Abort in Kernel during Filer_Boot
Jon Abbott (1421) 2652 posts |
I’ve not seen this occur on RC14, but have now had a few occurances of the Kernal Aborting when processing a Filer_Boot I’ve issued via OS_CLI on the 15-10-15 Pi build.
Sadly I can’t provide a register dump as the Wimp gets into an endless “Abort on Data Transfer” loop. The application being booted contains the following !Boot Obey script:
EDIT: It does occur on RC14 |
Jeffrey Lee (213) 6048 posts |
Are you able to connect a serial cable to your Pi? With the number of odd aborts you’re seeing it would be good if I could send you a build that would output a stack dump over the serial port whenever something goes wrong. |
Jon Abbott (1421) 2652 posts |
I think I have a USB serial port somewhere, or can certainly purchase one. Does it need to be a specific model? I’m not so certain I have another device to attach to the other end though, unless you can go Pi→Pi |
Jeffrey Lee (213) 6048 posts |
There are serial RX & TX lines on the expansion header – it’s those which you’ll need to connect to. If you have cables which will fit the pins on the expansion header then you could probably just connect the RX, TX, and GND of the two Pi’s together (remembering to cross over the RX & TX lines!). If not, you should be able to pick up a generic USB serial adapter and have the USB side connected to a PC or something. |
Colin (478) 2433 posts |
If you have the wires to connect the two pi serial pins I have a very basic (written in BASIC) serial terminal which runs in a taskwindow which reads serial output on a pi. I also have the same program modified to use a USB cp2102 ttl serial converter similar to this which I use to get output to my Iyonix or laptop. Just say if they would be any use. If you haven’t got the cables to directly connect the two PIs the cables that come with the above serial device would do the trick. |
Jon Abbott (1421) 2652 posts |
I’ve just had this occur on RC14, which reported “Not a Heap Block” prior to the endless Abort loop. So there’s possibly a bug around an OS_Module 7 in the Kernel somewhere. |