Hi again, Octavio Alvarez wrote:
> The problem is a full system lock-up on resume. The HD led stays on. > The 5-sec power-button cycle isn't enough to get the system back. > It requires a full mechanical power cycle from the power supply. > > I bisected the problem as requested by Alan Stern, and tracked it > down to the following commit: [...] > ACPICA: ACPI 5: Support for new FADT SleepStatus, SleepControl registers Looks similar to https://bugzilla.kernel.org/show_bug.cgi?id=43641 The purpose of commit 2feec47d4c5f was to add support for ACPI 5 registers. ACPI 5 was published in November, 2011. Your BIOS is from 2008 and uses a version 3 FADT. I guess we'll have to sift through 5 files changed, 475 insertions(+), 679 deletions(-) for the unintended change. Hmm. Could you try the following in order to make the hunt easier? Please enable ACPI_DEBUG and ACPI_DEBUG_FUNC_TRACE: scripts/config --enable ACPI_DEBUG scripts/config --enable ACPI_DEBUG_FUNC_TRACE make deb-pkg; # maybe with -j4 reboot Set debug_layer to ACPI_HARDWARE and debug_level to 0xffffffff: echo 2 >/sys/module/acpi/parameters/debug_layer echo 0xffffffff >/sys/module/acpi/parameters/debug_level and get a log of what happens during the suspend and resume that locks up. You'll probably need to use a serial console or netconsole to get the log (boot with no_console_suspend to make sure log output continues to flow during the suspend process). [1] and [2] explain how to set up a serial console or netconsole. ([3] and [4] describe ACPI debugging and no_console_suspend.) It's possible and even likely that the kernel won't be awake enough to say anything interesting at the crucial moment, but I figure it's worth a try (and at least it can say a little about what happened during suspend). Thanks for your patience. Jonathan [1] Documentation/serial-console.txt [2] Documentation/networking/netconsole.txt [3] Documentation/acpi/debug.txt [4] Documentation/power/basic-pm-debugging.txt -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120708073758.GC4625@burratino
