linux-kernel - Re: [PATCH v2 1/6] KVM: x86/mmu: add a new mmu zap helper to indicate memtype changes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 30 May 2023 16:51:25 -0700
From:   Sean Christopherson <seanjc@...gle.com>
To:     Yan Zhao <yan.y.zhao@...el.com>
Cc:     Chao Gao <chao.gao@...el.com>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org, pbonzini@...hat.com
Subject: Re: [PATCH v2 1/6] KVM: x86/mmu: add a new mmu zap helper to indicate
 memtype changes

On Tue, May 30, 2023, Yan Zhao wrote:
> On Thu, May 25, 2023 at 08:54:15AM -0700, Sean Christopherson wrote:
> And I combined the __kvm_mmu_honors_guest_mtrrs() into
> kvm_mmu_honors_guest_mtrrs(). Not sure if you like it :)

I would prefer to provide a separater inner helper, mainly so that the common
case callers don't need to pass %false.  I don't love passing bools, but it's
tolerable for a one-off use of an inner helper.

> +/*
> + * Returns if KVM honors guest MTRRs
> + * @override_vm_has_noncoherent_dma: Allow caller to override non-coherent DMA
> + *                                   status returned from
> + *                                   kvm_arch_has_noncoherent_dma()
> + */
> +bool kvm_mmu_honors_guest_mtrrs(struct kvm *kvm,
> +                               bool override_vm_has_noncoherent_dma)
> +{
> +       bool noncoherent_dma = override_vm_has_noncoherent_dma ? true :
> +                              kvm_arch_has_noncoherent_dma(kvm);

The "override" name is confusing, e.g. it won't be clear when it's safe/correct
for a new caller to override kvm_arch_has_noncoherent_dma().  If we go with a
single helper, I could live with:

bool kvm_mmu_honors_guest_mtrrs(struct kvm *kvm, bool stopping_noncoherent_dma)
{
	bool noncoherent_dma = stopping_noncoherent_dma ||
			       kvm_arch_has_noncoherent_dma(kvm);

	...
}

but that makes it awkward to use common code for start+stop assignment, and as
above there are three "normal" callers that would have to pass magic %false
values regardless of the name.