From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: pbonzini@redhat.com,
"Pierrick Bouvier" <pierrick.bouvier@linaro.org>,
"Philippe Mathieu-Daudé" <philmd@linaro.org>
Subject: Re: [PATCH v2 00/27] tcg/optimize: Track and use known 1's
Date: Mon, 23 Jun 2025 11:45:33 -0700 [thread overview]
Message-ID: <182c565e-1ba3-4798-aea5-5cefbef4c70e@linaro.org> (raw)
In-Reply-To: <20250603080908.559594-1-richard.henderson@linaro.org>
Ping.
On 6/3/25 01:08, Richard Henderson wrote:
> We already track and use known 0's; do the same for 1's.
> This actually simplifies some of the logical operations,
> where before required checking for a constant value.
>
> Changes for v2:
> - Some simplifications suggested by Paolo.
>
> - Merge affected bit handling into fold_masks_zosa.
> There's an optimization ordering preference for
> 1) constant result
> 2) copy result
> 3) computed result
> Having fold_affected_mask separate performed 2 before 1.
>
> - Rearrange the and->extract optimization, because it
> would perform 3 before 1 and 2.
>
>
> r~
>
>
> Richard Henderson (27):
> tcg/optimize: Introduce arg_const_val
> tcg/optimize: Add one's mask to TempOptInfo
> tcg/optimize: Introduce fold_masks_zosa
> tcg/optimize: Build and use o_bits in fold_and
> tcg/optimize: Build and use o_bits in fold_andc
> tcg/optimize: Build and use z_bits and o_bits in fold_eqv
> tcg/optimize: Build and use z_bits and o_bits in fold_nand
> tcg/optimize: Build and use z_bits and o_bits in fold_nor
> tcg/optimize: Build and use z_bits and o_bits in fold_not
> tcg/optimize: Build and use one and affected bits in fold_or
> tcg/optimize: Build and use zero, one and affected bits in fold_orc
> tcg/optimize: Build and use o_bits in fold_xor
> tcg/optimize: Build and use o_bits in fold_bswap
> tcg/optimize: Build and use o_bits in fold_deposit
> tcg/optimize: Build and use o_bits in fold_extract
> tcg/optimize: Build and use z_bits and o_bits in fold_extract2
> tcg/optimize: Build and use o_bits in fold_exts
> tcg/optimize: Build and use o_bits in fold_extu
> tcg/optimize: Build and use o_bits in fold_movcond
> tcg/optimize: Build and use o_bits in fold_sextract
> tcg/optimize: Build and use o_bits in fold_shift
> tcg/optimize: Use fold_and in do_constant_folding_cond[12]
> tcg/optimize: Fold and to extract during optimize
> tcg/optimize: Simplify fold_and constant checks
> tcg/optimize: Simplify fold_andc constant checks
> tcg/optimize: Simplify fold_orc constant checks
> tcg/optimize: Simplify fold_eqv constant checks
>
> tcg/optimize.c | 456 +++++++++++++++++++++++++++++--------------------
> 1 file changed, 274 insertions(+), 182 deletions(-)
>
prev parent reply other threads:[~2025-06-23 18:46 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-06-03 8:08 [PATCH v2 00/27] tcg/optimize: Track and use known 1's Richard Henderson
2025-06-03 8:08 ` [PATCH v2 01/27] tcg/optimize: Introduce arg_const_val Richard Henderson
2025-06-25 14:45 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 02/27] tcg/optimize: Add one's mask to TempOptInfo Richard Henderson
2025-06-25 14:44 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 03/27] tcg/optimize: Introduce fold_masks_zosa Richard Henderson
2025-06-25 14:44 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 04/27] tcg/optimize: Build and use o_bits in fold_and Richard Henderson
2025-06-25 14:46 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 05/27] tcg/optimize: Build and use o_bits in fold_andc Richard Henderson
2025-06-25 14:46 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 06/27] tcg/optimize: Build and use z_bits and o_bits in fold_eqv Richard Henderson
2025-06-25 14:47 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 07/27] tcg/optimize: Build and use z_bits and o_bits in fold_nand Richard Henderson
2025-06-25 14:47 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 08/27] tcg/optimize: Build and use z_bits and o_bits in fold_nor Richard Henderson
2025-06-25 14:48 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 09/27] tcg/optimize: Build and use z_bits and o_bits in fold_not Richard Henderson
2025-06-25 14:48 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 10/27] tcg/optimize: Build and use one and affected bits in fold_or Richard Henderson
2025-06-25 14:48 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 11/27] tcg/optimize: Build and use zero, one and affected bits in fold_orc Richard Henderson
2025-06-25 14:49 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 12/27] tcg/optimize: Build and use o_bits in fold_xor Richard Henderson
2025-06-25 14:49 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 13/27] tcg/optimize: Build and use o_bits in fold_bswap Richard Henderson
2025-06-25 14:52 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 14/27] tcg/optimize: Build and use o_bits in fold_deposit Richard Henderson
2025-06-25 14:53 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 15/27] tcg/optimize: Build and use o_bits in fold_extract Richard Henderson
2025-06-25 14:54 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 16/27] tcg/optimize: Build and use z_bits and o_bits in fold_extract2 Richard Henderson
2025-06-25 14:54 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 17/27] tcg/optimize: Build and use o_bits in fold_exts Richard Henderson
2025-06-25 14:54 ` Pierrick Bouvier
2025-06-03 8:08 ` [PATCH v2 18/27] tcg/optimize: Build and use o_bits in fold_extu Richard Henderson
2025-06-25 14:55 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 19/27] tcg/optimize: Build and use o_bits in fold_movcond Richard Henderson
2025-06-25 14:55 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 20/27] tcg/optimize: Build and use o_bits in fold_sextract Richard Henderson
2025-06-25 14:55 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 21/27] tcg/optimize: Build and use o_bits in fold_shift Richard Henderson
2025-06-25 14:56 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 22/27] tcg/optimize: Use fold_and in do_constant_folding_cond[12] Richard Henderson
2025-06-25 14:56 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 23/27] tcg/optimize: Fold and to extract during optimize Richard Henderson
2025-06-25 14:57 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 24/27] tcg/optimize: Simplify fold_and constant checks Richard Henderson
2025-06-25 14:57 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 25/27] tcg/optimize: Simplify fold_andc " Richard Henderson
2025-06-25 14:57 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 26/27] tcg/optimize: Simplify fold_orc " Richard Henderson
2025-06-25 14:57 ` Pierrick Bouvier
2025-06-03 8:09 ` [PATCH v2 27/27] tcg/optimize: Simplify fold_eqv " Richard Henderson
2025-06-25 14:57 ` Pierrick Bouvier
2025-06-23 18:45 ` Richard Henderson [this message]
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=182c565e-1ba3-4798-aea5-5cefbef4c70e@linaro.org \
--to=richard.henderson@linaro.org \
--cc=pbonzini@redhat.com \
--cc=philmd@linaro.org \
--cc=pierrick.bouvier@linaro.org \
--cc=qemu-devel@nongnu.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 NNTP newsgroup(s).