Dear Wolfgang Denk,
Dear Reinhard Meyer,
In message 4CC9E2A7.3080506@emk-elektronik.de you wrote:
Replace CONFIG_RELOC_FIXUP_WORKS by CONFIG_BROKEN_RELOC_FIXUPSeems a slight misnomer to me, and a bit hides what it really does. What about CONFIG_NEEDS_DATA_RELOC_FIXUP ?
Thinking about that, I feel the name is too long. And eventually not clear enough either.
How about CONFIG_NEEDS_MANUAL_RELOC ?
OK, since I had that in mind, too. I skipped it however since "manual" for me smells like really some extra manual labour. But I guess any full descriptive term would be too long. Fact is, that that sort of system wide "manual" reloc is needed for "initialized data with pointer contents" only. It would be nice if that fact would peek out to avoid conceptional misunderstandings.
Would it be possible for you to apply those 3 patches to a testing branch, I am having an odd error trying to git-am patch 2/3, however I can find no wrongs in the patch at the complained about places. 1/3 did apply without problems.
Best Regards, Reinhard