Received-SPF: pass (google.com: domain of linux-kernel+bounces-157656-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Date: Wed, 24 Apr 2024 13:06:44 -0700
In-Reply-To: <ZgyBckwbrijACeB1@google.com>
Precedence: bulk
Mime-Version: 1.0
References: <20240309012725.1409949-1-seanjc@google.com> <20240309012725.1409949-9-seanjc@google.com>
 <ZfRtSKcXTI/lAQxE@intel.com> <ZfSLRrf1CtJEGZw2@google.com>
 <1e063b73-0f9a-4956-9634-2552e6e63ee1@intel.com> <ZgyBckwbrijACeB1@google.com>
Message-ID: <ZilmVN0gbFlpnHO9@google.com>
Subject: Re: [PATCH v6 8/9] KVM: VMX: Open code VMX preemption timer rate mask
 in its accessor
From: Sean Christopherson <seanjc@google.com>
To: Xiaoyao Li <xiaoyao.li@intel.com>
Cc: Zhao Liu <zhao1.liu@intel.com>, Thomas Gleixner <tglx@linutronix.de>, 
	Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>, 
	Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org, 
	Paolo Bonzini <pbonzini@redhat.com>, Andy Lutomirski <luto@kernel.org>, 
	Peter Zijlstra <peterz@infradead.org>, linux-kernel@vger.kernel.org, kvm@vger.kernel.org, 
	Shan Kang <shan.kang@intel.com>, Kai Huang <kai.huang@intel.com>, Xin Li <xin3.li@intel.com>
Content-Type: text/plain; charset="us-ascii"

On Tue, Apr 02, 2024, Sean Christopherson wrote:
> On Mon, Apr 01, 2024, Xiaoyao Li wrote:
> > On 3/16/2024 1:54 AM, Sean Christopherson wrote:
> > > On Fri, Mar 15, 2024, Zhao Liu wrote:
> > > > On Fri, Mar 08, 2024 at 05:27:24PM -0800, Sean Christopherson wrote:
> > > > > Use vmx_misc_preemption_timer_rate() to get the rate in hardware_setup(),
> > > > > and open code the rate's bitmask in vmx_misc_preemption_timer_rate() so
> > > > > that the function looks like all the helpers that grab values from
> > > > > VMX_BASIC and VMX_MISC MSR values.
> > > 
> > > ...
> > > 
> > > > > -#define VMX_MISC_PREEMPTION_TIMER_RATE_MASK	GENMASK_ULL(4, 0)
> > > > >   #define VMX_MISC_SAVE_EFER_LMA			BIT_ULL(5)
> > > > >   #define VMX_MISC_ACTIVITY_HLT			BIT_ULL(6)
> > > > >   #define VMX_MISC_ACTIVITY_SHUTDOWN		BIT_ULL(7)
> > > > > @@ -162,7 +161,7 @@ static inline u32 vmx_basic_vmcs_mem_type(u64 vmx_basic)
> > > > >   static inline int vmx_misc_preemption_timer_rate(u64 vmx_misc)
> > > > >   {
> > > > > -	return vmx_misc & VMX_MISC_PREEMPTION_TIMER_RATE_MASK;
> > > > > +	return vmx_misc & GENMASK_ULL(4, 0);
> > > > >   }
> > > > 
> > > > I feel keeping VMX_MISC_PREEMPTION_TIMER_RATE_MASK is clearer than
> > > > GENMASK_ULL(4, 0), and the former improves code readability.
> > > > 
> > > > May not need to drop VMX_MISC_PREEMPTION_TIMER_RATE_MASK?
> > > 
> > > I don't necessarily disagree, but in this case I value consistency over one
> > > individual case.  As called out in the changelog, the motivation is to make
> > > vmx_misc_preemption_timer_rate() look like all the surrounding helpers.
> > > 
> > > _If_ we want to preserve the mask, then we should add #defines for vmx_misc_cr3_count(),
> > > vmx_misc_max_msr(), etc.
> > > 
> > > I don't have a super strong preference, though I think my vote would be to not
> > > add the masks and go with this patch.  These helpers are intended to be the _only_
> > > way to access the fields, i.e. they effectively _are_ the mask macros, just in
> > > function form.
> > > 
> > 
> > +1.
> > 
> > However, it seems different for vmx_basic_vmcs_mem_type() in patch 5, that I
> > just recommended to define the MASK.
> > 
> > Because we already have
> > 
> > 	#define VMX_BASIC_MEM_TYPE_SHIFT	50
> > 
> > and it has been used in vmx/nested.c,
> > 
> > static inline u32 vmx_basic_vmcs_mem_type(u64 vmx_basic)
> > {
> > 	return (vmx_basic & GENMASK_ULL(53, 50)) >>
> > 		VMX_BASIC_MEM_TYPE_SHIFT;
> > }
> > 
> > looks not intuitive than original patch.
> 
> Yeah, agreed, that's taking the worst of both worlds.  I'll update patch 5 to drop
> VMX_BASIC_MEM_TYPE_SHIFT when effectively "moving" it into vmx_basic_vmcs_mem_type().

Drat.  Finally getting back to this, dropping VMX_BASIC_MEM_TYPE_SHIFT doesn't
work because it's used by nested_vmx_setup_basic(), as is VMX_BASIC_VMCS_SIZE_SHIFT,
which is presumably why past me kept them around.

I'm leaning towards keeping things as proposed in this series.  I don't see us
gaining a third copy, or even a third user, i.e. I don't think we are creating a
future problem by open coding the shift in vmx_basic_vmcs_mem_type().  And IMO
code like this

	return (vmx_basic & VMX_BASIC_MEM_TYPE_MASK) >>
	       VMX_BASIC_MEM_TYPE_SHIFT;

is an unnecessary obfuscation when there is literally one user (the accessor).

Another idea would be to delete VMX_BASIC_MEM_TYPE_SHIFT and VMX_BASIC_VMCS_SIZE_SHIFT,
and either open code the values or use local const variables, but that also seems
like a net negative, e.g. splits the effective definitions over too many locations.