kernel/linux.git/drivers/scsi, branch v7.1.12

scsi: core: pair EH runtime PM get and put

2026-08-23T12:29:30+00:00

[ Upstream commit 872f486259ae0bc6b73ca4735a15d013241f73e9 ] shost->eh_noresume is currently consulted twice in one error handling iteration: once before scsi_autopm_get_host() and once again before scsi_autopm_put_host(). That is racy when a PM-triggered error path flips shost->eh_noresume while the SCSI EH thread is still running. The problem flow looks like this: PM path ufshcd_set_dev_pwr_mode() shost->eh_noresume = 1 ufshcd_execute_start_stop <-- trigger EH ... shost->eh_noresume = 0 EH path scsi_error_handler() if (!shost->eh_noresume) scsi_autopm_get_host() <-- skipped ... if (!shost->eh_noresume) scsi_autopm_put_host() <-- executed later In that case one EH iteration can skip autoresume on entry and still drop a runtime PM reference on exit. That leaves an unmatched runtime PM put and can trigger a runtime PM usage count underflow. Fix this by making eh_noresume a regular bool so it can be accessed with READ_ONCE() and WRITE_ONCE(). Snapshot it once per EH iteration and use that snapshot for both runtime PM get and put decisions. Fixes: ae0751ffc77e ("[SCSI] add flag to skip the runtime PM calls on the host") Signed-off-by: Hongjie Fang Reviewed-by: Bart Van Assche Reviewed-by: Peter Wang Link: https://patch.msgid.link/20260729111614.2407559-1-hongjiefang@asrmicro.com Signed-off-by: Martin K. Petersen (Oracle) Signed-off-by: Sasha Levin

scsi: scsi_debug: Negate wrapped memcmp() result

2026-08-19T16:20:31+00:00

commit c4f6916a99cf105c3ff340b6210fcbba3fa66b35 upstream. comp_write_worker() returns true when the compared data matches. memcmp() returns zero for equal data and non-zero for different data, so its result must be negated before it is stored in a bool. The first segment already uses !memcmp(), but the wrapped segment uses memcmp() directly, reversing the match result. Use !memcmp() there as well. Fixes: 38d5c8336e60 ("scsi_debug: add Report supported opcodes+tmfs; Compare and write") Cc: stable@vger.kernel.org Signed-off-by: Xu Rao Reviewed-by: John Garry Link: https://patch.msgid.link/E5AD42E9C0E18633+20260803095328.3445311-1-raoxu@uniontech.com Signed-off-by: Martin K. Petersen (Oracle) Signed-off-by: Greg Kroah-Hartman

scsi: scsi_debug: Fix REPORT ZONES alloc_len underflow OOB write

2026-08-09T18:26:48+00:00

commit 93dde0bf2f39a0f9f57fd610aa3201ce5b753433 upstream. resp_report_zones() sizes the reply buffer from the CDB allocation length. The v3 fix rounds alloc_len up with ALIGN() before deriving the descriptor count: rep_max_zones = (ALIGN((u64)alloc_len, RZONES_DESC_HD) - RZONES_DESC_HD) >> ilog2(RZONES_DESC_HD); arr_len = (u64)RZONES_DESC_HD * (rep_max_zones + 1); For alloc_len in 0xFFFFFFC1..0xFFFFFFFF, ALIGN() rounds up to 0x100000000, so arr_len is 4 GB. On 32-bit, kzalloc()'s size_t is 32-bit and truncates 0x100000000 to 0; kzalloc(0) returns ZERO_SIZE_PTR, which passes the !arr check, and desc = arr + 64 is then dereferenced in the loop -> out-of-bounds write / panic. Clamp rep_max_zones to devip->nr_zones. The loop already stops at sdebug_capacity (after nr_zones zones), so a report can never hold more than nr_zones descriptors; the clamp does not change the report, it only bounds arr_len to (nr_zones + 1) * RZONES_DESC_HD, a real device property that can never reach 0x100000000. Fixes: 7db0e0c8190a ("scsi: scsi_debug: Fix buffer size of REPORT ZONES command") Suggested-by: Damien Le Moal Cc: stable@vger.kernel.org Signed-off-by: Ibrahim Hashimov Assisted-by: AuditCode-AI:2026.07 Reviewed-by: Damien Le Moal Reviewed-by: Bart Van Assche Link: https://patch.msgid.link/20260712183739.83915-1-security@auditcode.ai Signed-off-by: Martin K. Petersen Signed-off-by: Greg Kroah-Hartman

scsi: libsas: terminate deferred commands on time out

2026-08-09T18:26:48+00:00

commit de202d2251bc181c6019b1ad3c0ba8133e5ec68d upstream. 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 additional SCSI EH wake up time with the same method as implemented in libata-scsi, using the eh_timed_out SCSI host template operation. The function sas_eh_timed_out() implements this operation and executes the function ata_scsi_retry_deferred_qc() for SATA devices. Co-developed-by: Igor Pylypiv Signed-off-by: Igor Pylypiv Fixes: 0ea84089dbf6 ("ata: libata-scsi: avoid Non-NCQ command starvation") Cc: stable@vger.kernel.org Signed-off-by: Damien Le Moal Reviewed-by: John Garry Reviewed-by: Hannes Reinecke Tested-by: Igor Pylypiv Reviewed-by: Niklas Cassel Reviewed-by: Martin K. Petersen Signed-off-by: Greg Kroah-Hartman

scsi: mpi3mr: Fix potential deadlock in mpi3mr_fault_uevent_emit

2026-08-09T18:26:35+00:00

[ Upstream commit ccff8c92571500fcfed21281e33daaf645bf692f ] mpi3mr_fault_uevent_emit() runs from the fault watchdog and reset paths where host I/O may already be blocked. GFP_KERNEL allocations here, both the local kzalloc_obj() and the ones inside kobject_uevent_env() itself, can trigger reclaim that waits on that blocked I/O and deadlock. Use memalloc_noio_save()/restore() to cover the whole call instead of just the local allocation. Fixes: ec54b348f274 ("scsi: mpi3mr: Record and report controller firmware faults") Reported-by: sashiko-bot Closes: https://sashiko.dev/#/patchset/20260724164630.924288-1-chandrakanth.patil%40broadcom.com Signed-off-by: Chandrakanth Patil Link: https://patch.msgid.link/20260724175231.935192-1-chandrakanth.patil@broadcom.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: libsas: Fix HA resume deadlock and hisi_sas disk-wake race

2026-08-09T18:26:28+00:00

[ Upstream commit 3dbbbf656b850c9c8de05df6ad4a1dfc6ff02845 ] Commit fbefe22811c3 ("scsi: libsas: Don't always drain event workqueue for HA resume") introduced sas_resume_ha_no_sync() to avoid a deadlock: the PHYE_RESUME_TIMEOUT handler, running on the HA event workqueue, calls sas_deform_port() -> sas_destruct_devices(), which removes SCSI devices and waits for the host to become runtime-active. But the host cannot resume until sas_resume_ha() -> sas_drain_work() returns, and the drain is blocked on that very handler. However skipping the drain reintroduces a race: hisi_sas returns from resume before all PHY UP work and libsas discovery work finish. The controller may then autosuspend while disks are still waking up. The disks issue IO to a suspended controller, the IO fails, and the disks get disabled. Fix the deadlock at its source by moving the PHYE_RESUME_TIMEOUT notification to after sas_drain_work(). By then the host resume is about to complete, so device removal through device_link no longer blocks on the resume and the cycle is broken. With the deadlock gone, restore sas_resume_ha() (the draining variant) in hisi_sas and remove sas_resume_ha_no_sync(). The reorder is safe for the other libsas consumers (isci, pm8001, aic94xx, mvsas). During suspend, sas_suspend_devices() calls sas_notify_lldd_dev_gone() for each device, which sets dev->lldd_dev to NULL. When scsi_unblock_requests re-enables I/O in resume, any I/O to a timed-out phy's disk is immediately rejected by the LLDD before reaching hardware: isci returns SAS_DEVICE_UNKNOWN (mapped to DID_BAD_TARGET), and pm8001 returns SAS_PHY_DOWN (mapped to DID_NO_CONNECT). Both complete directly via scsi_done() without entering SCSI EH. This is identical in both the old and new ordering since lldd_dev_gone runs during suspend, before resume. The reorder only affects when the PHYE_RESUME_TIMEOUT handler runs (synchronized by sas_drain_work() vs. asynchronous after resume returns), not whether I/O can reach the device. aic94xx and mvsas do not register any PM ops and never reach this code path. Fixes: fbefe22811c3 ("scsi: libsas: Don't always drain event workqueue for HA resume") Signed-off-by: Xingui Yang Reviewed-by: John Garry Link: https://patch.msgid.link/20260716081145.3950172-1-yangxingui@huawei.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: libiscsi_tcp: Bound SCSI Response data segment to the connection buffer

2026-08-09T18:26:28+00:00

[ Upstream commit c1dea15f819cded9b3faf58f8bec72323568b6e6 ] iscsi_tcp_hdr_dissect() receives the data segment of several PDU types into the fixed-size conn->data buffer, which is allocated for ISCSI_DEF_MAX_RECV_SEG_LEN (8192) bytes. For the LOGIN_RSP, TEXT_RSP, REJECT and ASYNC_EVENT opcodes the dissect path already rejects a PDU whose DataSegmentLength exceeds that buffer. The SCSI Command Response (ISCSI_OP_SCSI_CMD_RSP) path also copies its data segment (sense/response data) into conn->data via iscsi_tcp_data_recv_prep(), but it does so without the same check. The only upstream bound on in.datalen is conn->max_recv_dlength, the initiator's advertised MaxRecvDataSegmentLength, which is commonly negotiated well above 8192 (open-iscsi defaults to 262144). A target that returns a SCSI Response with a DataSegmentLength between 8193 and max_recv_dlength therefore overflows the 8192-byte conn->data buffer. Once the same bound applies, ISCSI_OP_SCSI_CMD_RSP is handled exactly like those responses: bound the data segment, receive it into conn->data when present, and otherwise complete the PDU with no data. Fold the opcode into that case group rather than duplicating the check. Fixes: a081c13e39b5 ("[SCSI] iscsi_tcp: split module into lib and lld") Suggested-by: Chris Leech Assisted-by: Claude:claude-opus-4-8 Signed-off-by: HyeongJun An Acked-by: Chris Leech Link: https://patch.msgid.link/20260716065848.1653431-1-sammiee5311@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: libiscsi: Fix stale-data leak into the SCSI sense buffer

2026-08-09T18:26:28+00:00

[ Upstream commit 98b87885de4b7f605533a2860685f5689fce8e82 ] iscsi_scsi_cmd_rsp() copies the sense data of a SCSI Response from the target-supplied data segment. The segment carries a 2-byte sense length followed by the sense bytes, so it must hold 2 + senselen bytes, but the bounds check only requires datalen >= senselen: senselen = get_unaligned_be16(data); if (datalen < senselen) goto invalid_datalen; memcpy(sc->sense_buffer, data + 2, min_t(uint16_t, senselen, SCSI_SENSE_BUFFERSIZE)); A target that returns a SCSI Response whose datalen equals senselen (with senselen <= SCSI_SENSE_BUFFERSIZE) makes the memcpy() from data + 2 read up to two bytes past the received data. Those bytes are stale conn->data contents and end up in the command's sense buffer, which is returned to userspace. Account for the 2-byte sense length prefix in the check. Fixes: 7996a778ff8c ("[SCSI] iscsi: add libiscsi") Suggested-by: Sashiko AI Assisted-by: Claude:claude-opus-4-8 Signed-off-by: HyeongJun An Acked-by: Chris Leech Link: https://patch.msgid.link/20260714104934.1404423-1-sammiee5311@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: core: wake eh reliably when using scsi_schedule_eh

2026-08-03T09:25:21+00:00

[ Upstream commit dccf3b1798b70f94e958b3d00b83010399e6fb05 ] Drivers which use the scsi_schedule_eh function to run the error handler currently risk the error handler thread never waking once all commands are timed out or inactive. There is no enforced memory order between setting the host into error recovery state and counting busy commands. This can result in a race with scsi_dec_host_busy where neither CPU sees both conditions of all commands inactive and the host error state to request waking the error handler. To fix this, run the scsi_schedule_eh's scsi_eh_wakeup from a new work item which will use rcu to ensure scsi_schedule_eh's call to scsi_host_busy will occur after the error state is globally visible and will be seen by any current scsi_dec_host_busy callers. Fixes: 6eb045e092ef ("scsi: core: avoid host-wide host_busy counter for scsi_mq") Signed-off-by: David Jeffery Link: https://patch.msgid.link/20260615174630.11492-1-djeffery@redhat.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: elx: efct: Fix I/O leak on unsupported additional CDB

2026-07-24T14:21:12+00:00

commit 9cb2d5291dbfe7bed565ead3337047dee9ed1064 upstream. efct_dispatch_fcp_cmd() allocates an efct_io before dispatching an unsolicited FCP command. If the command has an unsupported additional CDB, the function returns -EIO before handing the IO to the SCSI layer. Free the allocated IO before returning from this error path. Fixes: f45ae6aac0a0 ("scsi: elx: efct: Unsolicited FC frame processing routines") Cc: stable@vger.kernel.org Signed-off-by: Haoxiang Li Reviewed-by: Daniel Wagner Link: https://patch.msgid.link/20260622075844.832871-1-haoxiang_li2024@163.com Signed-off-by: Martin K. Petersen Signed-off-by: Greg Kroah-Hartman