Mani,
Pillai, Manikandan wrote:
Hi all,
Pls ignore the patches which have been sent for single binary image for NAND and OneNAND flashes. I will generate them again against the common branch and resend them.
In this patch, I am fixing some compilation bugs for jffs2 arising out of the upstream changes in nand header files and adding a default mux configuration for
EVM that support some display configuration.
Pls send in your comments for the same.
Thanks for your patches, they are not forgotten
Some remarks, though:
* In general, Steve and I like to wait with applying (new) patches to
Steve's git until basic OMAP3 support is accepted in upstream U-Boot.
Everything else would result in a lot merge work and confusion. The
next U-Boot merge window is supposed to open at mid of october, so
hopefully OMAP3 support is merged then. Hope it is ok for you to wait
some additional days. Maybe we can work on the topics below until then
* It would be easier to review and understand your patches if you could
a) split the patches in small logical parts with one functionality
change/bugfix in one patch. Having a quick look at changes below this
patch contains at least 3 logical independent parts (mux changes,
minor fixes/clean up, nand & jffs changes)
b) don't mix mails, subjects and patches. In this mail I count 4 (!)
different subjects:
[PATCH] - Fixing some MUX and compilation issues on the OMAP3 EVM
board(Patch1/1)
[PATCH] - Single binary image for NAND/OneNAND based OMAP3 board(Patch1/1)
Patch for fixing the nand hw mode write - Patch(1/1)
[PATCH] Changes for NAND compilation and EVM pin mux.
Quite confusing
c) send OMAP3 independent patches directly to U-Boot list for inclusion.
* Regarding mux.h I'd like to discuss how we can do this more generic
with less code.
Yes, having a #define MUX_DEFAULT_xxx() for each board in mux.h is the
easiest way. But adding additional mux settings for additional boards
will result in a quite large mux.h, soon. So we should discuss how we
can replace the existing #define MUX_DEFAULT_xxx() with a more generic
and scalable solution.
I think we should have something like one global default mux setting
for all boards and then overwrite the (hopefully small) *different*
board specific parts with something in the board files. E.g. something
like we already do with the "const omap3_sysinfo sysinfo" logic we
introduced recently.
Unfortunately, I don't have an idea how this could look in real code,
yet Any proposals?
Thanks
Dirk