| Message ID | ZcdNpLWC4B5JMuqM@tucnak |
|---|---|
| State | New |
| Headers | show |
| Series | [RFC] i386: Enable _BitInt support on ia32 | expand |
On Sat, Feb 10, 2024 at 11:19:16AM +0100, Jakub Jelinek wrote: > And bitint-24.c case seems to be miscompilation of __floatbitintxf > when built with -O2 -m32, it works fine when compiled with -O0 -m32. > Will address that next week. Ugh, -O2 -m32 -march=i686 works fine, -O2 -m32 -march=i686 -mtune=generic misbehaves, -O2 -m32 -march=i686 -mtune=generic -fno-ivopts works fine. Will not be fun to debug, the function is large. Jakub
--- gcc/config/i386/i386.cc.jj 2024-02-09 11:02:15.193830702 +0100 +++ gcc/config/i386/i386.cc 2024-02-09 16:30:28.568240299 +0100 @@ -25757,13 +25757,11 @@ ix86_get_excess_precision (enum excess_p bool ix86_bitint_type_info (int n, struct bitint_info *info) { - if (!TARGET_64BIT) - return false; if (n <= 8) info->limb_mode = QImode; else if (n <= 16) info->limb_mode = HImode; - else if (n <= 32) + else if (n <= 32 || (!TARGET_64BIT && n > 64)) info->limb_mode = SImode; else info->limb_mode = DImode;
Hi! Given the https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113837#c9 comment, the following patch just attempts to implement what I think is best for ia32. Compared to https://gitlab.com/x86-psABIs/i386-ABI/-/issues/5 , like that patch for _BitInt(64) or smaller it uses the smallest containing {,un}signed {char,short,int,long long} for passing/returning and layout of variables including in structures for alignment/size, with any extra bits unspecified. Unlike the above proposal, for larger _BitInt (i.e. _BitInt(65)+), it uses passing/returning/layout/alignment of structure containing minimum needed number of 32-bit limbs, again with the extra bits unspecified. This is because most operations (except copy or bitwise ops) on _BitInts aren't really vectorizable and will be under the hood implemented in loops over 32-bit limbs anyway (using 64-bit limbs under the hood would mean often using library implementation for the basic operations) and because ia32 doesn't align even long long/double in structures to 64-bit I think it is better to just use 32-bit alignment for that. And I don't see a reason to waste 32-bit bits say for _BitInt(224) or _BitInt(288) on ia32. So, effectively it is like the x86-64 _BitInt ABI with everything divided by 2, the only exception is that in x86-64 psABI _BitInt(128) is said to be already a structure of 2 limbs, which happens to be passed mostly the same as __int128 (except for alignment). Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? There are still 2 regressions: +FAIL: gcc.dg/pr113693.c at line 13 (test for warnings, line ) +FAIL: gcc.dg/torture/bitint-24.c -O0 execution test +FAIL: gcc.dg/torture/bitint-24.c -O2 execution test For pr113693.c I think we'll need to do something with the test, either make it x86_64 specific, add say -msse2 option for ia32, move to gcc.dg/vect/ And bitint-24.c case seems to be miscompilation of __floatbitintxf when built with -O2 -m32, it works fine when compiled with -O0 -m32. Will address that next week. 2024-02-10 Jakub Jelinek <jakub@redhat.com> * config/i386/i386.cc (ix86_bitint_type_info): Add support for !TARGET_64BIT. Jakub