Finding the serial (UART) pinout on KF
- Thread starter csholmq
- Start date
Command line of the kernel is "console=ttyO2,115200n8 mem=463M@0x80000000 init=/init vram=5M omapfb.vram=0:5M" .
I change it to yours and set"Always use the default kernel command string" and build again.
Last edited:
If it's no output totally , I think the problem is the boot.img . The kernel not be started at all.
I think he got output from the boot loader. He posted it in his post the other day. Maybe you could just switch out the kernel in a know-to-work boot.img?
Depends on the rom. He needs to use the cm7 ramdisk if using cm7 and the cm9 ramdisk if using cm9
What I meant was, he should take a boot.img from an ICS build that is known to work. And then replace the kernel in it. But if the boot loader is already giving output, then there's really no issue here.
He isn't getting output from the bootloader. At least not with the 3.0+ kernel. Look at the output he posted in that post:
The output he posted was from before trying the 3.0 kernel.
Code:
...
Starting kernel ...
Uncompressing Linux... done, booting the kernel.
[ 0.000000] Initializing cgroup subsys cpu
[ 0.000000] Linux version 2.6.35.7+ (ubuntu@ip-10-174-42-153) (gcc version 4.4.1 (Sourcery G++ Lite 2010q1-202) ) #1 SMP PREEMPT Tue Jan 10 06:22:28 UTC 2012
...The output he posted was from before trying the 3.0 kernel.
Last edited:
Let the games begin:
http://pastebin.com/6tEB3PBr
http://pastebin.com/6tEB3PBr
It means the uart is working. So now we can finally get a working 3.X kernel
Sent from my HTC Sensation XE with Beats Audio using XDA App
Give it time, this all just happened like 5 hours ago. Have patience
Sent from my SCH-I500 using xda premium
This is great news! I feel I won't even get a chance to test my backport before a 3.0 kernel is released! 
Sent from my Galaxy Nexus using XDA App
Sent from my Galaxy Nexus using XDA App
It was 2.30 am local time for Hashcode when he tried it. He needs to build the 3.0 with enabled U-boot, UART etc before debugging.
awsome work guys.
wish I could help, the only way i can is if you need something testing, I'm willing to try anything risky.
thanks for your work, you guys sound like wizards.
wish I could help, the only way i can is if you need something testing, I'm willing to try anything risky.
thanks for your work, you guys sound like wizards.
Similar threads
- Locked
Top Liked Posts
-
There are no posts matching your filters.
-
19Hey, can you give us any infos on the progress?
is everything working as expected, does the serial console really work and is helpful? would be interesting to hear if
Sure!
The UART Serial is working it's magic and up until a couple of days ago, we were able to see how the kernel was unpacked and the bootloader handing it off with
but after that it went blank. Hashcode then enabled CONFIG_DEBUG_LL and the kernel starting spewing it's gutsCode:Starting kernel … Uncompressing Linux… done, booting the kernel.
So piece by piece, the kernel is booting further and further. No screen yet, so the UART is still neccessary for development. Atm I believe it's the charging regulator causing most problems. So that needs to be tackled.
Just a reminder: No, there's no ETA. When there's something to show, it'll be posted.188
Thanks to jamez70's, we now know where the serial pins are.
Here is where the UART pins sits on the back of the motherboard.
(Courtesy of ifixit photo and jamez70's illustration)
The pins from 1-4 are:
RX TX GND GND (where #4 is marked)
From probing, we can detect the 1V8 level signal and the baud rate of 115200.
(Courtesy of jamez70)
If you want to hook this up, I suggest you solder on some wires to the pads and run the wires to the front, making life easier to re-assemble the KF.
I suggest you get something like the Breakout Board for FT232RL USB to Serial for hooking this up. Note that the KF is 1V8, so when you want to connect this unit (default 3V3), you may want to use a level shifter, or simply find the VCC on the KF and loop that back that through the VCCIO on the Sparkfun-converter. Sending 3V3 directly into the KF could possibly seriously hurt it!
If you want to construct a simple voltage divider, simply get two resistors and get your soldering gun out. According to Vout = Vin * Rb/(Ra+Rb)
Voltage divider circuit (courtesy of calculatoredge.com)
- Vin: 3V3
- Vout: 1V8
- Ra: 10 kOhm
- Rb: 12 kOhm
If you need more UART details, here are some tips:
- Baud: 115200
- Bits: 8
- Parity: None
- Stop Bits: 1
In the work to get a booting 3.0 kernel on the KF we need some low level access to debug since the USB initializes way later in the kernel sequence. I did some research on the OMAP4430 (the CPU in our dear KF) to try and find out where the UART interface might be accessed.
Looking at the datasheet we get a feeling of all the ball points ("pins") on the CPU. (Observe that this is seen from under the CPU)
OMAP4430 ball bottom layout
By then looking at the UART table TX pin numbers I try to get a feel where they might be located on the chip.
UART 1, 2, 3, 4 table
Finally, I cross referenced the UART table with TX pin and physical position on the board. The colors represent UART 1 TX, UART 2 TX and UART 4 TX. UART 3 doesn't seem likely to be mapped.
UART 1, 2 and 4 TX pin guesses
Observe that these are all estimates, but should narrow the search down (unless they're mapped somewhere way outside of the CPU area that is). I haven't had a chance of testing any of my guesses yet, but I thought I'd at least post my thoughts to get the ball running.
