At Thu, 03 Jan 2008 15:34:05 -0800, Tobin Davis wrote:
Summary: hda_intel.c use gcap register to determine number of streams supported by southbridge.
This patch removes hardcoded values for the number of streams supported by the southbridge in most chipsets, and reads these values from the chipset directly. Most systems are hardwired for 4 streams in each direction, but newer chipsets change that capability.
Signed off by Tobin Davis tdavis@dsl-only.net [2 hda-gcap.patch <text/x-patch; UTF-8 (7bit)>] diff -r 1227a1c12325 pci/hda/hda_intel.c --- a/pci/hda/hda_intel.c Mon Dec 24 14:40:56 2007 +0100 +++ b/pci/hda/hda_intel.c Thu Jan 03 15:27:10 2008 -0800 @@ -1709,12 +1709,13 @@ static int __devinit azx_create(struct s { struct azx *chip; int err;
unsigned short gcap; static struct snd_device_ops ops = { .dev_free = azx_dev_free, };
*rchip = NULL;
- err = pci_enable_device(pci); if (err < 0) return err;
@@ -1790,10 +1791,19 @@ static int __devinit azx_create(struct s chip->capture_index_offset = ATIHDMI_CAPTURE_INDEX; break; default:
chip->playback_streams = ICH6_NUM_PLAYBACK;chip->capture_streams = ICH6_NUM_CAPTURE;chip->playback_index_offset = ICH6_PLAYBACK_INDEX;chip->capture_index_offset = ICH6_CAPTURE_INDEX;
/* read number of streams from GCAP register instead of using* hardcoded value*/gcap = azx_readw(chip, GCAP);if(!gcap) {snd_printk(KERN_ERR "Device has no streams \n");goto errout;};
I think it's safer to assign old ICH6_* values in this case (just to be sure) after a warning message.
chip->playback_streams = (gcap&(0xF<<12))>>12;chip->capture_streams = (gcap&(0xF<<8))>>8;chip->playback_index_offset = (gcap&(0xF<<12))>>12;chip->capture_index_offset = 0;- break; } chip->num_streams = chip->playback_streams + chip->capture_streams;
The patch has some coding style issues, for example, no space around operators, etc. Try to run $LINUX/scripts/checkpatch.pl, fix the errors and the issue above, and please repost the patch again.
Thanks!
Takashi