craig_comstock at yahoo.com
Fri Apr 17 00:51:41 UTC 2015
Craig Comstock <craig_comstock <at> yahoo.com> writes:
> Hi all,
> Was starting to work on making a usable phone again and wondered about the
status of nuttx-bb git on osmocom.org versus nuttx source. It seems nuttx
source is more up-to-date and has configs for my device: compal_e86/c139.
Wondering if there is something in nuttx-bb that is unique that I should pay
attention to? Should nuttx-bb be rebased from nuttx upstream?
> I was able to cobble together a nuttx.bin but on loading it always stops
at 88% with the chainloader. The nuttx.bin is 148K which would seem plenty
of room since the C139 has 4000K flash (32MBit) (am I doing that calculation
> I had to fiddle with the MEMORY section of the ld.script in order to make
space for nuttx, I'm not too familiar with this sort of thing so may have
done something wrong...
> /* E86 stacked flash 32mbit flash, 4mbit sram, DBB internal 256kb SRAM
*/ /* 0x800000-0x87ffff */ /* bump up because we have 32mbit instead of
16mbit */ /* compal-loaded binary: our text, initialized data */ LRAM
(rw) : ORIGIN = 0x00800000, LENGTH = 0x00020000 TRAM (rw) : ORIGIN =
0x00820000, LENGTH = 0x00040000 /* compal-loaded binary: our unitialized
data, stacks, heap */ IRAM (rw) : ORIGIN = 0x00860000, LENGTH = 0x00020000
> Originally TRAM was 20000 long and gave me an error on building. Not sure
if I can fiddle with these values or not.
> Attachment (diff2): application/octet-stream, 4233 bytes
> Attachment (osmocon.log): text/x-log, 3339 bytes
I did some debugging on osmocon and it seems it is getting stuck on the
serial read not returning anything, maybe blocking forever waiting for data
coming back. I wasn't sure how the chainloader program was loaded... there
seemed to be 30-40 bytes of code in an array maybe? Maybe the chainloader is
being corrupted? Other chainloaded programs such as rssi work just fine so
maybe my nuttx.bin is causing a problem somehow.
More information about the baseband-devel