On 13 January 2015 at 02:35, Robert Nelson <[email protected]> wrote:

> but we will see the bugs when users start plugging them in. ;)


The revision that's in the DM814x certainly has some fun ones...

   - *Advisory 3.0.6* - Data is corrupted is burst accesses (e.g. EDMA) are
   performed that do not start and end on a 16-byte aligned boundary.  Any
   access which can't satisfy that constraint must be performed using single
   accesses.  (Doing that with EDMA is still possible but requires some
   creative self-chaining configuration.)
   - *Advisory 3.0.7* - A failed non-posted transaction doesn't set the
   status bits it's supposed to. You can still find the raw PCIe header of the
   response in some register and parse it yourself.
   - *Advisory 3.0.10* - PL_ACKTIMER register may only be changed while in
   Gen1 mode. Changes made in Gen2 mode corrupt contents.

-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
"BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to