Skip to contentSkip to conversation

Bug in i810 kernel module

DISCUSSION Stephane Dudzinski 3 participants 3 messages Feb 12, 2001 Sound

lore
Share & tools
lore
#0 lore raw

Hello,

I would like to submit the following bug to the list.

I've compiled the brand new 2.4.1 with the i810 module (the one which plays mp3s 15% faster than the usual speed). I did a hack modifying this line into :
/usr/src/linux/drivers/sound/i810_audio.c and modified the clocking as follows :

static unsigned int clocking=41194;

I don't think this why i got this behavior but using xmms sucked 98% of the CPU time reading either mp3s or ogg files. Something is definitly wrong with this driver as ALSA ones work just fine (xmms then uses only 0.5% of the CPU time).

hope it is useful.
Cheers
Steph

Signature · 10 lines
Email sent with RH 7.0 running kernel 241
"Go away or I will replace you with a very small shell script" - ThinkGeek
Mailto:[email protected]
Phone:8586009

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://vger.kernel.org/lkml/
#1 lore raw
Quoted 11 lines

I've compiled the brand new 2.4.1 with the i810 module (the one which plays mp3s 15% faster than the usual speed). I did a hack modifying this line into :
/usr/src/linux/drivers/sound/i810_audio.c and modified the clocking as follows :

static unsigned int clocking=41194;

I don't think this why i got this behavior but using xmms sucked 98% of the CPU time reading either mp3s or ogg files. Something is definitly wrong with this driver as ALSA ones work just fine (xmms then uses only 0.5% of the CPU time).

Or something is wrong with xmms. Right now I dont know which. I can tell you the speed thing is unrelated. Certain vendors wire the ac97 codecs to the wrong clock to save board space/crystals - notably dell laptops. That is why the clocking value is a module option and you can select 'fix the sound on dell' mode with the module option ftsodell=1

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://vger.kernel.org/lkml/

#2 lore raw
In reply to Stephane Dudzinski
Quoted 7 lines · [email protected] (Alan Cox) writes:
Quoted 4 lines

I don't think this why i got this behavior but using xmms sucked 98% of the CPU time reading either mp3s or ogg files. Something is definitly wrong with this driver as ALSA ones work just fine (xmms then uses only 0.5% of the CPU time).

Or something is wrong with xmms. Right now I dont know which. I can tell you the speed thing is unrelated.

My patch which was included in 2.2 seems to be missing from 2.4.1. The idea was taken from trident.c, and it keeps xmms at normal CPU usage. Same patch against 2.4.1 would look like this, but it's totally untested:

sound/i810_audio.c renamed from sound/i810_audio.c.orig +6 -2
@@ -879,7 +879,9 @@
 					dmabuf->endcleared = 1;
 				}
 			}			
-			wake_up(&dmabuf->wait);
+			if (dmabuf->count < (signed)dmabuf->dmasize/2) {
+				wake_up(&dmabuf->wait);
+			}
 		}
 	}
 	/* error handling and process wake up for DAC */
@@ -899,7 +901,9 @@
 				printk("DMA overrun on send\n");
 				dmabuf->error++;
 			}
-			wake_up(&dmabuf->wait);
+			if (dmabuf->count < (signed)dmabuf->dmasize/2) {
+				wake_up(&dmabuf->wait);
+			}
 		}
 	}
 }

Also I'll have to ask you to please go and play some mirrormagic as you promised when I sent the 2.2-patch. ;-)
As of 2.2.18 the sound in most games still sounds like crap on my 82443MX-100-based laptop with the i810_audio driver, while alsa works quite nicely.

//Marcus

Signature · 9 lines
---------------------------------+---------------------------------
         Marcus Sundberg         |      Phone: +46 707 452062
   Embedded Systems Consultant   |     Email: [email protected]
        Cendio Systems AB        |      http://www.cendio.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://vger.kernel.org/lkml/