Hi All,

I don't know if it's useful, or not, but it still fails if I boot
using "nosmp", but it's ok if I built a kernel without CONFIG_SMP.
It's also ok with CONFIG_SMP and CONFIG_PREEMPT_RTB.

So it's something related to PPC64 + SMP + PREEMPT_RT_FULL


/Arnaud.

2012/7/13 Arnaud B. [off-list ref]:

Thanks anyway ;)

I made another try : same test case fail on 3.2.22-rt35
So the regression is between 3.0 and 3.2 !

Is there any ppc64 user here ?

/Arnaud.

2012/7/12 Frank Rowand [off-list ref]:

quoted

On 07/12/12 06:11, Arnaud B. wrote:

quoted

Here are 2 crash logs. One is normal booting, the other is with
CONFIG_DEBUG_RT_MUTEXES choosen.
for sure w in rt_mutex_top_waiter, return by plist_first_entry is
corrupted. There is even ASCII in it from time to time :p

< snip >

quoted

2012/7/11 Frank Rowand [off-list ref]:

quoted

On 07/11/12 09:02, Arnaud B. wrote:

quoted

Hi all,

< snip >

quoted

quoted

quoted

If you need an more information, I got all this boards in front of me :)

Can you show a little bit more of the messages aound the crash?  For example
the register contents.  Disassembly of the instruction at pc, and a few before
that.

-Frank

Thanks for the extra info.

I was looking at a plist corruption issue on a 2.6.29 rt kernel, but it does not
look like your situation at all.  So I don't have anything useful to contribute
about your case.

-Frank