Module ali-ircc doesn't load after going to kernel 2.6.24. The problem results from a conflict with the second serial port 0x2f8. I could fix it generating a kernel with only one 8250 serial port. Another solution could be to use nr_uarts=1 as an option to 8250 module. Didn't test it however (no time yet ... :-( ) This didn't happen in k 2.4.22-suspend2-r2, at least.
Well reopen with emerge --info, kernel .config and dmesg output when ready. Thanks.
(In reply to comment #1) > Well reopen with emerge --info, kernel .config and dmesg output when ready. > Thanks. > I have no dmesg output. To get it again I needed to rebuild the kernel with more than 1 serial port to force the error situation. Here is the output extract I got from messages: Feb 16 02:25:07 H2 kernel: NET: Registered protocol family 10 Feb 16 02:25:08 H2 irattach: executing: '/sbin/modprobe irda0' Feb 16 02:25:08 H2 kernel: ali-ircc, driver loaded (Benjamin Kong) Feb 16 02:25:08 H2 kernel: ali_ircc_open(), can't get iobase of 0x2f8 Feb 16 02:25:08 H2 irattach: + FATAL: Error inserting ali_ircc (/lib/modules/2.6.24-tuxonice-r2/kernel/drivers/net/irda/ali-ircc.ko): No such device Feb 16 02:25:08 H2 irattach: Trying to load module irda0 exited with status 1 Feb 16 02:25:08 H2 irattach: executing: 'echo H2 > /proc/sys/net/irda/devname' Feb 16 02:25:08 H2 irattach: executing: 'echo 1 > /proc/sys/net/irda/discovery' Feb 16 02:25:08 H2 irattach: Starting device irda0 Feb 16 02:25:08 H2 kernel: ali-ircc, driver loaded (Benjamin Kong) Feb 16 02:25:08 H2 kernel: ali_ircc_open(), can't get iobase of 0x2f8 Feb 16 02:25:08 H2 irattach: ioctl(SIOCGIFFLAGS): No such device Feb 16 02:25:08 H2 irattach: Stopping device irda0 Feb 16 02:25:08 H2 kernel: ali-ircc, driver loaded (Benjamin Kong) Feb 16 02:25:08 H2 kernel: ali_ircc_open(), can't get iobase of 0x2f8 Feb 16 02:25:08 H2 irattach: ioctl(SIOCGIFFLAGS): No such device Feb 16 02:25:08 H2 irattach: exiting ... This is the expected output: Feb 16 16:51:18 H2 kernel: NET: Registered protocol family 10 Feb 16 16:51:20 H2 irattach: executing: '/sbin/modprobe irda0' Feb 16 16:51:20 H2 kernel: ali-ircc, driver loaded (Benjamin Kong) Feb 16 16:51:20 H2 kernel: IrDA: Registered device irda0 Feb 16 16:51:20 H2 kernel: ali_ircc_open(), ali-ircc, Found dongle: HP HSDL-3600 Feb 16 16:51:20 H2 irattach: executing: 'echo H2 > /proc/sys/net/irda/devname' Feb 16 16:51:20 H2 irattach: executing: 'echo 1 > /proc/sys/net/irda/discovery' Feb 16 16:51:20 H2 irattach: Starting device irda0
Created attachment 144000 [details] emerge --info
Created attachment 144002 [details] kernel .config
(In reply to comment #0) So which kernel worked for you? 2.4.22-suspend2-r2 mentioned in comment #0 never existed, that's just bogus version.
(In reply to comment #5) > (In reply to comment #0) > So which kernel worked for you? 2.4.22-suspend2-r2 mentioned in comment #0 > never existed, that's just bogus version. > Of course I mistyped it ... The kernel was (and is) 2.6.22-suspend2-r2. This is the kernel I have been using until now. I am moving to 2.6.24-tuxonice-r2. Sorry.
(In reply to comment #0) > The problem results from a conflict with the second serial port 0x2f8. > I could fix it generating a kernel with only one 8250 serial port. > Another solution could be to use nr_uarts=1 as an option to 8250 module. Can you detail how you reached those ideas? That would help us hunt down the changes made in the serial code since 2.6.22 which might have caused this.
(In reply to comment #7) > (In reply to comment #0) ... > > Can you detail how you reached those ideas? That would help us hunt down the > changes made in the serial code since 2.6.22 which might have caused this. > Look at comment #2 Feb 16 02:25:08 H2 kernel: ali_ircc_open(), can't get iobase of 0x2f8 Generating a kernel with only one 8250 serial port, 0x2f8 is not used. So ali-ircc works. Note that the default for the four 8250 serial ports are: 0x3f8, 0x2f8, 0x3e8 and 0x2e8 by this order. Somehow the kernel is geting all 4 ports not allowing ali_ircc to use 0x2f8. Generating a kernel with 1 port only, it only gets 0x3f8 leaving 0x2f8 for others to use. HTH
It's been awhile. Is this still an issue with later kernels?
(In reply to comment #9) > It's been awhile. Is this still an issue with later kernels? > I am sorry but I have a new laptop and cannot test it anymore.
Ok, sorry this was not resolved.
