On Tue, Apr 10, 2012 at 12:47 AM, Tae Jun Ham <[email protected]> wrote:

> Hello,
>
>
> I have observed following happening :
>
>
> When I do fast-fwd for multicore Marss86,
>
>
> Each core gets per_cpu_fast_fwd instructions assigned.
>
>
> For a case when all cores are being utilized (not idle(or halt)),  one of
> those cores finishes its assigned instructions early and finishes its
> ptl_simpoint_reached(env->cpuindex) call in qemu/cpu-exec.c.
>
>
> Then, when it reaches cpu_loop_exit(env) in the same code (line 808 of
> qemu/cpu-exec.c), everything stops and fast-fwd never finishes.
>
>
> Let me know if you have an idea for this issue.
>
>
> I have found a bug in the code that doesn't check the flag based idle CPU.
Please can you test the attached patch if it solves this issue or not.

- Avadh

> Thank you
>
>
> -Tae Jun Ham
>
>
>
>
> _______________________________________________
> http://www.marss86.org
> Marss86-Devel mailing list
> [email protected]
> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>
>

Attachment: fast-fwd-fix.patch
Description: Binary data

_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel

Reply via email to