On Thu, Oct 24, 2019 at 9:04 AM Yin Fengwei <fengwei.yin@xxxxxxxxx> wrote:
>
> In function acpi_idle_do_entry(), an ioport access is used for
> dummy wait to guarantee hardware behavior. But it could trigger
> unnecessary VMexit if kernel is running as guest in virtualization
> environment.
>
> If it's in virtualization environment, the deeper C state enter
> operation (inb()) will trap to hypervisor. It's not needed to do
> dummy wait after the inb() call. So we could just remove the
> dummy io port access to avoid unnecessary VMexit.
>
> And keep dummy io port access to maintain timing for native
> environment.
>
> Signed-off-by: Yin Fengwei <fengwei.yin@xxxxxxxxx>
> ---
> ChangeLog:
> v3 -> v4:
> - Drop overengineered function pointer and do check whether
> we are in guest before dummy inl call.
>
> v2 -> v3:
> - Remove dummy io port access totally for virtualization env.
>
> v1 -> v2:
> - Use ndelay instead of dead loop for dummy delay.
>
> drivers/acpi/processor_idle.c | 21 +++++++++++++++------
> 1 file changed, 15 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
> index ed56c6d20b08..2ae95df2e74f 100644
> --- a/drivers/acpi/processor_idle.c
> +++ b/drivers/acpi/processor_idle.c
> @@ -642,6 +642,19 @@ static int acpi_idle_bm_check(void)
> return bm_status;
> }
>
> +static void wait_for_freeze(void)
> +{
> +#ifdef CONFIG_X86
> + /* No delay is needed if we are in guest */
> + if (boot_cpu_has(X86_FEATURE_HYPERVISOR))
> + return;
> +#endif
> + /* Dummy wait op - must do something useless after P_LVL2 read
> + because chipsets cannot guarantee that STPCLK# signal
> + gets asserted in time to freeze execution properly. */
> + inl(acpi_gbl_FADT.xpm_timer_block.address);
> +}
> +
> /**
> * acpi_idle_do_entry - enter idle state using the appropriate method
> * @cx: cstate data
> @@ -658,10 +671,7 @@ static void __cpuidle acpi_idle_do_entry(struct acpi_processor_cx *cx)
> } else {
> /* IO port based C-state */
> inb(cx->address);
> - /* Dummy wait op - must do something useless after P_LVL2 read
> - because chipsets cannot guarantee that STPCLK# signal
> - gets asserted in time to freeze execution properly. */
> - inl(acpi_gbl_FADT.xpm_timer_block.address);
> + wait_for_freeze();
> }
> }
>
> @@ -682,8 +692,7 @@ static int acpi_idle_play_dead(struct cpuidle_device *dev, int index)
> safe_halt();
> else if (cx->entry_method == ACPI_CSTATE_SYSTEMIO) {
> inb(cx->address);
> - /* See comment in acpi_idle_do_entry() */
> - inl(acpi_gbl_FADT.xpm_timer_block.address);
> + wait_for_freeze();
> } else
> return -ENODEV;
> }
> --
Applying as 5.5 material, thanks!
