linux-kernel - Re: [PATCH] sbitmap: fix permanent io blocking caused by insufficient wakeup times

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Sat, 17 Sep 2022 23:06:41 +0800
From:   Liu Song <liusong@...ux.alibaba.com>
To:     Yu Kuai <yukuai1@...weicloud.com>, axboe@...nel.dk
Cc:     linux-kernel@...r.kernel.org, linux-block@...r.kernel.org,
        "yukuai (C)" <yukuai3@...wei.com>
Subject: Re: [PATCH] sbitmap: fix permanent io blocking caused by insufficient
 wakeup times


在 2022/9/17 18:20, Yu Kuai 写道:
> Hi,
>
> 在 2022/09/17 10:33, Liu Song 写道:
>> From: Liu Song <liusong@...ux.alibaba.com>
>>
>> In "sbitmap_queue_clear_batch", a batch of completed requests may be
>> processed at once, but "wait_cnt" is only reduced once by
>> "sbitmap_queue_wake_up".
>> In our environment, if "/sys/block/nvme0n1/nr_requests" is adjusted to
>> 32, it is easily because no tag and then enter the waiting situation,
>> if continue change "nr_requests" to 1023 at this time, it will basically
>> fall into the situation of permanent blocking. Because there will be
>> "blk_freeze_queue" in "blk_mq_update_nr_requests", which will prevent
>> any new requests, but due to insufficient wake-up, there are tasks
>> waiting for wake-up, but no new wake-up opportunities will be generated
>> at this time, so this situation needs to be repaired.
>>
>
> It seems Keith already fixed this recently. Can you have a check at his
> patch?
>
Hi

It's a bit embarrassing. My code is 6.0-rc3. Yesterday, there was a 
problem with
the Tsinghua mirror station. The fetched code could not be the latest, 
so I submitted it.
According to your suggestion, I checked and it has been fixed, thank you 
for your suggestion.

Thanks

> Thanks,
> Kuai
>> Signed-off-by: Liu Song <liusong@...ux.alibaba.com>
>> ---
>>   include/linux/sbitmap.h |  8 ++++++++
>>   lib/sbitmap.c           | 22 ++++++++++++++++++----
>>   2 files changed, 26 insertions(+), 4 deletions(-)
>>
>> diff --git a/include/linux/sbitmap.h b/include/linux/sbitmap.h
>> index 8f5a86e..153382e 100644
>> --- a/include/linux/sbitmap.h
>> +++ b/include/linux/sbitmap.h
>> @@ -579,6 +579,14 @@ static inline struct sbq_wait_state 
>> *sbq_wait_ptr(struct sbitmap_queue *sbq,
>>   void sbitmap_queue_wake_up(struct sbitmap_queue *sbq);
>>     /**
>> + * sbitmap_queue_wake_up_batch() - Attempt to wake up waiters in 
>> batches
>> + * on a &struct sbitmap_queue.
>> + * @sbq: Bitmap queue to wake up.
>> + * @nr: The number of attempts to wake the waiter.
>> + */
>> +void sbitmap_queue_wake_up_batch(struct sbitmap_queue *sbq, int nr);
>> +
>> +/**
>>    * sbitmap_queue_show() - Dump &struct sbitmap_queue information to 
>> a &struct
>>    * seq_file.
>>    * @sbq: Bitmap queue to show.
>> diff --git a/lib/sbitmap.c b/lib/sbitmap.c
>> index 29eb048..f2aa1da 100644
>> --- a/lib/sbitmap.c
>> +++ b/lib/sbitmap.c
>> @@ -600,7 +600,7 @@ static struct sbq_wait_state *sbq_wake_ptr(struct 
>> sbitmap_queue *sbq)
>>       return NULL;
>>   }
>>   -static bool __sbq_wake_up(struct sbitmap_queue *sbq)
>> +static bool __sbq_wake_up(struct sbitmap_queue *sbq, int *nr)
>>   {
>>       struct sbq_wait_state *ws;
>>       unsigned int wake_batch;
>> @@ -610,6 +610,7 @@ static bool __sbq_wake_up(struct sbitmap_queue *sbq)
>>       if (!ws)
>>           return false;
>>   +again:
>>       wait_cnt = atomic_dec_return(&ws->wait_cnt);
>>       if (wait_cnt <= 0) {
>>           int ret;
>> @@ -632,10 +633,14 @@ static bool __sbq_wake_up(struct sbitmap_queue 
>> *sbq)
>>           if (ret == wait_cnt) {
>>               sbq_index_atomic_inc(&sbq->wake_index);
>>               wake_up_nr(&ws->wait, wake_batch);
>> -            return false;
>> +            if (!nr || *nr <= 0)
>> +                return false;
>>           }
>>             return true;
>> +    } else if (nr && *nr) {
>> +        (*nr)--;
>> +        goto again;
>>       }
>>         return false;
>> @@ -643,11 +648,20 @@ static bool __sbq_wake_up(struct sbitmap_queue 
>> *sbq)
>>     void sbitmap_queue_wake_up(struct sbitmap_queue *sbq)
>>   {
>> -    while (__sbq_wake_up(sbq))
>> +    while (__sbq_wake_up(sbq, NULL))
>>           ;
>>   }
>>   EXPORT_SYMBOL_GPL(sbitmap_queue_wake_up);
>>   +void sbitmap_queue_wake_up_batch(struct sbitmap_queue *sbq, int nr)
>> +{
>> +    int i = SBQ_WAIT_QUEUES;
>> +
>> +    while (__sbq_wake_up(sbq, &nr) && --i)
>> +        ;
>> +}
>> +EXPORT_SYMBOL_GPL(sbitmap_queue_wake_up_batch);
>> +
>>   static inline void sbitmap_update_cpu_hint(struct sbitmap *sb, int 
>> cpu, int tag)
>>   {
>>       if (likely(!sb->round_robin && tag < sb->depth))
>> @@ -683,7 +697,7 @@ void sbitmap_queue_clear_batch(struct 
>> sbitmap_queue *sbq, int offset,
>>           atomic_long_andnot(mask, (atomic_long_t *) addr);
>>         smp_mb__after_atomic();
>> -    sbitmap_queue_wake_up(sbq);
>> +    sbitmap_queue_wake_up_batch(sbq, nr_tags);
>>       sbitmap_update_cpu_hint(&sbq->sb, raw_smp_processor_id(),
>>                       tags[nr_tags - 1] - offset);
>>   }
>>