On 2017/9/5 下午2:43, Christoph Hellwig wrote:
> On Tue, Sep 05, 2017 at 01:30:04AM +0800, Coly Li wrote:
>>
>> When you mentioned "whole chunk of code", do you mean the following
>> block of code ?
>>
>>
>> 1960 if (IS_ERR(bdev)) {
>> ========= start of whole chunk of code ============
>> 1961 if (bdev == ERR_PTR(-EBUSY)) {
>> 1962 bdev = lookup_bdev(strim(path));
>> 1963 mutex_lock(&bch_register_lock);
>> 1964 if (!IS_ERR(bdev) && bch_is_open(bdev))
>> 1965 err = "device already registered";
>> 1966 else
>> 1967 err = "device busy";
>> 1968 mutex_unlock(&bch_register_lock);
>> 1969 if (!IS_ERR(bdev))
>> 1970 bdput(bdev);
>> 1971 if (attr == &ksysfs_register_quiet)
>> 1972 goto out;
>> 1973 }
>> ========= end of whole chunk of code ============
>> 1974 goto err;
>> 1975 }
>>
>> I don't mind to remove it, just double check I don't misunderstand what
>> you meant.
>
> Yes, that's the problematic block.
>
Hi Christoph,
I tested the code, and my patch which removes the above code. The result
is, I feel the above chunk of code is useful. When I tried to register a
device and it was already registered. Without the above code, I only saw
"failed to open device", didn't realize it was because this device is
registered already. After adding the above code back, I knew where the
problem was.
The above chunk of code improves user experience and provides more
detailed diagnose information, it is useful. Then I suggest to keep the
code here and pick up Jan's patch.
Thanks.
--
Coly Li
--
To unsubscribe from this list: send the line "unsubscribe linux-bcache" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
