Re: [PATCH v3] x86/speculation/l1tf: Add KConfig for setting the L1D cache flush mode

From: Abhishek Bhardwaj <abhishekbh@google.com>
To: Waiman Long <longman@redhat.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>,
	Doug Anderson <dianders@google.com>,
	Anthony Steinhauser <asteinhauser@google.com>,
	LKML <linux-kernel@vger.kernel.org>,
	Borislav Petkov <bp@alien8.de>, "H. Peter Anvin" <hpa@zytor.com>,
	Ingo Molnar <mingo@redhat.com>, Jim Mattson <jmattson@google.com>,
	Joerg Roedel <joro@8bytes.org>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	Mark Gross <mgross@linux.intel.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Pawan Gupta <pawan.kumar.gupta@linux.intel.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Sean Christopherson <sean.j.christopherson@intel.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Tony Luck <tony.luck@intel.com>,
	Vitaly Kuznetsov <vkuznets@redhat.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	kvm@vger.kernel.org, x86 <x86@kernel.org>
Subject: Re: [PATCH v3] x86/speculation/l1tf: Add KConfig for setting the L1D cache flush mode
Date: Wed, 8 Jul 2020 12:32:27 -0700	[thread overview]
Message-ID: <CA+noqoh90h3Mtk_3++bnMGMYyEGmUUq09R1k2WzCUbKX1jgQXQ@mail.gmail.com> (raw)
In-Reply-To: <f8e84764-1cc7-c4e5-4e4f-4b907204a374@redhat.com>

On Mon, Jul 6, 2020 at 8:14 AM Waiman Long <longman@redhat.com> wrote:
>
> On 7/5/20 5:51 PM, Abhishek Bhardwaj wrote:
>
> That is why I said a comment will have to be added to highlight this
> dependency. For instance,
>
> +/*
> + * Three of the enums are explicitly assigned as the KVM_VMENTRY_L1D_FLUSH
> + * config entry in arch/x86/kvm/Kconfig depends on these values.
> + */
>   enum vmx_l1d_flush_state {
>          VMENTER_L1D_FLUSH_AUTO,
> -       VMENTER_L1D_FLUSH_NEVER,
> -       VMENTER_L1D_FLUSH_COND,
> -       VMENTER_L1D_FLUSH_ALWAYS,
> +       VMENTER_L1D_FLUSH_NEVER = 1,
> +       VMENTER_L1D_FLUSH_COND = 2,
> +       VMENTER_L1D_FLUSH_ALWAYS = 3,
>          VMENTER_L1D_FLUSH_EPT_DISABLED,
>          VMENTER_L1D_FLUSH_NOT_REQUIRED,
>   };
>
> Of course, this is just a suggestion.
>
> I'd rather avoid this dependency. However, if people are okay with the
> CONFIG option then I am happy to oblige with whatever people agree on.
> Can a maintainer chime in ? Waiman if you're the final authority on
> this, will you accept the patch if I incorporated your suggestion ?
>
> It is fine if you think this is not what you want.
>
> BTW, I am not a maintainer. However, no maintainer will give pre-approval. At least, you have to give the reason why this patch is needed in the patch itself. Before that happens, I don't see how it will get merged upstream.

I just updated the patch with the reasoning in the commit message -
https://lkml.org/lkml/2020/7/8/1325


>
> Cheers,
> Longman


-- 
Abhishek