The suspend/resume functions in this driver seem to have multiple problems,
the latest one just got introduced by a bugfix:
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c: In function '_suspend_v3_hw':
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c:5142:39: error: 'struct dev_pm_info' has no member named 'usage_count'
5142 | if (atomic_read(&device->power.usage_count)) {
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c: In function '_suspend_v3_hw':
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c:5142:39: error: 'struct dev_pm_info' has no member named 'usage_count'
5142 | if (atomic_read(&device->power.usage_count)) {
As far as I can tell, the 'usage_count' is not meant to be accessed by
device drivers at all, though I don't know what the driver is supposed to
do instead.
Another problem is the use of the deprecated UNIVERSAL_DEV_PM_OPS(), and
marking functions as __maybe_unused to avoid warnings about unused
functions. This should probably be changed to using
DEFINE_RUNTIME_DEV_PM_OPS().
Both changes require actually understanding what the driver needs to do,
and being able to test this, so instead here is the simplest patch to make
it pass the randconfig builds instead.
Fixes: e368d38cb9 ("scsi: hisi_sas: Exit suspend state when usage count is greater than 0")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230405083611.3376739-1-arnd@kernel.org
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When if_type equals zero and pci_resource_start(pdev, PCI_64BIT_BAR4)
returns false, drbl_regs_memmap_p is not remapped. This passes a NULL
pointer to iounmap(), which can trigger a WARN() on certain arches.
When if_type equals six and pci_resource_start(pdev, PCI_64BIT_BAR4)
returns true, drbl_regs_memmap_p may has been remapped and
ctrl_regs_memmap_p is not remapped. This is a resource leak and passes a
NULL pointer to iounmap().
To fix these issues, we need to add null checks before iounmap(), and
change some goto labels.
Fixes: 1351e69fc6 ("scsi: lpfc: Add push-to-adapter support to sli4")
Signed-off-by: Shuchang Li <lishuchang@hust.edu.cn>
Link: https://lore.kernel.org/r/20230404072133.1022-1-lishuchang@hust.edu.cn
Reviewed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
John Garry <john.g.garry@oracle.com> says:
It's easy to get scsi_debug to error on throughput testing when we have
multiple shosts:
$ lsscsi
[7:0:0:0] disk Linux scsi_debug 0191
[0:0:0:0] disk Linux scsi_debug 0191
$ fio --filename=/dev/sda --filename=/dev/sdb --direct=1 --rw=read
--bs=4k --iodepth=256 --runtime=60 --numjobs=40 --time_based --name=jpg
--eta-newline=1 --readonly --ioengine=io_uring --hipri --exitall_on_error
jpg: (g=0): rw=read, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=io_uring, iodepth=256
...
fio-3.28
Starting 40 processes
[ 27.521809] hrtimer: interrupt took 33067 ns
[ 27.904660] sd 7:0:0:0: [sdb] tag#171 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s
[ 27.904660] sd 0:0:0:0: [sda] tag#58 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s
fio: io_u error [ 27.904667] sd 0:0:0:0: [sda] tag#58 CDB: Read(10) 28 00 00 00 27 00 00 01 18 00
on file /dev/sda[ 27.904670] sd 0:0:0:0: [sda] tag#62 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s
The issue is related to how the driver manages submit queues and tags. A
single array of submit queues - sdebug_q_arr - with its own set of tags is
shared among all shosts. As such, for occasions when we have more than one
host it is possible to overload the submit queues and run out of tags.
Another separate issue that we may reduce the shost submit queue depth,
sdebug_max_queue, dynamically causing the shost to be overloaded. How many
IOs which the shost may be sent is fixed at can_queue at init time, which
is the same initial value for sdebug_max_queue. So reducing
sdebug_max_queue means that the shost may be sent more IOs than it is
configured to handle, causing overloading.
This series removes the scsi_debug submit queue concept and uses
pre-existing APIs to manage and examine tags, like scsi_block_requests()
and blk_mq_tagset_busy_iter(). Using standard APIs makes the driver more
maintainable and extensible in future.
A restriction is also added to allow sdebug_max_queue only be modified when
no shosts are present, i.e. we need to remove shosts, modify
sdebug_max_queue, and then re-add the shosts.
Link: https://lore.kernel.org/r/20230327074310.1862889-1-john.g.garry@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
It's easy to get scsi_debug to error on throughput testing when we have
multiple shosts:
$ lsscsi
[7:0:0:0] disk Linux scsi_debug 0191
[0:0:0:0] disk Linux scsi_debug 0191
$ fio --filename=/dev/sda --filename=/dev/sdb --direct=1 --rw=read --bs=4k
--iodepth=256 --runtime=60 --numjobs=40 --time_based --name=jpg
--eta-newline=1 --readonly --ioengine=io_uring --hipri --exitall_on_error
jpg: (g=0): rw=read, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=io_uring, iodepth=256
...
fio-3.28
Starting 40 processes
[ 27.521809] hrtimer: interrupt took 33067 ns
[ 27.904660] sd 7:0:0:0: [sdb] tag#171 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s
[ 27.904660] sd 0:0:0:0: [sda] tag#58 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s
fio: io_u error [ 27.904667] sd 0:0:0:0: [sda] tag#58 CDB: Read(10) 28 00 00 00 27 00 00 01 18 00
on file /dev/sda[ 27.904670] sd 0:0:0:0: [sda] tag#62 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s
The issue is related to how the driver manages submit queues and tags. A
single array of submit queues - sdebug_q_arr - with its own set of tags is
shared among all shosts. As such, for occasions when we have more than one
shost it is possible to overload the submit queues and run out of tags.
The struct sdebug_queue is to manage tags and hold the associated
queued command entry pointer (for that tag).
Since the tagset iters are now used for functions like
sdebug_blk_mq_poll(), there is no need to manage these queues. Indeed,
blk-mq already provides what we need for managing tags and queues.
Drop sdebug_queue and all its usage in the driver.
Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230327074310.1862889-12-john.g.garry@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The shost->can_queue value is initially used to set per-HW queue context
tag depth in the block layer. This ensures that the shost is not sent too
many commands which it can deal with. However lowering sdebug_max_queue
separately means that we can easily overload the shost, as in the following
example:
$ cat /sys/bus/pseudo/drivers/scsi_debug/max_queue
192
$ cat /sys/class/scsi_host/host0/can_queue
192
$ echo 100 > /sys/bus/pseudo/drivers/scsi_debug/max_queue
$ cat /sys/class/scsi_host/host0/can_queue
192
$ fio --filename=/dev/sda --direct=1 --rw=read --bs=4k --iodepth=256
--runtime=1200 --numjobs=10 --time_based --group_reporting
--name=iops-test-job --eta-newline=1 --readonly --ioengine=io_uring
--hipri --exitall_on_error
iops-test-job: (g=0): rw=read, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=io_uring, iodepth=256
...
fio-3.28
Starting 10 processes
[ 111.269885] scsi_io_completion_action: 400 callbacks suppressed
[ 111.269885] blk_print_req_error: 400 callbacks suppressed
[ 111.269889] I/O error, dev sda, sector 440 op 0x0:(READ) flags 0x1200000 phys_seg 1 prio class 2
[ 111.269892] sd 0:0:0:0: [sda] tag#132 FAILED Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK cmd_age=0s
[ 111.269897] sd 0:0:0:0: [sda] tag#132 CDB: Read(10) 28 00 00 00 01 68 00 00 08 00
[ 111.277058] I/O error, dev sda, sector 360 op 0x0:(READ) flags 0x1200000 phys_seg 1 prio class 2
[...]
Ensure that this cannot happen by allowing sdebug_max_queue be modified
only when we have no shosts. As such, any shost->can_queue value will match
sdebug_max_queue, and sdebug_max_queue cannot be modified separately.
Since retired_max_queue is no longer set, remove support.
Continue to apply the restriction that sdebug_host_max_queue cannot be
modified when sdebug_host_max_queue is set. Adding support for that would
mean extra code, and no one has complained about this restriction
previously.
A command like the following may be used to remove a shost:
echo -1 > /sys/bus/pseudo/drivers/scsi_debug/add_host
Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230327074310.1862889-11-john.g.garry@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The functions to update ndelay and delay value first check whether we have
any in-flight IO for any host. It does this by checking if any tag is used
in the global submit queues.
We can achieve the same by setting the host as blocked and then ensuring
that we have no in-flight commands with scsi_host_busy().
Note that scsi_host_busy() checks SCMD_STATE_INFLIGHT flag, which is only
set per command after we ensure that the host is not blocked, i.e. we see
more commands active after the check for scsi_host_busy() returns 0.
Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230327074310.1862889-10-john.g.garry@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Eventually we will drop the sdebug_queue struct as it is not really
required, so start with making the sdebug_queued_cmd dynamically allocated
for the lifetime of the scsi_cmnd in the driver.
As an interim measure, make sdebug_queued_cmd.sd_dp a pointer to struct
sdebug_defer. Also keep a value of the index allocated in
sdebug_queued_cmd.qc_arr in struct sdebug_queued_cmd.
To deal with an races in accessing the scsi cmnd allocated struct
sdebug_queued_cmd, add a spinlock for the scsi command in its priv area.
Races may be between scheduling a command for completion, aborting a
command, and the command actually completing and freeing the struct
sdebug_queued_cmd.
[mkp: typo fix]
Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230327074310.1862889-7-john.g.garry@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
There is no reason that calls to block_unblock_all_queues() from different
context can't race with one another, so protect with the
sdebug_host_list_mutex. There's no need for a more fine-grained per shost
locking here (and we don't have a per-host lock anyway).
Also simplify some touched code in sdebug_change_qdepth().
Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230327074310.1862889-5-john.g.garry@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In clear_luns_changed_on_target(), we iter all devices for all shosts to
conditionally clear the SDEBUG_UA_LUNS_CHANGED flag in the per-device
uas_bm.
One condition to see whether we clear the flag is to test whether the host
for the device under consideration is the same as the matching device's
(devip) host. This check will only ever pass for devices for the same
shost, so only iter the devices for the matching device shost.
We can now drop the spinlock'ing of the sdebug_host_list_lock in the same
function. This will allow us to use a mutex instead of the spinlock for the
global shost lock, as clear_luns_changed_on_target() could be called in
non-blocking context, in scsi_debug_queuecommand() -> make_ua() ->
clear_luns_changed_on_target() (which is why required a spinlock).
Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230327074310.1862889-3-john.g.garry@oracle.com
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When the current status of the host controller is suspended, enabling a
local PHY just after disabling all local PHYs in expander environment, a
hang as follows occurs:
[ 486.854655] INFO: task kworker/u256:1:899 blocked for more than 120 seconds.
[ 486.862207] Not tainted 6.1.0-rc4+ #1
[ 486.870545] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 486.878893] task:kworker/u256:1 state:D stack:0 pid:899 ppid:2 flags:0x00000008
[ 486.887745] Workqueue: 0000:74:02.0_disco_q sas_discover_domain [libsas]
[ 486.894704] Call trace:
[ 486.897400] __switch_to+0xf0/0x170
[ 486.901146] __schedule+0x3e4/0x1160
[ 486.904970] schedule+0x64/0x104
[ 486.908442] rpm_resume+0x158/0x6a0
[ 486.912163] __pm_runtime_resume+0x5c/0x84
[ 486.916489] smp_execute_task_sg+0x1f8/0x264 [libsas]
[ 486.921773] sas_discover_expander.part.0+0xbc/0x720 [libsas]
[ 486.927750] sas_discover_root_expander+0x90/0x154 [libsas]
[ 486.933552] sas_discover_domain+0x444/0x6d0 [libsas]
[ 486.938826] process_one_work+0x1e0/0x450
[ 486.943057] worker_thread+0x150/0x44c
[ 486.947015] kthread+0x114/0x120
[ 486.950447] ret_from_fork+0x10/0x20
[ 486.954292] INFO: task kworker/u256:2:1780 blocked for more than 120 seconds.
[ 486.961637] Not tainted 6.1.0-rc4+ #1
[ 486.966087] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 486.974356] task:kworker/u256:2 state:D stack:0 pid:1780 ppid:2 flags:0x00000208
[ 486.983141] Workqueue: 0000:74:02.0_event_q sas_port_event_worker [libsas]
[ 486.990252] Call trace:
[ 486.992930] __switch_to+0xf0/0x170
[ 486.996645] __schedule+0x3e4/0x1160
[ 487.000439] schedule+0x64/0x104
[ 487.003886] schedule_timeout+0x17c/0x1c0
[ 487.008102] wait_for_completion+0x7c/0x160
[ 487.012488] __flush_workqueue+0x104/0x3e0
[ 487.016782] sas_porte_bytes_dmaed+0x414/0x454 [libsas]
[ 487.022203] sas_port_event_worker+0x38/0x60 [libsas]
[ 487.027449] process_one_work+0x1e0/0x450
[ 487.031645] worker_thread+0x150/0x44c
[ 487.035594] kthread+0x114/0x120
[ 487.039017] ret_from_fork+0x10/0x20
[ 487.042828] INFO: task bash:11488 blocked for more than 121 seconds.
[ 487.049366] Not tainted 6.1.0-rc4+ #1
[ 487.053746] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 487.061953] task:bash state:D stack:0 pid:11488 ppid:10977 flags:0x00000204
[ 487.070698] Call trace:
[ 487.073355] __switch_to+0xf0/0x170
[ 487.077050] __schedule+0x3e4/0x1160
[ 487.080833] schedule+0x64/0x104
[ 487.084270] schedule_timeout+0x17c/0x1c0
[ 487.088474] wait_for_completion+0x7c/0x160
[ 487.092851] __flush_workqueue+0x104/0x3e0
[ 487.097137] drain_workqueue+0xb8/0x160
[ 487.101159] __sas_drain_work+0x50/0x90 [libsas]
[ 487.105963] sas_suspend_ha+0x64/0xd4 [libsas]
[ 487.110590] suspend_v3_hw+0x198/0x1e8 [hisi_sas_v3_hw]
[ 487.115989] pci_pm_runtime_suspend+0x5c/0x1d0
[ 487.120606] __rpm_callback+0x50/0x150
[ 487.124535] rpm_callback+0x74/0x80
[ 487.128204] rpm_suspend+0x110/0x640
[ 487.131955] rpm_idle+0x1f4/0x2d0
[ 487.135447] __pm_runtime_idle+0x58/0x94
[ 487.139538] queue_phy_enable+0xcc/0xf0 [libsas]
[ 487.144330] store_sas_phy_enable+0x74/0x100
[ 487.148770] dev_attr_store+0x20/0x34
[ 487.152606] sysfs_kf_write+0x4c/0x5c
[ 487.156437] kernfs_fop_write_iter+0x120/0x1b0
[ 487.161049] vfs_write+0x2d0/0x36c
[ 487.164625] ksys_write+0x70/0x100
[ 487.168194] __arm64_sys_write+0x24/0x30
[ 487.172280] invoke_syscall+0x50/0x120
[ 487.176186] el0_svc_common.constprop.0+0x168/0x190
[ 487.181214] do_el0_svc+0x34/0xc0
[ 487.184680] el0_svc+0x2c/0xb4
[ 487.187879] el0t_64_sync_handler+0xb8/0xbc
[ 487.192205] el0t_64_sync+0x19c/0x1a0
We find that when all local PHYs are disabled, all the devices will be
removed, the ->runtime_suspend() callback suspend_v3_hw() directly execute
since the controller usage count drop to 0. On the other side, the first
local PHY is enabled through the sysfs interface, and ensures that function
phy_up_v3_hw() is completed due to suspend_v3_hw()->
interrupt_disable_v3_hw(). In the expander scenario,
sas_discover_root_expander() is executed in event work
DISCE_DISCOVER_DOMAIN, which will increases the controller usage count and
carry out a resume and sends SMPIO, it cannot be completed because the
runtime PM status of the controller is RPM_SUSPENDING. At the same time,
the ->runtime_suspend() callback suspend_v3_hw() also cannot complete the
process because of drain libsas event queue in sas_suspend_ha(), so hung
occurs.
(thread 1) | (thread 2)
... |
rpm_idle() |
... |
__update_runtime_status(RPM_SUSPENDING)|
... | ...
suspend_v3_hw() | smp_execute_task_sg()
... | ...
interrupt_disable_v3_hw() | pm_runtime_get_sync()
| ...
... | rpm_resume() //RPM_SUSPENDING
|
__sas_drain_work() |
To fix this, check if the current runtime PM status of the controller
allows to be suspended continue after interrupt_disable_v3_hw(), return
immediately if not.
Signed-off-by: Yihang Li <liyihang9@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hislicon.com>
Link: https://lore.kernel.org/r/1679283265-115066-5-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For the controller reset operation, hisi_sas_phy_enable() is executed for
each enabled local PHY, and refresh the port id of each device based on the
latest hisi_sas_phy->port_id after 1 second sleep, hisi_sas_phy->port_id is
configured in the interrupt processing function phy_up_v3_hw(). However, in
directly attached scenario, for some SATA disks the amount of time for
phyup more than 1s sometimes. In this case, incorrect port id may be
configured in hisi_sas_refresh_port_id(). As a result, all the internal
IOs fail and disk lost, such as follows:
[10717.666565] hisi_sas_v3_hw 0000:74:02.0: phyup: phy1 link_rate=10(sata)
[10718.826813] hisi_sas_v3_hw 0000:74:02.0: erroneous completion iptt=63
task=00000000c1ab1c2b dev id=200 addr=5000000000000501 CQ hdr: 0x8000007 0xc8003f 0x0
0x0 Error info: 0x0 0x0 0x0 0x0
[10718.843428] sas: TMF task open reject failed 5000000000000501
[10718.849242] hisi_sas_v3_hw 0000:74:02.0: erroneous completion iptt=64
task=00000000c1ab1c2b dev id=200 addr=5000000000000501 CQ hdr: 0x8000007 0xc80040 0x0
0x0 Error info: 0x0 0x0 0x0 0x0
[10718.865856] sas: TMF task open reject failed 5000000000000501
[10718.871670] hisi_sas_v3_hw 0000:74:02.0: erroneous completion iptt=65
task=00000000c1ab1c2b dev id=200 addr=5000000000000501 CQ hdr: 0x8000007 0xc80041 0x0
0x0 Error info: 0x0 0x0 0x0 0x0
[10718.888284] sas: TMF task open reject failed 5000000000000501
[10718.894093] sas: executing TMF for 5000000000000501 failed after 3 attempts!
[10718.901114] hisi_sas_v3_hw 0000:74:02.0: ata disk 5000000000000501 reset failed
[10718.908410] hisi_sas_v3_hw 0000:74:02.0: controller reset complete
.....
[10773.298633] ata216.00: revalidation failed (errno=-19)
[10773.303753] ata216.00: disable device
So the time of waitting for PHYs up is 1s which may be not enough. To solve
the issue, running hisi_sas_phy_enable() in parallel through async
operations and use wait_for_completion_timeout() to wait for PHYs come up
instead of directly sleep for 1 second.
Signed-off-by: Yihang Li <liyihang9@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Link: https://lore.kernel.org/r/1679283265-115066-4-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If an NCQ error occurs when the IPTT is valid and slot->abort flag is set
in completion path, sas_task_abort() will be called to abort only one NCQ
command now, and the host would be set to SHOST_RECOVERY state. But this
may not kick-off EH Immediately until other outstanding QCs timeouts. As a
result, the host may remain in the SHOST_RECOVERY state for up to 30
seconds, such as follows:
[7972317.645234] hisi_sas_v3_hw 0000:74:04.0: erroneous completion iptt=3264 task=00000000466116b8 dev id=2 sas_addr=0x5000000000000502 CQ hdr: 0x1883 0x20cc0 0x40000 0x20420000 Error info: 0x0 0x0 0x200000 0x0
[7972341.508264] sas: Enter sas_scsi_recover_host busy: 32 failed: 32
[7972341.984731] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 32 tries: 1
All NCQ commands that are in the queue should be aborted when an NCQ error
occurs in this scenario.
Fixes: 05d91b557a ("scsi: hisi_sas: Directly trigger SCSI error handling for completion errors")
Signed-off-by: Xingui Yang <yangxingui@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Link: https://lore.kernel.org/r/1679283265-115066-3-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Some USB-SATA adapters have broken behavior when an unsupported VPD page is
probed: Depending on the VPD page number, a 4-byte header with a valid VPD
page number but with a 0 length is returned. Currently, scsi_vpd_inquiry()
only checks that the page number is valid to determine if the page is
valid, which results in receiving only the 4-byte header for the
non-existent page. This error manifests itself very often with page 0xb9
for the Concurrent Positioning Ranges detection done by sd_read_cpr(),
resulting in the following error message:
sd 0:0:0:0: [sda] Invalid Concurrent Positioning Ranges VPD page
Prevent such misleading error message by adding a check in
scsi_vpd_inquiry() to verify that the page length is not 0.
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Link: https://lore.kernel.org/r/20230322022211.116327-1-damien.lemoal@opensource.wdc.com
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
mptlan_probe() calls mpt_register_lan_device() which initializes the
&priv->post_buckets_task workqueue. A call to
mpt_lan_wake_post_buckets_task() will subsequently start the work.
During driver unload in mptlan_remove() the following race may occur:
CPU0 CPU1
|mpt_lan_post_receive_buckets_work()
mptlan_remove() |
free_netdev() |
kfree(dev); |
|
| dev->mtu
| //use
Fix this by finishing the work prior to cleaning up in mptlan_remove().
[mkp: we really should remove mptlan instead of attempting to fix it]
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Link: https://lore.kernel.org/r/20230318081635.796479-1-zyytlz.wz@163.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
