Re: [Qemu-devel] [PATCH] target-arm: Minimal implementation of performance counters

[Aurelien Jarno <aurelien@aurel32.net>]

On Mon, May 16, 2011 at 05:37:03PM +0100, Peter Maydell wrote:
> On 16 May 2011 17:10, Paul Brook <paul@codesourcery.com> wrote:
> >> > I just spoke with Paul on IRC about this. In summary:
> >> >  * for a helper to cause an exception then it has (a) to make sure CPU
> >> >
> >> > state (pc, condflags) is sync'd before the call to the helper and (b)
> >> > the helper has to be in a file with access to global env, because it
> >> > needs to call cpu_loop_exit()
> >>
> >> I don't think (a) is true. It is possible to use the same way as for
> >> load/store operations, that is call cpu_restore_state() before calling
> >> cpu_loop_exit().
> >
> > To call cpu_restore_state you need to know searched_pc.  To find that you need
> > to unwind the host stack all the way back to translated code.
> 
> You can do this by calling GETPC() from the top level helper function
> though, right? [OK, we'd need to move the definition out of dyngen-exec.h.]

No we don't need to move it out of dyngen-exec.h. dyngen-exec.h is
included in target-*/exec.h, as the softmmu helpers, which are included
in target-*/op_helper.c, call cpu_restore_state().

For an actual usage of cpu_restore_state() outside of the softmmu
helpers, you can have a look at target-sh4/op-helper.c, which uses this
technique for raising most exceptions, and especially the FPU ones.

-- 
Aurelien Jarno                          GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net