Re: [PATCH] x86/msr: add idle version of wrmsr trace

[Peter Zijlstra]

Dear FirstName LastName, pleae fix your MUA.

On Tue, Aug 23, 2022 at 04:43:53PM -0700, FirstName LastName wrote:
> From: Peter Shier <pshier@xxxxxxxxxx>
> 
> With commit bf5835bcdb963 ("intel_idle: Disable IBRS during long idle"),
> enabling wrmsr trace with CONFIG_LOCKDEP causes "suspicious
> rcu_dereference_check() usage" warning because do_trace_write_msr does not
> use trace_write_msr_rcuidle.
> 

No, the right thing here is to not do tracing at all.

*sigh* I should go finish this series:

  https://lore.kernel.org/lkml/20220608142723.103523089@xxxxxxxxxxxxx/

---
diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
index 3e101719689a..6e82b2df29cb 100644
--- a/drivers/idle/intel_idle.c
+++ b/drivers/idle/intel_idle.c
@@ -187,12 +187,12 @@ static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev,
 	int ret;
 
 	if (smt_active)
-		wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+		__native_wrmsr(MSR_IA32_SPEC_CTRL, 0);
 
 	ret = __intel_idle(dev, drv, index);
 
 	if (smt_active)
-		wrmsrl(MSR_IA32_SPEC_CTRL, spec_ctrl);
+		__native_wrmsr(MSR_IA32_SPEC_CTRL, spec_ctrl);
 
 	return ret;
 }