crash 8.0.6 -> 9.0.2 + crash-memory-driver by RobertBerger · Pull Request #1056 · openembedded/meta-openembedded

RobertBerger · 2026-06-09T07:00:15Z

Updated crash recipe from 8.0.6 to 9.0.2

Slightly different approach, which seems to be a bit less hacky
Some patches were removed since they seem to be obsolete now
crash-native builds and runs
crash-cross-canadian-aarch64 builds—I did not build an SDK to see if it runs.
crash for aarch64 target builds and runs

# target binary:
crash --buildinfo
build_command: crash
build_data: reproducible build
build_target: ARM64
build_version: 9.0.2
compiler version: aarch64-resy-linux-gcc (GCC) 15.2.0

# crash-native:
crash-native/9.0.2/recipe-sysroot-native/usr/bin/crash --buildinfo
build_command: crash
build_data: reproducible build
build_target: X86_64
build_version: 9.0.2
compiler version: gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0

Added recipe for crash memory driver

memory driver readme

kraj · 2026-06-09T16:59:31Z

I think its using 0001-hacked-Makefilefor-module.bbclass.patch but this patch is missing. Its failing as a result see

https://autobuilder.yoctoproject.org/valkyrie/api/v2/logs/7106116/raw_inline

OldManYellsAtCloud

I have added a couple notes and questions, mostly small nitpicks.

OldManYellsAtCloud · 2026-06-09T17:11:01Z

+-# This program is distributed in the hope that it will be useful,
+-# but WITHOUT ANY WARRANTY; without even the implied warranty of
+-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+-# GNU General Public License for more details.


Can we just leave the license part as it is?

Yes, we can keep the license of the Makefile: 7998a78

OldManYellsAtCloud · 2026-06-09T17:16:57Z

Looking at the referenced commit (crash-utility/crash@772fbb1), it seems to be included starting version 9.0.0 - though it wasn't included in the new patch for gdb 16.2.
Is this patch still needed?
I got the same question for 0005-Fix-build-failure-on-32bit-machine-i686.patch file.

To be honest, I have no idea. I kept the patches, which were cleanly applied.

Looking at it closer, both of these are dangling patches, neither of them are used in the recipes. They can be dropped.

OldManYellsAtCloud · 2026-06-09T17:22:47Z

+HOMEPAGE = "https://github.com/crash-utility/crash/tree/master/memory_driver"
+RECIPE_MAINTAINER = "Robert Berger <Robert.Berger@ReliableEmbeddedSystems.com>"
+LICENSE = "GPL-2.0-only"
+LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/GPL-2.0-only;md5=801f80980d171dd6425610833a22dbe6"


Not a very strong opinion, but I think referencing the source file's license declaration would be better: https://github.com/crash-utility/crash/blob/master/memory_driver/crash.c#L9
Also, it seems to be gpl-2-or-later, not only

I can reference the source file's license declaration. It looks like gpl-2-or-later, but the last line says MODULE_LICENSE("GPL"); which means it's GPL-2.0-only. This is needed if a kernel module needs to have access to all kernel symbols.

Haha, yeah, ambiguous license is always great.
Personally I consider the license's text itself to be stronger than the macro (I believe the kernel also only cares about being compatible with gpl2, though I might have dreamed that up).
I won't die on this hill though.

note to myself: done locally

OldManYellsAtCloud · 2026-06-09T17:23:17Z

+           file://0001-hacked-Makefilefor-module.bbclass.patch;striplevel=2 \
+           file://0002-add-sparse-support.patch;striplevel=2 \
+"
+PV = "9.0.2+git${SRCPV}"


SRCPV is not needed anymore, it can be dropped

note to myself: done locally

OldManYellsAtCloud · 2026-06-09T17:24:37Z

+SUMMARY = "Kernel analysis utility for live systems and core dumpfiles"
+DESCRIPTION = "The core analysis suite is a self-contained tool that can be used to \
+investigate either live systems, kernel core dumps created from kdump, or mcore dumpfiles."
+HOMEPAGE = "https://github.com"


Probably https://github.com/crash-utility/crash/ or https://crash-utility.github.io/ would be better for homepage

note to myself: done locally

OldManYellsAtCloud · 2026-06-09T17:27:38Z

+export LDFLAGS
+export AR
+export LD
+export RANLIB


Are these exports needed explicitly? Most (all?) of these are exported by default. Or am I missing some context?

note to myself: done locally (removed all the exports not just line 73 to 76. My test builds went through without the exports.)

OldManYellsAtCloud · 2026-06-09T17:28:54Z

+#  build_target: X86_64
+#  build_version: 9.0.2
+#  compiler version: gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
+# <--


Was this comment left here accidentally? Or is it useful for something?

It proves that 0001-Use-CC-env-var-to-get-compiler-version.patch is applied and so the right compiler is being picked up. But I can remove the comments.
note to myself: done locally

OldManYellsAtCloud · 2026-06-09T17:29:56Z

-# http://errors.yoctoproject.org/Errors/Details/186964/
-COMPATIBLE_HOST:libc-musl = 'null'
+ERROR_QA:remove = "buildpaths"
+WARN_QA:append = " buildpaths"


If this is really unavoidable, could you please add some notes about it in the commit message (or even as a comment in the recipe), why this is needed?

I added the following:

# WARNING: crash-9.0.2-r0 do_package_qa: QA Issue: File /usr/bin/crash in package crash contains reference to TMPDIR [buildpaths] # # WARNING: crash-cross-canadian-aarch64-9.0.2-r0 do_package_qa: QA Issue: File /opt/phytec-ampliphy-resy-systemd/6.0.98-devel/sysroots/x86_64-resysdk-linux/usr/bin/aarch64-resy-linux/crash in package crash-cross-canadian-aarch64 contains reference to TMPDIR [buildpaths] # ERROR_QA:remove = "buildpaths" WARN_QA:append = " buildpaths"

note to myself: done locally

I mean more like, why does this error happen now? Can it be corrected, instead of being ignored? It happens with some recipes that there is no choice but to ignore the QA error, but in case of this recipe this is a regression, it wasn't needed so far. The real question is, what happened that made this needed?

As a sidenote, I suspect the --with-sysroot=${STAGING_DIR_TARGET} is causing this in EXTRA_OEMAKE. Should this rather be --with-build-sysroot=${STAGING_DIR_TARGET}? (This is mostly a blind guess after taking a quick glance at the config file and the error. It can be totally wrong.)

OldManYellsAtCloud · 2026-06-09T17:46:50Z

+           ${GNU_MIRROR}/gdb/gdb-16.2.tar.gz;name=gdb;subdir=gdb \
           file://7003cross_ranlib.patch \
-           file://0001-cross_add_configure_option.patch \
-           file://donnot-extract-gdb-during-do-compile.patch \


Not all of these patch files were removed from the tree (e.g. this donnot...patch) - could you please also drop all the files that are not needed anymore?

I am not sure if they are needed or not. We will see once people start using the new crash version. That's the reason I left them there. But sure, I can remove them for now and keep only the ones that I use.
note to myself: done locally

OldManYellsAtCloud · 2026-06-09T17:50:52Z

+
+    # 5. Return to the active workspace build root to merge downstream additions
+    cd ${B}
+    if [ -f "gdb-16.2.patch" ]; then


I think this is overly defensive, and this if could be dropped: If upstream renames this patch after an update, it is easy to miss, because the lack of patch is swallowed silently. I think we should complain loud if this file doesn't exist, so the recipe can be updated also.

note to myself: done locally

kraj · 2026-06-11T15:35:18Z

@RobertBerger please squash all changes into one commit.

Security fixes: ================= - Remove import-time loading of timezone offset data from pickle to prevent unsafe deserialization from packaged data - Replace eval() use when parsing no_word_spacing with strict boolean parsing to prevent code execution from locale metadata (openembedded#1056) New features: ============= - Add support for expressions like "N {interval} from now" in English (#1271) - Add support for the en-US locale (#1222) Fixes: ======== - Honor REQUIRE_PARTS for ambiguous month-number inputs by retrying with a year-biased DATE_ORDER (#1298) - Fix parsing word-number relative phrases such as "two days later" (#1316) - Allow md5hash to work in FIPS environments (#1267) Improvements: ============= - Add Bosnian Cyrillic (ijekavica) date translations (#1293) - Add a new browser-based demo to the project documentation (#1306) - Update installation documentation to replace setup.py install guidance - Add a project security policy Signed-off-by: Wang Mingyu <wangmy@fujitsu.com> Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>

1) crash-memory-driver: fixed typo to include proper patch 2) crash-memory-driver: keep license,... patch against comments on pull request: 1) HOMEPAGE updated 2) UPSTREAM_CHECK_URI updated 3) exports removed 4) if [ -f "gdb-16.2.patch" ]; then... removed 5) notes about QA Issue: reference to TMPDIR [buildpaths] added 6) notes ./crash --buildinfo removed 7) LIC_FILES_CHKSUM: use license text in crash.c 8) PV = "9.0.2+git${SRCPV}" -> SRCPV removed 9) removed unused patches 10) crash: LICENSE = GPL-3.0-only

RobertBerger · 2026-06-13T05:52:47Z

@RobertBerger please squash all changes into one commit.

I tried to squash it all into a single commit. Please check.

OldManYellsAtCloud suggested changes Jun 9, 2026

View reviewed changes

RobertBerger requested a review from OldManYellsAtCloud June 10, 2026 09:52

RobertBerger force-pushed the crash-9.0.2 branch from 80c07da to 089c28d Compare June 13, 2026 05:50

Conversation

RobertBerger commented Jun 9, 2026

Uh oh!

kraj commented Jun 9, 2026

Uh oh!

OldManYellsAtCloud left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

OldManYellsAtCloud Jun 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

OldManYellsAtCloud Jun 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

OldManYellsAtCloud Jun 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kraj commented Jun 11, 2026

Uh oh!

RobertBerger commented Jun 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

OldManYellsAtCloud Jun 9, 2026 •

edited

Loading

OldManYellsAtCloud Jun 9, 2026 •

edited

Loading

OldManYellsAtCloud Jun 9, 2026 •

edited

Loading