linux-kernel - Re: [PATCH 11/12] perf, x86: implement AMD IBS event configuration

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: <20100421092145.GV11907@erda.amd.com>
Date:	Wed, 21 Apr 2010 11:21:45 +0200
From:	Robert Richter <robert.richter@....com>
To:	Stephane Eranian <eranian@...gle.com>
CC:	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Ingo Molnar <mingo@...e.hu>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 11/12] perf, x86: implement AMD IBS event configuration

On 21.04.10 11:02:42, Stephane Eranian wrote:
> On Wed, Apr 21, 2010 at 10:47 AM, Robert Richter <robert.richter@....com> wrote:
> > On 20.04.10 18:05:57, Robert Richter wrote:
> >> > What is the problem with directly using the period here, rejecting
> >> > any value that is off range or with bottom 4 bits set?
> >>
> >> Yes, I will create an updated version of this patch.
> >
> > Stephane, do you think having the lower 4 bits set is worth an EINVAL?
> > I would rather ignore them since the accuracy is not really necessary
> > compared to a range lets say from 100000 cycles? Otherwise this will
> > make the setup of ibs much more complicated. The check could be moved
> > to userland and generate a waring or so.
> 
> Explain why you think it would be more complicated?
> If I recall there is already a function to validate the attrs:
> amd_pmu_hw_config().
> But may be you are talking about userland setup.
> 
> Here is one argument why this might be important. Some people like to
> know exactly
> the sampling period because they use a particular value, like a prime
> number. You
> chopping off the bottom 4 bits could break this logic silently.

Ok, I see your point. I was thinking of some decimal value used to set
the sample period. You will then have to check if the lower 4 bits are
set or not by doing a dec to hex conversion and so on. But I realized
that multiples of 10000 can be devided by 16 and thus all lower 4 bits
are always cleared.

So, I will check the lower 4 bits and return an error if they are set.

Thanks,

-Robert

-- 
Advanced Micro Devices, Inc.
Operating System Research Center
email: robert.richter@....com

--
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/