Re: [PATCH v2] powerpc: slightly improve cache helpers

From: Christophe Leroy
Date: Tue Jul 09 2019 - 01:04:48 EST




Le 08/07/2019 à 21:14, Nathan Chancellor a écrit :
On Mon, Jul 08, 2019 at 11:19:30AM +1000, Michael Ellerman wrote:
On Fri, 2019-05-10 at 09:24:48 UTC, Christophe Leroy wrote:
Cache instructions (dcbz, dcbi, dcbf and dcbst) take two registers
that are summed to obtain the target address. Using 'Z' constraint
and '%y0' argument gives GCC the opportunity to use both registers
instead of only one with the second being forced to 0.

Suggested-by: Segher Boessenkool <segher@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxx>

Applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/6c5875843b87c3adea2beade9d1b8b3d4523900a

cheers

This patch causes a regression with clang:

Is that a Clang bug ?

Do you have a disassembly of the code both with and without this patch in order to compare ?

Segher, any idea ?

Christophe


https://travis-ci.com/ClangBuiltLinux/continuous-integration/jobs/213944668

I've attached my local bisect/build log.

Cheers,
Nathan