On Thu, Dec 05, 2002 at 11:40:40PM +0900, Takahiro Kambe wrote:
> In message <20021205030054.GA1272@krishna>
> 	on Thu, 5 Dec 2002 12:00:54 +0900,
> 	Bang Jun-Young <junyoung@netbsd.org> wrote:
> > On Tue, Nov 26, 2002 at 12:15:45AM +0900, Takahiro Kambe wrote:
> > > cpu0 at mainbus0: (uniprocessor)
> > > cpu0: IDT Pentium Pro compatible (686-class), 533.39 MHz
> > > cpu0: features 803035<FPU,DE,TSC,MSR,MTRR>
> > > cpu0: features 803035<PGE,MMX>
> > 
> > Could you test with the following patch?
> Here's result.  It seems OK about kind of CPU.
> 
> NetBSD 1.6K (NS) #102: Thu Dec  5 23:34:23 JST 2002
>     taca@ns.back-street.net:/sys/arch/i386/compile/NS
> total memory = 253 MB
> avail memory = 232 MB
> using 3271 buffers containing 13084 KB of memory
> BIOS32 rev. 0 found at 0xfb370
> PCI BIOS rev. 2.1 found at 0xfb3a0
> pcibios: config mechanism [1][x], special cycles [1][x], last bus 1
> PCI IRQ Routing Table rev. 1.0 found at 0xfdc70, size 112 bytes (5 entries)
> PCI Interrupt Router at 000:00:0
> PCI Exclusive IRQs: 10 11
> mainbus0 (root)
> cpu0 at mainbus0: (uniprocessor)
> cpu0: VIA C3 Samuel 2/Ezra (686-class), 533.38 MHz
> cpu0: features 803035<FPU,DE,TSC,MSR,MTRR>
> cpu0: features 803035<PGE,MMX>

Seems okay.

> 
> What's differ about "3DNOW" cpu feature?
> 
> >> cpu0: VIA C3 Samuel 2 ("CentaurHauls" 686-class) 533 MHz
> >> cpu0: FPU,DE,TSC,MSR,MTRR,PGE,MMX,3DNOW
> (OpenBSD 3.1)

It's the limitation of our cpuid routine which is Intel biased. My
Athlon processor has the same problem. I think it should be addressed
separately from this PR, since it's beyond the scope of the PR and it
may take some time to fix it.

Can I close the PR & could you file a new PR for incomplete cpu feature
bits with non-Intel processors?

Jun-Young

-- 
Bang Jun-Young <junyoung@netbsd.org>