Hi all,
I've been working at cleaning up some lingering loose ends on the i386 architecture. This patch series cleans up the bulk of them. This series in itself is really a number of smaller incemental steps needed to get to where I wanted to be.
Patches 1 though 6 only effect i386 specific code (cpu/i386, lib_i386 etc) these patches are needed to get my development platform (eNET board) up and running again (mainline has drifted away a little since I last worked on this board and a new gcc caused some compile issues as well)
Patches 7 through 9 effect drivers outside the i386 build tree, however as far as I can tell, these drivers are only used by the sc520_cdp and sc520_spunk boards
Patches 10 and 11 fix up the sc520_cdp and sc520_spunk boards. I did not include these fixes in the above patches because: a) The boards were already broken, and; b) It made more sense to me for this patch series to be a series of little steps towards the end fix rather than one big jumbled jump
As of patch 11, all i386 boards build clean. sc520_cdp and sc520_spunk could probably work now, but without either board available to test, I cannot confirm.
Patches 12 and 13 bring a couple of i386 components inline with u-boot design philosophy
Patch 14 adds PCI support to the eNET board in preparation for getting networking up and running
At this point, I noticed that my board would sometimes hang during Flash initialisation - looked like probing the Boot Flash was conflicting with the execute in place. I don't really know why it started to happen, but I decided it was about time to move u-boot into RAM which is what patches 15 through 17 do. I have not implemented proper relocation. Sorry Wolfgang, I know you'll hate what I have done, but it is the next step towards what will hopefully one day be real relocation.