sof-logger: add checksums to -d(ump_file) option

[marc-hb]

3 commits

Sample output:

logger ABI Version is 5:3:0
ldc_file ABI Version is 5:3:0
ldc_file src checksum 0x07d4f1ad <=== NEW!
Loaded FW expects chksum 0x07d4f1ad <=== NEW!
Components uuid dictionary size: 2400 bytes
Components uuid base address: 0x1fffa000
...

[marc-hb]

Everything is green except this checkpatch warning which was already there before this PR:

WARNING: quoted string split across lines
#74: FILE: tools/logger/logger.c:47:
+	fprintf(stdout, "%s:\t -v ver_file\t\tUse ver_file instead of "
+ "/sys/kernel/debug/sof/fw_version\n", APP_NAME);

I suspect the purpose of this warning is to make sure error messages can be git greped. Not really an issue here.

[ktrzcinx]

What motivation stands behind this change ?

[marc-hb]

Thanks @ktrzcinx for reviewing this PR, really appreciated.

This is what allows the -d ldc_file option to ALSO show the checksum expected by the firmware - when and only when a firmware is running. See sample output in the description. I suggest looking at each commit independently if you have looked only at the entire diff for now.

[ktrzcinx]

Ok, I've got it!

[ujfalusi]

Does the duplicated checksum gives anything for the user?

ldc_file src checksum 0x07d4f1ad <=== NEW!
Loaded FW expects chksum 0x07d4f1ad <=== NEW!

Would not be better to print a single line of the two checksum matches:

ldc checksum: 0x07d4f1ad

If there is a mismatch between the two checksum then:

lcd_file src checksum is 0x07d4f1ad while the loaded firmware expects 0x1b67cba9

[marc-hb]

Would not be better to print a single line of the two checksum matches:

I don't think so because a single line is also printed when the firmware is not loaded.

If there is a mismatch between the two checksum then:
lcd_file src checksum is 0x07d4f1ad while the loaded firmware expects 0x1b67cba9

I personally prefer having the checksums close to each other:

ldc_file src checksum 0x07d4f1ad
Loaded FW expects chksum 0x1b67cba9

... and this makes the code simpler because it does not have to compare them.

[lgirdwood]

C style

[marc-hb]

Wait, I thought #5046 meant this is allowed now?

[lgirdwood]

Fine in the SPDX, but my pattern matching needs C style in the code.

[marc-hb]

Done.

[marc-hb]

Unusual, isolated and totally unrelated timeout in https://sof-ci.01.org/sofpr/PR5110/build11443/devicetest/?model=ADLP_RVP_NOCODEC&testcase=check-pause-resume-capture-10, everything else is green.