2.6.34-rc3: Can't insmod after make, because versions now differ?!
Hi!
I guess version check is needed, but this is excessive:
Apr 5 07:33:16 amd kernel: udlfb: version magic
'2.6.34-rc3-00345-ge8240f9-dirty SMP mod_unload CORE2 ' should be '2.6.34-rc3-00344-g548fc0a-dirty SMP mod_unload CORE2 '
Yes, I played with git and recompiled my kernel in order to get new module version, but... not even insmod -f works.
Having to reboot every time I recompile the module or play with git makes modules useless for development :-(.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi,
Quoted 2 lines
Having to reboot every time I recompile the module or play with git makes modules useless for development :-(.
Confirmed. -rc3 (from 2.6.33 - sorry) was the only kernel I ever observed this on (I simply changed a driver and redid make modules_install, boom, version string (/lib/modules/...) changed from 2.6.34-rc3 to 2.6.34-rc3-...-dirty, despite having kept my git unupdated and on the 2.6.34-rc3 tag.
Andreas Mohr
Quoted 9 lines · Pavel Machek
Apr 5 07:33:16 amd kernel: udlfb: version magic'2.6.34-rc3-00345-ge8240f9-dirty SMP
mod_unloadCORE2 ' should be '2.6.34-rc3-00344-g548fc0a-dirty SMPmod_unloadCORE2 'Yes, I played with git and recompiled my kernel in order to get new module version, but... not even insmod -f works.
Having to reboot every time I recompile the module or play with git makes modules useless for development :-(.
Disable CONFIG_LOCALVERSION_AUTO?
Quoted 5 lines
Quoted 4 lines · Pavel Machek
Apr 5 07:33:16 amd kernel: udlfb: version magic'2.6.34-rc3-00345-ge8240f9-dirty SMP
mod_unloadCORE2 ' should be '2.6.34-rc3-00344-g548fc0a-dirty SMPmod_unloadCORE2 '
Quoted 7 lines
Quoted 5 lines
Yes, I played with git and recompiled my kernel in order to get new module version, but... not even insmod -f works.
Having to reboot every time I recompile the module or play with git makes modules useless for development :-(.
Disable
CONFIG_LOCALVERSION_AUTO?
Yes, but... should module version checking be fixed so that it does not interfere like that?
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Quoted 2 lines · Pavel Machek
Yes, but... should module version checking be fixed so that it does not interfere like that?
I don't think so. Essentially you are getting what you configured: an extremely fine-grained kernel version. It's up to you to decide whether the benefits of that outweigh the limitations it brings.
Stripping parts of the kernel version for module compatibility checks could interfere with suffixes explicitly added by distros to specify ABI compatibility.
Cheers,
FJP
Quoted 7 lines · Frans Pop
Quoted 2 lines · Pavel Machek
Yes, but... should module version checking be fixed so that it does not interfere like that?
I don't think so. Essentially you are getting what you configured: an extremely fine-grained kernel version. It's up to you to decide whether the benefits of that outweigh the limitations it brings.
Yes, I want descriptive version strings. And yes, I'd like to use modules, and do development.
Unfortunately, these two seem mutually exclusive. Given that Linus made noises (iirc) about 'everyone should set localversion_auto'...
Quoted 3 lines
Stripping parts of the kernel version for module compatibility checks could interfere with suffixes explicitly added by distros to specify ABI compatibility.
So split the version into regular parts and very detailed parts, putting explicit user options into 'regular'?
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html