aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2017-08-18skd: Remove driver version informationBart Van Assche1-2/+0
Remove the driver version information because this information is not useful in an upstream kernel driver. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Bump driver versionBart Van Assche1-12/+5
Bump the driver version. Remove the build ID because build IDs do not make sense for an upstream kernel driver. Keep the driver version in the module information but do not report it during every load, unload or probe. Signed-off-by: Bart Van Assche <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Optimize lockingBart Van Assche1-6/+15
Only take skdev->lock if necessary. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove several local variablesBart Van Assche1-30/+7
This patch does not change any functionality but makes the code more brief. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Reduce memory usageBart Van Assche1-37/+108
Every single coherent DMA memory buffer occupies at least one page. Reduce memory usage by switching from coherent buffers to streaming DMA for I/O requests (struct skd_fitmsg_context) and S/G-lists (struct fit_sg_descriptor[]). Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove skd_device.in_flightBart Van Assche1-16/+21
Since skd_device.in_flight is only used to display the number of in-flight requests in debug messages, remove that member and introduce skd_in_flight(). That last function relies on the block layer to determine the number of in flight requests. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Switch to block layer timeout mechanismBart Van Assche1-86/+31
Remove the timeout slot variables and rely on the block layer to detect request timeouts. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Convert to blk-mqBart Van Assche1-138/+91
Introduce a tag set and a blk_mq_ops structure. Set .cmd_size such that struct request and struct skd_request_context are allocated through a single allocation. Remove the skd_request_context.req pointer. Make queue starting asynchronous such that this can occur safely from interrupt context. Use locking to protect skdev->skmsg and *skdev->skmsg against concurrent access from concurrent .queue_rq() calls. Introduce the functions skd_init_request() and skd_exit_request() to set up / clean up the per-request S/G-list. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Coalesce struct request and struct skd_request_contextBart Van Assche1-120/+54
Set request_queue.cmd_size, introduce skd_init_rq() and skd_exit_rq() and remove skd_device.skreq_table. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Move skd_free_sg_list() upBart Van Assche1-13/+12
Issue a warning if a NULL argument is passed to skd_free_sg_list(). Move this function up to make the blk-mq conversion patch easier to read. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Split skd_recover_requests()Bart Van Assche1-17/+22
This patch does not change any functionality but makes the blk-mq conversion patch easier to read. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Introduce skd_process_request()Bart Van Assche1-118/+119
The only functional change in this patch is that the skd_fitmsg_context in which requests are accumulated is changed from a local variable into a member of struct skd_device. This patch will make the blk-mq conversion easier. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Convert several per-device scalar variables into atomicsBart Van Assche1-31/+37
Convert the per-device scalar variables that are protected by the queue lock into atomics such that it becomes safe to access these variables without holding the queue lock. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Enable request tags for the block layer queueBart Van Assche1-194/+73
Use the request tag when allocating a skd_fitmsg_context or skd_request_context such that the lists used to track free elements can be eliminated. Swap the skd_end_request() and skd_release_req() calls to avoid triggering a use-after-free. Remove skd_fitmsg_context.state and .outstanding because FIT messages are shared among requests and because updating a FIT message after a request has finished whould trigger a use-after-free. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Initialize skd_special_context.req.n_sg to oneBart Van Assche1-0/+1
The debug code in skd_send_special_fitmsg() assumes that req.n_sg represents the number of S/G descriptors. However, skd_construct() initializes that member variable to zero. Set req.n_sg to one such that the debugging code in skd_send_special_fitmsg() works as expected. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove dead codeBart Van Assche1-12/+0
Removing the SG IO code also removed the code that sets SKD_REQ_STATE_ABORTED. Hence also remove the code that checks for this state. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove SG IO supportBart Van Assche1-1069/+2
The skd SG IO support duplicates the functionality of the bsg driver. Hence remove it. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Convert explicit skd_request_fn() callsBart Van Assche1-5/+5
This will make it easier to convert this driver to the blk-mq approach. This patch also reduces interrupt latency by moving skd_request_fn() calls out of the skd_isr() interrupt. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Rework request failing code pathBart Van Assche1-10/+8
Move the skd_fail_all_pending() call out of skd_request_fn_not_online() such that this function can be reused in the blk-mq code path. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Move a function definitionBart Van Assche1-43/+41
This patch does not change any functionality but makes the next patch in this series easier to read. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skb: Use symbolic names for SCSI opcodesBart Van Assche1-6/+9
This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Use kcalloc() instead of kzalloc() with multiplyBart Van Assche1-13/+17
This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove superfluous occurrences of the 'volatile' keywordBart Van Assche1-26/+22
mem_map[i] is accessed through readl() / writel() hence declaring mem_map as volatile is not necessary. Remove the volatile declarations from struct fit_completion_entry_v1 pointers and struct fit_comp_error_info since reading these structures multiple times is safe. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove a redundant init_timer() callBart Van Assche1-1/+0
Since setup_timer() invokes init_timer(), invoking init_timer() just before setup_timer() is redundant. Hence remove the init_timer() call. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Use for_each_sg()Bart Van Assche1-6/+6
This change makes skd_preop_sg_list() support chained sg-lists. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Drop second argument of skd_recover_requests()Bart Van Assche1-14/+9
Since all callers pass zero as second argument to skd_recover_requests(), drop that second argument. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove superfluous initializations from skd_isr_completion_posted()Bart Van Assche1-6/+6
The value of skcmp, cmp_cntxt etc. is overwritten during every loop iteration and is not used after the loop has finished. Hence initializing these variables outside the loop is not necessary. This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Simplify the code for handling data directionBart Van Assche1-16/+9
Use DMA_FROM_DEVICE and DMA_TO_DEVICE directly instead of introducing driver-private constants with the same numerical value. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Use ARRAY_SIZE() where appropriateBart Van Assche1-3/+2
Use ARRAY_SIZE() instead of open-coding it. This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Make the skd_isr() code more briefBart Van Assche1-2/+1
This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Use __packed only when neededBart Van Assche1-5/+1
Since needless use of __packed slows down access to data structures, only use __packed when needed. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Check structure sizes at build timeBart Van Assche1-0/+5
This patch will help to verify the changes made by the next patch. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Use a structure instead of hardcoding structure offsetsBart Van Assche1-19/+22
This change makes the source code easier to read. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Simplify the code for allocating DMA message buffersBart Van Assche2-14/+7
dma_alloc_coherent() guarantees alignment on a page boundary so no explicit alignment is needed to align on a 64 byte boundary. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Simplify the code for deciding whether or not to send a FIT msgBart Van Assche1-24/+5
Due to the previous patch it is guaranteed that the FIT msg contains at least one request after the for-loop has finished. Use this to simplify the code. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Reorder the code in skd_process_request()Bart Van Assche1-33/+9
Prepare the S/G-list before allocating a FIT msg such that the FIT msg always contains at least one request after the for-loop is finished. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Fix size argument in skd_free_skcomp()Bart Van Assche1-7/+2
Pass the correct size to pci_free_consistent() in skd_free_skcomp(). Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Introduce SKD_SKCOMP_SIZEBart Van Assche1-14/+8
This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Introduce the symbolic constant SKD_MAX_REQ_PER_MSGBart Van Assche1-2/+8
This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Document locking assumptionsBart Van Assche1-0/+8
Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Fix endianness annotationsBart Van Assche2-17/+15
Ensure that sparse does not report any warnings when building the skd driver with sparse verification enabled (C=1 or C=2). Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Switch from the pr_*() to the dev_*() logging functionsBart Van Assche1-521/+391
Use dev_err() and dev_info() instead of pr_err() and pr_info(). Since dev_dbg() is able to report file name and line number information, remove __FILE__ and __LINE__ from the dev_dbg() calls. Remove the struct skd_device members and the function (skd_name()) that became superfluous due to these changes. This patch removes the device name and serial number from log statements. An example of the old log line format: (skd0:STM000196603:[0000:00:09.0]): Driver state STARTING(3)=>ONLINE(4) An example of the new log line format: skd:0000:00:09.0: Driver state STARTING(3)=>ONLINE(4) Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove useless barrier() callsBart Van Assche1-32/+10
The purpose of barrier() is to prevent reordering by the compiler. Since the compiler does not reorder calls to non-pure functions, remove the barrier() calls from skd_reg_{read,write}{32,64}(). Since pr_debug() is able to report file name and line number information, remove __FILE__ and __LINE__ from the pr_debug() calls. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove a set-but-not-used variable from struct skd_deviceBart Van Assche1-3/+0
This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove set-but-not-used local variablesBart Van Assche1-10/+1
These variables have been detected by building with W=1. Declare 'acc' as __maybe_unused because most access_ok() implementations ignore their first argument. This patch does not change any functionality. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Fix a function name in a commentBart Van Assche1-1/+1
There is no function skd_completion_posted_isr() in the skd driver but there is a function called skd_isr_completion_posted(). Fix the function name in the comment. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Fix spelling in a source code commentBart Van Assche1-1/+1
Change "ptimal" into "optimal" and remove the misleading reference to sysfs. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Avoid that gcc 7 warns about fall-through when building with W=1Bart Van Assche1-1/+1
Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove unnecessary blank linesBart Van Assche1-16/+5
This patch does not change any functionality but makes the skd driver source code more uniform with that of other kernel drivers. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
2017-08-18skd: Remove ESXi codeBart Van Assche1-14/+0
Since the code guarded by #ifdef SKD_VMK_POLL_HANDLER / #endif is never built on Linux systems, remove it. Signed-off-by: Bart Van Assche <[email protected]> Cc: Christoph Hellwig <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Johannes Thumshirn <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.