Bug 12360
| Summary: | with 2.6.28 undock/dock work bad or freeze computer - Dell D430 | ||
|---|---|---|---|
| Product: | ACPI | Reporter: | Daniel Smolik (marvin) |
| Component: | Config-Hotplug | Assignee: | Shaohua (shaohua.li) |
| Status: | CLOSED CODE_FIX | ||
| Severity: | normal | CC: | acpi-bugzilla, holger |
| Priority: | P1 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | 2.6.28 | Subsystem: | |
| Regression: | Yes | Bisected commit-id: | |
| Attachments: |
dmesg output
patch which may be solve my problem dmesg output when dock/undock is OK Move bailing out down the function Patch containing possible fix and some debug output dmesg when notebook is docked dmesg when notebook is undocked Don't eval _STA on every sysfs read |
||
|
Description
Daniel Smolik
2009-01-04 12:29:58 UTC
Created attachment 19649 [details]
dmesg output
Sorry Dell is D430 > After this I lost keyboard and mouse
is this always reproducible during the docking?
can you login the computer via network?
what do you see in the screen?
please attach the dmesg output when the laptop is not frozen after docking.
Created attachment 19664 [details]
patch which may be solve my problem
Created attachment 19665 [details]
dmesg output when dock/undock is OK
Not always, sometime I must dock/undock two times to get to this situation. I don't know I am on mountains and have't another computer,but next week I can test it. Screen is no corrupted I see normal KDE desktop. I try diff dock.c from 2.6.26 to 2.6.28 and may be I found workaround. With this small patch system work well. Dmesg output attached. Created attachment 19672 [details]
Move bailing out down the function
Could you please also test this patch? It's quite the same except that it cares about the uevent being sent.
With your patch docking/undocking stop working:-( After pressing undock button it starts flashing and in dmesg nothing is displayed. But if undocking work button must light off not flash. I got a report that the Dell D620 has the same problem. Fortunately I've got such a system in my office, so I can do some debugging on monday. Do you have full-fledged userspace running when trying to undock? Anyway, I'm going to attach a patch for you to test. It contains one possible fix and some additional debug output. Please test and report back with attaching your dmesg. Thanks. Created attachment 19785 [details]
Patch containing possible fix and some debug output
Yes, it looks like that patch solve problem. A attach dmesg outputs. Please look to it. Created attachment 19824 [details]
dmesg when notebook is docked
dmesg when notebook is docked
Created attachment 19825 [details]
dmesg when notebook is undocked
dmesg when notebook is undocked
Created attachment 19826 [details]
Don't eval _STA on every sysfs read
That's the patch without the debugging output I will send upstream if it proves ok. Please just answer the question if you have a full userspace operating system, even with desktop, running while testing this. Especially, if the hal daemon was active. Thanks.
Yes I run Debian Sid with KDE and HAL running. I it tested with console too. Shaohua, I sent the patch to linux-acpi: [PATCH] ACPI: dock: Don't eval _STA on every show_docked sysfs read Please close the bug. Thanks. note that bugzilla.kernel.org sightings are RESOLVED when a patch is available, and are CLOSED only when the patch ships upstream. patch in comment #15 applied to acpi tree fc5a9f8841ee87d93376ada5d73117d4d6a373ea (ACPI: dock: Don't eval _STA on every show_docked sysfs read) shipped in 2.6.29-rc4 closed. |