Latest working kernel version: 2.6.23 Earliest failing kernel version: 2.6.24 Distribution: Ubuntu/Debian Lenny/Fedora Hardware Environment: laptop, LG LW75 Software Environment: Problem Description: Since 2.6.24 my laptop (LG LW75) hangs hard when pressing TAB once or twice. Sysrq is not functional after that. The hang does not happen when the boot parameter acpi=off is used. If I use one of the following: acpi=noirq pci=noacpi acpi_irq_balance acpi_irq_nobalance acpi_oldboot it still hangs. Reproducible on 2.6.25-rc5 please let me know what additional information you might need. Steps to reproduce: Boot in single user mode and press TAB twice.
According to the description it seems that it is a regression. Will you please attach the output of acpidump and lspci -vxxx? Will you please confirm whether the system still hangs if TAB is not pressed? Will you please attach the output of dmesg on the working kernel(2.6.23)? It will be great if you can get the output of dmesg through serial console on the kernel 2.6.25-rc5. ( If there is no serial console, please capture the picture and attach it) Thanks.
Created attachment 15479 [details] requested files acpidumps, lspci's and dmesgs of working and non-working systems: 2.6.23.17 and 2.6.25-rc7-git3
And yes, the system seems to hang randomkly later, if TAB is not pressed. The hangs seem to come when some key is pressed. For instance, once trying to (by accident) produce $ in shell hung the system. But Generally the system stays alive for a while before dying - like 1 to 5 minutes or so. The hang produces no console output. When running Debian, the system hung quite systematically at the same point in the boot sequence, saying something like "starting lircd" or so. But only 90% of the time, every now and then it would freeze at a different point.
This is a bug in the pcspkr driver. The system hangs systematically if for instance command-line program "beep" is used, or tab pressed twice, when the pcspkr-module is loaded. When pcspkr-module is not loaded, system does not hang.
See: http://lkml.org/lkml/2008/11/22/207 Care to try the fix ?
Yes, this fixes the bug for 2.6.28-rc6-git1. Without the patch the hang is reproducible. With the same config and same kernel version, with the patch applied, the system no more hangs. Thanks a lot!