On Sat, 2009-09-19 at 10:37 -0500, Nishanth Menon wrote:
Peter Tyser said the following on 09/19/2009 09:03 AM:
On Fri, 2009-09-18 at 21:21 -0500, Nishanth Menon wrote:
This is questionable if this is really required as the av_ static initalized values should have been loaded to sdram as part of the boot process and initialization should have been done.
Is there a reason you need to do this fixup? Based on your commit message its unclear if this patch is really needed...
Essentially, the loaded memory from the NOR looks all corrupted. I am unable to convince myself why the SDRAM is not updated with the static default inits - SDRAM corruption would have cracked everything else and scope measurement looks good too.
Signed-off-by: Nishanth Menon <nm void mem_malloc_init(ulong start, ulong size) {
- u8 i;
- av_[0] = av_[1] = 0;
- for (i = 0; i < 128; i++)
av_[2 + i * 2] = av_[2 + i * 2 + 1] = bin_at(i);- mem_malloc_start = start; mem_malloc_end = start + size; mem_malloc_brk = start;
If you are going to do this fixup, av_ should not be initialized with values (you're currently doing the same initialization 2 times). In general, we could probably shave a bit off of U-Boot's size by leaving av_ uninitialized and implementing your manual calculation of av_ above,
yep.. missed finishing that out.. :(
but I'm not sure why this change should be included in this patch series.
This patch is need for booting SDP3430 from NOR flash.
Once the initialized values for av_ are removed, this patch would be useful for everyone. It looks like it removes a few hundred bytes of code size, so I'm all for it. I'd be a bit concerned about why your board wasn't loading the initialized av_ table though. Seems like it would be indicative of a larger problem...
So in any case, I like the patch because it reduces code size, but I don't think the explanation or commit message of "for an unknown reason it makes my board work" commit message is the best. Its affecting every board, so at a minimum it should describe how it benefits them - eg changing static initialization into dynamic initialization saves some space.
That's my $.02, feel free to proceed as you see fit as I don't have any real power here:)
Best, Peter