Re: Weird interrupt issues with -current

[Kurt Schreiner <ks%ub.uni-mainz.de@localhost>]

On Wed, Jun 04, 2008 at 04:36:14AM +0200, Scott Ellis wrote:
> I updated my kernel from May 28th to today (June 3rd), and found that at
> least two of my NICs didn't seem to communicate properly.  The "msk0"
> didn't spit out any error messages, but it didn't pass traffic either.
> :-)  My "ex0" kept printing "device timeout", and didn't seem to pass
> traffic at all.
> 
> The before and after dmesg's are attached, but diff is:
> 
> 2c7
> < NetBSD 4.99.64 (INTREPID.P5W) #1: Tue Jun  3 19:19:09 PDT 2008
> ---
>  > NetBSD 4.99.63 (INTREPID.P5W) #0: Wed May 28 18:13:50 PDT 2008
> 57c62
> < , Yukon-2 EC rev. A3 (0x2): ioapic0 pin 16
> ---
>  > , Yukon-2 EC rev. A3 (0x2): ioapic0 pin 19
> 90c95
> < ex0: interrupting at ioapic0 pin 17
> ---
>  > ex0: interrupting at ioapic0 pin 22
> 128c133
> < timecounter: Timecounter "TSC" frequency 2703547170 Hz quality 3000
> ---
>  > timecounter: Timecounter "TSC" frequency 2703544650 Hz quality 3000
> 
> 
> I'm not sure why the different interrupt routing happened, but it
> doesn't seem good. :-)
Same here on a IBM T43p notebook:

8c8
< netbsd: NetBSD 4.99.64 (PADDI) #283: Tue Jun  3 16:54:36 MEST 2008
---
> netbsd: NetBSD 4.99.64 (PADDI) #284: Tue Jun  3 22:57:37 MEST 2008
14,15c14,15
< netbsd: multiboot: Symbol table at 0xc0547e34, length 280064 bytes
< netbsd: multiboot: String table at 0xc058c434, length 280871 bytes
---
> netbsd: multiboot: Symbol table at 0xc0547e34, length 280448 bytes
> netbsd: multiboot: String table at 0xc058c5b4, length 281398 bytes
56c56
< netbsd: acpitz0 at acpi0 (THM0): critical 99.0C passive 49.0C, passive cooling
---
> netbsd: acpitz0 at acpi0 (THM0): critical 99.0C passive 64.0C, passive cooling
82c82
< netbsd: bge0: interrupting at ioapic0 pin 16
---
> netbsd: bge0: interrupting at ioapic0 pin 20
113c113
< netbsd: ath0: interrupting at ioapic0 pin 21
---
> netbsd: ath0: interrupting at ioapic0 pin 11
120c120
< netbsd: cbb0: interrupting at ioapic0 pin 16
---
> netbsd: cbb0: interrupting at ioapic0 pin 22

and later in /var/log/messages:

Jun  3 22:59:29 ipaddi /netbsd: bge0: watchdog timeout -- resetting
Jun  3 23:00:40 ipaddi /netbsd: bge0: watchdog timeout -- resetting
Jun  3 23:00:40 ipaddi /netbsd: bge0: block failed to stop: reg 0x1400, bit 
0x00000002

indicating bge0 doesn't work.


Another really serious problem w/ kernel #284:

Temperature on sensor TMP3 (no idea, what this mesures) goes up very
fast: it shows normally at most ~80 degC and goes up to > 115! Some
seconds later the notebook is powered down really ungracefully by the
BIOS! This can even be accelerated by switching from X to text console.


Kurt