Date: Fri, 27 Nov 1998 23:41:09 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: dillon@apollo.backplane.com (Matthew Dillon) Cc: tlambert@primenet.com, joelh@gnu.org, DBECK@ludens.elte.hu, freebsd-hackers@FreeBSD.ORG Subject: Re: SYSV Semaphores & mmap problems Message-ID: <199811272341.QAA24991@usr02.primenet.com> In-Reply-To: <199811240147.RAA11543@apollo.backplane.com> from "Matthew Dillon" at Nov 23, 98 05:47:24 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> The difference between mmap() and SysV shared memory isn't so bad. The > mmap()'d file will basically get synced every 30 seconds or so by the > syncer daemon. madvise() can be used to pre-fault any pages already in > the cache. > > I think all that is needed is a way to flag a file so the syncer doesn't > touch it under normal circumstances, instead allowing its pages to be > synced by normal paging activity. You can then madvise(... MADV_FREE) > the pages after you are through with the shared memory segment to throw > them away. The question to ask is "Why did Oracle pay John Dyson to make FreeBSD's SYSV SHM faster, and reject mmap(), even though the Oriacle source code supports using mmap()?". Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. 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?199811272341.QAA24991>