mwester at dls.net
Sun Feb 22 18:07:37 CET 2009
Paul Fertser wrote:
> I flashed moko10 and did some experimenting before i wrote that
> line. 1000ms looks incorrect, for sure, but the reason for that is
> simple: just before that i assert and de-assert push-button and 500ms
> for that is always enough (i decided to be on the safe side and not
> tried to lower it further). If not for the bug in firmware, that'd be
> all is needed to be done inside the driver. But if we want to
> workaround that nasty issue, we have to wait a bit for the modem to
> boot. I decided to consolidate additional delay needed in this
> workaround line, so the workaround can be easily dropped altogether,
> minimizing the time needed for power_on.
> So, that's not a time needed for the calypso to "feel" the line,
> that's mostly the time needed for it to boot. 500ms was not enough.
Is there something we can trigger on to know that the calypso has
booted? Can we look for a change in the serial port control lines?
I'm thinking that if we forced flow-control on, and then waited for
either a maximum amount of time (1 sec) _or_ the appearance of the GSM
interrupt that would signal that the calypso just sent its startup
message, then we could handle the case of old firmware, but optimize the
delay for the latest fixed firmware.
More information about the openmoko-kernel