kms reinit...

mobi phil mobi at
Thu Apr 15 01:30:22 CEST 2010

I mixed up a bit email streams with the other list...

The bug where VRAM didn't get released when a client exits has been
> fixed for quite a while now, so I'm a bit curious as to why you'd want
> to do this.

no other reason... I had the impression, as I have some strange behaviour as
reported in my email to the other list...

> It's theoretically possible for the kernel to force all objects to be
> deleted of course,
> but you can't really just pull the carpet out from underneath the
> applications like that.
Sometimes I would pull the carpet for slow apps :)...

what  was reporting in the email to the prev list is that directfb apps are
blocking (apparently after the patch for "~150 shifted pixels ". Here is the
end of the strace. Fd 9 is apparently /dev/fb0

fstat64(8, {st_mode=S_IFCHR|0660, st_rdev=makedev(29, 0), ...}) = 0
ioctl(8, FBIOPUT_CON2FBMAP, 0xbed0e4c0) = 0
ioctl(9, VIDIOC_G_COMP or VT_ACTIVATE, 0x8) = 0

mobi phil

being mobile, but including technology
-------------- next part --------------
An HTML attachment was scrubbed...

More information about the openmoko-kernel mailing list