[PATCH v2 7/7] arm64/sve: Don't zero non-FPSIMD register state on syscall by default

Mark Brown broonie at kernel.org
Wed Jul 20 07:31:34 PDT 2022


On Wed, Jul 20, 2022 at 10:29:56AM +0100, Marc Zyngier wrote:
> Mark Brown <broonie at kernel.org> wrote:

> > it.  Marc Zyngier has previously noted publicly the current behaviour
> > being a consideration in the context of discusion of optimisation ideas
> > like this one, I was a bit surprised that he commented on an earlier
> > patch in the series but not this one.

> Just because I don't repeat myself doesn't mean I changed my mind. I
> still don't buy the ABI change, as there is already a large(-ish) body
> of SW that assumes the current semantics.

Your previous comments (or at least the ones that I saw) weren't *quite*
that definitive.

I'm unconvinced about us actually having a large body of software at
this point, it's only potentially things that directly do syscalls in
the middle of SVE using functions and never get run with qemu user mode
- that could definitely be a thing, but let's not overstate it.
-------------- 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/20220720/987e3f2f/attachment.sig>


More information about the linux-arm-kernel mailing list