aleksander at aleksander.es
Fri Jun 13 07:22:54 PDT 2014
On Fri, Jun 13, 2014 at 10:24 AM, Bjørn Mork <bjorn at mork.no> wrote:
> Eyyh, thinking some more about this I wonder if cdc-wdm play well with
> this strategy? We cannot expect the firmware to notify us again about
> the remaining bytes. Which means that we won't read them until the next
> time the firmware has something to tell us. Which is probably what
> screws up MM here? The delay between the initial 256 bytes and the rest
> is possibly infinite.
> This is probably food for a generic cdc-wdm fixup: We should resubmit
> the read URB if the last read returned a full buffer.
Yeah, if we don't get a IN event in the socket we won't try to read
anything, so in this case we read up to the first 256 bytes and then
MM was just sitting there waiting for more events to come.
Is this something quick to fix in the driver?
More information about the ModemManager-devel