Message-ID: <ba4420b1-ea8c-d75d-2c4a-f0e7a6d01c6b@intel.com>
Date:   Wed, 25 Jan 2023 08:15:16 -0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.6.1
Subject: Re: [PATCH] x86: enable Data Operand Independent Timing Mode
Content-Language: en-US
To:     Eric Biggers <ebiggers@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        "H . Peter Anvin" <hpa@zytor.com>, x86@kernel.org
Cc:     linux-crypto@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-hardening@vger.kernel.org,
        Peter Zijlstra <peterz@infradead.org>,
        Roxana Bradescu <roxabee@chromium.org>,
        Adam Langley <agl@google.com>,
        Ard Biesheuvel <ardb@kernel.org>,
        "Jason A . Donenfeld" <Jason@zx2c4.com>
References: <20230125012801.362496-1-ebiggers@kernel.org>
 <14506678-918f-81e1-2c26-2b347ff50701@intel.com>
From:   Dave Hansen <dave.hansen@intel.com>
In-Reply-To: <14506678-918f-81e1-2c26-2b347ff50701@intel.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 1/25/23 07:29, Dave Hansen wrote:
> There's another part here which I think was recently added to the
> documentation:
> 
> 	Intel expects the performance impact of this mode may be
> 	significantly higher on future processors. 
> 
> That's _meant_ to be really scary and keep folks from turning this on by
> default, aka. what this patch does.  Your new CPU will be really slow if
> you turn this on!  Boo!

*If* we go forward with this patch's approach in the kernel, I think we
should at least consider what the kernel will do in a future where there
are two classes of systems:

	1. Ice Lake era ones where DOITM=1 is cheap enough that it can
	   on by default.
	2. "Future processors" where DOITM=1 by default is too costly.

Maybe for #2 we set DOITM=0 in the kernel.  Maybe we add per-task
controls.

But, there *is* DOITM cost where the large fleets are going to be
tempted to turn it off somehow, somewhere.  The kernel will be better
off if we can design that in now.