linux-kernel - Re: PCI MSI breaks when booting with nosmp

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: <200804211340.21713.jbarnes@virtuousgeek.org>
Date:	Mon, 21 Apr 2008 13:40:20 -0700
From:	Jesse Barnes <jbarnes@...tuousgeek.org>
To:	Jean Delvare <jdelvare@...e.de>
Cc:	Andi Kleen <andi@...stfloor.org>,
	linux-pci@...ey.karlin.mff.cuni.cz,
	LKML <linux-kernel@...r.kernel.org>, Pavel Machek <pavel@....cz>,
	Tejun Heo <htejun@...il.com>,
	Tom Long Nguyen <tom.l.nguyen@...el.com>,
	Randy Dunlap <rdunlap@...otime.net>,
	Jeff Garzik <jgarzik@...ox.com>
Subject: Re: PCI MSI breaks when booting with nosmp

On Monday, April 21, 2008 1:20 pm Jean Delvare wrote:
> Hi Andi,
>
> Le lundi 21 avril 2008, Andi Kleen a écrit :
> > Jesse Barnes <jbarnes@...tuousgeek.org> writes:
> > > Ok, I see this too on my desktop machine.  It looks like we're not
> > > getting interrupts setup correctly in the nosmp case.  Still digging
> > > through to see why though...
> >
> > NoSMP disables the io-apic and a lot of modern systems don't work without
> > APIC.
> >
> > If you just want to run with a single cpu for testing etc. always use
> > maxcpus=1 (not 0, that will disable the APIC too)
>
> In my case, booting with noapic works fine, so I don't think that the
> problem is related to APIC being disabled.

One difference between noapic and nosmp is that in the nosmp case even the 
local APIC setup won't occur in native_smp_prepare_cpus(), due to 
smp_sanity_check() returning -1 in the setup_max_cpus == 0 case.

So we either need smp_sanity_check to do a little more APIC setup if max_cpus 
== 0 or shuffle things around in native_smp_prepare_cpus().  Since the former 
is already done for 32 bit builds for the NMI vector, maybe we should just 
make it unconditional so that MSI works?

Jesse
--
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/