https://bugzilla.novell.com/show_bug.cgi?id=228344 ------- Comment #25 from bk@novell.com 2007-02-09 09:58 MST ------- Just a quick update: I have found that the execution of set_fixed_ranges() which sets up the fixed-range MTRRs cause the sata_sil driver to fail by misreading the size of sda in the resume phase while suspending to disk by returning from it without setting up the fixed-range MTRRs. The reset while resuming from disk while initializing the CPU (I think 2nd core, have to check to logs closer) which happened with the vanilla 2.6.20-rc5 is also gone when set_fixed_ranges() is changed to not set the fixed-range MTRRs. As Andi suggested, I'll try to nail down which of the MTRRs causes the problem. Another idea which comes to my mind: As the MTRR are AFAICS are initially set-up by the BIOS, I guess a look for an BIOS upgrade could also be useful, but in case the BIOS does not allow downgrading, I will not be able to further debug the issue then, so I am a bit hestitant to that. However, it would be good to know if the BIOS changelog has something, I guess. I do not yet really understand what the kernel does with the fixed-range MTRRs: I have not seen any possibility in the kernel to change the setup of the fixed-range MTRRs, so it looks to me that set_fixed_ranges() should normally do nothing to the MTRRs, as nobody ever asks to change something, so if everything is correct, it should not try to write to them, it should normally only verify and restore them in case the initial setup from the BIOS is lost sometime. I hope that we can sched some light on this.... For a start, I have added some debug printks to the function to see more what it's doing and where it's causing the reset and the sata_sil trouble. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.