Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
Also force pm-utils to disable powersaving features across a suspend/resume.
|
|
|
|
|
|
|
|
|
|
Replace all occurences of ; not only the first one
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
|
|
The native quirks DB separates alternatives with ';'. For regexp matching, turn
those into '|', so that they will actually match.
|
|
98-video-quirk-db-handler expects find_native() to write the quirks to stdout,
so actually do write them there. This makes the native quirks work.
|
|
canonicalize_dmivar() refused the valid property "system.firmware.release_date"
because of a missing '_' in the character filter.
|
|
Certian newer nvidia binary drivers do not chvt in kernelspace, which
leads to reported lockups. Go ahead and have them chvt in userspace again.
|
|
This can be useful for testing to see if your system requires an unusual set
of quirks, but can also make your system crash in new and exciting ways across
suspend/resume. You have been warned.
|
|
Some longstanding typos in the quirks were fixed in 98smart-kernel-video.
Update the new quirk handling machinery to take those fixes into account.
|
|
Conflicts:
pm/sleep.d/98smart-kernel-video
|
|
- fix wrong function calls to pci_restore and pci_save which did
not match the actual defined functions (restore_pci and save_pci)
- use the correct quirk parameter in "remove_paramters"
(use --quirk-save-pci instead of --quirk-pci-save)
|
|
We have been using vga-mode3 while the standard says vga-mode-3.
This might fix (or even break) some resume issues on hardware which did
not have any video quirks, since the option is set in
fdi/information/10freedesktop/99-video-quirk-default.fdi from hal-info.
http://bugs.freedesktop.org/show_bug.cgi?id=25334
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Signed-off-by: Victor Lowther <victor.lowther@gmail.com>
|
|
|
|
the last known working quirk.
|
|
Also add the functionality the documentation mentions.
|
|
|
|
|
|
script.
Integrating video quirk db handling into pm-utils requires splitting out
database querying from database importing anyways.
|
|
|
|
Add video-quirk-db-handler, which knows how to read the video .fdi files,
translate them into a native format, and use that native format to find the
appropriate quirks for a given system.
This hook replaces 00auto-quirk and 98smart-kernel-video.
|
|
We still need to consider the case where i915 is used with modesetting
disabled. So revert the have_smart_intel removal in
504faf0a0c31cbdbc03a608cf633d58f12e49eb7. Thanks, mbiebl!
|
|
kpowersave pops up a window telling that an error "1" has occurred when
suspending to ram. The error "1" message is annoying and lacks meaning
to the user.
It's caused by maybe_deallocvt returning the ret_val of state_exists
when the tested (here: console) file isn't found. This absence of this file
seems not to do any harm therefore the escalation of the ret_val 1 is
deprecated.
Solution: just tell maybe_deallocvt to return 0 in the above case.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=554959
|
|
|
|
It turns out that the method we were using to detect kernel modesetting
support was not very accurate. It turns out there is no bulletproof way
for us to tell if KMS is being used, but Michael Biebl found a way that
sucks less than out current method. Hopefully the framebuffer drivers
will grow a flag in sysfs somewhere that tells us that KMS is in use.
|
|
|
|
Suprisingly, there are not that many.
|
|
The function is correctly called add_parameters.
Debian bug http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=524243
|
|
|
|
Also factor some things out to make 98-smart-kernel-video shorter.
|
|
This prepares for adding kernel modesetting support.
|
|
|
|
This will make pm-utils slightly easier to debug.
Thanks, Michael Biebl for the idea.
|
|
|
|
ntpd operates normally over a sleep/wake cycle. Stoppping and
restarting it in a sleep hook is too adhoc. If it is essential
that ntpd be stopped and restarted, then having NetworkManager
do the job when interfaces go offline and come back online is
a better way to do things.
|
|
alsa drivers that support suspend/resume will save their state without it.
alsa drivers that don't will need to be unloaded and reloaded as well as
having their state saved and restored to work properly.
See thread at http://mailman.alsa-project.org/pipermail/alsa-devel/2008-November/013030.html
|
|
Most systems do not need it. Add NEED_CLOCK_SYNC, and only run hwclock
if it is set. This will probably go away entirely in a later release.
|
|
HAL has handled rescanning the batteries since mid 2006.
|
|
This function works just like dbus-send, but returns $NA if the command
fails for any reason. Also modified hooks to use dbus_send where applicable.
|
|
If no --quirk-s3-* option is passed to pm-suspend, set acpi_video_flags
to 0. This allows to reliably test different set of quirks and clears
any flags that were set via the kernel boot option. We deliberately
ignore the kernel acpi_sleep option, quirks should be set via HAL.
|
|
|
|
|
|
|
|
|
|
Instead, save them in /etc/hal/fdi/information.
Closes an airline-plot potential DoS attack.
|
|
|
|
|