[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <a3f6c88f-5b0f-0da0-c5ad-dd40ea36f4a5@linuxfoundation.org>
Date: Fri, 2 Oct 2020 17:50:09 -0600
From: Shuah Khan <skhan@...uxfoundation.org>
To: Jens Axboe <axboe@...nel.dk>, idryomov@...il.com,
dongsheng.yang@...ystack.cn
Cc: ceph-devel@...r.kernel.org,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Kees Cook <keescook@...omium.org>,
Shuah Khan <skhan@...uxfoundation.org>
Subject: Re: drivers/block/rbd.c: atomic_inc_return_safe() &
atomic_dec_return_safe()
On 10/2/20 5:44 PM, Jens Axboe wrote:
> On 10/2/20 4:34 PM, Shuah Khan wrote:
>> All,
>>
>> I came across these atomic_inc_return_safe() & atomic_dec_return_safe()
>> functions that hold the counters at safe values.
>>
>> atomic_inc_return_safe()
>>
>> If the counter is already 0 it will not be incremented.
>> If the counter is already at its maximum value returns
>> -EINVAL without updating it.
>>
>> atomic_dec_return_safe()
>>
>> Decrement the counter. Return the resulting value, or -EINVAL
>>
>> These two routines are static and only used in rbd.c.
>>
>> Can these become part of atomic_t ops?
>
> I think you just want to use refcount_t for this use case. They
> have safe guards for under/overflow.
>
Makes sense. Guess these came before refcount_t.
I will track this for refcount_t conversion.
thanks,
-- Shuah
Powered by blists - more mailing lists