Date: Sun, 17 May 1998 18:57:57 +0200 From: Mark Murray <mark@grondar.za> To: "Justin T. Gibbs" <gibbs@plutotech.com> Cc: gibbs@FreeBSD.ORG, scsi@FreeBSD.ORG Subject: Re: CAM broken with SMP/current/AHC Message-ID: <199805171657.SAA03015@greenpeace.grondar.za>
next in thread | raw e-mail | index | archive | help
"Justin T. Gibbs" wrote: > I started hunting around for where the system launches the other CPUs, but > I haven't found it yet. My guess is that just at the point of startup, sys/i386/i386/mp_machdep.c > these CPUs are vulnerable to receiving interrupts at the same time as CPU > 0. I've done some analysis with Russell Cattelan on this, and the only > way I can see for this error to occur is if one of the interrupt driver > lists is corrupted by unexpected interrupt re-entrance. Sounds reasonable. Unfortunately I know very little about the kernel. I'll play.. > Why wouldn't this affect the old SCSI code? CAM continues with the boot > process while still configuring the devices on the bus, but after > registering the presence of the devices to the system. This allows the > boot to block only during the configuration of the devices it cares about > (like your root disk). But, this does open the window for I/O to occur > while extra CPUs are started. Could slpx() and friends be the answer? M -- Mark Murray Join the anti-SPAM movement: http://www.cauce.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805171657.SAA03015>