linux-kernel - Re: [PATCH] Revert "x86/tsc: Consolidate init code"

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: <20180911121548.GW5565@intel.com>
Date:   Tue, 11 Sep 2018 15:15:48 +0300
From:   Ville Syrjälä <ville.syrjala@...ux.intel.com>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Dou Liyang <douly.fnst@...fujitsu.com>,
        Pasha Tatashin <Pavel.Tatashin@...rosoft.com>,
        "H. Peter Anvin" <hpa@...or.com>,
        Peter Zijlstra <peterz@...radead.org>
Subject: Re: [PATCH] Revert "x86/tsc: Consolidate init code"

On Mon, Sep 10, 2018 at 06:53:54PM +0200, Thomas Gleixner wrote:
> On Mon, 10 Sep 2018, Ville Syrjälä wrote:
> 
> Good: 1718674.70 BogoMIPS (lpj=2863311530)
> Bad:   859455.59 BogoMIPS (lpj=1431852151)
> 
> while both kernels agree on the CPU frequency of 996MHz. This pretty much
> smells like the 32bit LPJ conversion bug which got fixed in rc3. Does the
> problem persist with rc3?

Indeed looks to be fixed by commit 17f6bac22493 ("x86/tsc:
Prevent result truncation on 32bit"). I both cherry-picked that
on top of rc2 to make sure it really is that commit, and also
tested plain rc3 to make sure it still works.

And comparing the bogomips between the three relevant commits makes
the bug pretty obvious in hindsight:
Calibrating delay loop (skipped), value calculated using timer frequency.. 1718674.70 BogoMIPS (lpj=2863311530)
Calibrating delay loop (skipped), value calculated using timer frequency.. 859455.59 BogoMIPS (lpj=1431852151)
Calibrating delay loop (skipped), value calculated using timer frequency.. 1994.50 BogoMIPS (lpj=3322410)

I suppose we just got very lucky with older kernels.

-- 
Ville Syrjälä
Intel