linux-stable.git/drivers/scsi, branch v4.19.4

SCSI: fix queue cleanup race before queue initialization is done

2018-11-21T08:19:18+00:00

commit 8dc765d438f1e42b3e8227b3b09fad7d73f4ec9a upstream.

c2856ae2f315d ("blk-mq: quiesce queue before freeing queue") has
already fixed this race, however the implied synchronize_rcu()
in blk_mq_quiesce_queue() can slow down LUN probe a lot, so caused
performance regression.

Then 1311326cf4755c7 ("blk-mq: avoid to synchronize rcu inside blk_cleanup_queue()")
tried to quiesce queue for avoiding unnecessary synchronize_rcu()
only when queue initialization is done, because it is usual to see
lots of inexistent LUNs which need to be probed.

However, turns out it isn't safe to quiesce queue only when queue
initialization is done. Because when one SCSI command is completed,
the user of sending command can be waken up immediately, then the
scsi device may be removed, meantime the run queue in scsi_end_request()
is still in-progress, so kernel panic can be caused.

In Red Hat QE lab, there are several reports about this kind of kernel
panic triggered during kernel booting.

This patch tries to address the issue by grabing one queue usage
counter during freeing one request and the following run queue.

Fixes: 1311326cf4755c7 ("blk-mq: avoid to synchronize rcu inside blk_cleanup_queue()")
Cc: Andrew Jones 
Cc: Bart Van Assche 
Cc: linux-scsi@vger.kernel.org
Cc: Martin K. Petersen 
Cc: Christoph Hellwig 
Cc: James E.J. Bottomley 
Cc: stable 
Cc: jianchao.wang 
Signed-off-by: Ming Lei 
Signed-off-by: Jens Axboe 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Initialize port speed to avoid setting lower speed

2018-11-21T08:19:17+00:00

commit f635e48e866ee1a47d2d42ce012fdcc07bf55853 upstream.

This patch initializes port speed so that firmware does not set lower
operating speed. Setting lower speed in firmware impacts WRITE perfomance.

Fixes: 726b85487067 ("qla2xxx: Add framework for async fabric discovery")
Cc: 
Signed-off-by: Quinn Tran 
Signed-off-by: Himanshu Madhani 
Tested-by: Laurence Oberman 
Reviewed-by: Ewan D. Milne 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Fix driver hang when FC-NVMe LUNs are configured

2018-11-21T08:19:14+00:00

commit 39553065f77c297239308470ee313841f4e07db4 upstream.

This patch fixes multiple call for qla_nvme_unregister_remote_port() as part
of qlt_schedule_session_for_deletion(), Do not call it again during
qla_nvme_delete()

Fixes: e473b3074104 ("scsi: qla2xxx: Add FC-NVMe abort processing")
Cc: 
Reviewed-by: Ewan D. Milne 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Fix duplicate switch database entries

2018-11-21T08:19:14+00:00

commit 732ee9a912cf2d9a50c5f9c4213cdc2f885d6aa6 upstream.

The response data buffer used in switch scan is reused 4 times.  (For example,
for commands GPN_FT, GNN_FT for FCP and FC-NVME) Before driver reuses this
buffer, clear it to prevent duplicate entries in our database.

Fixes: a4239945b8ad1 ("scsi: qla2xxx: Add switch command to simplify fabric discovery"
Cc: 
Signed-off-by: Quinn Tran 
Reviewed-by: Ewan D. Milne 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Fix NVMe Target discovery

2018-11-21T08:19:14+00:00

commit db186382af21e926e90df19499475f2552192b77 upstream.

This patch fixes issue when remoteport registers itself as both FCP and
FC-NVMe with the switch, driver will pick FC-NVMe personality as default when
scanning for targets.

Driver was using comaprative operator instead of bitwise operator to check for
fc4_type for both FCP and FC-NVME.

Fixes: 2b5b96473efc ("scsi: qla2xxx: Fix FC-NVMe LUN discovery")
Cc: 
Signed-off-by: Quinn Tran 
Reviewed-by: Ewan D. Milne 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Fix NVMe session hang on unload

2018-11-21T08:19:14+00:00

commit f7d61c995df74d6bb57bbff6a2b7b1874c4a2baa upstream.

Send aborts only when chip is active.

Fixes: 623ee824e579 ("scsi: qla2xxx: Fix FC-NVMe IO abort during driver reset")
Cc:  # 4.14
Signed-off-by: Quinn Tran 
Reviewed-by: Ewan D. Milne 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Fix for double free of SRB structure

2018-11-21T08:19:14+00:00

commit bcc71cc3cde1468958a3ea859276d8d1a1a68265 upstream.

This patch fixes issue during switch command query where driver was freeing
SRB resources multiple times

Following stack trace will be seen
[  853.436234] BUG: unable to handle kernel NULL pointer dereference at
0000000000000001
[  853.436348] IP: [] kmem_cache_alloc+0x74/0x1e0
[  853.436476] PGD 0
[  853.436601] Oops: 0000 [#1] SMP

[  853.454700]  [] ? mod_timer+0x14a/0x220
[  853.455543]  [] mempool_alloc_slab+0x15/0x20
[  853.456395]  [] mempool_alloc+0x69/0x170
[  853.457257]  [] ? internal_add_timer+0x32/0x70
[  853.458136]  [] qla2xxx_queuecommand+0x29b/0x3f0 [qla2xxx]
[  853.459024]  [] scsi_dispatch_cmd+0xaa/0x230
[  853.459923]  [] scsi_request_fn+0x4df/0x680
[  853.460829]  [] ? __switch_to+0xd7/0x510
[  853.461747]  [] __blk_run_queue+0x33/0x40
[  853.462670]  [] blk_delay_work+0x25/0x40
[  853.463603]  [] process_one_work+0x17a/0x440
[  853.464546]  [] worker_thread+0x126/0x3c0
[  853.465501]  [] ? manage_workers.isra.24+0x2a0/0x2a0
[  853.466447]  [] kthread+0xcf/0xe0
[  853.467379]  [] ? insert_kthread_work+0x40/0x40
[  853.470172] Code: db e2 7e 49 8b 50 08 4d 8b 20 49 8b 40 10 4d 85 e4 0f 84 20
01 00 00 48 85 c0 0f 84 17 01 00 00 49 63 46 20 48 8d 4a 01 4d 8b 06 <49> 8b 1c
04 4c 89 e0 65 49 0f c7 08 0f 94 c0 84 c0 74 ba 49 63
[  853.472072] RIP  [] kmem_cache_alloc+0x74/0x1e0
[  853.472971]  RSP 

Fixes: 726b85487067 ("qla2xxx: Add framework for async fabric discovery")
Cc: 
Signed-off-by: Giridhar Malavali 
Reviewed-by: Ewan D. Milne 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Fix re-using LoopID when handle is in use

2018-11-21T08:19:13+00:00

commit 5c6400536481d9ef44ef94e7bf2c7b8e81534db7 upstream.

This patch fixes issue where driver clears NPort ID map instead of marking
handle in use. Once driver clears NPort ID from the database, it can reuse
the same NPort ID resulting in a PLOGI failure.

[mkp: fixed Himanshu's SoB]

Fixes: a084fd68e1d2 ("scsi: qla2xxx: Fix re-login for Nport Handle in use")
Cc: 
Signed-of-by: Quinn Tran 
Reviewed-by: Ewan D. Milne 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: Reject bsg request if chip is down.

2018-11-21T08:19:13+00:00

commit 56d942de59ebfa2e970a6cd33299d1984710b6c0 upstream.

Reject bsg request if chip is down.  This prevent erroneous timeout.

Fixes: d051a5aa1c23 ("[SCSI] qla2xxx: Add an "is reset active" helper.")
Cc: stable@vger.kernel.org # 4.10
Signed-off-by: Quinn Tran 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman

scsi: qla2xxx: shutdown chip if reset fail

2018-11-21T08:19:13+00:00

commit 1e4ac5d6fe0a4af17e4b6251b884485832bf75a3 upstream.

If chip unable to fully initialize, use full shutdown sequence to clear out
any stale FW state.

Fixes: e315cd28b9ef ("[SCSI] qla2xxx: Code changes for qla data structure refactoring")
Cc: stable@vger.kernel.org  #4.10
Signed-off-by: Quinn Tran 
Signed-off-by: Himanshu Madhani 
Signed-off-by: Martin K. Petersen 
Signed-off-by: Greg Kroah-Hartman