Repository navigation
sof_ri_info: new 'reproducible.ri' feature + test - #4889
Conversation
615bed7 to
81f1e58
Compare
Pure comments, no code change. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Also convert 'differ' to all caps 'DIFFER' to highlight different files. No other functional change. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
This broke parsing any extension past the first one. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Fixes the adsp_mft_cse_entry Component added by commit 8885b46 ("tools: sof-ri-info: move parsing adsp mft outside cse mft"). `adsp_mft_cse_entry.entry_file_offset` did not account for the extended manifest. Also fix the large `reader.ext_mft_length + entry_offset` copy/paste/diverge That variable did not seem to be used anywhere yet but it's wrong value was confusing and time-consuming when debugging. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
File offsets were critical in finding the serious begin_off bug fixed in the same series. The information is already there so just show it. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
stderr is often highlighted with a different color. This also makes the usual >/dev/null redirection possible. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
These changes are visible only in -v mode. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Add: ... to reduce pylint warnings to a usable amount. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Signature and date are (for now) the only variable fields when re-signing with the same tool so their locations matter. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
sof_ri_info.py --erase_vars reproducible.ri sof.ri replaces variable fields (signature and date right now) by constants. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Update the scripts/test-repro-build.sh test accordingly. Finally, different people and CI systems can make sure they're building the same firmware. reproducible.ri is sof.ri with variable parts removed. Signed-off-by: Marc Herbert <marc.herbert@intel.com>
81f1e58 to
0c7b95e
Compare
|
I wrote a quick and dirty https://sof-ci.01.org/sofpr/PR4889/build10762/devicetest/?model=CML_RVP_SDW&testcase=check-suspend-resume-with-capture-5 is the usual TIMEOUT thesofproject/sof-test#650. Everything else is green. QB is still running the last force-push but a previous version passed everything: https://sof-ci.01.org/sof-pr-viewer/#/build/PR4889/build7677288 EDIT: https://sof-ci.01.org/sof-pr-viewer/#/build/PR4889/build7679835 is all green too. |
|
@bardliao did you try I tried enabling IPC4 in Kconfig (interesting compiler warnings...) but that alone makes no difference with respect to rimage and the image file, does it? |
@marc-hb Yes, and I do see some difference when I set |
reproducible.ri is sof.ri with variable parts removed.
Finally, different people and CI systems can make sure they're running the same firmware.