Date: Tue, 16 Oct 2001 10:10:05 -0700 (PDT) From: Matthew Dillon <dillon@apollo.backplane.com> To: Maxime Henrion <mux@qualys.com> Cc: cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG Subject: Re: cvs commit: src/sys/vm vnode_pager.c Message-ID: <200110161710.f9GHA5933464@apollo.backplane.com> References: <200110121817.f9CIHYU38714@freefall.freebsd.org> <xzpelo8ii94.fsf@flood.ping.uio.no> <200110122330.f9CNU1u36917@earth.backplane.com> <xzpelo4r90f.fsf@flood.ping.uio.no> <200110152135.f9FLZpg56816@earth.backplane.com> <20011016172843.A469@nebula.cybercable.fr> <200110161618.f9GGIpM31430@apollo.backplane.com> <20011016190340.A465@nebula.cybercable.fr>
next in thread | previous in thread | raw e-mail | index | archive | help
:
:Matthew Dillon wrote:
:>
:> It sure looks like qsync() is the responsible party in this
:> profile, and looking at the code I can see why. It's an
:> insanely expensive procedure. It's vget()ing and vput()ing every
:> single vnode in the vnode cache and doing other nasty things.
:
:
:I tried to mount the partitions without quotas, and I unfortunately
:experience exactly the same problems.
:
:Maxime
:--
:Don't be fooled by cheap finnish imitations ; BSD is the One True Code
Do a mount -v to make sure you mounted without quotas. If you
did mount without quotas and have the same problem, do another
profile.
I think the only solution to both problems is to propogate the
various modified flags up to the vnode v_flag level so the
mount scan can check them there and loop-up without having to
get the vnode lock or the vnode interlock.
-Matt
Matthew Dillon
<dillon@backplane.com>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200110161710.f9GHA5933464>