kernel/drivers/ata/libata-eh.c, branch linux-6.12.y

ata: libata-eh: Fix detection of deferred qc timeouts

2026-03-13T16:20:49Z

commit ee0e6e69a772d601e152e5368a1da25d656122a8 upstream. If the ata_qc_for_each_raw() loop finishes without finding a matching SCSI command for any QC, the variable qc will hold a pointer to the last element examined, which has the tag i == ATA_MAX_QUEUE - 1. This qc can match the port deferred QC (ap->deferred_qc). If that happens, the condition qc == ap->deferred_qc evaluates to true despite the loop not breaking with a match on the SCSI command for this QC. In that case, the error handler mistakenly intercepts a command that has not been issued yet and that has not timed out, and thus erroneously returning a timeout error. Fix the problem by checking for i < ATA_MAX_QUEUE in addition to qc == ap->deferred_qc. The problem was found by an experimental code review agent based on gemini-3.1-pro while reviewing backports into v6.18.y. Assisted-by: Gemini:gemini-3.1-pro Fixes: eddb98ad9364 ("ata: libata-eh: correctly handle deferred qc timeouts") Signed-off-by: Guenter Roeck [cassel: modified commit log as suggested by Damien] Reviewed-by: Damien Le Moal Signed-off-by: Niklas Cassel Signed-off-by: Greg Kroah-Hartman

ata: libata: cancel pending work after clearing deferred_qc

2026-03-13T16:20:49Z

commit aac9b27f7c1f2b2cf7f50a9ca633ecbbcaf22af9 upstream. Syzbot reported a WARN_ON() in ata_scsi_deferred_qc_work(), caused by ap->ops->qc_defer() returning non-zero before issuing the deferred qc. ata_scsi_schedule_deferred_qc() is called during each command completion. This function will check if there is a deferred QC, and if ap->ops->qc_defer() returns zero, meaning that it is possible to queue the deferred qc at this time (without being deferred), then it will queue the work which will issue the deferred qc. Once the work get to run, which can potentially be a very long time after the work was scheduled, there is a WARN_ON() if ap->ops->qc_defer() returns non-zero. While we hold the ap->lock both when assigning and clearing deferred_qc, and the work itself holds the ap->lock, the code currently does not cancel the work after clearing the deferred qc. This means that the following scenario can happen: 1) One or several NCQ commands are queued. 2) A non-NCQ command is queued, gets stored in ap->deferred_qc. 3) Last NCQ command gets completed, work is queued to issue the deferred qc. 4) Timeout or error happens, ap->deferred_qc is cleared. The queued work is currently NOT canceled. 5) Port is reset. 6) One or several NCQ commands are queued. 7) A non-NCQ command is queued, gets stored in ap->deferred_qc. 8) Work is finally run. Yet at this time, there is still NCQ commands in flight. The work in 8) really belongs to the non-NCQ command in 2), not to the non-NCQ command in 7). The reason why the work is executed when it is not supposed to, is because it was never canceled when ap->deferred_qc was cleared in 4). Thus, ensure that we always cancel the work after clearing ap->deferred_qc. Another potential fix would have been to let ata_scsi_deferred_qc_work() do nothing if ap->ops->qc_defer() returns non-zero. However, canceling the work when clearing ap->deferred_qc seems slightly more logical, as we hold the ap->lock when clearing ap->deferred_qc, so we know that the work cannot be holding the lock. (The function could be waiting for the lock, but that is okay since it will do nothing if ap->deferred_qc is not set.) Reported-by: syzbot+bcaf842a1e8ead8dfb89@syzkaller.appspotmail.com Fixes: 0ea84089dbf6 ("ata: libata-scsi: avoid Non-NCQ command starvation") Fixes: eddb98ad9364 ("ata: libata-eh: correctly handle deferred qc timeouts") Reviewed-by: Igor Pylypiv Reviewed-by: Damien Le Moal Signed-off-by: Niklas Cassel Signed-off-by: Greg Kroah-Hartman

ata: libata-eh: correctly handle deferred qc timeouts

2026-03-13T16:20:49Z

commit eddb98ad9364b4e778768785d46cfab04ce52100 upstream. A deferred qc may timeout while waiting for the device queue to drain to be submitted. In such case, since the qc is not active, ata_scsi_cmd_error_handler() ends up calling scsi_eh_finish_cmd(), which frees the qc. But as the port deferred_qc field still references this finished/freed qc, the deferred qc work may eventually attempt to call ata_qc_issue() against this invalid qc, leading to errors such as reported by UBSAN (syzbot run): UBSAN: shift-out-of-bounds in drivers/ata/libata-core.c:5166:24 shift exponent 4210818301 is too large for 64-bit type 'long long unsigned int' ... Call Trace: __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x100/0x190 lib/dump_stack.c:120 ubsan_epilogue+0xa/0x30 lib/ubsan.c:233 __ubsan_handle_shift_out_of_bounds+0x279/0x2a0 lib/ubsan.c:494 ata_qc_issue.cold+0x38/0x9f drivers/ata/libata-core.c:5166 ata_scsi_deferred_qc_work+0x154/0x1f0 drivers/ata/libata-scsi.c:1679 process_one_work+0x9d7/0x1920 kernel/workqueue.c:3275 process_scheduled_works kernel/workqueue.c:3358 [inline] worker_thread+0x5da/0xe40 kernel/workqueue.c:3439 kthread+0x370/0x450 kernel/kthread.c:467 ret_from_fork+0x754/0xd80 arch/x86/kernel/process.c:158 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245 Fix this by checking if the qc of a timed out SCSI command is a deferred one, and in such case, clear the port deferred_qc field and finish the SCSI command with DID_TIME_OUT. Reported-by: syzbot+1f77b8ca15336fff21ff@syzkaller.appspotmail.com Fixes: 0ea84089dbf6 ("ata: libata-scsi: avoid Non-NCQ command starvation") Signed-off-by: Damien Le Moal Reviewed-by: Hannes Reinecke Reviewed-by: Igor Pylypiv Signed-off-by: Greg Kroah-Hartman

ata: libata-scsi: avoid Non-NCQ command starvation

2026-03-13T16:20:23Z

[ Upstream commit 0ea84089dbf62a92dc7889c79e6b18fc89260808 ] When a non-NCQ command is issued while NCQ commands are being executed, ata_scsi_qc_issue() indicates to the SCSI layer that the command issuing should be deferred by returning SCSI_MLQUEUE_XXX_BUSY. This command deferring is correct and as mandated by the ACS specifications since NCQ and non-NCQ commands cannot be mixed. However, in the case of a host adapter using multiple submission queues, when the target device is under a constant load of NCQ commands, there are no guarantees that requeueing the non-NCQ command will be executed later and it may be deferred again repeatedly as other submission queues can constantly issue NCQ commands from different CPUs ahead of the non-NCQ command. This can lead to very long delays for the execution of non-NCQ commands, and even complete starvation for these commands in the worst case scenario. Since the block layer and the SCSI layer do not distinguish between queueable (NCQ) and non queueable (non-NCQ) commands, libata-scsi SAT implementation must ensure forward progress for non-NCQ commands in the presence of NCQ command traffic. This is similar to what SAS HBAs with a hardware/firmware based SAT implementation do. Implement such forward progress guarantee by limiting requeueing of non-NCQ commands from ata_scsi_qc_issue(): when a non-NCQ command is received and NCQ commands are in-flight, do not force a requeue of the non-NCQ command by returning SCSI_MLQUEUE_XXX_BUSY and instead return 0 to indicate that the command was accepted but hold on to the qc using the new deferred_qc field of struct ata_port. This deferred qc will be issued using the work item deferred_qc_work running the function ata_scsi_deferred_qc_work() once all in-flight commands complete, which is checked with the port qc_defer() callback return value indicating that no further delay is necessary. This check is done using the helper function ata_scsi_schedule_deferred_qc() which is called from ata_scsi_qc_complete(). This thus excludes this mechanism from all internal non-NCQ commands issued by ATA EH. When a port deferred_qc is non NULL, that is, the port has a command waiting for the device queue to drain, the issuing of all incoming commands (both NCQ and non-NCQ) is deferred using the regular busy mechanism. This simplifies the code and also avoids potential denial of service problems if a user issues too many non-NCQ commands. Finally, whenever ata EH is scheduled, regardless of the reason, a deferred qc is always requeued so that it can be retried once EH completes. This is done by calling the function ata_scsi_requeue_deferred_qc() from ata_eh_set_pending(). This avoids the need for any special processing for the deferred qc in case of NCQ error, link or device reset, or device timeout. Reported-by: Xingui Yang Reported-by: Igor Pylypiv Fixes: bdb01301f3ea ("scsi: Add host and host template flag 'host_tagset'") Cc: stable@vger.kernel.org Signed-off-by: Damien Le Moal Reviewed-by: Niklas Cassel Reviewed-by: Martin K. Petersen Reviewed-by: John Garry Tested-by: Igor Pylypiv Tested-by: Xingui Yang Signed-off-by: Sasha Levin

ata: libata: Introduce ata_port_eh_scheduled()

2026-03-13T16:20:23Z

[ Upstream commit 7aae547bbe442affc4afe176b157fab820a12437 ] Introduce the inline helper function ata_port_eh_scheduled() to test if EH is pending (ATA_PFLAG_EH_PENDING port flag is set) or running (ATA_PFLAG_EH_IN_PROGRESS port flag is set) for a port. Use this helper in ata_port_wait_eh() and __ata_scsi_queuecmd() to replace the hardcoded port flag tests. No functional changes. Signed-off-by: Damien Le Moal Reviewed-by: Niklas Cassel Link: https://lore.kernel.org/r/20250704104601.310643-1-dlemoal@kernel.org Signed-off-by: Niklas Cassel Stable-dep-of: 0ea84089dbf6 ("ata: libata-scsi: avoid Non-NCQ command starvation") Signed-off-by: Sasha Levin

ata: libata: avoid long timeouts on hot-unplugged SATA DAS

2026-03-04T12:21:20Z

[ Upstream commit 151cabd140322205e27dae5c4bbf261ede0056e3 ] When a SATA DAS enclosure is connected behind a Thunderbolt PCIe switch, hot-unplugging the whole enclosure causes pciehp to tear down the PCI hierarchy before the SCSI layer issues SYNCHRONIZE CACHE and START STOP UNIT for the disks. libata still queues these commands and the AHCI driver tries to access the HBA registers even though the PCI channel is already offline. This results in a series of timeouts and error recovery attempts, e.g.: [ 824.778346] pcieport 0000:00:07.0: pciehp: Slot(14): Link Down [ 891.612720] ata8.00: qc timeout after 5000 msecs (cmd 0xec) [ 902.876501] ata8.00: qc timeout after 10000 msecs (cmd 0xec) [ 934.107998] ata8.00: qc timeout after 30000 msecs (cmd 0xec) [ 936.206431] sd 7:0:0:0: [sda] Synchronize Cache(10) failed: Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK ... [ 1006.298356] ata1.00: qc timeout after 5000 msecs (cmd 0xec) [ 1017.561926] ata1.00: qc timeout after 10000 msecs (cmd 0xec) [ 1048.791790] ata1.00: qc timeout after 30000 msecs (cmd 0xec) [ 1050.890035] sd 0:0:0:0: [sdb] Synchronize Cache(10) failed: Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK With this patch applied, the same hot-unplug looks like: [ 59.965496] pcieport 0000:00:07.0: pciehp: Slot(14): Link Down [ 60.002502] sd 7:0:0:0: [sda] Synchronize Cache(10) failed: Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK ... [ 60.103050] sd 0:0:0:0: [sdb] Synchronize Cache(10) failed: Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK In this test setup with two disks, the hot-unplug sequence shrinks from about 226 seconds (~3.8 minutes) between the Link Down event and the last SYNCHRONIZE CACHE failure to under a second. Without this patch the total delay grows roughly with the number of disks, because each disk gets its own SYNCHRONIZE CACHE and qc timeout series. If the underlying PCI device is already gone, these commands cannot succeed anyway. Avoid issuing them by introducing ata_adapter_is_online(), which checks pci_channel_offline() for PCI-based hosts. It is used from ata_scsi_find_dev() to return NULL, causing the SCSI layer to fail new commands with DID_BAD_TARGET immediately, and from ata_qc_issue() to bail out before touching the HBA registers. Since such failures would otherwise trigger libata error handling, ata_adapter_is_online() is also consulted from ata_scsi_port_error_handler(). When the adapter is offline, libata skips ap->ops->error_handler(ap) and completes error handling using the existing path, rather than running a full EH sequence against a dead adapter. With this change, SYNCHRONIZE CACHE and START STOP UNIT commands issued during hot-unplug fail quickly once the PCI channel is offline, without qc timeout spam or long libata EH delays. Suggested-by: Damien Le Moal Signed-off-by: Henry Tseng Signed-off-by: Damien Le Moal Signed-off-by: Sasha Levin

ata: libata-eh: Do not use ATAPI DMA for a device limited to PIO mode

2025-04-20T08:15:19Z

[ Upstream commit 91ec84f8eaddbc93d7c62e363d68aeb7b89879c7 ] atapi_eh_request_sense() currently uses ATAPI DMA if the SATA controller has ATA_FLAG_PIO_DMA (PIO cmds via DMA) set. However, ATA_FLAG_PIO_DMA is a flag that can be set by a low-level driver on a port at initialization time, before any devices are scanned. If a controller detects a connected device that only supports PIO, we set the flag ATA_DFLAG_PIO. Modify atapi_eh_request_sense() to not use ATAPI DMA if the connected device only supports PIO. Reported-by: Philip Pemberton Closes: https://lore.kernel.org/linux-ide/c6722ee8-5e21-4169-af59-cbbae9edc02f@philpem.me.uk/ Tested-by: Philip Pemberton Reviewed-by: Damien Le Moal Link: https://lore.kernel.org/r/20250221015422.20687-2-cassel@kernel.org Signed-off-by: Niklas Cassel Signed-off-by: Sasha Levin

ata: libata: Set DID_TIME_OUT for commands that actually timed out

2024-10-24T09:14:00Z

When ata_qc_complete() schedules a command for EH using ata_qc_schedule_eh(), blk_abort_request() will be called, which leads to req->q->mq_ops->timeout() / scsi_timeout() being called. scsi_timeout(), if the LLDD has no abort handler (libata has no abort handler), will set host byte to DID_TIME_OUT, and then call scsi_eh_scmd_add() to add the command to EH. Thus, when commands first enter libata's EH strategy_handler, all the commands that have been added to EH will have DID_TIME_OUT set. Commit e5dd410acb34 ("ata: libata: Clear DID_TIME_OUT for ATA PT commands with sense data") clears this bogus DID_TIME_OUT flag for all commands that reached libata's EH strategy_handler. libata has its own flag (AC_ERR_TIMEOUT), that it sets for commands that have not received a completion at the time of entering EH. ata_eh_worth_retry() has no special handling for AC_ERR_TIMEOUT, so by default timed out commands will get flag ATA_QCFLAG_RETRY set, and will be retried after the port has been reset (ata_eh_link_autopsy() always triggers a port reset if any command has AC_ERR_TIMEOUT set). For a command that has ATA_QCFLAG_RETRY set, while also having an error flag set (e.g. AC_ERR_TIMEOUT), ata_eh_finish() will not increment scmd->allowed, so the command will at most be retried scmd->allowed number of times (which by default is set to 3). However, scsi_eh_flush_done_q() will only retry commands for which scsi_noretry_cmd() returns false. For a command that has DID_TIME_OUT set, while also having either the FAILFAST flag set, or the command being a passthrough command, scsi_noretry_cmd() will return true. Thus, such a command will never be retried. Thus, make sure that libata sets SCSI's DID_TIME_OUT flag for commands that actually timed out (libata's AC_ERR_TIMEOUT flag), such that timed out commands will once again not be retried if they are also a FAILFAST or passthrough command. Cc: stable@vger.kernel.org Fixes: e5dd410acb34 ("ata: libata: Clear DID_TIME_OUT for ATA PT commands with sense data") Reported-by: Lai, Yi Closes: https://lore.kernel.org/linux-ide/ZxYz871I3Blsi30F@ly-workstation/ Reviewed-by: Damien Le Moal Link: https://lore.kernel.org/r/20241023105540.1070012-2-cassel@kernel.org Signed-off-by: Niklas Cassel

ata: libata: avoid superfluous disk spin down + spin up during hibernation

2024-10-09T14:21:19Z

A user reported that commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") introduced a spin down + immediate spin up of the disk both when entering and when resuming from hibernation. This behavior was not there before, and causes an increased latency both when entering and when resuming from hibernation. Hibernation is done by three consecutive PM events, in the following order: 1) PM_EVENT_FREEZE 2) PM_EVENT_THAW 3) PM_EVENT_HIBERNATE Commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") modified ata_eh_handle_port_suspend() to call ata_dev_power_set_standby() (which spins down the disk), for both event PM_EVENT_FREEZE and event PM_EVENT_HIBERNATE. Documentation/driver-api/pm/devices.rst, section "Entering Hibernation", explicitly mentions that PM_EVENT_FREEZE does not have to be put the device in a low-power state, and actually recommends not doing so. Thus, let's not spin down the disk on PM_EVENT_FREEZE. (The disk will instead be spun down during the subsequent PM_EVENT_HIBERNATE event.) This way, PM_EVENT_FREEZE will behave as it did before commit aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop"), while PM_EVENT_HIBERNATE will continue to spin down the disk. This will avoid the superfluous spin down + spin up when entering and resuming from hibernation, while still making sure that the disk is spun down before actually entering hibernation. Cc: stable@vger.kernel.org # v6.6+ Fixes: aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") Reviewed-by: Damien Le Moal Link: https://lore.kernel.org/r/20241008135843.1266244-2-cassel@kernel.org Signed-off-by: Niklas Cassel

ata: libata: Clear DID_TIME_OUT for ATA PT commands with sense data

2024-09-10T23:03:43Z

When ata_qc_complete() schedules a command for EH using ata_qc_schedule_eh(), blk_abort_request() will be called, which leads to req->q->mq_ops->timeout() / scsi_timeout() being called. scsi_timeout(), if the LLDD has no abort handler (libata has no abort handler), will set host byte to DID_TIME_OUT, and then call scsi_eh_scmd_add() to add the command to EH. Thus, when commands first enter libata's EH strategy_handler, all the commands that have been added to EH will have DID_TIME_OUT set. libata has its own flag (AC_ERR_TIMEOUT), that it sets for commands that have not received a completion at the time of entering EH. Thus, libata doesn't really care about DID_TIME_OUT at all, and currently clears the host byte at the end of EH, in ata_scsi_qc_complete(), before scsi_eh_finish_cmd() is called. However, this clearing in ata_scsi_qc_complete() is currently only done for commands that are not ATA passthrough commands. Since the host byte is visible in the completion that we return to user space for ATA passthrough commands, for ATA passthrough commands that got completed via EH (commands with sense data), the user will incorrectly see: ATA pass-through(16): transport error: Host_status=0x03 [DID_TIME_OUT] Fix this by moving the clearing of the host byte (which is currently only done for commands that are not ATA passthrough commands) from ata_scsi_qc_complete() to the start of EH (regardless if the command is ATA passthrough or not). While at it, use the proper helper function to clear the host byte, rather than open coding the clearing. This will make sure that we: -Correctly clear DID_TIME_OUT for both ATA passthrough commands and commands that are not ATA passthrough commands. -Do not needlessly clear the host byte for commands that did not go via EH. ata_scsi_qc_complete() is called both for commands that are completed normally (without going via EH), and for commands that went via EH, however, only commands that went via EH will have DID_TIME_OUT set. Fixes: 24aeebbf8ea9 ("scsi: ata: libata: Change ata_eh_request_sense() to not set CHECK_CONDITION") Reported-by: Igor Pylypiv Closes: https://lore.kernel.org/linux-ide/ZttIN8He8TOZ7Lct@google.com/ Signed-off-by: Niklas Cassel Tested-by: Igor Pylypiv Reviewed-by: Hannes Reinecke Signed-off-by: Damien Le Moal