RE: FYI: Merge in progress...

[YhLu]

Why enable cache in amd_early_mtrr_init?

That causes CPU sleep 5s again.

After get rid of that, the new auto.c stage still cost more 1s than the code
two weeks ago.

Regards

YH

-----Original Message-----
From: YhLu 
Sent: Tuesday, October 19, 2004 6:30 PM
To: [e-mail]
Cc: Li-Ta Lo; 'Ronald G. Minnich'; 'LinuxBIOS'
Subject: RE: FYI: Merge in progress...

Suggest you to add two options
1. premem 64 bit support? PCI_MEM_PREF_64BIT_SUPPORT
2. pci mmio optimizing?  PCI_MEM_BASE_OPTIMIZATION

Regards

YH

-----Original Message-----
From: [e-mail] [mailto:[e-mail]] 
Sent: Tuesday, October 19, 2004 4:37 PM
To: YhLu
Cc: Li-Ta Lo; '[e-mail]'; 'Ronald G. Minnich'; 'LinuxBIOS'
Subject: Re: FYI: Merge in progress...

YhLu  writes:
> Find out the problem, should be typo error.
>             /* Now place the memory as high up as it will go */
>             mem2->base = resource_max(mem2);
>             mem1->limit = mem2->base - 1;
>             mem1->base = resource_max(mem2);
>
> ------>
>                 /* Now place the memory as high up as it will go */
>                 mem2->base = resource_max(mem2);
>                 mem1->limit = mem2->base - 1;
>                 mem1->base = resource_max(mem1);
>
> regards

Thanks.  That was a definite blind spot, on my side.

I have now committed the fix.

Eric

_______________________________________________
Linuxbios mailing list
[e-mail]
http://www.clustermatic.org/mailman/listinfo/linuxbios

FYI: Merge in progress...

YhLu  writes:
> Why enable cache in amd_early_mtrr_init?

Usually that is a speed up.  I think all I changed was using the C implementation.
Previously the assembly was called from Config.lb.
> That causes CPU sleep 5s again.

That is peculiar.
> After get rid of that, the new auto.c stage still cost more 1s than the code
> two weeks ago.

Hmm.

This is certainly worth fixing.  And it tends to explain why I have
not yet seen the 5s speed up.

Eric