Re: 5.2-RELEASE TODO

From: Jesse Guardiani <jesse_at_wingnet.net>
Date: Mon, 15 Dec 2003 13:15:37 -0500
Munehiro Matsuda wrote:

> From: Peter Schultz <pmes_at_bis.midco.net>
> Date: Mon, 15 Dec 2003 09:47:58 -0600
> ::Robert Watson wrote:
> ::> Testing focuses for 5.2-RELEASE
> ::> 
> ::>  +---------------------------------------------------------------------------+
> ::>  |    Issue     |   Status    |Responsible|           Description     
> ::>  |          |
> ::>  |--------------+-------------+-----------+----------------------------------|
> ::>  |              |             |           |The PCM audio framework and
> ::>  |              |             |           |device| drivers have been
> ::>  |              |             |           |locked and free |
> ::>  |PCM locking   |             |           |of Giant for quite a while,
> ::>  |but   |
> ::>  |and           |Needs testing|--         |LOR problems persist along
> ::>  |with   |
> ::>  |performance   |             |           |reports of poor audio
> ::>  |performance |
> ::>  |issues        |             |           |under load. These problems
> ::>  |are    |
> ::>  |              |             |           |believed to have been
> ::>  |              |             |           |corrected,  |
> ::>  |              |             |           |but more testing is
> ::>  |              |             |           |desired.      |
> ::>  |--------------+-------------+-----------+----------------------------------|
> :: >
> ::This longstanding problem has now been fixed as far as I'm concerned.
> ::Xmms hasn't played mp3s without a hicup for me on -CURRENT since at
> ::least a year, if not two or more.  It's been so long I can't recall.
> ::
> ::I'm testing with a soundblaster live! value card:
> ::pcm0: <Creative EMU10K1> port 0xef20-0xef3f irq 17 at device 17.0 on
> ::pci0 pcm0: <SigmaTel STAC9721/23 AC97 Codec>
> ::
> ::I'm running a -CURRENT SMP kernel on a Tyan S1832DL w/dual PII 350s.  I
> ::only assume it's working with 5.2 RC1.
> ::
> ::Pete...
> 
> Hi,
> 
> I still get LOR with my sound card and some hicup under disk load, with
> today's 5.2-CURRENT. It's a UP system with following sound card.
> 
> pcm0: <Yamaha DS-1E (YMF744)> port 0xfcac-0xfcaf,0xfc00-0xfc3f mem
> 0xfecf0000-0xfecf7fff irq 9 at device 9.0 on pci0 pcm0: <Asahi Kasei
> AK4543 AC97 Codec>
> 
> Also, here's hand written LOR message:
> 
> lock order reversal
>  1st 0xc3405440 pcm0 (sound softc) _at_ dev/sound/pci/ds1.c:734
>  2nd 0xc3405080 pcm0:play:0 (pcm play channel) _at_
>  dev/sound/pcm/channel.c:480
> Stack backtrace:
> ....
> witness_lock(c3405080,0,c07a8705,1e0,c33f7300) at witness_lock+0x672
> _mtx_lock_flags(c3405080,0,c07a8705,1e0,0) at _mtx_lock_flags+0xba
> chn_intr(c33f7300,104,4,4,c3405240) at chn_intr+0x2f
> ds_intr(c340b4000,0,c07b3baa,21f,c334c1c4) at ds_intr+0xf5
> ithread_loop(c1856900,d03a7d48,c07b3a24,311,3473635f) at
> ithread_loop+0x192 fork_exit(c059da00,c1856900,d03a7d48) at fork_exit+0xb4
> fork_trampoline() at fork_trampoline+0x8


Hmmm... I'm not much of a kernel hacker, but that's certainly different from
the duplicate lock Matthew Kanner fixed here (which fixed my sound problems):

    http://article.gmane.org/gmane.os.freebsd.current/34708

For what it's worth, I've also noticed a few hiccups in XMMS, but it's nothing
like what it was before Matthew Kanner's channel.c patch. I notice it only
under extremely high load and only once a day or week, rather than once every
5 minutes or so like it was in 5.1-RELEASE. Unfortunately I haven't seen an
LOR or dup lock the last few times it happened, so I thought the hiccups might
not be caused by the sound code... but maybe I was wrong... :)

To recap, my hardware is working great with 5.2-BETA (boot -v output below):

    pcm0: <Intel ICH3 (82801CA)> port 0x18c0-0x18ff,0x1c00-0x1cff irq 5 at device 31.5 on pci0
    pcm0: <Cirrus Logic CS4299 AC97 Codec (id = 0x43525936)>
    pcm0: Codec features headphone, 20 bit DAC, 18 bit ADC, 6 bit master volume, Crystal Semi 3D Stereo Enhancement
    pcm0: Primary codec extended features variable rate PCM, AMAP
    pcm0: sndbuf_setmap 261000, 4000; 0xe38ec000 -> 261000
    pcm0: sndbuf_setmap 27f000, 4000; 0xe38f0000 -> 27f000

    pcm0: measured ac97 link rate at 47998 Hz, will use 48000 Hz

-- 
Jesse Guardiani, Systems Administrator
WingNET Internet Services,
P.O. Box 2605 // Cleveland, TN 37320-2605
423-559-LINK (v)  423-559-5145 (f)
http://www.wingnet.net
Received on Mon Dec 15 2003 - 09:15:49 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:34 UTC