OK, I may be slightly dense here, but I did your trick of using setserial to make sure the serial driver left ttyS2 alone, and I did a "lsmod | grep serial" as root, which came up empty. The progress is that when i now did "ifconfig irda0 up", i only had this in my syslog:
found SMC SuperIO Chip (devid=0x0a rev=00 base=0x00e0): FDC37N971 SMC IrDA Controller found IrCC version 2.0, firport 0x100, sirport 0x3e8 dma=2, irq=3 IrDA: Registered device irda0 which to me seems right. Then I went and did a "echo 1 > /proc/net/irda/discovery" and a "tail -f /proc/net/irda/discovery" to get the log of what the port finds. This left me with "IrLMP: Discovery log:" and no more if I held my Palm Vx or Ericsson t39 in front of the IR port. (Yes, 10-15 cm, no flourescent lighting, semi dark room, lit candles in a hexagon with a dead chicken in the middle and full moon, etc). Nothing more in the discovery log. Next pointer? :) ttfn, A Bj�rn Mork <[EMAIL PROTECTED]> writes: > Alexander Hoogerhuis <[EMAIL PROTECTED]> writes: > > > Tried it here, and the ifconfig-bit seems to run fine, although the > > output from smc-ircc's init is printed twice in /var/log/messages: > > found SMC SuperIO Chip (devid=0x0a rev=00 base=0x00e0): FDC37N971 > > SMC IrDA Controller found > > IrCC version 2.0, firport 0x100, sirport 0x3e8 dma=2, irq=3 > > found SMC SuperIO Chip (devid=0x0a rev=00 base=0x00e0): FDC37N971 > > SMC IrDA Controller found > > IrCC version 2.0, firport 0x100, sirport 0x3e8 dma=2, irq=3 > > IrDA: Registered device irda0 > > > > And if I now type pppd /dev/ircomm0 the machine freezes instantly... > > > > FYI I run 2.4.17 with Robert Loves preempt-patch, and apart from that > > it is a stock RedHat 7.2 system. > > > > Any pointers? > > Ugh, I just saw the same yesterday when I tricked a colleague to try > it ;-) Didn't make the connection then, but I think I do now... > > The thing that happens is that your serial driver probably have found > and taken control over /dev/ttyS2 (which really is the IrDA port). This > causes smc-ircc to fail the first time, but succeed the second time > you (or the system) try to load it. So you get a double set of > messages. > > This happened without my patch too, but I introduced an ugly bug in > ircc_close so it became a bit more fatal than it should be. Impressive > given the low number of changed lines, eh? Which proves that Jean was > perfectly right not to accept any patch without proper testing. New > version attached. > > But you should really run 'setserial /dev/ttyS2 uart none' before > loading smc-ircc in any case. I don't know too much about about > RedHat, but there is probably some startup script where such things > should be. In Debian you would put it in /etc/serial.conf. > > > Bj�rn > > -- -- Alexander Hoogerhuis | [EMAIL PROTECTED] CCNP - CCDP - MCNE - CCSE | +47 905 76 198 "You have zero privacy anyway. Get over it." --Scott McNealy _______________________________________________ Linux-IrDA mailing list - [EMAIL PROTECTED] http://www.pasta.cs.UiT.No/mailman/listinfo/linux-irda
