Gilles Chanteperdrix wrote:
> Jan Kiszka wrote:
>> Marc Brown wrote:
>>> Hi,
>>> It seems that there are still structure alignment problems in RTDM (while 
>>> running xenotest), as already 
>>> talked about with Philippe. Compiled with a GCC 4.0.3-2 (on x86), here is 
>>> the 
>>> problem (output trace and xenomai /proc are provided herewith):
>> I rather suspect that you built all testing devices into the kernel or
>> loaded xeno_switchtest before xeno_timerbench so that the latency test
>> happens to open the switchtest device. Try "-D1" with latency, or check
>> /proc/xenomai/rtdm/rttestX/information for the actual mapping.
>> [We need to resolve this more gracefully...]
> A way to solve this may be to use the same method as switchtest: open
> all rttest devices in turn until the first ioctl succeeds.

Yeah, I'm going to check something in this direction (while still
keeping -D to override any automatism).

Attachment: signature.asc
Description: OpenPGP digital signature

Xenomai-core mailing list

Reply via email to