Date: Tue, 16 Feb 1999 09:33:30 -0500 (EST) From: "John S. Dyson" <dyson@iquest.net> To: dillon@apollo.backplane.com (Matthew Dillon) Cc: dyson@iquest.net, freebsd-hackers@FreeBSD.ORG Subject: Re: inode / exec_map interlock ? (follow up) Message-ID: <199902161433.JAA01411@y.dyson.net> In-Reply-To: <199902160748.XAA21828@apollo.backplane.com> from Matthew Dillon at "Feb 15, 99 11:48:09 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Matthew Dillon said: > :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 original version of your code appeared to have the problem, and if > :you have fixed it, or you aren't queueing more than a few pageouts at > :a time (perhaps <10 or so), then you are okay. Note that "pageouts" > :often imply the need for "pageins", and often the same disk for "pageouts" > > Excuse me, John, you aren't listening. > > The ORIGINAL VM CODE. Do I need to repeat that? The *ORIGINAL* VM CODE > does not have one single line of source to prevent excessive queueing > of I/O for pageout ops. > You are wrong. Please look at the code. I will point the code out to you if you want, but I suspect that you don't want to know. > > You keep on accusing me of doing something wrong. I keep on telling you > that I haven't done what you seem to think I have done. If you are going > to continue to accuse me, then point to the code that you feel is broken. > You are wrong. It is entertaining to me to see you make such strong responses to simple, truthful assertions, and proof is simple. That makes me suspect that you really didn't understand the code, and is why I am quite worried about the changes that you have been making. I truely believed that you understood how the original swap pager worked, and now, I suspect that you only parroted the design, missing some important points. If there wasn't a tsleep or two in the code in strategic areas, I could have possibly understood your position. :-). The data structure improvements that you made in the swap pager were valuable, however you also stripped out important functionality. Already, there is anecdotal information regarding performance regressions. -- 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?199902161433.JAA01411>