linux-kernel - Re: [GIT PULL] x86/boot changes for v3.13

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: <20131112113208.GC12849@pd.tnic>
Date:	Tue, 12 Nov 2013 12:32:08 +0100
From:	Borislav Petkov <bp@...en8.de>
To:	Ingo Molnar <mingo@...nel.org>
Cc:	Yinghai Lu <yinghai@...nel.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	"H. Peter Anvin" <hpa@...or.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [GIT PULL] x86/boot changes for v3.13

On Tue, Nov 12, 2013 at 12:23:38PM +0100, Ingo Molnar wrote:
> So I suspect what Yinghai tried to say if CPU0 and CPU1 are not on the
> same node we do the printout incorrectly.

I hope your translation is correct :) I'd still like to get a
confirmation from him though.

> Arguably this was a pre-existing condition, but would be nice to fix
> it now that this code has emerged out of steady bitrot! :-)
>
> How difficult would it be in your opinion?

Well, I did try a weird, non-existant configuration:

kvm ... -smp 6 -numa node,nodeid=0,cpus=0\;2\;3 -numa node,nodeid=1,cpus=1\;4\;5

and what I get is:

[    0.068574] x86: Booting SMP configuration:
[    0.069006] .... node  #1, CPUs:      #1
[    0.147005] .... node  #0, CPUs:   #2 #3 #4 #5
[    0.445273] x86: Booted up 2 nodes, 6 CPUs

Before my cleanup and after removing the "fixing up alternatives"
message which hid things, the output looked like:

[    0.069621] smpboot: Booting Node   1, Processors  #   1 OK
[    0.146006] smpboot: Booting Node   0, Processors  #   2 #   3 #   4 #   5 OK
[    0.448320] Brought up 6 CPUs

The problem is not the indentation but that the current code slaps all
cpus on the last node, in this case node 0, because announce_cpu gets
the cores one by one.

A possible fix would be to collect the topology and dump it *only*
*after* the last core has been announced.

But before we do anything, I'd like to get a concrete system which has
that issue and we can talk about it then.

> Btw., while staring at that code once more I noticed the following
> small nit, there's a pre-existing weird way of writing the -1 literal:
>
>                           if (current_node > (-1))

It came from 2eaad1fddd7450a48ad464229775f97fbfe8af36.

-- 
Regards/Gruss,
    Boris.

Sent from a fat crate under my desk. Formatting is fine.
--
--
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/