M2v-mx_se asus : s3 sleeping broken since rev4625 ruik

Hello,

I did some testing on my Asus M2v-mx_se and I discover that s3 sleeping state in broken since revision 4625 from Rudolf Marek for the motherboard.
It enters in s3 sleeping state, the LED is blinking, but during the wakeup it simply reboots instead of jumping to the boot segment.

Does anyone encounter the same problem ?
I don't know if it's an ACPI problem or the coreboot v3 features added that causes the trouble.

Hope someone could help or upgrade coreboot for M2v-mx_se.

Thanks in advance.

____________________________________________________

Tables en fête, recettes parfaites, sorties en goguette tout est sur http://evenementiel.voila.fr/Noel

[Rudolf Marek]

Hi,

Yeah I know. I think it is broken since  Stephan changes to wakeup core and the 
memory manager for the "blocks" stuff. There is one patch for consolidating the 
integrated UMA stuff - this patch needs to be adjusted plus there must be some 
patches for the zero copy logic. I had no time  for this. I can try  to fix it 
tomorrow if I find some semi finished patch. But mine time is limited so it is 
difficult.

You mentioned the patch 4625 broke it. Can you try 4624 and/or 4627? I suspect 
this change is unrelated.

Rudolf

[e-mail] napsal(a):
> Hello,
>
> I did some testing on my Asus M2v-mx_se and I discover that s3 sleeping state in broken since revision 4625 from Rudolf Marek for the motherboard.
> It enters in s3 sleeping state, the LED is blinking, but during the wakeup it simply reboots instead of jumping to the boot segment.
>
> Does anyone encounter the same problem ?
> I don't know if it's an ACPI problem or the coreboot v3 features added that causes the trouble.
>
> Hope someone could help or upgrade coreboot for M2v-mx_se.
>
> Thanks in advance.
>
> Tables en fête, recettes parfaites, sorties en goguette tout est sur http://evenementiel.voila.fr/Noel

[Rudolf Marek]

Hi,

I think you can get it to work again if you simply comment out in lib/cbmem.c
line 186  //acpi_slp_type == 0;

and in acpi.c

a return on line 592 this will let the stuff running even when no resume area 
was defined/filled. The region is still marked reserved although I think you 
will need bit more fix to Kconfig.

See attached patch. Totally untested but I think it could work.

Rudolf

[Rudolf Marek]

Hi,

I can confirm the patch works fine. The board reserves the memory on 31MB as 
hole. So, the saving stuff can be left unimplemented for now.

[    0.000000]  BIOS-e820: 0000000000000000 - 000000000009f000 (usable)
[    0.000000]  BIOS-e820: 000000000009f000 - 00000000000a0000 (reserved)
[    0.000000]  BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved)
[    0.000000]  BIOS-e820: 0000000000100000 - 0000000001f00000 (usable)
[    0.000000]  BIOS-e820: 0000000001f00000 - 0000000002000000 (reserved)
[    0.000000]  BIOS-e820: 0000000002000000 - 0000000017fe7000 (usable)
[    0.000000]  BIOS-e820: 0000000017fe7000 - 0000000018000000 (reserved)
[    0.000000]  BIOS-e820: 00000000e0000000 - 00000000f0000000 (reserved)

High Tables Base is 17ff0000, but that little missing stuff 0000000017fe7000 - 
17ff0000 is from SeaBIOS I think.

I used classic buildtarget/abuild method beccause I think Kconfig has bit 
different configuration.

I'm glad that you have chosen this board! It was a lot of work to get it running ;)

Thanks,
Rudolf