[Alsa-user] cs4281 driver regression - thinkpad x21

[Kai V. <kai...@wa...>]

Hi, 

this isn't a critical issue, but just in case someone is struggling
with the same problem.

Shortly put, with ALSA 0.5.12a, the cs4281 driver works ok (tested
with full-duplex operation in both native and OSS-emul modes). But
when upgrading to 0.9.x, card is not identified anymore. The error message 
printed is:

--cut--
Jun 28 15:25:36 localhost kernel: ALSA ../alsa-kernel/pci/cs4281.c:675: AC'97 read problem (ACSTS_VSTS), reg = 0x0
Jun 28 15:25:36 localhost kernel: ALSA ../alsa-kernel/pci/cs4281.c:675: AC'97 read problem (ACSTS_VSTS), reg = 0x7c
Jun 28 15:25:36 localhost kernel: ALSA ../alsa-kernel/pci/cs4281.c:675: AC'97 read problem (ACSTS_VSTS), reg = 0x7e
--cut--

And here's some hw info (IBM Thinkpad X21):

--cut--
00:0b.0 Multimedia audio controller: Cirrus Logic Crystal CS4281 PCI Audio (rev 01)
        Subsystem: IBM: Unknown device 0183
        Flags: bus master, medium devsel, latency 64, IRQ 11
        Memory at f4012000 (32-bit, non-prefetchable) [size=4K]
        Memory at f4000000 (32-bit, non-prefetchable) [size=64K]
        Capabilities: [40] Power Management version 2
--cut--

The AC97 codec is 'Cirrus Logic CS4297A rev B'. With the OSS-driver of
2.4.18 kernel, the playback works ok, but recording fails to wake up the
user-space (ie. only push-model use like 'cat </dev/dsp >/dev/dsp' works).
So currently ALSA 0.5.12a is the best driver for this specific chip-combo.

--
 http://www.eca.cx
 Audio software for Linux!