From: Fabian Vogt <fvogt@suse.de>
To: libc-alpha@sourceware.org
Cc: Noah Goldstein <goldstein.w.n@gmail.com>,
"H.J. Lu" <hjl.tools@gmail.com>,
aurelien@aurel32.net
Subject: Re: [PATCH] x86: include OSXSAVE in x86-64-v3 level
Date: Wed, 14 Dec 2022 16:23:53 +0100 [thread overview]
Message-ID: <4794883.31r3eYUQgx@linux-e202.suse.de> (raw)
In-Reply-To: <CAMe9rOqHQ7kf7CRiPKrYcGgDPv4BrtG-Wau5ojyonCi-8bEVQQ@mail.gmail.com>
Am Freitag, 9. Dezember 2022, 23:52:37 CET schrieb H.J. Lu:
> On Fri, Dec 9, 2022 at 1:09 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> >
> > On Fri, Dec 9, 2022 at 12:26 PM H.J. Lu via Libc-alpha
> > <libc-alpha@sourceware.org> wrote:
> > >
> > > On Fri, Dec 9, 2022 at 2:36 AM Fabian Vogt <fvogt@suse.de> wrote:
> > > >
> > > > For some reason the initial x86-64-v3 detection code was missing checks for
> > > > BMI, BMI2 and OSXSAVE, which are all required for that level to be met.
> > > > BMI and BMI2 got added recently, but OSXSAVE was still missing. Add it.
> > > >
> > > > Signed-off-by: Fabian Vogt <fvogt@suse.de>
> > > > ---
> > > > sysdeps/x86/get-isa-level.h | 3 ++-
> > > > 1 file changed, 2 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/sysdeps/x86/get-isa-level.h b/sysdeps/x86/get-isa-level.h
> > > > index 5b4dd5f062..d62bf92cde 100644
> > > > --- a/sysdeps/x86/get-isa-level.h
> > > > +++ b/sysdeps/x86/get-isa-level.h
> > > > @@ -52,7 +52,8 @@ get_isa_level (const struct cpu_features *cpu_features)
> > > > && CPU_FEATURE_USABLE_P (cpu_features, F16C)
> > > > && CPU_FEATURE_USABLE_P (cpu_features, FMA)
> > > > && CPU_FEATURE_USABLE_P (cpu_features, LZCNT)
> > > > - && CPU_FEATURE_USABLE_P (cpu_features, MOVBE))
> > > > + && CPU_FEATURE_USABLE_P (cpu_features, MOVBE)
> > > > + && CPU_FEATURE_USABLE_P (cpu_features, OSXSAVE))
> > >
> > > If OSXSAVE isn't usable, all AVX/AVX512 features shouldn't be usable.
> > > Am I missing something?
> >
> > Seems like better to err on side of caution here.
>
> We will have serious issues in many places if we don't get it correctly.
> CPU_FEATURE_USABLE_P is set according to the states supported
> by OSXSAVE. Checking OSXSAVE isn't needed here.
I see, you explicitly unset some flags if OSXSAVE is not available.
In that case the patch is indeed a noop, but also won't hurt. What about
adding a comment instead?
> > Skipping the extra check seems prone to bugs like BZ #29611.
next prev parent reply other threads:[~2022-12-14 15:23 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-09 10:36 Fabian Vogt
2022-12-09 17:14 ` Noah Goldstein
2022-12-09 23:53 ` Florian Weimer
2022-12-09 20:25 ` H.J. Lu
2022-12-09 21:09 ` Noah Goldstein
2022-12-09 22:52 ` H.J. Lu
2022-12-14 15:23 ` Fabian Vogt [this message]
2022-12-14 16:11 ` H.J. Lu
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=4794883.31r3eYUQgx@linux-e202.suse.de \
--to=fvogt@suse.de \
--cc=aurelien@aurel32.net \
--cc=goldstein.w.n@gmail.com \
--cc=hjl.tools@gmail.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).