* [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
@ 2021-10-13 12:19 Ye Bin
2021-10-20 3:12 ` yebin
` (2 more replies)
0 siblings, 3 replies; 8+ messages in thread
From: Ye Bin @ 2021-10-13 12:19 UTC (permalink / raw)
To: rafael, len.brown, pavel, linux-pm; +Cc: tytso, linux-kernel, jack, Ye Bin
We got follow issue:
[ 89.266592] ------------[ cut here ]------------
[ 89.267427] kernel BUG at fs/buffer.c:3020!
[ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
[ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
[ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
[ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
[ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
[ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
[ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
[ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
[ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
[ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
[ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
[ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 89.290368] Call Trace:
[ 89.290842] write_mmp_block+0x2ca/0x510
[ 89.292218] kmmpd+0x433/0x9a0
[ 89.294902] kthread+0x2dd/0x3e0
[ 89.296268] ret_from_fork+0x22/0x30
[ 89.296906] Modules linked in:
We can reproduce this issue as follow:
1. mkfs.ext4 -O mmp /dev/sda -b 1024
2. mount /dev/sda /home/test
3. echo "/dev/sda" > /sys/power/resume
4. wait a moment we will get exception
The sequence of issue is as follows:
Thread1 Thread2
mount /dev/sda /home/test
get s_mmp_bh --> has mapped flag
start kmmpd thread
echo "/dev/sda" > /sys/power/resume
resume_store
software_resume
swsusp_check
set_blocksize
truncate_inode_pages_range
truncate_cleanup_page
block_invalidatepage
discard_buffer --> clean mapped flag
write_mmp_block
submit_bh
submit_bh_wbc
BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
To solve this issue, get block device exclusively when do swsusp_check.
Signed-off-by: Ye Bin <yebin10@huawei.com>
---
kernel/power/swap.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/kernel/power/swap.c b/kernel/power/swap.c
index 9ec418955556..26c0bd2a50da 100644
--- a/kernel/power/swap.c
+++ b/kernel/power/swap.c
@@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
int swsusp_check(void)
{
int error;
+ void *holder;
hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
- FMODE_READ, NULL);
+ FMODE_READ | FMODE_EXCL, &holder);
if (!IS_ERR(hib_resume_bdev)) {
set_blocksize(hib_resume_bdev, PAGE_SIZE);
clear_page(swsusp_header);
@@ -1545,7 +1546,7 @@ int swsusp_check(void)
put:
if (error)
- blkdev_put(hib_resume_bdev, FMODE_READ);
+ blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
else
pr_debug("Image signature found, resuming\n");
} else {
--
2.31.1
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
2021-10-13 12:19 [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check Ye Bin
@ 2021-10-20 3:12 ` yebin
2021-10-20 14:26 ` yebin
2021-10-21 11:00 ` Rafael J. Wysocki
2 siblings, 0 replies; 8+ messages in thread
From: yebin @ 2021-10-20 3:12 UTC (permalink / raw)
To: rafael, len.brown, pavel, linux-pm; +Cc: tytso, linux-kernel, jack
On 2021/10/13 20:19, Ye Bin wrote:
> We got follow issue:
> [ 89.266592] ------------[ cut here ]------------
> [ 89.267427] kernel BUG at fs/buffer.c:3020!
> [ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
> [ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
> [ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
> [ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
> [ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
> [ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
> [ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
> [ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
> [ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
> [ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
> [ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
> [ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [ 89.290368] Call Trace:
> [ 89.290842] write_mmp_block+0x2ca/0x510
> [ 89.292218] kmmpd+0x433/0x9a0
> [ 89.294902] kthread+0x2dd/0x3e0
> [ 89.296268] ret_from_fork+0x22/0x30
> [ 89.296906] Modules linked in:
>
> We can reproduce this issue as follow:
> 1. mkfs.ext4 -O mmp /dev/sda -b 1024
> 2. mount /dev/sda /home/test
> 3. echo "/dev/sda" > /sys/power/resume
> 4. wait a moment we will get exception
>
> The sequence of issue is as follows:
> Thread1 Thread2
> mount /dev/sda /home/test
> get s_mmp_bh --> has mapped flag
> start kmmpd thread
> echo "/dev/sda" > /sys/power/resume
> resume_store
> software_resume
> swsusp_check
> set_blocksize
> truncate_inode_pages_range
> truncate_cleanup_page
> block_invalidatepage
> discard_buffer --> clean mapped flag
> write_mmp_block
> submit_bh
> submit_bh_wbc
> BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
>
> To solve this issue, get block device exclusively when do swsusp_check.
>
> Signed-off-by: Ye Bin <yebin10@huawei.com>
> ---
> kernel/power/swap.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
> index 9ec418955556..26c0bd2a50da 100644
> --- a/kernel/power/swap.c
> +++ b/kernel/power/swap.c
> @@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
> int swsusp_check(void)
> {
> int error;
> + void *holder;
>
> hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
> - FMODE_READ, NULL);
> + FMODE_READ | FMODE_EXCL, &holder);
> if (!IS_ERR(hib_resume_bdev)) {
> set_blocksize(hib_resume_bdev, PAGE_SIZE);
> clear_page(swsusp_header);
> @@ -1545,7 +1546,7 @@ int swsusp_check(void)
>
> put:
> if (error)
> - blkdev_put(hib_resume_bdev, FMODE_READ);
> + blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
> else
> pr_debug("Image signature found, resuming\n");
> } else {
ping...
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
2021-10-13 12:19 [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check Ye Bin
2021-10-20 3:12 ` yebin
@ 2021-10-20 14:26 ` yebin
2021-10-21 11:00 ` Rafael J. Wysocki
2 siblings, 0 replies; 8+ messages in thread
From: yebin @ 2021-10-20 14:26 UTC (permalink / raw)
To: rafael, len.brown, pavel, linux-pm; +Cc: tytso, linux-kernel, jack
On 2021/10/13 20:19, Ye Bin wrote:
> We got follow issue:
> [ 89.266592] ------------[ cut here ]------------
> [ 89.267427] kernel BUG at fs/buffer.c:3020!
> [ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
> [ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
> [ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
> [ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
> [ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
> [ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
> [ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
> [ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
> [ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
> [ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
> [ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
> [ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [ 89.290368] Call Trace:
> [ 89.290842] write_mmp_block+0x2ca/0x510
> [ 89.292218] kmmpd+0x433/0x9a0
> [ 89.294902] kthread+0x2dd/0x3e0
> [ 89.296268] ret_from_fork+0x22/0x30
> [ 89.296906] Modules linked in:
>
> We can reproduce this issue as follow:
> 1. mkfs.ext4 -O mmp /dev/sda -b 1024
> 2. mount /dev/sda /home/test
> 3. echo "/dev/sda" > /sys/power/resume
> 4. wait a moment we will get exception
>
> The sequence of issue is as follows:
> Thread1 Thread2
> mount /dev/sda /home/test
> get s_mmp_bh --> has mapped flag
> start kmmpd thread
> echo "/dev/sda" > /sys/power/resume
> resume_store
> software_resume
> swsusp_check
> set_blocksize
> truncate_inode_pages_range
> truncate_cleanup_page
> block_invalidatepage
> discard_buffer --> clean mapped flag
> write_mmp_block
> submit_bh
> submit_bh_wbc
> BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
>
> To solve this issue, get block device exclusively when do swsusp_check.
>
> Signed-off-by: Ye Bin <yebin10@huawei.com>
> ---
> kernel/power/swap.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
> index 9ec418955556..26c0bd2a50da 100644
> --- a/kernel/power/swap.c
> +++ b/kernel/power/swap.c
> @@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
> int swsusp_check(void)
> {
> int error;
> + void *holder;
>
> hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
> - FMODE_READ, NULL);
> + FMODE_READ | FMODE_EXCL, &holder);
> if (!IS_ERR(hib_resume_bdev)) {
> set_blocksize(hib_resume_bdev, PAGE_SIZE);
> clear_page(swsusp_header);
> @@ -1545,7 +1546,7 @@ int swsusp_check(void)
>
> put:
> if (error)
> - blkdev_put(hib_resume_bdev, FMODE_READ);
> + blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
> else
> pr_debug("Image signature found, resuming\n");
> } else {
ping...
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
2021-10-13 12:19 [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check Ye Bin
2021-10-20 3:12 ` yebin
2021-10-20 14:26 ` yebin
@ 2021-10-21 11:00 ` Rafael J. Wysocki
2021-10-21 11:38 ` yebin
2 siblings, 1 reply; 8+ messages in thread
From: Rafael J. Wysocki @ 2021-10-21 11:00 UTC (permalink / raw)
To: Ye Bin
Cc: Rafael J. Wysocki, Len Brown, Pavel Machek, Linux PM,
Ted Ts'o, Linux Kernel Mailing List, Jan Kara
On Wed, Oct 13, 2021 at 2:06 PM Ye Bin <yebin10@huawei.com> wrote:
>
> We got follow issue:
> [ 89.266592] ------------[ cut here ]------------
> [ 89.267427] kernel BUG at fs/buffer.c:3020!
> [ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
> [ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
> [ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
> [ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
> [ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
> [ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
> [ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
> [ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
> [ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
> [ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
> [ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
> [ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> [ 89.290368] Call Trace:
> [ 89.290842] write_mmp_block+0x2ca/0x510
> [ 89.292218] kmmpd+0x433/0x9a0
> [ 89.294902] kthread+0x2dd/0x3e0
> [ 89.296268] ret_from_fork+0x22/0x30
> [ 89.296906] Modules linked in:
>
> We can reproduce this issue as follow:
> 1. mkfs.ext4 -O mmp /dev/sda -b 1024
> 2. mount /dev/sda /home/test
> 3. echo "/dev/sda" > /sys/power/resume
> 4. wait a moment we will get exception
>
> The sequence of issue is as follows:
> Thread1 Thread2
> mount /dev/sda /home/test
> get s_mmp_bh --> has mapped flag
> start kmmpd thread
> echo "/dev/sda" > /sys/power/resume
> resume_store
> software_resume
> swsusp_check
> set_blocksize
> truncate_inode_pages_range
> truncate_cleanup_page
> block_invalidatepage
> discard_buffer --> clean mapped flag
> write_mmp_block
> submit_bh
> submit_bh_wbc
> BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
>
> To solve this issue, get block device exclusively when do swsusp_check.
And why exactly is this going to help?
>
> Signed-off-by: Ye Bin <yebin10@huawei.com>
> ---
> kernel/power/swap.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
> index 9ec418955556..26c0bd2a50da 100644
> --- a/kernel/power/swap.c
> +++ b/kernel/power/swap.c
> @@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
> int swsusp_check(void)
> {
> int error;
> + void *holder;
>
> hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
> - FMODE_READ, NULL);
> + FMODE_READ | FMODE_EXCL, &holder);
So you need to explain to me how this works.
> if (!IS_ERR(hib_resume_bdev)) {
> set_blocksize(hib_resume_bdev, PAGE_SIZE);
> clear_page(swsusp_header);
> @@ -1545,7 +1546,7 @@ int swsusp_check(void)
>
> put:
> if (error)
> - blkdev_put(hib_resume_bdev, FMODE_READ);
> + blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
> else
> pr_debug("Image signature found, resuming\n");
> } else {
> --
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
2021-10-21 11:00 ` Rafael J. Wysocki
@ 2021-10-21 11:38 ` yebin
2021-10-21 13:27 ` Rafael J. Wysocki
0 siblings, 1 reply; 8+ messages in thread
From: yebin @ 2021-10-21 11:38 UTC (permalink / raw)
To: Rafael J. Wysocki
Cc: Len Brown, Pavel Machek, Linux PM, Ted Ts'o,
Linux Kernel Mailing List, Jan Kara
On 2021/10/21 19:00, Rafael J. Wysocki wrote:
> On Wed, Oct 13, 2021 at 2:06 PM Ye Bin <yebin10@huawei.com> wrote:
>> We got follow issue:
>> [ 89.266592] ------------[ cut here ]------------
>> [ 89.267427] kernel BUG at fs/buffer.c:3020!
>> [ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
>> [ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
>> [ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
>> [ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
>> [ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
>> [ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
>> [ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
>> [ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
>> [ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
>> [ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
>> [ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
>> [ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> [ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> [ 89.290368] Call Trace:
>> [ 89.290842] write_mmp_block+0x2ca/0x510
>> [ 89.292218] kmmpd+0x433/0x9a0
>> [ 89.294902] kthread+0x2dd/0x3e0
>> [ 89.296268] ret_from_fork+0x22/0x30
>> [ 89.296906] Modules linked in:
>>
>> We can reproduce this issue as follow:
>> 1. mkfs.ext4 -O mmp /dev/sda -b 1024
>> 2. mount /dev/sda /home/test
>> 3. echo "/dev/sda" > /sys/power/resume
>> 4. wait a moment we will get exception
>>
>> The sequence of issue is as follows:
>> Thread1 Thread2
>> mount /dev/sda /home/test
>> get s_mmp_bh --> has mapped flag
>> start kmmpd thread
>> echo "/dev/sda" > /sys/power/resume
>> resume_store
>> software_resume
>> swsusp_check
>> set_blocksize
>> truncate_inode_pages_range
>> truncate_cleanup_page
>> block_invalidatepage
>> discard_buffer --> clean mapped flag
>> write_mmp_block
>> submit_bh
>> submit_bh_wbc
>> BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
>>
>> To solve this issue, get block device exclusively when do swsusp_check.
> And why exactly is this going to help?
If a block device is already mounted, then do resume, it will discard
buffer
which file system being used. That can cause the file system to crash and
may even cause the system to reset abnormally.
If the "/sys/power/resume" sysfs interface is unrestricted, it will provide
criminals with a way to attack the system.
>> Signed-off-by: Ye Bin <yebin10@huawei.com>
>> ---
>> kernel/power/swap.c | 5 +++--
>> 1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
>> index 9ec418955556..26c0bd2a50da 100644
>> --- a/kernel/power/swap.c
>> +++ b/kernel/power/swap.c
>> @@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
>> int swsusp_check(void)
>> {
>> int error;
>> + void *holder;
>>
>> hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
>> - FMODE_READ, NULL);
>> + FMODE_READ | FMODE_EXCL, &holder);
> So you need to explain to me how this works.
As we call 'blkdev_get_by_path' in 'mount_bdev' pass mode includes
FMODE_EXCL.
So if 'swsusp_check' try to get block device with mode includes
FMODE_EXCL will failed.
I think if a block device is used as a swap partition, it cannot be
mounted at the
same time. Conversely, if a block device is already mounted, it cannot
be used as
a swap partition.
>
>> if (!IS_ERR(hib_resume_bdev)) {
>> set_blocksize(hib_resume_bdev, PAGE_SIZE);
>> clear_page(swsusp_header);
>> @@ -1545,7 +1546,7 @@ int swsusp_check(void)
>>
>> put:
>> if (error)
>> - blkdev_put(hib_resume_bdev, FMODE_READ);
>> + blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
>> else
>> pr_debug("Image signature found, resuming\n");
>> } else {
>> --
> .
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
2021-10-21 11:38 ` yebin
@ 2021-10-21 13:27 ` Rafael J. Wysocki
2021-10-21 14:38 ` yebin
0 siblings, 1 reply; 8+ messages in thread
From: Rafael J. Wysocki @ 2021-10-21 13:27 UTC (permalink / raw)
To: yebin
Cc: Rafael J. Wysocki, Len Brown, Pavel Machek, Linux PM,
Ted Ts'o, Linux Kernel Mailing List, Jan Kara
On Thu, Oct 21, 2021 at 1:38 PM yebin <yebin10@huawei.com> wrote:
>
>
>
> On 2021/10/21 19:00, Rafael J. Wysocki wrote:
> > On Wed, Oct 13, 2021 at 2:06 PM Ye Bin <yebin10@huawei.com> wrote:
> >> We got follow issue:
> >> [ 89.266592] ------------[ cut here ]------------
> >> [ 89.267427] kernel BUG at fs/buffer.c:3020!
> >> [ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
> >> [ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
> >> [ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
> >> [ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
> >> [ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
> >> [ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
> >> [ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
> >> [ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
> >> [ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
> >> [ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
> >> [ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >> [ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
> >> [ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >> [ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >> [ 89.290368] Call Trace:
> >> [ 89.290842] write_mmp_block+0x2ca/0x510
> >> [ 89.292218] kmmpd+0x433/0x9a0
> >> [ 89.294902] kthread+0x2dd/0x3e0
> >> [ 89.296268] ret_from_fork+0x22/0x30
> >> [ 89.296906] Modules linked in:
> >>
> >> We can reproduce this issue as follow:
> >> 1. mkfs.ext4 -O mmp /dev/sda -b 1024
> >> 2. mount /dev/sda /home/test
> >> 3. echo "/dev/sda" > /sys/power/resume
> >> 4. wait a moment we will get exception
> >>
> >> The sequence of issue is as follows:
> >> Thread1 Thread2
> >> mount /dev/sda /home/test
> >> get s_mmp_bh --> has mapped flag
> >> start kmmpd thread
> >> echo "/dev/sda" > /sys/power/resume
> >> resume_store
> >> software_resume
> >> swsusp_check
> >> set_blocksize
> >> truncate_inode_pages_range
> >> truncate_cleanup_page
> >> block_invalidatepage
> >> discard_buffer --> clean mapped flag
> >> write_mmp_block
> >> submit_bh
> >> submit_bh_wbc
> >> BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
> >>
> >> To solve this issue, get block device exclusively when do swsusp_check.
> > And why exactly is this going to help?
> If a block device is already mounted, then do resume, it will discard
> buffer
> which file system being used. That can cause the file system to crash and
> may even cause the system to reset abnormally.
I see.
Is there a way to indicate in swsusp_check() that the access will be
read-only and so it doesn't care whether or not the block device is in
use by someone else?
> If the "/sys/power/resume" sysfs interface is unrestricted, it will provide
> criminals with a way to attack the system.
>
> >> Signed-off-by: Ye Bin <yebin10@huawei.com>
> >> ---
> >> kernel/power/swap.c | 5 +++--
> >> 1 file changed, 3 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
> >> index 9ec418955556..26c0bd2a50da 100644
> >> --- a/kernel/power/swap.c
> >> +++ b/kernel/power/swap.c
> >> @@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
> >> int swsusp_check(void)
> >> {
> >> int error;
> >> + void *holder;
> >>
> >> hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
> >> - FMODE_READ, NULL);
> >> + FMODE_READ | FMODE_EXCL, &holder);
> > So you need to explain to me how this works.
> As we call 'blkdev_get_by_path' in 'mount_bdev' pass mode includes
> FMODE_EXCL.
> So if 'swsusp_check' try to get block device with mode includes
> FMODE_EXCL will failed.
>
> I think if a block device is used as a swap partition, it cannot be
> mounted at the
> same time. Conversely, if a block device is already mounted, it cannot
> be used as
> a swap partition.
Well, what if a hibernation image is located in a swap file? This is
one of the cases that need to be supported.
> >
> >> if (!IS_ERR(hib_resume_bdev)) {
> >> set_blocksize(hib_resume_bdev, PAGE_SIZE);
> >> clear_page(swsusp_header);
> >> @@ -1545,7 +1546,7 @@ int swsusp_check(void)
> >>
> >> put:
> >> if (error)
> >> - blkdev_put(hib_resume_bdev, FMODE_READ);
> >> + blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
> >> else
> >> pr_debug("Image signature found, resuming\n");
> >> } else {
> >> --
> > .
> >
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
2021-10-21 13:27 ` Rafael J. Wysocki
@ 2021-10-21 14:38 ` yebin
2021-10-21 17:17 ` Rafael J. Wysocki
0 siblings, 1 reply; 8+ messages in thread
From: yebin @ 2021-10-21 14:38 UTC (permalink / raw)
To: Rafael J. Wysocki
Cc: Len Brown, Pavel Machek, Linux PM, Ted Ts'o,
Linux Kernel Mailing List, Jan Kara
On 2021/10/21 21:27, Rafael J. Wysocki wrote:
> On Thu, Oct 21, 2021 at 1:38 PM yebin <yebin10@huawei.com> wrote:
>>
>>
>> On 2021/10/21 19:00, Rafael J. Wysocki wrote:
>>> On Wed, Oct 13, 2021 at 2:06 PM Ye Bin <yebin10@huawei.com> wrote:
>>>> We got follow issue:
>>>> [ 89.266592] ------------[ cut here ]------------
>>>> [ 89.267427] kernel BUG at fs/buffer.c:3020!
>>>> [ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
>>>> [ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
>>>> [ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
>>>> [ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
>>>> [ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
>>>> [ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
>>>> [ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
>>>> [ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
>>>> [ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
>>>> [ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
>>>> [ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>> [ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
>>>> [ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>>>> [ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>>>> [ 89.290368] Call Trace:
>>>> [ 89.290842] write_mmp_block+0x2ca/0x510
>>>> [ 89.292218] kmmpd+0x433/0x9a0
>>>> [ 89.294902] kthread+0x2dd/0x3e0
>>>> [ 89.296268] ret_from_fork+0x22/0x30
>>>> [ 89.296906] Modules linked in:
>>>>
>>>> We can reproduce this issue as follow:
>>>> 1. mkfs.ext4 -O mmp /dev/sda -b 1024
>>>> 2. mount /dev/sda /home/test
>>>> 3. echo "/dev/sda" > /sys/power/resume
>>>> 4. wait a moment we will get exception
>>>>
>>>> The sequence of issue is as follows:
>>>> Thread1 Thread2
>>>> mount /dev/sda /home/test
>>>> get s_mmp_bh --> has mapped flag
>>>> start kmmpd thread
>>>> echo "/dev/sda" > /sys/power/resume
>>>> resume_store
>>>> software_resume
>>>> swsusp_check
>>>> set_blocksize
>>>> truncate_inode_pages_range
>>>> truncate_cleanup_page
>>>> block_invalidatepage
>>>> discard_buffer --> clean mapped flag
>>>> write_mmp_block
>>>> submit_bh
>>>> submit_bh_wbc
>>>> BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
>>>>
>>>> To solve this issue, get block device exclusively when do swsusp_check.
>>> And why exactly is this going to help?
>> If a block device is already mounted, then do resume, it will discard
>> buffer
>> which file system being used. That can cause the file system to crash and
>> may even cause the system to reset abnormally.
> I see.
>
> Is there a way to indicate in swsusp_check() that the access will be
> read-only and so it doesn't care whether or not the block device is in
> use by someone else?
The root reason is that call 'set_blocksize' in 'swsusp_check' which
maybe invalidate
block device's page.
Moreover, the file system and swap partition share storage space, which
will lead to
data confusion.
>> If the "/sys/power/resume" sysfs interface is unrestricted, it will provide
>> criminals with a way to attack the system.
>>
>>>> Signed-off-by: Ye Bin <yebin10@huawei.com>
>>>> ---
>>>> kernel/power/swap.c | 5 +++--
>>>> 1 file changed, 3 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
>>>> index 9ec418955556..26c0bd2a50da 100644
>>>> --- a/kernel/power/swap.c
>>>> +++ b/kernel/power/swap.c
>>>> @@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
>>>> int swsusp_check(void)
>>>> {
>>>> int error;
>>>> + void *holder;
>>>>
>>>> hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
>>>> - FMODE_READ, NULL);
>>>> + FMODE_READ | FMODE_EXCL, &holder);
>>> So you need to explain to me how this works.
>> As we call 'blkdev_get_by_path' in 'mount_bdev' pass mode includes
>> FMODE_EXCL.
>> So if 'swsusp_check' try to get block device with mode includes
>> FMODE_EXCL will failed.
>>
>> I think if a block device is used as a swap partition, it cannot be
>> mounted at the
>> same time. Conversely, if a block device is already mounted, it cannot
>> be used as
>> a swap partition.
> Well, what if a hibernation image is located in a swap file? This is
> one of the cases that need to be supported.
I don't know much about this, but I think the process can call
swsusp_check, the
hibernation image must be loaded into the block device.
I refer to the process of swapon. If the swap partition is a block
device, it is also
opened exclusively.
swapon
claim_swapfile
static int claim_swapfile(struct swap_info_struct *p, struct inode *inode)
{
int error;
if (S_ISBLK(inode->i_mode)) { -->If a block device is
specified, it is also opened exclusively here.
p->bdev = blkdev_get_by_dev(inode->i_rdev,
FMODE_READ | FMODE_WRITE |
FMODE_EXCL, p);
if (IS_ERR(p->bdev)) {
error = PTR_ERR(p->bdev);
p->bdev = NULL;
return error;
}
p->old_block_size = block_size(p->bdev);
error = set_blocksize(p->bdev, PAGE_SIZE);
if (error < 0)
return error;
/*
* Zoned block devices contain zones that have a sequential
* write only restriction. Hence zoned block devices
are not
* suitable for swapping. Disallow them here.
*/
if (blk_queue_is_zoned(p->bdev->bd_disk->queue))
return -EINVAL;
p->flags |= SWP_BLKDEV;
} else if (S_ISREG(inode->i_mode)) {
p->bdev = inode->i_sb->s_bdev;
}
return 0;
}
>>>> if (!IS_ERR(hib_resume_bdev)) {
>>>> set_blocksize(hib_resume_bdev, PAGE_SIZE);
>>>> clear_page(swsusp_header);
>>>> @@ -1545,7 +1546,7 @@ int swsusp_check(void)
>>>>
>>>> put:
>>>> if (error)
>>>> - blkdev_put(hib_resume_bdev, FMODE_READ);
>>>> + blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
>>>> else
>>>> pr_debug("Image signature found, resuming\n");
>>>> } else {
>>>> --
>>> .
>>>
> .
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check
2021-10-21 14:38 ` yebin
@ 2021-10-21 17:17 ` Rafael J. Wysocki
0 siblings, 0 replies; 8+ messages in thread
From: Rafael J. Wysocki @ 2021-10-21 17:17 UTC (permalink / raw)
To: yebin
Cc: Rafael J. Wysocki, Len Brown, Pavel Machek, Linux PM,
Ted Ts'o, Linux Kernel Mailing List, Jan Kara
On Thu, Oct 21, 2021 at 4:38 PM yebin <yebin10@huawei.com> wrote:
>
>
>
> On 2021/10/21 21:27, Rafael J. Wysocki wrote:
> > On Thu, Oct 21, 2021 at 1:38 PM yebin <yebin10@huawei.com> wrote:
> >>
> >>
> >> On 2021/10/21 19:00, Rafael J. Wysocki wrote:
> >>> On Wed, Oct 13, 2021 at 2:06 PM Ye Bin <yebin10@huawei.com> wrote:
> >>>> We got follow issue:
> >>>> [ 89.266592] ------------[ cut here ]------------
> >>>> [ 89.267427] kernel BUG at fs/buffer.c:3020!
> >>>> [ 89.268264] invalid opcode: 0000 [#1] SMP KASAN PTI
> >>>> [ 89.269116] CPU: 7 PID: 1750 Comm: kmmpd-loop0 Not tainted 5.10.0-862.14.0.6.x86_64-08610-gc932cda3cef4-dirty #20
> >>>> [ 89.273169] RIP: 0010:submit_bh_wbc.isra.0+0x538/0x6d0
> >>>> [ 89.277157] RSP: 0018:ffff888105ddfd08 EFLAGS: 00010246
> >>>> [ 89.278093] RAX: 0000000000000005 RBX: ffff888124231498 RCX: ffffffffb2772612
> >>>> [ 89.279332] RDX: 1ffff11024846293 RSI: 0000000000000008 RDI: ffff888124231498
> >>>> [ 89.280591] RBP: ffff8881248cc000 R08: 0000000000000001 R09: ffffed1024846294
> >>>> [ 89.281851] R10: ffff88812423149f R11: ffffed1024846293 R12: 0000000000003800
> >>>> [ 89.283095] R13: 0000000000000001 R14: 0000000000000000 R15: ffff8881161f7000
> >>>> [ 89.284342] FS: 0000000000000000(0000) GS:ffff88839b5c0000(0000) knlGS:0000000000000000
> >>>> [ 89.285711] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>>> [ 89.286701] CR2: 00007f166ebc01a0 CR3: 0000000435c0e000 CR4: 00000000000006e0
> >>>> [ 89.287919] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> >>>> [ 89.289138] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> >>>> [ 89.290368] Call Trace:
> >>>> [ 89.290842] write_mmp_block+0x2ca/0x510
> >>>> [ 89.292218] kmmpd+0x433/0x9a0
> >>>> [ 89.294902] kthread+0x2dd/0x3e0
> >>>> [ 89.296268] ret_from_fork+0x22/0x30
> >>>> [ 89.296906] Modules linked in:
> >>>>
> >>>> We can reproduce this issue as follow:
> >>>> 1. mkfs.ext4 -O mmp /dev/sda -b 1024
> >>>> 2. mount /dev/sda /home/test
> >>>> 3. echo "/dev/sda" > /sys/power/resume
> >>>> 4. wait a moment we will get exception
> >>>>
> >>>> The sequence of issue is as follows:
> >>>> Thread1 Thread2
> >>>> mount /dev/sda /home/test
> >>>> get s_mmp_bh --> has mapped flag
> >>>> start kmmpd thread
> >>>> echo "/dev/sda" > /sys/power/resume
> >>>> resume_store
> >>>> software_resume
> >>>> swsusp_check
> >>>> set_blocksize
> >>>> truncate_inode_pages_range
> >>>> truncate_cleanup_page
> >>>> block_invalidatepage
> >>>> discard_buffer --> clean mapped flag
> >>>> write_mmp_block
> >>>> submit_bh
> >>>> submit_bh_wbc
> >>>> BUG_ON(!buffer_mapped(bh)) --> trigger bug_on
> >>>>
> >>>> To solve this issue, get block device exclusively when do swsusp_check.
> >>> And why exactly is this going to help?
> >> If a block device is already mounted, then do resume, it will discard
> >> buffer
> >> which file system being used. That can cause the file system to crash and
> >> may even cause the system to reset abnormally.
> > I see.
> >
> > Is there a way to indicate in swsusp_check() that the access will be
> > read-only and so it doesn't care whether or not the block device is in
> > use by someone else?
> The root reason is that call 'set_blocksize' in 'swsusp_check' which
> maybe invalidate
> block device's page.
I see.
> Moreover, the file system and swap partition share storage space, which
> will lead to
> data confusion.
> >> If the "/sys/power/resume" sysfs interface is unrestricted, it will provide
> >> criminals with a way to attack the system.
> >>
> >>>> Signed-off-by: Ye Bin <yebin10@huawei.com>
> >>>> ---
> >>>> kernel/power/swap.c | 5 +++--
> >>>> 1 file changed, 3 insertions(+), 2 deletions(-)
> >>>>
> >>>> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
> >>>> index 9ec418955556..26c0bd2a50da 100644
> >>>> --- a/kernel/power/swap.c
> >>>> +++ b/kernel/power/swap.c
> >>>> @@ -1521,9 +1521,10 @@ int swsusp_read(unsigned int *flags_p)
> >>>> int swsusp_check(void)
> >>>> {
> >>>> int error;
> >>>> + void *holder;
> >>>>
> >>>> hib_resume_bdev = blkdev_get_by_dev(swsusp_resume_device,
> >>>> - FMODE_READ, NULL);
> >>>> + FMODE_READ | FMODE_EXCL, &holder);
> >>> So you need to explain to me how this works.
> >> As we call 'blkdev_get_by_path' in 'mount_bdev' pass mode includes
> >> FMODE_EXCL.
> >> So if 'swsusp_check' try to get block device with mode includes
> >> FMODE_EXCL will failed.
> >>
> >> I think if a block device is used as a swap partition, it cannot be
> >> mounted at the
> >> same time. Conversely, if a block device is already mounted, it cannot
> >> be used as
> >> a swap partition.
> > Well, what if a hibernation image is located in a swap file? This is
> > one of the cases that need to be supported.
> I don't know much about this, but I think the process can call
> swsusp_check, the
> hibernation image must be loaded into the block device.
Yes, it is invalid to call swsusp_check() without a hibernation image
in the target swap partition or swap file if that's what you mean.
However, in the case of a swap file, the image may be there in theory
even if the filesystem holding the swap file is mounted.
> I refer to the process of swapon. If the swap partition is a block
> device, it is also
> opened exclusively.
>
> swapon
> claim_swapfile
There are two cases, though, the S_ISBLK() one and the S_ISREG() one,
and the latter is the swap file case I believe.
> static int claim_swapfile(struct swap_info_struct *p, struct inode *inode)
> {
> int error;
>
> if (S_ISBLK(inode->i_mode)) { -->If a block device is
> specified, it is also opened exclusively here.
> p->bdev = blkdev_get_by_dev(inode->i_rdev,
> FMODE_READ | FMODE_WRITE |
> FMODE_EXCL, p);
> if (IS_ERR(p->bdev)) {
> error = PTR_ERR(p->bdev);
> p->bdev = NULL;
> return error;
> }
> p->old_block_size = block_size(p->bdev);
> error = set_blocksize(p->bdev, PAGE_SIZE);
> if (error < 0)
> return error;
> /*
> * Zoned block devices contain zones that have a sequential
> * write only restriction. Hence zoned block devices
> are not
> * suitable for swapping. Disallow them here.
> */
> if (blk_queue_is_zoned(p->bdev->bd_disk->queue))
> return -EINVAL;
> p->flags |= SWP_BLKDEV;
> } else if (S_ISREG(inode->i_mode)) {
> p->bdev = inode->i_sb->s_bdev;
Here, it is not exclusive.
Anyway, I think that the change made by the patch should be fine,
because it mostly affects restore and I'm not expecting anyone to
restore the system from an image in a swap file while the filesystem
holding it is mounted.
So I'm going to apply it with some changelog edits.
Thanks!
> }
>
> return 0;
> }
> >>>> if (!IS_ERR(hib_resume_bdev)) {
> >>>> set_blocksize(hib_resume_bdev, PAGE_SIZE);
> >>>> clear_page(swsusp_header);
> >>>> @@ -1545,7 +1546,7 @@ int swsusp_check(void)
> >>>>
> >>>> put:
> >>>> if (error)
> >>>> - blkdev_put(hib_resume_bdev, FMODE_READ);
> >>>> + blkdev_put(hib_resume_bdev, FMODE_READ | FMODE_EXCL);
> >>>> else
> >>>> pr_debug("Image signature found, resuming\n");
> >>>> } else {
> >>>> --
> >>> .
> >>>
> > .
> >
>
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2021-10-21 17:17 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-13 12:19 [PATCH -next] PM: hibernate: Get block device exclusively when do swsusp_check Ye Bin
2021-10-20 3:12 ` yebin
2021-10-20 14:26 ` yebin
2021-10-21 11:00 ` Rafael J. Wysocki
2021-10-21 11:38 ` yebin
2021-10-21 13:27 ` Rafael J. Wysocki
2021-10-21 14:38 ` yebin
2021-10-21 17:17 ` Rafael J. Wysocki
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).