kernel/linux.git/drivers/ata, branch v6.12.107

ata: libata-scsi: terminate deferred commands on time out

2026-08-23T12:22:46+00:00

[ Upstream commit 2e1d2e65e773d67dab163127f11a47dab0fbca9f ] If a command times out while we have deferred non-NCQ commands waiting to be issued, the SCSI EH task is not immediately woken up as the waiting deferred commands are never issued nor completed, thus leaving the SCSI host in a busy state (shost->host_failed != scsi_host_busy(shost)) which prevents the SCSI EH task from being woken up. Eventually, when the deferred commands also time out, the SCSI EH task is woken up and the timeout processing occurs. Avoid this unnecessary SCSI EH task wake-up additional time by scheduling a retry of all waiting deferred QCs, using the eh_timed_out SCSI host template operation. The function ata_scsi_eh_timed_out() is introduced to implement this operation. However, terminating deferred commands with DID_REQUEUE to force a retry by calling the function ata_scsi_requeue_deferred_qc() may still keep the SCSI host in a busy state because the block layer may immediately re-issue these commands. The solution to this is to schedule libata EH for the port which suffered the command timeout to prevent accepting any new command. ata_scsi_requeue_deferred_qc() is modified to add a call to ata_port_schedule_eh() for this purpose. In addition to this change, ata_scsi_requeue_deferred_qc() is also modified to take a new timedout_scmd scsi command argument which indicates the SCSI command that timed out. With this additional argument, ata_scsi_requeue_deferred_qc() can now also terminate with DID_TIME_OUT any timed out deferred qc, which simplifies ata_scsi_cmd_error_handler(). In this case, ata_scsi_requeue_deferred_qc() returns SCSI_EH_DONE, with this return value propagated back to the ata_scsi_eh_timed_out() operation to indicate to scsi_timeout() that the timed out command was handled and no further processing is needed. For non-timed out deferred qc that need to be retried, ata_scsi_requeue_deferred_qc() returns SCSI_EH_NOT_HANDLED, thus indicating to scsi_timeout() that the timed out command needs to go through the SCSI EH (and libata EH) processing by adding it to the EH work queue with scsi_eh_scmd_add(). One side effect of these changes is that the function atapi_qc_complete() needs to be modified to ensure that a deferred ATAPI command that needs to be retried is completed with DID_REQUEUE instead of the default SAM_STAT_GOOD status, and a command that timed out is completed with DID_TIME_OUT instead of SAM_STAT_CHECK_CONDITION. Fixes: 0ea84089dbf6 ("ata: libata-scsi: avoid Non-NCQ command starvation") Cc: stable@vger.kernel.org Signed-off-by: Damien Le Moal Reviewed-by: Igor Pylypiv Tested-by: Igor Pylypiv Reviewed-by: Niklas Cassel Reviewed-by: Martin K. Petersen Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

ata: pata_sl82c105: fix bridge revision use-after-free

2026-08-19T16:14:13+00:00

[ Upstream commit 7700a31039cdc6715cb6cce7e7a664ee4e945f67 ] pci_get_slot() returns a referenced PCI device. Commit 44c10138fd4b ("PCI: Change all drivers to use pci_device->revision") replaced a configuration-space read with direct access to the cached revision field, but left that access after pci_dev_put(). The bridge may therefore be freed before its revision is read. Read the revision before dropping the reference. Fixes: 44c10138fd4b ("PCI: Change all drivers to use pci_device->revision") Signed-off-by: Hongyan Xu Reviewed-by: Niklas Cassel Signed-off-by: Damien Le Moal Signed-off-by: Sasha Levin

ata: ahci: Make ahci_ignore_port() handle empty mask_port_map

2026-08-09T18:23:22+00:00

commit 130ff5c8b78e6fd05270a04985c50bce6a3de6c1 upstream. Commit 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive port numbers") added a skip to ahci_platform_enable_phys() for ports that are not in mask_port_map. The code in ahci_platform_get_resources(), will currently set mask_port_map for each child "port" node it finds in the device tree. However, device trees that do not have any child "port" nodes will not have mask_port_map set, and for non-device tree platforms mask_port_map will only exist as a quirk for specific PCI device + vendor IDs, or as a kernel module parameter, but will not be set by default. Therefore, the common thing is that mask_port_map is only set if you do not want to use all ports (as defined by Offset 0Ch: PI – Ports Implemented register), but instead only want to use the ports in mask_port_map. If mask_port_map is not set, all ports are available. Thus, ahci_ignore_port() must be able to handle an empty mask_port_map. Fixes: 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive port numbers") Fixes: 2c202e6c4f4d ("ata: libahci_platform: Do not set mask_port_map when not needed") Fixes: c9b5be909e65 ("ahci: Introduce ahci_ignore_port() helper") Reported-by: Marek Szyprowski Closes: https://lore.kernel.org/linux-ide/10b31dd0-d0bb-4f76-9305-2195c3e17670@samsung.com/ Tested-by: Marek Szyprowski Co-developed-by: Damien Le Moal Signed-off-by: Damien Le Moal Link: https://lore.kernel.org/r/20250225141612.942170-2-cassel@kernel.org Signed-off-by: Niklas Cassel Signed-off-by: Greg Kroah-Hartman

ata: libahci_platform: Do not set mask_port_map when not needed

2026-08-09T18:23:22+00:00

commit 2c202e6c4f4dd19d2e8c1dfac9df05170aa3934f upstream. Commit 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive port numbers") modified ahci_platform_get_resources() to allow identifying the ports of a controller that are defined as child nodes of the controller node in order to support non-consecutive port numbers (as defined by the platform device tree). However, this commit also erroneously sets bit 0 of hpriv->mask_port_map when the platform devices tree does not define port child nodes, to match the fact that the temporary default number of ports used in that case is 1 (which is also consistent with the fact that only index 0 of hpriv->phys[] is initialized with the call to ahci_platform_get_phy(). But doing so causes ahci_platform_init_host() to initialize and probe only the first port, even if this function determines that the controller has in fact multiple ports using the capability register of the controller (through a call to ahci_nr_ports()). This can be seen with the ahci_mvebu driver (Armada 385 SoC) with the second port declared as "dummy": ahci-mvebu f10a8000.sata: masking port_map 0x3 -> 0x1 ahci-mvebu f10a8000.sata: AHCI vers 0001.0000, 32 command slots, 6 Gbps, platform mode ahci-mvebu f10a8000.sata: 1/2 ports implemented (port mask 0x1) ahci-mvebu f10a8000.sata: flags: 64bit ncq sntf led only pmp fbs pio slum part sxs scsi host0: ahci-mvebu scsi host1: ahci-mvebu ata1: SATA max UDMA/133 mmio [mem 0xf10a8000-0xf10a9fff] port 0x100 irq 40 lpm-pol 0 ata2: DUMMY Fix this issue by removing setting bit 0 of hpriv->mask_port_map when the platform device tree does not define port child nodes. Reported-by: Klaus Kudielka Fixes: 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive port numbers") Tested-by: Klaus Kudielka Signed-off-by: Damien Le Moal Acked-by: Josua Mayer Link: https://lore.kernel.org/r/20250207232915.1439174-1-dlemoal@kernel.org Signed-off-by: Niklas Cassel Signed-off-by: Greg Kroah-Hartman

ata: libata-sata: fix ata_scsi_lpm_supported() iteration

2026-08-09T18:23:13+00:00

commit 3fd70e96914d761c17c376aadd0b0d1a3c9badba upstream. The inner loop of ata_scsi_lpm_supported() uses the wrong variable when iterating. It should obviously use the link that we are currently iterating over, rather than always using the host link. ata_scsi_lpm_supported() is used to control if a user should be allowed to change lpm policy (from the default) via sysfs. Thus, this bug could potentially disallow users to change the LPM policy for certain SATA devices via sysfs. Cc: stable@vger.kernel.org Fixes: 0060beec0bfa ("ata: libata-sata: Add link_power_management_supported sysfs attribute") Reported-by: Sashiko Link: https://lore.kernel.org/linux-ide/20260728112200.B99F21F000E9@smtp.kernel.org/ Signed-off-by: Niklas Cassel Signed-off-by: Damien Le Moal Signed-off-by: Greg Kroah-Hartman

ata: libata-eh: Increase STANDBY IMMEDIATE timeout

2026-08-09T18:23:13+00:00

commit 1e024d2b41ee32bc06818f7f09a3562c58842cf9 upstream. Correct a previous change (see Fixes) which reduced the standby timeout from 30 to 5 seconds. Increase it to 15 seconds. I was troubleshooting an error spotted during system suspend: [ 1217.152867] ata1.00: Entering standby power mode [ 1222.322948] ata1.00: qc timeout after 5000 msecs (cmd 0xe0) [ 1222.324010] ata1.00: STANDBY IMMEDIATE failed (err_mask=0x4) This drive is a Samsung 870 EVO SSD in good SMART standing, and I wasn't aware of any reason it should be taking so long to standby. The issue is intermittent, but I observed it sometimes taking 7 seconds to manually standby. I assume this was interruption of background maintenance after a power outage. As a desktop user, I would prefer to wait the extra 2 seconds at suspend to let the drive finish its business rather than drop the rails from under it. The change from 30 to 5 seconds was implicit when switching suspend from START STOP UNIT to an internal command with no timeout table entry. No reason was stated for the change. Fixes: aa3998dbeb3a ("ata: libata-scsi: Disable scsi device manage_system_start_stop") Cc: stable@vger.kernel.org Signed-off-by: Matt Vollrath Assisted-by: Claude:claude-5-fable Signed-off-by: Damien Le Moal Signed-off-by: Greg Kroah-Hartman

ata: ahci_ceva: fix error paths in ceva_ahci_platform_enable_resources()

2026-08-09T18:23:02+00:00

[ Upstream commit 4d99a91574c420decab56cc880fad0dc15b8a7a3 ] On phy_init() failure the error path fallsthrough to disable_rsts, which deasserts the controller reset and then enters disable_phys calling phy_power_off() on PHYs that were never powered on. That corrupts the PHY power_count and triggers an extra runtime PM put. Use a separate exit_phys path that unwinds with phy_exit() only and falls through to disable_clks while the controller remains in reset. Reserve phy_power_off() for the phy_power_on() failure path only, and skip masked-out ports in both unwind loops. On phy_power_on() failure re-assert the controller reset before disabling clocks and regulators, matching the teardown order used by ahci_platform_enable_resources() and ahci_platform_disable_resources(). Fixes: 26c8404e162b ("ata: ahci_ceva: fix error handling for Xilinx GT PHY support") Signed-off-by: Radhey Shyam Pandey Signed-off-by: Damien Le Moal Signed-off-by: Sasha Levin

ahci: Introduce ahci_ignore_port() helper

2026-08-09T18:23:02+00:00

[ Upstream commit c9b5be909e6595547ed5d45aef39fd65948aa342 ] libahci and AHCI drivers may ignore some ports if the port is invalid (its ID does not correspond to a valid physical port) or if the user explicitly requested the port to be ignored with the mask_port_map ahci module parameter. Such port that shall be ignored can be identified by checking that the bit corresponding to the port ID is not set in the mask_port_map field of struct ahci_host_priv. E.g. code such as: "if (!(hpriv->mask_port_map & (1 << portid)))". Replace all direct use of the mask_port_map field to detect such port with the new helper inline function ahci_ignore_port() to make the code more readable/easier to understand. The comment describing the mask_port_map field of struct ahci_host_priv is also updated to be more accurate. Signed-off-by: Damien Le Moal Reviewed-by: Niklas Cassel Stable-dep-of: 4d99a91574c4 ("ata: ahci_ceva: fix error paths in ceva_ahci_platform_enable_resources()") Signed-off-by: Sasha Levin

ata: libahci_platform: support non-consecutive port numbers

2026-08-09T18:23:01+00:00

[ Upstream commit 8c87215dd3a2c814dcffc0bafe8c80c8f98f2574 ] So far ahci_platform relied on number of child nodes in firmware to allocate arrays and expected port numbers to start from 0 without holes. This number of ports is then set in private structure for use when configuring phys and regulators. Some platforms may not use every port of an ahci controller. E.g. SolidRUN CN9130 Clearfog uses only port 1 but not port 0, leading to the following errors during boot: [ 1.719476] ahci f2540000.sata: invalid port number 1 [ 1.724562] ahci f2540000.sata: No port enabled Update all accessesors of ahci_host_priv phys and target_pwrs arrays to support holes. Access is gated by hpriv->mask_port_map which has a bit set for each enabled port. Update ahci_platform_get_resources to ignore holes in the port numbers and enable ports defined in firmware by their reg property only. When firmware does not define children it is assumed that there is exactly one port, using index 0. Signed-off-by: Josua Mayer Reviewed-by: Hans de Goede Signed-off-by: Damien Le Moal Stable-dep-of: 4d99a91574c4 ("ata: ahci_ceva: fix error paths in ceva_ahci_platform_enable_resources()") Signed-off-by: Sasha Levin

ata: sata_mv: accept 1 or 2 resources in platform probe

2026-08-09T18:23:01+00:00

[ Upstream commit ef19a9cf037957fe3a35df8355c76ff0a63a0436 ] Board files in arch/arm/plat-orion, arch/arm/mach-dove, arch/arm/mach-mv78xx0 and arch/arm/mach-orion5x still register the "sata_mv" device with two resources (IORESOURCE_MEM plus IORESOURCE_IRQ). Those devices are rejected with -EINVAL, so SATA no longer probes on legacy Marvell Orion/Kirkwood-style boards. Accept both 1 resource (DT, IRQ fetched via platform_get_irq()) and 2 resources (legacy, IRQ supplied as a second resource) so both probing paths work. Fixes: b3b2bec9646e ("ata: sata_mv: Fixes expected number of resources now IRQs are gone") Assisted-by: opencode:big-pickle Signed-off-by: Rosen Penev Signed-off-by: Damien Le Moal Signed-off-by: Sasha Levin