linux-kernel - Re: Network problem with 2.6.18-mm1 ?

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: <m1ejtui73g.fsf@ebiederm.dsl.xmission.com>
Date:	Fri, 29 Sep 2006 17:30:27 -0600
From:	ebiederm@...ssion.com (Eric W. Biederman)
To:	"Jesse Brandeburg" <jesse.brandeburg@...il.com>
Cc:	"Sukadev Bhattiprolu" <sukadev@...ibm.com>,
	"Auke Kok" <auke-jan.h.kok@...el.com>,
	linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
	"Andrew Morton" <akpm@...l.org>
Subject: Re: Network problem with 2.6.18-mm1 ?

"Jesse Brandeburg" <jesse.brandeburg@...il.com> writes:

> On 9/28/06, Sukadev Bhattiprolu <sukadev@...ibm.com> wrote:
>> $ cat /proc/interrupts
>>
>>            CPU0       CPU1
>>  28:          0          0   IO-APIC-fasteoi  eth0
>> NMI:         96         35
>> LOC:      18251      18524
>> ERR:          0
>
> you should be getting an interrupt every two seconds from the eth0
> (e1000) driver.  You are having interrupt delivery problems probably
> due to something screwing up interrupt routing in the kernel.
> Normally these issues are associated with MSI interrupts but your
> adapter doesn't support those and is using generic IRQ
>
> I'm guessing that if you somehow enable interrupts on your vga card on
> the same bus as e1000 (bus 3) it will have interrupt delivery problems
> as well.  Maybe try xorg?

To summarize.

We have an e1000 plugged into a pci-x slot on an Opteron system with
an amd chipset.

That motherboard has 3 ioapics.  (One on each PCI-X bridge and
one on the 8111 for handling everything else.

We know we are getting interrupts through the 8111 ioapic.

We don't know which ioapic the pci-x bus is hooked to.

So either the ioapics on the 8131 are having problems.
Or we have a problem parsing the irq routing tables.

We see in dmesg.
[    0.000000] I/O APIC #2 at 0xFEC00000.
[    0.000000] I/O APIC #3 at 0xE8000000.
[    0.000000] I/O APIC #4 at 0xE8001000.
...
[   97.410411] PCI: Cannot allocate resource region 0 of device 0000:00:0a.1
[   97.423945] PCI: Cannot allocate resource region 0 of device 0000:00:0b.1

We see in lspci
> 
> 0000:00:0a.1 PIC: Advanced Micro Devices [AMD] AMD-8131 PCI-X IOAPIC (rev 01)
> (prog-if 10 [IO-APIC])
> 	Subsystem: Advanced Micro Devices [AMD] AMD-8131 PCI-X IOAPIC
> 	Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
> Stepping- SERR- FastB2B-
> 	Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort-
> <TAbort- <MAbort- >SERR- <PERR-
> 	Latency: 0
> 	Region 0: Memory at 88100000 (64-bit, non-prefetchable) [size=4K]
> 
> 0000:00:0b.1 PIC: Advanced Micro Devices [AMD] AMD-8131 PCI-X IOAPIC (rev 01)
> (prog-if 10 [IO-APIC])
> 	Subsystem: Advanced Micro Devices [AMD] AMD-8131 PCI-X IOAPIC
> 	Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
> Stepping- SERR- FastB2B-
> 	Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort-
> <TAbort- <MAbort- >SERR- <PERR-
> 	Latency: 0
> 	Region 0: Memory at 88101000 (64-bit, non-prefetchable) [size=4K]

So it looks like the kernel moved the ioapics.

The following patch in 2.6.18-mm1 is known to have that effect.
x86_64-mm-insert-ioapics-and-local-apic-into-resource-map

Can you please try reverting that one patch?

There is a fix an updated version of that patch I think in -mm2
but I haven't had a chance to see if it fixes the problem yet.

Eric
-
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/