A400-6X always crashes after a while
Using the 2.4.18-pa26 kernel on my A400-6X, the machine doesn't want to stay alive for more than about a day. The console log shows this:
~~~
Kernel panic: sba_iommu.c: I/O MMU @ fffffffffed22000 is out of mapping resources
<6>sym53c896-3-<15,0>: extraneous data discarded. sym53c896-3-<15,0>: COMMAND FAILED (89 0) @0000000016c95800.
Kernel panic: sba_iommu.c: I/O MMU @ fffffffffed22000 is out of mapping resources
In interrupt handler - not syncing
~~~
Does anyone know what might be causing this?
tia,
Duraid
Quoted 4 lines · Duraid Madina
<6>sym53c896-3-<15,0>: extraneous data discarded.
sym53c896-3-<15,0>: COMMAND FAILED (89 0) @0000000016c95800.Kernel panic: sba_iommu.c: I/O MMU @ fffffffffed22000 is out of mappingresources
That's my code.
It's possible the sym53c896 driver isn't releasing mapping resources in the above error path.
Under workload, can you
1) run the following shell script:
while : do date >> ~/Astro.out cat /proc/bus/runway/Astro | fgrep PDIR >> ~/Astro.out sleep 10 done
2) provide matching /var/log/syslog for the same timeframe.
3) explain briefly what the workload is.
The idea here is to monitor PDIR resource over time as whatever workload you are running exercises the system. In the meantime, I'll look at sym53c8xx driver and see if it's obviously not releasing the mapping resource.
grant