There are a couple of more ways. You can specify that a checkpoint be
taken at the end of the simulation. Or you can instrument application code
with gem5's pseudo instructions.
--
Nilay
On Mon, 6 Feb 2012, Madhav Suresh wrote:
How do you take a checkpoint without the shell?
Madhav
On Mon, Feb 6, 2012 at 10:49 AM, Nilay Vaish <[email protected]> wrote:
Then it might be that the processor is in an idle loop, and is not doing
anything. You can try taking checkpoint at some other points, may be very
early in the boot process and check if pc value changes.
--
Nilay
On Mon, 6 Feb 2012, Madhav Suresh wrote:
Oh sorry, yes exactly, I'm able to connect to the terminal and issue
commands.
Madhav
On Mon, Feb 6, 2012 at 10:04 AM, Nilay Vaish <[email protected]> wrote:
You did not answer my question. What does 'run correctly' mean?
On Mon, 6 Feb 2012, Madhav Suresh wrote:
Yea exactly, I'm able to load the checkpoint and have it run correctly
with
bogus PC and NPC values. I was under the impression that this was the
actual value that's loaded after the checkpoint? After a checkpoint is
loaded, what is the codepath? I know that m5.simulate() is called in
Simulation.py, but I'm unsure as to what happens after that.
Apart from M5, the RIP that is saved is 0x400880. Where exactly is this
RIP
in the context of an x86 system?
Thanks,
Madhav
On Mon, Feb 6, 2012 at 8:32 AM, Nilay Vaish <[email protected]> wrote:
What do you mean by checkpoint resumes properly? Are you able to
connect
to the terminal and issue some commands? IIRC, there are two places in
the
checkpoint where the PC and NPC are stored.
--
Nilay
On Sat, 4 Feb 2012, Madhav Suresh wrote:
I'm using the AtomicSimpleCPU. Right for the _pc and _npc fields in
the
checkpoint file I put zero, but the checkpoint still resumes properly.
Here
are the first couple fetches:
1611728894411000: system.cpu: Fetch: PC:0x400880
1611728894412500: system.cpu: Fetch: PC:0x40056f
This happens every time I restore any checkpoint from this specific
kernel
image.
Nilay Vaish
Fri, 03 Feb 2012 10:28:00 -0800
On Fri, 3 Feb 2012, Madhav Suresh wrote:
Hey Guys,
I'm a little confused as to how exactly the checkpoint load works.
I've
noticed that on a restore the PC is always at 0x400880. No matter what
the
_pc is, M5 will always resume to that address, and then eventually get
back
into normal execution. What is the exact code path the the simulator
follows when the "-r" flag is set?
That sounds really strange. What CPU model are you using?
--
Nilay
______________________________******_________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/******mailman/listinfo/gem5-users<http://m5sim.org/cgi-bin/****mailman/listinfo/gem5-users>
<ht**tp://m5sim.org/cgi-bin/****mailman/listinfo/gem5-users<http://m5sim.org/cgi-bin/**mailman/listinfo/gem5-users>
<ht**tp://m5sim.org/cgi-bin/****mailman/listinfo/gem5-users<http://m5sim.org/cgi-bin/**mailman/listinfo/gem5-users>
<ht**tp://m5sim.org/cgi-bin/**mailman/listinfo/gem5-users<http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users>
______________________________****_________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/****mailman/listinfo/gem5-users<http://m5sim.org/cgi-bin/**mailman/listinfo/gem5-users>
<ht**tp://m5sim.org/cgi-bin/**mailman/listinfo/gem5-users<http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users>
______________________________**_________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/**mailman/listinfo/gem5-users<http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users>
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
