Age | Commit message (Collapse) | Author |
|
While re-writing result history for
tcwg_gnu_native_fast_check_gcc--master-aarch64
we have detected a corrupted sumfiles/xfails.xfail file,
which previous versions of validate_failures.py happened
to accept. Current version of validate_failures.py complains
on this result, and we need to remove this result from the history.
We are going to do this by failing with $EXTERNAL_FAIL when
validate_failures.py can't process results.
Change-Id: Id6b3abbf9904ab2ddb041a683b11c405e053e5b9
|
|
... so that history re-write can remove result entries that
no longer pass due to previously-removed flaky entries.
Change-Id: I83e826fb69d81bfff35887f2695723a17da48b8f
|
|
and re-generate LNT dashboard files.
Change-Id: I6bb1b814d2cd9d88faeda3dea7de853d23591af0
|
|
Change-Id: Ic4d03097ab214335f44cc736501cce73099fd431
|
|
xfails.xfail
We emit a header for the first two sections of xfails.xfail ('New
flaky tests', if any, and 'Known flaky tests'), but not for the 'Known
failures' although it is mentioned in the beginning of the file.
Add a separator so that this section is easier to find.
Change-Id: I3e04676ec660ef310d73e571993c04aa5dcc7aad
|
|
Change-Id: Idb48c13ee41107cef10c6a219ec45089af9b1367
|
|
Change-Id: I383587d0023327b89af212f42d0379ace2c4807f
|
|
Change-Id: I02b325d14bc419360e7bf9c06e4595d969feab11
|
|
Change-Id: I8d414fb0b4933d3198251b7fc23fb17d71ddb18e
|
|
Change-Id: I01633b5da182edc6cbbeb13b64e49f66e28e19ad
|
|
We unconditionally enabled the read1 library in runs of the GDB testsuite
with Abe commit b6a7eb1e0c82 ("make.sh: Run GDB testsuite with read1
library"). Since Abe commit 1777ba845544 ("lib/make.sh (make_check):
Support various output buffering workarounds"), we need to pass a flag to
enable it, so do that.
Change-Id: Iaf7351fd4227937a000413d45be12fd198e64299
|
|
Change-Id: I4fb639053642ec42ac617cf346305ec123fc1ca9
|
|
... which is now merged upstream.
Change-Id: Ia9a61cc10db987ba1889407ce2f222c3e3957b5c
|
|
Old results were removed together with flaky entries they provided
for the more recent results. Rewrite history to delete results
that depend on the removed flaky entries.
We are going to fix this properly by adding redundant flaky entries
whenever we see a test being flaky.
Change-Id: I56528331c7eecbb8870b14a794b646fc2c1ebb22
|
|
To avoid having useless global variables and tricky overriding rules,
define and use a new gnu_data[] array containing the appropriate
values.
This is the same idea as we use in tcwg_bmk-config.sh with bmk_data[].
Change-Id: I743a036df03004143a412008c245d0877c3fc3e6
|
|
Upgrade only minor version, all artifact history should be compatible
with current version of the scripts. No revision should be droped in
that case.
Change-Id: I2dfa376591fc40cb676e3a6ecf5a491c433a6416
|
|
distro debug info"
This reverts commit 731e34001dff4459f01af90717aa242ea60188b0.
The patch was just committed upstream so we don't need to apply it locally
anymore:
https: //sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=740ce35025a5a37d78b0ee2a363c35534eaa1a2a
Change-Id: I443b8631bf145420befc449811fe2079bd4e5584
|
|
These variables are not used in tcwg_gnu-build.sh, but
settings_for_ci_project_and_config checks they exist.
They are currently defined and used only in round-robin-notify.sh.
Change-Id: Id44bd554e4386d1ffc065a72ef878b20f2d0a46a
|
|
tcwg_gnu-build.sh
that we can add some features to it, and use it from round-robin-notify.sh later
Change-Id: Ib1033df8b68bcbc3fb080c54f071a9b0fc4299ad
|
|
from tcwg_gnu-build.sh"
This reverts commit de4cfe2e7675296e27cf3ee6df4f519f8e7d734a.
Reason for revert: build failure
Change-Id: I7edb188e371e879a6815943def313539f4a87513
|
|
tcwg_gnu-build.sh
So that we can add some features to it, and use it from round-robin-notify.sh later
Change-Id: Ib38ff9e02cf4ba4aface040a1a1bdc0663029050
|
|
... with links to details.txt files in interesting-commits. This
will help with checking whether failures listed in the old reports
are still present in the results.
Change-Id: I518135669241357c131b2f3c37ed00a315350f27
|
|
Change-Id: Ib8aed14f51054596c5a3a6082f9dd949802e74e8
|
|
Use 'crypto-neon-fp-armv8' as 'auto' is not a valid value for
--with-fpu=XXX configure option.
(although -mfpu=auto is a valid GCC option)
Change-Id: I4e505a57ce10aaf626dbc52e281a12dc50122051
|
|
To make validation coverage more consistent for tcwg_gnu_embed:
- use --with-float / -mfloat-abi everywhere (so that we test FP when
our -march/-mtune enables it)
- use -march in target_board options instead of -mcpu because many
tests check if -march is overridden, but do not check -mcpu
- use -mtune mainly as a reminder of which cpu we aim to test with a
given -march
- use -mfpu=auto so that the compiler automatically selects the
appropriate FPU for the given -march
We also support a new arm_v7a_soft_eabi configuration (dropping
arm_v7a_softfp_eabi later), so that we test -mfloat-abi=soft for
A-profile.
Change-Id: I5d2844f91db7751ec1fa8074dc996620ff6b154c
|
|
... for tcwg_gcc_check--master-arm to test history rewrite.
Rewriting got to a point where an old result was removed, and that
removed result had a flaky test detected. Consequently, some
subsequent results now fail to identify that result as flaky,
and fail. Increasing rr[major] will cause those results to be
removed from history.
Change-Id: I19362af03af266ae93f94aa313b27a24e1baa74c
|
|
We cannot use both --with-cpu and --with-arch, so drop
--with-arch=armv8-m.main+dsp for consistency with the other
configurations.
Change-Id: Idc413138131504db1b6941391b7c9106605338d1
|
|
Change-Id: Iebe85747cdf764e91b72ff2580b1dfa18d2dfbe2
|
|
Complete the list of configurations we want to monitor in our CI.
When we use --with-mode=thumb/-mthumb, the config is named thumb*_eabi
instead of arm*_eabi.
For M-profile, we use the processor name as a shortcut, but this makes
it less obvious which architecture level we are testing.
We add the equivalent of the following target-boards:
-mthumb/-march=armv8-a/-mfpu=crypto-neon-fp-armv8/-mfloat-abi=hard
-mthumb/-mcpu=cortex-m0
-mthumb/-mcpu=cortex-m3
-mthumb/-mcpu=cortex-m7
-mthumb/-mcpu=cortex-m23
-mthumb/-mcpu=cortex-m33/-march=armv8-m.main+dsp
-mthumb/-march=armv8.1-m.main+mve/-mfloat-abi=hard
Change-Id: I632c4617e2356566b66ebadfeb175cb97f4663b8
|
|
The string arm_v7a_softfp is a shortcut to mean testing an arm-eabi toolchain with:
-marm -march=armv7-a -mfpu=vfpv3-d16 -mfloat-abi=softfp
We could build the toolchain with the default configuration flags, but
that would mean building a lot of useless multilibs, since we want to
force the above flags when running the tests.
So we also configure & build GCC with
--disable-multilib --with-arch=armv7-a --with-fpu=vfpv3-d16 --with-mode=arm --with-float=softfp
Change-Id: Ieb6c40ad56a3ae3795ad5a74faeb5d0add16a977
|
|
Change-Id: I04d1e77262fd4eea40800a046cff05898c3cd205
|
|
Otherwise, when re-writing history, rr[components] is taken
from the old manifest, and may be out-of-date.
Change-Id: I3169cf12ff0b3193baec0455538d1c5da1d7a3bc
|
|
debug info
When debugging arm targets, GDB needs to find the ld.so debug information
to determine whether it needs to put a Thumb or regular breakpoint
instruction in order to detect shared library loads and unloads.
The problem is that it's currently not possible to install GDB in a custom
installation directory (prefix) and at the same time ask it to look for
debug info in /usr/lib when running the GDB testsuite.
A patch has been posted upstream¹ to add this ability. This commit applies
that patch.
This fixes tests on tcwg_gnu_native_check_gdb on arm that use the distro's
compiler instead of a compiler built by ABE, such as the tests for Ada.
¹ https://inbox.sourceware.org/gdb-patches/20230606233118.146167-1-thiago.bauermann@linaro.org/
Change-Id: Iba02cf98ec314e07815289471cfc3e068c6475a8
|
|
On arm builds this fixes a big number of tests because it allows GDB
finding the distro's ld.so debug info but it should improve results a bit
on aarch64 as well, since a few GDB tests assume presence of libc debug
info.
Change-Id: I1bad4d984bd98207bb11fd5e355d687975aaafff
|
|
Change-Id: If47e57c9d7bde51c8870322f5129b8057875b04b
|
|
Change-Id: I605453907754b876642b16206a75bf243d6561ff
|
|
Change-Id: I7fc821b258bd31ae58b638ee3957797256b84e2a
|
|
... to round-robin-notify.sh.
Change-Id: Ieabec24beb75abd49f44b5e61b500e23462fa56b
|
|
... rather than rely on "build_abe check". This allows
re-generation of results in round-robin-baseline.sh using
up-to-date history of base-artifacts.git.
Change-Id: I19b61782b5080dcc56e292592efc2a8accf2b6d4
|
|
... which causes differences in manifests when rewriting history.
Also, problems with using relative directories should be fixed
by now.
Change-Id: I2b3597bbce9eca07fbb94c112010d3be4aab7b50
|
|
This reverts commit f0e452758d3f10a69bf8afe2ee437113071dc9a7.
The patch has just been pushed upstream.
Change-Id: I61372d4862840cc1a48e38f05d9e9dfe069e501c
|
|
This fixes an infinite loop we are hitting on that job. The patch being
applied is:
https://inbox.sourceware.org/gdb-patches/20230605232844.80533-1-thiago.bauermann@linaro.org/
We can revert this commit after the patch is applied upstream.
Change-Id: I6936c8ea4f76343b022bafa9a03c14ee490858e6
|
|
Move code for update_baseline() from round-robin.sh to its own
script -- round-robin-baseline.sh. This allows us to move
update_baseline() step after notification step, which simplifies
notification logic. For the main workflow round-robin-baseline.sh
is now called from tcwg/round-robin.yaml.inc for successful builds.
The main reason for this reorganization is to simplify re-processing
of results.
Change-Id: I038bfc1ac68b1d6bb6dfd6f36dfb559d8e481583
|
|
tcwg_gnu_embed
Change-Id: I1b38879ce41e31ab6505be66a694e85dfeec76e7
|
|
so that we get less verbose notifications, thus removing lines like:
Manifest: /..../sumfiles/xfails.xfail
Getting actual results from build directory /.....
/.../gcc.sum
[..]
Change-Id: I07a6781955c99df61535dc59b970e9a9c9269680
|
|
"eabi" is used for arm toolchains, while aarch64 toolchains use "elf"
for the bare-metal toolchains, so it's better not to use "gnu_eabi" in
preparation for future jobs monitoring aarch64 bare-metal toolchains.
Since "bare-metal" is a bit long, we use "embed" instead (this name is
also forwarded to patchwork, which imposes a limit on name length)
Change-Id: Id5a5a3225507d8981db1e501a2fc2897f9b3f1ca
|
|
... to cherry-pick a patch for testing. Glibc needs this
to produce dejagnu-style test results.
Change-Id: I338eb1b739e040e3d937aac6990cce3e80a99fad
|
|
... which build glibc using system tools.
Change-Id: Ic77e50777357d35185c2f2384d0d50aeddfce119
|
|
Change-Id: I426097f73ac9a110b49838b449955806cf570624
|
|
Change-Id: I333f9ab734aa71fb5aa6bad52c8038018777201d
|