[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6bb64a49-a8db-3431-537f-913982886ba9@huaweicloud.com>
Date: Tue, 29 Apr 2025 09:43:12 +0800
From: Yu Kuai <yukuai1@...weicloud.com>
To: John Garry <john.g.garry@...cle.com>, Yu Kuai <yukuai1@...weicloud.com>,
hch@...radead.org, axboe@...nel.dk, xni@...hat.com, agk@...hat.com,
snitzer@...nel.org, mpatocka@...hat.com, song@...nel.org, cl@...ux.com,
nadav.amit@...il.com, ubizjak@...il.com, akpm@...ux-foundation.org
Cc: linux-block@...r.kernel.org, linux-kernel@...r.kernel.org,
dm-devel@...ts.linux.dev, linux-raid@...r.kernel.org, yi.zhang@...wei.com,
yangerkun@...wei.com, johnny.chenyi@...wei.com,
"yukuai (C)" <yukuai3@...wei.com>
Subject: Re: [PATCH v2 3/9] block: WARN if bdev inflight counter is negative
Hi,
在 2025/04/28 22:06, John Garry 写道:
> On 27/04/2025 09:29, Yu Kuai wrote:
>> From: Yu Kuai <yukuai3@...wei.com>
>>
>> Which means there is a BUG
>
> nit: to me, BUG means symbol BUG(), and not a software bug (which I
> think that you mean)
Actually, I mean the bio-based disk driver or blk-mq messed up the IO
accounting, IO done is more than IO start, and this is a bug.
>
>> for related bio-based disk driver, or blk-mq
>> for rq-based disk, it's better not to hide the BUG.
>
> AFICS, this check was not present for mq, so is it really required now?
> I suppose that the code is simpler to always have the check. I find it
> an odd check to begin with...
This check do present for mq, for example, diskstats_show() and
update_io_ticks().
Thanks,
Kuai
>
>>
>> Signed-off-by: Yu Kuai <yukuai3@...wei.com>
>> ---
>> block/genhd.c | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/block/genhd.c b/block/genhd.c
>> index f671d9ee00c4..d158c25237b6 100644
>> --- a/block/genhd.c
>> +++ b/block/genhd.c
>> @@ -136,9 +136,9 @@ static void part_in_flight_rw(struct block_device
>> *part,
>> inflight[0] += part_stat_local_read_cpu(part, in_flight[0],
>> cpu);
>> inflight[1] += part_stat_local_read_cpu(part, in_flight[1],
>> cpu);
>> }
>> - if ((int)inflight[0] < 0)
>> + if (WARN_ON_ONCE((int)inflight[0] < 0))
>> inflight[0] = 0;
>> - if ((int)inflight[1] < 0)
>> + if (WARN_ON_ONCE((int)inflight[1] < 0))
>> inflight[1] = 0;
>> }
>
>
> .
>
Powered by blists - more mailing lists