Luis Chamberlain wrote on 1/19/23 5:08 PM:
As of v5.10 we now have merged commit 14721add58ef267344bee254bc276c9139b7b665
("module: Add more error message for failed kernel module loading")
which helps expand on this being more friendly:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=14721add58ef267344bee254bc276c9139b7b665
Thanks Luis! Can we interpret these commit insertions in relation to this error from 4.19? So "Exec format error" is really meaning "bad WRITE|EXEC flags"? or are we still stuck in the dark?
That's from apply_relocate_add(). The locs are supposed to be
0 and they are not so it bails. That relocation type is 1 so
R_X86_64_64.
https://en.wikipedia.org/wiki/Relocation_(computing)
I can't say I understand much of this, does it mean anything to you, in relation to these particular errors?
Did you really build the module yourself? It doesn't even seem
you have kernel headers.
Well, I have to say that GNU make built the module, after I issued `make` command, is that what you mean? `linux-headers-amd64` package is installed and /usr/src/linux-headers-4.19.0-5-amd64/ and `make /usr/src/linux-headers-4.19.0-5-common/ are there. Plus `make` is entering "/usr/src/linux-headers-4.19.0-5-amd64" to compile the modules
It would be nice to know the actual cause, probably stemming from me and something I did, since: The previous version of the modules were loaded and working great, in fact system uptime is 741 days now. I unloaded/removed those modules and uninstalled the old software, then configured, built, and installed the latest version of the driver/modules. These new modules were also inserted fine several days ago, producing no errors. In order to help out another user with some troubleshooting, I removed the new modules
