On Wed, Aug 24, 2011 at 03:50:53PM -0400, Jim Faulkner wrote: > On Mon, Aug 22, 2011 at 05:36:02PM -0700, David Miller wrote: > > > Unfortunately we need the very first crash message that happens, but > > even your first snapshot is after other crashes have occurred. > > > > You'll need to build a kernel with CONFIG_BOOT_PRINTK_DELAYED enabled > > and then you can boot with boot_delay=xxx (where "xxx" is something > > like 100, it's in milliseconds) to slow down the message printing so > > that you can capture the first message cleanly. > > OK, I've attached several images of the kernel panic w/ boot_delay to > the bug: > > https://bugzilla.kernel.org/show_bug.cgi?id=41522 https://bugzilla.kernel.org/attachment.cgi?id=70042 (the third image) looks like it first finds two address space collisions with ali7101... but they're both listed at 0000:00:03.0 (i.e. not a collision of two different PCI devices but one and the same), and that quirk code in drivers/pci/quirks.c is ancient (2005, 2007). The next problem is a WARN about 0000:00:07.0 being added twice, but latest change in that sabre_* function seems to be a generic OF matching code change, so that's not immediately suspicious either. It looks as if something is generally screwing with the PCI detection logic from the get-go. Do you happen to have a working kernel to fall back to, and then do a git bisect? -- 2. That which causes joy or happiness. -- To unsubscribe from this list: send the line "unsubscribe sparclinux" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
