[PATCH 05/10] powerpc/booke64: Use SPRG7 for VDSO

mihai.caraman at freescale.com mihai.caraman at freescale.com
Tue Mar 18 01:25:04 EST 2014

> -----Original Message-----
> From: Wood Scott-B07421
> Sent: Friday, March 14, 2014 2:01 AM
> To: Benjamin Herrenschmidt
> Cc: linuxppc-dev at lists.ozlabs.org; Kumar Gala; Tiejun Chen; Wood Scott-
> B07421; Caraman Mihai Claudiu-B02008; Anton Blanchard; Paul Mackerras;
> kvm-ppc at vger.kernel.org
> Subject: [PATCH 05/10] powerpc/booke64: Use SPRG7 for VDSO
> Previously SPRG3 was marked for use by both VDSO and critical
> interrupts (though critical interrupts were not fully implemented).
> In commit 8b64a9dfb091f1eca8b7e58da82f1e7d1d5fe0ad ("powerpc/booke64:
> Use SPRG0/3 scratch for bolted TLB miss & crit int"), Mihai Caraman
> made an attempt to resolve this conflict by restoring the VDSO value
> early in the critical interrupt, but this has some issues:

>  - It forces critical exceptions to be a special case handled
>    differently from even machine check and debug level exceptions.

Yes, this was ugly.

> Since we cannot use SPRG4-7 for scratch without corrupting the state of
> a KVM guest, move VDSO to SPRG7 on book3e.

At that time I thought that information exposed through SPRN_USPRG3 is
part of the ABI which can't be changed, and this lead to complicated
gymnastic. But since VSDO's __kernel_getcpu() function hides SPRN_USPRGx
access, I think your unified approach is the right way.


More information about the Linuxppc-dev mailing list