This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: RedBoot in Little Endian Mode


Mark Salter wrote:

Yeah, CPU marked IXP425AD.
But i can't understand, really it is impossible to fix it(LE/BE conversion) in the NPE-ethernet driver?
Why not?



The B-step and later CPUs support a "data coherent" attribute bit in page table entries. This bit controls how the CPU views the LE memory area covered by the page table section entry. The networking support for NPE relies on this special view into memory, but the A-step CPUs cannot provide this.

I wouldn't say it is impossible to get around this, but it
isn't worth the effort to support the original run of CPUs.

--Mark



Thanks for nice explaining!

But now i has some problem with loading little-endian kernel.
I has 2.4.27 patched for IXP425 CPU and IXDP425 board support.
This kernel work nice in Big Endian Mode(compiled with armv5b-softfloat-linux-gcc 3.4.3),
but now i build kernel with arm-softfloat-linux-gnu-gcc 3.4.3. And when i try execute, booting
is hanged right after uncomressing(see below).


RedBoot> load -r -v -b 0x01600000 zImage.l
Using default protocol (TFTP)
-
Raw file loaded 0x01600000-0x0168e1bf, assumed entry at 0x01600000
RedBoot> exec -b 0x01600000 -l 0 0x01600000
Uncompressing Linux............................................. done, booting the kernel.


--
Nickolay


-- Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]