Since kernel major version 2.6.36 my Samsung R40 Cinoso notebook sometimes looses the ability to use the keyboard. Most times this happens when the screensaver is active and I cannot enter anything for the password. Remote login is possible and dmesg is giving tons of these errors: Jan 1 18:54:49 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Jan 1 18:54:49 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Jan 1 18:54:49 (none) kernel: ACPI Exception: AE_TIME, Returned by Handler for [EmbeddedControl] (20100702/evregion-474) Jan 1 18:54:49 (none) kernel: ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.BAT1._BST] (Node ffff880077820208), AE_TIME Jan 1 18:54:49 (none) kernel: ACPI Exception: AE_TIME, Evaluating _BST (20100702/battery-441) Jan 1 18:54:50 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Jan 1 18:54:50 (none) kernel: ACPI Exception: AE_TIME, Returned by Handler for [EmbeddedControl] (20100702/evregion-474) Jan 1 18:54:50 (none) kernel: ACPI Error (psparse-0537): Method parse/execution failed [\_SB_.BAT1._STA] (Node ffff880077820230), AE_TIME Jan 1 18:54:50 (none) kernel: ACPI Exception: AE_ERROR, Evaluating _STA (20100702/battery-382) Jan 1 18:54:50 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Any idea? Do you need more logs? -Andy
The failure occurs in 2.6.36.2, but does not occur in 2.6.36? Please attach the output from acpidump. does it happen only when the screen saver kicks in? please attach the complete dmesg
Created attachment 42412 [details] clean dmesg
When I remember right I had it also once when the screen was turning off and even before I had to enter a password the desktop didn't accept any input. My distribution doesn't ship acpidump. I will build it myself from source. No problem. Can you tell me when I should run that command? Is it needed to run it when the notebook is looked via ssh or can I run it anytime?
Created attachment 45462 [details] acpidump acpidump added. so far it has not happened again, now running 2.6.37. btw: LCD backlight always stays on in screensaver mode. maybe this is another issue.
close the bug report as it can not be reproduced in 2.6.37. please feel free to re-open it if you see the problem again in the latest upstream kernel.
This is still a problem in 2.6.39.2. I'm running now Awesome desktop without any screensaver. It just blanks the screen after some time out, guess this comes from the bios. No keyboard input is possible then. Mouse is still working so I can shutdown the notebook. Still the same messages in dmesg when this happens: Jul 6 21:01:01 (none) /USR/SBIN/CROND[3051]: (root) CMD (run-parts /etc/cron.hourly) Jul 6 21:15:05 (none) -- MARK -- Jul 6 21:21:54 (none) ntpd[966]: skew change 10.833 exceeds limit Jul 6 21:30:17 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Jul 6 21:30:17 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Jul 6 21:30:17 (none) kernel: ACPI Exception: AE_TIME, Returned by Handler for [EmbeddedControl] (20110316/evregion-476) Jul 6 21:30:17 (none) kernel: ACPI Error: Method parse/execution failed [\_SB_.BAT1._BST] (Node ffff880074a3d208), AE_TIME (20110316/psparse-536) Jul 6 21:30:17 (none) kernel: ACPI Exception: AE_TIME, Evaluating _BST (20110316/battery-455) Jul 6 21:30:18 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Jul 6 21:30:18 (none) kernel: ACPI Exception: AE_TIME, Returned by Handler for [EmbeddedControl] (20110316/evregion-476) Jul 6 21:30:18 (none) kernel: ACPI Error: Method parse/execution failed [\_SB_.BAT1._BST] (Node ffff880074a3d208), AE_TIME (20110316/psparse-536) Jul 6 21:30:18 (none) kernel: ACPI Exception: AE_TIME, Evaluating _BST (20110316/battery-455) Jul 6 21:30:18 (none) kernel: ACPI: EC: input buffer is not empty, aborting transaction Do you need any further information?
It's great that kernel bugzilla is back. Actually, I do not have any idea about this problem so far. can you please verify if the problem still exists in the latest upstream kernel first?
This Bug is still present from time to time in 3.2.1.
Situation has become worse meanwhile with recent kernels. Whith old kernels this bug seems only triggered by lid close events. Tested 3.6.2 kernel and the bug happened quickly at regular desktop use without any special ACPI event. This make new kernel unusable with this notebook.
Can you apply the following patch and attach the output of dmesg? This patch is to open ec debug info and will be helpful to get more clues. https://bugzilla.kernel.org/attachment.cgi?id=95921
Too late. I don't own that device anymore. Feel free to close this one for now.
Ok. Sorry for late reply.