From: "H.J. Lu" <hjl.tools@gmail.com>
To: Florian Weimer <fweimer@redhat.com>
Cc: "H.J. Lu via Libc-alpha" <libc-alpha@sourceware.org>
Subject: Re: [PATCH 1/4] x86: Initialize CPU info via IFUNC relocation [BZ 26203]
Date: Mon, 28 Sep 2020 06:48:20 -0700 [thread overview]
Message-ID: <CAMe9rOo_OUmsikKWj_pXiy-G+NJZGGJwzn2_YTczXPvd09FaRA@mail.gmail.com> (raw)
In-Reply-To: <87imby6obw.fsf@oldenburg2.str.redhat.com>
On Mon, Sep 28, 2020 at 6:08 AM Florian Weimer <fweimer@redhat.com> wrote:
>
> * H. J. Lu via Libc-alpha:
>
> > X86 CPU features in ld.so are initialized by init_cpu_features, which is
> > invoked by DL_PLATFORM_INIT from _dl_sysdep_start. But when ld.so is
> > loaded by static executable, DL_PLATFORM_INIT is never called. Also
> > x86 cache info in libc.o and libc.a is initialized by a constructor
> > which may be called too late. Since _rtld_global_ro in ld.so is
> > initialized by dynamic relocation, we can also initialize x86 CPU
> > features in _rtld_global_ro in ld.so and cache info in libc.so by
> > initializing dummy function pointers in ld.so and libc.so via IFUNC
> > relocation.
>
> _rtld_global_ro is *partially* initialized by relocation. Most of it is
> not initialized (see the need for rtld_active).
>
> Please make this a little bit clearer in the commit message.
Will do
> > diff --git a/sysdeps/i386/dl-machine.h b/sysdeps/i386/dl-machine.h
> > index 0f08079e48..5e22e795cc 100644
> > --- a/sysdeps/i386/dl-machine.h
> > +++ b/sysdeps/i386/dl-machine.h
> > @@ -25,7 +25,6 @@
> > #include <sysdep.h>
> > #include <tls.h>
> > #include <dl-tlsdesc.h>
> > -#include <cpu-features.c>
> >
> > /* Return nonzero iff ELF header is compatible with the running host. */
> > static inline int __attribute__ ((unused))
> > @@ -250,7 +249,7 @@ dl_platform_init (void)
> > #if IS_IN (rtld)
> > /* init_cpu_features has been called early from __libc_start_main in
> > static executable. */
> > - init_cpu_features (&GLRO(dl_x86_cpu_features));
> > + _dl_x86_init_cpu_features ();
>
> Is the commented outdated?
I will fix it.
> > diff --git a/sysdeps/x86/cacheinfo.c b/sysdeps/x86/cacheinfo.c
> > index 217c21c34f..7a325ab70e 100644
> > --- a/sysdeps/x86/cacheinfo.c
> > +++ b/sysdeps/x86/cacheinfo.c
>
> > + assert (cpu_features->basic.kind != arch_kind_unknown);
>
> Why doesn't this assert fire occasionally? How do you ensure that
See
https://sourceware.org/bugzilla/show_bug.cgi?id=26203
It only happens in dlopen from a static executable.
> relocation processing is correctly ordered?
cpu_features is also initialized by IFUNC relocation in ld.so which
is relocated before libc.so.
> > +/* NB: Call init_cacheinfo by initializing a dummy function pointer via
> > + IFUNC relocation. */
> > +extern void __x86_cacheinfo (void) attribute_hidden;
> > +const void (*__x86_cacheinfo_p) (void) attribute_hidden
> > + = __x86_cacheinfo;
> > +
> > +__ifunc (__x86_cacheinfo, __x86_cacheinfo, NULL, void, init_cacheinfo);
> > #endif
>
> Please expand the comment and mention that is used to initialize the
> dormant copy of ld.so after static dlopen. The comment in
> sysdeps/x86/dl-get-cpu-features.c is good.
Will do.
> > diff --git a/sysdeps/x86/dl-get-cpu-features.c b/sysdeps/x86/dl-get-cpu-features.c
> > index 5f9e46b0c6..da4697b895 100644
> > --- a/sysdeps/x86/dl-get-cpu-features.c
> > +++ b/sysdeps/x86/dl-get-cpu-features.c
> > @@ -1,4 +1,4 @@
> > -/* This file is part of the GNU C Library.
> > +/* Initialize CPU feature data via IFUNC relocation.
> > Copyright (C) 2015-2020 Free Software Foundation, Inc.
> >
> > The GNU C Library is free software; you can redistribute it and/or
> > @@ -18,6 +18,29 @@
> >
> > #include <ldsodefs.h>
> >
> > +#ifdef SHARED
> > +# include <cpu-features.c>
> > +
> > +/* NB: Normally, DL_PLATFORM_INIT calls init_cpu_features to initialize
> > + CPU features. But when loading ld.so inside of static executable,
> > + DL_PLATFORM_INIT isn't called. Call init_cpu_features by initializing
> > + a dummy function pointer via IFUNC relocation for ld.so. */
> > +extern void __x86_cpu_features (void) attribute_hidden;
> > +const void (*__x86_cpu_features_p) (void) attribute_hidden
> > + = __x86_cpu_features;
> > +
> > +void
> > +_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);
> > +}
> > +
> > +__ifunc (__x86_cpu_features, __x86_cpu_features, NULL, void,
> > + _dl_x86_init_cpu_features);
> > +#endif
> > +
> > #undef __x86_get_cpu_features
>
> Why do we need both the conditional check and the function pointer hack?
Because _dl_x86_init_cpu_features is called both indirectly and by IFUNC
reloc in dynamic executable, but it is only called by IFUNC reloc when
dlopen in static executable.
> I expect that one of the function pointers can go, probably the one
> here. The cache hierarchy data might be used by a string function that
> has not been selected by IFUNC.
>
There are one IFUNC reloc in ld.so and the other in libc.so. We need
both.
--
H.J.
next prev parent reply other threads:[~2020-09-28 13:48 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-18 16:07 V2 [PATCH 0/4] ld.so: Add --list-tunables to print tunable values H.J. Lu
2020-09-18 16:07 ` [PATCH 1/4] x86: Initialize CPU info via IFUNC relocation [BZ 26203] H.J. Lu
2020-09-28 13:08 ` Florian Weimer
2020-09-28 13:48 ` H.J. Lu [this message]
2020-09-28 14:05 ` Florian Weimer
2020-09-28 14:20 ` H.J. Lu
2020-09-28 14:22 ` Florian Weimer
2020-09-28 14:39 ` H.J. Lu
2020-09-28 14:47 ` Florian Weimer
2020-09-28 17:54 ` V3 [PATCH] " H.J. Lu
2020-09-29 7:53 ` Florian Weimer
2020-09-29 11:44 ` H.J. Lu
2020-10-01 8:46 ` Florian Weimer
2020-10-01 19:50 ` V4 " H.J. Lu
2020-10-08 13:22 ` PING: " H.J. Lu
2020-10-15 12:53 ` PING^2: " H.J. Lu
2022-05-02 13:59 ` Sunil Pandey
2022-05-03 18:51 ` Sunil Pandey
2020-09-18 16:07 ` [PATCH 2/4] Set tunable value as well as min/max values H.J. Lu
2020-09-28 13:35 ` Florian Weimer
2020-09-28 13:53 ` H.J. Lu
2020-09-28 14:03 ` Florian Weimer
2020-09-28 17:30 ` Siddhesh Poyarekar
2020-09-29 4:00 ` V3 [PATCH] " H.J. Lu
2020-09-29 4:45 ` Siddhesh Poyarekar
2020-09-29 4:47 ` Siddhesh Poyarekar
2020-09-29 12:30 ` V4 " H.J. Lu
2020-09-29 13:50 ` Siddhesh Poyarekar
2020-09-29 14:54 ` V5 " H.J. Lu
2020-09-29 15:58 ` Siddhesh Poyarekar
2020-09-18 16:07 ` [PATCH 3/4] x86: Move x86 processor cache info to cpu_features H.J. Lu
2020-09-18 16:07 ` [PATCH 4/4] ld.so: Add --list-tunables to print tunable values H.J. Lu
2020-09-21 8:25 ` Florian Weimer
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=CAMe9rOo_OUmsikKWj_pXiy-G+NJZGGJwzn2_YTczXPvd09FaRA@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).