Re: M68k tree updates

[Maxim Kuvyrkov]

Geert Uytterhoeven wrote:
> On Mon, Nov 30, 2009 at 14:43, Maxim Kuvyrkov <maxim@xxxxxxxxxxxxxxxx> wrote:
> > Geert Uytterhoeven wrote:
> > > It took me a while, but finally I managed to update
> > > http://git.kernel.org/?p=linux/kernel/git/geert/linux-m68k.git again.
> > ...
> > > If anything is missing, please tell me. Thx!
> > How about the NPTL patch (0001-Add-NPTL-support-for-m68k.patch,
> > http://marc.info/?l=linux-m68k&m=125447760017098&w=2)?
>
> As you're adding syscalls (which are shared with m68knommu), I'll have
> to ask whether
> you will make it work on nommu, too?

Yes, that's quite straight-forward so I'll try to post the patch tomorrow.

> Furthermore, when building for Sun3:
>
> arch/m68k/kernel/sys_m68k.c:706: warning: passing argument 1 of
> 'kunmap' from incompatible pointer type
> arch/m68k/kernel/sys_m68k.c:714: warning: passing argument 1 of
> 'kunmap' from incompatible pointer type
>
> Which looks like a problem with the Sun 3 version of pte_unmap().

Right.  Sun3's implementation assumes that one can cast between (pte_t 
*) and pgtable_t interchangeably; fwiw, I don't know if this is the case.

> I'm gonna add it to the m68k tree though, unless someone complains.

Thanks!

> > Also, this trivial fix (0004-Fix-m68k-s-asm-swab.h-for-ColdFire.patch,
> > http://marc.info/?l=linux-m68k&m=125447494213740&w=2) can go in.
>
> I want to run this one by the nommu people as well...

To illustrate the problem more clearly, here is the code that the patch 
fixes:

#if defined (__mcfisaaplus__) || defined (__mcfisac__)
static inline __attribute_const__ __u32 __arch_swab32(__u32 val)
{
	__asm__("byterev %0" : "=d" (val) : "0" (val));
	return val;
}

#define __arch_swab32 __arch_swab32
#elif !defined(__uClinux__)

static inline __attribute_const__ __u32 __arch_swab32(__u32 val)
{
	__asm__("rolw #8,%0; swap %0; rolw #8,%0" : "=d" (val) : "0" (val));
	return val;
}
#define __arch_swab32 __arch_swab32
#endif

As far as I know, all non-ColdFire CPUs support rolw instruction, so 
enabling it for uClinux should not be a problem.  Regarding the ColdFire 
 CPUs, none of the CF ISAs include rolw, so the patch fixes the case of 
building the kernel for MMU Linux ISA_B CPU, e.g., mcf547x/8x family.

--
Maxim Kuvyrkov
CodeSourcery
maxim@xxxxxxxxxxxxxxxx
(650) 331-3385 x724
--
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/