On Mon, Dec 11, 2006 at 04:18:12PM -0800, Jay Lan wrote: > We found in testing that certain PCI devices don't handle the > kexec'ed kernel reboot gracefully on IA64 Altix. Some GbE card > would get into a bad state permanently after one reboot, and > resulted in MCA on the machines when booting kexec'ed kernel > thereafter. The MCA was a result of a PIO read to the Expansion > ROM of the card. We also saw a USB card issuing stale DMA's > which was queued up prior to the kexec reboot. > > Do you observe similar problems on other platforms? Are the drivers > supposed to reset the cards when it re-attach to the card on > the kexec reboot? I think that the simple answer to that is yes. However isn't it the case that this simply isn't possible for some hardware? There was talk of passing a boot parameter to the second kernel so it knew it was a second kernel and drivers could take evasive action as needed. I'm not sure what happened to that idea. -- Horms H: http://www.vergenet.net/~horms/ W: http://www.valinux.co.jp/en/ - To unsubscribe from this list: send the line "unsubscribe linux-ia64" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
|