Date:   Thu, 26 Jan 2023 22:01:41 +0000
From:   Sean Christopherson <seanjc@google.com>
To:     "Huang, Kai" <kai.huang@intel.com>
Cc:     "zhi.wang.linux@gmail.com" <zhi.wang.linux@gmail.com>,
        "sean.j.christopherson@intel.com" <sean.j.christopherson@intel.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "Shahar, Sagi" <sagis@google.com>,
        "isaku.yamahata@gmail.com" <isaku.yamahata@gmail.com>,
        "Aktas, Erdem" <erdemaktas@google.com>,
        "dmatlack@google.com" <dmatlack@google.com>,
        "kvm@vger.kernel.org" <kvm@vger.kernel.org>,
        "Yamahata, Isaku" <isaku.yamahata@intel.com>,
        "pbonzini@redhat.com" <pbonzini@redhat.com>
Subject: Re: [PATCH v11 030/113] KVM: x86/mmu: Replace hardcoded value 0 for
 the initial value for SPTE
Message-ID: <Y9L4RQXuTJ4RTVcF@google.com>
References: <cover.1673539699.git.isaku.yamahata@intel.com>
 <dee30f0562d8be0102547d8eb9fc77736eae679d.1673539699.git.isaku.yamahata@intel.com>
 <20230125112434.0000512a@gmail.com>
 <Y9Fj/vgPEzfU1eof@google.com>
 <0be55c001aa1a538a02055aa244c655262228ce4.camel@intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <0be55c001aa1a538a02055aa244c655262228ce4.camel@intel.com>
Precedence: bulk

On Thu, Jan 26, 2023, Huang, Kai wrote:
> On Wed, 2023-01-25 at 17:22 +0000, Sean Christopherson wrote:
> > I agree that handling this in the common code would be cleaner, but repurposing
> > gfp_zero gets kludgy because it would require a magic value to say "don't initialize
> > the data", e.g. x86's mmu_shadowed_info_cache isn't pre-filled.

...

> > @@ -400,6 +405,13 @@ int __kvm_mmu_topup_memory_cache(struct kvm_mmu_memory_cache *mc, int capacity,
> >  		if (WARN_ON_ONCE(!capacity))
> >  			return -EIO;
> >  
> > +		/*
> > +		 * Custom init values can be used only for page allocations,
> > +		 * and obviously conflict with __GFP_ZERO.
> > +		 */
> > +		if (WARN_ON_ONCE(mc->init_value && (mc->kmem_cache || mc->gfp_zero)))
> > +			return -EIO;
> > +
> >  		mc->objects = kvmalloc_array(sizeof(void *), capacity, gfp);
> >  		if (!mc->objects)
> >  			return -ENOMEM;
> > 
> > base-commit: 503f0315c97739d3f8e645c500d81757dfbf76be
> 
> init_value and gfp_zone is kinda redundant.  How about removing gfp_zero
> completely?
> 
> 	mmu_memory_cache_alloc_obj(...)
> 	{
> 		...
> 		if (!mc->init_value)
> 			gfp_flags |= __GFP_ZERO;
> 		...
> 	}
> 
> And in kvm_mmu_create() you initialize all caches' init_value explicitly.

No, as mentioned above there's also a "don't initialize the data" case.  Leaving
init_value=0 means those users would see unnecessary zeroing, and again I don't
want to use a magic value to say "don't initialize".