Re: [PATCH v2 1/3] context_tracking: Split guest_enter/exit_irqoff

Christian Borntraeger <borntraeger@xxxxxxxxxx> · Tue, 13 Apr 2021 09:48:32 +0200






On 13.04.21 09:38, Wanpeng Li wrote:
On Tue, 13 Apr 2021 at 15:35, Christian Borntraeger
<borntraeger@xxxxxxxxxx> wrote:



On 13.04.21 09:16, Wanpeng Li wrote:
[...]

@@ -145,6 +155,13 @@ static __always_inline void guest_exit_irqoff(void)
   }

   #else
THis is the else part


+static __always_inline void context_guest_enter_irqoff(void)
+{
+     instrumentation_begin();

2nd on
+     rcu_virt_note_context_switch(smp_processor_id());
+     instrumentation_end();
2nd off
+}
+
   static __always_inline void guest_enter_irqoff(void)
   {
       /*
@@ -155,10 +172,13 @@ static __always_inline void guest_enter_irqoff(void)
       instrumentation_begin();

first on
       vtime_account_kernel(current);
       current->flags |= PF_VCPU;
-     rcu_virt_note_context_switch(smp_processor_id());
       instrumentation_end();

first off
+
+     context_guest_enter_irqoff();
here we call the 2nd on and off.

So we now do instrumentation_begin 2 times?

Similar to context_guest_enter_irqoff() ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN.

For the
ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN part
context_guest_enter_irqoff()
does not have instrumentation_begin/end.

Or did I miss anything.






