block: fix NULL pointer dereference in account statistics with IDE
PATCH 1/1 block: fix NULL pointer dereference in account statistics with IDE
The IDE driver creates some passthru requests which never get submitted to the block layer in such a way that blk_account_io_start() gets called. However, the driver still calls __blk_mq_end_request() in ide_end_rq() which will call blk_account_io_completion() which tries to dereferences req->part which is never set. See ide_prep_sense() for an example of where these requests come from.
To fix this, blk_account_io_completion() and blk_account_io_done() should do nothing if req->part is not set.
The back trace of this bug is:
BUG: kernel NULL pointer dereference, address: 000002ac
#PF: supervisor write access in kernel mode
#PF: error_code(0x0002) - not-present page
*pde = 00000000
Oops: 0002 [#1]
CPU: 0 PID: 237 Comm: kworker/0:1H Not tainted
5.4.0-rc2-00011-g48d9b0d43105e #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1
04/01/2014
Workqueue: kblockd drive_rq_insert_work
EIP: blk_account_io_completion+0x7a/0xf0
Code: 89 54 24 08 31 d2 89 4c 24 04 31 c9 c7 04 24 02 00 00 00 c1 ee
09 e8 f5 21 a6 ff e8 70 5c a7 ff 8b 53 60 8d 04 bd 00 00 00 00 <01> b4
02 ac 02 00 00 8b 9a 88 02 00 00 85 db 74 11 85 d2 74 51 8b
EAX: 00000000 EBX: f5b80000 ECX: 00000000 EDX: 00000000
ESI: 00000000 EDI: 00000000 EBP: f3031e70 ESP: f3031e54
DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 EFLAGS: 00010046
CR0: 80050033 CR2: 000002ac CR3: 03c25000 CR4: 000406d0
Call Trace:
<IRQ>
blk_update_request+0x85/0x420
ide_end_rq+0x38/0xa0
ide_complete_rq+0x3d/0x70
cdrom_newpc_intr+0x258/0xba0
ide_intr+0x135/0x250
__handle_irq_event_percpu+0x3e/0x250
handle_irq_event_percpu+0x1f/0x50
handle_irq_event+0x32/0x60
handle_level_irq+0x6c/0x110
handle_irq+0x72/0xa0
</IRQ>
do_IRQ+0x45/0xad
common_interrupt+0x115/0x11c--- block/blk-core.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
block/blk-core.c +3 -2
@@ -1310,7 +1310,7 @@ EXPORT_SYMBOL_GPL(blk_rq_err_bytes);void blk_account_io_completion (struct request *req,unsigned int bytes) { -if (blk_do_io_stat (req)) { +if (req->part &&blk_do_io_stat (req)) {const int sgrp =op_stat_group (req_op (req));struct hd_struct *part;
@@ -1328,7 +1328,8 @@ void blk_account_io_done(struct request *req, u64 now)* normal IO on queueing nor completion. Accounting the * containing request is enough. */ -if (blk_do_io_stat (req) && !(req->rq_flags &RQF_FLUSH_SEQ )) { +if (req->part &&blk_do_io_stat (req) && + !(req->rq_flags &RQF_FLUSH_SEQ )) {const int sgrp =op_stat_group (req_op (req));struct hd_struct *part;
Signature · 1 line
2.20.1
Jens Axboe
maintainerQuoted 37 lines · Logan Gunthorpe
The IDE driver creates some passthru requests which never get submitted to the block layer in such a way that
blk_account_io_start()gets called. However, the driver still calls__blk_mq_end_request()inide_end_rq()which will callblk_account_io_completion()which tries to dereferencesreq->partwhich is never set. Seeide_prep_sense()for an example of where these requests come from.To fix this,
blk_account_io_completion()andblk_account_io_done()should do nothing ifreq->partis not set.The back trace of this bug is:
BUG: kernel NULL pointer dereference, address: 000002ac #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page *pde = 00000000 Oops: 0002 [#1] CPU: 0 PID: 237 Comm: kworker/0:1H Not tainted 5.4.0-rc2-00011-g48d9b0d43105e #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Workqueue: kblockd drive_rq_insert_work EIP: blk_account_io_completion+0x7a/0xf0 Code: 89 54 24 08 31 d2 89 4c 24 04 31 c9 c7 04 24 02 00 00 00 c1 ee 09 e8 f5 21 a6 ff e8 70 5c a7 ff 8b 53 60 8d 04 bd 00 00 00 00 <01> b4 02 ac 02 00 00 8b 9a 88 02 00 00 85 db 74 11 85 d2 74 51 8b EAX: 00000000 EBX: f5b80000 ECX: 00000000 EDX: 00000000 ESI: 00000000 EDI: 00000000 EBP: f3031e70 ESP: f3031e54 DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 EFLAGS: 00010046 CR0: 80050033 CR2: 000002ac CR3: 03c25000 CR4: 000406d0 Call Trace: <IRQ> blk_update_request+0x85/0x420 ide_end_rq+0x38/0xa0 ide_complete_rq+0x3d/0x70 cdrom_newpc_intr+0x258/0xba0 ide_intr+0x135/0x250
Quoted 8 lines
__handle_irq_event_percpu+0x3e/0x250
handle_irq_event_percpu+0x1f/0x50
handle_irq_event+0x32/0x60
handle_level_irq+0x6c/0x110
handle_irq+0x72/0xa0
</IRQ>
do_IRQ+0x45/0xad
common_interrupt+0x115/0x11cWhy not just:
block/blk.h +1 -1
@@ -250,7 +250,7 @@ int blk_dev_init(void);*/ static inline bool blk_do_io_stat (struct request *rq) { -return rq->rq_disk && (rq->rq_flags &RQF_IO_STAT ); +return rq->part && rq->rq_disk && (rq->rq_flags &RQF_IO_STAT ); }static inline void req_set_nomerge (struct request_queue *q,struct request *req)
Signature · 1 line
Jens Axboe
Quoted 54 lines · Jens Axboe
Quoted 37 lines · Logan Gunthorpe
The IDE driver creates some passthru requests which never get submitted to the block layer in such a way that
blk_account_io_start()gets called. However, the driver still calls__blk_mq_end_request()inide_end_rq()which will callblk_account_io_completion()which tries to dereferencesreq->partwhich is never set. Seeide_prep_sense()for an example of where these requests come from.To fix this,
blk_account_io_completion()andblk_account_io_done()should do nothing ifreq->partis not set.The back trace of this bug is:
BUG: kernel NULL pointer dereference, address: 000002ac #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page *pde = 00000000 Oops: 0002 [#1] CPU: 0 PID: 237 Comm: kworker/0:1H Not tainted 5.4.0-rc2-00011-g48d9b0d43105e #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Workqueue: kblockd drive_rq_insert_work EIP: blk_account_io_completion+0x7a/0xf0 Code: 89 54 24 08 31 d2 89 4c 24 04 31 c9 c7 04 24 02 00 00 00 c1 ee 09 e8 f5 21 a6 ff e8 70 5c a7 ff 8b 53 60 8d 04 bd 00 00 00 00 <01> b4 02 ac 02 00 00 8b 9a 88 02 00 00 85 db 74 11 85 d2 74 51 8b EAX: 00000000 EBX: f5b80000 ECX: 00000000 EDX: 00000000 ESI: 00000000 EDI: 00000000 EBP: f3031e70 ESP: f3031e54 DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 EFLAGS: 00010046 CR0: 80050033 CR2: 000002ac CR3: 03c25000 CR4: 000406d0 Call Trace: <IRQ> blk_update_request+0x85/0x420 ide_end_rq+0x38/0xa0 ide_complete_rq+0x3d/0x70 cdrom_newpc_intr+0x258/0xba0 ide_intr+0x135/0x250Quoted 8 lines
__handle_irq_event_percpu+0x3e/0x250 handle_irq_event_percpu+0x1f/0x50 handle_irq_event+0x32/0x60 handle_level_irq+0x6c/0x110 handle_irq+0x72/0xa0 </IRQ> do_IRQ+0x45/0xad common_interrupt+0x115/0x11cWhy not just:
block/blk.h +0
-0
block/blk.h int blk_dev_init(void); · line 254
Quoted 6 lines
block/blk.h +1
-1@@ -250,7 +250,7 @@ int blk_dev_init(void);*/ static inline bool blk_do_io_stat (struct request *rq) { -return rq->rq_disk && (rq->rq_flags &RQF_IO_STAT ); +return rq->part && rq->rq_disk && (rq->rq_flags &RQF_IO_STAT ); }
Because blk_account_io_start() also checks blk_do_io_stat() and, in that case, rq->part will never be set (seeing that's the function that typically sets it); thus that solution would disable stats entirely.
Logan
Jens Axboe
maintainerQuoted 52 lines · Logan Gunthorpe
Quoted 49 lines · Jens Axboe
Quoted 37 lines · Logan Gunthorpe
The IDE driver creates some passthru requests which never get submitted to the block layer in such a way that
blk_account_io_start()gets called. However, the driver still calls__blk_mq_end_request()inide_end_rq()which will callblk_account_io_completion()which tries to dereferencesreq->partwhich is never set. Seeide_prep_sense()for an example of where these requests come from.To fix this,
blk_account_io_completion()andblk_account_io_done()should do nothing ifreq->partis not set.The back trace of this bug is:
BUG: kernel NULL pointer dereference, address: 000002ac #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page *pde = 00000000 Oops: 0002 [#1] CPU: 0 PID: 237 Comm: kworker/0:1H Not tainted 5.4.0-rc2-00011-g48d9b0d43105e #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Workqueue: kblockd drive_rq_insert_work EIP: blk_account_io_completion+0x7a/0xf0 Code: 89 54 24 08 31 d2 89 4c 24 04 31 c9 c7 04 24 02 00 00 00 c1 ee 09 e8 f5 21 a6 ff e8 70 5c a7 ff 8b 53 60 8d 04 bd 00 00 00 00 <01> b4 02 ac 02 00 00 8b 9a 88 02 00 00 85 db 74 11 85 d2 74 51 8b EAX: 00000000 EBX: f5b80000 ECX: 00000000 EDX: 00000000 ESI: 00000000 EDI: 00000000 EBP: f3031e70 ESP: f3031e54 DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 EFLAGS: 00010046 CR0: 80050033 CR2: 000002ac CR3: 03c25000 CR4: 000406d0 Call Trace: <IRQ> blk_update_request+0x85/0x420 ide_end_rq+0x38/0xa0 ide_complete_rq+0x3d/0x70 cdrom_newpc_intr+0x258/0xba0 ide_intr+0x135/0x250Quoted 8 lines
__handle_irq_event_percpu+0x3e/0x250 handle_irq_event_percpu+0x1f/0x50 handle_irq_event+0x32/0x60 handle_level_irq+0x6c/0x110 handle_irq+0x72/0xa0 </IRQ> do_IRQ+0x45/0xad common_interrupt+0x115/0x11cWhy not just:
Quoted 15 lines
block/blk.h int blk_dev_init(void); · line 254
Quoted 11 lines
block/blk.h +1
-1@@ -250,7 +250,7 @@ int blk_dev_init(void);*/ static inline bool blk_do_io_stat (struct request *rq) { -return rq->rq_disk && (rq->rq_flags &RQF_IO_STAT ); +return rq->part && rq->rq_disk && (rq->rq_flags &RQF_IO_STAT ); }Because
blk_account_io_start()also checksblk_do_io_stat()and, in that case,rq->partwill never be set (seeing that's the function that typically sets it); thus that solution would disable stats entirely.
Gotcha. I'm fine with the patch you posted in that case.
Signature · 1 line
Jens Axboe