Re: [PATCH V5 8/9] null_blk: serialize configfs attribute stores with device setup

在 2026/7/25 5:13, Bart Van Assche 写道:
> On 7/13/26 9:18 PM, Zizhi Wo wrote:
>>       ret = nullb_device_##TYPE##_attr_store(&new_value, page, count);\
>>       if (ret < 0)                            \
>>           return ret;                        \
>> +    ret = 0;                            \
>> +    mutex_lock(&lock);                        \
>>       if (apply_fn)                            \
>>           ret = apply_fn(dev, new_value);                \
>>       else if (test_bit(NULLB_DEV_FL_CONFIGURED, &dev->flags))     \
>>           ret = -EBUSY;                        \
>> +    if (!ret)                            \
>> +        dev->NAME = new_value;                    \
>> +    mutex_unlock(&lock);                        \
>>       if (ret < 0)                            \
>>           return ret;                        \
>> -    dev->NAME = new_value;                        \
>>       return count;                            \
>>   }                                    \
> 
> If guard()() would be used instead of explicit mutex_lock() and
> mutex_unlock() calls, the if (ret) test wouldn't have to be duplicated,
> isn't it?

Yes.

> 
> Additionally, the above changes modify the behavior if apply_fn()
> returns a strictly positive value (> 0).
> 
> Bart.

Yes, I made this change because the current apply_fn() never return a
value greater than 0. Thanks for the suggestion. I'll switch to using
guard() in the next version.

Thanks,
Zizhi Wo