Re: [Gems-users] Magic instruction fails to stop cpu & enter debugger

[Dan Gibson <degibson@xxxxxxxx>]

The error you are seeing happens because the target OS cannot deliver an interrupt to a cpu (in this case cpu7). The OS thinks the CPU is malfunctioning, and panics.

Usually, when I see this error, it is because someone has disabled a CPU. This sort of thing is unrelated to simple magic breakpoints. 

Are you disabling CPUs? If so, you are leaving them disabled for too long.

Are you running with a checkpoint you took while Ruby was loaded? If so, you must manually edit the checkpoint to set all CPU stall_times to zero.

Regards,

Dan

On Mon, Jun 7, 2010 at 3:52 AM, lopamudra chatterjee <lopamudralopa@xxxxxxxxxxxxxx> wrote:

Hi,

I have used the MAGIC(0x40000) in code before start of parallel section. 
Now if before loading ruby I do magic-break-enable & start the programs then I am getting the following error :


send mondo timeout [4836707 NACK 0 BUSY]
IDSR 0x80 cpuids:0x7
panic:failed to stop cpu7.
panic[cpu15]/thread=30002255000:send mondo set :timeout
Then it shows message telling that it fails to stop all the processors.& then it enters a debugger.


But this does not occur if I first load ruby & then start the processes. 
My target is a 16 core abisko machine running solaris 10 & benchmark is water.

Any idea how this error could be prevented?


Thanks & Regards,
Lopamudra Chatterjee

_______________________________________________
Gems-users mailing list
Gems-users@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/gems-users
Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/" to your search.

--
http://www.cs.wisc.edu/~gibson [esc]:wq!