From: Josh Poimboeuf <jpoimboe@redhat.com>
To: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
the arch/x86 maintainers <x86@kernel.org>,
Andy Lutomirski <luto@kernel.org>,
Steven Rostedt <rostedt@goodmis.org>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@kernel.org>,
Thomas Gleixner <tglx@linutronix.de>,
Linus Torvalds <torvalds@linux-foundation.org>,
Masami Hiramatsu <mhiramat@kernel.org>,
Jason Baron <jbaron@akamai.com>, Jiri Kosina <jkosina@suse.cz>,
David Laight <David.Laight@aculab.com>,
Borislav Petkov <bp@alien8.de>
Subject: Re: [RFC PATCH 1/3] static_call: Add static call infrastructure
Date: Fri, 9 Nov 2018 08:55:21 -0600 [thread overview]
Message-ID: <20181109145521.2nfypucjsaq6bvxx@treble> (raw)
In-Reply-To: <CAKv+Gu_K-=rhV3fkE=OzFM=QipP8wbYhOeMZ8Pa6m=ZCnSOLew@mail.gmail.com>
On Fri, Nov 09, 2018 at 10:51:16AM +0100, Ard Biesheuvel wrote:
> Hi Josh,
>
> Thanks a lot for looking into this.
>
> On 8 November 2018 at 22:15, Josh Poimboeuf <jpoimboe@redhat.com> wrote:
> > Add a static call infrastructure. Static calls use code patching to
> > hard-code function pointers into direct branch instructions. They give
> > the flexibility of function pointers, but with improved performance.
> > This is especially important for cases where retpolines would otherwise
> > be used, as retpolines can significantly impact performance.
> >
> > This code is heavily inspired by the jump label code (aka "static
> > jumps"), as some of the concepts are very similar.
> >
> > There are three implementations, depending on arch support:
> >
> > 1) optimized: patched call sites (CONFIG_HAVE_STATIC_CALL_OPTIMIZED)
> > 2) unoptimized: patched trampolines (CONFIG_HAVE_STATIC_CALL_UNOPTIMIZED)
>
> Could we move to an idiom like inline/out of line? The unoptimized
> variant may be perfectly adequate for many arches, and 'unoptimized'
> sounds like there is something wrong with it.
Yeah, inline/out-of-line would be better. It's both more descriptive
and less derogatory :-)
> > + * Usage example:
> > + *
> > + * # Start with the following functions (with identical prototypes):
> > + * int func_a(int arg1, int arg2);
> > + * int func_b(int arg1, int arg2);
> > + *
> > + * # Define a 'my_key' reference, associated with func_a() by default
> > + * DEFINE_STATIC_CALL(my_key, func_a);
> > + *
> > + * # Call func_a()
> > + * static_call(my_key, arg1, arg2);
> > + *
> > + * # Update 'my_key' to point to func_b()
> > + * static_call_update(my_key, func_b);
> > + *
> > + * # Call func_b()
> > + * static_call(my_key, arg1, arg2);
> > + *
>
> Any way we can revert to the default implementation? That would be
> useful for, e.g., unloading modules that provide an alternative
> implementation.
I saw your original implementation had a "reset to default", but from
what I can tell, most users wouldn't need that.
Assuming the caller knows what the original dest function was, can't
they just call static_call_update(my_key, orig_func) directly?
> > diff --git a/include/linux/static_call_types.h b/include/linux/static_call_types.h
> > new file mode 100644
> > index 000000000000..7dd4b3d7ec2b
> > --- /dev/null
> > +++ b/include/linux/static_call_types.h
> > @@ -0,0 +1,19 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +#ifndef _STATIC_CALL_TYPES_H
> > +#define _STATIC_CALL_TYPES_H
> > +
> > +#include <linux/stringify.h>
> > +
> > +#define STATIC_CALL_TRAMP_PREFIX ____static_call_tramp_
> > +#define STATIC_CALL_TRAMP_PREFIX_STR __stringify(STATIC_CALL_TRAMP_PREFIX)
> > +
> > +#define STATIC_CALL_TRAMP(key) STATIC_CALL_TRAMP_PREFIX##key
> > +#define STATIC_CALL_TRAMP_STR(key) __stringify(STATIC_CALL_TRAMP(key))
> > +
>
> I needed to apply
>
> diff --git a/include/linux/static_call_types.h
> b/include/linux/static_call_types.h
> index 7dd4b3d7ec2b..6859b208de6e 100644
> --- a/include/linux/static_call_types.h
> +++ b/include/linux/static_call_types.h
> @@ -7,7 +7,7 @@
> #define STATIC_CALL_TRAMP_PREFIX ____static_call_tramp_
> #define STATIC_CALL_TRAMP_PREFIX_STR __stringify(STATIC_CALL_TRAMP_PREFIX)
>
> -#define STATIC_CALL_TRAMP(key) STATIC_CALL_TRAMP_PREFIX##key
> +#define STATIC_CALL_TRAMP(key) __PASTE(STATIC_CALL_TRAMP_PREFIX, key)
> #define STATIC_CALL_TRAMP_STR(key) __stringify(STATIC_CALL_TRAMP(key))
>
> /* The static call site table is created by objtool. */
>
> or I end up with
>
> 0000000000000000 <STATIC_CALL_TRAMP_PREFIXfoo>:
> 0: 14000000 b a8 <foo_a>
> 4: d503201f nop
Ha, oops. That was part of a last minute cleanup to share the macros
with objtool.
--
Josh
next prev parent reply other threads:[~2018-11-09 14:55 UTC|newest]
Thread overview: 57+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-08 21:15 [PATCH RFC 0/3] Static calls Josh Poimboeuf
2018-11-08 21:15 ` [RFC PATCH 1/3] static_call: Add static call infrastructure Josh Poimboeuf
2018-11-09 9:51 ` Ard Biesheuvel
2018-11-09 14:55 ` Josh Poimboeuf [this message]
2018-11-09 13:39 ` Ard Biesheuvel
2018-11-09 15:10 ` Josh Poimboeuf
2018-11-09 15:14 ` Ard Biesheuvel
2018-11-09 17:25 ` Ard Biesheuvel
2018-11-09 17:31 ` Josh Poimboeuf
2018-11-09 17:33 ` Ard Biesheuvel
2018-11-09 17:46 ` Josh Poimboeuf
2018-11-09 17:52 ` Ard Biesheuvel
2018-11-09 17:53 ` Ard Biesheuvel
2018-11-09 19:03 ` Josh Poimboeuf
2018-11-09 19:12 ` Ard Biesheuvel
2018-11-09 17:33 ` Josh Poimboeuf
2018-11-09 18:33 ` Steven Rostedt
2018-11-09 19:35 ` Josh Poimboeuf
2018-11-09 19:57 ` Steven Rostedt
2018-11-09 20:34 ` Josh Poimboeuf
2018-11-10 5:10 ` Steven Rostedt
2018-11-10 11:58 ` Ard Biesheuvel
2018-11-10 13:09 ` Steven Rostedt
2018-11-12 3:07 ` Josh Poimboeuf
2018-11-12 4:39 ` Ard Biesheuvel
2018-11-12 4:56 ` Josh Poimboeuf
2018-11-12 5:02 ` Ard Biesheuvel
2018-11-10 11:56 ` Ard Biesheuvel
2018-11-08 21:15 ` [RFC PATCH 2/3] x86/static_call: Add x86 unoptimized static call implementation Josh Poimboeuf
2018-11-08 21:15 ` [RFC PATCH 3/3] x86/static_call: Add optimized static call implementation for 64-bit Josh Poimboeuf
2018-11-08 21:24 ` [PATCH RFC 0/3] Static calls Josh Poimboeuf
2018-11-09 7:28 ` Ingo Molnar
2018-11-09 7:50 ` Ingo Molnar
2018-11-09 13:50 ` Ard Biesheuvel
2018-11-09 15:20 ` Josh Poimboeuf
2018-11-10 23:20 ` Peter Zijlstra
2018-11-11 13:42 ` Ard Biesheuvel
2018-11-11 14:25 ` Peter Zijlstra
2018-11-09 14:45 ` Josh Poimboeuf
2018-11-12 5:02 ` Ingo Molnar
2018-11-12 5:30 ` Josh Poimboeuf
2018-11-12 9:39 ` Ard Biesheuvel
2018-11-12 22:52 ` Josh Poimboeuf
2018-11-12 17:03 ` Steven Rostedt
2018-11-12 22:56 ` Josh Poimboeuf
2018-11-12 5:34 ` Andy Lutomirski
2018-11-09 15:16 ` Andy Lutomirski
2018-11-09 15:21 ` Josh Poimboeuf
2018-11-09 16:41 ` Josh Poimboeuf
2018-11-09 18:42 ` Steven Rostedt
2018-11-09 19:05 ` Andy Lutomirski
2018-11-09 19:37 ` Steven Rostedt
2018-11-09 19:44 ` Josh Poimboeuf
2018-11-09 19:59 ` Steven Rostedt
2018-11-09 20:36 ` Josh Poimboeuf
2018-11-10 15:13 ` Masami Hiramatsu
2018-11-09 20:53 ` Rasmus Villemoes
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=20181109145521.2nfypucjsaq6bvxx@treble \
--to=jpoimboe@redhat.com \
--cc=David.Laight@aculab.com \
--cc=ard.biesheuvel@linaro.org \
--cc=bp@alien8.de \
--cc=jbaron@akamai.com \
--cc=jkosina@suse.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=mhiramat@kernel.org \
--cc=mingo@kernel.org \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=x86@kernel.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 NNTP newsgroup(s).