summaryrefslogtreecommitdiff
path: root/.gitlab-ci/deqp-runner.sh
diff options
context:
space:
mode:
authorEric Anholt <eric@anholt.net>2019-11-04 10:54:41 -0800
committerEric Anholt <eric@anholt.net>2019-11-12 12:54:04 -0800
commitf08c810028ba0ef264fb48d5ef0ddac08feef773 (patch)
tree94fdff594b5ae41f1c5463c2a21fee021ac91396 /.gitlab-ci/deqp-runner.sh
parent7f52df7fc9bbfcd862320be8fbab4b82649aadf3 (diff)
ci: Use cts_runner for our dEQP runs.
This runner is a little project by Bas, written in C++, that spawns threads that then loop grabbing chunks of the (randomly shuffled but consistently so) test list and hand it to a dEQP instance. As the remaining list gets shorter, so do the chunks, so hopefully the threads all complete effectively at once. It also handles restarting after crashes automatically. I've extended the runner a bit to do what I was doing in the bash scripts before, like the skip list and expected failures handling. This project should also be a good baseline for extending to handle retesting of intermittent failures. By switching to it, we can have the swrast tests just take up one job slot on the shared runners and keep their allotment of CPUs busy, instead of taking up job slots with single-threaded dEQP jobs. It will also let us (eventually, once I reprovision) switch the freedreno runners over to threading within the job instead of running concurrent jobs, so that memory scribbles in one pipeline don't affect unrelated pipelines, and I can experiment with their parallelism (particularly on a306 where we are frequently backed up) without trashing other people's jobs. What we lose in this process is per-test output in the log (not a big loss, I think, since we summarize fails at the end and reducing log length keeps chrome from choking on our logs so badly). We also drop the renderer sanity checking, since it's not saving qpa files for us to go poke through. Given that all the drivers involved have fail lists, if we got the wrong renderer somehow, we'd get a job failure anyway. v2: Rebase on droppong of the autoscale cluster and the arm64 build/test split. Use a script to deduplicate the cts-runner build. v3: Rebase on the amd64 build/test container split. Acked-by: Daniel Stone <daniels@collabora.com> (v1) Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com> (v2)
Diffstat (limited to '.gitlab-ci/deqp-runner.sh')
-rwxr-xr-x.gitlab-ci/deqp-runner.sh101
1 files changed, 22 insertions, 79 deletions
diff --git a/.gitlab-ci/deqp-runner.sh b/.gitlab-ci/deqp-runner.sh
index 822665ca66e..2e93a643a36 100755
--- a/.gitlab-ci/deqp-runner.sh
+++ b/.gitlab-ci/deqp-runner.sh
@@ -6,8 +6,6 @@ DEQP_OPTIONS=(--deqp-surface-width=256 --deqp-surface-height=256)
DEQP_OPTIONS+=(--deqp-surface-type=pbuffer)
DEQP_OPTIONS+=(--deqp-gl-config-name=rgba8888d24s8ms0)
DEQP_OPTIONS+=(--deqp-visibility=hidden)
-DEQP_OPTIONS+=(--deqp-log-images=disable)
-DEQP_OPTIONS+=(--deqp-crashhandler=enable)
# It would be nice to be able to enable the watchdog, so that hangs in a test
# don't need to wait the full hour for the run to time out. However, some
@@ -26,20 +24,7 @@ if [ -z "$DEQP_SKIPS" ]; then
exit 1
fi
-# Prep the expected failure list
-if [ -n "$DEQP_EXPECTED_FAILS" ]; then
- export DEQP_EXPECTED_FAILS=`pwd`/artifacts/$DEQP_EXPECTED_FAILS
-else
- export DEQP_EXPECTED_FAILS=/tmp/expect-no-failures.txt
- touch $DEQP_EXPECTED_FAILS
-fi
-sort < $DEQP_EXPECTED_FAILS > /tmp/expected-fails.txt
-
-# Fix relative paths on inputs.
-export DEQP_SKIPS=`pwd`/artifacts/$DEQP_SKIPS
-
-# Be a good citizen on the shared runners.
-export LP_NUM_THREADS=4
+ARTIFACTS=`pwd`/artifacts
# Set up the driver environment.
export LD_LIBRARY_PATH=`pwd`/install/lib/
@@ -52,19 +37,9 @@ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
RESULTS=`pwd`/results
mkdir -p $RESULTS
-cd /deqp/modules/$DEQP_VER
-
# Generate test case list file
cp /deqp/mustpass/$DEQP_VER-master.txt /tmp/case-list.txt
-# Note: not using sorted input and comm, becuase I want to run the tests in
-# the same order that dEQP would.
-while read -r line; do
- if echo "$line" | grep -q '^[^#]'; then
- sed -i "/$line/d" /tmp/case-list.txt
- fi
-done < $DEQP_SKIPS
-
# If the job is parallel, take the corresponding fraction of the caselist.
# Note: N~M is a gnu sed extension to match every nth line (first line is #1).
if [ -n "$CI_NODE_INDEX" ]; then
@@ -76,61 +51,29 @@ if [ ! -s /tmp/case-list.txt ]; then
exit 1
fi
-# Cannot use tee because dash doesn't have pipefail
-touch /tmp/result.txt
-tail -f /tmp/result.txt &
-
-./deqp-$DEQP_VER "${DEQP_OPTIONS[@]}" --deqp-log-filename=$RESULTS/results.qpa --deqp-caselist-file=/tmp/case-list.txt >> /tmp/result.txt
-DEQP_EXITCODE=$?
-
-sed -ne \
- '/StatusCode="Fail"/{x;p}; s/#beginTestCaseResult //; T; h' \
- $RESULTS/results.qpa \
- > /tmp/unsorted-fails.txt
-
-# Scrape out the renderer that the test run used, so we can validate that the
-# right driver was used.
-if grep -q "dEQP-.*.info.renderer" /tmp/case-list.txt; then
- # This is an ugly dependency on the .qpa format: Print 3 lines after the
- # match, which happens to contain the result.
- RENDERER=`sed -n '/#beginTestCaseResult dEQP-.*.info.renderer/{n;n;n;p}' $RESULTS/results.qpa | sed -n -E "s|<Text>(.*)</Text>|\1|p"`
-
- echo "GL_RENDERER for this test run: $RENDERER"
-
- if [ -n "$DEQP_RENDERER_MATCH" ]; then
- echo $RENDERER | grep -q $DEQP_RENDERER_MATCH > /dev/null
- fi
+if [ -n "$DEQP_EXPECTED_FAILS" ]; then
+ XFAIL="--xfail-list $ARTIFACTS/$DEQP_EXPECTED_FAILS"
fi
-if grep -q "dEQP-.*.info.version" /tmp/case-list.txt; then
- # This is an ugly dependency on the .qpa format: Print 3 lines after the
- # match, which happens to contain the result.
- VERSION=`sed -n '/#beginTestCaseResult dEQP-.*.info.version/{n;n;n;p}' $RESULTS/results.qpa | sed -n -E "s|<Text>(.*)</Text>|\1|p"`
- echo "Driver version tested: $VERSION"
-fi
+set +e
+
+vulkan-cts-runner \
+ --deqp /deqp/modules/$DEQP_VER/deqp-$DEQP_VER \
+ --output $RESULTS/cts-runner-results.txt \
+ --caselist /tmp/case-list.txt \
+ --exclude-list $ARTIFACTS/$DEQP_SKIPS \
+ $XFAIL \
+ --job ${DEQP_PARALLEL:-1} \
+ -- \
+ "${DEQP_OPTIONS[@]}"
+DEQP_EXITCODE=$?
if [ $DEQP_EXITCODE -ne 0 ]; then
- exit $DEQP_EXITCODE
-fi
-
-sort < /tmp/unsorted-fails.txt > $RESULTS/fails.txt
-
-comm -23 $RESULTS/fails.txt /tmp/expected-fails.txt > /tmp/new-fails.txt
-if [ -s /tmp/new-fails.txt ]; then
- echo "Unexpected failures:"
- cat /tmp/new-fails.txt
- exit 1
-else
- echo "No new failures"
-fi
-
-sort /tmp/case-list.txt > /tmp/sorted-case-list.txt
-comm -12 /tmp/sorted-case-list.txt /tmp/expected-fails.txt > /tmp/expected-fails-in-caselist.txt
-comm -13 $RESULTS/fails.txt /tmp/expected-fails-in-caselist.txt > /tmp/new-passes.txt
-if [ -s /tmp/new-passes.txt ]; then
- echo "Unexpected passes, please update $DEQP_EXPECTED_FAILS (or add flaky tests to $DEQP_SKIPS):"
- cat /tmp/new-passes.txt
- exit 1
-else
- echo "No new passes"
+ echo "Some unexpected results found (see cts-runner-results.txt in artifacts for full results):"
+ cat $RESULTS/cts-runner-results.txt | \
+ grep -v ",Pass" | \
+ grep -v ",Skip" | \
+ grep -v ",ExpectedFail" | \
+ head -n 50
+ exit $DEQP_EXITCODE
fi