The mainline kernel you installed is maintained by the Ubuntu kernel team. It is intended for testing only.
The kernel team did not enable the gma500_gfx module in the kernel .config , so it is not enabled.
You need to file a bug report with the kernel team and ask them to enable the gma500_gfx for the mainline kernel build or build a custom kernel yourself.
IMO a better option is to test the 3.4 kernel wither on 12.10 or Fedora 17.
Keep in mind we do not support custom kernels.
https://help.ubuntu.com/community/Kernel/CompileBuilding and using a custom kernel will make it very difficult to get support for your system.
While it is a learning experience to compile your own kernel, you will not be allowed to file bugs on the custom-built kernel (if you do, they will be Rejected without further explanation).
And
So while you are free to build kernels and learn to do so, support is limited at best. If the kernel team is not offering support ...The mainline kernels builds are produced for debugging purposes and therefore come with no support. Use them at your own risk.
I think you are better off filing a bug report.
Why are you trying to use the 3.4 kernel ? Why are your building a custom kernel ?
https://wiki.ubuntu.com/Kernel/MainlineBuilds
There are two mistakes one can make along the road to truth...not going all the way, and not starting.
--Prince Gautama Siddharta
#ubuntuforums web interface
Hello,
On my computer, this works very fine (and faster) with 3.4 from Quantal. I followed these instructions :
http://www.upubuntu.com/2012/05/how-...stable-on.html
N.B : just after that, I removed Quantal lines from my sources.list to avoid unwanted updates… I think it would be possible to tune apt preferences, but I don't think all that is really supported.
Hope this will help.
Nice day!
I have finally achieved success in installing Quantal kernel too and another problem I want to solve now is that the backlight brightness on my t91mt is still not easy to change.
There are some script workarounds but what about a hardcoded variant?
I have rebuilt gsd-backlight-helper with hardcoded psb-bl and it's really working!!!
So, for those who are fed up with all these scripts and console brightness...
download the file, unzip it and move the program with something like
It will be immidiately available for changing brightness and make you happyCode:sudo cp gsd-backlight-helper /usr/lib/gnome-settings-daemon
Hi!
Have you suspend working under 3.4? For me there is a regression in this kernel...
An explanation for my workaround to make backlight be set natively with the keys:
Get the sources of gnome-settings-daemon:
It's probably in your home folder, go there and in plugins/power you will find gsd-backlight-helper.c.Code:apt-get source gnome-settings-daemon
That's the file that should be edited, so open it with gedit and you will see c-code.
There are two functions which should give the program path to the backlight changer. I can't fix them to work properly for everyone as I don't know C/C++ but i was able to replace them with simply a return to the needed psb-bl, which is enough for all the users of gma500 who have the problem.
Than, compilation as usual and the only file we need is the one I provided previously that should be placed in the folder with actually working gsd.
If anyone knows what to do with these functions to make a proper fix which can be uploaded as a patch, that would be better... but for now it's enough to replace it once and fel like everything is fixed.
Original code:
Code:static gchar * gsd_backlight_helper_get_type (GList *devices, const gchar *type) { const gchar *type_tmp; GList *d; for (d = devices; d != NULL; d = d->next) { type_tmp = g_udev_device_get_sysfs_attr (d->data, "type"); if (g_strcmp0 (type_tmp, type) == 0) return g_strdup (g_udev_device_get_sysfs_path (d->data)); } return NULL; } static gchar * gsd_backlight_helper_get_best_backlight () { gchar *path = NULL; GList *devices; GUdevClient *client; client = g_udev_client_new (NULL); devices = g_udev_client_query_by_subsystem (client, "backlight"); if (devices == NULL) { g_warning ("failed to find any devices"); goto out; } /* search the backlight devices and prefer the types: * firmware -> platform -> raw */ path = gsd_backlight_helper_get_type (devices, "firmware"); if (path != NULL) goto out; path = gsd_backlight_helper_get_type (devices, "platform"); if (path != NULL) goto out; path = gsd_backlight_helper_get_type (devices, "raw"); if (path != NULL) goto out; out: g_object_unref (client); g_list_foreach (devices, (GFunc) g_object_unref, NULL); g_list_free (devices); return path; }
Hmm, when I choose 3.2.0 to boot it suspends with the fix about quirk-vbemode-restore
But with 3.4 it's just hanging with black screen and bright backlight and won't go suspend... and even resume from this state
hi i am using a sony vaio vpc-p
my backlight/display is 70% of the time flickering
resolution is 1600x768
i dont know if its backlight or the pixels, but changing of brightness doesnt affect flickering.
if you get close enough you can see diagonale lines on bright areas.
i am not sure if they are related to the flickering.
i am using ubuntu 12.04
it is extremely hard to find information for my laptop
I wanted to report the state of my Sony Vaio P series VGN-P788K. I tested with a 12.04 32-bit install with all updates as of 2012-05-28 applied. The kernels used were 3.2.0-24 from precise and 3.4.0-3 from quantal.
Graphics
- 3.2: Boots to black screen. Switching to another tty then back to tty7 brings up graphics. They are always broken; only the upper half of the screen is used.
- 3.2 with console=tty1: Switching to another tty then back to tty7 brings up graphics. They are usually correct, but sometimes only the left two-thirds of the screen is used.
- 3.4: Switching to another tty then back to tty7 brings up graphics. They are always correct.
- 3.4 with console=tty1: No different than 3.4 without this option.
Suspend and Resume
- 3.2: Suspend and resume work out of the box.
- 3.4: Suspend does not work. Although the screen goes blank, the laptop never suspends. I tried all the possible quirk options to pm-suspend but none changed this behavior.
Bookmarks