aboutsummaryrefslogtreecommitdiff
path: root/drivers/infiniband/core
AgeCommit message (Collapse)Author
2006-09-22RDMA: iWARP Connection Manager.Tom Tucker
Add an iWARP Connection Manager (CM), which abstracts connection management for iWARP devices (RNICs). It is a logical instance of the xx_cm where xx is the transport type (ib or iw). The symbols exported are used by the transport independent rdma_cm module, and are available also for transport dependent ULPs. Signed-off-by: Tom Tucker <tom@opengridcomputing.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB: Whitespace fixesRoland Dreier
Remove some trailing whitespace that has snuck in despite the best efforts of whitespace=error-all. Also fix a few other whitespace bogosities. Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/cm: Randomize starting comm IDSean Hefty
Randomize the starting local comm ID to avoid getting a rejected connection due to a stale connection after a system reboot or reloading of the ib_cm. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/mad: Remove unused includesJames Lentini
The ib_mad module does not use a kthread function, but mad_priv.h includes <linux/kthread.h>. mad_rmpp.c does not do any DMA-related stuff, but includes <linux/dma-mapping.h>. Remove the unused includes. Signed-off-by: James Lentini <jlentini@netapp.com> Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/mad: Add support for dual-sided RMPP transfers.Sean Hefty
The implementation assumes that any RMPP request that requires a response uses DS RMPP. Based on the RMPP start-up scenarios defined by the spec, this should be a valid assumption. That is, there is no start-up scenario defined where an RMPP request is followed by a non-RMPP response. By having this assumption we avoid any API changes. In order for a node that supports DS RMPP to communicate with one that does not, RMPP responses assume a new window size of 1 if a DS ACK has not been received. (By DS ACK, I'm referring to the turn-around ACK after the final ACK of the request.) This is a slight spec deviation, but is necessary to allow communication with nodes that do not generate the DS ACK. It also handles the case when a response is sent after the request state has been discarded. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/cm: Use correct reject code for invalid GIDSean Hefty
Set the reject code properly when rejecting a request that contains an invalid GID. A suitable GID is returned by the IB CM in the additional reject information (ARI). This is a spec compliancy issue. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/cm: Enable atomics along with RDMA readsSean Hefty
Enable atomic operations along with RDMA reads if a local RDMA read/atomic depth is provided by the user. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/uverbs: Pass userspace data to modify_srq and modify_qp methodsRalph Campbell
Pass a struct ib_udata to the low-level driver's ->modify_srq() and ->modify_qp() methods, so that it can get to the device-specific data passed in by the userspace driver. Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/uverbs: Allow resize CQ operation to return driver-specific dataRalph Campbell
Add a ib_uverbs_resize_cq_resp.driver_data field so that low-level drivers can return data from a resize CQ operation to userspace. Have ib_uverbs_resize_cq() only copy the cqe field, to avoid having to bump the userspace ABI. Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/uverbs: Fix lockdep warning when QP is created with 2 CQsRoland Dreier
Lockdep warns when userspace creates a QP that uses different CQs for send completions and receive completions, because both CQs are locked and their mutexes belong to the same lock class. However, we know that the mutexes are distinct and the nesting is safe (there is no possibility of AB-BA deadlock because the mutexes are locked with down_read()), so annotate the situation with SINGLE_DEPTH_NESTING to get rid of the lockdep warning. Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-22IB/uverbs: Use idr_read_cq() where appropriateRoland Dreier
There were two functions that open-coded idr_read_cq() in terms of idr_read_uobj() rather than using the helper. Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-09-14RDMA/cma: Increase the IB CM retry count in CMAMichael S. Tsirkin
3 seems like a low number of IB Communication Manager retries to set; we see connections failing under stress, and in any case 3 just looks like an arbitrary number. 15 is the max value allowed by the InfiniBand spec. Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Acked-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-08-16IB/core: Fix SM LID/LID change with client reregister setJack Morgenstein
After commit 12bbb2b7be7f5564952ebe0196623e97464b8ac5, when SM LID change or LID change MAD also has a client reregistration bit set, only CLIENT_REREGISTER event is generated. As a result, the sa_query module and the cache module don't update the port information, and ULPs (e.g. IPoIB) stop working. This is the regression we observe as compared to 2.6.17. Rather than generate multiple events (which would have negative performance impact), let us simply let cache and SA query respond to reregister event in the same way as to LID and SM change events. 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>
2006-08-03IB/uverbs: Avoid a crash on device hot removeJack Morgenstein
Wait until all users have closed their device context before allowing device unregistration to complete. This prevents a crash caused by referring to stale data structures. A better solution would be to have a way to revoke contexts rather than waiting for userspace to close the context, but that's a much bigger change that will have to wait. For now let's at least avoid the crash. 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>
2006-08-03IB/cm: Fix error handling in ib_send_cm_reqSean Hefty
Report error code rather than success (0) on failure allocating timewait_info in ib_send_cm_req(). Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-08-02[NET] infiniband: Cleanup ib_addr module to use the neteventsTom Tucker
Signed-off-by: Tom Tucker <tom@opengridcomputing.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-24IB/mad: Validate MADs for spec complianceSean Hefty
Validate MADs sent by userspace clients for spec compliance with C13-18.1.1 (prevent duplicate requests and responses sent on the same port). Without this, RMPP transactions get aborted because of duplicate packets. This patch is similar to that provided by Jack Morgenstein. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-07-23IB/uverbs: Fix lockdep warningsRoland Dreier
Lockdep warns because uverbs is trying to take uobj->mutex when it already holds that lock. This is because there are really multiple types of uobjs even though all of their locks are initialized in common code. Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-07-23IB/uverbs: Fix unlocking in error pathsMichael S. Tsirkin
ib_uverbs_create_ah() and ib_uverbs_create_srq() did not release the PD's read lock in their error paths, which lead to deadlock when destroying the PD. Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-07-14[PATCH] IB/core: use correct gfp_mask in sa_queryMichael S. Tsirkin
Avoid bogus out of memory errors: fix sa_query to actually pass gfp_mask supplied by the user to idr_pre_get. Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Acked-by: "Sean Hefty" <mshefty@ichips.intel.com> Acked-by: "Roland Dreier" <rdreier@cisco.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-14[PATCH] fmr pool: remove unnecessary pointer dereferenceMichael S. Tsirkin
ib_fmr_pool_map_phys gets the virtual address by pointer but never writes there, and users (e.g. srp) seem to assume this and ignore the value returned. This patch cleans up the API to get the VA by value, and updates all users. Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Acked-by: Roland Dreier <rolandd@cisco.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-14[PATCH] IB/cm: set private data length for reject messagesIra Weiny
Set private data length for reject messages to the correct size. Fix from openib svn r8483. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Cc: Roland Dreier <rolandd@cisco.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-14[PATCH] IB/addr: gid structure alignment fixMichael S. Tsirkin
The device address contains unsigned character arrays, which contain raw GID addresses. The GIDs may not be naturally aligned, so do not cast them to structures or unions. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Cc: Roland Dreier <rolandd@cisco.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-14[PATCH] IB/cm: drop REQ when out of memoryMichael S. Tsirkin
If a user of the IB CM returns -ENOMEM from their connection callback, simply drop the incoming REQ - do not attempt to send a reject. This should allow the sender to retry the request. Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Signed-off-by: Sean Hefty <sean.hefty@intel.com> Cc: Roland Dreier <rolandd@cisco.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-30IB/core: Set alternate port number when initializing QP attributesSean Hefty
Set alternate port number when initializing QP attributes. This bug is OpenFabrics bugzilla bug #160. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-30IB/uverbs: Set correct user handle for user SRQsRoland Dreier
Store away the user handle passed in from userspace when creating an SRQ, so that the kernel can return the correct handle when an SRQ asynchronous event occurs. (A 0 was incorrectly stored as the user handle as part of the changes in 9ead190b, "IB/uverbs: Don't serialize with ib_uverbs_idr_mutex") Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-26[PATCH] Array overrun in drivers/infiniband/core/cma.cEric Sesterhenn
This was spotted by coverity #id 1300. Since the array has only four elements, we should just use those four. Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de> Acked-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-26[PATCH] drivers: use list_move()Akinobu Mita
This patch converts the combination of list_del(A) and list_add(A, B) to list_move(A, B) under drivers/. Acked-by: Corey Minyard <minyard@mvista.com> Cc: Ben Collins <bcollins@debian.org> Acked-by: Roland Dreier <rolandd@cisco.com> Cc: Alasdair Kergon <dm-devel@redhat.com> Cc: Gerd Knorr <kraxel@bytesex.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Frank Pavlic <fpavlic@de.ibm.com> Acked-by: Matthew Wilcox <matthew@wil.cx> Cc: Andrew Vasquez <linux-driver@qlogic.com> Cc: Mikael Starvik <starvik@axis.com> Cc: Greg Kroah-Hartman <greg@kroah.com> Signed-off-by: Akinobu Mita <mita@miraclelinux.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-25Merge branch 'for-linus' of ↵Linus Torvalds
master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband * 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband: IB/iser: iSER Kconfig and Makefile IB/iser: iSER handling of memory for RDMA IB/iser: iSER RDMA CM (CMA) and IB verbs interaction IB/iser: iSER initiator iSCSI PDU and TX/RX IB/iser: iSCSI iSER transport provider high level code IB/iser: iSCSI iSER transport provider header file IB/uverbs: Remove unnecessary list_del()s IB/uverbs: Don't free wr list when it's known to be empty
2006-06-23[PATCH] VFS: Permit filesystem to override root dentry on mountDavid Howells
Extend the get_sb() filesystem operation to take an extra argument that permits the VFS to pass in the target vfsmount that defines the mountpoint. The filesystem is then required to manually set the superblock and root dentry pointers. For most filesystems, this should be done with simple_set_mnt() which will set the superblock pointer and then set the root dentry to the superblock's s_root (as per the old default behaviour). The get_sb() op now returns an integer as there's now no need to return the superblock pointer. This patch permits a superblock to be implicitly shared amongst several mount points, such as can be done with NFS to avoid potential inode aliasing. In such a case, simple_set_mnt() would not be called, and instead the mnt_root and mnt_sb would be set directly. The patch also makes the following changes: (*) the get_sb_*() convenience functions in the core kernel now take a vfsmount pointer argument and return an integer, so most filesystems have to change very little. (*) If one of the convenience function is not used, then get_sb() should normally call simple_set_mnt() to instantiate the vfsmount. This will always return 0, and so can be tail-called from get_sb(). (*) generic_shutdown_super() now calls shrink_dcache_sb() to clean up the dcache upon superblock destruction rather than shrink_dcache_anon(). This is required because the superblock may now have multiple trees that aren't actually bound to s_root, but that still need to be cleaned up. The currently called functions assume that the whole tree is rooted at s_root, and that anonymous dentries are not the roots of trees which results in dentries being left unculled. However, with the way NFS superblock sharing are currently set to be implemented, these assumptions are violated: the root of the filesystem is simply a dummy dentry and inode (the real inode for '/' may well be inaccessible), and all the vfsmounts are rooted on anonymous[*] dentries with child trees. [*] Anonymous until discovered from another tree. (*) The documentation has been adjusted, including the additional bit of changing ext2_* into foo_* in the documentation. [akpm@osdl.org: convert ipath_fs, do other stuff] Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: Al Viro <viro@zeniv.linux.org.uk> Cc: Nathan Scott <nathans@sgi.com> Cc: Roland Dreier <rolandd@cisco.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-22IB/uverbs: Remove unnecessary list_del()sRoland Dreier
In ib_uverbs_cleanup_ucontext(), when iterating through the lists of objects, there's no reason to do list_del() to remove the objects, since both the objects and the lists that contain them are about to be freed anyway. Since list_del() is a moderately big inline function, getting rid of this extra work saves quite a bit of .text: add/remove: 0/0 grow/shrink: 1/2 up/down: 3/-217 (-214) function old new delta ib_uverbs_comp_handler 225 228 +3 ib_uverbs_async_handler 256 255 -1 ib_uverbs_close 905 689 -216 Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-22IB/uverbs: Don't free wr list when it's known to be emptyKrishna Kumar
In ib_uverbs_post_send(), move the "out:" label after the loop that frees the list of work requests, since the only place that jumps there is before any work requests could possibly be added to the list. This removes a compile warning: "is_ud might be used uninitialized in this function". Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/uverbs: Don't serialize with ib_uverbs_idr_mutexRoland Dreier
Currently, all userspace verbs operations that call into the kernel are serialized by ib_uverbs_idr_mutex. This can be a scalability issue for some workloads, especially for devices driven by the ipath driver, which needs to call into the kernel even for datapath operations. Fix this by adding reference counts to the userspace objects, and then converting ib_uverbs_idr_mutex into a spinlock that only protects the idrs long enough to take a reference on the object being looked up. Because remove operations may fail, we have to do a slightly funky two-step deletion, which is described in the comments at the top of uverbs_cmd.c. This also still leaves ib_uverbs_idr_lock as a single lock that is possibly subject to contention. However, the lock hold time will only be a single idr operation, so multiple threads should still be able to make progress, even if ib_uverbs_idr_lock is being ping-ponged. Surprisingly, these changes even shrink the object code: add/remove: 23/5 grow/shrink: 4/21 up/down: 633/-693 (-60) Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/uverbs: Factor out common idr codeRoland Dreier
Factor out common code for adding a userspace object to an idr into a function idr_add_uobj(). This shrinks both the source and object code: add/remove: 1/0 grow/shrink: 0/6 up/down: 57/-220 (-163) function old new delta idr_add_uobj - 57 +57 ib_uverbs_create_ah 543 512 -31 ib_uverbs_create_srq 662 630 -32 ib_uverbs_reg_mr 737 699 -38 ib_uverbs_create_cq 639 600 -39 ib_uverbs_alloc_pd 485 446 -39 ib_uverbs_create_qp 1020 979 -41 Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/uverbs: Don't decrement usecnt on error pathsRoland Dreier
In error paths when destroying an object, uverbs should not decrement associated objects' usecnt, since ib_dereg_mr(), ib_destroy_qp(), etc. already do that. Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/uverbs: Release lock on error pathGanapathi CH
If ibdev->alloc_ucontext() fails then ib_uverbs_get_context() does not unlock file->mutex before returning error. Signed-off by: Ganapathi CH <cganapathi@novell.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/cm: Use address handle helpersSean Hefty
Use new ib_init_ah_from_wc() and ib_init_ah_from_path() helper functions to clean up the IB CM. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/sa: Add ib_init_ah_from_path()Sean Hefty
Add a call to initialize address handle attributes given a path record. This is used by the CM, and would be useful for users of UD QPs. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB: Add ib_init_ah_from_wc()Sean Hefty
Add a function to initialize address handle attributes from a work completion. This functionality is duplicated by both verbs and the CM. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/ucm: Get rid of duplicate P_Key parameterSean Hefty
The P_Key is provided into a SIDR REQ in two places, once as a parameter, and again in the path record. Remove the P_Key as a parameter and always use the one given in the path record. This change has no practical effect on ABI functionality. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/fmr: Use device's max_map_map_per_fmr attribute in FMR pool.Or Gerlitz
When creating a FMR pool, query the IB device and use the returned max_map_map_per_fmr attribute as for the max number of FMR remaps. If the device does not suport querying this attribute, use the original IB_FMR_MAX_REMAPS (32) default. Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/mad: Check GID/LID when matching requestsJack Morgenstein
Check GID/LID for requester side when searching for request which matches received response. This is in order to guarantee uniqueness if the same TID is used when requesting via multiple source LIDs (when LMC is not zero). Use ports' cached LMC to perform the check. Further, do not perform LID check for direct-routed packets, since the permissive LID makes a proper check impossible. Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il> Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB: Add caching of ports' LMCJack Morgenstein
Add an LMC cache to struct ib_device, and add a function ib_get_cached_lmc() to query the cache. Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il> Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/cm: remove unneeded flush_workqueueMichael S. Tsirkin
destroy_workqueue() already does flush_workqueue(). Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il> Signed-off-by: Sean Hefty <sean.hefty@intel.com>
2006-06-17IB/ucm: convert semaphore to mutexSean Hefty
Convert semaphore in ib_ucm_file to a real mutex. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB: Make needlessly global ib_mad_cache staticRoland Dreier
Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB: IP address based RDMA connection managerSean Hefty
Kernel connection management agent over InfiniBand that connects based on IP addresses. The agent defines a generic RDMA connection abstraction to support clients wanting to connect over different RDMA devices. The agent also handles RDMA device hotplug events on behalf of clients. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB: address translation to map IP toIB addresses (GIDs)Sean Hefty
Add an address translation service that maps IP addresses to InfiniBand GID addresses using IPoIB. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB/cm: Match connection requests based on private dataSean Hefty
Extend matching connection requests to listens in the InfiniBand CM to include private data checks. This allows applications to listen on the same service identifier, with private data directing the request to the appropriate application. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2006-06-17IB: common handling for marshalling parameters to/from userspaceSean Hefty
Provide common handling for marshalling data between userspace clients and kernel InfiniBand drivers. Signed-off-by: Sean Hefty <sean.hefty@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
generated by cgit v1.2.3 (git 2.25.1) at 2024-08-15 03:19:43 +0000