22 Nov
2019
22 Nov
'19
9:29 p.m.
On Fri, Nov 22, 2019 at 08:50:15AM -0600, Pierre-Louis Bossart wrote:
With this proposal from Jaroslav, users of the older firmware will not see the panic but they are still facing potential underflows.
Underflows do seem better and more recoverable than panics.
We may also need to look at different ways to identify the firmware, in this case the problem is not due to the ABI proper but a change in the timing sequences, we may need a different sort of ID here?
Yeah. I guess it depends how widespread the different behaviours of the racy firmware are and what they're influenced by?