The patch below does not apply to the 5.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable@xxxxxxxxxxxxxxx>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.4.y git checkout FETCH_HEAD git cherry-pick -x 066c5b46b6eaf2f13f80c19500dbb3b84baabb33 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable@xxxxxxxxxxxxxxx>' --in-reply-to '2023123002-vocalist-unmoved-5bd9@gregkh' --subject-prefix 'PATCH 5.4.y' HEAD^.. Possible dependencies: 066c5b46b6ea ("scsi: core: Always send batch on reset or error handling command") bf23e619039d ("scsi: core: Use a structure member to track the SCSI command submitter") aa8e25e5006a ("scsi: core: Use scsi_cmd_to_rq() instead of scsi_cmnd.request") d2c945f01d23 ("scsi: core: Make scsi_get_lba() return the LBA") f0f214fe8cd3 ("scsi: core: Introduce scsi_get_sector()") 7ba46799d346 ("scsi: core: Add scsi_prot_ref_tag() helper") 2ceda20f0a99 ("scsi: core: Move command size detection out of the fast path") 3a8dc5bbc8c0 ("scsi: core: Remove scsi_init_cmd_errh") 2a242d59d6b9 ("scsi: core: Add limitless cmd retry support") 15f73f5b3e59 ("blk-mq: move failure injection out of blk_mq_complete_request") 818dbde78e0f ("Merge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi") thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 066c5b46b6eaf2f13f80c19500dbb3b84baabb33 Mon Sep 17 00:00:00 2001 From: Alexander Atanasov <alexander.atanasov@xxxxxxxxxxxxx> Date: Fri, 15 Dec 2023 14:10:08 +0200 Subject: [PATCH] scsi: core: Always send batch on reset or error handling command In commit 8930a6c20791 ("scsi: core: add support for request batching") the block layer bd->last flag was mapped to SCMD_LAST and used as an indicator to send the batch for the drivers that implement this feature. However, the error handling code was not updated accordingly. scsi_send_eh_cmnd() is used to send error handling commands and request sense. The problem is that request sense comes as a single command that gets into the batch queue and times out. As a result the device goes offline after several failed resets. This was observed on virtio_scsi during a device resize operation. [ 496.316946] sd 0:0:4:0: [sdd] tag#117 scsi_eh_0: requesting sense [ 506.786356] sd 0:0:4:0: [sdd] tag#117 scsi_send_eh_cmnd timeleft: 0 [ 506.787981] sd 0:0:4:0: [sdd] tag#117 abort To fix this always set SCMD_LAST flag in scsi_send_eh_cmnd() and scsi_reset_ioctl(). Fixes: 8930a6c20791 ("scsi: core: add support for request batching") Cc: <stable@xxxxxxxxxxxxxxx> Signed-off-by: Alexander Atanasov <alexander.atanasov@xxxxxxxxxxxxx> Link: https://lore.kernel.org/r/20231215121008.2881653-1-alexander.atanasov@xxxxxxxxxxxxx Reviewed-by: Ming Lei <ming.lei@xxxxxxxxxx> Signed-off-by: Martin K. Petersen <martin.petersen@xxxxxxxxxx> diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c index c67cdcdc3ba8..1223d34c04da 100644 --- a/drivers/scsi/scsi_error.c +++ b/drivers/scsi/scsi_error.c @@ -1152,6 +1152,7 @@ static enum scsi_disposition scsi_send_eh_cmnd(struct scsi_cmnd *scmd, scsi_log_send(scmd); scmd->submitter = SUBMITTED_BY_SCSI_ERROR_HANDLER; + scmd->flags |= SCMD_LAST; /* * Lock sdev->state_mutex to avoid that scsi_device_quiesce() can @@ -2459,6 +2460,7 @@ scsi_ioctl_reset(struct scsi_device *dev, int __user *arg) scsi_init_command(dev, scmd); scmd->submitter = SUBMITTED_BY_SCSI_RESET_IOCTL; + scmd->flags |= SCMD_LAST; memset(&scmd->sdb, 0, sizeof(scmd->sdb)); scmd->cmd_len = 0;
