On Thu, Oct 2, 2014 at 5:00 AM, Paul Bolle pebolle@tiscali.nl wrote:
Reported-by: Paul Bolle pebolle@tiscali.nl
- I do not have fedpkg installed so "make prep" failed. Neither did I
test this any other way. As a consequence this is not Signed-off-by, but only Reported-by. I think that testing that nothing breaks is the real work here. I do hope it turns out not to be too much work.
Nope, worked fine and no differences showed in the resulting configs.
- I stumbled on kernel-arm64.patch. That was fun! Anyhow, if one
applies this patch and runs the perl script I just sent, one should see: config-arm64:111: No Kconfig symbol matches 'CONFIG_PCI_XGENE' config-arm64:120: No Kconfig symbol matches 'CONFIG_SBSAUART_TTY' config-generic:5205: No Kconfig symbol matches 'CONFIG_SBSAUART_TTY'
unless one is building for arm64.
Yeah, that one is special.
- Please note that this also takes into account Kconfig symbols added
by patches currently in master. References to their macros were not removed!
Excellent work! I've applied it locally and it should be pushed later today. One question below.
diff --git a/config-arm-generic b/config-arm-generic index 9ddb1f0ccec8..1f07efa9b2fc 100644 --- a/config-arm-generic +++ b/config-arm-generic @@ -79,15 +79,12 @@ CONFIG_GENERIC_CPUFREQ_CPU0=m # Device tree CONFIG_DTC=y CONFIG_DMA_OF=y -CONFIG_PROC_DEVICETREE=y
I remember this change happening upstream, but there are quite a few defconfig files that still have this included. They also include some of the other symbols you've identified, like CONFIG_MTD_CHAR. Have you sent patches upstream to remove these symbols from the arch defconfigs as well?
josh