ghans wrote:The only guy really working on it is gsh (so far i know) .
He is a Broadcom employee and i believe that keeps him busy.
Therefore his USB fixing is on voluntary basis and he
can't promise steady progress , even if he is the only one
with access to Synopsys docs and RTL .
Actually that's not true anymore... I'm a full time Raspberry Pi employee now! Have been working full time on the USB issues over the last couple of weeks and have got much further with my fixes... Basically I'm implementing the whole split transaction completion code in the FIQ which will make all the keyboard / serial type issues go away.
Of course it still needs a few bug fixes and some time for testing and then the slow release procedure (i.e. release it to a few people before we're happy to push it for everyone.)
But, I'm not sure it'll fix your specific problem, you'll have to check to see what changes... I've still not been able to look at any isochronous comms issues so if it is something there it may well still be around, but at least I'll have a bit of breathing space to go and check...