Re: removing Unfilled ROM holes

well but it is eboot... I have to devide eboot into 2, and the first half of
it fit in the xip region of nand flash. which is 2K in size.
it's 27x, btw
right now i am putting boot reason, memory, mmu init, copy code to sdram and
continue, into the first half.
the second half is the C code, which contain ethernet stuff.

anyway targeting code 2K in size, and my code around 10K now, still trimming
code. but the memory hole is 8K!?! no matter how hard I do, the 8K hole and
the 4K jump will kill the 2K xip region anyway!

I am still contacting msys's fae, his sample code for another arm based
platform is exactly 2K and don't have the 4K jump. I wonder why...
"Steve Maillet (eMVP)" <nospam1@xxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:u9DHJVzdFHA.2548@xxxxxxxxxxxxxxxxxxxxxxx
> Executable code needs to be aligned on specific boundaries so the ROMIMAGE
> tool aligns them properly, but that can often leave holes in the memory
> used if the code doesn't match the exact size of the alignment. So,
> ROMIMAGE and the ROM filesystem will fill in those holes with data from
> data files as best it can. So the presence of holes is not something that
> you can change and you can't do anything about how the tools fill them
> either. At best you can combine Dlls together and link code statically
> instead of using a DLL to change how it fits in memory reducing the sizes
> of the holes.
>
> --
> Steve Maillet
> EmbeddedFusion
> www.EmbeddedFusion.com
> smaillet at EmbeddedFusion dot com
>
>


.

Relevant Pages

  • Re: JCL
    ... The refrigerator sized cabinet not only housed this core memory, ... Speaking of paper tape, anyone ever have to bootstrap a machine by entering ... Another new fangled invention by IBM: 96 column cards! ... Not the standard rectangular 80 column ones with rectangular holes ...
    (comp.sys.hp.mpe)
  • Re: removing Unfilled ROM holes
    ... I have to devide eboot into 2, ... > right now i am putting boot reason, memory, mmu init, copy code to sdram ... but that can often leave holes in the memory ... >> smaillet at EmbeddedFusion dot com ...
    (microsoft.public.windowsce.platbuilder)
  • Re: [PATCH 2/2] cciss: disable dma prefetch for P600
    ... falling off into one the holes on IPF and AMD. ... It doesn't happen on Proliant because the last 4kB of memory is ... prefetching was walking off the end of real mmeory and into the AGP region ... There is a bug in the DMA engine that that may result in prefetching ...
    (Linux-Kernel)
  • Re: hash table sizes
    ... >> NUMA with memory holes between nodes at the moment, ... > holes just below 4GB regardless. ... do - were you seeing this on x440 or NUMA-Q? ... send the line "unsubscribe linux-kernel" in ...
    (Linux-Kernel)
  • Re: Kernel Dump
    ... > IA64 and Sparc systems usually had sparse memory configurations and the ... I reused most of the ia64 code. ... > The problem is that x86 machines are increasinly having memory holes. ...
    (freebsd-arch)