- 09 May, 2020 40 commits
-
-
Chaitanya Kulkarni authored
Now that we have a generic type to name map for configfs, get rid of the nvmet_ana_state_names structure and replace it with newly added nvmet_type_name_map. Signed-off-by:
Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Signed-off-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
Jens Axboe <axboe@kernel.dk>
-
Chaitanya Kulkarni authored
Right now nvmet_addr_adrfam_[store|show]() uses switch and if else ladder for address family to string and reverse mapping which also repeats the strings in show and store function. With addition of generic nvmet_type_name_map structure we can now get rid of the switch and if else ladder and string duplication. Also, we add a newline in before found label in nvmet_addr_trtype_store() which keeps goto label code consistent with nvmet_allowed_hosts_drop_link(), nvmet_port_subsys_drop_link() and nvmet_ana_group_ana_state_store(). Signed-off-by:
-
Chaitanya Kulkarni authored
This patch adds a new type to name mapping generic structure. It replaces nvmet_transport_name with new generic mapping structure nvmet_transport. Signed-off-by:
-
Christoph Hellwig authored
nvme-multipath already uses the gendisk private data, not need to also set up the request_queue queuedata and use it in one place only. Signed-off-by:
Sagi Grimberg <sagi@grimberg.me> Signed-off-by:
-
Sagi Grimberg authored
Today, nvme-tcp automatically schedules a send request to a workqueue context, which is 1 more than we'd need in case the socket buffer is wide open. However, because we have async send activity (as a result of r2t, or write_space callbacks), we need to synchronize sends from possibly multiple contexts (ideally all running on the same cpu though). Thus, we only try to send directly from queue_rq in cases: 1. the send_list is empty 2. we can send it synchronously (i.e. not from the RX path) 3. we run on the same cpu as the queue->io_cpu to avoid contention on the send operation. Proposed-by:
Mark Wunderlich <mark.wunderlich@intel.com> Signed-off-by:
-
Sagi Grimberg authored
When the user runs polled I/O, we shouldn't have to trigger the workqueue to generate the receive work upon the .data_ready upcall. This prevents a redundant context switch when the application is already polling for completions. Proposed-by:
-
Sagi Grimberg authored
data_ready may be invoked from send context or from softirq, so need bh locking for that. Fixes: 3f2304f8 ("nvme-tcp: add NVMe over TCP host driver") Signed-off-by:
-
Weiping Zhang authored
Since commit 147b27e4 ("nvme-pci: allocate device queues storage space at probe"), nvme_alloc_queue does not alloc the nvme queues itself anymore. If the write/poll_queues module parameters are changed at runtime to values larger than the number of allocated queues in nvme_probe, nvme_alloc_queue will access unallocated memory. Add a new nr_allocated_queues member to struct nvme_dev to record how many queues were alloctated in nvme_probe to avoid using more than the allocated queues after a reset following a change to the write/poll_queues module parameters. Also add nr_write_queues and nr_poll_queues members to allow refreshing the number of write and poll queues based on a change to the module parameters when resetting the controller. Fixes: 147b27e4 ("nvme-pci: allocate device queues storage space at probe") Signed-off-by:
Weiping Zhang <zhangweiping@didiglobal.com> Reviewed-by:
Keith Busch <kbusch@kernel.org> Reviewed-by:
Max Gurtovoy <maxg@mellanox.com> [hch: add nvme_max_io_queues, update the commit message] Signed-off-by:
-
Keith Busch authored
The nvme driver does not have enough tags to wrap the queue, and blk-mq will no longer call commit_rqs() when there are no new submissions to notify. Signed-off-by:
-
Keith Busch authored
The completion queue entry is not volatile once the phase is confirmed. Remove the volatile keywords and check the phase using the appropriate READ_ONCE() accessor, allowing the compiler to optimize the remaining completion path. Signed-off-by:
-
Keith Busch authored
If a passthrough command causes the namespace inventory or capabilities to change, flush the scan work that handles these changes so the driver synchronizes with the user command's effects before returning the result to user space. Signed-off-by:
-
Christoph Hellwig authored
Use a common label for putting the nshead if needed and only convert nvme status codes for the one case where it actually is needed. Signed-off-by:
-
Arnd Bergmann authored
When CONFIG_ARCH_NO_SG_CHAIN is set, op->sgl[0] cannot be dereferenced, as gcc-10 now points out: drivers/nvme/host/fc.c: In function 'nvme_fc_init_request': drivers/nvme/host/fc.c:1774:29: warning: array subscript 0 is outside the bounds of an interior zero-length array 'struct scatterlist[0]' [-Wzero-length-bounds] 1774 | op->op.fcp_req.first_sgl = &op->sgl[0]; | ^~~~~~~~~~~ drivers/nvme/host/fc.c:98:21: note: while referencing 'sgl' 98 | struct scatterlist sgl[NVME_INLINE_SG_CNT]; | ^~~ I don't know if this is a legitimate warning or a false-positive. If this is just a false alarm, the warning is easily suppressed by interpreting the array as a pointer. Fixes: b1ae1a23 ("nvme-fc: Avoid preallocating big SGL for data") Signed-off-by:Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Anthony Iliopoulos authored
Add support for detecting capacity changes on nvmet blockdev and file backed namespaces. This allows for emulating and testing online resizing of nvme devices and filesystems on top. Signed-off-by:
Anthony Iliopoulos <ailiop@suse.com> [chaitanya: Fix comments posted on V1] Signed-off-by:
-
Keith Busch authored
The stream parameters indicating optimal io settings were just getting overwritten later. Rearrange the settings so the streams parameters can be preserved if provided. Signed-off-by:
-
Keith Busch authored
The stream parameters are based on the currently formatted logical block size. Recheck these parameters on namespace revalidation so the registered constraints will be accurate. Signed-off-by:
-
Keith Busch authored
Move the quirked chunk_sectors setting to the same location as noiob so one place registers this setting. And since the noiob value is only used locally, remove the member from struct nvme_ns. Signed-off-by:
-
Keith Busch authored
If the namespace identifiers have changed, skip updating the disk information, as that will register parameters from a mismatched namespace. Signed-off-by:
-
Keith Busch authored
The queues' backing device info capabilities don't change with each namespace revalidation. Set it only when each path's request_queue is initially added to a multipath queue. Signed-off-by:
-
Keith Busch authored
Reject a new shared namespace if a duplicate unshared namespace exists. Signed-off-by:
-
Keith Busch authored
Even if a namespace reports it is not capable of sharing, search the subsystem for a matching namespace head. If found, the driver should reject that namespace since it's coming from an invalid configuration. Signed-off-by:
-
Keith Busch authored
If a namespace identification does not match the subsystem's head for that NSID, release the reference that was taken when the matching head was initially found. Signed-off-by:
-
Keith Busch authored
The driver had been unlinking the namespace head from the subsystem's list only after the last reference was released, and outside of the list's subsys->lock protection. There is no reason to track an empty head, so unlink the entry from the subsystem's list when the last namespace using that head is removed and with the mutex lock protecting the list update. The next namespace to attach reusing the previous NSID will allocate a new head rather than find the old head with mismatched identifiers. Signed-off-by:
-
Christoph Hellwig authored
Replace it with a value derived from the identify data and nsid sizes. Signed-off-by:
-
Christoph Hellwig authored
The namespace lists are 0-terminated, so we don't really need the NN value execept for the legacy sequential scan. Signed-off-by:
-
Christoph Hellwig authored
Factor out a piece of deeply indented and logicaly separate code from nvme_scan_ns_list into a new helper. Signed-off-by:
-
Christoph Hellwig authored
Move the check for the supported CNS value into nvme_scan_ns_list, and limit the life time of the identify controller allocation. Signed-off-by:
-
Christoph Hellwig authored
Add a helper to check if we can use Identify CNS values > 1, and refine the Qemu quirk to not apply to reported versions larger than 1.1, as the Qemu implementation had been fixed by then. Signed-off-by:
-
James Smart authored
The kbuild tst robot flagged the following 3 issues: Case 1) >> drivers/nvme/target/fc.c:1201:37: warning: Either the condition >> '!assoc' is redundant or there is possible null pointer dereference: >> assoc. [nullPointerRedundantCheck] >> struct nvmet_fc_tgtport *tgtport = assoc->tgtport; ^ >> drivers/nvme/target/fc.c:1853:7: note: Assuming that condition '!assoc' >> is not redundant >> if (!assoc) ^ >> drivers/nvme/target/fc.c:1850:37: note: Assignment >> 'assoc=nvmet_fc_find_target_assoc(tgtport,be64_to_cpu( >> rqst->associd.association_id))', assigned value is 0 >> assoc = nvmet_fc_find_target_assoc(tgtport, ^ >> drivers/nvme/target/fc.c:1896:31: note: Calling function >> 'nvmet_fc_delete_target_assoc', 1st argument 'assoc' value is 0 >> nvmet_fc_delete_target_assoc(assoc); ^ The tool isn't smart enough to see that line 1854 sets a ret value which thereafter causes the routine to exit. This occurs before any of the assoc references, so it is not an issue. There are 2 more reportings of this same failure. To quiet the tool - rework the if test that does the exit to also reference assoc. No change in logic otherwise. Case 2) drivers/nvme/target/fc.c:1202:29: warning: The scope of the variable 'queue' can be reduced. [variableScope] struct nvmet_fc_tgt_queue *queue; ^ The tool is requesting the variable be declared within the code block that utilizes it. Ignoring this report as existing code style is fine. Case 3) drivers/nvme/target/fc.c:1137:16: warning: Variable 'needrandom' is assigned a value that is never used. [unreadVariable] needrandom = true; ^ Another parsing issue with the tool. Given that parens were not used with the list_for_each_entry() check, it inadvertantly thinks the break exited the outer while loop not the inner for loop. This is not an error. But, added parens to the inner list_for_each_entry() to quiet the tool and as it is better coding style. -- james Signed-off-by:James Smart <jsmart2021@gmail.com> Reported-by:
kbuild test robot <lkp@intel.com> CC: kbuild test robot <lkp@intel.com> CC: Christoph Hellwig <hch@lst.de> Signed-off-by:
-
Max Gurtovoy authored
In order to save resource allocation and utilize the completion locality in a better way (compared to SRQ per device that exist today), allocate Shared Receive Queues (SRQs) per completion vector. Associate each created QP/CQ with an appropriate SRQ according to the queue index. This association will reduce the lock contention in the fast path (compared to SRQ per device solution) and increase the locality in memory buffers. Add new module parameter for SRQ size to adjust it according to the expected load. User should make sure the size is >= 256 to avoid lack of resources. Also reduce the debug level of "last WQE reached" event that is raised when a QP is using SRQ during destruction process to relief the log. Signed-off-by:
-
Keith Busch authored
nvme_alloc_ns_head() doesn't use the 'struct nvme_id_ns' parameter. Remove it, and update caller accordingly. Signed-off-by:
-
Keith Busch authored
Various nvme commands use a zeroes based number of dwords field. Create a helper function to convert byte lengths to this format. Signed-off-by:
-
James Smart authored
Now that common helpers exist, add the ability to Send an NVME LS Request and to Abort an outstanding LS Request to the nvmet side of the driver. Signed-off-by:
Paul Ely <paul.ely@broadcom.com> Signed-off-by:
-
James Smart authored
As the nvmet layer does not have the concept of a remoteport object, which can be used to identify the entity on the other end of the fabric that is to receive an LS, the hosthandle was introduced. The driver passes the hosthandle, a value representative of the remote port, with a ls request receive. The LS request will create the association. The transport will remember the hosthandle for the association, and if there is a need to initiate a LS request to the remote port for the association, the hosthandle will be used. When the driver loses connectivity with the remote port, it needs to notify the transport that the hosthandle is no longer valid, allowing the transport to terminate associations related to the hosthandle. This patch adds support to the driver for the hosthandle. The driver will use the ndlp pointer of the remote port for the hosthandle in calls to nvmet_fc_rcv_ls_req(). The discovery engine is updated to invalidate the hosthandle whenever connectivity with the remote port is lost. Signed-off-by:
Hannes Reinecke <hare@suse.de> Signed-off-by:
-
James Smart authored
Now that common helpers exist, add the ability to receive NVME LS requests to the driver. New requests will be delivered to the transport by nvme_fc_rcv_ls_req(). In order to complete the LS, add support for Send LS Response and send LS response completion handling to the driver. Signed-off-by:
-
James Smart authored
Currently, the ability to send an NVME LS response is limited to the nvmet (controller/target) side of the driver. In preparation of both the nvme and nvmet sides supporting Send LS Response, rework the existing send ls_rsp and ls_rsp completion routines such that there is common code that can be used by both sides. Signed-off-by:
-
James Smart authored
Send LS Abort support is needed when Send LS Request is supported. Currently, the ability to abort an NVME LS request is limited to the nvme (host) side of the driver. In preparation of both the nvme and nvmet sides supporting Send LS Abort, rework the existing ls_req abort routines such that there is common code that can be used by both sides. While refactoring it was seen the logic in the abort routine was incorrect. It attempted to abort all NVME LS's on the indicated port. As such, the routine was reworked to abort only the NVME LS request that was specified. Signed-off-by:
-
James Smart authored
Currently, the ability to send an NVME LS request is limited to the nvme (host) side of the driver. In preparation of both the nvme and nvmet sides support Send LS Request, rework the existing send ls_req and ls_req completion routines such that there is common code that can be used by both sides. Signed-off-by:
-
James Smart authored
In preparation for supporting both intiator mode and target mode receiving NVME LS's, commonize the existing NVME LS request receive handling found in the base driver and in the nvmet side. Using the original lpfc_nvmet_unsol_ls_event() and lpfc_nvme_unsol_ls_buffer() routines as a templates, commonize the reception of an NVME LS request. The common routine will validate the LS request, that it was received from a logged-in node, and allocate a lpfc_async_xchg_ctx that is used to manage the LS request. The role of the port is then inspected to determine which handler is to receive the LS - nvme or nvmet. As such, the nvmet handler is tied back in. A handler is created in nvme and is stubbed out. Signed-off-by:
-
James Smart authored
The last step of commonization is to remove the 'T' suffix from state and flag field definitions. This is minor, but removes the mental association that it solely applies to nvmet use. Signed-off-by:
-
