linux-kernel - Re: 4.5.0+ panic when setup loop device

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: <20160318075115.GD6344@twins.programming.kicks-ass.net>
Date:	Fri, 18 Mar 2016 08:51:15 +0100
From:	Peter Zijlstra <peterz@...radead.org>
To:	Mike Galbraith <umgwanakikbuti@...il.com>
Cc:	Xiong Zhou <jencce.kernel@...il.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Ingo Molnar <mingo@...nel.org>, Borislav Petkov <bp@...en8.de>,
	Andreas Herrmann <aherrmann@...e.com>
Subject: Re: 4.5.0+ panic when setup loop device

On Fri, Mar 18, 2016 at 05:11:54AM +0100, Mike Galbraith wrote:
> On Thu, 2016-03-17 at 10:52 +0100, Peter Zijlstra wrote:
> 
> > Andreas; Borislav said to Cc you since you wrote all this.
> > The issue is that Linux assumes:
> > 
> > 	> nr_logical_cpus = nr_cores * nr_siblings
> 
> It also seems to now assume that if SMT is possible, it's enabled.

Urgh..

What I think, with my pre wakeup brain, happens is that the CPUID
topology muck still reports 2 siblings (it tends to do that).

But the BIOS only reports APIC-IDs for all your cores, so our
nr_cpu_ids is reduced, while the nr_siblings count it not.

And them *boom*.

This is the same old problem that is nearly impossible to tell if HT is
enabled or not -- complete and utter trainwreck :/

I need to go make wake-up-juice and ponder wth to do about this.