-->
Page 9 of 9

Re: Firmware Dump

PostPosted: Mon Sep 08, 2014 5:32 pm
by tinhead
on Windows i did used Trace32 (but seems that i need some scripts to get proper disasm).

There is as well msys binutils with support fo all architectures

http://7shi.hateblo.jp/entry/2013/07/30/011348

the included all-objdump is generating proper disasm.

Re: Firmware Dump

PostPosted: Mon Sep 08, 2014 5:36 pm
by kongo
sounds wrote:Can you explain how you disassembled the rom dump? Just xt-objdump -S?


Well, I tried to convert the binary to an ELF with objcopy and then add the symbols, but objdump didn't really want to disassemble at the symbols anyway, so I got tired and went for a pragmatic approach involving a python script and a synthesized assembly file, which could be assembled and then disassembled :)
It can probably be done the right way with a little patience and some command line tuning.

It seems like xt-objdump requires a '.xt.prop' section that tells it what regions contains instructions/literals/cruft, and won't try to disassemble function symbols otherwise, so I added
Code: Select all.section .xt.prop
.int 0, 0x8328, 2

Re: Firmware Dump

PostPosted: Fri Sep 12, 2014 7:47 am
by tinhead
attached latest fw build, 00160910. As far i know the latest modules (v091) have 00150900 installed.
Since i have no sources of 00150900 i can't tell what are the changes, so feel free to find out ^^

Re: Firmware Dump

PostPosted: Thu Sep 25, 2014 3:50 pm
by Bert
I received mine today, they indeed came with 00150900.

I'll check if my 00150900 matches yours and I'll try 00160910 and report soon.

DNS lookup is neat, but it has a bad typo in the failure response:
Code: Select allAT+CIPSTART="TCP","does.not.exist",80
DNS Fail/r/n


Edit: got AT+CIPSERVER to work, forgot to AT+CIPMUX=1 first.