From: Richard Sandiford <richard.sandiford@arm.com>
To: Christophe Lyon <christophe.lyon@linaro.org>
Cc: gcc-patches@gcc.gnu.org, Kyrylo.Tkachov@arm.com
Subject: Re: [PATCH] aarch64: Add pattern for bswap + rotate [PR 110039]
Date: Wed, 31 May 2023 11:22:16 +0100 [thread overview]
Message-ID: <mpty1l44x7r.fsf@arm.com> (raw)
In-Reply-To: <CAPS5khYP_RWi_zd66KCzU0TOQ4c8tB7+HH7nEsHBKZ5=jwr_6A@mail.gmail.com> (Christophe Lyon's message of "Wed, 31 May 2023 12:08:20 +0200")
Christophe Lyon <christophe.lyon@linaro.org> writes:
> On Wed, 31 May 2023 at 11:49, Richard Sandiford <richard.sandiford@arm.com>
> wrote:
>
>> Christophe Lyon <christophe.lyon@linaro.org> writes:
>> > After commit g:d8545fb2c71683f407bfd96706103297d4d6e27b, we missed a
>> > pattern to match the new GIMPLE form.
>> >
>> > With this patch, gcc.target/aarch64/rev16_2.c passes again.
>> >
>> > 2023-05-31 Christophe Lyon <christophe.lyon@linaro.org>
>> >
>> > PR target/110039
>> > gcc/
>> > * config/aarch64/aarch64.md (aarch64_rev16<mode>2_alt3): New
>> > pattern.
>> > ---
>> > gcc/config/aarch64/aarch64.md | 10 ++++++++++
>> > 1 file changed, 10 insertions(+)
>> >
>> > diff --git a/gcc/config/aarch64/aarch64.md
>> b/gcc/config/aarch64/aarch64.md
>> > index 8b8951d7b14..663353791fd 100644
>> > --- a/gcc/config/aarch64/aarch64.md
>> > +++ b/gcc/config/aarch64/aarch64.md
>> > @@ -6267,6 +6267,16 @@
>> > [(set_attr "type" "rev")]
>> > )
>> >
>> > +;; Similar pattern to mache (rotate (bswap) 16)
>> > +(define_insn "aarch64_rev16<mode>2_alt3"
>> > + [(set (match_operand:GPI 0 "register_operand" "=r")
>> > + (rotate:GPI (bswap:GPI (match_operand:GPI 1 "register_operand"
>> "r"))
>> > + (const_int 16)))]
>> > + ""
>> > + "rev16\\t%<w>0, %<w>1"
>> > + [(set_attr "type" "rev")]
>> > +)
>> > +
>>
>> Doesn't this have to be :SI only? The rtl expression and the
>> instruction are different for :DI.
>>
> Do you mean the other two examples in the testcase?
> ( __rev16_64_alt, __rev16_64)
> They currently use aarch64_rev16di2_alt1 and aarch64_rev16di2_alt2
> respectively.
I meant more that the new pattern would generate wrong code if someone
wrote a 64-bit bswap followed by a 64-bit rotate left.
Richard
next prev parent reply other threads:[~2023-05-31 10:22 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-31 9:28 Christophe Lyon
2023-05-31 9:49 ` Richard Sandiford
2023-05-31 10:08 ` Christophe Lyon
2023-05-31 10:22 ` Richard Sandiford [this message]
2023-05-31 13:13 ` [PATCH v2] " Christophe Lyon
2023-05-31 13:15 ` Richard Sandiford
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=mpty1l44x7r.fsf@arm.com \
--to=richard.sandiford@arm.com \
--cc=Kyrylo.Tkachov@arm.com \
--cc=christophe.lyon@linaro.org \
--cc=gcc-patches@gcc.gnu.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).