2015-06-11 23:25 GMT+09:00 Joe Hershberger <joe.hershber...@gmail.com>: > Hi Masahiro-san, > > On Thu, Jun 11, 2015 at 5:16 AM, Masahiro Yamada > <yamada.masah...@socionext.com> wrote: >> Commit 9ba9e85f3f1c (net: Fix NET_RANDOM_ETHADDR dependencies) >> accidentally dropped CONFIG_LIB_RAND defines for 14 Blackfin boards. >> >> Prior to that commit, those boards defined CONFIG_LIB_RAND, but not >> CONFIG_NET_RANDOM_ETHADDR. So, commit 9ba9e85f3f1c should not have >> touched them, but in fact it ripped CONFIG_LIB_RAND off from all the >> header files, which caused undefined reference to srand and rand. >> CONFIG_LIB_RAND=y must be revived for such boards. >> >> BTW, this commit indeed makes it better, but even with this fix, >> three boards (bf533-stamp, bf538f-ezkit, cm-bf548) still can not >> build due to region 'ram' overflowed error. This was cause by >> commit 6eed3786c68c (net: Move the CMD_NET config to defconfigs) >> because CMD_NET selects NET, and NET selects REGEX. Eventually, >> some boards were newly enabled with CONFIG_REGEX, increasing the >> memory footprint. A patch is expected to fix the build error. >> >> Signed-off-by: Masahiro Yamada <yamada.masah...@socionext.com> >> --- > > This is surprising. I would have expected moveconfig.py to take care > of this... but I guess since the LIB_RAND option was not selectable it > only dropped the headers and did not add it to the defconfigs. I guess > there's a lesson to learn there about how to use the tool... making > sure to make the option selectable when running it, and only making it > hidden if nothing selects it. I should have have thought of that when > cleaning up Michal's patch. Apologies.
Another lesson to learn: Please do build test before sending/applying a patch. -- Best Regards Masahiro Yamada _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
