Re: [Qemu-devel] [PATCH v2 11/12] cpus: call cpu_update_icount on read

[Paolo Bonzini <pbonzini@redhat.com>]

----- Original Message -----
> From: "Alex Bennée" <alex.bennee@linaro.org>
> To: "Paolo Bonzini" <pbonzini@redhat.com>
> Cc: dovgaluk@ispras.ru, rth@twiddle.net, "peter maydell" <peter.maydell@linaro.org>, qemu-devel@nongnu.org,
> mttcg@greensocs.com, "fred konrad" <fred.konrad@greensocs.com>, "a rigo" <a.rigo@virtualopensystems.com>,
> cota@braap.org, "bobby prani" <bobby.prani@gmail.com>, nikunj@linux.vnet.ibm.com, "Peter Crosthwaite"
> <crosthwaite.peter@gmail.com>
> Sent: Friday, April 7, 2017 7:35:29 PM
> Subject: Re: [PATCH v2 11/12] cpus: call cpu_update_icount on read
> 
> 
> Paolo Bonzini <pbonzini@redhat.com> writes:
> 
> > On 05/04/2017 15:25, Alex Bennée wrote:
> >> This ensures each time the vCPU thread reads the icount we update the
> >> master timer_state.qemu_icount field. This way as long as updates are
> >> in BQL protected sections (which they should be) the main-loop can
> >> never come to update the log and find time has gone backwards.
> >>
> >> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
> >> ---
> >>  cpus.c | 6 ++----
> >>  1 file changed, 2 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/cpus.c b/cpus.c
> >> index ff75af449a..63de033cc8 100644
> >> --- a/cpus.c
> >> +++ b/cpus.c
> >> @@ -246,19 +246,17 @@ void cpu_update_icount(CPUState *cpu)
> >>
> >>  int64_t cpu_get_icount_raw(void)
> >>  {
> >> -    int64_t icount;
> >>      CPUState *cpu = current_cpu;
> >>
> >> -    icount = atomic_read(&timers_state.qemu_icount);
> >>      if (cpu && cpu->running) {
> >>          if (!cpu->can_do_io) {
> >>              fprintf(stderr, "Bad icount read\n");
> >>              exit(1);
> >>          }
> >>          /* Take into account what has run */
> >> -        icount += cpu_get_icount_executed(cpu);
> >> +        cpu_update_icount(cpu);
> >>      }
> >> -    return icount;
> >> +    return atomic_read(&timers_state.qemu_icount);
> >>  }
> >>
> >>  /* Return the virtual CPU time, based on the instruction counter.  */
> >>
> >
> > Maybe the update should be done in gen_io_start instead.  There
> > shouldn't be any interference between vCPU and I/O threads except
> > between gen_io_start and gen_io_end.
> 
> I'm not sure I follow. gen_io_start is a translation time thing. At
> least here we ensure we update whenever the value is read.

Sorry, I meant we should generate TCG opcodes for the translation in
gen_io_start.

Paolo