Date: Tue, 16 Feb 1999 09:21:36 -0500 (EST) From: "John S. Dyson" <dyson@iquest.net> To: syssgm@detir.qld.gov.au (Stephen McKay) Cc: freebsd-hackers@freebsd.org, syssgm@detir.qld.gov.au, dyson@iquest.net, dillon@apollo.backplane.com, julian@whistle.com Subject: Re: inode / exec_map interlock ? (follow up) Message-ID: <199902161421.JAA01397@y.dyson.net> In-Reply-To: <199902161213.WAA28362@nymph.detir.qld.gov.au> from Stephen McKay at "Feb 16, 99 10:13:26 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Stephen McKay said:
>
> This is a danger, but with the code becoming easier to understand, I expect
> that changes that are performance damaging will be repaired. Some small
> errors have already been reverted. Others will need to be first recognised,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> then deliberately repaired, as below:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> In a separate message on Monday, 15th February 1999, "John S. Dyson" wrote:
>
> >Are you blocking on excessively large numbers of output requests? You
> >know *exactly* the issue at hand, and it has to do with the backpressure
> >needed to keep the pageout daemon from doing an evil nasty on all of the
> >pages in the system.
>
> The pagedaemon on a test machine of mine used to spend much time waiting on
> "swpfre". Now, under 4.0, the paging rate has shot up (about 2x as a guess)
> and it is much less responsive. Of course it has only 16Mb of ram, and I
> thrash it. But I favour John's view that the new swap pager has a deficiency
> that must be rectified before it can be considered better (in all cases) than
> the previous version.
>
The very sad thing is that I told him about it a long time ago (before the
commit.) He hasn't acknowleged it. I knew it before even testing the code,
because I even understand what HE is doing, because I read code. The old
code did block, and he noted that and somehow believed in his infinite wisdom
that it shouldn't.
The old code wouldn't have blocked if it would have worked better. This is
one example, but the vast number of fast changes will increase entropy
significantly.
Maybe the code is easier to read, because some of the functionality and
quality is being stripped out, trading simplicity?
--
John | Never try to teach a pig to sing,
dyson@iquest.net | it makes one look stupid
jdyson@nc.com | and it irritates the pig.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199902161421.JAA01397>