Re: [discuss] Re: 2.6.19-rc4: known unfixed regressions (v3)

[Linus Torvalds]

On Wed, 8 Nov 2006, Adrian Bunk wrote:
> > 
> > Anyway, I do not consider this a regression. MMCONFIG has _never_ worked 
> > reliably. It has always been a case of "we can make it work on some 
> > machines by making it break on others".
> 
> It is a serious regression:
> 
> The problem is that with the default CONFIG_PCI_GOANY, MMCONFIG is the 
> _first_ method tried.

No. That was a bug at some point, but it's not that way now. See

	pci_access_init(void)

which checks the pci_direct_probe() first, and only _then_ calls 
pci_mmcfg_init(). And pci_mmcfg_init() will refuse to even use MMCONFIG 
unless either the direct probe failed _or_ the MMCONFIG area is marked 
entirely reserved in the e820 tables. Exactly because MMCONFIG generally 
doesn't _work_.

Now, if that is indeed broken, then yes, we need to fix it. 

Jeff - when you enable "direct PCI access", what is the printout? You 
should get

	PCI: Using configuration type 1

and the kernel should never have used MMCONFIG if the area wasn't marked 
as reserved in e820..

		Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/