2.5.42: IrDA issues
Quoted 2 lines · Alessandro Suardi
Quoted 1 line
Jean Tourrilhes wrote:
[snip]
Quoted 15 lines
Quoted 15 lines
09:03:01.369859 xid:rsp
589c38b5<bb700000S=6 s=5 Nokia 6310 hint=b125 [ PnP Modem Fax Telephony IrCOMM IrOBEX ] (27) 09:03:01.382763 xid:cmd589c38b5> ffffffff S=6 s=* dolphin hint=0400 [ Computer ] (23)
09:03:01.422691 snrm:cmd ca=fe pf=1589c38b5>bb700000new-ca=66 (33) 09:03:01.524835 ua:rsp ca=66 pf=1589c38b5<bb700000(31) 09:03:01.524997 rr:cmd > ca=66 pf=1 nr=0 (2)
09:03:01.774694 rr:cmd > ca=66 pf=1 nr=0 (2)
09:03:02.274609 rr:cmd > ca=66 pf=1 nr=0 (2)
09:03:02.774533 rr:cmd > ca=66 pf=1 nr=0 (2)
--------------------------------------------------------Then hangs (all this is with the 20021007 driver).
Quoted 8 lines
Quoted 3 lines
Hope it's helpful, ciao,
--alessandro
That's a speed problem (1152000 is MIR). You are supposed to
select the proper module parameters to get speed changes to MIR/FIR to work properly (check IrDA mailing list messages from Daniele). Another solution is to limit the speed of the stack to 115200 (115k - see my web page for details).
OK... found finally time to re-test, I ran today the 10/30 smsc-ircc2 driver
under RedHat 8.0, kernel 2.5.45 and the 115200 limit (after patching the few remaining uses of __FUNCTION__ in smsc-ircc.c), and the GPRS link stays up instead of going down after < 3 minutes. Speed is definitely slower (using the GPRS link) than the current 2.4.20-rc1 / smc-ircc combo but given the record of GPRS performance 'round here I do know it's not meaningful yet - I'll repost with irdadump logs if performance stays on the low end (my "fast" transfers peak at 3.9KB/s, the slowness of today's tests was about 0.2KB/s).
Under 2.5 I keep being spammed by
Nov 2 17:02:52 dolphin kernel: IrLAP, no activity on link! Nov 2 17:03:37 dolphin last message repeated 2 times Nov 2 17:04:52 dolphin last message repeated 8 times Nov 2 17:05:58 dolphin last message repeated 5 times Nov 2 17:07:03 dolphin last message repeated 10 times Nov 2 17:08:08 dolphin last message repeated 4 times Nov 2 17:09:11 dolphin last message repeated 6 times Nov 2 17:10:28 dolphin last message repeated 5 times Nov 2 17:12:37 dolphin last message repeated 8 times Nov 2 17:13:43 dolphin last message repeated 3 times
while a substantially lower amount of spam under 2.4 tells me
Nov 2 21:28:31 dolphin kernel: NETDEV WATCHDOG: irda0: transmit timed out Nov 2 21:28:31 dolphin kernel: irda0: transmit timed out Nov 2 21:28:38 dolphin kernel: NETDEV WATCHDOG: irda0: transmit timed out Nov 2 21:28:38 dolphin kernel: irda0: transmit timed out Nov 2 21:29:05 dolphin kernel: NETDEV WATCHDOG: irda0: transmit timed out Nov 2 21:29:05 dolphin kernel: irda0: transmit timed out
Thanks for now & ciao,
--alessandro
Quoted 9 lines · ALESSANDRO.SUARDI
while a substantially lower amount of spam under 2.4 tells me
Nov 2 21:28:31 dolphin kernel: NETDEV WATCHDOG: irda0: transmit timed out Nov 2 21:28:31 dolphin kernel: irda0: transmit timed out Nov 2 21:28:38 dolphin kernel: NETDEV WATCHDOG: irda0: transmit timed out Nov 2 21:28:38 dolphin kernel: irda0: transmit timed out Nov 2 21:29:05 dolphin kernel: NETDEV WATCHDOG: irda0: transmit timed out Nov 2 21:29:05 dolphin kernel: irda0: transmit timed out
Yep, I have the same problem when I try irport. It seems that
somehow we drop/miss a Tx-done interrupt (UART_IIR_THRI). I tried to investigate that but came empty handed.
I looked a the new smsc-ircc2 driver, and I suspect that what
you see is another manifestation of the exact same problem.
Quoted 3 lines
Thanks for now & ciao,
--alessandro
Have fun... Jean