hi,
just fixed the issue from 29c3, where flashing of phones did not work anymore. if no one complains, i will apply it.
it is also possible to flash c155 phones. the original compal boot loader will start firmware at 0x20000 instead of 0x2000 (c123), so i modified the linker scripts for compal_e99. i am not sure, if all c155 phones behave the same. to check this, i dumped some pages of the flash memory:
run the loader: $ host/osmocon/osmocon -p /dev/ttyUSB0 -m c155 target/firmware/board/compal_e99/loader.compalram.bin
start the phone and dump some flash: $ host/osmocon/osmoload memdump 0x000000 0x30000 dump
a hexdump of that dump showed me where the actual firmware starts:
... 0000800 4f42 544f 392e 2e30 3530 0000 0000 0000 0000810 3031 3330 0101 0000 ffff ffff ffff ffff 0000820 ffff ffff ffff ffff ffff ffff ffff ffff * 0020000 4f43 4544 392e 2e30 3130 0000 0000 0000 0020010 4c46 5845 392e 2e30 3130 302e 0031 0000 0020020 5352 4b50 392e 2e30 3130 302e 0031 0000 ...
i would like to know, if this is true for other c155 phones.
regards,
andreas
just fixed the issue from 29c3, where flashing of phones did not work anymore. if no one complains, i will apply it.
1) Can you make those patch over my testing branch instead ? The build system was completely reworked and the merge is going to suck if you change the makefile in master
2) Your patch enables TX and flash write by default, that's bad.
Cheers,
Sylvain
And wrt to the msgb_pull patch :
- I'm not sure it's the correct way to fix it. Having mgb_pull return a pointer to the end of the bytes we just removed seems a very weird API ... maybe the bug is msgb_pull itself.
- The addition of the handler for KEY_MENU should be a separate patch, not in that one.
Sylvain Munaut wrote:
- I'm not sure it's the correct way to fix it. Having mgb_pull return
a pointer to the end of the bytes we just removed seems a very weird API ... maybe the bug is msgb_pull itself.
for all msgb_pull_u* harald took this into account, so i think there is a reason for that. but i don't like this way either.
- The addition of the handler for KEY_MENU should be a separate
patch, not in that one.
again i missed to remove that when creating diffs...
- I'm not sure it's the correct way to fix it. Having mgb_pull return
a pointer to the end of the bytes we just removed seems a very weird API ... maybe the bug is msgb_pull itself.
Ok, I guess changing the api of pull wouldn't be a great idea because of other users and symmetry to _put ...
But then, I would introduce a msgb_pull_buffer() or _data that returns a pointer to the buffer/data zone that's just been pulled.
Cheers,
Sylvain
Sylvain Munaut wrote:
- Can you make those patch over my testing branch instead ? The build
system was completely reworked and the merge is going to suck if you change the makefile in master
can do that.
- Your patch enables TX and flash write by default, that's bad.
forgot to remove it.
i will send you both patches after finishing.
baseband-devel@lists.osmocom.org
-
Andreas Eversberg -
jolly
-
Sylvain Munaut