[En-Nut-Discussion] Merge of CortexM3 development back to the trunk
harald.kipp at egnite.de
Thu May 10 12:06:14 CEST 2012
On 22.04.2012 19:24, Ole Reinhardt wrote:
> If you (and all others too) could do further testing of the branch on
> real hardware would be very appreciated.
I've tested several AVR and ARM boards with this branch. Nothing seems to be broken, which wasn't already broken in the trunk.
The major problem is, that I still didn't get the IP checksum routine right. The problem is in the trunk.
On the AVR there is no specific alignment at all. Depending on the heap allocation, the IP header may exist at any address, which is not correctly handled by my new checksum routine. I do have a fix, but would like to do more testing and apply it directly to the trunk, after the merge.
So far, I didn't test any AVR32 hardware. However, nothing essential has changed on that part, right?
I didn't test any CM3 hardware either. But as this is a new feature of the branch, I'd vote for doing the merge ASAP.
More information about the En-Nut-Discussion