There have been at least a few cases where an exception has occurred in a
thread and resulted in buildman hanging: running out of disk space and
getting a unicode error.
Handle these by collecting a list of exceptions, printing them out and
reporting failure if any are found. Add a test for this.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present even if only a single thread is in use, buildman still uses
threading.
For some debugging it is helpful to do everything in the main process.
Allow -T0 to support this.
Signed-off-by: Simon Glass <sjg@chromium.org>
The isAlive() method was deprecated in Python 3.8 and has been removed in
Python 3.9. See https://bugs.python.org/issue37804. Use is_alive() instead.
Since Python 2.6 is_alive() has been a synonym for isAlive(). So there
should be no problems for users using elder Python 3 versions.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
This patch makes buildman create linked working trees instead of clones
of the source repository, but keeps updating the older clones of the
repository that might already exist. These worktrees share "everything
except working directory specific files such as HEAD, index, etc." with
the source repository. See the git-worktree(1) manual page for more
information.
If git-worktree isn't available, silently falls back to cloning the
repository.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
It is interesting to note the number of builds completed per second to
track machine performance and build speed. Add a 'rate' value at the end
of the build to show this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Currently, the following scenario will rebuild the first commit even
though it is not really necessary - the commit sha or the position in the
patchset did not change:
$ git am <local-patch-0001>
$ tools/buildman/buildman -P -E -W -b master mx6
<do some more development work>
$ git am <local-patch-0002>
$ tools/buildman/buildman -P -E -W -b master mx6 <- will rebuild the first
commit as well, even
though nothing has
changed about it.
This is due to the fact that previous results directories get removed
when the number of commits change. By removing the _of_#_ part of the
directory path, the commits will be rebuilt only if the commit sha or the
position in the patchset changes. Also, update the testcase to reflect this
change.
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
At present patman sets the python path on startup so that it can access
the libraries it needs. If we convert to use absolute imports this is not
necessary.
Move patman to use absolute imports. This requires changes in tools which
use the patman libraries (which is most of them).
Signed-off-by: Simon Glass <sjg@chromium.org>
At present buildman sets the python path on startup so that it can access
the libraries it needs. If we convert to use absolute imports this is not
necessary.
Move buildman to use absolute imports. Also adjust moveconfig.py too since
it uses some buildman modules and cannot work without this.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present buildman does not write its own output files (err, done, the
environment) when using -w. However this is useful for when the build is
run with -s to check it.
In fact ProduceResultSummary() reads the result from those files rather
than using the 'result' info directly. So ProcessResult() does not work
with -w at present. It does not print any output.
Fix this by writing output files even when -w is used.
Signed-off-by: Simon Glass <sjg@chromium.org>
These are becoming more common now. They cause boards to show warnings
which can be mistaking for compiler warnings.
Add a buildman option to ignore them. This option works only with the
summary option (-s). It does not affect the build process.
Signed-off-by: Simon Glass <sjg@chromium.org>
Unfortunately the plague of device-tree warnings has not lifted. These
warnings infiltrate almost every build, adding noise and confusion.
Add a buildman option to ignore them. This option works only with the
summary option (-s). It does not affect the build process.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present buildman defaults to running 'mrproper' on every thread before
it starts building commits for each board. This can add a delay of about 5
seconds to the start of the process, since the tools and other invariants
must be rebuilt.
In particular, a build without '-b', to build current source, runs much
slower without -I, since any existing build is removed, thus losing the
possibility of an incremental build.
Partly this behaviour was to avoid strange build-system problems caused by
running 'make defconfig' for one board and then one with a different
architecture. But these problems were fixed quite a while ago.
The -I option (which disabled mrproper) was introduced four years ago and
does not seem to cause any problems with builds.
So make -I the default and deprecate the option. To allow use of
'mrproper', add a new -m flag.
Signed-off-by: Simon Glass <sjg@chromium.org>
When buildman finishes it leaves the last summary line visible, which
shows the number of successful builds, builds with warnings and builds
with errors.
It is useful also to see how many builds were done in total along with
the time taken. Show these on a separate line before buildman finishes.
Signed-off-by: Simon Glass <sjg@chromium.org>
If a progress message is longer than the terminal line it will scroll the
terminal. Limit the messages to the terminal width.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is nice to see the actual number of builds remaining to complete. Add
this in the progress message, using a different colour.
Drop the unnecessary 'name' variable while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
The commit counter is a hangover from when buildman processed each board
for a commit. Now buildman processes each commit for a board, so this
output is never triggered.
Delete it.
Signed-off-by: Simon Glass <sjg@chromium.org>
Fetching updated versions of a repo can take time. At present buildman
gives no indication that it is doing this.
Add a message to explain the delay.
Tidy up a few other messages while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the board names shown with -l are separated by commas. This
makes it hard to double-click to select a particular board. Also it is not
possible to select all boards and paste them as arguments to a subsequent
buildman run, since buildman requires spaces to separate the list on the
command line, not commas.
Change the output format to use spaces instead of commas.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is quite hard to see the list of board for each error line since the
colour is the same as the actual error line. Show the board list in
magenta so that it is easier to distinguish them.
There is no point in checking the colour of the overall line, since there
are now multiple colours. So drop those tests.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the string for each error line is created in _CalcErrorDelta()
and used to create the summary output. This is inflexible since all the
information (error/warning character, error line, list of boards with that
error line) is munged together in a string.
Create an object to hold this information and only convert it to a string
when printing the actual output.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present warnings are shown in yellow in the summary (-s) but magenta in
the detail listing (-e). Use yellow in both.
Signed-off-by: Simon Glass <sjg@chromium.org>
This is useful in some situations, in particular with -w and when building
in-tree. Now that we are more careful about what we remove in
_PrepareOutputSpace(), it should be safe to relax this restriction.
Update the progress information also so it is clear what buildman is
doing. Remove files can take a long time.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present buildman removes any directory it doesn't intend to write
output into. This is overly expansive since if the output directory
happens to be somewhere with existing files, they may be removed. Using
an existing directory for buildman is not a good practice, but since the
result might be catastrophic, it is best to guard against it.
A previous commit[1] fixed this by refusing to write to a subdirectory
of the current directory, assumed to have U-Boot source code. But we can
do better by only removing directories that look like the ones buildman
creates.
Update the code to do this and add a test.
Signed-off-by: Simon Glass <sjg@chromium.org>
[1] 409fc029c4 tools: buildman: Don't use the working dir as build dir
This help is a bit ambiguous. It only does anything if asked to show size
changes with -S. Update the help and the function comments.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is useful to run a simple build and put all the output in a single
directory. Add a -w option to support this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Today when parsing the .sizes files we get a warning about an invalid
line in the file as it's blank. Solve this by checking that we have a
non-blank line prior to processing.
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The 'done' files created by buildman may end up being empty if buildman
runs out of disk space while writing them. At present buildman dies with
an exception when using -s to check the build status. Fix this.
Seriesl-cc: trini
Signed-off-by: Simon Glass <sjg@chromium.org>
At present we should boards with warnings in the same way as those with
errors. This is not ideal. Add a new 'warn' state and show these listed
in yellow to match the actual warning lines printing with -e.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present we don't distinguish between errors and warnings when printing
the architecture summary. Rename the variables to better describe their
purpose.
'Worse' at present means we got an error, so use that as the name.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present messages from the device-tree compiler like this:
arch/arm/dts/socfpga_arria10_socdk_sdmmc.dtb: Warning
(avoid_unnecessary_addr_size): /clocks: unnecessary
#address-cells/#size-cells without "ranges" or child "reg" property
are detected as errors since they don't match the gcc warning regex. Add a
new one for dtc to fix this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Something has changed in the last several month such that when buildman
builds U-Boot incrementally and a new CONFIG option has been added to the
Kconfig, the build hanges waiting for input:
Test new config (NEW_CONFIG) [N/y/?] (NEW)
Since binamn does not connect the build's stdin to anything this waits on
stdin to the build thread, which never comes. Eventually I suspect all the
threads end up in this state and the build does not progress.
Fix this by passing /dev/null as input to the build. That way, if there is
a new CONFIG, the build will stop (and fail):
Test new config (NEW_CONFIG) [N/y/?] (NEW)
Error in reading or end of file.
Signed-off-by: Simon Glass <sjg@chromium.org>
When summarising the builds, add the -U option to emit delta lines for
the default environment built into U-Boot at each commit.
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
Add a new option '-E' for treating all compiler warnings as errors.
Eventually this will pass 'KCFLAGS=-Werror' to Kbuild.
Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
In the case where a new build only decreases sizes and does not increase
any size we still want to report what functions have been dropped when
doing a bloat comparison.
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
One of these has crept in in this commit:
40a808f1 ARCv2: SLC: Make sure busy bit is set properly on SLC flushing
Adjust buildman to handle it.
Signed-off-by: Simon Glass <sjg@chromium.org>
When using #define CONFIG_SOME_OPTION, the value it set to '1'. When using
defconfig (i.e. CONFIG_SOME_OPTION=y) the value is set to 'y'. This results
in differences showing up with -K. These differences are seldom useful.
Adjust buildman to suppress these differences by default.
Signed-off-by: Simon Glass <sjg@chromium.org>
Normally buildman does a full build of a board. This includes creating the
u-boot.cfg file which contains all the configuration options. Buildman uses
this file with the -K option, to show differences in effective configuration
for each commit.
Doing a full build of U-Boot just to create the u-boot.cfg file is wasteful.
Add a -D option which causes buildman to only create the configuration. This
is enough to support use of -K and can be done much more quickly (typically
5-10 times faster).
Signed-off-by: Simon Glass <sjg@chromium.org>
When Ctrl-C is pressed, just exited quietly. There is no sense in displaying
a stack trace since buildman will always be in the same place: waiting for
threads to complete building all the jobs on the queue.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is annoying that buildman does not respond cleanly to Ctrl-C or SIGINT,
particularly on machines with lots of CPUS. Unfortunately queue.join()
blocks the main thread and does not allow it to see the signal. Use a
separate thread instead,
Signed-off-by: Simon Glass <sjg@chromium.org>
Make it clear when buildman actually starts building. This happens when it
has prepared the threads, working directory and output directories.
Signed-off-by: Simon Glass <sjg@chromium.org>
When buildman starts, it prepares its output directory by removing any old
build directories which will not be used this time. This can happen if a
previous build left directories around for commit hashes which are no-longer
part of the branch.
This can take quite a while, so print a message to indicate what is going
on.
Signed-off-by: Simon Glass <sjg@chromium.org>
On a machine with a lot of CPUs this prints a lot of useless lines of the
form:
Cloning repo for thread <n>
Adjust the output so that these all appear on one line, and disappear when
the cloning is complete.
Note: This cloning is actually unnecessary and very wasteful on disk space
(about 3.5GB each time). It would be better to create symlinks.
Signed-off-by: Simon Glass <sjg@chromium.org>
One use-case for buildman is to continually run it interactively after
each small step in a large refactoring operation. This gives more
immediate feedback than making a number of commits and then going back and
testing them. For this to work well, buildman needs to be extremely fast.
At present, a couple issues prevent it being as fast as it could be:
1) Each time buildman runs "make %_defconfig", it runs "make mrproper"
first. This throws away all previous build results, requiring a
from-scratch build. Optionally avoiding this would speed up the build, at
the cost of potentially causing or missing some build issues.
2) A build tree is created per thread rather than per board. When a thread
switches between building different boards, this often causes many files
to be rebuilt due to changing config options. Using a separate build tree
for each board would avoid this. This does put more strain on the system's
disk cache, but it is worth it on my system at least.
This commit adds two command-line options to implement the changes
described above; -I ("--incremental") turns of "make mrproper" and -P
("--per-board-out-dir") creats a build directory per board rather than per
thread.
Tested:
./tools/buildman/buildman.py tegra
./tools/buildman/buildman.py -I -P tegra
./tools/buildman/buildman.py -b tegra_dev tegra
./tools/buildman/buildman.py -b tegra_dev -I -P tegra
... each once after deleting the buildman result/work directory, and once
"incrementally" after a previous identical invocation.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Simon Glass <sjg@chromium.org> # v1
Tested-by: Simon Glass <sjg@chromium.org> # v1
Acked-by: Simon Glass <sjg@chromium.org>
Simon Glass