Re: 2.6.20-rc2: kernel BUG at include/asm/dma-mapping.h:110!

[Stefan Richter <stefanr@s5r6.in-berlin.de>]

I wrote on 2007-01-02:
> Kyuma Ohta wrote:
> ...
>> Now,I'm testing 2.6.20-rc3 for x86_64, submitted patch for this issue;
>> "Fault has happened  in 'cleanuped' sbp2/1394 module in *not 32bit*
>> architecture hardwares ."
>>
>> As result of, sbp2 driver in 2.6.20-rc3 is seems to running 
>> w/o any faults,but communication both host and harddrive(s) 
>> was seems to be unstable yet :-(
>> Sometimes confuse packets,such as *very* older 1394 driver :-(
> 
> That is, sbp2 on 2.6.20-rc3 works less stable for you than on 2.6.19? Or
> which previous kernel is the basis of your comparison? Are there any log
> messages or other diagnostics? And what hardware do you have?
> 
> If you can tell which kernel was good for you, I could create a set of
> patches for you which allows to revert sbp2 while keeping the rest of
> the kernel at the level of 2.6.20-rc3, so that you could find the
> destabilizing change (if it happened in sbp2, not somewhere else).
[...]

So, how about it? Is there an actual regression? If so, we should find
the cause and fix before 2.6.20 is released.

Note, sbp2's optional parameter serialize_io=0 does not work correctly
yet with some devices (it never did), therefore use sbp2 with anything
than default parameters if there are problems.
-- 
Stefan Richter
-=====-=-=== ---= =--=-
http://arcgraph.de/sr/