On 04/23/2019 07:42 AM, Nguyen, Tom wrote:
Per question, "So doesn't this also cause us to bring the modem offline when we
wake up from resume?"
Yes, it does. While this modem is capable of waking the host CPU via GPIO upon SMS
reception, it doesn't have a handshake mechanism for the host to signal that it's
ready to communicate with the modem, which would allow the host to prevent the modem from
sending the QMI indication too early. Frustratingly, this handshake does exist if the
modem is used with the AT interface. I tried to convince the vendor that the handshake
should be applied to the QMI interface too, but...
That sounds horrible.
Per comment, "So any state data must be allocated as part of the driver data."
Thanks for the insight. Are you referring to "struct sms_data"?
Yes, correct. Anything that contains state must be allocated on the
heap. For oFono drivers this is typically the driver specific data
structure setup in the .probe method.
I've copied my other email, so hopefully my future responses
won't break this chain.
Also, thank you for helping with the IMEI patch.
No worries. Just FYI, when responding to the list, please refrain from
top-posting. The common open source mailing list etiquette is to