Hi,
Le 09/05/2012 19:21, Daniel Stodden a écrit :
On Wed, 2012-05-09 at 08:25 -0500, David Purdy wrote:
On Wed, May 9, 2012 at 3:00 AM, Daniel Stodden daniel.stodden@googlemail.com wrote:
Have you tried to put the bootrom in the debug mode ?
https://doukki.net/doku.php?id=wiki:tutoriels:u-boot.uart#boot_from_uart0
Réfs : http://www.marvell.com/embedded-processors/kirkwood/assets/FS_88F6180_9x_628...
May be it can helps to understand the 1.11 and 1.21 differences ?
I never played with it. But according to the documentation, it's rather orthogonal to image upload. You get a command line interface. However, I see no command references in the docs.
Might support entering it in kwboot in the future, but I don't see how it's would help with the protocol issues.
Daniel
@ Daniel: Just 2 thoughts...
- (less important...) I wonder if it is possible to query the
BootROM version via UART... if possible, on detecting 111 @ offset ff00003c, have it print "Not Supported on V1.11 BootROM"
Unless s/o can point me at documentation for the debug mode, I'm not sure whether that's going to work smoothly. Sounds more prone to future breakage, as the device firmware evolves.
Iff people would really like to see that supported, I'd rather suggest to make it some --test command. Not run ahead of the regular boot mode. Which then could match against a black/white list of firmware versions tested.
I do not know the importance that it may be have. I have done a dump from one of my sheevaplug bootrom (v1.11) and it's big-endianness. The GoFlex found [1] is little-endianness.
I can send the binary dump and results disasm if needed.
- (more importantly) : I am actually _just_fine__ w/ the level of
function in your current V2. You've already done a good service to those work on Kirkwood stuff, and would want to press anyone into fixing something that they/we didn't break and possibly can't be fixed. [There may well be a reason no one has gotten anything w/ the V1.11 BootROM to work with UART-booting.]
For me, your V2 already feels "mature". Its a big step forward from where we were.
Thanks.
I'll fix the select() glitch mentioned before and send out v3.
Daniel