s390/dasd: use blk_mq_rq_from_pdu for per request data
[ Upstream commit
f0f59a2fab8e52b9d582b39da39f22230ca80aee ]
Dasd uses completion_data from struct request to store per request
private data - this is problematic since this member is part of a
union which is also used by IO schedulers.
Let the block layer maintain space for per request data behind each
struct request.
Fixes crashes on block layer timeouts like this one:
Unable to handle kernel pointer dereference in virtual kernel address space
Failing address:
0000000000000000 TEID:
0000000000000483
Fault in home space mode while using kernel ASCE.
AS:
0000000001308007 R3:
00000000fffc8007 S:
00000000fffcc000 P:
000000000000013d
Oops: 0004 ilc:2 [#1] PREEMPT SMP
Modules linked in: [...]
CPU: 0 PID: 1480 Comm: kworker/0:2H Not tainted 4.17.0-rc4-00046-gaa3bcd43b5af #203
Hardware name: IBM 3906 M02 702 (LPAR)
Workqueue: kblockd blk_mq_timeout_work
Krnl PSW :
0000000067ac406b 00000000b6960308 (do_raw_spin_trylock+0x30/0x78)
R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:3 CC:2 PM:0 RI:0 EA:3
Krnl GPRS:
0000000000000c00 0000000000000000 0000000000000000 0000000000000001
0000000000b9d3c8 0000000000000000 0000000000000001 00000000cf9639d8
0000000000000000 0700000000000000 0000000000000000 000000000099f09e
0000000000000000 000000000076e9d0 000000006247bb08 000000006247bae0
Krnl Code:
00000000001c159c:
b90400c2 lgr %r12,%r2
00000000001c15a0:
a7180000 lhi %r1,0
#
00000000001c15a4:
583003a4 l %r3,932
>
00000000001c15a8:
ba132000 cs %r1,%r3,0(%r2)
00000000001c15ac:
a7180001 lhi %r1,1
00000000001c15b0:
a784000b brc 8,1c15c6
00000000001c15b4:
c0e5004e72aa brasl %r14,b8fb08
00000000001c15ba: 1812 lr %r1,%r2
Call Trace:
([<
0700000000000000>] 0x700000000000000)
[<
0000000000b9d3d2>] _raw_spin_lock_irqsave+0x7a/0xb8
[<
000000000099f09e>] dasd_times_out+0x46/0x278
[<
000000000076ea6e>] blk_mq_terminate_expired+0x9e/0x108
[<
000000000077497a>] bt_for_each+0x102/0x130
[<
0000000000774e54>] blk_mq_queue_tag_busy_iter+0x74/0xd8
[<
000000000076fea0>] blk_mq_timeout_work+0x260/0x320
[<
0000000000169dd4>] process_one_work+0x3bc/0x708
[<
000000000016a382>] worker_thread+0x262/0x408
[<
00000000001723a8>] kthread+0x160/0x178
[<
0000000000b9e73a>] kernel_thread_starter+0x6/0xc
[<
0000000000b9e734>] kernel_thread_starter+0x0/0xc
INFO: lockdep is turned off.
Last Breaking-Event-Address:
[<
0000000000b9d3cc>] _raw_spin_lock_irqsave+0x74/0xb8
Kernel panic - not syncing: Fatal exception: panic_on_oops
Signed-off-by: Sebastian Ott <sebott@linux.ibm.com>
Reviewed-by: Stefan Haberland <sth@linux.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Sasha Levin <alexander.levin@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>