RE: [ACPI] Re: Radeon PM Support

[Pallipadi, V. <ven...@in...>]

>-----Original Message-----
>From: acp...@li...=20
>[mailto:acp...@li...] On Behalf Of=20
>Pavel Machek
>Sent: Thursday, August 19, 2004 4:17 AM
>To: Stefan Seyfried
>Cc: acp...@li...
>Subject: Re: [ACPI] Re: Radeon PM Support
>
>Hi!
>
>> > > Yes. If this patch works for tohers, then we can move it into
>> > > pci_default_resume during the cleanup.
>> >=20
>> >=20
>> > I'd rather not do that. Some cards are already posted after
>> > S3, some even preserve state over S3. It is okay to do VGAPOST
>> > for radeon and any card that needs it, but I'd rather not=20
>do it generically.
>>=20
>> what about "acpi_sleep=3Dvgapost"?
>
>Could work, too.
>
>BTW I tried vgapost here, and it did not work on 2 radeon machines I
>have here. Could you possibly do some testing on bigger number of
>machines to find one where it actually works?
>								Pavel
>--=20

I guess we need to do something like this "acpi_sleep=3Dvgapost",=20
rather than doing a generic vgapost for all display devices.=20
As the generic solution seems to have broken few systems=20
which were working before.

I have two systems with almost identical video setup, with=20
Radeon 9000M AGP adapter. vgapost works on one system and=20
fails on another. The failing system never returns=20
from real mode vgapost call. I also found that=20
the VBIOS is not same on these two systems. Probably=20
different versions of the firmware. I am digging further=20
to find out why it works on one Radeon 9000M and not=20
on another one.=20

-Venki