Hi,In message <b5a77dec51.martin@xxxxxxxxxxxxxxxxxxx>, Martin Wuerthner <public@xxxxxxxxxxxxxxx> writes
The aborting instruction would be the one two instructions ahead of the BLT, but as you say, it cannot be invalid since it is a PC relative LDR with an aligned offset. Still, if you look at the register contents, you can see that the LDR R6 three instructions before the BLT has indeed happened whereas the LDR R5 after it has not, so everything points to this instruction causing the abort.
So?I have roughly assumed this is a load from inside the memory space of the program, and that such memory is read/write allowed. Perhaps I need to be more precise.
This is code that is used a lot, it is inside the main poll loop. It will have been executed thousands of times before it fails.
It has occurred to me that the bug is over writing this area, and then the corrupted code crashes.
Can we actually grab the memory after the crash.Chris - thanks for all the help. The person who reported the bug, said it was happening immediately after boot, as you say a clean machine.
-- David Pilling email: david@xxxxxxxxxxxxxxxxxxx web: http://www.davidpilling.net post: David Pilling, P.O. Box 22, Thornton-Cleveleys, Blackpool. FY5 1LR. UK To unsubscribe or subscribe goto: //www.freelists.org/list/davidpilling
