linux-kernel - Re: [PATCH] x86/signal: Deinline get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 28 Sep 2015 15:43:11 +0200
From:	Denys Vlasenko <dvlasenk@...hat.com>
To:	"H. Peter Anvin" <hpa@...or.com>, Ingo Molnar <mingo@...nel.org>
CC:	Andy Lutomirski <luto@...capital.net>,
	Borislav Petkov <bp@...en8.de>,
	Brian Gerst <brgerst@...il.com>, x86@...nel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] x86/signal: Deinline get_sigframe, save 240 bytes

On 09/28/2015 02:33 PM, H. Peter Anvin wrote:
> This one makes me slightly nervous, because it isn't clear
> that these aren't potentially performance sensitive.

CALL instruction is not a crime :)
It costs about the same as one read-modify-write
operation on a memory operand.

This function is used in signal delivery code.
If performance critical code uses massive numbers
of signals, it already has a problem, Unix signals
are too inefficient. That's why we have futexes etc...

> On September 28, 2015 5:23:57 AM PDT, Denys Vlasenko <dvlasenk@...hat.com> wrote:
>> This function compiles to 277 bytes of machine code and has 4
>> callsites.

I must correct myself: there are two callsites, not four.
(There are four calls in the source, but two of them are
in 32-bit code and two are in 64-bit).

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/