linux-kernel - Re: [PATCH 7/8] wbt: add general throttling mechanism

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

Message-ID: <5720D90F.6000609@fb.com>
Date:	Wed, 27 Apr 2016 09:21:51 -0600
From:	Jens Axboe <axboe@...com>
To:	xiakaixu <xiakaixu@...wei.com>
CC:	<linux-kernel@...r.kernel.org>, <linux-fsdevel@...r.kernel.org>,
	<linux-block@...r.kernel.org>, <jack@...e.cz>,
	<dchinner@...hat.com>, <sedat.dilek@...il.com>,
	"miaoxie (A)" <miaoxie@...wei.com>, Huxinwei <huxinwei@...wei.com>,
	Bintian <bintian.wang@...wei.com>
Subject: Re: [PATCH 7/8] wbt: add general throttling mechanism

On 04/27/2016 06:06 AM, xiakaixu wrote:
>> +void __wbt_done(struct rq_wb *rwb)
>> +{
>> +	int inflight, limit = rwb->wb_normal;
>> +
>> +	/*
>> +	 * If the device does write back caching, drop further down
>> +	 * before we wake people up.
>> +	 */
>> +	if (rwb->wc && !atomic_read(&rwb->bdi->wb.dirty_sleeping))
>> +		limit = 0;
>> +	else
>> +		limit = rwb->wb_normal;
>> +
>> +	/*
>> +	 * Don't wake anyone up if we are above the normal limit. If
>> +	 * throttling got disabled (limit == 0) with waiters, ensure
>> +	 * that we wake them up.
>> +	 */
>> +	inflight = atomic_dec_return(&rwb->inflight);
>> +	if (limit && inflight >= limit) {
>> +		if (!rwb->wb_max)
>> +			wake_up_all(&rwb->wait);
>> +		return;
>> +	}
>> +
> Hi Jens,
>
> Just a little confused about this. The rwb->wb_max can't be 0 if the variable
> 'limit' does not equal to 0. So the if (!rwb->wb_max) branch maybe does not
> make sense.

You are right, it doesn't make a lot of sense. I think it suffers from 
code shuffling. How about the attached? The important part is that we 
wake up waiters, if wbt got disabled while we had tracked IO in flight.

-- 
Jens Axboe


View attachment "wbt-disable-wakeup.patch" of type "text/x-patch" (913 bytes)