Merge branch 'main' into gfx950-mfma-rocdl

Author: jungpark-mlir

.ci/metrics/metrics.py

@@ -26,7 +26,67 @@ class JobMetrics:
     workflow_id: int
 
 
-def get_metrics(github_repo: github.Repository, workflows_to_track: dict[str, int]):
+@dataclass
+class GaugeMetric:
+    name: str
+    value: int
+    time_ns: int
+
+
+def get_sampled_workflow_metrics(github_repo: github.Repository):
+    """Gets global statistics about the Github workflow queue
+
+    Args:
+      github_repo: A github repo object to use to query the relevant information.
+
+    Returns:
+      Returns a list of GaugeMetric objects, containing the relevant metrics about
+      the workflow
+    """
+
+    # Other states are available (pending, waiting, etc), but the meaning
+    # is not documented (See #70540).
+    # "queued" seems to be the info we want.
+    queued_workflow_count = len(
+        [
+            x
+            for x in github_repo.get_workflow_runs(status="queued")
+            if x.name in WORKFLOWS_TO_TRACK
+        ]
+    )
+    running_workflow_count = len(
+        [
+            x
+            for x in github_repo.get_workflow_runs(status="in_progress")
+            if x.name in WORKFLOWS_TO_TRACK
+        ]
+    )
+
+    workflow_metrics = []
+    workflow_metrics.append(
+        GaugeMetric(
+            "workflow_queue_size",
+            queued_workflow_count,
+            time.time_ns(),
+        )
+    )
+    workflow_metrics.append(
+        GaugeMetric(
+            "running_workflow_count",
+            running_workflow_count,
+            time.time_ns(),
+        )
+    )
+    # Always send a hearbeat metric so we can monitor is this container is still able to log to Grafana.
+    workflow_metrics.append(
+        GaugeMetric("metrics_container_heartbeat", 1, time.time_ns())
+    )
+    return workflow_metrics
+
+
+def get_per_workflow_metrics(
+    github_repo: github.Repository, workflows_to_track: dict[str, int]
+):
     """Gets the metrics for specified Github workflows.
 
     This function takes in a list of workflows to track, and optionally the
@@ -43,14 +103,14 @@ def get_metrics(github_repo: github.Repository, workflows_to_track: dict[str, in
       Returns a list of JobMetrics objects, containing the relevant metrics about
       the workflow.
     """
-    workflow_runs = iter(github_repo.get_workflow_runs())
-
     workflow_metrics = []
 
     workflows_to_include = set(workflows_to_track.keys())
 
-    while len(workflows_to_include) > 0:
-        workflow_run = next(workflow_runs)
+    for workflow_run in iter(github_repo.get_workflow_runs()):
+        if len(workflows_to_include) == 0:
+            break
+
         if workflow_run.status != "completed":
             continue
 
@@ -139,12 +199,27 @@ def upload_metrics(workflow_metrics, metrics_userid, api_key):
       metrics_userid: The userid to use for the upload.
       api_key: The API key to use for the upload.
     """
+
+    if len(workflow_metrics) == 0:
+        print("No metrics found to upload.", file=sys.stderr)
+        return
+
     metrics_batch = []
     for workflow_metric in workflow_metrics:
-        workflow_formatted_name = workflow_metric.job_name.lower().replace(" ", "_")
-        metrics_batch.append(
-            f"{workflow_formatted_name} queue_time={workflow_metric.queue_time},run_time={workflow_metric.run_time},status={workflow_metric.status} {workflow_metric.created_at_ns}"
-        )
+        if isinstance(workflow_metric, GaugeMetric):
+            name = workflow_metric.name.lower().replace(" ", "_")
+            metrics_batch.append(
+                f"{name} value={workflow_metric.value} {workflow_metric.time_ns}"
+            )
+        elif isinstance(workflow_metric, JobMetrics):
+            name = workflow_metric.job_name.lower().replace(" ", "_")
+            metrics_batch.append(
+                f"{name} queue_time={workflow_metric.queue_time},run_time={workflow_metric.run_time},status={workflow_metric.status} {workflow_metric.created_at_ns}"
+            )
+        else:
+            raise ValueError(
+                f"Unsupported object type {type(workflow_metric)}: {str(workflow_metric)}"
+            )
 
     request_data = "\n".join(metrics_batch)
     response = requests.post(
@@ -176,16 +251,21 @@ def main():
     # Enter the main loop. Every five minutes we wake up and dump metrics for
     # the relevant jobs.
     while True:
-        current_metrics = get_metrics(github_repo, workflows_to_track)
-        if len(current_metrics) == 0:
-            print("No metrics found to upload.", file=sys.stderr)
-            continue
+        current_metrics = get_per_workflow_metrics(github_repo, workflows_to_track)
+        current_metrics += get_sampled_workflow_metrics(github_repo)
+        # Always send a hearbeat metric so we can monitor is this container is still able to log to Grafana.
+        current_metrics.append(
+            GaugeMetric("metrics_container_heartbeat", 1, time.time_ns())
+        )
 
         upload_metrics(current_metrics, grafana_metrics_userid, grafana_api_key)
         print(f"Uploaded {len(current_metrics)} metrics", file=sys.stderr)
 
         for workflow_metric in reversed(current_metrics):
-            workflows_to_track[workflow_metric.job_name] = workflow_metric.workflow_id
+            if isinstance(workflow_metric, JobMetrics):
+                workflows_to_track[
+                    workflow_metric.job_name
+                ] = workflow_metric.workflow_id
 
         time.sleep(SCRAPE_INTERVAL_SECONDS)
 

.github/CODEOWNERS

@@ -27,7 +27,6 @@
 /llvm/lib/Transforms/Scalar/MemCpyOptimizer.cpp @nikic
 /llvm/lib/Transforms/InstCombine/ @nikic
 
-/clang/include/clang/Sema/Sema.h @Endilll
 /clang/test/CXX/drs/ @Endilll
 /clang/www/cxx_dr_status.html @Endilll
 /clang/www/make_cxx_dr_status @Endilll

.github/workflows/build-ci-container.yml

@@ -23,7 +23,9 @@ jobs:
     runs-on: depot-ubuntu-22.04-16
     outputs:
       container-name: ${{ steps.vars.outputs.container-name }}
+      container-name-agent: ${{ steps.vars.outputs.container-name-agent }}
       container-name-tag: ${{ steps.vars.outputs.container-name-tag }}
+      container-name-agent-tag: ${{ steps.vars.outputs.container-name-agent-tag }}
       container-filename: ${{ steps.vars.outputs.container-filename }}
     steps:
       - name: Checkout LLVM
@@ -36,19 +38,22 @@ jobs:
           tag=`date +%s`
           container_name="ghcr.io/$GITHUB_REPOSITORY_OWNER/ci-ubuntu-22.04"
           echo "container-name=$container_name" >> $GITHUB_OUTPUT
+          echo "container-name-agent=$container_name-agent" >> $GITHUB_OUTPUT
           echo "container-name-tag=$container_name:$tag" >> $GITHUB_OUTPUT
+          echo "container-name-agent-tag=$container_name-agent:$tag" >> $GITHUB_OUTPUT
           echo "container-filename=$(echo $container_name:$tag  | sed -e 's/\//-/g' -e 's/:/-/g').tar" >> $GITHUB_OUTPUT
       - name: Build container
         working-directory: ./.github/workflows/containers/github-action-ci/
         run: |
-          podman build -t ${{ steps.vars.outputs.container-name-tag }} .
+          podman build --target ci-container -t ${{ steps.vars.outputs.container-name-tag }} .
+          podman build --target ci-container-agent -t ${{ steps.vars.outputs.container-name-agent-tag }} .
 
       # Save the container so we have it in case the push fails.  This also
       # allows us to separate the push step into a different job so we can
       # maintain minimal permissions while building the container.
       - name: Save container image
         run: |
-          podman save  ${{ steps.vars.outputs.container-name-tag }} >  ${{ steps.vars.outputs.container-filename }}
+          podman save  ${{ steps.vars.outputs.container-name-tag }} ${{ steps.vars.outputs.container-name-agent-tag }} >  ${{ steps.vars.outputs.container-filename }}
 
       - name: Upload container image
         uses: actions/upload-artifact@v4
@@ -86,3 +91,7 @@ jobs:
           podman login -u ${{ github.actor }} -p $GITHUB_TOKEN ghcr.io
           podman push ${{ needs.build-ci-container.outputs.container-name-tag }}
           podman push ${{ needs.build-ci-container.outputs.container-name }}:latest
+
+          podman tag ${{ needs.build-ci-container.outputs.container-name-agent-tag }} ${{ needs.build-ci-container.outputs.container-name-agent }}:latest
+          podman push ${{ needs.build-ci-container.outputs.container-name-agent-tag }}
+          podman push ${{ needs.build-ci-container.outputs.container-name-agent }}:latest

.github/workflows/containers/github-action-ci-windows/Dockerfile

@@ -108,7 +108,7 @@ RUN choco install -y handle
 
 RUN pip3 install pywin32 buildbot-worker==2.8.4
 
-ARG RUNNER_VERSION=2.319.1
+ARG RUNNER_VERSION=2.321.0
 ENV RUNNER_VERSION=$RUNNER_VERSION
 
 RUN powershell -Command \

.github/workflows/containers/github-action-ci/Dockerfile

@@ -13,7 +13,8 @@ RUN apt-get update && \
     ninja-build \
     python3 \
     git \
-    curl
+    curl \
+    zlib1g-dev
 
 RUN curl -O -L https://github.com/llvm/llvm-project/archive/refs/tags/llvmorg-$LLVM_VERSION.tar.gz && tar -xf llvmorg-$LLVM_VERSION.tar.gz
 
@@ -38,7 +39,7 @@ RUN cmake -B ./build -G Ninja ./llvm \
 
 RUN ninja -C ./build stage2-clang-bolt stage2-install-distribution && ninja -C ./build install-distribution
 
-FROM base
+FROM base as ci-container
 
 COPY --from=stage1-toolchain $LLVM_SYSROOT $LLVM_SYSROOT
 
@@ -91,4 +92,15 @@ RUN adduser gha sudo
 RUN echo '%sudo ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers
 
 USER gha
+WORKDIR /home/gha
+
+FROM ci-container as ci-container-agent
+
+ENV GITHUB_RUNNER_VERSION=2.321.0
+
+RUN mkdir actions-runner && \
+    cd actions-runner && \
+    curl -O -L https://github.com/actions/runner/releases/download/v$GITHUB_RUNNER_VERSION/actions-runner-linux-x64-$GITHUB_RUNNER_VERSION.tar.gz && \
+    tar xzf ./actions-runner-linux-x64-$GITHUB_RUNNER_VERSION.tar.gz && \
+    rm ./actions-runner-linux-x64-$GITHUB_RUNNER_VERSION.tar.gz
 

.github/workflows/premerge.yaml

@@ -30,6 +30,8 @@ jobs:
           fetch-depth: 2
       - name: Setup ccache
         uses: hendrikmuhs/[email protected]
+        with:
+          max-size: "2000M"
       - name: Build and Test
         # Mark the job as a success even if the step fails so that people do
         # not get notified while the new premerge pipeline is in an

.github/workflows/release-binaries-all.yml

@@ -83,7 +83,6 @@ jobs:
       matrix:
         runs-on:
           - ubuntu-22.04
-          - windows-2022
           - macos-13
           - macos-14