From: "H.J. Lu" <hjl.tools@gmail.com>
To: Florian Weimer <fweimer@redhat.com>
Cc: GNU C Library <libc-alpha@sourceware.org>
Subject: Re: [PATCH v2 5/6] x86: Add x86-64-vN check to early startup
Date: Fri, 14 Jan 2022 10:40:29 -0800 [thread overview]
Message-ID: <CAMe9rOodDWV=u8xns6yPuYkfnYuhhvZwnmsCJyeR-LQ3nZh1Cg@mail.gmail.com> (raw)
In-Reply-To: <703f911ba1d0d946e4978830e5d3161674bfc77c.1642179009.git.fweimer@redhat.com>
On Fri, Jan 14, 2022 at 8:53 AM Florian Weimer <fweimer@redhat.com> wrote:
>
> This ISA level covers the glibc build itself. <dl-hwcap-check.h>
> cannot be used because this check (by design) happens before
> DL_PLATFORM_INIT and the x86 CPU flags initialization.
> ---
> v2: Reflect renamed Makefile variable.
> sysdeps/x86/Makefile | 1 +
> sysdeps/x86/dl-get-cpu-features.c | 31 ++++++++++++++++++++++++++++++-
> 2 files changed, 31 insertions(+), 1 deletion(-)
>
> diff --git a/sysdeps/x86/Makefile b/sysdeps/x86/Makefile
> index 402986ff68..6cf708335c 100644
> --- a/sysdeps/x86/Makefile
> +++ b/sysdeps/x86/Makefile
> @@ -7,6 +7,7 @@ sysdep_routines += get-cpuid-feature-leaf
> sysdep-dl-routines += dl-get-cpu-features
> sysdep_headers += sys/platform/x86.h bits/platform/x86.h
>
> +CFLAGS-dl-get-cpu-features.os += $(rtld-early-cflags)
> CFLAGS-get-cpuid-feature-leaf.o += $(no-stack-protector)
>
> tests += tst-get-cpu-features tst-get-cpu-features-static \
> diff --git a/sysdeps/x86/dl-get-cpu-features.c b/sysdeps/x86/dl-get-cpu-features.c
> index 6339c9df4e..4ec0e5d2af 100644
> --- a/sysdeps/x86/dl-get-cpu-features.c
> +++ b/sysdeps/x86/dl-get-cpu-features.c
> @@ -20,6 +20,7 @@
>
> #ifdef SHARED
> # include <cpu-features.c>
> +# include <gcc-macros.h>
>
> /* NB: Normally, DL_PLATFORM_INIT calls init_cpu_features to initialize
> CPU features in dynamic executable. But when loading ld.so inside of
> @@ -36,7 +37,35 @@ _dl_x86_init_cpu_features (void)
> {
> struct cpu_features *cpu_features = __get_cpu_features ();
> if (cpu_features->basic.kind == arch_kind_unknown)
> - init_cpu_features (cpu_features);
> + {
> + init_cpu_features (cpu_features);
> +
> +# if IS_IN (rtld)
> + /* See isa-level.c. */
> +# if defined GCCMACRO__GCC_HAVE_SYNC_COMPARE_AND_SWAP_16 \
> + && defined HAVE_X86_LAHF_SAHF && defined GCCMACRO__POPCNT__ \
> + && defined GCCMACRO__SSE3__ && defined GCCMACRO__SSSE3__ \
> + && defined GCCMACRO__SSE4_1__ && defined GCCMACRO__SSE4_2__
> + if (!(cpu_features->isa_1 & GNU_PROPERTY_X86_ISA_1_V2))
> + _dl_fatal_printf ("\
> +Fatal glibc error: CPU does not support x86-64-v%d\n", 2);
> +# if defined GCCMACRO__AVX__ && defined GCCMACRO__AVX2__ \
> + && defined GCCMACRO__F16C__ && defined GCCMACRO__FMA__ \
> + && defined GCCMACRO__LZCNT__ && defined HAVE_X86_MOVBE
> + if (!(cpu_features->isa_1 & GNU_PROPERTY_X86_ISA_1_V3))
> + _dl_fatal_printf ("\
> +Fatal glibc error: CPU does not support x86-64-v%d\n", 3);
> +# if defined GCCMACRO__AVX512F__ && defined GCCMACRO__AVX512BW__ \
> + && defined GCCMACRO__AVX512CD__ && defined GCCMACRO__AVX512DQ__ \
> + && defined GCCMACRO__AVX512VL__
> + if (!(cpu_features->isa_1 & GNU_PROPERTY_X86_ISA_1_V4))
> + _dl_fatal_printf ("\
> +Fatal glibc error: CPU does not support x86-64-v%d\n", 4);
> +# endif /* ISA level 4 */
> +# endif /* ISA level 3 */
> +# endif /* ISA level 2 */
> +# endif /* IS_IN (rtld) */
> + }
> }
>
> __ifunc (__x86_cpu_features, __x86_cpu_features, NULL, void,
> --
> 2.34.1
>
>
LGTM.
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Thanks.
--
H.J.
next prev parent reply other threads:[~2022-01-14 18:41 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-01-14 16:51 [PATCH v2 0/6] Reliable CPU compatibility diagnostics in ld.so Florian Weimer
2022-01-14 16:52 ` [PATCH v2 1/6] elf/Makefile: Reflow and sort most variable assignments Florian Weimer
2022-01-14 18:39 ` H.J. Lu
2022-01-14 16:52 ` [PATCH v2 2/6] elf: Split dl-printf.c from dl-misc.c Florian Weimer
2022-01-14 18:40 ` H.J. Lu
2022-01-14 16:52 ` [PATCH v2 3/6] Add --with-rtld-early-cflags configure option Florian Weimer
2022-01-14 18:38 ` H.J. Lu
2022-01-14 16:53 ` [PATCH v2 4/6] powerpc64le: Use <gcc-macros.h> in early HWCAP check Florian Weimer
2022-01-14 18:51 ` Carlos O'Donell
2022-01-14 16:53 ` [PATCH v2 5/6] x86: Add x86-64-vN check to early startup Florian Weimer
2022-01-14 18:40 ` H.J. Lu [this message]
2022-01-14 16:53 ` [PATCH 6/6] s390x: Use <gcc-macros.h> in early HWCAP check Florian Weimer
2022-01-14 18:52 ` Carlos O'Donell
2022-01-14 17:42 ` [PATCH v2 0/6] Reliable CPU compatibility diagnostics in ld.so Joseph Myers
2022-01-14 17:59 ` Florian Weimer
2022-01-14 18:51 ` Carlos O'Donell
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAMe9rOodDWV=u8xns6yPuYkfnYuhhvZwnmsCJyeR-LQ3nZh1Cg@mail.gmail.com' \
--to=hjl.tools@gmail.com \
--cc=fweimer@redhat.com \
--cc=libc-alpha@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).