Date: Wed, 5 Mar 1997 15:45:04 +1100 From: Bruce Evans <bde@zeta.org.au> To: dg@root.com, freebsd-bugs@freefall.freebsd.org Subject: Re: kern/2875: vinvalbuf() clobbers b_vnbufs chain Message-ID: <199703050445.PAA07046@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
> >> Hmmm, the fix seems overly complicated. How about this instead: > > > >Your fix inhibit disk I/O interruption for more long time. If this > >is not a problem, I prefer your fix. > > I don't think it's a problem. The actual delay is fairly low in practice > (probably less than 50us), and letting an interrupt through prior to the 50us is fairly large on a 386, and very large on a P5. Lite2 locks only the B_BUSY test and set here. The whole loop is not locked in any case because of the explicit sleep in the loop and the implicit sleep in VOP_BWRITE(). Sort-of-locking of the whole loop together with the bug was introduced without special comment in rev.1.59. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199703050445.PAA07046>