From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 1778) id A5AEA3858401; Mon, 3 Oct 2022 22:02:26 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org A5AEA3858401 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1664834546; bh=mUw0ZhkYt6k6lGHqJzmei4feI8NXESgLTjmMRoS2KeQ=; h=From:To:Subject:Date:From; b=lsg7RSbmRC5cSkBc2mauEgm5E3qT5ROzD2tXIh0S5O2les0gWSc/Wi0p2H2N38gut e5AXR/5AgyGyp6Chubh0qSYz7PrnrV0T2iAAw1wqYFk3dq9OwJsZ13LGHItNC281Qc qOa47wQ/8xeZOt2GvpZqy3LM0HYcXTZha/8R6iko= Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit From: Aurelien Jarno To: glibc-cvs@sourceware.org Subject: [glibc/release/2.34/master] x86-64: Require BMI2 for AVX2 wcs(n)cmp implementations X-Act-Checkin: glibc X-Git-Author: Aurelien Jarno X-Git-Refname: refs/heads/release/2.34/master X-Git-Oldrev: b9cbb8dd48b545f3b36b5d411481dc0bd118ee94 X-Git-Newrev: 67e863742d98c990b3d3b814b80042c0fa0d50a5 Message-Id: <20221003220226.A5AEA3858401@sourceware.org> Date: Mon, 3 Oct 2022 22:02:26 +0000 (GMT) List-Id: https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=67e863742d98c990b3d3b814b80042c0fa0d50a5 commit 67e863742d98c990b3d3b814b80042c0fa0d50a5 Author: Aurelien Jarno Date: Mon Oct 3 23:16:46 2022 +0200 x86-64: Require BMI2 for AVX2 wcs(n)cmp implementations The AVX2 wcs(n)cmp implementations use the 'bzhi' instruction, which belongs to the BMI2 CPU feature. NB: It also uses the 'tzcnt' BMI1 instruction, but it is executed as BSF as BSF if the CPU doesn't support TZCNT, and produces the same result for non-zero input. Partially fixes: b77b06e0e296 ("x86: Optimize strcmp-avx2.S") Partially resolves: BZ #29611 Reviewed-by: Noah Goldstein (cherry picked from commit f31a5a884ed84bd37032729d4d1eb9d06c9f3c29) Diff: --- sysdeps/x86_64/multiarch/ifunc-impl-list.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/sysdeps/x86_64/multiarch/ifunc-impl-list.c b/sysdeps/x86_64/multiarch/ifunc-impl-list.c index 34d5f6efe5..e76a991dc6 100644 --- a/sysdeps/x86_64/multiarch/ifunc-impl-list.c +++ b/sysdeps/x86_64/multiarch/ifunc-impl-list.c @@ -693,10 +693,12 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array, /* Support sysdeps/x86_64/multiarch/wcscmp.c. */ IFUNC_IMPL (i, name, wcscmp, IFUNC_IMPL_ADD (array, i, wcscmp, - CPU_FEATURE_USABLE (AVX2), + (CPU_FEATURE_USABLE (AVX2) + && CPU_FEATURE_USABLE (BMI2)), __wcscmp_avx2) IFUNC_IMPL_ADD (array, i, wcscmp, (CPU_FEATURE_USABLE (AVX2) + && CPU_FEATURE_USABLE (BMI2) && CPU_FEATURE_USABLE (RTM)), __wcscmp_avx2_rtm) IFUNC_IMPL_ADD (array, i, wcscmp, @@ -709,10 +711,12 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array, /* Support sysdeps/x86_64/multiarch/wcsncmp.c. */ IFUNC_IMPL (i, name, wcsncmp, IFUNC_IMPL_ADD (array, i, wcsncmp, - CPU_FEATURE_USABLE (AVX2), + (CPU_FEATURE_USABLE (AVX2) + && CPU_FEATURE_USABLE (BMI2)), __wcsncmp_avx2) IFUNC_IMPL_ADD (array, i, wcsncmp, (CPU_FEATURE_USABLE (AVX2) + && CPU_FEATURE_USABLE (BMI2) && CPU_FEATURE_USABLE (RTM)), __wcsncmp_avx2_rtm) IFUNC_IMPL_ADD (array, i, wcsncmp,