Re: [PATCH] futex: Optimise the size check get_futex_key()

From: Peter Zijlstra <peterz@infradead.org>
To: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: "K Prateek Nayak" <kprateek.nayak@amd.com>,
	"Arnd Bergmann" <arnd@arndb.de>,
	"Thomas Gleixner" <tglx@kernel.org>,
	"Ingo Molnar" <mingo@redhat.com>,
	"Borislav Petkov" <bp@alien8.de>,
	"Dave Hansen" <dave.hansen@linux.intel.com>,
	x86@kernel.org, "Catalin Marinas" <catalin.marinas@arm.com>,
	"Will Deacon" <will@kernel.org>, "Paul Walmsley" <pjw@kernel.org>,
	"Palmer Dabbelt" <palmer@dabbelt.com>,
	"Albert Ou" <aou@eecs.berkeley.edu>,
	"Heiko Carstens" <hca@linux.ibm.com>,
	"Vasily Gorbik" <gor@linux.ibm.com>,
	"Alexander Gordeev" <agordeev@linux.ibm.com>,
	"Darren Hart" <dvhart@infradead.org>,
	"Davidlohr Bueso" <dave@stgolabs.net>,
	"André Almeida" <andrealmeid@igalia.com>,
	linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org,
	"Samuel Holland" <samuel.holland@sifive.com>,
	"Charlie Jenkins" <thecharlesjenkins@gmail.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-riscv@lists.infradead.org, linux-s390@vger.kernel.org,
	"H. Peter Anvin" <hpa@zytor.com>,
	"Thomas Huth" <thuth@redhat.com>,
	"Sean Christopherson" <seanjc@google.com>,
	"Jisheng Zhang" <jszhang@kernel.org>,
	"Alexandre Ghiti" <alex@ghiti.fr>,
	"Christian Borntraeger" <borntraeger@linux.ibm.com>,
	"Sven Schnelle" <svens@linux.ibm.com>
Subject: Re: [PATCH] futex: Optimise the size check get_futex_key()
Date: Thu, 2 Jul 2026 10:59:21 +0200	[thread overview]
Message-ID: <20260702085921.GD49951@noisy.programming.kicks-ass.net> (raw)
In-Reply-To: <20260701161736.xYYizA0e@linutronix.de>

On Wed, Jul 01, 2026 at 06:17:36PM +0200, Sebastian Andrzej Siewior wrote:
> The futex address must be naturally aligned and this is checked via
> "address % size" where `address' is the supplied address and `size' is
> the expected size of futex. It is guaranteed that `size' is power of two
> but the compiler does not see it and creates here a `div' operation
> (x86, arm, gcc-15).

That's pretty stupid, since size = futex_size() := 1 << (blah); so it
has every opportunity to actually see this. Also, see below, clang does!

> We can take advantage of the pow2 property and rewrite it as
> "address & (size-1)".
> 
> As per testing, the command
> |perf bench futex hash -f 1 -b 16384 -t 1 -r 30
> 
> improved from
> | [thread  0] futex: 0x5619f931f740 [ 7001583 ops/sec ]
> to
> | [thread  0] futex: 0x55da173e5740 [ 7376137 ops/sec ]
> 
> or by 5.3%
> 
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
> 
> Could someone verify this, please? The 5% look a bit high. This is on
> top of the series (but not worsen by the series).

Bah, I tried to reproduce and couldn't. Then I noticed I did a clang
build and that is in fact clever enough to do this optimization itself.

/me tries again with a GCC build.

pre:  [thread  0] futex: 0x561f14430680 [ 9021408 ops/sec ]
post: [thread  0] futex: 0x55feadbbb680 [ 8977527 ops/sec ]

(and this seems to be well inside the error threshold of this test).

So I see the GCC build do the DIV, and no longer with his patch applied,
but for some reason I cannot get the runtime performance to actually
improve much of anything on my system.