Battery-related regression between 2.6.17-git3 and 2.6.17-git6

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

(ie. after reverting them hald works again).

Quoted 28 lines · Rafael J. Wysocki

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

Quoted 1 line

(ie. after reverting them hald works again).

Ick, that should not cause any problems, as sysfs should look identical to how it was before those patches. Except that the /sys/class/usb/ stuff is now symlinks instead of real directories, but HAL has had to handle that for a long time now (and it's even documented in Documentation/ABI/testing/sysfs-class)

Can you tell me exactly which of the above patches breaks HAL?

Which version of HAL are you using? I have 0.5.7 here and it works just fine.

And why would they even matter? The battery is not a USB device...

Quoted 29 lines · Greg KH

Quoted 28 lines · Rafael J. Wysocki

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

Quoted 7 lines

Quoted 1 line

(ie. after reverting them hald works again).

Ick, that should not cause any problems, as sysfs should look identical to how it was before those patches. Except that the /sys/class/usb/ stuff is now symlinks instead of real directories, but HAL has had to handle that for a long time now (and it's even documented in Documentation/ABI/testing/sysfs-class)

Well, apparently one of them happens to trigger a buffer overflow in "my" version of hal. ;-)

Quoted 1 line

Can you tell me exactly which of the above patches breaks HAL?

That would be quite a bit of testing and now I'm sure it's a hal issue.

Quoted 2 lines

Which version of HAL are you using? I have 0.5.7 here and it works just fine.

0.5.4 :-(

Quoted 41 lines · Rafael J. Wysocki

Quoted 29 lines · Greg KH

Quoted 28 lines · Rafael J. Wysocki

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

Quoted 7 lines

Quoted 1 line

(ie. after reverting them hald works again).

Ick, that should not cause any problems, as sysfs should look identical to how it was before those patches. Except that the /sys/class/usb/ stuff is now symlinks instead of real directories, but HAL has had to handle that for a long time now (and it's even documented in Documentation/ABI/testing/sysfs-class)

Well, apparently one of them happens to trigger a buffer overflow in "my" version of hal. ;-)

Quoted 3 lines

Quoted 1 line

Can you tell me exactly which of the above patches breaks HAL?

That would be quite a bit of testing and now I'm sure it's a hal issue.

git bisect would help out a lot. Or just ask the HAL developers, they might know.

Quoted 4 lines

Quoted 2 lines

Which version of HAL are you using? I have 0.5.7 here and it works just fine.

0.5.4 :-(

Can you upgrade to a newer version? SuSE 10.1 is out which should work just fine...

Quoted 46 lines · Greg KH

Quoted 44 lines · Rafael J. Wysocki

Quoted 29 lines · Greg KH

Quoted 28 lines · Rafael J. Wysocki

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

Quoted 7 lines

Quoted 1 line

(ie. after reverting them hald works again).

Ick, that should not cause any problems, as sysfs should look identical to how it was before those patches. Except that the /sys/class/usb/ stuff is now symlinks instead of real directories, but HAL has had to handle that for a long time now (and it's even documented in Documentation/ABI/testing/sysfs-class)

Well, apparently one of them happens to trigger a buffer overflow in "my" version of hal. ;-)

Quoted 1 line

Can you tell me exactly which of the above patches breaks HAL?

That would be quite a bit of testing and now I'm sure it's a hal issue.

Quoted 2 lines

git bisect would help out a lot. Or just ask the HAL developers, they might know.

Anyway I'd have to compile and test at least a couple of kernels. [For the record: I'm quite sure that 36679ea59846d8f34a48f71ca1a37671ca0ad3c5 and 9bde7497e0b54178c317fac47a18be7f948dd471 together break hal on my system; this seems to be related to endpoints' paths in sysfs.]

Quoted 47 lines · Rafael J. Wysocki

Quoted 46 lines · Greg KH

Quoted 44 lines · Rafael J. Wysocki

Quoted 29 lines · Greg KH

Quoted 28 lines · Rafael J. Wysocki

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

Quoted 7 lines

Quoted 1 line

(ie. after reverting them hald works again).

Ick, that should not cause any problems, as sysfs should look identical to how it was before those patches. Except that the /sys/class/usb/ stuff is now symlinks instead of real directories, but HAL has had to handle that for a long time now (and it's even documented in Documentation/ABI/testing/sysfs-class)

Well, apparently one of them happens to trigger a buffer overflow in "my" version of hal. ;-)

Quoted 1 line

Can you tell me exactly which of the above patches breaks HAL?

That would be quite a bit of testing and now I'm sure it's a hal issue.

Quoted 7 lines

Quoted 2 lines

git bisect would help out a lot. Or just ask the HAL developers, they might know.

Anyway I'd have to compile and test at least a couple of kernels. [For the record: I'm quite sure that 36679ea59846d8f34a48f71ca1a37671ca0ad3c5 and 9bde7497e0b54178c317fac47a18be7f948dd471 together break hal on my system; this seems to be related to endpoints' paths in sysfs.]

I don't understand why that would break HAL, we are just adding new devices to the sysfs device tree, which the kernel is free to do at any time. HAL should not care about that.

Oh, and 36679ea59846d8f34a48f71ca1a37671ca0ad3c5 is just an internal api change, it does not affect userspace in any way. So I don't see how that would have anything to do with HAL at all.

Quoted 56 lines · Greg KH

Quoted 47 lines · Rafael J. Wysocki

Quoted 46 lines · Greg KH

Quoted 44 lines · Rafael J. Wysocki

Quoted 29 lines · Greg KH

Quoted 28 lines · Rafael J. Wysocki

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

Quoted 7 lines

Quoted 1 line

(ie. after reverting them hald works again).

Ick, that should not cause any problems, as sysfs should look identical to how it was before those patches. Except that the /sys/class/usb/ stuff is now symlinks instead of real directories, but HAL has had to handle that for a long time now (and it's even documented in Documentation/ABI/testing/sysfs-class)

Well, apparently one of them happens to trigger a buffer overflow in "my" version of hal. ;-)

Quoted 1 line

Can you tell me exactly which of the above patches breaks HAL?

That would be quite a bit of testing and now I'm sure it's a hal issue.

Quoted 7 lines

Quoted 2 lines

git bisect would help out a lot. Or just ask the HAL developers, they might know.

Anyway I'd have to compile and test at least a couple of kernels. [For the record: I'm quite sure that 36679ea59846d8f34a48f71ca1a37671ca0ad3c5 and 9bde7497e0b54178c317fac47a18be7f948dd471 together break hal on my system; this seems to be related to endpoints' paths in sysfs.]

Quoted 7 lines

I don't understand why that would break HAL, we are just adding new devices to the sysfs device tree, which the kernel is free to do at any time. HAL should not care about that.

Oh, and 36679ea59846d8f34a48f71ca1a37671ca0ad3c5 is just an internal api change, it does not affect userspace in any way. So I don't see how that would have anything to do with HAL at all.

Could you please have a look at the end of the attached output of 'strace -f /usr/sbin/hald --daemon=yes --retain-privileges --verbose=yes' (produced on vanilla 2.6.17-git4)?

I'm not sure what exactly happens there, but I think hal crashes due to a buffer overflow.

Quoted 65 lines · Rafael J. Wysocki

Quoted 56 lines · Greg KH

Quoted 47 lines · Rafael J. Wysocki

Quoted 46 lines · Greg KH

Quoted 44 lines · Rafael J. Wysocki

Quoted 29 lines · Greg KH

Quoted 28 lines · Rafael J. Wysocki

Hi,

Quoted 7 lines · Rafael J. Wysocki

Quoted 6 lines · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10) the powersave demon is apparently unable to get the battery status, although the data in /proc/acpi/battery/BAT0 seem to be correct. As a result, battery status notification via kpowersave doesn't work and it's hard to notice when the battery is low/critical.

Quoted 7 lines

Quoted 5 lines

So far I have verified that this feature works fine with 2.6.17-git3 and doesn't work with 2.6.17-git6 (-git5 doesn't compile here).

I'll try to get more information tomorrow (unless someone in the know has an idea of what's up ;-) ).

I've verified that the problem first appeared in 2.6.17-git4.

Apparently this happens because powersaved takes the battery status information from hald and the following kernel changes make hald crash on my system:

http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=43104f1da88f5335e9a45695df92a735ad550dda
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=bd00949647ddcea47ce4ea8bb2cfcfc98ebf9f2a
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=c182274ffe1277f4e7c564719a696a37cacf74ea
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9bde7497e0b54178c317fac47a18be7f948dd471
http://kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=36679ea59846d8f34a48f71ca1a37671ca0ad3c5

Quoted 7 lines

Quoted 1 line

(ie. after reverting them hald works again).

Ick, that should not cause any problems, as sysfs should look identical to how it was before those patches. Except that the /sys/class/usb/ stuff is now symlinks instead of real directories, but HAL has had to handle that for a long time now (and it's even documented in Documentation/ABI/testing/sysfs-class)

Well, apparently one of them happens to trigger a buffer overflow in "my" version of hal. ;-)

Quoted 1 line

Can you tell me exactly which of the above patches breaks HAL?

That would be quite a bit of testing and now I'm sure it's a hal issue.

Quoted 7 lines

Quoted 2 lines

git bisect would help out a lot. Or just ask the HAL developers, they might know.

Anyway I'd have to compile and test at least a couple of kernels. [For the record: I'm quite sure that 36679ea59846d8f34a48f71ca1a37671ca0ad3c5 and 9bde7497e0b54178c317fac47a18be7f948dd471 together break hal on my system; this seems to be related to endpoints' paths in sysfs.]

Quoted 7 lines

I don't understand why that would break HAL, we are just adding new devices to the sysfs device tree, which the kernel is free to do at any time. HAL should not care about that.

Oh, and 36679ea59846d8f34a48f71ca1a37671ca0ad3c5 is just an internal api change, it does not affect userspace in any way. So I don't see how that would have anything to do with HAL at all.

Quoted 6 lines

Could you please have a look at the end of the attached output of 'strace -f /usr/sbin/hald --daemon=yes --retain-privileges --verbose=yes' (produced on vanilla 2.6.17-git4)?

I'm not sure what exactly happens there, but I think hal crashes due to a buffer overflow.

Yes, that looks like what is happening. Perhaps one of the HAL developers can point you at a patch that you can apply to your version of HAL to get it working.

Either way, this is not a kernel bug, as it could have happened with any very long depth device tree, you were just lucky it didn't happen sooner.

Quoted 4 lines · Greg KH

Quoted 3 lines · Rafael J. Wysocki

Quoted 2 lines · Greg KH

Quoted 1 line · Rafael J. Wysocki

On Monday 03 July 2006 21:44, Greg KH wrote:

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines

Quoted 8 lines · Rafael J. Wysocki

Quoted 7 lines · Greg KH

Quoted 6 lines · Rafael J. Wysocki

Quoted 5 lines

Hi,

Quoted 2 lines · Rafael J. Wysocki

Quoted 1 line · Rafael J. Wysocki

With the recent -git on my box (Asus L5D, x86_64 SUSE 10)