linux-kernel - Re: [patch] x86_64: fix boot hang caused by CALGARY_IOMMU_ENABLED_BY

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]

Date:	Wed, 20 Dec 2006 20:09:53 +0200
From:	Muli Ben-Yehuda <muli@...ibm.com>
To:	Ingo Molnar <mingo@...e.hu>
Cc:	Linus Torvalds <torvalds@...l.org>, linux-kernel@...r.kernel.org,
	Dave Jones <davej@...hat.com>, Andrew Morton <akpm@...l.org>
Subject: Re: [patch] x86_64: fix boot hang caused by CALGARY_IOMMU_ENABLED_BY_DEFAULT

On Wed, Dec 20, 2006 at 05:23:38PM +0100, Ingo Molnar wrote:

> > I think that it makes sense to have it default y for the mainline 
> > kernel and default n for the distro kernels, which is why I added the 
> > option to make it possible to compile Calgary in but only enable it if 
> > you want to use it. Previously if you compiled it in it would be used, 
> > period. You may disagree, but fundamentally I think the mainline 
> > kernel should be fairly experimental, which means enabling new code by 
> > default.
> 
> that's a totally wrong attitude - the mainline kernel is /not/
> experimental. A distro might or might not enable the new option, but
> we just dont enable experimental platform support code via "default
> y"...

I disagree, it seems to me most "experimental platform support code"
is simply enabled because it doesn't even have a CONFIG option (c.f.,
recent genirq and IO-APIC breakage on x86-64). With regards to this
specific option, you might even say that not defaulting to 'y' here
would be a regression in behaviour against previous released kernels,
which used Calgary if it was compiled in, no questions asked. So at
least in that sense, instructing the user to select y if unsure and
default y are appropriate.

> The other problem is that the changelog entry says that it's off by
> default, while in reality the new option switched this code on for
> my box, and broke it.

Sorry about that (both the wrong changelog entry and the fact that it
broke your box).

> > As to what actually happened, I'm betting your machine has both 
> > Calgary and CalIOC2, the PCI-e version of Calgary, which is not yet 
> > supported by pci-calgary.c. [...]
> 
> no, what happened is what i described in my second patch. That 'new 
> code' which was default-enabled had a bug which locked up my box.

Yes, I realized that once I've read your other mail.

Cheers,
Muli
-
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/