All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Laight <David.Laight@ACULAB.COM>
To: "'Clément Léger'" <cleger@rivosinc.com>,
	"Conor Dooley" <conor@kernel.org>
Cc: "Paul Walmsley" <paul.walmsley@sifive.com>,
	"Palmer Dabbelt" <palmer@dabbelt.com>,
	"Albert Ou" <aou@eecs.berkeley.edu>,
	"Atish Patra" <atishp@rivosinc.com>,
	"Andrew Jones" <ajones@ventanamicro.com>,
	"Evan Green" <evan@rivosinc.com>,
	"Björn Topel" <bjorn@rivosinc.com>,
	"linux-riscv@lists.infradead.org"
	<linux-riscv@lists.infradead.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"Ron Minnich" <rminnich@gmail.com>,
	"Daniel Maslowski" <cyrevolt@googlemail.com>
Subject: RE: [PATCH 0/7] Add support to handle misaligned accesses in S-mode
Date: Wed, 4 Oct 2023 14:10:40 +0000	[thread overview]
Message-ID: <caa22904d2d7469595dcc5cb1f11be89@AcuMS.aculab.com> (raw)
In-Reply-To: <6ac16274-3816-4b46-9068-e97af66c815a@rivosinc.com>

...
> Saying that you support misaligned access does not mean that they are
> going to be efficient, just that they are supported (in fact, the uABI
> state that they may perform poorly). The compiler is actually not so
> stupid and will try to do as much aligned access as possible in what it
> generates (unless forced by some assembly, cast or whatever that can
> screw up alignment accesses). This is already the case and it will most
> probably not change.

I did a quick check.

https://godbolt.org/z/j3e9drv4e

The code generated by both clang and gcc for misaligned reads is horrid.
Gcc does a better job if the alignment is known but generates much
the same as the clang code when it isn't.

The C code is much shorter.
Even though both gcc and clang add a (different) instruction to it

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
_______________________________________________
linux-riscv mailing list
linux-riscv@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-riscv

WARNING: multiple messages have this Message-ID (diff)
From: David Laight <David.Laight@ACULAB.COM>
To: "'Clément Léger'" <cleger@rivosinc.com>,
	"Conor Dooley" <conor@kernel.org>
Cc: "Paul Walmsley" <paul.walmsley@sifive.com>,
	"Palmer Dabbelt" <palmer@dabbelt.com>,
	"Albert Ou" <aou@eecs.berkeley.edu>,
	"Atish Patra" <atishp@rivosinc.com>,
	"Andrew Jones" <ajones@ventanamicro.com>,
	"Evan Green" <evan@rivosinc.com>,
	"Björn Topel" <bjorn@rivosinc.com>,
	"linux-riscv@lists.infradead.org"
	<linux-riscv@lists.infradead.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"Ron Minnich" <rminnich@gmail.com>,
	"Daniel Maslowski" <cyrevolt@googlemail.com>
Subject: RE: [PATCH 0/7] Add support to handle misaligned accesses in S-mode
Date: Wed, 4 Oct 2023 14:10:40 +0000	[thread overview]
Message-ID: <caa22904d2d7469595dcc5cb1f11be89@AcuMS.aculab.com> (raw)
In-Reply-To: <6ac16274-3816-4b46-9068-e97af66c815a@rivosinc.com>

...
> Saying that you support misaligned access does not mean that they are
> going to be efficient, just that they are supported (in fact, the uABI
> state that they may perform poorly). The compiler is actually not so
> stupid and will try to do as much aligned access as possible in what it
> generates (unless forced by some assembly, cast or whatever that can
> screw up alignment accesses). This is already the case and it will most
> probably not change.

I did a quick check.

https://godbolt.org/z/j3e9drv4e

The code generated by both clang and gcc for misaligned reads is horrid.
Gcc does a better job if the alignment is known but generates much
the same as the clang code when it isn't.

The C code is much shorter.
Even though both gcc and clang add a (different) instruction to it

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)

  reply	other threads:[~2023-10-04 14:10 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-26 15:03 [PATCH 0/7] Add support to handle misaligned accesses in S-mode Clément Léger
2023-09-26 15:03 ` Clément Léger
2023-09-26 15:03 ` [PATCH 1/7] riscv: remove unused functions in traps_misaligned.c Clément Léger
2023-09-26 15:03   ` Clément Léger
2023-09-26 15:03 ` [PATCH 2/7] riscv: add support for misaligned handling in S-mode Clément Léger
2023-09-26 15:03   ` Clément Léger
2023-09-26 15:03 ` [PATCH 3/7] riscv: report perf event for misaligned fault Clément Léger
2023-09-26 15:03   ` Clément Léger
2023-09-26 15:03 ` [PATCH 4/7] riscv: add floating point insn support to misaligned access emulation Clément Léger
2023-09-26 15:03   ` Clément Léger
2023-09-26 15:03 ` [PATCH 5/7] riscv: add support for sysctl unaligned_enabled control Clément Léger
2023-09-26 15:03   ` Clément Léger
2023-09-26 15:03 ` [PATCH 6/7] riscv: report misaligned accesses emulation to hwprobe Clément Léger
2023-09-26 15:03   ` Clément Léger
2023-09-26 21:57   ` Evan Green
2023-09-26 21:57     ` Evan Green
2023-09-28  7:46     ` Clément Léger
2023-09-28  7:46       ` Clément Léger
2023-09-28 16:51       ` Evan Green
2023-09-28 16:51         ` Evan Green
2023-10-03  9:50         ` Atish Kumar Patra
2023-10-03  9:50           ` Atish Kumar Patra
2023-09-29  1:02   ` kernel test robot
2023-09-29  1:02     ` kernel test robot
2023-09-26 15:03 ` [PATCH 7/7] riscv: add support for PR_SET_UNALIGN and PR_GET_UNALIGN Clément Léger
2023-09-26 15:03   ` Clément Léger
2023-09-26 21:43 ` [PATCH 0/7] Add support to handle misaligned accesses in S-mode Evan Green
2023-09-26 21:43   ` Evan Green
2023-09-28  7:49   ` Clément Léger
2023-09-28  7:49     ` Clément Léger
2023-09-28 16:48     ` Evan Green
2023-09-28 16:48       ` Evan Green
2023-10-03  8:40       ` Atish Kumar Patra
2023-10-03  8:40         ` Atish Kumar Patra
2023-10-03 15:39         ` ron minnich
2023-10-03 15:39           ` ron minnich
2023-09-30  9:23 ` Conor Dooley
2023-09-30  9:23   ` Conor Dooley
2023-10-02  7:40   ` Clément Léger
2023-10-02  7:40     ` Clément Léger
2023-10-02 10:49     ` Conor Dooley
2023-10-02 10:49       ` Conor Dooley
2023-10-02 11:18       ` Clément Léger
2023-10-02 11:18         ` Clément Léger
2023-10-02 15:32         ` ron minnich
2023-10-02 15:32           ` ron minnich
2023-10-02 22:22           ` Jessica Clarke
2023-10-02 22:22             ` Jessica Clarke
2023-10-03 15:37             ` ron minnich
2023-10-03 15:37               ` ron minnich
2023-10-03  8:12           ` Clément Léger
2023-10-03  8:12             ` Clément Léger
2023-10-04  8:26     ` David Laight
2023-10-04  8:26       ` David Laight
2023-10-04 10:03       ` Clément Léger
2023-10-04 10:03         ` Clément Léger
2023-10-04 14:10         ` David Laight [this message]
2023-10-04 14:10           ` David Laight

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=caa22904d2d7469595dcc5cb1f11be89@AcuMS.aculab.com \
    --to=david.laight@aculab.com \
    --cc=ajones@ventanamicro.com \
    --cc=aou@eecs.berkeley.edu \
    --cc=atishp@rivosinc.com \
    --cc=bjorn@rivosinc.com \
    --cc=cleger@rivosinc.com \
    --cc=conor@kernel.org \
    --cc=cyrevolt@googlemail.com \
    --cc=evan@rivosinc.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=palmer@dabbelt.com \
    --cc=paul.walmsley@sifive.com \
    --cc=rminnich@gmail.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.