[PATCH] Azure/GitLab: Use buildman -T0 for binman, etc, test
There are times where buildman seems to get stuck in this job when in CI. Forcing single-threaded here allows us to complete and move on.
Signed-off-by: Tom Rini trini@konsulko.com --- .azure-pipelines.yml | 2 +- .gitlab-ci.yml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/.azure-pipelines.yml b/.azure-pipelines.yml index 013b170bf251..506e0c0618e5 100644 --- a/.azure-pipelines.yml +++ b/.azure-pipelines.yml @@ -148,7 +148,7 @@ jobs: export UBOOT_TRAVIS_BUILD_DIR=/tmp/sandbox_spl export PYTHONPATH=${UBOOT_TRAVIS_BUILD_DIR}/scripts/dtc/pylibfdt export PATH=${UBOOT_TRAVIS_BUILD_DIR}/scripts/dtc:${PATH} - ./tools/buildman/buildman -o ${UBOOT_TRAVIS_BUILD_DIR} -w sandbox_spl + ./tools/buildman/buildman -T0 -o ${UBOOT_TRAVIS_BUILD_DIR} -w sandbox_spl ./tools/binman/binman --toolpath ${UBOOT_TRAVIS_BUILD_DIR}/tools test ./tools/buildman/buildman -t ./tools/dtoc/dtoc -t diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index 65996a66f0b2..9e0ef8bca9a7 100644 --- a/.gitlab-ci.yml +++ b/.gitlab-ci.yml @@ -170,7 +170,7 @@ Run binman, buildman, dtoc, Kconfig and patman testsuites: export UBOOT_TRAVIS_BUILD_DIR=/tmp/sandbox_spl; export PYTHONPATH="${UBOOT_TRAVIS_BUILD_DIR}/scripts/dtc/pylibfdt"; export PATH="${UBOOT_TRAVIS_BUILD_DIR}/scripts/dtc:${PATH}"; - ./tools/buildman/buildman -o ${UBOOT_TRAVIS_BUILD_DIR} -w sandbox_spl; + ./tools/buildman/buildman -T0 -o ${UBOOT_TRAVIS_BUILD_DIR} -w sandbox_spl; ./tools/binman/binman --toolpath ${UBOOT_TRAVIS_BUILD_DIR}/tools test; ./tools/buildman/buildman -t; ./tools/dtoc/dtoc -t;
On Thu, Mar 04, 2021 at 05:03:52PM -0500, Tom Rini wrote:
There are times where buildman seems to get stuck in this job when in CI. Forcing single-threaded here allows us to complete and move on.
Signed-off-by: Tom Rini trini@konsulko.com
Applied to u-boot/master, thanks!
Hi Tom,
On Fri, Mar 5, 2021 at 6:04 AM Tom Rini trini@konsulko.com wrote:
There are times where buildman seems to get stuck in this job when in CI. Forcing single-threaded here allows us to complete and move on.
Is this a potential buildman bug? +Simon
My concern is this will increase the overall CI time
Signed-off-by: Tom Rini trini@konsulko.com
.azure-pipelines.yml | 2 +- .gitlab-ci.yml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
Regards, Bin
On Fri, Mar 05, 2021 at 10:05:16AM +0800, Bin Meng wrote:
Hi Tom,
On Fri, Mar 5, 2021 at 6:04 AM Tom Rini trini@konsulko.com wrote:
There are times where buildman seems to get stuck in this job when in CI. Forcing single-threaded here allows us to complete and move on.
Is this a potential buildman bug? +Simon
Kinda? Simon and I were talking about it on IRC, and taking this was his suggestion.
My concern is this will increase the overall CI time
It shouldn't, since we only build a single board here. Further, it fixes or papers over CI failures where the job gets stuck.
Hi Tom, Bin,
On Thu, 4 Mar 2021 at 22:51, Tom Rini trini@konsulko.com wrote:
On Fri, Mar 05, 2021 at 10:05:16AM +0800, Bin Meng wrote:
Hi Tom,
On Fri, Mar 5, 2021 at 6:04 AM Tom Rini trini@konsulko.com wrote:
There are times where buildman seems to get stuck in this job when in CI. Forcing single-threaded here allows us to complete and move on.
Is this a potential buildman bug? +Simon
Kinda? Simon and I were talking about it on IRC, and taking this was his suggestion.
My concern is this will increase the overall CI time
It shouldn't, since we only build a single board here. Further, it fixes or papers over CI failures where the job gets stuck.
Yes it should not affect performance.
It is a bit sad though. I think 'papers over' is accurate. I was actually expecting to get a Python error and then be able to debug it. I was not expecting that the bug would just go away.
I have not been able to repeat the problem locally. It is as if buildman hangs somewhere in the builder thread and/or the output never makes it back up to the main thread.
Regards, Simon
participants (3)
-
Bin Meng -
Simon Glass -
Tom Rini