Date: Sat, 23 Sep 2017 18:18:01 +0200 From: Andreas Longwitz <longwitz@incore.de> To: Don Lewis <truckman@FreeBSD.org> Cc: kostikbel@gmail.com, mckusick@mckusick.com, freebsd-fs@freebsd.org Subject: Re: fsync: giving up on dirty on ufs partitions running vfs_write_suspend() Message-ID: <59C68939.2040509@incore.de> In-Reply-To: <201709222220.v8MMKEW5085371@gw.catspoiler.org> References: <201709222220.v8MMKEW5085371@gw.catspoiler.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Don Lewis wrote:
>> Patch against HEAD:
>> --- vfs_default.c.orig 2017-09-22 11:56:26.950084000 +0200
>> +++ vfs_default.c 2017-09-22 11:58:33.211196000 +0200
>> @@ -690,6 +690,8 @@
>> bremfree(bp);
>> bawrite(bp);
>> }
>> + if( maxretry < 1000)
>> + DELAY(1000); /* 1 ms */
>> BO_LOCK(bo);
>> goto loop2;
>> }
>
> Do you need to use a busy loop here, or can you yield the cpu by using
> something like pause(9)?
>
No, I don't need a busy loop, it is even bad if we have only one CPU. I
have assumed that the person doing eventually a commit for this problem
will much better know than I how to wait 1 ms in the kernel.
I have repested my tests with "pause("dirty", hz/1000)" instead of
"DELAY(1000)" and got the same results in maximal loop numbers and spent
times as before.
--
Andreas Longwitz
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?59C68939.2040509>