Re: APIC causes vr0 watchdog timeouts and Interrupt storms

On Sunday 29 August 2004 06:49, Justin Settle wrote:

> Hello,
>
> I recently installed 5.3-BETA.  After booting, I found was unable to use
> my usb mouse or my ethernet (Via-rhine).  I started tinkering with the
> kernel and found that if I disabled APIC, this system worked without
> ethernet issues or mouse isssues.  After talking with some people they
> suggested I post boot -v, dmesg's of the different kernels so here they
> are:
>
> The first thing I tried was a GENERIC with SMP commented out.  It had
> the same issues as GENERIC:
>
> Copyright (c) 1992-2004 The FreeBSD Project.
> Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
>  The Regents of the University of California. All rights reserved.
> FreeBSD 5.3-BETA2 #0: Sat Aug 28 23:18:57 EDT 2004
>     root_at_amon.quaggaspace.org:/usr/obj/usr/src/sys/GENERICNOSMP

[snip]

> acpi link set: _CRS failed for link \\_SB_.PCI0.ALKB - AE_NULL_ENTRY
> acpi link set: curr irq 0 != 21 for \\_SB_.PCI0.ALKB (ignoring)
> unknown: _SRS failed, irq 21 via \\_SB_.PCI0.ALKB
> found-> vendor=0x1106, dev=0x3038, revid=0x80
>  bus=0, slot=16, func=0
>  class=0c-03-00, hdrtype=0x00, mfdev=1
>  cmdreg=0x0007, statreg=0x0210, cachelnsz=8 (dwords)
>  lattimer=0x20 (960 ns), mingnt=0x00 (0 ns), maxlat=0x00 (0 ns)
>  intpin=a, irq=11
>  powerspec 2  supports D0 D1 D2 D3  current D0
>  map[20]: type 4, range 32, base 0000d400, size  5, enabled
> pcib0: matched entry for 0.16.INTB (src \\_SB_.PCI0.ALKB)
> pcib0: slot 16 INTB is already routed to irq 21
> found-> vendor=0x1106, dev=0x3038, revid=0x80
>  bus=0, slot=16, func=1
>  class=0c-03-00, hdrtype=0x00, mfdev=1
>  cmdreg=0x0007, statreg=0x0210, cachelnsz=8 (dwords)
>  lattimer=0x20 (960 ns), mingnt=0x00 (0 ns), maxlat=0x00 (0 ns)
>  intpin=b, irq=21
>  powerspec 2  supports D0 D1 D2 D3  current D0
>  map[20]: type 4, range 32, base 0000d800, size  5, enabled
> pcib0: matched entry for 0.16.INTC (src \\_SB_.PCI0.ALKB)
> pcib0: slot 16 INTC is already routed to irq 21
> found-> vendor=0x1106, dev=0x3038, revid=0x80
>  bus=0, slot=16, func=2
>  class=0c-03-00, hdrtype=0x00, mfdev=1
>  cmdreg=0x0007, statreg=0x0210, cachelnsz=8 (dwords)
>  lattimer=0x20 (960 ns), mingnt=0x00 (0 ns), maxlat=0x00 (0 ns)
>  intpin=c, irq=21
>  powerspec 2  supports D0 D1 D2 D3  current D0
>  map[10]: type 1, range 32, base ee000000, size  8, enabled
> pcib0: matched entry for 0.16.INTD (src \\_SB_.PCI0.ALKB)
> pcib0: slot 16 INTD is already routed to irq 21
> found-> vendor=0x1106, dev=0x3104, revid=0x82
>  bus=0, slot=16, func=3
>  class=0c-03-20, hdrtype=0x00, mfdev=0
>  cmdreg=0x0007, statreg=0x0210, cachelnsz=8 (dwords)
>  lattimer=0x20 (960 ns), mingnt=0x00 (0 ns), maxlat=0x00 (0 ns)
>  intpin=d, irq=21
>  powerspec 2  supports D0 D1 D2 D3  current D0

[snip]

> Interrupt storm detected on "irq21: uhci1 uhci2"; throttling interrupt
> source

[snip]

I had similar USB problems on a Via KT600 board. In my case, manual irq 
routing fixed the interrupt strom on irq 21.     

# cat /boot/device.hints | grep acpi
hw.acpi.pci.link.0.16.0.irq=21
hw.acpi.pci.link.0.16.1.irq=21
hw.acpi.pci.link.0.16.2.irq=21
hw.acpi.pci.link.0.16.3.irq=21

USB works perfectly now with ACPI and with options SMP/APIC defined in my 
kernelconfig and ACPI. 

Nate Lawson comitted some major improvements on the acpi_pci_link.c code 
regarding the _CRS and _SRS objects and KTxxx bords in the last weeks, so I 
think it's a good idea cc'ing him.     

I can't tell you much about the vr (VT6102) because it's irq routing is 
different on the KT600 board (device 0.18.0 --> irq 5 ) and the chip 
initializes without problems here. 

Cheers, 
ch 

-- 
Christian Hiris <4711_at_chello.at> | OpenPGP KeyID 0x941B6B0B 
OpenPGP-Key at hkp://wwwkeys.eu.pgp.net and http://pgp.mit.edu