qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
* [Qemu-devel]  qemu icount mode timer accuracy
@ 2019-08-19 21:14 Wu, Wentong
  0 siblings, 0 replies; 3+ messages in thread
From: Wu, Wentong @ 2019-08-19 21:14 UTC (permalink / raw)
  To: Wu, Wentong, qemu-devel


Could anyone please give some comments? Thanks in advance!



Hi,

Recently I'm working to enable Qemu icount mode with TCG, with source code review I found that Qemu can give deterministic execution for guest code timeout. But for exact time point for guest OS, I have a question:

For armv7m_systick.c example, guest OS will use systick_read which will call "t = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); " to calculate his exact time point, and qemu_clock_get_ns will use qemu_icount. But from qemu_tcg_rr_cpu_thread_fn { prepare_icount_for_run(cpu); r = tcg_cpu_exec(cpu); process_icount_data(cpu);}, we know qemu just update qemu_icount value after tcg_cpu_exec, so for each tcg_cpu_exec execution there is the same qemu_icount value, and then guest code will get the same time point for that one tcg execution. Can someone confirm that?

^ permalink raw reply	[flat|nested] 3+ messages in thread

* [Qemu-devel]  qemu icount mode timer accuracy
@ 2019-08-19 20:49 Wu, Wentong
  0 siblings, 0 replies; 3+ messages in thread
From: Wu, Wentong @ 2019-08-19 20:49 UTC (permalink / raw)
  To: Wu, Wentong, qemu-devel


Could you please give some comments about this? Thanks a lot!


^ permalink raw reply	[flat|nested] 3+ messages in thread

* [Qemu-devel] qemu icount mode timer accuracy
@ 2019-08-09  4:00 Wu, Wentong
  0 siblings, 0 replies; 3+ messages in thread
From: Wu, Wentong @ 2019-08-09  4:00 UTC (permalink / raw)
  To: qemu-devel


Hi,

Recently I'm working to enable Qemu icount mode with TCG, with source code review I found that Qemu can give deterministic execution for guest code timeout. But for exact time point for guest OS, I have a question:

For armv7m_systick.c example, guest OS will use systick_read which will call "t = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); " to calculate his exact time point, and qemu_clock_get_ns will use qemu_icount. But from qemu_tcg_rr_cpu_thread_fn { prepare_icount_for_run(cpu); r = tcg_cpu_exec(cpu); process_icount_data(cpu);}, we know qemu just update qemu_icount value after tcg_cpu_exec, so for each tcg_cpu_exec execution there is the same qemu_icount value, and then guest code will get the same time point for that one tcg execution. Can someone confirm that?



Thanks

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2019-08-19 21:15 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-08-19 21:14 [Qemu-devel] qemu icount mode timer accuracy Wu, Wentong
  -- strict thread matches above, loose matches on Subject: below --
2019-08-19 20:49 Wu, Wentong
2019-08-09  4:00 Wu, Wentong

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).