The function kzalloc() in detached_dev_do_request() can fail, so its
return value should be checked.

Fixes: bc082a55d25c (bcache: fix inaccurate io state for detached bcache devices)
Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
---
 drivers/md/bcache/request.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index d15aae6c51c1..1b5ccfa93b8c 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -1107,6 +1107,8 @@ static void detached_dev_do_request(struct bcache_device *d, struct bio *bio,
 	 * which would call closure_get(&dc->disk.cl)
 	 */
 	ddip = kzalloc(sizeof(struct detached_dev_io_private), GFP_NOIO);
+	if (!ddip)
+		return;
 	ddip->d = d;
 	/* Count on the bcache device */
 	ddip->orig_bdev = orig_bdev;
-- 
2.17.1

On 2/25/22 4:20 PM, Jia-Ju Bai wrote:
> The function kzalloc() in detached_dev_do_request() can fail, so its
> return value should be checked.
>
> Fixes: bc082a55d25c (bcache: fix inaccurate io state for detached bcache devices)
> Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
> Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
> ---
>   drivers/md/bcache/request.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
> index d15aae6c51c1..1b5ccfa93b8c 100644
> --- a/drivers/md/bcache/request.c
> +++ b/drivers/md/bcache/request.c
> @@ -1107,6 +1107,8 @@ static void detached_dev_do_request(struct bcache_device *d, struct bio *bio,
>   	 * which would call closure_get(&dc->disk.cl)
>   	 */
>   	ddip = kzalloc(sizeof(struct detached_dev_io_private), GFP_NOIO);

Yes this is a potential problem, may result NULL pointer deference.


> +	if (!ddip)
> +		return;

The fix here is incorrect. A proper way to fix might be something like this,

     if (!ddip) {

         bio->bi_status = BLK_STS_RESOURCE;

         bio->bi_end_io(bio);

         return;

     }


The orignal patch will make I/O task be in sleep forever.

>   	ddip->d = d;
>   	/* Count on the bcache device */
>   	ddip->orig_bdev = orig_bdev;

On 2022/3/2 18:11, Coly Li wrote:
> On 2/25/22 4:20 PM, Jia-Ju Bai wrote:
>> The function kzalloc() in detached_dev_do_request() can fail, so its
>> return value should be checked.
>>
>> Fixes: bc082a55d25c (bcache: fix inaccurate io state for detached 
>> bcache devices)
>> Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
>> Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
>> ---
>>   drivers/md/bcache/request.c | 2 ++
>>   1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
>> index d15aae6c51c1..1b5ccfa93b8c 100644
>> --- a/drivers/md/bcache/request.c
>> +++ b/drivers/md/bcache/request.c
>> @@ -1107,6 +1107,8 @@ static void detached_dev_do_request(struct 
>> bcache_device *d, struct bio *bio,
>>        * which would call closure_get(&dc->disk.cl)
>>        */
>>       ddip = kzalloc(sizeof(struct detached_dev_io_private), GFP_NOIO);
>
> Yes this is a potential problem, may result NULL pointer deference.
>
>
>> +    if (!ddip)
>> +        return;
>
> The fix here is incorrect. A proper way to fix might be something like 
> this,
>
>     if (!ddip) {
>
>         bio->bi_status = BLK_STS_RESOURCE;
>
>         bio->bi_end_io(bio);
>
>         return;
>
>     }
>
>
> The orignal patch will make I/O task be in sleep forever.
>
>>       ddip->d = d;
>>       /* Count on the bcache device */
>>       ddip->orig_bdev = orig_bdev;
>

Hi Coly,

Thanks for the advice!
I will send a V2 patch.


Best wishes,
Jia-Ju Bai