Re: [PATCH v1] KVM: VMX: Remove unnecessary parentheses

[Sean Christopherson <seanjc@google.com>]

On Fri, Mar 06, 2026, BillXiang wrote:
> On 3/6/2026 7:54 AM, Xin Li (Intel) wrote:
> > From: Xin Li <xin@zytor.com>
> > 
> > Drop redundant parentheses; & takes precedence over &&.
> 
> I would not recommend relying on default operator precedence.

Eh, in practice, the kernel heavily relies on operator precedence all over the
place, so I'm not worried about correctness.  But as you note below, judicious
use of parentheses can help readability.

> > Signed-off-by: Xin Li <xin@zytor.com>
> > ---
> >   arch/x86/kvm/vmx/capabilities.h | 10 +++++-----
> >   1 file changed, 5 insertions(+), 5 deletions(-)
> > 
> > diff --git a/arch/x86/kvm/vmx/capabilities.h b/arch/x86/kvm/vmx/capabilities.h
> > index 4e371c93ae16..0dad9e7c4ff4 100644
> > --- a/arch/x86/kvm/vmx/capabilities.h
> > +++ b/arch/x86/kvm/vmx/capabilities.h
> > @@ -107,7 +107,7 @@ static inline bool cpu_has_load_perf_global_ctrl(void)
> >   
> >   static inline bool cpu_has_load_cet_ctrl(void)
> >   {
> > -	return (vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_CET_STATE);
> > +	return vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_CET_STATE;
> >   }
> >   
> >   static inline bool cpu_has_save_perf_global_ctrl(void)
> > @@ -162,7 +162,7 @@ static inline bool cpu_has_vmx_ept(void)
> >   static inline bool vmx_umip_emulated(void)
> >   {
> >   	return !boot_cpu_has(X86_FEATURE_UMIP) &&
> > -	       (vmcs_config.cpu_based_2nd_exec_ctrl & SECONDARY_EXEC_DESC);
> > +	       vmcs_config.cpu_based_2nd_exec_ctrl & SECONDARY_EXEC_DESC;
> >   }
> >   
> >   static inline bool cpu_has_vmx_rdtscp(void)
> > @@ -376,9 +376,9 @@ static inline bool cpu_has_vmx_invvpid_global(void)
> >   
> >   static inline bool cpu_has_vmx_intel_pt(void)
> >   {
> > -	return (vmcs_config.misc & VMX_MISC_INTEL_PT) &&
> > -		(vmcs_config.cpu_based_2nd_exec_ctrl & SECONDARY_EXEC_PT_USE_GPA) &&
> > -		(vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_IA32_RTIT_CTL);
> > +	return vmcs_config.misc & VMX_MISC_INTEL_PT &&
> > +	       vmcs_config.cpu_based_2nd_exec_ctrl & SECONDARY_EXEC_PT_USE_GPA &&
> > +	       vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_IA32_RTIT_CTL;
> >   }
> 
> Removing the parentheses could significantly reduce code readability here.

+1.

For cases like cpu_has_load_cet_ctrl() where it's a single statement, I'm 100%
in favor of dropping the parentheses because they're pure noise.  E.g. putting
them in the return statement is kinda like doing this:

	if ((vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_CET_STATE))

But for code where there are multiple checks, IMO it's totally fine to use
parentheses, and probably even encourage.

Can you send a v2 to just clean up cpu_has_load_cet_ctrl()?  Ah, and the changelog
will need to be different, because there's no &&, which is another argument for
treating "return x & y;" differently.