Ticket #2231 (closed defect: fixed)
interrupt not always generated when flowcontrolled=1 and modem wants to talk
|Reported by:||lindi||Owned by:||joerg|
|Blocking:||Estimated Completion (week):|
This issue has been discussed on IRC and a mailing list thread
"Occasional fail to initiate resume by incoming call, easy workaround
proposed" but I have not seen a bug report with clear steps to
reproduce. In that thread moko11-beta1 firmware is suggested as a
workaround so anybody who upgrades to that will want to verify that it
really fixes the issue for them with these steps.
Steps to reproduce: 1) start socat - file:/dev/ttySAC0,crtscts,crnl,echo=0,b115200 in another terminal and keep it running during steps 1-9 2) echo 0 > /sys/class/i2c-adapter/i2c-0/0-0073/neo1973-pm-gsm.0/power_on 3) echo 1 > /sys/class/i2c-adapter/i2c-0/0-0073/neo1973-pm-gsm.0/power_on 4) register to network manually by typing the following to socat (mind your own PIN though!): ATE1 AT+CFUN=1 AT+CPIN="0000" AT+CPIN? AT+COPS AT+COPS? 4) grep modem /proc/interrupts 5) call the phone from some other phone 6) let caller hangup 7) echo 1 > /sys/class/i2c-adapter/i2c-0/0-0073/neo1973-pm-gsm.0/flowcontrolled 8) call the phone from some other phone 9) grep modem /proc/interrupts Expected results: 4-9) When flowcontrolled=1 an interrupt is generated when GSM has something to send. Actual results: 4-9) Interrupt count does not increase: 17: 5 s3c-ext0 modem 17: 5 s3c-ext0 modem and thus if the phone is put to suspend it won't wake up either. More info: 1) I am running moko8 firmware on GTA02V5 AT+CGMR +CGMR: "HW: GTA02BV5, GSM: gsm_ac_gp_fd_pu_em_cph_ds_vc_cal35_ri_36_amd8_ts0-Moko8" $ grep Revision /proc/cpuinfo Revision : 0350 2) I am using andy-tracking kernel b8b36e5ec3db71d5 from beginning of January 2009.