Date: Mon, 10 May 1999 17:22:50 -0700 (PDT) From: Cliff Skolnick <cliff@steam.com> To: freebsd-stable@FreeBSD.ORG Cc: luoqi@FreeBSD.ORG Subject: Re: vm_fault deadlock and PR 8416 (was Re: Strange reboot saga ) Message-ID: <Pine.BSF.4.10.9905101717560.453-100000@lazlo.internal.steam.com> In-Reply-To: <199905102111.OAA50817@passer.osg.gov.bc.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
After removing the patch from my system, poof - deadlock. I did the ctl-printscrean thing, and everything looked ok. There were a bunch of locks, shared, but I could not make much sense beyond the "ps" and doing a trace which did not give me much info since I could not figure out how to trace a specific process. :( Anyways this is a production email and dns system, so I can't futz with it too much. I've added the patch in and expect my system will run again until *I* take it down. Usually it locks up within 24 hours without the patch, and it has run 4+ days with no lockup with the patch. After 4 days I took the system down to remote the patch and add debugging to my kernel, monday rolls around and poof - locked up. I would urge someone who has more of a clue about the locking issues to take a look at this. Cliff On Mon, 10 May 1999, Cy Schubert wrote: > In message <3.0.5.32.19990510151933.017f1b90@staff.sentex.ca>, Mike > Tancsa writ > es: > > At 12:11 PM 5/10/99 -0700, Cy Schubert wrote: > > >I remember trying this exploit about a month ago when it was first > > >announced. It locked the system up solid within seconds. After > > >applying the patches from PR 8416 I currently have four copies of > > >find and one tripwire running for over an hour and a half, with no > > >hang. The patches from PR 8416 appear to have fixed the hang. > > > > > > That is most excellent... Is there any reason why this patch has not been > > applied to the tree ? i.e. does it have any adverse side effects > > preventing it from being committed ? > > > > ---Mike > > > > Version 1.24 of kern_lock.c does have the fix, however it relies on > macros that only exist in -current. Not being a FreeBSD kernel > wizard, I can only guess that the version of the fix in PR 8416 may > cause other problems or it might just be that version 1.24 of > kern_lock.c is simply a more elegant solution. If it is the > latter, then temporarily committing the PR 8416 version into > -stable might provide some relief until a more permanent solution > based on version 1.24 can be engineered into -stable. > > Version 1.24 of kern_lock.c only fixes read()/mmap() deadlocks. > Write()/mmap() deadlocks still have the potential of hanging the > system, which may answer why the solution hasn't been MFC'd yet, as > it's only a partial solution. [IMO a partial solution is better > than nothing, but then again I'm not a FreeBSD kernel wizard.] > > > Regards, Phone: (250)387-8437 > Cy Schubert Fax: (250)387-5766 > Open Systems Group Internet: Cy.Schubert@uumail.gov.bc.ca > ITSD Cy.Schubert@gems8.gov.bc.ca > Province of BC > "e**(i*pi)+1=0" > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-stable" in the body of the message > -- Cliff Skolnick | "They that can give up essential liberty to obtain Steam Tunnel Operations | a little temporary safety deserve neither liberty cliff@steam.com | nor safety." http://www.steam.com/ | -- Benjamin Franklin, 1759 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.10.9905101717560.453-100000>