| Message ID | 20230206234344.2433950-3-mcgrof@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | selftests: bump timeout for firmware and kmod | expand |
diff --git a/tools/testing/selftests/firmware/settings b/tools/testing/selftests/firmware/settings index 085e664ee093..75773074af35 100644 --- a/tools/testing/selftests/firmware/settings +++ b/tools/testing/selftests/firmware/settings @@ -5,4 +5,10 @@ # Additionally, fw_fallback may take 5 seconds for internal timeouts in each # of the 3 configs, so at least another 15 seconds are needed. Add another # 10 seconds for each testing config: 120 + 15 + 30 -timeout=165 +# +# That's 165.. but we also now have some other batched tests, we get the +# current timeout value by running manually withtout the timeout: +# +# time ./tools/testing/selftests/firmware/fw_run_tests.sh +# then add give or take about 50 seconds. +timeout=230
Bump the timeout up to what we can empirally need on a q35 8MiB system / 8 vcpus as supported easily now with kdevops [0]. To test firmware_loader with kdevops: make menuconfig # enable dedicated selftests and firmware test make make bringup make linux make selftests-firmware In practice this test now takes about 170 seconds so let's give it a bit more breathing room we end up with 230. Note: I'm seeing a failure only on Android setups running this test in a loop, we eventually OOM on linux-next tag next-20230119. [0] https://github.com/linux-kdevops/kdevops Signed-off-by: Luis Chamberlain <mcgrof@kernel.org> --- tools/testing/selftests/firmware/settings | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)