Notes on linux on the Sony Vaio PCG-GR214MP

I will add new information (if any) to the top of this page. Each added section will be dated at the beginning.

2nd January 2002

• Useful information re ACPI support found here on the acpi4linux mailing list. This appears to suggest that the GR214MP will work fairly sensibly with 2.4.17 and the latest acpi patches from the development site.

  I haven't had a chance to check this out myself yet.

1st January 2002

• Spoke too soon - while the i810_audio.c version (0.13) mentioned below works fine with xmms (for instance), it fails horribly with timidity. The symptom is that around half a second of sound is played, followed by half a second of silence, followed by half a second of sound ...

  I haven't yet determined if this is a problem with the driver or the application.

30th December 2001

• Sound didn't start working on this until version 0.13 of the i810_audio.c module (see http://people.redhat.com/dledford/i810_audio.c.gz). I'm running this with a 2.4.16 kernel.
• to avoid agpgart core-dumping, you need a version that knows about the intel 830MP agp chipset. This hasn't yet been integrated into the 2.4 kernel sources (at version 2.4.17 at any rate), but the relevant patch was posted to the linux-kernel list. It should be available at http://marc.theaimsgroup.com/?l=linux-kernel&m=100686414511876&w=2.
• Once you've got agpgart to work, you can get DRI working sort of if you get hold of a CVS version of XFree86 . With 4.1.99.1, DRI works, with the following three gotchas:
  • when particularly intensive GL operations are happening, noise appears across the whole of the display (not just the direct rendered window).
  • certain operations temporarily corrupt the display (the particular example observed is the enlightenment window manager doing an animated iconify of a large window, which leaves a sequence of black outlines of the window across the screen).
  • the Xserver has a habit of crashing the machine when it (the Xserver) is shut down.
  I'm waiting for XFree86 4.2 to come out, but in the meantime, if you discourage the agpgart module from loading, XFree86 will not attempt to do DRI, and will run much more stably ...
• the eepro100 driver doesn't work well with the 82801CAM/ICH3 Intel ethernet controller (the driver locks up under moderate to heavy network load, and requires the interface to be shut down and restarted to clear the problem). I believe that a development version is available that fixes the problem, but in the meantime, Intel's driver does the job well.

There's some evidence in the boot logs that suggests that the Sony BIOS is lying about IRQ's, but I haven't yet found any evidence that it's causing breakage in practice (all three usb sockets appear to work fine, which is what I'd expect to break, given the reported errors).

I'm currently using a kernel compiled with ACPI rather than APM support. ACPI works, but support is very minimal. In particular, the only practical functionality appears to be that the power button will trigger a clean shutdown. I haven't yet had time to pursue ACPI to see if this can be improved on ...

USB works fine.

I haven't (yet) investigated whether it copes with driving an external monitor and (if so) whether this can be toggled using the Fn-F7 key combination.

The sonypi driver (for jog dial support, among other things) appears to work ok. For details of how to use it, look in Documentation/sonypi.txt in the kernel source tree.