| Message ID | 20230327074310.1862889-1-john.g.garry@oracle.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Fix shost command overloading issues | expand |
John, > It's easy to get scsi_debug to error on throughput testing when we have > multiple shosts: Applied to 6.4/scsi-staging, thanks!
On Mon, 27 Mar 2023 07:42:59 +0000, John Garry wrote: > It's easy to get scsi_debug to error on throughput testing when we have > multiple shosts: > > $ lsscsi > [7:0:0:0] disk Linux scsi_debug 0191 > [0:0:0:0] disk Linux scsi_debug 0191 > > [...] Applied to 6.4/scsi-queue, thanks! [01/11] scsi: scsi_debug: Fix check for sdev queue full https://git.kernel.org/mkp/scsi/c/6500d2045d52 [02/11] scsi: scsi_debug: Don't iter all shosts in clear_luns_changed_on_target() https://git.kernel.org/mkp/scsi/c/00f9d622e8b2 [03/11] scsi: scsi_debug: Change shost list lock to a mutex https://git.kernel.org/mkp/scsi/c/0aaa3fad4fd9 [04/11] scsi: scsi_debug: Protect block_unblock_all_queues() with mutex https://git.kernel.org/mkp/scsi/c/25b80b2c7582 [05/11] scsi: scsi_debug: Use scsi_block_requests() to block queues https://git.kernel.org/mkp/scsi/c/a0473bf31df5 [06/11] scsi: scsi_debug: Dynamically allocate sdebug_queued_cmd https://git.kernel.org/mkp/scsi/c/1107c7b24ee3 [07/11] scsi: scsi_debug: Use blk_mq_tagset_busy_iter() in sdebug_blk_mq_poll() https://git.kernel.org/mkp/scsi/c/600d9ead3936 [08/11] scsi: scsi_debug: Use blk_mq_tagset_busy_iter() in stop_all_queued() https://git.kernel.org/mkp/scsi/c/9c559c9b4748 [09/11] scsi: scsi_debug: Use scsi_host_busy() in delay_store() and ndelay_store() https://git.kernel.org/mkp/scsi/c/12f3eef016ea [10/11] scsi: scsi_debug: Only allow sdebug_max_queue be modified when no shosts https://git.kernel.org/mkp/scsi/c/57f7225a4fe2 [11/11] scsi: scsi_debug: Drop sdebug_queue https://git.kernel.org/mkp/scsi/c/f1437cd1e535
It's easy to get scsi_debug to error on throughput testing when we have multiple shosts: $ lsscsi [7:0:0:0] disk Linux scsi_debug 0191 [0:0:0:0] disk Linux scsi_debug 0191 $ fio --filename=/dev/sda --filename=/dev/sdb --direct=1 --rw=read --bs=4k --iodepth=256 --runtime=60 --numjobs=40 --time_based --name=jpg --eta-newline=1 --readonly --ioengine=io_uring --hipri --exitall_on_error jpg: (g=0): rw=read, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=io_uring, iodepth=256 ... fio-3.28 Starting 40 processes [ 27.521809] hrtimer: interrupt took 33067 ns [ 27.904660] sd 7:0:0:0: [sdb] tag#171 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s [ 27.904660] sd 0:0:0:0: [sda] tag#58 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s fio: io_u error [ 27.904667] sd 0:0:0:0: [sda] tag#58 CDB: Read(10) 28 00 00 00 27 00 00 01 18 00 on file /dev/sda[ 27.904670] sd 0:0:0:0: [sda] tag#62 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s The issue is related to how the driver manages submit queues and tags. A single array of submit queues - sdebug_q_arr - with its own set of tags is shared among all shosts. As such, for occasions when we have more than one host it is possible to overload the submit queues and run out of tags. Another separate issue that we may reduce the shost submit queue depth, sdebug_max_queue, dynamically causing the shost to be overloaded. How many IOs which the shost may be sent is fixed at can_queue at init time, which is the same initial value for sdebug_max_queue. So reducing sdebug_max_queue means that the shost may be sent more IOs than it is configured to handle, causing overloading. This series removes the scsi_debug submit queue concept and uses pre-existing APIs to manage and examine tags, like scsi_block_requests() and blk_mq_tagset_busy_iter(). Using standard APIs makes the driver more maintainable and extensible in future. A restriction is also added to allow sdebug_max_queue only be modified when no shosts are present, i.e. we need to remove shosts, modify sdebug_max_queue, and then re-add the shosts. Difference to v2: - Make sdebug_alloc_queued_cmd() static Differences to v1: - Add patch to fix sdev queue full test - Add restriction to allow sdebug_max_queue only be modified for no shosts John Garry (11): scsi: scsi_debug: Fix check for sdev queue full scsi: scsi_debug: Don't iter all shosts in clear_luns_changed_on_target() scsi: scsi_debug: Change shost list lock to a mutex scsi: scsi_debug: Protect block_unblock_all_queues() with mutex scsi: scsi_debug: Use scsi_block_requests() to block queues scsi: scsi_debug: Dynamically allocate sdebug_queued_cmd scsi: scsi_debug: Use blk_mq_tagset_busy_iter() in sdebug_blk_mq_poll() scsi: scsi_debug: Use blk_mq_tagset_busy_iter() in stop_all_queued() scsi: scsi_debug: Use scsi_host_busy() in delay_store() and ndelay_store() scsi: scsi_debug: Only allow sdebug_max_queue be modified when no shosts scsi: scsi_debug: Drop sdebug_queue drivers/scsi/scsi_debug.c | 783 ++++++++++++++++++-------------------- 1 file changed, 361 insertions(+), 422 deletions(-)