kernel/git/stable/linux.git/drivers/scsi/lpfc, branch linux-6.8.y

scsi: lpfc: Use a dedicated lock for ras_fwlog state

2024-05-17T10:14:37+00:00

[ Upstream commit f733a76ea0a9a84aee4ac41b81fad4d610ecbd8e ] To reduce usage of and contention for hbalock, a separate dedicated lock is used to protect ras_fwlog state. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240305200503.57317-8-justintee8345@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: lpfc: Release hbalock before calling lpfc_worker_wake_up()

2024-05-17T10:14:37+00:00

[ Upstream commit ded20192dff31c91cef2a04f7e20e60e9bb887d3 ] lpfc_worker_wake_up() calls the lpfc_work_done() routine, which takes the hbalock. Thus, lpfc_worker_wake_up() should not be called while holding the hbalock to avoid potential deadlock. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240305200503.57317-7-justintee8345@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: lpfc: Replace hbalock with ndlp lock in lpfc_nvme_unregister_port()

2024-05-17T10:14:36+00:00

[ Upstream commit d11272be497e48a8e8f980470eb6b70e92eed0ce ] The ndlp object update in lpfc_nvme_unregister_port() should be protected by the ndlp lock rather than hbalock. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240305200503.57317-6-justintee8345@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: lpfc: Update lpfc_ramp_down_queue_handler() logic

2024-05-17T10:14:36+00:00

[ Upstream commit bb011631435c705cdeddca68d5c85fd40a4320f9 ] Typically when an out of resource CQE status is detected, the lpfc_ramp_down_queue_handler() logic is called to help reduce I/O load by reducing an sdev's queue_depth. However, the current lpfc_rampdown_queue_depth() logic does not help reduce queue_depth. num_cmd_success is never updated and is always zero, which means new_queue_depth will always be set to sdev->queue_depth. So, new_queue_depth = sdev->queue_depth - new_queue_depth always sets new_queue_depth to zero. And, scsi_change_queue_depth(sdev, 0) is essentially a no-op. Change the lpfc_ramp_down_queue_handler() logic to set new_queue_depth equal to sdev->queue_depth subtracted from number of times num_rsrc_err was incremented. If num_rsrc_err is >= sdev->queue_depth, then set new_queue_depth equal to 1. Eventually, the frequency of Good_Status frames will signal SCSI upper layer to auto increase the queue_depth back to the driver default of 64 via scsi_handle_queue_ramp_up(). Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240305200503.57317-5-justintee8345@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: lpfc: Remove IRQF_ONESHOT flag from threaded IRQ handling

2024-05-17T10:14:36+00:00

[ Upstream commit 4623713e7ade46bfc63a3eade836f566ccbcd771 ] IRQF_ONESHOT is found to mask HBA generated interrupts when thread_fn is running. As a result, some EQEs/CQEs miss timely processing resulting in SCSI layer attempts to abort commands due to io_timeout. Abort CQEs are also not processed leading to the observations of hangs and spam of "0748 abort handler timed out waiting for aborting I/O" log messages. Remove the IRQF_ONESHOT flag. The cmpxchg and xchg atomic operations on lpfc_queue->queue_claimed already protect potential parallel access to an EQ/CQ should the thread_fn get interrupted by the primary irq handler. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240305200503.57317-4-justintee8345@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: lpfc: Move NPIV's transport unregistration to after resource clean up

2024-05-17T10:14:36+00:00

[ Upstream commit 4ddf01f2f1504fa08b766e8cfeec558e9f8eef6c ] There are cases after NPIV deletion where the fabric switch still believes the NPIV is logged into the fabric. This occurs when a vport is unregistered before the Remove All DA_ID CT and LOGO ELS are sent to the fabric. Currently fc_remove_host(), which calls dev_loss_tmo for all D_IDs including the fabric D_ID, removes the last ndlp reference and frees the ndlp rport object. This sometimes causes the race condition where the final DA_ID and LOGO are skipped from being sent to the fabric switch. Fix by moving the fc_remove_host() and scsi_remove_host() calls after DA_ID and LOGO are sent. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240305200503.57317-3-justintee8345@gmail.com Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: lpfc: Fix possible memory leak in lpfc_rcv_padisc()

2024-04-13T11:10:04+00:00

[ Upstream commit 2ae917d4bcab80ab304b774d492e2fcd6c52c06b ] The call to lpfc_sli4_resume_rpi() in lpfc_rcv_padisc() may return an unsuccessful status. In such cases, the elsiocb is not issued, the completion is not called, and thus the elsiocb resource is leaked. Check return value after calling lpfc_sli4_resume_rpi() and conditionally release the elsiocb resource. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240131185112.149731-3-justintee8345@gmail.com Reviewed-by: Himanshu Madhani Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin

scsi: lpfc: Correct size for wqe for memset()

2024-04-03T13:32:50+00:00

commit 28d41991182c210ec1654f8af2e140ef4cc73f20 upstream. The wqe is of type lpfc_wqe128. It should be memset with the same type. Fixes: 6c621a2229b0 ("scsi: lpfc: Separate NVMET RQ buffer posting from IO resources SGL/iocbq/context") Signed-off-by: Muhammad Usama Anjum Link: https://lore.kernel.org/r/20240304090649.833953-1-usama.anjum@collabora.com Reviewed-by: AngeloGioacchino Del Regno Reviewed-by: Justin Tee Signed-off-by: Martin K. Petersen Signed-off-by: Greg Kroah-Hartman

scsi: lpfc: Correct size for cmdwqe/rspwqe for memset()

2024-04-03T13:32:50+00:00

commit 16cc2ba71b9f6440805aef7f92ba0f031f79b765 upstream. The cmdwqe and rspwqe are of type lpfc_wqe128. They should be memset() with the same type. Fixes: 61910d6a5243 ("scsi: lpfc: SLI path split: Refactor CT paths") Signed-off-by: Muhammad Usama Anjum Link: https://lore.kernel.org/r/20240304091119.847060-1-usama.anjum@collabora.com Reviewed-by: Justin Tee Signed-off-by: Martin K. Petersen Signed-off-by: Greg Kroah-Hartman

scsi: lpfc: Use unsigned type for num_sge

2024-02-05T21:19:11+00:00

LUNs going into "failed ready running" state observed on >1T and on even numbers of size (2T, 4T, 6T, 8T and 10T). The issue occurs when DIF is enabled at the host. The kernel logs: Cannot setup S/G List for HBAIO segs 1/1 SGL 512 SCSI 256: 3 0 The host lpfc driver is failing to setup scatter/gather list (protection data) for the I/Os. The return type lpfc_bg_setup_sgl()/lpfc_bg_setup_sgl_prot() causes the compiler to remove the most significant bit. Use an unsigned type instead. Signed-off-by: Hannes Reinecke [dwagner: added commit message] Signed-off-by: Daniel Wagner Link: https://lore.kernel.org/r/20231220162658.12392-1-dwagner@suse.de Signed-off-by: Martin K. Petersen