Date: Wed, 22 Mar 2000 16:18:27 -0800 (PST) From: Matthew Dillon <dillon@apollo.backplane.com> To: Paul Saab <paul@mu.org> Cc: current@FreeBSD.ORG Subject: Re: Buf cache, swap spl, and NFS patches need review. Message-ID: <200003230018.QAA94491@apollo.backplane.com> References: <200003221713.JAA91710@apollo.backplane.com> <20000322161606.A19237@elvis.mu.org>
next in thread | previous in thread | raw e-mail | index | archive | help
:Content-Type: text/plain; charset=us-ascii : :Matthew Dillon (dillon@apollo.backplane.com) wrote: :> swap-SPL: :> :> Required splvm()'s had to be added to two places in the swap subsystem :> (but I don't think the lack of these spl's is the cause of the :> recently reported crashes, though it is possible). : :There are a few more that I still get if I dont add spl protection. :I need to commit the SPLASSERT stuff but I'm scared about how badly it :will break for people that run with INVARIANTS. : :I'll attach the whole SPLASSERT to the VM here. : :Also, I went through vfs_bio.c and added SPLASSERT there and it :goes kaboom really quick. bdwrite calls everything with no spl :proctection when it really should. My full SPLASSERT patch is at :http://people.freebsd.org/~ps/splassert.diff. I should just go ahead and :submit this. This patch doesn't have any of the vm patches we discussed :last month. Lets do this one step at a time. Let me get in my (relatively minor) spl fixes, then I'll take a look at your vfs_bio.c stuff -- there's got to be some cockpit trouble there because for the most part vfs_bio.c does *NOT* need to run at splbio(). Only the buffer queue adds and drops and hash table ops need to run at splbio(). Once you have a buffer in hand and locked you own it. -Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200003230018.QAA94491>