From owner-freebsd-stable Fri Mar 23 17:28:21 2001 Delivered-To: freebsd-stable@freebsd.org Received: from earth.backplane.com (earth-nat-cw.backplane.com [208.161.114.67]) by hub.freebsd.org (Postfix) with ESMTP id 750B537B71B for ; Fri, 23 Mar 2001 17:28:18 -0800 (PST) (envelope-from dillon@earth.backplane.com) Received: (from dillon@localhost) by earth.backplane.com (8.11.2/8.9.3) id f2O1S2V87488; Fri, 23 Mar 2001 17:28:02 -0800 (PST) (envelope-from dillon) Date: Fri, 23 Mar 2001 17:28:02 -0800 (PST) From: Matt Dillon Message-Id: <200103240128.f2O1S2V87488@earth.backplane.com> To: Dave Tweten Cc: freebsd-stable@FreeBSD.ORG Subject: Re: 4.3-RC Kernel Buffer Corruption (Was: 4.3-BETA makeworld of current STABLE Fails) References: <200103232356.f2NNuq601634@gilmore.nas.nasa.gov> Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG : FreeBSD 4.3-RC (FLOATER) #2: Thu Mar 22 23:19:11 PST 2001 : :and installed it, I tried to buildworld again. As before, it stopped in :"stage 4: building libraries." As before, the stopper was source file :(/usr/src/lib/libc/stdio/setvbuf.c) corruption in kernel buffers by binary :garbage. It started on a 1024-byte boundary (4096 bytes in) and extending for :1024 bytes. Again, the file itself was not corrupted. : :For review, Floater is an NEC Versa 6050MX with 48 megabytes of main memory :and 2 4-Gig disk drives. A kernel I built on February 16 is the last one that :doesn't exhibit buffer corruption. The problem has been consistent across a :previous kernel config file and the current one. : :The previous config file fell victim to my blowing away all of /usr/src and :re-cvsupping before I realized that the corruption wasn't on the disk. This :explanation is by way of saying that the attached config file is a work in :progress, but it couldn't have caused the identical buffer corruption observed :with its predecessor. As a kicker, the previous config file was used to build :the February 16 kernel that works. : :I'd call this a release show-stopper. A bunch of questions: * Is the corruption in the same file every time you try the buildworld? * Is the corruption at the same offset/length? * How are you monitoring the corruption? ktrace? cat? vi? * How definitive a kernel -stable date can you lock the corruption down at? Judging from this and prior messages, somewhere between Feb16 and Mar1 ? * Do you have softupdates enabled? If so, try turning them off. * Are you using any special sysctl's ? -Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message