Sansa Fuze+

[kerobaros]

Okay, I have compiled a bootloader.elf, and am trying to pack it into the .sb file following the instructions on the wiki. However, elf2sb segfaults every time I run it.

From my dmesg:

[ 3185.897354] elftosb[13167]: segfault at 30303044 ip 0804cb60 sp bf8db8b8 error 4 in elftosb[8048000+7000]
[ 3206.972947] elftosb[13229]: segfault at 30303044 ip 0804cb60 sp bfffbf48 error 4 in elftosb[8048000+7000]
[ 3227.578542] elftosb[13292]: segfault at 30303044 ip 0804cb60 sp bfe53768 error 4 in elftosb[8048000+7000]
[ 3279.296148] elftosb[13445]: segfault at 30303044 ip 0804cb60 sp bf90be38 error 4 in elftosb[8048000+7000]

Running Arch Linux x86, compiled the toolchain using the automatic method from the wiki. Help?

[pamaury]

Strange,
normally elftosb is compiled in debug mode by default, can you use gdb or valgrind to see where it segfaults ?
If you don't know how to do this or just can't, could you send me both the bootloader.elf file and the elftosb binary ?

[kerobaros]

(gdb) run bootloader.db fuze+_key_file.txt bootloader.sb
Starting program: /home/kerobaros/fuze/elftosb bootloader.db fuze+_key_file.txt bootloader.sb

Program received signal SIGSEGV, Segmentation fault.
0x0804cb60 in elf_get_nr_sections (params=0x8050018) at elf.c:467
467           sec = sec->next;

Does that help at all? (Also: I didn't know how to use gdb, but apparently it's time to learn!)

[pamaury]

That's really strange, I don't how it could segfault at this point.
The simplest solution is that you send me all the related files (the elf files which come from sbtoelf and bootloader.elf) since it will take you lots of time to debug it if you don't know how the program works.

[kerobaros]

Alright. I sent you a message on IRC; could we discuss this there?

[pamaury]

I fixed the bug, there was an uninitialized structure in the code. I think you can retry now

[metaphys]

Ok it's working. Should we test anything specific? For what I can see REL and Abs values are correct and there is also a value changing when pushing volume key

[pamaury]

I have committed somes changse, now there is a graphical feedback! For now that's just for fun / testing that the code works on all devices. I will now move on to the storage code.

[kerobaros]

Just played with your test code for a few minutes. So nice to see something on that screen that isn't the Sandisk firmware! Works perfectly, pamaury, keep up the good work.

[metaphys]

Very nice indeed  . On my device the upper part of the red square is unreachable (like 0,5 cm) And on the bottom part the cursor go out of the screen... Don't know if this it important, so I prefer to mention it...

Great job anyway! 

[pamaury]

So the red rectangle fits in the screen but the cursor goes out of the rectangle ?
Normally that's impossible :-/ Can you copy the values displayed on the screen (and in particular x_max and y_max) ?

[metaphys]

RMI: id=1533 p=B7 s=E
xmax=3009 ymax=1974 res=82

I retested it precisly:
I unable to exced about 1700 (record 1696) on the vertical axe (by wich value I'm already at the end of the touchpad).
between 50-70 the point disapear from screen (througt not reaching the red line) but I see it's still tracked down because of the vertical value that are still droping to nearly zero
By the way by the time I wrote I've tested again and now this does not happen anymore (for the bottom part) but unstead, I reached the end of touchpad (no more tracking of value) between value 80-100.
I'm the one who take his device appart, so pehraps I just broked it partly? I does feel any problem when using it anyway...

EDIT:
Altogether I would say the cursor covert pretty well the whole surface except for the upper part of the red rectangle

[pamaury]

I don't reach the extreme values either but normally the cursor CAN'T escape the rectangle by design (the value reported are always below the maximum) so I found it a bit strange

[metaphys]

no no it does not! But being all the way down and disapearing leave the impression it's gone out of the screen... Plus the fact value are still tracked down (but I can't reproduce this right now...)

[pamaury]

Hello,
just a quick post to say that I'm still alive but as I said previously, I don't have much time right now. I've tried to make the storage work (it's SD based) but without success for now...Perhaps I'll make some progress at the devcon next week-end