A user powers off the device by writing 0 to `./config/nullb/nullb1/power`
via the nullb configfs interface, and later powers it on by writing 1
to the same file.

When the power-off command is executed, the zones created by the previous
task are not freed before new zones are created upon powering on again;
this results in the memory leak described in [1].

Add an operation to free zones within nullb_device_power_store().

[1]
BUG: memory leak
unreferenced object 0xffff888126990000 (size 56000):
  backtrace (crc 5c9573f8):
    null_init_zoned_dev+0x108/0x520 drivers/block/null_blk/zoned.c:94
    null_add_dev+0x8a6/0xae0 drivers/block/null_blk/main.c:1993
    nullb_device_power_store+0xf9/0x1f0 drivers/block/null_blk/main.c:501
    flush_write_buffer fs/configfs/file.c:207 [inline]
    configfs_write_iter+0x10d/0x1a0 fs/configfs/file.c:229
    new_sync_write fs/read_write.c:595 [inline]
    vfs_write+0x3ac/0x720 fs/read_write.c:687

Fixes: cedcafad8277 ("nullb: add interface to power on disk")
Reported-by: syzbot+f57d67a92532a98a562d@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=f57d67a92532a98a562d
Tested-by: syzbot+f57d67a92532a98a562d@syzkaller.appspotmail.com
Signed-off-by: Edward Adam Davis <eadavis@qq.com>
---
 drivers/block/null_blk/main.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/block/null_blk/main.c b/drivers/block/null_blk/main.c
index f8c0fd57e041..fb001d002d17 100644
--- a/drivers/block/null_blk/main.c
+++ b/drivers/block/null_blk/main.c
@@ -511,6 +511,7 @@ static ssize_t nullb_device_power_store(struct config_item *item,
 		if (test_and_clear_bit(NULLB_DEV_FL_UP, &dev->flags)) {
 			dev->power = newp;
 			null_del_dev(dev->nullb);
+			null_free_zoned_dev(dev);
 		}
 		clear_bit(NULLB_DEV_FL_CONFIGURED, &dev->flags);
 	}
-- 
2.43.0

Hi, Edward

在 2026/8/14 16:21, Edward Adam Davis 写道:
> A user powers off the device by writing 0 to `./config/nullb/nullb1/power`
> via the nullb configfs interface, and later powers it on by writing 1
> to the same file.
> 
> When the power-off command is executed, the zones created by the previous
> task are not freed before new zones are created upon powering on again;
> this results in the memory leak described in [1].
> 
> Add an operation to free zones within nullb_device_power_store().
> 
> [1]
> BUG: memory leak
> unreferenced object 0xffff888126990000 (size 56000):
>    backtrace (crc 5c9573f8):
>      null_init_zoned_dev+0x108/0x520 drivers/block/null_blk/zoned.c:94
>      null_add_dev+0x8a6/0xae0 drivers/block/null_blk/main.c:1993
>      nullb_device_power_store+0xf9/0x1f0 drivers/block/null_blk/main.c:501
>      flush_write_buffer fs/configfs/file.c:207 [inline]
>      configfs_write_iter+0x10d/0x1a0 fs/configfs/file.c:229
>      new_sync_write fs/read_write.c:595 [inline]
>      vfs_write+0x3ac/0x720 fs/read_write.c:687
> 
> Fixes: cedcafad8277 ("nullb: add interface to power on disk")
> Reported-by: syzbot+f57d67a92532a98a562d@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=f57d67a92532a98a562d
> Tested-by: syzbot+f57d67a92532a98a562d@syzkaller.appspotmail.com
> Signed-off-by: Edward Adam Davis <eadavis@qq.com>
> ---
>   drivers/block/null_blk/main.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/block/null_blk/main.c b/drivers/block/null_blk/main.c
> index f8c0fd57e041..fb001d002d17 100644
> --- a/drivers/block/null_blk/main.c
> +++ b/drivers/block/null_blk/main.c
> @@ -511,6 +511,7 @@ static ssize_t nullb_device_power_store(struct config_item *item,
>   		if (test_and_clear_bit(NULLB_DEV_FL_UP, &dev->flags)) {
>   			dev->power = newp;
>   			null_del_dev(dev->nullb);
> +			null_free_zoned_dev(dev);
>   		}
>   		clear_bit(NULLB_DEV_FL_CONFIGURED, &dev->flags);
>   	}

Thanks for the patch. This issue has already been addressed in my
null_blk series posted back in July:

https://lore.kernel.org/all/20260725022509.714271-6-
wozizhi@huaweicloud.com/

The series has already collected some Reviewed-by tags but hasn't been
merged yet. After it, null_free_zoned_dev() no longer needs to be called
from null_free_dev(), so this fix is handled differently there.

Thanks,
Zizhi Wo