On 10/9/24 10:52 AM, Palmer Dabbelt wrote:
On Tue, 08 Oct 2024 16:43:13 PDT (-0700), jeffreya...@gmail.com wrote:
On 10/7/24 11:33 PM, Tsung Chun Lin wrote:
That M implies Zmmul.
gcc/ChangeLog:
* common/config/riscv/riscv-common.cc: M implies Zmmul.
THanks. I've pushed this to the trunk.
jeff
ps. Quite a discussion on this topic in the zmmul public discussion
from a couple years ago. But it looks like the consensus was that M
should imply Zmmul.
Wacky timing: I just sat down to try and reply to the "how do C and Zca
relate" thread in binutils, only to find this one instead. I think
we've now got the same issue that Jan is pointing out in binutils:
basically +M and -M aren't inverses any more (it would leave on Zmmul).
Not 100% sure there.
FWIW that M/Zmmul thread basically made me give up so I don't really
care that much either way, IMO anyone depending on these fine-grained
extension things is in for a pile of brokenness. So as long as we
document what we do it's fine with me.
Yea, when I read it my thought was this was a waste of time, but that I
didn't care enough to fight it.
Unfortunately it looks like the patches didn't actually get tested and
is causing a few hundred regressions. Once I confirm I'll officially
revert and ask Tsung Chun to post an update after fixing the testing
failures.
Jeff
