[Qemu-devel] Execute QEMU "the same"

[Qemu-devel] Execute QEMU "the same"

[Lyu Mitnick]

[-- Attachment #1: Type: text/plain, Size: 561 bytes --]

Hello all,

I am wondering to know whether QEMU could run "the same" twice. ie. the same
of execution trace.
I read the user manual and find options like:  "-clock dynticks -rtc
base=2006-06-17T16:01:21,clock=vm -icount 1".
Would QEMU run "the same" with these options??

PS. I also found the function cpu_io_recompile() with comments:
      /* in deterministic execution mode, instructions doing device I/Os
         must be at the end of the TB */
      But I don't know the means of deterministic execution mode and how to
turn on it.

thanks a lot

Mitnick

[-- Attachment #2: Type: text/html, Size: 774 bytes --]

Re: [Qemu-devel] Execute QEMU "the same"

[Edgar E. Iglesias]

On Tue, Aug 23, 2011 at 07:04:43PM +0800, Lyu Mitnick wrote:
> Hello all,
> 
> I am wondering to know whether QEMU could run "the same" twice. ie.
> the same of execution trace.
> I read the user manual and find options like:  "-clock dynticks -rtc
> base=2006-06-17T16:01:21,clock=vm -icount 1".
> Would QEMU run "the same" with these options??
> 
> PS. I also found the function cpu_io_recompile() with comments:
>       /* in deterministic execution mode, instructions doing device I/Os
>          must be at the end of the TB */
>       But I don't know the means of deterministic execution mode and how to
> turn on it.
> 
> thanks a lot

Hi,

I think so, yes. But it really depends on what you are doing.
I've gotten good results, but only tested it on CRIS with models of boards
that dont have peripherals that interact with the outside. For example,
the board in question has a CRIS and a couple of devices that compute
stuff or generate data but do not interact with the outside world.
External IO (such as you typing stuff into a simulated uart) will
ofcourse not be deterministic.

Cheers

Re: [Qemu-devel] Execute QEMU "the same"

[陳韋任]

> I read the user manual and find options like:  "-clock dynticks -rtc
> base=2006-06-17T16:01:21,clock=vm -icount 1".
> Would QEMU run "the same" with these options??
> 
> PS. I also found the function cpu_io_recompile() with comments:
>       /* in deterministic execution mode, instructions doing device I/Os
>          must be at the end of the TB */
>       But I don't know the means of deterministic execution mode and how to
> turn on it.

  Perhaps you can find the answer of what is the deterministic execution
mode in the thesis below,
  http://www.few.vu.nl/~vdkouwe/doc/msc-thesis-cs-erik-van-der-kouwe.pdf

Regards,
chenwj

-- 
Wei-Ren Chen (陳韋任)
Computer Systems Lab, Institute of Information Science,
Academia Sinica, Taiwan (R.O.C.)
Tel:886-2-2788-3799 #1667