Gabe, the only side effect seems to be the invalidation of the ITB and DTB
entries. Other than that any change in a register value affects only the
registers themselves. If copyRegs() is being called when the registers are
in a consistent state amongst themselves, I don't think any checks need to
be carried out or side effects need to be taken care of. And as I
understand, we can always proactively invalidate ITB and DTB entries when
copying registers.
--
Nilay
On Fri, 1 Jul 2011, Gabriel Michael Black wrote:
That's more complicated, and in the mean time there should probably be a
panic there. If it's quick, then the panic would only be there for a short
time, and if it's not then there isn't a gotcha dangling there waiting to
catch somebody.
If you want to implement that function (which would be nice, so go for it)
you'll need to look at what effects and checks the various miscregs have in
isa.cc. Then you'll need to determine what order they need to be updated in
so that nothing ends up broken, no panics are triggered, all the right side
effects end up in effect, etc. It shouldn't be -too- nasty because the side
effects I think are relatively minor in most cases, but it still needs to be
done carefully.
Gabe
Quoting Nilay Vaish <[email protected]>:
On 2011-07-01 15:02:18, Nilay Vaish wrote:
Gabe, I took care of the things you had pointed out. Do you think we
should
commit this woithout taking care of the misc. registers?
Gabe Black wrote:
This change itself is for the better since it gets copyRegs itself to
work. The question is whether to bump the panic from there up to
copyMiscRegs, and since I really don't think it will work as is the answer
is probably yes.
Well, that would not solve the actual problem. Suppose I were to implement
it, where should I start from?
- Nilay
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.m5sim.org/r/759/#review1377
-----------------------------------------------------------
On 2011-07-01 15:00:08, Nilay Vaish wrote:
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.m5sim.org/r/759/
-----------------------------------------------------------
(Updated 2011-07-01 15:00:08)
Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt, and
Nathan Binkert.
Summary
-------
x86: Implements copyRegs() function
The copyRegs() function for x86 is currently unimplemented. This patch
provides an implementation.
Apart from this patch, I have another question. In the function
copyMiscRegs(), a comment appears that 'it has been implemented naively.'
Why the comment? What would an accurate implementation look like?
Diffs
-----
src/arch/x86/utility.cc 559ef3da5dac
Diff: http://reviews.m5sim.org/r/759/diff
Testing
-------
Thanks,
Nilay
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev
