gcc/gcc
gcc/gcc
Watch
43
Fork
You've already forked gcc
 58
Code Issues Pull requests 58 Wiki Activity Actions

Remove the restrictions on instruction operand count #223

Open
feedable wants to merge 7 commits from feedable/gcc-TEST:operand-cnt-test into trunk
pull from: feedable/gcc-TEST:operand-cnt-test
merge into: gcc:trunk
Conversation 10 Commits 7 Files changed 33 +894 -531
feedable commented 2026-08-20 20:36:45 +00:00
Member
Copy link

Changes since v1:

  • Formatting changes in vec.h.
  • Supplemental table handling fixes in ira.cc.
  • Increase in reload limit in lra-constraints.cc.

This patch series adjusts the RTL infrastructure so that it can recognize
and handle an arbitrary amount of instruction operands.

This is mostly done by replacing the usages of C-style arrays with a fixed
bound in favor of SOO-vectors and array_views.
MAX_RECOG_OPERANDS is retained for compatibility, but now it mostly
serves as a hint for inline buffer size for SOO-vectors.

The only observable effect for most targets is the ability to express
asm blocks with more than 30 operands without issues, but this PR is
also a prerequisite for register allocation for targets with instructions
that have an arbitrary operand count.

The patchset is bootstrapped and tested on x86_64. I've done some
benchmarking, and the patch seems to cause a ~4% slowdown across
the testsuite.

CC: Richard Biener rguenther@suse.de, Richard Sandiford rdsandiford@googlemail.com, Jeffrey Law jeffrey.law@oss.qualcomm.com, Vladimir Makarov vmakarov@redhat.com, Robin Dapp rdapp.gcc@gmail.com

Changes since v1: - Formatting changes in `vec.h`. - Supplemental table handling fixes in `ira.cc`. - Increase in reload limit in `lra-constraints.cc`. This patch series adjusts the RTL infrastructure so that it can recognize and handle an arbitrary amount of instruction operands. This is mostly done by replacing the usages of C-style arrays with a fixed bound in favor of SOO-vectors and `array_view`s. `MAX_RECOG_OPERANDS` is retained for compatibility, but now it mostly serves as a hint for inline buffer size for SOO-vectors. The only observable effect for most targets is the ability to express asm blocks with more than 30 operands without issues, but this PR is also a prerequisite for register allocation for targets with instructions that have an arbitrary operand count. The patchset is bootstrapped and tested on x86_64. I've done some benchmarking, and the patch seems to cause a ~4% slowdown across the testsuite. CC: Richard Biener <rguenther@suse.de>, Richard Sandiford <rdsandiford@googlemail.com>, Jeffrey Law <jeffrey.law@oss.qualcomm.com>, Vladimir Makarov <vmakarov@redhat.com>, Robin Dapp <rdapp.gcc@gmail.com>
const_derived_container seems to accept its arguments by-value, and then
tries to forward them, which doesn't make sense. Accept values by
universal reference instead.

gcc/ChangeLog:

	* iterator-utils.h: Pass constructor args by reference.
forge-bot commented 2026-08-20 20:36:47 +00:00
Member
Copy link

Hi @feedable. This bot helps send your PR as a patch series to the mailing list.

Next steps

  1. Write a clear PR description. It will be used as the cover letter.
  2. To CC extra people, add lines like this at the end of the PR description: CC: Reviewer1 <revi.ewer@example.com>, Reviewer2 <an.other@example.com>.
  3. To preview the emails, comment with /preview.
  4. When ready, request review in the Reviewers section. Or, if you just want to send the series, comment with /submit.
  5. Use the cover-letter links below to follow replies.
  6. For a new patch series, push your changes, update the PR description with a Changes since vX section, and comment with /submit again.

See Submitting Patches and UsingForge.

Consider joining the gcc and gcc-patches mailing lists..
For real time communication, check the gcc irc channels.

Or join #overseers on Libera Chat, particularly if this automation is not working (stay online to get replies, IRC does not save messages if people are not online).

<!-- pr-welcome --> Hi @feedable. This bot helps send your PR as a patch series to the mailing list. ## Next steps 1. Write a clear PR description. It will be used as the cover letter. 2. To CC extra people, add lines like this **at the end** of the PR description: `CC: Reviewer1 <revi.ewer@example.com>, Reviewer2 <an.other@example.com>`. 3. To preview the emails, comment with `/preview`. 4. When ready, request review in the Reviewers section. Or, if you just want to send the series, comment with `/submit`. 5. Use the cover-letter links below to follow replies. 6. For a new patch series, push your changes, update the PR description with a `Changes since vX` section, and comment with `/submit` again. See [Submitting Patches](https://gcc.gnu.org/contribute.html#patches) and [UsingForge](https://gcc.gnu.org/wiki/UsingForge). Consider joining the [gcc and gcc-patches mailing lists.](https://gcc.gnu.org/lists.html). For real time communication, check the [gcc irc channels](https://gcc.gnu.org/wiki/GCConIRC). Or join [`#overseers`](https://web.libera.chat/#overseers) on Libera Chat, particularly if this automation is not working (stay online to get replies, IRC does not save messages if people are not online).
feedable force-pushed operand-cnt-test from 722672098f
Some checks failed
/ format-checks (pull_request) Failing after 1m40s
Details
to a7986cd385
Some checks failed
/ format-checks (pull_request) Failing after 1m35s
Details
2026-08-20 22:55:36 +00:00 Compare
feedable force-pushed operand-cnt-test from a7986cd385
Some checks failed
/ format-checks (pull_request) Failing after 1m35s
Details
to 2fbcaae196
Some checks failed
/ format-checks (pull_request) Failing after 1m37s
Details
2026-08-20 23:10:04 +00:00 Compare
feedable force-pushed operand-cnt-test from 2fbcaae196
Some checks failed
/ format-checks (pull_request) Failing after 1m37s
Details
to ec830f12ed
Some checks failed
/ format-checks (pull_request) Failing after 3m44s
Details
2026-08-21 10:55:37 +00:00 Compare
feedable force-pushed operand-cnt-test from ec830f12ed
Some checks failed
/ format-checks (pull_request) Failing after 3m44s
Details
to 9c54b27001
Some checks failed
/ format-checks (pull_request) Failing after 1m40s
Details
2026-08-21 11:00:37 +00:00 Compare
feedable force-pushed operand-cnt-test from 9c54b27001
Some checks failed
/ format-checks (pull_request) Failing after 1m40s
Details
to 331da0018a
Some checks failed
/ format-checks (pull_request) Failing after 2m19s
Details
2026-08-21 15:13:15 +00:00 Compare
feedable force-pushed operand-cnt-test from 331da0018a
Some checks failed
/ format-checks (pull_request) Failing after 2m19s
Details
to 7a55203f10
Some checks failed
/ format-checks (pull_request) Failing after 2m30s
Details
2026-08-21 17:21:09 +00:00 Compare
feedable force-pushed operand-cnt-test from 7a55203f10
Some checks failed
/ format-checks (pull_request) Failing after 2m30s
Details
to c0214189f7
Some checks failed
/ format-checks (pull_request) Failing after 1m34s
Details
2026-08-21 17:28:01 +00:00 Compare
feedable commented 2026-08-21 17:34:27 +00:00
Author
Member
Copy link

/submit

/submit
forge-bot commented 2026-08-21 17:35:00 +00:00
Member
Copy link

Version 1 of this pull request has been stored. It includes the following commits:

  • iterator-utils.h: Fixup const_derived_container constructor - ad88464fe4
  • vec.h: Add a convenient way to construct an invalid array_slice - 94ccda57ad
  • vec.h: Add utility functions to array_slice - 5d2323cff3
  • vec.h: Implement copy&move construction&assignment for auto_vec - b11e19daac
  • i386: Don't use operands array as storage - aec89fa152
  • rtl: Make operand count dynamic - bda6f75a1d
  • rtl: Adjust generation machinery to work with arbitrary amount of operands - c0214189f7
<!-- pr-new-version --> Version 1 of this pull request has been stored. It includes the following commits: - iterator-utils.h: Fixup const_derived_container constructor - ad88464fe42ab9e83f7e0a372cfef20aec882541 - vec.h: Add a convenient way to construct an invalid array_slice - 94ccda57ada63f503d3d4a7357ff76189fdef5ad - vec.h: Add utility functions to array_slice - 5d2323cff34aaa09d110bce2142c02bedc1c2278 - vec.h: Implement copy&move construction&assignment for auto_vec - b11e19daac158f814ed6c49bf1c68d8466ab0c3c - i386: Don't use operands array as storage - aec89fa152446460ad967ac2e1bacfc94b4bb95e - rtl: Make operand count dynamic - bda6f75a1d0f6fe09ec928cb5b32c24e4cff830b - rtl: Adjust generation machinery to work with arbitrary amount of operands - c0214189f7e69062c6ca3c4b78a405bab2fcf541
forge-bot commented 2026-08-21 17:35:34 +00:00
Member
Copy link

Sent patch series version 1 containing 7 patches to gcc-patches mailing list gcc-patches@gcc.gnu.org and cc'd Richard Biener rguenther@suse.de, Richard Sandiford rdsandiford@googlemail.com, Jeffrey Law jeffrey.law@oss.qualcomm.com, Vladimir Makarov vmakarov@redhat.com, Robin Dapp rdapp.gcc@gmail.com.
Cover letter

Sent patch series version 1 containing 7 patches to gcc-patches mailing list <gcc-patches@gcc.gnu.org> and cc'd Richard Biener <rguenther@suse.de>, Richard Sandiford <rdsandiford@googlemail.com>, Jeffrey Law <jeffrey.law@oss.qualcomm.com>, Vladimir Makarov <vmakarov@redhat.com>, Robin Dapp <rdapp.gcc@gmail.com>. [Cover letter](https://inbox.sourceware.org/gcc-patches/bmm.hlk1r3usv8.gcc.gcc.feedable.223.1.0@forge-stage.sourceware.org)
feedable changed title from WIP: Remove the restrictions on instruction operand count to Remove the restrictions on instruction operand count 2026-08-21 17:44:18 +00:00
feedable force-pushed operand-cnt-test from c0214189f7
Some checks failed
/ format-checks (pull_request) Failing after 1m34s
Details
to e3184d0fec
Some checks failed
/ format-checks (pull_request) Failing after 1m33s
Details
2026-08-22 14:09:43 +00:00 Compare
feedable force-pushed operand-cnt-test from e3184d0fec
Some checks failed
/ format-checks (pull_request) Failing after 1m33s
Details
to 1fdebfa779
Some checks failed
/ format-checks (pull_request) Failing after 1m38s
Details
2026-08-22 17:28:08 +00:00 Compare
feedable force-pushed operand-cnt-test from 1fdebfa779
Some checks failed
/ format-checks (pull_request) Failing after 1m38s
Details
to de27c80098
Some checks failed
/ format-checks (pull_request) Failing after 1m37s
Details
2026-08-24 22:42:30 +00:00 Compare
feedable force-pushed operand-cnt-test from de27c80098
Some checks failed
/ format-checks (pull_request) Failing after 1m37s
Details
to c81f66d609
Some checks failed
/ format-checks (pull_request) Failing after 1m36s
Details
2026-08-24 23:01:45 +00:00 Compare
feedable force-pushed operand-cnt-test from c81f66d609
Some checks failed
/ format-checks (pull_request) Failing after 1m36s
Details
to a6ca7e094b
Some checks failed
/ format-checks (pull_request) Failing after 1m35s
Details
2026-08-24 23:10:12 +00:00 Compare
feedable commented 2026-08-24 23:21:55 +00:00
Author
Member
Copy link

/preview

/preview
forge-bot commented 2026-08-24 23:22:04 +00:00
Member
Copy link

Sent patch series version 2 containing 7 patches to @feedable's email address.
Message ID is <bmm.hlnmel6rs6.gcc.gcc.feedable.223.2.0@forge-stage.sourceware.org>.

Sent patch series version 2 containing 7 patches to @feedable's email address. Message ID is `<bmm.hlnmel6rs6.gcc.gcc.feedable.223.2.0@forge-stage.sourceware.org>`.
feedable commented 2026-08-24 23:24:44 +00:00
Author
Member
Copy link

/submit

/submit
forge-bot commented 2026-08-24 23:25:17 +00:00
Member
Copy link

Version 2 of this pull request has been stored. It includes the following commits:

  • iterator-utils.h: Fixup const_derived_container constructor - ad88464fe4
  • vec.h: Add a convenient way to construct an invalid array_slice - 94ccda57ad
  • vec.h: Add utility functions to array_slice - 469ce43cb8
  • vec.h: Implement copy&move construction&assignment for auto_vec - 348c0e4c9f
  • i386: Don't use operands array as storage - c049d0fc9c
  • rtl: Make operand count dynamic - 1b4d679adf
  • rtl: Adjust generation machinery to work with arbitrary amount of operands - a6ca7e094b
<!-- pr-new-version --> Version 2 of this pull request has been stored. It includes the following commits: - iterator-utils.h: Fixup const_derived_container constructor - ad88464fe42ab9e83f7e0a372cfef20aec882541 - vec.h: Add a convenient way to construct an invalid array_slice - 94ccda57ada63f503d3d4a7357ff76189fdef5ad - vec.h: Add utility functions to array_slice - 469ce43cb8d99a81598164b535b83f5b9a5ae2ad - vec.h: Implement copy&move construction&assignment for auto_vec - 348c0e4c9fc190047f757393d9cee0714997a0c4 - i386: Don't use operands array as storage - c049d0fc9cdce1500dca52f15119f8620051ed3b - rtl: Make operand count dynamic - 1b4d679adfe4d87a0e0a562d2e311201930ff429 - rtl: Adjust generation machinery to work with arbitrary amount of operands - a6ca7e094b4d57ac4c8661911c328b5d6823ea5f
forge-bot commented 2026-08-24 23:25:18 +00:00
Member
Copy link

Pull Request versions:

# Base Head
1 844c665f17 c0214189f7 (diff)
2 844c665f17 a6ca7e094b (diff)

In order to compare , clone this repository and run

PR=223
git fetch origin "refs/versioned_pull/${PR}/*:refs/versioned_pull/${PR}/*"
PRV1=1
PRV2=2
git range-diff "refs/versioned_pull/${PR}/${PRV1}/base..refs/versioned_pull/${PR}/${PRV1}/head"  "refs/versioned_pull/${PR}/${PRV2}/base..refs/versioned_pull/${PR}/${PRV2}/head"
<!-- pr-versions --> Pull Request versions: | # | Base | Head | | | - | ---- | ---- | - | | 1 | 844c665f1701e889b4883df65004dfc310445422 | c0214189f7e69062c6ca3c4b78a405bab2fcf541 | [(diff)](https://forge.sourceware.org/gcc/gcc.git/compare/844c665f1701e889b4883df65004dfc310445422...c0214189f7e69062c6ca3c4b78a405bab2fcf541) | | 2 | 844c665f1701e889b4883df65004dfc310445422 | a6ca7e094b4d57ac4c8661911c328b5d6823ea5f | [(diff)](https://forge.sourceware.org/gcc/gcc.git/compare/844c665f1701e889b4883df65004dfc310445422...a6ca7e094b4d57ac4c8661911c328b5d6823ea5f) | In order to compare , clone this repository and run ``` PR=223 git fetch origin "refs/versioned_pull/${PR}/*:refs/versioned_pull/${PR}/*" PRV1=1 PRV2=2 git range-diff "refs/versioned_pull/${PR}/${PRV1}/base..refs/versioned_pull/${PR}/${PRV1}/head" "refs/versioned_pull/${PR}/${PRV2}/base..refs/versioned_pull/${PR}/${PRV2}/head" ```
forge-bot commented 2026-08-24 23:25:51 +00:00
Member
Copy link

Sent patch series version 2 containing 7 patches to gcc-patches mailing list gcc-patches@gcc.gnu.org and cc'd Richard Biener rguenther@suse.de, Richard Sandiford rdsandiford@googlemail.com, Jeffrey Law jeffrey.law@oss.qualcomm.com, Vladimir Makarov vmakarov@redhat.com, Robin Dapp rdapp.gcc@gmail.com.
Cover letter

Sent patch series version 2 containing 7 patches to gcc-patches mailing list <gcc-patches@gcc.gnu.org> and cc'd Richard Biener <rguenther@suse.de>, Richard Sandiford <rdsandiford@googlemail.com>, Jeffrey Law <jeffrey.law@oss.qualcomm.com>, Vladimir Makarov <vmakarov@redhat.com>, Robin Dapp <rdapp.gcc@gmail.com>. [Cover letter](https://inbox.sourceware.org/gcc-patches/bmm.hlnmhdo3u6.gcc.gcc.feedable.223.2.0@forge-stage.sourceware.org)
feedable force-pushed operand-cnt-test from a6ca7e094b
Some checks failed
/ format-checks (pull_request) Failing after 1m35s
Details
to 7d66940ce3
Some checks failed
/ format-checks (pull_request) Failing after 1m21s
Details
2026-09-12 16:05:24 +00:00 Compare
Some checks failed
/ format-checks (pull_request) Failing after 1m21s
Details
This pull request has changes conflicting with the target branch.
  • gcc/genextract.cc
View command line instructions

Manual merge helper

Use this merge commit message when completing the merge manually.

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u operand-cnt-test:feedable-operand-cnt-test
git switch feedable-operand-cnt-test

Merge

Merge the changes and update on Forgejo.
git switch trunk
git merge --no-ff feedable-operand-cnt-test
git switch feedable-operand-cnt-test
git rebase trunk
git switch trunk
git merge --ff-only feedable-operand-cnt-test
git switch feedable-operand-cnt-test
git rebase trunk
git switch trunk
git merge --no-ff feedable-operand-cnt-test
git switch trunk
git merge --squash feedable-operand-cnt-test
git switch trunk
git merge --ff-only feedable-operand-cnt-test
git switch trunk
git merge feedable-operand-cnt-test
git push origin trunk
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels   
Compat/Breaking
Breaking change that won't be backward compatible

  
Frontend/ada

   
Frontend/c

   
Frontend/c++

   
Frontend/rust
Code affecting the Rust front end

  
General/forge
Forgejo infrastructure

  
Kind/Bug
Something is not working

  
Kind/Documentation
Documentation changes

  
Kind/Enhancement
Improve existing functionality

  
Kind/Feature
New functionality

  
Kind/Security
This is security issue

  
Kind/Testing
Issue or pull request related to testing

  
Library/libgcc

   
Library/libstdc++

   
Midend/gimple

   
Midend/rtl

   
Midend/tree
  
Priority
Critical
 The priority is critical

  
Priority
High
 The priority is high

  
Priority
Low
 The priority is low

  
Priority
Medium
 The priority is medium

  
Reviewed
Confirmed
 Issue has been confirmed

  
Reviewed
Duplicate
 This issue or pull request already exists

  
Reviewed
Invalid
 Invalid issue

  
Reviewed
Won't Fix
 This issue won't be fixed

  
Status
Abandoned
 Somebody has started to work on this but abandoned work

  
Status
Blocked
 Something is blocking this issue or pull request

  
Status
Need More Info
 Feedback is required to reproduce issue or to continue work

  
Target/aarch64

   
Target/arm

   
Target/i386
  
bug
Something is not working

  
duplicate
This issue or pull request already exists

  
enhancement
New feature

  
help wanted
Need some help

  
invalid
Something is wrong

  
question
More information is needed

  
wontfix
This won't be fixed

No labels
Compat/Breaking
Frontend/ada
Frontend/c
Frontend/c++
Frontend/rust
General/forge
Kind/Bug
Kind/Documentation
Kind/Enhancement
Kind/Feature
Kind/Security
Kind/Testing
Library/libgcc
Library/libstdc++
Midend/gimple
Midend/rtl
Midend/tree
Priority
Critical
Priority
High
Priority
Low
Priority
Medium
Reviewed
Confirmed
Reviewed
Duplicate
Reviewed
Invalid
Reviewed
Won't Fix
Status
Abandoned
Status
Blocked
Status
Need More Info
Target/aarch64
Target/arm
Target/i386
bug
duplicate
enhancement
help wanted
invalid
question
wontfix
Milestone
Clear milestone
No items
No milestone
Assignees
Clear assignees
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
gcc/gcc!223
No description provided.