[Build environment broken?]

Thanks Igor,

I built preempt model 4 ok and booted it on OPi+2e

but my RT app does not think the result is really an RT kernel

the hint is  from  uname -a...

Linux orangepiplus2e 3.4.113-rt143-sun8i #2 SMP PREEMPT Thu May 18 08:04:38 ICT 2017 armv7l GNU/Linux

 

it does not say "PREEMPT RT" so apps wont work ( and many linuxes are PREEMPT out of the box )

 

I was able to load and run 'cyclictest' from https://www.kernel.org/pub/linux/utils/rt-tests/

but the jitter was way too high (500 to 2500uS) on the few tests i tried (https://rt.wiki.kernel.org/index.php/Cyclictest)

So, I dont have very tight timing bounds. I'd like to get 200uS, but its not christmas today ;-)

 

Thanks for your help, the fix is beyond my skills. I will watch for news though, its a great little board.

TomP

 

 

[Build environment broken?]

Hello Christos & Igor Pecovnik,

Thanks very much for all this great work!

 

I am very NOOB to building kernels

I wanted RT kernel to use with linuxcnc, specifically linuxcnc's 'hal' system

 

Before building, I repaired 2 patch files that are missing 0x0a at end of file

( they produce err saying the file ends in middle of line, actually meaning at EOF )

friendlyarm_m1_gigabit.patch ( ends in 0x3a)

ov5640-enhanced-device-with-vfe-drc-table-fix.patch (ends in 0x29)

 

I repair by simply opening in nano and immediately saving.

This action adds 1 0x0a byte to each file, and changes the timestamp.

 

But, changing the timestamp creates another problem,

the compile reports that it cannot update due to the change.

 

I chose to continue

Is this correct to add the missing 0x0a?

Can I ignore the errors about ending in middle of line?

 

 

I am trying to make preempt model 5,

Christos said "Did a fresh OrangePi Plus2E image, legacy, Debian jessie, desktop, RT, but the SD boot gives this"

Which preempt model did you succeed with?

 

regards

tomp