Joern, Lacking the modem hardware, I can't do any useful testing myself. But might it help if: The version 1.0.14rc2 libraries in addition to the snd-hda-intel driver set. The slmodemd was next compiled anew, thus using the new libasound2.* code. The hope here is that using "pure" version 1.0.14rc2 ALSA code may help. MarvS On 1/18/07, Joern Wuestenfeld <j.wuestenfeld@xxxxxx> wrote:
Hello Lajos, MarvS and all others interested, I did some debugging of the ALSA driver while trying to get the modem working with alsa > 1.0.12. Mine is a SI3054 on an Fujitsu Siemens Lifebook S7020. Here are the results! Jan 16 20:03:46 YYYY kernel: ALSA XXX/alsa-driver/pci/hda/../../alsa-kernel/pci/hda/hda_intel.c:555: hda_intel: azx_get_response timeout, switching to polling mode... Vanishes if enable_msi=1 set in /etc/modprobe.d/sound as option to snd-hda-codec If starting actual ALSA driver from repository(today!): Jan 18 13:31:45 YYYY kernel: PCI: Setting latency timer of device 0000:00:1b.0 to 64 Jan 18 13:31:45 YYYY kernel: ALSA XX/alsa-driver/pci/hda/../../alsa-kernel/pci/hda/hda_intel.c:696: codec_mask = 0x3 Jan 18 13:31:45 YYYY kernel: ALSA XXX/alsa-driver/pci/hda/hda_codec.c:1764: hda_codec: model 'fujitsu' is selected for config 10cf:1326 (Fujitsu S702X) Starting slmodemd and dialing gives: # ./slmodemd -a -c GERMANY hw:0,6 SmartLink Soft Modem: version 2.9.11 Jan 1 2007 16:52:00 symbolic link `/dev/ttySL0' -> `/dev/pts/9' created. modem `hw:0,6' created. TTY is `/dev/pts/9' Use `/dev/ttySL0' as modem device, Ctrl+C for termination. error: cannot setup hw params for playback: Invalid argument And in syslog: an 18 13:32:48 YYYY kernel: ALSA XXX/alsa-driver/pci/hda/hda_codec.c:627: hda_codec_setup_stream: NID=0x1, stream=0x0, channel=0, format=0x0 Jan 18 13:32:48 YYYY kernel: ALSA XXX/alsa-driver/pci/hda/hda_codec.c:1400: invalid rate 9600 Jan 18 13:32:48 YYYY kernel: ALSA XXX/alsa-driver/pci/hda/../../alsa-kernel/pci/hda/hda_intel.c:1153: hda-intel: invalid format_val, rate=9600, ch=1, format=2 The problem here is the stram number 0x0! It should be (according to documentation) something between 0x1 and 0xf. Seems to be a bug somewhere! Still looking for the details! Joern
