Date: Thu, 12 Oct 2000 07:10:03 -0700 (PDT) From: "Justin T. Gibbs" <gibbs@scsiguy.com> To: freebsd-bugs@FreeBSD.org Subject: Re: kern/21915: Machine dies sig 12 in ahc driver (Freebsd4.1.1) Message-ID: <200010121410.HAA19529@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/21915; it has been noted by GNATS. From: "Justin T. Gibbs" <gibbs@scsiguy.com> To: Josef Karthauser <joe@pavilion.net> Cc: systems@pavilion.net, FreeBSD-gnats-submit@freebsd.org Subject: Re: kern/21915: Machine dies sig 12 in ahc driver (Freebsd4.1.1) Date: Wed, 11 Oct 2000 14:51:23 +0000 >Had another go today, incorporated Justin's MFC of last night. >The problem still exists though. Is the stack trace identical? >My suspicion is that there's a race condition somewhere which is >causing the bus to crash - somewhere between the ahc and the fxp >drivers. I've not investigated deeper though, not having any >familiarity with the device driver code. If the buffer pointer is still null in this latest instance, one thing to try would be to instrument both dastrategy and dastart in sys/cam/scsi/scsi_da.c. dastrategy is the entry point to the CAM system for buffers headed for a disk. If bp->b_data is NULL at that point, we've pretty much exonerated the da driver, the CAM subsystem and ahc. If you don't hit that panic, then dastart is the transfer point from the DA driver to the XPT and then ahc driver. It would be interesting to know if the buffer pointer has been zero'ed by that point. -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200010121410.HAA19529>