linux-kernel - Re: [PATCH 0/4] perf, amd: Enable AMD family 15h northbridge counters

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]

Message-ID: <20121112143630.GB24433@rric.localhost>
Date:	Mon, 12 Nov 2012 15:36:30 +0100
From:	Robert Richter <rric@...nel.org>
To:	Stephane Eranian <eranian@...gle.com>
Cc:	Jacob Shin <jacob.shin@....com>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Paul Mackerras <paulus@...ba.org>,
	Ingo Molnar <mingo@...hat.com>,
	Arnaldo Carvalho de Melo <acme@...stprotocols.net>,
	Thomas Gleixner <tglx@...utronix.de>,
	"H. Peter Anvin" <hpa@...or.com>, x86 <x86@...nel.org>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 0/4] perf, amd: Enable AMD family 15h northbridge counters

On 11.11.12 19:17:07, Stephane Eranian wrote:
> On Sat, Nov 10, 2012 at 12:50 PM, Robert Richter <rric@...nel.org> wrote:
> > Peter's main concerns were that my patch set is not in the
> > Intel-uncore style. I started reworking this but was not able to
> > finish my work. This concerns still exist.
> >
> That was my concern too. I don't recall exactly why it could not
> be totally disconnected from the core PMU. I think hardware-wise,
> it was possible. Could you refresh my memory?

Current implementation only allows the use of a single x86 pmu.
Multiple instances of x86 pmus in a system like an additional pmu for
nb counters requires a complete rework. You need to remove global
variables and extend function interfaces by arguments pointing to a
struct x86_pmu. All this without adding overhead (e.g. pointer
chasing) compared to the existing code.

This is also the reason why the intel-uncore implemenation is
basically a copy-and-paste of generic x86 pmu code plus uncore
specific changes.

Avoiding code cuplication and a complex rework were the reasons for
simply extending the family 10h implementation to also support family
15h nb counters. This seemed to me the best approach that time.

-Robert
--
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/