> From: Peter Maydell [mailto:peter.mayd...@linaro.org]
> On 18 June 2015 at 08:12, Pavel Dovgaluk <pavel.dovga...@ispras.ru> wrote:
> >> From: Aurelien Jarno [mailto:aurel...@aurel32.net]
> >> Looking at how icount work, I see it's basically a variable in the CPU
> >> state (icount_decr.u16.low), which is already accessed from the TB.
> >> Couldn't we adjust it using additional code before generating an
> >> exception, when in icount mode.
> >>
> >> For example for MIPS, we can add some code before generate_exception
> >> which use the value from s->gen_opc_icount[j] to adjust
> >> the variable icount_decr.u16.low.
> >
> > It is possible, but it will incur additional overhead, because we will
> > have to update icount every time the exception might be generated.
> > We'll have to update icount value before and after every helper call,
> > that can cause an exception:
> >
> > icount -= n
> > ...
> > instr_k
> > icount += n - k
> > helper
> > icount -= n - k
> > ...
> >
> > And this overhead will slowdown the code even if no exception occur.
> 
> Right, this is a tradeoff: in some cases it's faster to assume
> no exception and handle state resync by doing a retranslate.
> In some cases it's faster to assume there will be an exception
> and do a manual sync. Guest load/store is obviously in the
> first category. Guest doing an instruction which always takes
> an exception (like syscall insns) is in the second category.
> For other cases there's a choice. We need to support both
> approaches; obviously you can argue for any particular case
> whether it should be approach 1 or approach 2.

Syscall and non-implemented instructions are in third category - they
always take an exception. In this case the translation should be stopped
without any additional actions.

By the way, I implemented this 'third category' approach for mips
and measured the performance. It does not show any performance degradation
when compared to original unfixed version.
All other exception-generating helpers and instructions use approach 1.

Pavel Dovgalyuk


Reply via email to