[PATCH v3 3/3] kselftest/arm64: Add lsfe to the hwcaps test
Mark Brown
broonie at kernel.org
Tue Sep 16 14:54:29 PDT 2025
On Tue, Sep 16, 2025 at 10:16:11PM +0100, Will Deacon wrote:
> On Mon, Aug 18, 2025 at 08:21:20PM +0100, Mark Brown wrote:
> > + /* LDFADD H0, H0, [X0] */
> > + asm volatile(".inst 0x7c200000"
> > + : "+r" (memp)
> Doesn't this corrupt H0 without the compiler knowing? It's probably
> easier to use STFADD.
Yeah, that's more correct and easier than specifying constraints. In
practice it should be safe as the compiler is unlikely to use FP in the
instructions it generates and it's a caller saved register. The program
is in general not careful with constraints.
> > + :
> > + : "cc", "memory");
> Why do you need the "cc" clobber?
It's overkill.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20250916/4bee9fbe/attachment-0001.sig>
More information about the linux-arm-kernel
mailing list