Bump LLVM to 19.1.7 #1653

sunshaoce · 2025-01-09T09:14:13Z

No description provided.

TommyMurphyTM1234 · 2025-01-09T10:46:46Z

Shouldn't .gitmodules also be updated?

Lines 49 to 53 in 43536ac

    
           [submodule "llvm"] 
        
           	path = llvm 
        
           	url = https://github.com/llvm/llvm-project.git 
        
           	branch = release/18.x 
        
           	shallow = true

sunshaoce · 2025-01-09T10:48:49Z

Shouldn't .gitmodules also be updated?

riscv-gnu-toolchain/.gitmodules

Lines 49 to 53 in 43536ac

[submodule "llvm"]

path = llvm

url = https://github.com/llvm/llvm-project.git

branch = release/18.x

shallow = true

Done. Thanks!

TommyMurphyTM1234 · 2025-01-09T10:54:06Z

The CI failure seems odd and not directly related to this change?

https://github.com/riscv-collab/riscv-gnu-toolchain/actions/runs/12688396333/job/35364834786?pr=1653#step:6:43

sunshaoce · 2025-01-10T07:10:48Z

It should be an issue with my local Git repository, but I don't know exactly where the problem lies.

TommyMurphyTM1234 · 2025-01-10T08:11:10Z

It should be an issue with my local Git repository, but I don't know exactly where the problem lies.

I don't think that that's it - the failure is on the GitHub CI side so should have nothing to do with your repos... I'll try a build later to see if I can spot anything.

sunshaoce · 2025-01-10T08:12:40Z

It seems to be affected by the shallow option of git clone, making it unable to checkout ca371cf9c48186716d26c8e10d442affaf002c80.

TommyMurphyTM1234 · 2025-01-13T10:32:39Z

I don't think that simply removing the shallow clone/checkout is the correct solution here as it undoes changes recently applied here:

Update build workflow #1608
- efc7cc9

Maybe @mickflemm might have some better idea?

TommyMurphyTM1234

Hope this helps!

TommyMurphyTM1234 · 2025-01-13T17:16:43Z

.github/workflows/build.yaml

@@ -50,7 +50,7 @@ jobs:

      - name: Checkout required submodules
        if: steps.smcache.outputs.cache-hit != 'true'
-        run: git submodule update --init -j $(nproc) --depth 1 $(echo ${submodule_paths} | sed '$d' | tr '\n' ' ')
+        run: git submodule update --init -j $(nproc) $(echo ${submodule_paths} | sed '$d' | tr '\n' ' ')


As mentioned in the PR conversation I don't think that this is the correct solution since the shallow clone/checkout was deliberately added recently for a specific purpose:

Update build workflow #1608

efc7cc9

Ditto for the removal of shallow = true in .gitmodules below.

TommyMurphyTM1234 · 2025-01-13T17:16:48Z

.gitmodules

 [submodule "llvm"]
 	path = llvm
 	url = https://github.com/llvm/llvm-project.git
-	branch = release/18.x
-	shallow = true
+	branch = release/19.x


Branch release/19.x:

https://github.com/llvm/llvm-project/tree/release/19.x

is commit 61c9f97:

61c9f97

But this PR changes the submodule commit from 3b5b5c to e21dc4b.

Isn't this a mismatch?

The branch release/19.x has tag llvmorg-19.1.6 ( e21dc4b, https://github.com/llvm/llvm-project/tree/llvmorg-19.1.6 ) . When I submitted (including now), there was no llvmorg-19.1.7 (61c9f97) related tag. I will update it as soon as it becomes available.

TommyMurphyTM1234 · 2025-01-13T17:17:52Z

.gitmodules

@@ -2,56 +2,44 @@
 	path = binutils
 	url = https://sourceware.org/git/binutils-gdb.git
 	branch = binutils-2_43-branch
-	shallow = true


See earlier review comment about removing the shallow clone/checkout functionality.

TommyMurphyTM1234 · 2025-01-13T18:00:29Z

I tried to bump LLVM locally to see if I could reproduce the CI failure but I hit other issues along the way...

git clone https://github.com/riscv-collab/riscv-gnu-toolchain pull-1653
Cloning into 'pull-1653'...
remote: Enumerating objects: 9508, done.
remote: Counting objects: 100% (415/415), done.
remote: Compressing objects: 100% (101/101), done.
remote: Total 9508 (delta 353), reused 316 (delta 314), pack-reused 9093 (from 2)
Receiving objects: 100% (9508/9508), 6.46 MiB | 6.68 MiB/s, done.
Resolving deltas: 100% (4729/4729), done.

cd pull-1653

# Edit .gitmodules to bump LLVM
# [submodule "llvm"]
#        path = llvm
#        url = https://github.com/llvm/llvm-project.git
#        branch = release/19.x
#        shallow = true

git submodule update

cd llvm

git checkout -b release/19.x
Switched to a new branch 'release/19.x'

git pull https://github.com/llvm/llvm-project
remote: Enumerating objects: 6217989, done.
remote: Counting objects: 100% (8728/8728), done.
remote: Compressing objects: 100% (1221/1221), done.
remote: Total 6217989 (delta 8321), reused 7510 (delta 7507), pack-reused 6209261 (from 2)
Receiving objects: 100% (6217989/6217989), 1.58 GiB | 10.10 MiB/s, done.
Resolving deltas: 100% (5154153/5154153), done.
From https://github.com/llvm/llvm-project
 * branch                      HEAD       -> FETCH_HEAD
hint: You have divergent branches and need to specify how to reconcile them.
hint: You can do so by running one of the following commands sometime before
hint: your next pull:
hint:
hint:   git config pull.rebase false  # merge
hint:   git config pull.rebase true   # rebase
hint:   git config pull.ff only       # fast-forward only
hint:
hint: You can replace "git config" with "git config --global" to set a default
hint: preference for all repositories. You can also pass --rebase, --no-rebase,
hint: or --ff-only on the command line to override the configured default per
hint: invocation.
fatal: Need to specify how to reconcile divergent branches.

I've tried various other approaches to updating the submodule to the release/19.x commit but get the same or similar problems so I'm not sure what's the correct way to do this...

I don't understand why the CI build fails because it cannot find the relevant Binutils commit when the shallow clone/checkout is used. And I can't see how bumping LLVM would trigger this.

FWIW I cannot reproduce the problem with the "missing/unavailable" Binutils commit if I just clone, configure and build as usual from the repo master/head.

sunshaoce · 2025-01-14T02:14:05Z

Many thanks to @TommyMurphyTM1234 for the patient response. Indeed, the shallow=true and --depth=1 code should not be removed in this PR. However, after trying this approach https://github.com/riscv-collab/riscv-gnu-toolchain/actions/runs/12743602742, it seems to resolve the issue.

TommyMurphyTM1234 · 2025-01-14T11:39:01Z

However, after trying this approach https://github.com/riscv-collab/riscv-gnu-toolchain/actions/runs/12743602742, it seems to resolve the issue.

I don't undestand this. Is the link correct? It seems to link to a CI run status. What approach are you referring to?

sunshaoce · 2025-01-15T14:28:27Z

I don't undestand this. Is the link correct? It seems to link to a CI run status. What approach are you referring to?

Sorry, I didn't express myself clearly. At that time, I removed shallow and depth=1, and then the CI passed.

cmuellner · 2025-01-15T15:29:28Z

I don't undestand this. Is the link correct? It seems to link to a CI run status. What approach are you referring to?

Sorry, I didn't express myself clearly. At that time, I removed shallow and depth=1, and then the CI passed.

I've just created #1655 (as a quick workaround).

TommyMurphyTM1234 · 2025-01-15T18:10:00Z

The latest failure:

https://github.com/riscv-collab/riscv-gnu-toolchain/actions/runs/12790520055/job/35656455751?pr=1653#step:6:46

presumably depends on this PR being merged first?

dejagnu: Bump version from ca371cf9c48 to 935a51f3c66e #1655

cmuellner · 2025-01-15T19:56:24Z

Please rebase this PR to fix the CI issue.

This reverts commit f1d0e21.

This reverts commit 28c48ae.

cmuellner

LGTM

cmuellner · 2025-01-16T07:38:06Z

Thanks!

sunshaoce force-pushed the llvm-19 branch from 37eb2bb to c9d07df Compare January 10, 2025 07:09

TommyMurphyTM1234 mentioned this pull request Jan 13, 2025

git submodule update fails #1654

Open

TommyMurphyTM1234 reviewed Jan 13, 2025

View reviewed changes

sunshaoce changed the title ~~Bump LLVM to 19.1.6~~ Bump LLVM to 19.1.7 Jan 15, 2025

sunshaoce added 6 commits January 16, 2025 09:19

Bump LLVM to 19.1.6

425105f

fixup! Disable shallow

1937acf

fixup! Remove depth=1

a7f254b

Revert "fixup! Disable shallow"

59adb68

This reverts commit f1d0e21.

Revert "fixup! Remove depth=1"

fb6d930

This reverts commit 28c48ae.

fixup! Bump LLVM to 19.1.7

9b1c1a2

sunshaoce force-pushed the llvm-19 branch from e67b7d3 to 9b1c1a2 Compare January 16, 2025 01:20

cmuellner approved these changes Jan 16, 2025

View reviewed changes

cmuellner merged commit 071541e into riscv-collab:master Jan 16, 2025
27 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Bump LLVM to 19.1.7 #1653

Bump LLVM to 19.1.7 #1653

sunshaoce commented Jan 9, 2025

TommyMurphyTM1234 commented Jan 9, 2025

sunshaoce commented Jan 9, 2025

TommyMurphyTM1234 commented Jan 9, 2025

sunshaoce commented Jan 10, 2025

TommyMurphyTM1234 commented Jan 10, 2025

sunshaoce commented Jan 10, 2025

TommyMurphyTM1234 commented Jan 13, 2025

TommyMurphyTM1234 left a comment

TommyMurphyTM1234 Jan 13, 2025

TommyMurphyTM1234 Jan 13, 2025

sunshaoce Jan 14, 2025 •

edited

Loading

TommyMurphyTM1234 Jan 13, 2025 •

edited

Loading

TommyMurphyTM1234 commented Jan 13, 2025 •

edited

Loading

sunshaoce commented Jan 14, 2025

TommyMurphyTM1234 commented Jan 14, 2025

sunshaoce commented Jan 15, 2025

cmuellner commented Jan 15, 2025

TommyMurphyTM1234 commented Jan 15, 2025

cmuellner commented Jan 15, 2025

cmuellner left a comment

cmuellner commented Jan 16, 2025

Bump LLVM to 19.1.7 #1653

Bump LLVM to 19.1.7 #1653

Conversation

sunshaoce commented Jan 9, 2025

TommyMurphyTM1234 commented Jan 9, 2025

sunshaoce commented Jan 9, 2025

TommyMurphyTM1234 commented Jan 9, 2025

sunshaoce commented Jan 10, 2025

TommyMurphyTM1234 commented Jan 10, 2025

sunshaoce commented Jan 10, 2025

TommyMurphyTM1234 commented Jan 13, 2025

TommyMurphyTM1234 left a comment

Choose a reason for hiding this comment

TommyMurphyTM1234 Jan 13, 2025

Choose a reason for hiding this comment

TommyMurphyTM1234 Jan 13, 2025

Choose a reason for hiding this comment

sunshaoce Jan 14, 2025 • edited Loading

Choose a reason for hiding this comment

TommyMurphyTM1234 Jan 13, 2025 • edited Loading

Choose a reason for hiding this comment

TommyMurphyTM1234 commented Jan 13, 2025 • edited Loading

sunshaoce commented Jan 14, 2025

TommyMurphyTM1234 commented Jan 14, 2025

sunshaoce commented Jan 15, 2025

cmuellner commented Jan 15, 2025

TommyMurphyTM1234 commented Jan 15, 2025

cmuellner commented Jan 15, 2025

cmuellner left a comment

Choose a reason for hiding this comment

cmuellner commented Jan 16, 2025

sunshaoce Jan 14, 2025 •

edited

Loading

TommyMurphyTM1234 Jan 13, 2025 •

edited

Loading

TommyMurphyTM1234 commented Jan 13, 2025 •

edited

Loading