kernel/linux.git/drivers/nvme/host/fc.c, branch v6.6.131

nvme-fc: release admin tagset if init fails

2026-02-11T12:39:12+00:00

[ Upstream commit d1877cc7270302081a315a81a0ee8331f19f95c8 ] nvme_fabrics creates an NVMe/FC controller in following path: nvmf_dev_write() -> nvmf_create_ctrl() -> nvme_fc_create_ctrl() -> nvme_fc_init_ctrl() nvme_fc_init_ctrl() allocates the admin blk-mq resources right after nvme_add_ctrl() succeeds. If any of the subsequent steps fail (changing the controller state, scheduling connect work, etc.), we jump to the fail_ctrl path, which tears down the controller references but never frees the admin queue/tag set. The leaked blk-mq allocations match the kmemleak report seen during blktests nvme/fc. Check ctrl->ctrl.admin_tagset in the fail_ctrl path and call nvme_remove_admin_tag_set() when it is set so that all admin queue allocations are reclaimed whenever controller setup aborts. Reported-by: Yi Zhang Reviewed-by: Justin Tee Signed-off-by: Chaitanya Kulkarni Signed-off-by: Keith Busch Signed-off-by: Sasha Levin

nvme-pci: do not directly handle subsys reset fallout

2026-01-30T09:27:42+00:00

[ Upstream commit 210b1f6576e8b367907e7ff51ef425062e1468e4 ] Scheduling reset_work after a nvme subsystem reset is expected to fail on pcie, but this also prevents potential handling the platform's pcie services may provide that might successfully recovering the link without re-enumeration. Such examples include AER, DPC, and power's EEH. Provide a pci specific operation that safely initiates a subsystem reset, and instead of scheduling reset work, read back the status register to trigger a pcie read error. Since this only affects pci, the other fabrics drivers subscribe to a generic nvmf subsystem reset that is exactly the same as before. The loop fabric doesn't use it because nvmet doesn't support setting that property anyway. And since we're using the magic NSSR value in two places now, provide a symbolic define for it. Reported-by: Nilay Shroff Reviewed-by: Christoph Hellwig Signed-off-by: Keith Busch Stable-dep-of: 0edb475ac0a7 ("nvme: fix PCIe subsystem reset controller state transition") Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

nvme-fc: rename free_ctrl callback to match name pattern

2026-01-30T09:27:42+00:00

[ Upstream commit 205fb5fa6fde1b5b426015eb1ff69f2ff25ef5bb ] Rename nvme_fc_nvme_ctrl_freed to nvme_fc_free_ctrl to match the name pattern for the callback. Reviewed-by: Christoph Hellwig Reviewed-by: Sagi Grimberg Reviewed-by: Hannes Reinecke Signed-off-by: Daniel Wagner Signed-off-by: Keith Busch Stable-dep-of: 0edb475ac0a7 ("nvme: fix PCIe subsystem reset controller state transition") Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

nvme-fc: don't hold rport lock when putting ctrl

2026-01-11T14:21:53+00:00

[ Upstream commit b71cbcf7d170e51148d5467820ae8a72febcb651 ] nvme_fc_ctrl_put can acquire the rport lock when freeing the ctrl object: nvme_fc_ctrl_put nvme_fc_ctrl_free spin_lock_irqsave(rport->lock) Thus we can't hold the rport lock when calling nvme_fc_ctrl_put. Justin suggested use the safe list iterator variant because nvme_fc_ctrl_put will also modify the rport->list. Cc: Justin Tee Reviewed-by: Christoph Hellwig Signed-off-by: Daniel Wagner Signed-off-by: Keith Busch Signed-off-by: Sasha Levin

nvme: nvme-fc: Ensure ->ioerr_work is cancelled in nvme_fc_delete_ctrl()

2025-12-01T10:41:42+00:00

commit 0a2c5495b6d1ecb0fa18ef6631450f391a888256 upstream. nvme_fc_delete_assocation() waits for pending I/O to complete before returning, and an error can cause ->ioerr_work to be queued after cancel_work_sync() had been called. Move the call to cancel_work_sync() to be after nvme_fc_delete_association() to ensure ->ioerr_work is not running when the nvme_fc_ctrl object is freed. Otherwise the following can occur: [ 1135.911754] list_del corruption, ff2d24c8093f31f8->next is NULL [ 1135.917705] ------------[ cut here ]------------ [ 1135.922336] kernel BUG at lib/list_debug.c:52! [ 1135.926784] Oops: invalid opcode: 0000 [#1] SMP NOPTI [ 1135.931851] CPU: 48 UID: 0 PID: 726 Comm: kworker/u449:23 Kdump: loaded Not tainted 6.12.0 #1 PREEMPT(voluntary) [ 1135.943490] Hardware name: Dell Inc. PowerEdge R660/0HGTK9, BIOS 2.5.4 01/16/2025 [ 1135.950969] Workqueue: 0x0 (nvme-wq) [ 1135.954673] RIP: 0010:__list_del_entry_valid_or_report.cold+0xf/0x6f [ 1135.961041] Code: c7 c7 98 68 72 94 e8 26 45 fe ff 0f 0b 48 c7 c7 70 68 72 94 e8 18 45 fe ff 0f 0b 48 89 fe 48 c7 c7 80 69 72 94 e8 07 45 fe ff <0f> 0b 48 89 d1 48 c7 c7 a0 6a 72 94 48 89 c2 e8 f3 44 fe ff 0f 0b [ 1135.979788] RSP: 0018:ff579b19482d3e50 EFLAGS: 00010046 [ 1135.985015] RAX: 0000000000000033 RBX: ff2d24c8093f31f0 RCX: 0000000000000000 [ 1135.992148] RDX: 0000000000000000 RSI: ff2d24d6bfa1d0c0 RDI: ff2d24d6bfa1d0c0 [ 1135.999278] RBP: ff2d24c8093f31f8 R08: 0000000000000000 R09: ffffffff951e2b08 [ 1136.006413] R10: ffffffff95122ac8 R11: 0000000000000003 R12: ff2d24c78697c100 [ 1136.013546] R13: fffffffffffffff8 R14: 0000000000000000 R15: ff2d24c78697c0c0 [ 1136.020677] FS: 0000000000000000(0000) GS:ff2d24d6bfa00000(0000) knlGS:0000000000000000 [ 1136.028765] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 1136.034510] CR2: 00007fd207f90b80 CR3: 000000163ea22003 CR4: 0000000000f73ef0 [ 1136.041641] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 1136.048776] DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400 [ 1136.055910] PKRU: 55555554 [ 1136.058623] Call Trace: [ 1136.061074] [ 1136.063179] ? show_trace_log_lvl+0x1b0/0x2f0 [ 1136.067540] ? show_trace_log_lvl+0x1b0/0x2f0 [ 1136.071898] ? move_linked_works+0x4a/0xa0 [ 1136.075998] ? __list_del_entry_valid_or_report.cold+0xf/0x6f [ 1136.081744] ? __die_body.cold+0x8/0x12 [ 1136.085584] ? die+0x2e/0x50 [ 1136.088469] ? do_trap+0xca/0x110 [ 1136.091789] ? do_error_trap+0x65/0x80 [ 1136.095543] ? __list_del_entry_valid_or_report.cold+0xf/0x6f [ 1136.101289] ? exc_invalid_op+0x50/0x70 [ 1136.105127] ? __list_del_entry_valid_or_report.cold+0xf/0x6f [ 1136.110874] ? asm_exc_invalid_op+0x1a/0x20 [ 1136.115059] ? __list_del_entry_valid_or_report.cold+0xf/0x6f [ 1136.120806] move_linked_works+0x4a/0xa0 [ 1136.124733] worker_thread+0x216/0x3a0 [ 1136.128485] ? __pfx_worker_thread+0x10/0x10 [ 1136.132758] kthread+0xfa/0x240 [ 1136.135904] ? __pfx_kthread+0x10/0x10 [ 1136.139657] ret_from_fork+0x31/0x50 [ 1136.143236] ? __pfx_kthread+0x10/0x10 [ 1136.146988] ret_from_fork_asm+0x1a/0x30 [ 1136.150915] Fixes: 19fce0470f05 ("nvme-fc: avoid calling _nvme_fc_abort_outstanding_ios from interrupt context") Cc: stable@vger.kernel.org Tested-by: Marco Patalano Reviewed-by: Justin Tee Signed-off-by: Ewan D. Milne Signed-off-by: Keith Busch Signed-off-by: Greg Kroah-Hartman

nvme: nvme-fc: move tagset removal to nvme_fc_delete_ctrl()

2025-12-01T10:41:42+00:00

commit ea3442efabd0aa3930c5bab73c3901ef38ef6ac3 upstream. Now target is removed from nvme_fc_ctrl_free() which is the ctrl->ref release handler. And even admin queue is unquiesced there, this way is definitely wrong because the ctr->ref is grabbed when submitting command. And Marco observed that nvme_fc_ctrl_free() can be called from request completion code path, and trigger kernel warning since request completes from softirq context. Fix the issue by moveing target removal into nvme_fc_delete_ctrl(), which is also aligned with nvme-tcp and nvme-rdma. Patch originally proposed by Ming Lei, then modified to move the tagset removal down to after nvme_fc_delete_association() after further testing. Cc: Marco Patalano Cc: Ewan Milne Cc: James Smart Cc: Sagi Grimberg Signed-off-by: Ming Lei Cc: stable@vger.kernel.org Tested-by: Marco Patalano Reviewed-by: Justin Tee Signed-off-by: Ewan D. Milne Signed-off-by: Keith Busch Signed-off-by: Greg Kroah-Hartman

nvme-fc: use lock accessing port_state and rport state

2025-11-24T09:29:21+00:00

[ Upstream commit 891cdbb162ccdb079cd5228ae43bdeebce8597ad ] nvme_fc_unregister_remote removes the remote port on a lport object at any point in time when there is no active association. This races with with the reconnect logic, because nvme_fc_create_association is not taking a lock to check the port_state and atomically increase the active count on the rport. Reported-by: Shinichiro Kawasaki Closes: https://lore.kernel.org/all/u4ttvhnn7lark5w3sgrbuy2rxupcvosp4qmvj46nwzgeo5ausc@uyrkdls2muwx Signed-off-by: Daniel Wagner Reviewed-by: Hannes Reinecke Signed-off-by: Keith Busch Signed-off-by: Sasha Levin

nvme-fc: rely on state transitions to handle connectivity loss

2025-03-22T19:50:50+00:00

commit f13409bb3f9140dad7256febcb478f0c9600312c upstream. It's not possible to call nvme_state_ctrl_state with holding a spin lock, because nvme_state_ctrl_state calls cancel_delayed_work_sync when fastfail is enabled. Instead syncing the ASSOC_FLAG and state transitions using a lock, it's possible to only rely on the state machine transitions. That means nvme_fc_ctrl_connectivity_loss should unconditionally call nvme_reset_ctrl which avoids the read race on the ctrl state variable. Actually, it's not necessary to test in which state the ctrl is, the reset work will only scheduled when the state machine is in LIVE state. In nvme_fc_create_association, the LIVE state can only be entered if it was previously CONNECTING. If this is not possible then the reset handler got triggered. Thus just error out here. Fixes: ee59e3820ca9 ("nvme-fc: do not ignore connectivity loss during connecting") Closes: https://lore.kernel.org/all/denqwui6sl5erqmz2gvrwueyxakl5txzbbiu3fgebryzrfxunm@iwxuthct377m/ Reported-by: Shinichiro Kawasaki Tested-by: Shin'ichiro Kawasaki Reviewed-by: Sagi Grimberg Signed-off-by: Daniel Wagner Signed-off-by: Keith Busch Signed-off-by: Greg Kroah-Hartman

nvme-fc: do not ignore connectivity loss during connecting

2025-03-22T19:50:40+00:00

[ Upstream commit ee59e3820ca92a9f4307ae23dfc7229dc8b8d400 ] When a connectivity loss occurs while nvme_fc_create_assocation is being executed, it's possible that the ctrl ends up stuck in the LIVE state: 1) nvme nvme10: NVME-FC{10}: create association : ... 2) nvme nvme10: NVME-FC{10}: controller connectivity lost. Awaiting Reconnect nvme nvme10: queue_size 128 > ctrl maxcmd 32, reducing to maxcmd 3) nvme nvme10: Could not set queue count (880) nvme nvme10: Failed to configure AEN (cfg 900) 4) nvme nvme10: NVME-FC{10}: controller connect complete 5) nvme nvme10: failed nvme_keep_alive_end_io error=4 A connection attempt starts 1) and the ctrl is in state CONNECTING. Shortly after the LLDD driver detects a connection lost event and calls nvme_fc_ctrl_connectivity_loss 2). Because we are still in CONNECTING state, this event is ignored. nvme_fc_create_association continues to run in parallel and tries to communicate with the controller and these commands will fail. Though these errors are filtered out, e.g in 3) setting the I/O queues numbers fails which leads to an early exit in nvme_fc_create_io_queues. Because the number of IO queues is 0 at this point, there is nothing left in nvme_fc_create_association which could detected the connection drop. Thus the ctrl enters LIVE state 4). Eventually the keep alive handler times out 5) but because nothing is being done, the ctrl stays in LIVE state. There is already the ASSOC_FAILED flag to track connectivity loss event but this bit is set too late in the recovery code path. Move this into the connectivity loss event handler and synchronize it with the state change. This ensures that the ASSOC_FAILED flag is seen by nvme_fc_create_io_queues and it does not enter the LIVE state after a connectivity loss event. If the connectivity loss event happens after we entered the LIVE state the normal error recovery path is executed. Signed-off-by: Daniel Wagner Reviewed-by: Hannes Reinecke Reviewed-by: Sagi Grimberg Signed-off-by: Keith Busch Signed-off-by: Sasha Levin

nvme-fc: go straight to connecting state when initializing

2025-03-22T19:50:40+00:00

[ Upstream commit d3d380eded7ee5fc2fc53b3b0e72365ded025c4a ] The initial controller initialization mimiks the reconnect loop behavior by switching from NEW to RESETTING and then to CONNECTING. The transition from NEW to CONNECTING is a valid transition, so there is no point entering the RESETTING state. TCP and RDMA also transition directly to CONNECTING state. Reviewed-by: Sagi Grimberg Reviewed-by: Hannes Reinecke Reviewed-by: Christoph Hellwig Signed-off-by: Daniel Wagner Signed-off-by: Keith Busch Signed-off-by: Sasha Levin