IssueTracker

eclass: Don't patch LLVM with new branch workflows

There's no particularly easy way to extract whether
we're on a specific sha with the manifest, or whether
we're on an upstream branch. The easiest way seems to
be to just say later versions use this new workflow.

BUG=b:368393962
TEST=sudo emerge sys-devel/llvm on new branch
TEST=CQ

Change-Id: I8cbc5c620d3ba429576b6cba02f761760e2e5ceb
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/5905716
Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com>
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>
Reviewed-by: George Burgess <gbiv@chromium.org>

Files:

• M eclass/cros-llvm.eclass

Hash: f553966e5e54820caa194785da5cbd5335976cc3
Date:  Mon Oct 07 19:40:07 2024

llvm_tools: Add --keep-non-patch for git_am calls

We want to preserve the leading tags like [libc++] and
[lldb], and this flag lets us do that.

BUG=b:368393962
TEST=Applied patch stack and uploaded

Change-Id: Ic452e4c0a03291dd6530f0c23bead9a6a883d470
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/5923000
Commit-Queue: George Burgess <gbiv@chromium.org>
Auto-Submit: Jordan Abrahams-Whitehead <ajordanr@google.com>
Tested-by: George Burgess <gbiv@chromium.org>
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>
Commit-Queue: ChromeOS Auto Runner <chromeos-auto-runner@chromeos-bot.iam.gserviceaccount.com>
Reviewed-by: George Burgess <gbiv@chromium.org>

Files:

• M llvm_tools/patch_utils.py

Hash: 729fc0df0e556eda0fbbcecd905b3a8ab347ac55
Date:  Thu Oct 10 21:33:42 2024

llvm_tools: Update llvm-next revision and SHA

This commit does 2 things:
1. Move LLVM Next officially to r547379
2. Disable warnings for nightly SDK builds

BUG=b:368393962
TEST=CQ

Change-Id: I8a73f94c1c77892da79d266fc81c907d27d50982
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/5923001
Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com>
Reviewed-by: George Burgess <gbiv@chromium.org>
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>

Files:

• M llvm_tools/llvm_next.py

Hash: eff66fbbb676e266959c94286e2ecb460438c54e
Date:  Thu Oct 10 22:36:41 2024

llvm_tools: Fetch external remote with verify_patch_consistency

Before we were fetching "origin", which I'm not actually sure went
anywhere? But running this in a clean chroot seems to do what we want.
The `git fetch cros main` above this seems to get us the latest
upstream llvm-project commits, it's that we don't find the right SHA
during translation.

Also add some error logging/helpful messaging for when patches fail to
apply.

BUG=b:368393962
TEST=verify_patch_consistency --cl 5902466

Change-Id: Id65ac1e1805778bf26cc34e19d344fc0691082df
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/5926252
Reviewed-by: George Burgess <gbiv@chromium.org>
Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com>
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>

Files:

• M llvm_tools/verify_patch_consistency.py

Hash: 9375ed17fff353aca9956c5b5bb18c8c2fc17996
Date:  Fri Oct 11 17:05:00 2024

llvm_tools: Add verified label command to verify_patch_consistency

This step was missing in the approval commands.

It's a bit messy of an addition but works fine.

BUG=b:368393962
TEST=Ran it and looked at the printed output

Change-Id: Iedf3b3d1cbf4860c50ec9f2f7ea554f86f222919
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/5926253
Reviewed-by: George Burgess <gbiv@chromium.org>
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>
Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com>

Files:

• M llvm_tools/verify_patch_consistency.py

Hash: d924cc28b63e3baaa250a8966f2f07b5ad5a43c6
Date:  Fri Oct 11 17:57:49 2024

Android Bringup: See http://go/android-fw-sync
 eclass: Don't patch LLVM with new branch workflows

There's no particularly easy way to extract whether
we're on a specific sha with the manifest, or whether
we're on an upstream branch. The easiest way seems to
be to just say later versions use this new workflow.

BUG=b:368393962
TEST=sudo emerge sys-devel/llvm on new branch
TEST=CQ

(cherry picked from commit f553966e5e54820caa194785da5cbd5335976cc3)

Change-Id: I8cbc5c620d3ba429576b6cba02f761760e2e5ceb
Original-Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/5905716
Original-Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com>
Original-Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>
Original-Reviewed-by: George Burgess <gbiv@chromium.org>
GitOrigin-RevId: f553966e5e54820caa194785da5cbd5335976cc3
Cr-Build-Id: 8734524681159170321
Cr-Build-Url: https://cr-buildbucket.appspot.com/build/8734524681159170321
Copybot-Job-Name: android-main-chromiumos-overlay-copybot-downstream
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/5918276
Reviewed-by: Aaron Massey <aaronmassey@google.com>
Tested-by: Jonathon Murphy <jpmurphy@google.com>
Reviewed-by: Eran Mitrani <mitrani@google.com>

Files:

• M eclass/cros-llvm.eclass

Hash: 05d8701b817fdbdc182a482d9919c92853f2c704
Date:  Mon Oct 07 19:40:07 2024

Android Bringup: See http://go/android-fw-sync
 sys-devel: Update llvm patches

NOTE!
This patch does not affect either the current LLVM or
actually the future LLVM. This may be confusing, but
this has no connection to current state of the cros/llvm-r547379-1
branch. The patches there are already uploaded.

The PATCHES.json file should be overwritten and eventually moved
to toolchain-utils. But while we are trying to land the cros-llvm.eclass
hack later in the chain, we do need this to prevent patches from
diverging from the PATCHES.json and the branch itself--at least in
the intermin while we conduct the move.

This commit limits one patch and adds several patches:

1. Creates a new patch version for dereferencing a nullptr.
2. Creates a new patch version to disable vfork events in lldb.
3. Disables value tracking handle patch for newer LLVMs.

BUG=b:368393962
TEST=CQ

(cherry picked from commit 29cece6a67380286b4534443b2d3a76e034c30fa)

Change-Id: I1d8ee8775606f797ba5ed2c8b4bd28514902653d
Original-Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/5905715
Original-Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com>
Original-Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>
Original-Reviewed-by: George Burgess <gbiv@chromium.org>
GitOrigin-RevId: 29cece6a67380286b4534443b2d3a76e034c30fa
Cr-Build-Id: 8734289971778225713
Cr-Build-Url: https://cr-buildbucket.appspot.com/build/8734289971778225713
Copybot-Job-Name: android-main-chromiumos-overlay-copybot-downstream
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/5927150
Tested-by: ChromeOS Prod (Robot) <chromeos-ci-prod@chromeos-bot.iam.gserviceaccount.com>
Reviewed-by: Jonathon Murphy <jpmurphy@google.com>

Files:

• A sys-devel/llvm/files/Disable-vfork-events-v2.patch
• M sys-devel/llvm/files/PATCHES.json
• A sys-devel/llvm/files/v5-Crash-when-dereferencing-nullopt-for-std-optional.patch

Hash: 4c7b8456e0bcc7d56b35f00a3c26cc3f799d4b9b
Date:  Mon Sep 30 23:51:30 2024

llvm_tools: Add -c arg to ready_llvm_branch

This argument, also known as 'continue-on-patch-failure', will
let ready_llvm_branch to skip patches that fail to apply.
This is useful for bisecting issues while still applying the
local CrOS Base Commit for -9999 ebuilds.

BUG=b:373947346, b:368393962
TEST=Bisected LLVM with '-c'

Change-Id: I0b2effe0eef7d674ebe09b55f39b222e4d2edbe9
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/5941557
Reviewed-by: George Burgess <gbiv@chromium.org>
Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com>
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>

Files:

• M llvm_tools/ready_llvm_branch.py

Hash: 5eba11ca4575dac5050e1ef5cfc024f0e1ff9c65
Date:  Wed Oct 16 23:07:59 2024

chromeos-base/ec-utils: Extend the lifetime of the result string 
 
With llvm-r547379-1, the following warning is issued: 
 
    util/ectool.cc:1911:16: error: object backing the pointer error_str will be destroyed 
          at the end of the full-expression [-Werror,-Wdangling-assignment-gsl] 
     1911 |                         error_str = frame_cmd->ResultString().c_str(); 
          |                                     ^~~~~~~~~~~~~~~~~~~~~~~~~ 
    1 error generated. 
 
To ensure error_str doesn't point to invalid memory outside of the 
if block, retrieve ResultString into a std::string with lifetime 
long enough to still exist when the string is actually used. 
 
BUG=b:368393962 
TEST=emerge-jacuzzi ec-utils 
 
Change-Id: Ib9ee399e581a6461c5ac2dcc1465675a9e37bce6 
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/ec/+/6070888 
Reviewed-by: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Code-Coverage: Zoss <zoss-cl-coverage@prod.google.com> 
Tested-by: Andrej Shadura <andrew.shadura@collabora.com> 
Reviewed-by: Fabio Baltieri <fabiobaltieri@google.com> 
Commit-Queue: Andrej Shadura <andrew.shadura@collabora.com>

Files:

• M util/ectool.cc

Hash: de72c179cb99025bb71df293ba1c4226c8120efb
Date:  Wed Dec 04 17:24:35 2024

chromeos-base/net-base: Drop unused "ipv4" variables 
 
With llvm-r547379-1, the following warning is issued: 
 
    .../net-base/ip_address.cc:107:18: error: variable 'ipv4' set but not used [-Werror,-Wunused-but-set-variable] 
      107 |   if (const auto ipv4 = std::get_if<IPv4Address>(&address_)) { 
          |                  ^ 
    .../net-base/ip_address.cc:287:18: error: variable 'ipv4' set but not used [-Werror,-Wunused-but-set-variable] 
      287 |   if (const auto ipv4 = std::get_if<IPv4CIDR>(&cidr_)) { 
          |                  ^ 
    2 errors generated. 
 
BUG=b:368393962 
TEST=emerge-jacuzzi net-base 
 
Change-Id: I02c11c7ae26b6971e1a6875cf9a6d691f9e85c80 
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform2/+/6070207 
Reviewed-by: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Reviewed-by: Jie Jiang <jiejiang@chromium.org> 
Reviewed-by: Chih-Yu Huang <akahuang@chromium.org> 
Reviewed-by: Hugo Benichi <hugobenichi@google.com> 
Commit-Queue: Andrej Shadura <andrew.shadura@collabora.com> 
Tested-by: Andrej Shadura <andrew.shadura@collabora.com>

Files:

• M net-base/ip_address.cc

Hash: f76bda6db5de77f2397f2fb6503977960ba7aff6
Date:  Wed Dec 04 17:44:31 2024

floss: Fix -Wmissing-template-arg-list-after-template-kw warning 
 
With llvm-r547379-1, the following warning is issued: 
 
    In file included from .../ext/filters/client_channel/channel_connectivity.cc:31: 
    In file included from .../ext/filters/client_channel/client_channel.h:40: 
    In file included from .../ext/filters/client_channel/client_channel_factory.h:24: 
    In file included from .../ext/filters/client_channel/subchannel.h:35: 
    In file included from .../ext/filters/client_channel/connector.h:31: 
    In file included from .../lib/transport/transport.h:58: 
    In file included from .../lib/promise/pipe.h:43: 
    In file included from .../lib/promise/seq.h:23: 
    .../lib/promise/detail/basic_seq.h:474:38: error: a template argument 
    list is expected after a name prefixed by the template keyword 
    [-Wmissing-template-arg-list-after-template-kw] 
      474 |                     Traits::template CallSeqFactory(f_, *cur_, std::move(arg))); 
          |                                      ^ 
    1 error generated. 
 
Apply an upstream patch from the embedded grpc library: 
 
    Clang now requires a template argument list after the use of the 
    template keyword. Edit this instance to remove the template keyword 
    since there are no template arguments. 
 
    See https://github.com/llvm/llvm-project/pull/80801. 
 
See https://github.com/grpc/grpc/pull/36805 
 
Change-Id: I54d8680f34338e5faea7278a0f4a99469f406e60 
 
BUG=b:368393962 
TEST=emerge-jacuzzi third-party-crates-src floss 
 
Change-Id: I5f222e78132f2a7fae0ad106250688c095e294e0 
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/rust_crates/+/6075068 
Tested-by: Andrej Shadura <andrew.shadura@collabora.com> 
Reviewed-by: Allen Webb <allenwebb@google.com> 
Commit-Queue: Andrej Shadura <andrew.shadura@collabora.com> 
Reviewed-by: George Burgess <gbiv@chromium.org>

Files:

• A patches/grpcio-sys-0.13.0+1.56.2-patched/0004-fix-Wmissing-template-arg-list-after-template-kw.patch
• M vendor/grpcio-sys-0.13.0+1.56.2-patched/.cargo-checksum.json
• M vendor/grpcio-sys-0.13.0+1.56.2-patched/grpc/src/core/lib/promise/detail/basic_seq.h
• M vendor_artifacts/std/annotations.json

Hash: 41c8ba04571ab843573e6fc5daa70cdb9339bec7
Date:  Thu Dec 05 10:45:44 2024

llvm_tools: Disable afdo for LLVM Next Testing 
 
This commit adds the autofdo disable CL to the list of 
LLVM Next CLs so we can determine if anything is broken 
beyond autofdo. 
 
BUG=b:380093598, b:368393962 
TEST=CQ 
 
Change-Id: Idfa73b749eb777d5c8ab5e8822e3f7223c9daaf0 
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/6077457 
Reviewed-by: Bob Haarman <inglorion@chromium.org> 
Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>

Files:

• M llvm_tools/llvm_next.py

Hash: 60512b5a03ce9a60c0ea532c90db13f39e35186c
Date:  Fri Dec 06 22:33:32 2024

arc: Include missing vector header 
 
With the latest LLVM, we encountered this missing <vector> 
include that should have been introduced along with 
https://chromium-review.googlesource.com/c/aosp/platform/system/keymaster/+/6058651 
 
Caught because we encountered the error here: 
https://paste.googleplex.com/5661780354007040 
 
BUG=b:368393962 
TEST=Built locally for Brya 
 
Change-Id: I20e8c48763957423f8dca11d9d3f7dc959861433 
Reviewed-on: https://chromium-review.googlesource.com/c/aosp/platform/system/keymaster/+/6196346 
Commit-Queue: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Reviewed-by: Vaibhav Raheja <vraheja@chromium.org>

Files:

• M include/keymaster/km_openssl/openssl_utils.h

Hash: d260a9ead5c237efde712203b19f828bbf71e543
Date:  Sat Jan 25 00:19:55 2025

llvm_tools: Update LLVM_NEXT_TESTING_CL_URLS 
 
Manually updates these, fixing previous merge conflicts 
and setting the right cLs. 
 
BUG=b:368393962 
TEST=CQ 
 
Change-Id: I0238d10a17426b2e6bb0f795f7b2ae6a1b56bdcd 
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/6237014 
Commit-Queue: George Burgess <gbiv@chromium.org> 
Auto-Submit: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Reviewed-by: George Burgess <gbiv@chromium.org> 
Commit-Queue: ChromeOS Auto Runner <chromeos-auto-runner@chromeos-bot.iam.gserviceaccount.com>

Files:

• M llvm_tools/llvm_next.py

Hash: 71a4615c57e18dcd090efec812d6a722f6da40d5
Date:  Wed Feb 05 21:51:56 2025

llvm_tools: Update llvm_next cls 
 
The intel-openvino CL has been merged and is no longer needed. 
 
BUG=b:368393962 
TEST=pytest 
 
Change-Id: I2c1f4b20169304a72b10b3479e3c46a78addd8f2 
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/6253871 
Reviewed-by: George Burgess <gbiv@chromium.org> 
Auto-Submit: Jordan Abrahams-Whitehead <ajordanr@google.com> 
Commit-Queue: George Burgess <gbiv@chromium.org> 
Tested-by: Jordan Abrahams-Whitehead <ajordanr@google.com>

Files:

• M llvm_tools/llvm_next.py

Hash: 0a2a14256ce33d801f8d3e91306e80993836f3ca
Date:  Tue Feb 11 21:26:24 2025