> On 11 Jul 2025, at 16:48, Richard Sandiford <richard.sandif...@arm.com> wrote: > > Kyrylo Tkachov <ktkac...@nvidia.com> writes: >>> On 10 Jul 2025, at 11:12, Kyrylo Tkachov <ktkac...@nvidia.com> wrote: >>> >>> >>> >>>> On 10 Jul 2025, at 10:40, Richard Sandiford <richard.sandif...@arm.com> >>>> wrote: >>>> >>>> Kyrylo Tkachov <ktkac...@nvidia.com> writes: >>>>> Hi all, >>>>> >>>>> While the SVE2 NBSL instruction accepts MOVPRFX to add more flexibility >>>>> due to its tied operands, the destination of the movprfx cannot be also >>>>> a source operand. But the offending pattern in aarch64-sve2.md tries >>>>> to do exactly that for the "=?&w,w,w" alternative and gas warns for the >>>>> attached testcase. >>>>> >>>>> This patch just removes that alternative causing RA to emit a normal extra >>>>> move. >>>>> So for the testcase in the patch we now generate: >>>>> nor_z: >>>>> nbsl z1.d, z1.d, z2.d, z1.d >>>>> mov z0.d, z1.d >>>>> ret >>>>> >>>>> instead of the previous: >>>>> nor_z: >>>>> movprfx z0, z1 >>>>> nbsl z0.d, z0.d, z2.d, z0.d >>>>> ret >>>>> >>>>> which generated a gas warning. >>>> >>>> Shouldn't we instead change it to: >>>> >>>> [ ?&w , w , w ; yes ] movprfx\t%0, %1\;nbsl\t%0.d, >>>> %0.d, %2.d, %1.d >>>> >>>> ? The "&" ensures that %1 is still valid in the NBSL. >>>> >>>> (That's OK if it works.) >>> >>> Yes, that seems to work, thanks. >>> I’ll push this version after some more testing. >>> >> >> Shall I backport this for GCC 15.2 as well? >> The test case uses C operators which were enabled in GCC 15, though I >> suppose one could construct a pure ACLE intrinsics testcase too. > > Sounds good to me. It's fixing wrong code, even if the gas warning > makes it somewhat noisy wrong code. >
Looks like there’s a simple merge conflicts due to trunk also having http://gcc.gnu.org/g:f260146bc05f6fba7b2a67a62063c770588b769d Author: Richard Earnshaw <rearn...@arm.com> Date: Mon Apr 14 16:41:16 2025 +0100 aarch64: Fix up commutative and early-clobber markers on compact inns I’d like to backport that commit as well as it looks like a low-risk cleanup. Both commits bootstrap and test cleanly on the branch. Ok? Thanks, Kyrill > Thanks, > Richard
