In other news, I noticed recently that the IO-APIC on my MS-6163
BX-Master was being disabled at boot by the kernel, due to a recently
introduced blacklist:
{ apm_kills_local_apic, "Microstar 6163",
{ MATCH(DMI_BOARD_VENDOR, "MICRO-STAR INTERNATIONAL CO., LTD"),
MATCH(DMI_BOARD_NAME, "MS-6163"),
NO_MATCH, NO_MATCH } },
Consulting mailing list archives indicated that there is some sort of
problem with the IO-APIC on the MS-6163 Pro and APM events. However,
this seems a rather clumsy fix to the problem, since it disabled the
IO-APIC on _all_ MS-6163 boards rather than just the Pro, and also
regardless of whether APM support is even enabled (I don't enable it and
don't use it at all).
The DMI string for my board is:
Board Information Block
Vendor: MICRO-STAR INTERNATIONAL CO., LTD
Product: MS-6163 (i440BX)
Version: 3.X
Serial Number:
It would seem that if a closer match were performed, using the version
number of the board (3.X in my case, likely 2.X in the case of the
broken Pro), it would be a better idea. Perhaps another alternative
solution would be to only disable the IO-APIC if CONFIG_APM is defined.
(?)
Thoughts?
--
Ryan Underwood, <nemesis at icequake.net>, icq=10317253
On Sad, 2003-06-14 at 02:46, Ryan Underwood wrote:
> It would seem that if a closer match were performed, using the version
> number of the board (3.X in my case, likely 2.X in the case of the
> broken Pro), it would be a better idea. Perhaps another alternative
> solution would be to only disable the IO-APIC if CONFIG_APM is defined.
> (?)
Its more likely to be about BIOS version than about the board itself
On Sat, Jun 14, 2003 at 09:19:14AM +0100, Alan Cox wrote:
> On Sad, 2003-06-14 at 02:46, Ryan Underwood wrote:
> > It would seem that if a closer match were performed, using the version
> > number of the board (3.X in my case, likely 2.X in the case of the
> > broken Pro), it would be a better idea. Perhaps another alternative
> > solution would be to only disable the IO-APIC if CONFIG_APM is defined.
> > (?)
>
> Its more likely to be about BIOS version than about the board itself
It would have been nice if the original submitter of the patch would
have specified what bios they were using. IO-APIC support on my board
seemed to be working nicely until eventually it got disabled. :)
I am using the 5.6 bios from their site for the BX Master -- the 6163
Pro (v2) has a totally different BIOS versions and I'm not sure if a
blanket black-list for this board would be the best idea.
Thanks,
--
Ryan Underwood, <nemesis at icequake.net>, icq=10317253
On Fri, 13 Jun 2003 20:46:46 -0500, Ryan Underwood wrote:
>In other news, I noticed recently that the IO-APIC on my MS-6163
>BX-Master was being disabled at boot by the kernel, due to a recently
>introduced blacklist:
>
>{ apm_kills_local_apic, "Microstar 6163",
> { MATCH(DMI_BOARD_VENDOR, "MICRO-STAR INTERNATIONAL CO., LTD"),
> MATCH(DMI_BOARD_NAME, "MS-6163"),
> NO_MATCH, NO_MATCH } },
>
>Consulting mailing list archives indicated that there is some sort of
>problem with the IO-APIC on the MS-6163 Pro and APM events. However,
>this seems a rather clumsy fix to the problem, since it disabled the
>IO-APIC on _all_ MS-6163 boards rather than just the Pro, and also
>regardless of whether APM support is even enabled (I don't enable it and
>don't use it at all).
I added this blacklist rule last spring, but I'm also about to
submit a patch to Linus and Marcelo to remove it and the AL440LX
blacklist rule. You can read the full explanation then; suffice
it to say that using APM's DISPLAY_BLANK with certain graphics
cards causes hangs if the local APIC is enabled. You can just
remove the MS-6163 Pro blacklist rule for now.
/Mikael