On Thu, May 23, 2019 at 01:35:02AM +0000, Geordan Neukum wrote: > On Wed, May 22, 2019 at 12:27 PM Greg Kroah-Hartman > <gregkh@xxxxxxxxxxxxxxxxxxx> wrote: > > depends on is better than select. There's a change to depend on UIO for > > this code already in my -linus branch which will show up in Linus's tree > > in a week or so. > > Noted on both accounts. Thanks for the feedback and sorry for the > inconvenience on the latter. > > > Are you sure we need MFD_CORE as well for this code? > > I noticed the build issue when working locally. I was doing > something along the lines of: 'make distclean && make x86_64_defconfig', > selecting 'CONFIG_KPC2000' and 'CONFIG_UIO' via menuconfig, then > running a good old 'make'. From make, I received an error along the > lines of: > > ERROR: "mfd_remove_devices" > [drivers/staging/kpc2000/kpc2000/kpc2000.ko] undefined! > ERROR: "mfd_add_devices" [drivers/staging/kpc2000/kpc2000/kpc2000.ko] undefined! > make[1]: *** [scripts/Makefile.modpost:91: __modpost] Error 1 > make: *** [Makefile:1290: modules] Error 2 > > which appears to indicate that those two symbols are undefined. When > I looked, it appeared that those symbols were exported from the > mfd-core which is why I also threw in a select for that Kconfig > symbol. Assuming that I didn't do something silly above, I'd be happy > to submit a new patch (with only a depends on for MFD_CORE) as I > continue trying to fix up the i2c driver. Yes, a depends for MFD_CORE would be good, can you base it against my staging-linus branch so that fix can go to Linus for this release? thanks, greg k-h _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
