All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Laight <david.laight.linux@gmail.com>
To: Mark Rutland <mark.rutland@arm.com>
Cc: vladimir.murzin@arm.com, peterz@infradead.org,
	catalin.marinas@arm.com, hca@linux.ibm.com,
	linux-kernel@vger.kernel.org, ruanjinjie@huawei.com,
	yang@os.amperecomputing.com, maz@kernel.org, will@kernel.org,
	ardb@kernel.org, linux-arm-kernel@lists.infradead.org
Subject: Re: [RFC PATCH 07/13] arm64: percpu: Implement preemptible read/write ops
Date: Tue, 28 Jul 2026 23:05:04 +0100	[thread overview]
Message-ID: <20260728230504.41e2861a@pumpkin> (raw)
In-Reply-To: <20260728123859.2911495-8-mark.rutland@arm.com>

On Tue, 28 Jul 2026 13:38:53 +0100
Mark Rutland <mark.rutland@arm.com> wrote:

> Use the PCPU GPR infrastructure to implement preemptible this_cpu_read()
> and this_cpu_write().
> 
> Note that this change means that this_cpu_read() will always use a plain
> LDR, even in LTO configurations where READ_ONCE() will use LDA[P]R. I
> am not sure whether that's ok, or whether we should similarly use
> LDA[P]R here.
> 
...
> Generated code after this patch:
> 
> | <outline_this_cpu_write_u64>:
> |        mrs     x2, sp_el0
> |        mov     w3, #0x7c60
> |        strh    w3, [x2, #20]
> |        mrs     x3, tpidr_el1
> |        str     x1, [x0, x3]
> |        strh    wzr, [x2, #20]
> |        ret

How is that likely to get scheduled on a typical arm64 cpu?
In particular does the reuse of [wx]3 create a register dependency
between the two instruction flows and stop then running in parallel.
A typical x86 cpu (that anyone cares about) does register renaming
so it wouldn't matter, but I don't know about arm.
So it might be worth using different register for w3 and x3 (although
that might cause a spill - which will be worse).

	David


WARNING: multiple messages have this Message-ID (diff)
From: David Laight <david.laight.linux@gmail.com>
To: Mark Rutland <mark.rutland@arm.com>
Cc: linux-arm-kernel@lists.infradead.org, ada.coupriediaz@arm.com,
	ardb@kernel.org, catalin.marinas@arm.com, hca@linux.ibm.com,
	linux-kernel@vger.kernel.org, maz@kernel.org,
	peterz@infradead.org, ruanjinjie@huawei.com,
	vladimir.murzin@arm.com, will@kernel.org,
	yang@os.amperecomputing.com
Subject: Re: [RFC PATCH 07/13] arm64: percpu: Implement preemptible read/write ops
Date: Tue, 28 Jul 2026 23:05:04 +0100	[thread overview]
Message-ID: <20260728230504.41e2861a@pumpkin> (raw)
In-Reply-To: <20260728123859.2911495-8-mark.rutland@arm.com>

On Tue, 28 Jul 2026 13:38:53 +0100
Mark Rutland <mark.rutland@arm.com> wrote:

> Use the PCPU GPR infrastructure to implement preemptible this_cpu_read()
> and this_cpu_write().
> 
> Note that this change means that this_cpu_read() will always use a plain
> LDR, even in LTO configurations where READ_ONCE() will use LDA[P]R. I
> am not sure whether that's ok, or whether we should similarly use
> LDA[P]R here.
> 
...
> Generated code after this patch:
> 
> | <outline_this_cpu_write_u64>:
> |        mrs     x2, sp_el0
> |        mov     w3, #0x7c60
> |        strh    w3, [x2, #20]
> |        mrs     x3, tpidr_el1
> |        str     x1, [x0, x3]
> |        strh    wzr, [x2, #20]
> |        ret

How is that likely to get scheduled on a typical arm64 cpu?
In particular does the reuse of [wx]3 create a register dependency
between the two instruction flows and stop then running in parallel.
A typical x86 cpu (that anyone cares about) does register renaming
so it wouldn't matter, but I don't know about arm.
So it might be worth using different register for w3 and x3 (although
that might cause a spill - which will be worse).

	David

  parent reply	other threads:[~2026-07-28 22:05 UTC|newest]

Thread overview: 70+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-28 12:38 [RFC PATCH 00/13] arm64: Preemptible this_cpu_*() operations Mark Rutland
2026-07-28 12:38 ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 01/13] arm64: preempt: Simplify and optimize __preempt_count_dec_and_test() Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-29  3:56   ` Jinjie Ruan
2026-07-29  3:56     ` Jinjie Ruan
2026-07-28 12:38 ` [RFC PATCH 02/13] arm64: preempt: Treat should_resched() as unlikely Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-29  4:00   ` Jinjie Ruan
2026-07-29  4:00     ` Jinjie Ruan
2026-07-28 12:38 ` [RFC PATCH 03/13] arm64: ptrace: Always inline pt_regs_[read,write}_reg() Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-29  4:02   ` Jinjie Ruan
2026-07-29  4:02     ` Jinjie Ruan
2026-07-28 12:38 ` [RFC PATCH 04/13] arm64: percpu: Factor out percpu offset asm Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-29  4:03   ` Jinjie Ruan
2026-07-29  4:03     ` Jinjie Ruan
2026-07-28 12:38 ` [RFC PATCH 05/13] arm64: gpr-num: Add wxN aliases for wN registers Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 06/13] arm64: percpu: Add infrastructure for preemptible this_cpu_*() ops Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 14:19   ` Peter Zijlstra
2026-07-28 14:19     ` Peter Zijlstra
2026-07-28 14:30     ` Peter Zijlstra
2026-07-28 14:30       ` Peter Zijlstra
2026-07-28 16:03       ` Mark Rutland
2026-07-28 16:03         ` Mark Rutland
2026-07-28 16:49         ` Peter Zijlstra
2026-07-28 16:49           ` Peter Zijlstra
2026-07-28 15:53     ` Mark Rutland
2026-07-28 15:53       ` Mark Rutland
2026-07-28 16:46       ` Peter Zijlstra
2026-07-28 16:46         ` Peter Zijlstra
2026-07-28 16:49       ` Peter Zijlstra
2026-07-28 16:49         ` Peter Zijlstra
2026-08-04 13:23         ` Mark Rutland
2026-08-04 13:23           ` Mark Rutland
2026-07-28 16:51   ` Heiko Carstens
2026-07-28 16:51     ` Heiko Carstens
2026-07-29  8:16     ` Heiko Carstens
2026-07-29  8:16       ` Heiko Carstens
2026-07-29 14:25   ` Vladimir Murzin
2026-07-29 14:25     ` Vladimir Murzin
2026-07-30 10:59     ` Mark Rutland
2026-07-30 10:59       ` Mark Rutland
2026-07-30 12:04       ` Vladimir Murzin
2026-07-30 12:04         ` Vladimir Murzin
2026-07-28 12:38 ` [RFC PATCH 07/13] arm64: percpu: Implement preemptible read/write ops Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 21:47   ` David Laight
2026-07-28 21:47     ` David Laight
2026-08-04 15:00     ` Mark Rutland
2026-08-04 15:00       ` Mark Rutland
2026-07-28 22:05   ` David Laight [this message]
2026-07-28 22:05     ` David Laight
2026-08-04 15:05     ` Mark Rutland
2026-08-04 15:05       ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 08/13] arm64: percpu: Implement preemptible void RMW ops Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 09/13] arm64: percpu: Implement preemptible return " Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 10/13] arm64: percpu: Implement preemptible XCHG ops Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 11/13] arm64: percpu: Implement preemptible CMPXCHG ops Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 12/13] arm64: percpu: Implement preemptible CMPXCHG128 ops Mark Rutland
2026-07-28 12:38   ` Mark Rutland
2026-07-28 12:38 ` [RFC PATCH 13/13] arm64: percpu: Remove _pcp_protect*() wrappers Mark Rutland
2026-07-28 12:38   ` Mark Rutland

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260728230504.41e2861a@pumpkin \
    --to=david.laight.linux@gmail.com \
    --cc=ardb@kernel.org \
    --cc=catalin.marinas@arm.com \
    --cc=hca@linux.ibm.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mark.rutland@arm.com \
    --cc=maz@kernel.org \
    --cc=peterz@infradead.org \
    --cc=ruanjinjie@huawei.com \
    --cc=vladimir.murzin@arm.com \
    --cc=will@kernel.org \
    --cc=yang@os.amperecomputing.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.