From: quale@saavik.cs.wisc.edu (Douglas E. Quale) Subject: Re: [comp.os.linux]: Re: File system issues! Date: Tue, 14 Jul 1992 05:06:22 GMT
In article <cy7uNB1w165w@ssg.com> unixsys@ssg.com (Rick Emerson) writes:
>news-daemon@optical.bms.com writes:
>
>'Splain me why a "fscache -writeback 0" (or something equivalent) is so
>distasteful? Please, let's not debate the pros and cons of write-back
>caching. In most cases it's the best thing since Christy Brinkley... well,
>almost. What I do not understand is why is a writeback disable switch is so
>awful that people fight it as if it were the plague come calling.
>
If you folks really, really, really have to have this, please make it a
mount parameter called synchronous (a la noexec, nodev, nosuid). This
is what is done in BSD. The mount(2) call then needs to be enhanced
to permit it update the mount flags of already mounted file systems (a
nice thing to have in general, so you can mount ro and later go to rw
without having to umount then mount again).
Also, although this has been ingrained in this thread since way back,
"writeback" is not the applicable term here. The type of caching done
by Unix is called "delayed write". "Write back" refers to memory caching,
the alternative flavor being "write through". Many Linuxer's have one
of these types of caching on their motherboards, but that is a hardware
issue.
-- Doug Quale quale@saavik.cs.wisc.edu