Re: [RTnet-developers] Kernel module for smc91111

[Hinko K. <hin...@i-...>]

On 09/20/2011 11:47 AM, Gilles Chanteperdrix wrote:
> On 09/20/2011 11:02 AM, Hinko Kocevar wrote:
>> I stand corrected - the problem is in the second part of that line, what
>> was proven by two more debug lines:
>>
>> rtdm_printk("%s: rtdm_clock_read() = %lld\n", __FUNCTION__,
>> rtdm_clock_read());
>> rtdm_printk("%s: *((nanosecs_abs_t *)skb->data) = %lld\n", __FUNCTION__,
>
> This is wrong: skb->data is a char *, you cast it to nanosecs_abs_t *
> which has a strongest alignment requirement, so you get...
>
>> [  226.206024] Xenomai: suspending kernel thread bf011030 ('rtnet-rtpc')
>> at 0xbf01e66c after exception #0x8
>
> ... an alignment trap.
>

What I've done is take the lines of code from the existing RTnet GIT 
source and just try to print out the value.
The offending line can be seen in icmp.c:241,

http://rtnet.git.sourceforge.net/git/gitweb.cgi?p=rtnet/rtnet;a=blob;f=stack/ipv4/icmp.c;h=2dbd3315dfa60e9fa02332cd8a203661130db8de;hb=a3bc0758ee33980836f1ca54ada9ccee762e24ad#l241


Maybe this is not an issue on non-arm (pxa255) targets?

Thank you!

-- 
Hinko Kocevar
Technical support software engineer
Instrumentation Technologies d.d.
Velika pot 22, SI-5250 Solkan - Slovenia
T:+386 5 3352600, F:+386 5 3352601
mailto: hin...@i-...

http://www.i-tech.si - When your users demand stability