoib aliases The routine count_ib_ports() now lspci's for QLogic SDR, DDR, and QDR HCAs to correctly configure ib_ipoib.conf aliases. The aliases are also corrected for upgrade installs as well as initial installs. Signed-off-by: Mike Marciniszyn commit a61000bf2180da935306819d981fdc00aa9386b3 Author: Alexander Schmidt Date: Mon Jan 3 14:10:14 2011 +0100 ehca: fixes The changes are tested with ./ofed_scripts/ofed_makedist.sh Signed-off-by: Alexander Schmidt commit 6855440dd8cfab0a0d82ab33c1d3979b4b5ce443 Merge: 03dc015 9887f6b Author: Vladimir Sokolovsky Date: Tue Jan 4 17:46:29 2011 +0200 Merge branch 'ofed_kernel_1_5' of git://git.openfabrics.org/~mmarciniszyn/linux-2.6.to_ofed into ofed_kernel_1_5 commit 9887f6b42a367e7a19534aa8f24bbe12c6191a9f Author: Mitko Haralanov Date: Mon Jan 3 15:40:43 2011 -0500 truescale.cmds: balance IRQs for kernel contexts across CPUs Balance the IRQs of the kernel receive contexts across the available CPUs by assigning IRQ affinities in a round-robin fashion. Signed-off-by: Mike Marciniszyn commit eeb4091b076e3bfd224a2ba70b376150fa9b64c8 Author: Mitko Haralanov Date: Mon Jan 3 15:38:49 2011 -0500 truescale.cmds: Fix setting SERDES Tx values on HP Mezz If the mezz card is in the second mezz slot, both ports are enabled. Therefore, we need to set the correct Tx settings for the second port. Signed-off-by: Mike Marciniszyn commit 7971d9be1cdd03b2c8f0928cbc11e24fd8e3e699 Author: Mitko Haralanov Date: Mon Jan 3 15:34:30 2011 -0500 truescale.cmds: Fix invalid 'if' statement syntax Delete leading -a. Signed-off-by: Mike Marciniszyn commit 16d386a4ed37befb3ccfcdd3d176a08017aeedf6 Author: Mike Marciniszyn Date: Mon Jan 3 15:30:18 2011 -0500 truescale.cmds: make the mount of /ipathfs more robust and quiet The 'start' section of truescale.cmds isn't terribly robust against being called multiple times or having a slightly wonky environment. Device testers in particular were getting messages from the failing mkdir. Signed-off-by: Mike Marciniszyn commit 03dc0154a1aa21b335d5df251010bace7bfc62a3 Author: Amir Vadai Date: Sun Jan 2 13:31:12 2011 +0200 sdp: Allow bind to address with family AF_INETx_SDP Signed-off-by: Amir Vadai commit 46e20629818ba040ce1ea83df701496844aa9c23 Author: Amir Vadai Date: Thu Dec 30 14:38:48 2010 +0200 sdp: Use %pI4 + %pI6 in new kernels A patch in this subject was in the wrong backport. Moving it to a dedicated patch. Signed-off-by: Amir Vadai commit 954d37f69074eb23f70e848f11fd2af089f800ba Author: Amir Vadai Date: Wed Dec 29 17:23:07 2010 +0200 sdp: fix connect to IPv4 over IPv6 Signed-off-by: Amir Vadai commit de706b6e1c3d9503e650411878b25cf13d51b519 Author: Amir Vadai Date: Wed Dec 29 12:40:52 2010 +0200 sdp: Fix some issues in ipv6 support Signed-off-by: Amir Vadai commit 0ff5467c40c0ec3e5e25e46946900ea98a04d279 Author: Amir Vadai Date: Mon Dec 27 11:58:26 2010 +0200 sdp: ipv6 support Signed-off-by: Amir Vadai commit 2e0ab2acaef2528a9ca94692af08e550fedac691 Author: Mike Marciniszyn Date: Mon Dec 27 15:04:30 2010 -0500 IB/qib: Code style issues found during kernel.org submission testing These are fixes to style issues detected by checkpatch.pl. Signed-off-by: Mike Marciniszyn commit cf8b6ca317a31ccee24e0058415db36a59242eb6 Merge: 3435de1 b17abc2 Author: Vladimir Sokolovsky Date: Sun Dec 26 09:39:46 2010 +0200 Merge branch 'ofed_kernel_1_5' of git://git.openfabrics.org/~mmarciniszyn/linux-2.6.to_ofed into ofed_kernel_1_5 commit b17abc2b2164948531f8fc591a9ff6e3d971cdd1 Author: Mike Marciniszyn Date: Wed Dec 22 12:59:51 2010 -0500 IB/qib: Add return value testing for work queue creation The code ignored potentially NULL rvals. This patch adds testing for NULL and the appropriate failure branches at the end of qib_init(). Signed-off-by: Mike Marciniszyn commit 53027eea12923aaa3ba11961a486af80d9c0ce73 Author: Mike Marciniszyn Date: Wed Dec 22 12:56:40 2010 -0500 IB/qib: Allow driver to load if PCIe advanced error reporting fails Some PCIe root complex chip sets don't support advanced error reporting. Allow the driver to load OK if pci_enable_pcie_error_reporting() fails. Signed-off-by: Mike Marciniszyn commit cbb1b059540f802f0c65573629bf41fdf3a2eb2e Author: Mike Marciniszyn Date: Wed Dec 22 12:53:25 2010 -0500 IB/qib: clean up properly if pci_set_consistent_dma_mask() fails Clean up properly when pci_set_consistent_dma_mask() fails. Signed-off-by: Mike Marciniszyn commit 56fde9902048fec37e81dc8be3fb7f774d0dbc8f Author: Mike Marciniszyn Date: Wed Dec 22 12:48:01 2010 -0500 IB/qib: Fix QIB early error reporting prints Replace incorrect use of dev_info with dev_err. Signed-off-by: Mike Marciniszyn commit 6a1024778af0b7c4b7f1e0075548fe1a75867f26 Author: Mike Marciniszyn Date: Wed Dec 22 12:43:41 2010 -0500 IB/qib: Improve SERDES tunning on QMH boards Improve the QMH SERDES tunning on initial driver load by removing the delay after SERDES init and having the driver go through a link state change. Signed-off-by: Mike Marciniszyn commit 4714a7ae7734059bd6b63186bb7bccd1a6cde441 Author: Mike Marciniszyn Date: Wed Dec 22 12:39:31 2010 -0500 IB/qib: Add a delay to allow SERDES to initialize Adding a small delay would allow the SERDES to adapt better, which would improve the link negotiation. Signed-off-by: Mike Marciniszyn commit 825356532db07e34632a9aa9a64451b75201eeb1 Author: Mike Marciniszyn Date: Wed Dec 22 12:32:45 2010 -0500 IB/qib: Un-necessary delayed completions on RC connection Currently on receipt of a response message (ACKs, RDMA Response, Atomic Responses etc.) if the SDMA completion counter is not advanced the driver delays the completion of the WQE. In most cases this is overly pessimistic as the response (ACK) to a previously transmitted send implies that the send is complete. Ensure that SDMA queue is progressed appropriately before determining if a send has delayed completions. Signed-off-by: Mike Marciniszyn commit deac6a9658f1e4121fa0e26899702323c715e8f2 Author: Mike Marciniszyn Date: Wed Dec 22 12:25:33 2010 -0500 IB/qib: Issue pre-emptive NAKs on eager buffer overflow Under congestion resulting in eager buffer overflow attempt to send pre-emptive NAKs if header queue entries with TID errors are generated and a valid header is present. This prevents long timeouts and flow restarts if a trailing set of packets are dropped due to eager overflows. Pre-emptive NAKs are currently only supported for RDMA Writes. Signed-off-by: Mike Marciniszyn commit e2f3b28bbe27e34a1abfa7bccb8377c42535c910 Author: Mike Marciniszyn Date: Wed Dec 22 12:16:48 2010 -0500 IB/qib: RDMA lkey/rkey validation is inefficient for large MRs The current code loops during rkey/lkey validiation to isolate the mr for the RDMA, which is expensive when the current operation is inside a very large memory region. This fix optimizes rkey/lkey validation routines for user memory regions and fast memory regions. The mr entry can be isolated by shifts/mods instead of looping. The existing loop is preserved for phys memory regions for now. Signed-off-by: Mike Marciniszyn commit 9ad35a43699e7d77c1a1081e184afb79c0df0dd5 Author: Mike Marciniszyn Date: Wed Dec 22 12:07:50 2010 -0500 IB/qib: change qpn increment Changing from +1 to +2 allows for better QP distribution across receive contexts. Signed-off-by: Mike Marciniszyn commit 55333e094dd97f75e4e007f03183f71527bdf962 Author: Mike Marciniszyn Date: Wed Dec 22 11:33:43 2010 -0500 IB/qib: backport changes for last two commits Signed-off-by: Mike Marciniszyn commit 07933e2d064eb0f0eadee8f4fa7b634334207518 Author: Mike Marciniszyn Date: Wed Dec 22 11:29:43 2010 -0500 IB/qib: adding fix missing from earlier patch 0669da697c7779955806ef3dae3db4ecf68618cc was missing part of a fix that makes merges impossible. Signed-off-by: Mike Marciniszyn commit 0e0ccb83fb114887d2abfafffbb5bb5a576ec2f0 Author: Mike Marciniszyn Date: Wed Dec 22 11:28:56 2010 -0500 IB/qib: change rcv queue/qpn selection The basic idea is that on SusieQ, the difficult part of mapping QPN to context is handled by the mapping registers so the generic QPN allocation doesn't need to worry about chip specifics. For Monty and Linda, there is no mapping table so the qpt->mask (same as dd->qpn_mask), is used to see if the QPN to context falls within [zero..dd->n_krcv_queues). Signed-off-by: Mike Marciniszyn commit 8f448d7d90a50ab0cc32a06e0811c702c5007c78 Author: Mike Marciniszyn Date: Tue Dec 21 13:01:59 2010 -0500 IB/qib: interrupt mitigation fix For SusieQ we need to write to the interrupt timer register before updating the header queue head with interrupt count. This is to ensure that the timer is enabled properly and a receive available interrupt is delivered else this interrupt can be lost if the receiver header/eager queues are full before the timer is enabled. Signed-off-by: Mike Marciniszyn commit 684272b5a7a72508a4c70023b2099793fb7948c5 Author: Mike Marciniszyn Date: Tue Dec 21 12:40:12 2010 -0500 IB/qib: Avoid duplicate writes to the rcv head register Avoid duplicate writes to the head register as this can lead to lost interrupts if the context goes full before the second write is done. Signed-off-by: Mike Marciniszyn commit bd512a70873e18d098ab940472f81693af71b578 Author: Mike Marciniszyn Date: Tue Dec 21 12:35:31 2010 -0500 IB/qib: fix RDMA write with immediate The immediate word for RDMA_WRITE_ONLY_WITH_IMMEDIATE was being extracted from the wrong location in the header. Thanks to Jason Gunthorpe for finding this. Signed-off-by: Mike Marciniszyn commit 7bc68c54f2fa74632b15225a1cd946b463b90e77 Author: Mike Marciniszyn Date: Tue Dec 21 12:29:48 2010 -0500 IB/qib: Add a few new SERDES tunings Add new SERDES tuning to aid manufacturing. Signed-off-by: Mike Marciniszyn commit 1b4acc545ee7065acaf9e5f56e290a24bcf54467 Author: Mike Marciniszyn Date: Tue Dec 21 12:22:32 2010 -0500 IB/qib: Reset packet list after freeing Reset the list pointers after freeing the SDMA packet list. This is done to any potential double-free cases. Signed-off-by: Mike Marciniszyn commit 3252ff781558691d6e09f567456db55912099404 Author: Mike Marciniszyn Date: Tue Dec 21 12:15:00 2010 -0500 IB/qib: New SERDES init routine and improvements to SI quality Implement new SERDES initialization routine and improvements to SI - disable LE1 adaptation, disable LOS after link-up, set better SERDES parameters. Signed-off-by: Mike Marciniszyn commit 37644c3c66a9c7cf6559aa0735d92b673731efa0 Author: Mike Marciniszyn Date: Tue Dec 21 12:03:46 2010 -0500 IB/qib: remove IB latency turnoff This is required for hardware testing. Signed-off-by: Mike Marciniszyn commit 55d63e9ee381cf842f56d57c950d3aa76cb6880b Author: Mike Marciniszyn Date: Tue Dec 21 10:37:43 2010 -0500 IB/qib: clear WAIT_SEND flags when setting QP to error state If these flags are set when the QP is transitioned to the error state, it will wait until the flags are cleared, which may never happen if the error transition is due to a link going down. Signed-off-by: Mike Marciniszyn commit f834687c62a99e884ff915725222fe7e2d69edc9 Author: Mike Marciniszyn Date: Tue Dec 21 10:36:04 2010 -0500 IB/qib: fix context allocation with multiple HCAs The driver was incorrectly choosing HCAs on which to allocate new user contexts based on overall count of usable ports regardless whether the usable port was on the currently selected HCA. Signed-off-by: Mike Marciniszyn commit c881784aa894d97a5e41ab18cdaa0a2a47fee178 Author: Mike Marciniszyn Date: Tue Dec 21 10:34:57 2010 -0500 IB/qib: Multi Florida HCA Hosts panic on reboot Add check when setting configured contexts that the value does not exceed the number of contexts allocated for the card. If the value exceeds the already allocated count, set it to what is already allocated. Signed-off-by: Mike Marciniszyn commit addaddd141739e6d4a69ff54c0ae12aa8ba4e110 Author: Mike Marciniszyn Date: Tue Dec 21 10:33:50 2010 -0500 IB/qib: Handle transitions from ACTIVE_DEFERRED to ACTIVE better When the link transitions from ACTIVE_DEFERRED to ACTIVE, the driver only sees the ACTIVE state. With this change, it will check whether the state was already ACTIVE and if so, it will not generated IB events and will not clear symbol error counts. Signed-off-by: Mike Marciniszyn commit fc430b5cc66fa0a23f3ba879a0c94c21938536fa Author: Mike Marciniszyn Date: Tue Dec 21 10:32:12 2010 -0500 IB/qib: UD send with immediate Rx completion has wrong size The code to generate receive completion entries for UD send with immediate contains the wrong payload length. This is because when the code to compute the payload size was moved, the value of hdrsize didn't get moved too. The fix is to update tlen directly. Signed-off-by: Mike Marciniszyn commit 0a9f81750d0c9909c4b95a2de2d430daf754d020 Author: Mike Marciniszyn Date: Tue Dec 21 10:30:40 2010 -0500 IB/qib: set port physical state even if other fields are invalid The IBTA vol. 1 release 1.2.1 spec. says: C14-24.2.1: If PortInfo:Portstate=Down, then a SubnSet(PortInfo) shall make any changes it specifies to PortInfo:PortPhysicalState; any other result is vendor-dependent. The patch changes the error handling so that the reply says there are invalid fields but still attempts to set fields that are in range including PortInfo:PortPhysicalState. Signed-off-by: Mike Marciniszyn commit 85dff982123cdb426b69b0a41e1101cf91227352 Author: Mike Marciniszyn Date: Tue Dec 21 10:29:28 2010 -0500 IB/qib: generate completion callback on errors According to IBTA vol. 1, C11-30.1.1, a notification callback is invoked if the CQ is armed for the next solicited completion event or an error completion. The error case wasn't being generated previously. Signed-off-by: Mike Marciniszyn commit 466c85baa925b006c659b0c26e109bf1e4151d99 Author: Mike Marciniszyn Date: Tue Dec 21 10:28:47 2010 -0500 IB/qib: add thresholds to VendorPortCounters PMA operation This adds thresholds to be specified in the VendorPortCounters PMA operation. Signed-off-by: Mike Marciniszyn commit 4b9e0ac717070c3e8628b9926b7e3a430c3f3541 Author: Mike Marciniszyn Date: Tue Dec 21 10:27:02 2010 -0500 IB/qib: add support for the new QME7362 card This patch adds support to recognize another board variation named QME7362. Signed-off-by: Mike Marciniszyn commit fc6238f5dfcc0e5e369a5a65ec46708cebafa895 Author: Mike Marciniszyn Date: Tue Dec 21 10:24:59 2010 -0500 IB/qib: add receive header queue size module parameters The receive header queue sizes need to modified to support performance tuning. Three module parameters are added to support this. Signed-off-by: Mike Marciniszyn commit 5be565611feaf040e50142339000379cf4790dc3 Author: Mike Marciniszyn Date: Tue Dec 21 10:24:05 2010 -0500 IB/qib: set cfgctxts to number of CPUs by default Up to now, we have set the number of available user contexts based on the number of hardware contexts which is set according to the number of available CPUs. This was fine since most CPUs had a power of two number of cores and the chip supported 4, 8, or 16 user contexts. Now that some systems have 12 cores, the default isn't optimal and should be set to 12 even though 16 hardware contexts need to be enabled. Signed-off-by: Mike Marciniszyn