**** BEGIN LOGGING AT Tue Jan 02 03:00:02 2018
**** BEGIN LOGGING AT Tue Jan 02 08:55:02 2018
Jan 02 08:57:13 <Herrie|Laptop> Morning!
Jan 02 10:18:23 <Tofe> Morning!
Jan 02 10:18:32 <Tofe> elvispre: good!
Jan 02 11:45:09 <Herrie|Laptop> elvispre/Tofe: Weird... When I get elvispre's branch + modify the defconfig it fails for me still with the mismatch errors
**** BEGIN LOGGING AT Tue Jan 02 12:23:05 2018
Jan 02 18:47:50 <elvispre> Herrie: So do I now that I have reinstated one of the lifebook patches. Oh well.
Jan 02 18:49:37 <Herrie> elvispre: Hmmz weird :S
Jan 02 18:53:31 <elvispre> Herrie: Just issuing the bb command a second time runs into the C undefined reference problems rather than the section mismatch ones. That must have been (effectively) what I did last night.
Jan 02 18:56:06 <Herrie> elvispre: Hmmz
Jan 02 18:56:10 <Herrie> Weird stuff
Jan 02 19:03:56 <Herrie> elvispre: Seems I might have forgotten to bump SRCREV in my recipe :P
Jan 02 19:03:59 <Herrie> Trying again
Jan 02 19:04:14 <Herrie> 4 mins into compile, so far so good
Jan 02 19:04:20 <Herrie> and failed LOL
Jan 02 20:01:14 <Herrie> Tofe: You have any clues as to these errors? My Googling doesn't give me much useful info except for "try different GCC" :P
Jan 02 20:01:18 <Herrie> Which isn't really an option
Jan 02 20:03:31 <Herrie> On some Russian forum there was said the following: This is a typical section mismatch, caused by the fact that the function is not marked with an initializing function. Or, on the contrary, it is marked, but not correctly. The marking rules are set in include/linux/init.h. Accordingly, it is necessary to understand in essence and either to add to the function the marking, or to remove it
Jan 02 20:03:31 <Herrie> from the variables used in the function.
Jan 02 20:05:58 <elvispre> Herrie: That is, it is necessary to correct the kernel code. There are only 13 mismatches. Perhaps it won't be that hard. It seems odd that we are (nearly) the only people with the issue though.
Jan 02 20:06:21 <Herrie> elvispre: Nobody really uses new GCC with kernels for devices ;)
Jan 02 20:06:30 <Herrie> All Android devs use the Google supplied one usually ;)
Jan 02 20:06:59 <elvispre> Herrie: Nobody *else* :-)
Jan 02 20:11:48 <Herrie> elvispre: Weird thing is that the athene (The Motorola G4, though it uses 3.10 kernel) has very similar code for the bits where we have issues in our 3.4 kernel
Jan 02 20:11:59 <Herrie> So it coudl be something simple as a defconfig value that triggers it
Jan 02 20:12:32 <elvispre> That OnePlus defconfig is massive compared with all the others in that folder.
Jan 02 20:30:27 <Herrie> elvispre: You can add all possible lines but when you set them to "is not set" it doesn't mean much :P
Jan 02 23:30:48 <elvispre> If anyone is interested, my failing kernel build log with config section debugging turned on is here: https://pastebin.com/hbHG45cm
Jan 02 23:31:51 <elvispre> Or rather, https://pastebin.com/hbHG4Scm
**** ENDING LOGGING AT Wed Jan 03 03:00:00 2018