Hi, On Thu, Dec 18, 2008 at 01:58:43PM +0100, Alexander 'Leo' Bergolth wrote: > >> If I understand the issue correctly, the problem lies in a false > >> positive check on find_task_by_pid. Unfortunately when a check is > >> positive the verbositiy is low, see the output below. But the > >> test is probably hidden in configure somewhere: > > > > [...] > > > >> WARNING: Symbol version dump > >> /srv/atrpms.net/atrpms/kernelsrc/f9-x86_64/2.6.26.6-79.fc9-x86_64/Module.symvers > >> is missing; modules will have no dependencies and modversions. > > > > This is the culprit. If Module.symvers is not there, the test succeeds > > here too. > > > > Now, is it really missing in your kernel tree or is the configure check > > somehow broken?
It depends on the definition. Openafs seems to be the first package that actually probes Module.symvers. And you really only need it for GPL checks. So ATrpms was missing this one and will continure to miss it for some older distributions, but I changed the buildsystem to now use kernel-devel packages were available. > I've uploaded openafs-1.4.8-30.fc10.src.rpm to > [email protected]:openafs/. > > Axel: Could you try to build this? (I couldn't find any change in the > corresponding configure checks since 1.4.7 but maybe I missed something.) I'll try to do so right now. I finally got the buildsystem to use kernel-devel. Sorry for the delay, but the code changes were not trivial. > This version also includes a patch that should help debugging this > configure-check: > stderr and stdout of the find_task_by_pid check should be saved in > confdebug.out and confdebug.dir. In addition, make is called with "-d", > so the config.log output should also contain more information. > > Could you try compiling only the kernel-module (building the > "kmdl_userland 1" part afterwards would overwrite the BUILD directory > and replace the debugging output) and post the output of config.log, > confdebug.out and confdebug.dir? I hope this just builds through, let's see. If it does we can remove the debugging code later on. Thanks to both of you for helping nailing down the issue! -- Axel.Thimm at ATrpms.net
pgpxq7yyEbqrY.pgp
Description: PGP signature
_______________________________________________ atrpms-devel mailing list [email protected] http://lists.atrpms.net/mailman/listinfo/atrpms-devel
