The OpenVortex Project - Patches: patch #3948, patch to fix oopses on amd64
You are not allowed to post comments on this tracker with your current authentication level.
patch #3948: patch to fix oopses on amd64
Submitter: | Igor V. Kovalenko <kovalenko> | ||
Submitted: | Mon 25 Apr 2005 09:47:27 PM UTC | ||
Category: | None | Priority: | 5 - Normal |
Status: | None | Privacy: | Public |
Assigned to: | None | Open/Closed: | Open |
( Jump to the original submission )
Thu 02 Feb 2006 08:49:56 AM UTC, comment #227: |
Anonymous |
Tue 31 Jan 2006 03:58:33 PM UTC, comment #226: Available settings:
|
Anonymous |
Thu 26 Jan 2006 03:44:30 PM UTC, comment #225: FIFO_SIZE_BITS should be 6 instead of 5 for au8810
|
Anonymous |
Mon 23 Jan 2006 07:15:26 AM UTC, comment #224: At the same time, in block 264 audio processing peripheral 230 monitors the amount of data left in the FIFO buffer in order to determine if another DMA access is required. When a trigger level is reached in the FIFO buffer, a DMA access is requested by the FIFO controller. The FIFO buffer may be any suitable size capable of handling bus latencies while small enough such that the memory requirement is not overly burdensome. In one embodiment of the present invention, the FIFO buffer is 32 samples or words in depth, which would withstand a maximum bus latency of about 4 microseconds. By way of example, if 48 DMA channels trigger simultaneously, with a trigger level set at half the buffer size (16 samples), each DMA channel can wait 16 samples (20 microseconds per sample), or approximately 320 microseconds before running out of data. Thus, the 48th DMA channel can wait more than 4 microseconds for each of the 47 DMA channel request that comes before it.
Additionally, the trigger level may be set at any appropriate level to maintain enough data in the FIFO buffer to continue real-time operations while attempting to refill the FIFO buffer. In one embodiment, the trigger level is set at half the size of the FIFO buffer. Raising the trigger level allows us to withstand higher PCI bus latencies, but increases the percentage of PCI bus utilizatio |
Anonymous |
Sun 15 Jan 2006 03:18:52 PM UTC, comment #223: To enable au8810 modem
/* Enable codec channels 3 */
hwwrite(vortex->mmio, 0x291a0, 0x00000804); // GPIO
|
Anonymous |
Fri 13 Jan 2006 04:15:12 PM UTC, comment #222: Referring back to block 250 of FIG. 10, in one embodiment, when CPU 210 updates a sub-buffer table 380(0)-380(x), in relation to a particular DMA channel, CPU 210 checks mode register 382. In an alternate embodiment, CPU 210 also periodically checks sub-buffer tables 380(0)-380(x) in addition to any updates initiated by interrupts, as discussed in reference to FIGS. 10 and 11. By way of example, CPU 210 may update-sub-buffer tables 380(0)-380(x) routinely every 10 milliseconds. The periodic checks may be initiated by audio processing peripheral 230 by way of interrupts, or may be scheduled independently by the driver. In another embodiment, the CPU may check a current sub-buffer table (not shown) the indicates the current sub-buffer being utilized by the peripheral. Based upon the current sub-buffer and information about the previous value of the current sub-buffer table the CPU can determine that the peripheral has moved on to the next data segment. CPU 210 may then update the sub-buffer table. |
Anonymous |
Thu 12 Jan 2006 07:41:37 AM UTC, comment #221: The function sign_invert() is a side-effect of using U16 instead of S16.
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/au88x0/au88x0_eq.c?r1=1.7&r2=1.8
You make examine asp10.o in
|
Anonymous |
Wed 11 Jan 2006 09:19:59 AM UTC, comment #220: you mean signed instead of short?
s16 is signed, u16 is unsigned, both are short... |
Maarten Vanraes <alien999999999> |
Wed 11 Jan 2006 08:43:40 AM UTC, comment #219: Those equalizer gainstarget, gainscurrent, levels, CurrBypassGain, TimeConsts, States are short (i.e. s16) instead of u16 |
Anonymous |
Fri 06 Jan 2006 10:04:29 AM UTC, comment #218: #ifndef CHIP_AU8820
for (i = 0; i < 32; i++) {
|
Anonymous |
Fri 06 Jan 2006 09:41:50 AM UTC, comment #217: of course it doesn't, and what does that have to do with this problem???
this whole thing about flash audio, of course i was using a 32bit browser at that time... otherwise it'd be pretty evident that the flash would not work... |
Anonymous |
Fri 06 Jan 2006 02:56:29 AM UTC, comment #216: Flash Player does not work with 64bit browsers
|
Anonymous |
Thu 05 Jan 2006 02:03:48 AM UTC, comment #215: The routines vortex_src_addwtd, vortex_src_delwrd, vortex_mix_addwtd and vortex_mix_delwtd seem to maintain a linked list of active SampleRateConvertor and active mixer out in those MMIO registers SRC_CHNBASE, SRC_RTBASE, MIX_CHNBASE and MIX_RTBASE.
There are 16 SRC and 16 Mixer out.
The active SRC are added into the tail of the linked list where SRC_CHNBASE + (ch<< 2) store the first SRC
The second SRC will be put in SRC_RTBASE + ( first src << 2) and set bit8 of SRC_CHNBASE + ( ch << 2 )
The others SRC will be put in SRC_RTBASE + (previous src << 2) and set bit8 of previous SRC
If the resource management is correct, these routines will not return any error.
|
Anonymous |
Wed 04 Jan 2006 04:00:01 AM UTC, comment #214: Should add { } to enclose the hwwrite() in if statements.
How do macro writel(value,mmio) expand in AMD64 compiler ?
7b83
else {
|
Anonymous |
Mon 02 Jan 2006 11:09:46 AM UTC, comment #213: seems odd, that the if statement ends with ';' ? maybe that is just the problem?
maybe it has to unlock if the if statement is true? |
Anonymous |
Sun 01 Jan 2006 03:20:44 AM UTC, comment #212: You may need to check the 64-bits compiler generate the correct instruction in
"if (vortex_adbdma_bufshift(vortex, i)) ;"
if (source & IRQ_PCMOUT) {
|
Anonymous |
Thu 29 Dec 2005 04:58:47 PM UTC, comment #211: The GPIO3 of Turtle Beach Montego II/Diamond MX300 seem to be controlled by bit 6/7 of MMIO at 2205c |
Anonymous |
Thu 22 Dec 2005 03:15:44 AM UTC, comment #210: STAC9708 seem to have 2 version (LA2 and LA4).
The bad codec flag in au88x0 driver is refered to STAC9708 LA2.
|
Anonymous |
Mon 19 Dec 2005 04:26:32 PM UTC, comment #209: static int vortex_suspend(strut pci_dev *pci, pm_message_t state)
static int vortex_resume(struct pci_dev *pci)
|
Anonymous |
Mon 19 Dec 2005 11:47:03 AM UTC, comment #208: #ifdef CHIP_AU8830
|
Anonymous |
Sat 17 Dec 2005 03:05:07 PM UTC, comment #207: It seem that the initialisation of some DMA registers are missing in the au8820 driver
/* Init audio engine. */
|
Anonymous |
Thu 15 Dec 2005 07:22:18 AM UTC, comment #206: It seem that the FIFO register (17008) of au8810 is not initialised with correct values. ( 0x61 is wrong )
/* trigger... */
|
Anonymous |
Mon 12 Dec 2005 09:06:56 AM UTC, comment #205: static unsigned long vortex_wt_GetReg(vortex_t * vortex, char reg, int wt)
/* WT hardware abstraction layer generic register interface. */
static void
|
Anonymous |
Sat 10 Dec 2005 01:57:10 AM UTC, comment #204: static void vortex_codec_init(vortex_t * vortex,int power_state)
// power_state = 0 - Cold Reset
int i;
if ( power_state == 0 ) {
|
Anonymous |
Fri 09 Dec 2005 01:33:11 AM UTC, comment #203: Diamond MX300 with VIA chipset
|
Anonymous |
Mon 28 Nov 2005 09:33:13 AM UTC, comment #202: It seem that au8810 use different set of register for SPDIF 2a010 instead of 2205c and a different value to enable spdif when writing VORTEX_CODEC_EN 0x29190
This mean SQ1500 SPDIF will not work with current driver.
Do anyone with au8810 verify this by disable the SPDIF in the au8810 driver ? (Not sure IRQ reg error is caused by this bug.)
|
Anonymous |
Wed 23 Nov 2005 08:07:46 AM UTC, comment #201: The registers of au88x0 can also be accessed by io ports (BAR1) instead of mmio (BAR0)
chip->mmio = ioremap_nocache(pci_resource_start(pci, 0),
chip->io = pci_resource_start(pci, 1);
addr=0x2a804;
printk(KERN_INFO "mmio : %.8lx\n",hwread(vortex->mmio,addr)); |
Anonymous |
Thu 17 Nov 2005 03:43:47 PM UTC, comment #200: Your WM9701A do not have stereo enhancment
[root@queen RPM]# cat /proc/asound/au8810/codec97#0/ac97#0-0+regs
0:20 = 0000 <- general purpose
0:7c = 574d
|
Anonymous |
Thu 17 Nov 2005 03:36:46 AM UTC, comment #199: OffTopic
Why don't you post your sblive sb0220 and subsystem to alsa-devel ?
Some one with sb0220 with different subsystem id = 0x100a1102.
http://sourceforge.net/mailarchive/forum.php?thread_id=8989204&forum_id=1752
BTW, 1.0.10 final is going to release, when will your guys submit the AMD64 patch ? |
Anonymous |
Wed 16 Nov 2005 01:50:59 AM UTC, comment #198: http://www.alsa-project.org/alsa-doc/alsa-howto/c1706.htm
It seem that au88x0 SPDIF (device 1) is map to OSS /dev/adsp
|
Anonymous |
Tue 15 Nov 2005 01:00:09 PM UTC, comment #197: It seem that some au8820 only has 16-bits ADC/DAC , this means that the +6dB gain in hardware mixer may cause distortion
http://www.active-hardware.com/english/reviews/soundcards/magwave.htm
Those au8820 sound cards which has 18-bits codec is OK |
Anonymous |
Mon 14 Nov 2005 08:56:30 AM UTC, comment #196: http://cvs.sourceforge.net/viewcvs.py/*checkout*/alsa/alsa-kernel/include/mpu401.h?rev=1.14
struct _snd_mpu401 {
unsigned short hardware; /* MPU401_HW_XXXX */
}
int snd_mpu401_uart_new(snd_card_t * card, int device,
...
For I/O port cport = port + 1,
how about MMIO ?
|
Anonymous |
Fri 11 Nov 2005 08:52:48 AM UTC, comment #195: Some MMIO read/write are actually in 8-bits instead of 32-bits
asp10.o
|
Anonymous |
Fri 04 Nov 2005 03:33:52 AM UTC, comment #194:
port = (unsigned long)(vortex->mmio + VORTEX_MIDI_DATA); |
Anonymous |
Thu 03 Nov 2005 01:39:50 PM UTC, comment #193: i'll send a diff when i get home, but i didn't really look at gameport or mpu .
the problem i'm having with alsa hangs aren't related to cedega (i only used cedega for GTA3 once (it didn't work, btw)). |
Anonymous |
Thu 03 Nov 2005 10:22:20 AM UTC, comment #192: Have you fix the gameport and mpu401 bugs in the AMD64 patch (diff1.diff) ?
I guess cedega will properly try to use the gameport.
Please post your current code diff against ALSA CVS. |
Anonymous |
Wed 02 Nov 2005 09:24:41 AM UTC, comment #191: VIA Application note Serial #AN137 Date 07/05/00.
|
Anonymous |
Wed 02 Nov 2005 03:31:21 AM UTC, comment #190: /* Check for pause/resume support */
|
Anonymous |
Mon 31 Oct 2005 06:31:05 PM UTC, comment #189: that one doesn't resemble mine, i don't have the external black and spdif & aux internal connectors.
Today, i actually experienced something with snd-au8810 what probably is defined as the latency problem (i'd like someone to verify it):
- clicks in sound
It happened when i was busy with cedega (maybe it initializes a OSS socket or something)
Maybe i'd like to have the latency fixes, see if i get the problem again...
also the last few days i've experienced the sound-hanging while no flash browser was open, and nothing that used OSS (or sound for that matter) (Gaim was open, but it uses ALSA with hw; and i checked, within that time, no reason for sound was shown).
I do have it mostly with xine (when i watch some movie)...
anyone any idea? |
Maarten Vanraes <alien999999999> |
Sun 30 Oct 2005 09:40:48 AM UTC, comment #188: SQ1500
It look like using Sigmatec codec and two grenn jacks
|
Anonymous |
Fri 28 Oct 2005 05:20:26 PM UTC, comment #187: i think i'm using STAC
https://hive.homelinux.org/fotos/PC/emu10k1/finalized/
but last time i checked, this was the vortex mailinglist.
not to sound egotistic, but can't we get to the problem i've been having? |
Maarten Vanraes <alien999999999> |
Thu 27 Oct 2005 02:00:03 AM UTC, comment #186: SB0220 is different from other SBLive!
http://www.soomal.com/gb/001/007/00000005.htm
Do your SB0220 use EMicro 28028 or STAC9708 ?
https://bugtrack.alsa-project.org/alsa-bug/view.php?id=698
It seem sb0220 also need PCI latency fix.
|
Anonymous |
Wed 26 Oct 2005 05:09:03 PM UTC, comment #185: what is this about xine-ui ... i'm able to play music through xine-ui and amarok (using xine-lib).
I do not understand where you are getting at...
and i'm beginning to suspect that there are more than 1 person who replies here anonymously... |
Maarten Vanraes <alien999999999> |
Wed 26 Oct 2005 03:23:11 PM UTC, comment #184: The ALSA developer disables SNDRV_PCM_INFO_RESUME flag for drivers which does not support the full resume.
http://sourceforge.net/mailarchive/message.php?msg_id=12686297
It seem xine depends on this feature.
Have you patch your AMD64 version ?
#define ALSA_LOG 1
Where can I find the debug output of xine-lib ? (xine-ui)
|
Anonymous |
Wed 26 Oct 2005 02:36:50 AM UTC, comment #183: > 0:6c = 0000
It look like Erik 's STAC9708 is LA2 instead of LA4
|
Anonymous |
Tue 25 Oct 2005 09:49:41 AM UTC, comment #182: i don't know what my NIC has to do with anything, but the 3com is used for internet and the onboard one for LAN.
i do not get system hangs, i get just that all sound hangs...
it is so that some time ago, i got sound hangs that made my snd-au8810 have 210 dependencies, meaning that i could not rmmod it, thus if i wanted to have sound, i needed to reboot. and also the ps aux hung, etc...
but this is quite some time ago, i haven't had these things for a long time... maybe it was solved?
I'll try and discuss this with Toby Haynes, but i don't know if it is this problem... |
Maarten Vanraes <alien999999999> |
Tue 25 Oct 2005 04:53:34 AM UTC, comment #181: I guess you better discuss with Toby Haynes, he seem to have au8830 on K8T800 and VT8237
http://lists.nongnu.org/archive/html/openvortex-dev/2004-11/msg00000.html |
Anonymous |
Tue 25 Oct 2005 02:17:32 AM UTC, comment #180: |
Anonymous |
Tue 25 Oct 2005 01:50:50 AM UTC, comment #179: Sorry , URL should be
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/au88x0/au88x0.c?r1=1.8&r2=1.9 |
Anonymous |
Tue 25 Oct 2005 01:48:52 AM UTC, comment #178: Do you have two NIC ( on-board and 3COM ) ?
Which NIC is using to access Internet ?
Are the ACT(Tx/Rx) LED on your NIC/DSL modem still blinking when the system hangs ?
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/au88x0/au88x0.c?r1=1.11&r2=1.12
|
Anonymous |
Mon 24 Oct 2005 04:20:16 PM UTC, comment #177: i do not know what your reasoning is, or where you are getting at, but it shouldn't matter, no matter what, it should work...
slots:
and lotsa other stuff onboard... |
Maarten Vanraes <alien999999999> |
Mon 24 Oct 2005 04:20:06 PM UTC, comment #176: |
Anonymous |
Mon 24 Oct 2005 02:59:11 PM UTC, comment #175: Do your TV card alreay on slot 4 ? |
Anonymous |
Mon 24 Oct 2005 02:08:01 PM UTC, comment #174: why would i do that? you do have to think of IRQ sharing as well...
why MUST it be in PCI slot 2? |
Maarten Vanraes <alien999999999> |
Mon 24 Oct 2005 12:55:31 PM UTC, comment #173: Put your au8810 in PCI Slot 4 instead of Slot 2
http://lists.nongnu.org/archive/html/openvortex-dev/2005-09/pngAUTdhLPPsP.png |
Anonymous |
Mon 24 Oct 2005 07:57:25 AM UTC, comment #172: hahaha, indeed it is... ;-)
the "SB Live!" emu10k1 card is indeed below the au8810...
very well noticed! |
Maarten Vanraes <alien999999999> |
Mon 24 Oct 2005 06:58:20 AM UTC, comment #171: Refer to
http://lists.nongnu.org/archive/html/openvortex-dev/2005-09/pngqq5047afsi.png
It seem that your SB Live! is SB0220 if the card below your au8810 is your emu10k1. |
Anonymous |
Mon 24 Oct 2005 05:21:47 AM UTC, comment #170: onboard sound is disabled automatically when another sound card is put in... since i need 2 sound cards, i need 2 extra sound cards snd-au8810 and snd-emu10k1 ...
Last year I asked if I could make it available (the snd-via82xx) and there was a negative response.
I do not know the model of the "SB Live!" card and it is listed as unknown in alsamixer... I suspect not everything is correct with the snd-emu10k1, but then it doesn't have the problems my snd-au88x0 has...
does anyone know a method so i can determine if the fix listed below ( http://article.gmane.org/gmane.linux.alsa.user/10392 ) actually works for me? where do i need to change the code? |
Maarten Vanraes <alien999999999> |
Mon 24 Oct 2005 03:39:13 AM UTC, comment #169: What is the model of your SB Live! ?
|
Anonymous |
Mon 24 Oct 2005 02:46:49 AM UTC, comment #168: ASUS K8V
Chipset
VIA K8T800
Audio
ADI AD1980, 6-channel CODEC
www.analog.com/UploadedFiles/Data_Sheets/206585810AD1980_0.pdf
ALSA via82xx driver for your onboard sound
|
Anonymous |
Mon 24 Oct 2005 02:17:29 AM UTC, comment #167: struct _snd_mpu401 {
...
|
Anonymous |
Mon 24 Oct 2005 02:04:26 AM UTC, comment #166: About mpu401 in the patch, vortex->mmio is 64-bits address and cast to port (32-bits)
Do you need to patch snd_mpu401_uart_new() ?
- port = (unsigned long)(vortex->mmio + (VORTEX_MIDI_DATA >> 2));
|
Anonymous |
Sun 23 Oct 2005 04:20:01 PM UTC, comment #165: you should also note that i'm using a K8V 64bit motherboard!
some time ago, the via-fix option of the snd-au88x0 module didn't list my mobo as a faulty one...
of course, this may or may not be an issue now, i don't know...
this fix that i read, is this a one time only thing? or what? is this a fix that will help on my 64bit mobo? is there any way to be sure that this works? when should i do this? every time the sound fails? or just at modprobe time? or after?
about xine & flash websites:
procedure is simple:
|
Maarten Vanraes <alien999999999> |
Sun 23 Oct 2005 11:43:22 AM UTC, comment #164: Look like "Vortex: vortex_fifo_setadbctrl fail" related to VIA chipset.
|
Anonymous |
Sun 23 Oct 2005 07:31:07 AM UTC, comment #163: It seem to me that there is a bug in xine.
use xine (xine-lib-1.1.0) play music
music sometime paused (fftscope of xine also paused) for a while after clicking to load those web pages in http://www.macromedia.com/
It happen when xine using au88x0 and intel8x0 (alsa mode)
There is no such problem if I am using mplayer.
|
Anonymous |
Sat 22 Oct 2005 04:28:22 AM UTC, comment #162: The proposed PCM Volume Control per substream (subdevice)
http://lists.nongnu.org/archive/html/openvortex-dev/2005-08/msg00011.html
For 16-bits audio pass through au88x0 to 18-bits AC97 codec
10 * log 2 = 3 dB
In theory, it is possible to provide gain up to 54dB but it will cause distortion since AC97 codec is only 18-bits.
In practice,
au8810/au8830 front -48dB to 0dB
au8820 -48dB to 6dB
The 10-bands equalizer in au8810/au8830 use the 6dB difference
|
Anonymous |
Sat 22 Oct 2005 03:15:33 AM UTC, comment #161: What is the maufacturer and model of your motherboard ?
http://lists.nongnu.org/archive/html/openvortex-dev/2005-06/msg00010.html
It seem that your motherboard are using VIA chipset.
http://lists.nongnu.org/archive/html/openvortex-dev/2003-08/msg00014.html
There are reported incompatiblities case for Aureal sound card and VIA chipsets in VIA forum
|
Anonymous |
Fri 21 Oct 2005 12:18:23 PM UTC, comment #160: I don't think au8810 support APM suspend/resume
|
Anonymous |
Fri 21 Oct 2005 07:39:32 AM UTC, comment #159: If you have a procedure which can reproduce the system hang, you can ask the other AMD64 user to test.
au8820 - Igor V. Kovalenko
au8830 - Toby Haynes, Erik
|
Anonymous |
Fri 21 Oct 2005 07:08:49 AM UTC, comment #158: when the hang occurs and i'm using xine, then the visualisation also stops, cause alsa blocks.
at this moment, i can reproduce it easily enough, but i do not think that it is limited to only that, there are other (less reproducable) ways of triggering it.
play some random music with amarok and crossfading (i do not know yet if this has anything to do with it), that uses xinelib and alsa.
then browse using opera (don't know if this has anything to do with it either) to the official dido homepage and play some audio/video, click on links and generally browse the site, you get a hang within a 10min... |
Maarten Vanraes <alien999999999> |
Fri 21 Oct 2005 06:16:12 AM UTC, comment #157: Please note that the ALSA developer request the patch to use DMA_32BIT_MASK instead of VORTEX_DMA_MASK
http://sourceforge.net/mailarchive/message.php?msg_id=12965133 |
Anonymous |
Fri 21 Oct 2005 04:09:21 AM UTC, comment #156: Which subdevice of device 0 will be used for playback/capture is assigned by ALSA core, the au88x0 driver only dynamically dma, fifo, src and mixer to the substream.
There is one-to-one relationship between subdevice and dma.
However substream->number may not equal to stream->dma. |
Anonymous |
Fri 21 Oct 2005 03:59:41 AM UTC, comment #155: The freeze of using GAIM is not directly caused by au88x0.
The other emu10k1 also experience this freeze
http://sourceforge.net/mailarchive/forum.php?thread_id=8362982&forum_id=1752 |
Anonymous |
Fri 21 Oct 2005 03:30:28 AM UTC, comment #154: Add missing KERN_* suffix to printk.
http://sourceforge.net/mailarchive/forum.php?thread_id=8705967&forum_id=33141
Please make a diff of your au88x0 code against CVS or ALSA 1.0.9b and upload here |
Anonymous |
Fri 21 Oct 2005 01:43:07 AM UTC, comment #153: There is difference between aplay %s and aplay -D hw:0,0 %s
Try
aplay -v -D hw:0,0 foo.wav
and
aplay -v foo.wav
default is plughw
Do the hang always occure when you are running specification application (e.g flash player ) ?
Any step to reproduce the hang ?
For xine/mplayer, there are visualisation feature which display the waveform of the audio.
Is FFTScope of xine still running when there is no sound/system hang ?
Please post the sound portion of your modprobes.conf. |
Anonymous |
Thu 20 Oct 2005 08:12:16 PM UTC, comment #152: hw:0,0 is my default ...
aplay %s is my gaim sound setting... |
Maarten Vanraes <alien999999999> |
Thu 20 Oct 2005 02:18:03 PM UTC, comment #151: On gaim's preferences dialog, go to Sounds section, and choose Command on the Sound method setting. Then, on the Sound command textbox, enter:
aplay -D hw:0,0 %s
|
Anonymous |
Thu 20 Oct 2005 10:29:38 AM UTC, comment #150: |
Anonymous |
Thu 20 Oct 2005 10:15:57 AM UTC, comment #149: it's hard to say for sure if i actually have used it, at that time, it's my primary soundcard and it's for example also used by gaim to play a short beep tone, but i think that at that time, i only was playing music with amarok and browsing a flash website (that site has short tunes when pressing some buttons)
is there any other printk's i can add to give you more info on how to solve this problem? |
Maarten Vanraes <alien999999999> |
Thu 20 Oct 2005 02:08:08 AM UTC, comment #148: In your case,
Have you used au8810 between Oct 17 19:22:02 and Oct 17 19:23:16 ?
What applications are you running ?
it is strange that "Oct 17 19:23:18 queen kernel: vortex: start 0" appear after "Oct 17 19:23:18 queen kernel: vortex: stop 0".
Oct 17 19:23:16 queen kernel: direction: 0, type:0, subdevice: 0
|
Anonymous |
Wed 19 Oct 2005 06:46:01 PM UTC, comment #147: i'll remove them all, once this issue is resolved; and anyway, it's my own version, it's not like someone has any problems with it. |
Maarten Vanraes <alien999999999> |
Wed 19 Oct 2005 04:13:44 PM UTC, comment #146: No problem on my au8830 (32-bits) alsa-1.0.10rc1 + codec_init.patch + quad_front_rear_swap.patch
xine (xine-lib 1.1.0), mplayer(1.0Pre7Try2), ....
http://www.macromedia.com/support/flash/sound/sound_player/
You should comment all those unnecessary printk statements
|
Anonymous |
Wed 19 Oct 2005 07:46:22 AM UTC, comment #145: In the setting of flash player, there is an option to restrict the flash player do not use webcam and microphone.
This mean the capture device should not be opened by flash player when I had restricted the flash player do NOT use web cam and microphone.
However the /proc/asound/au8830/pcm0c/sub0 is used when flash playing sound. |
Anonymous |
Wed 19 Oct 2005 07:42:26 AM UTC, comment #144: xine has also a software solution (AMPLIFICATION), you can set it as standard mixer control, i believe alsa is working on implementing those hardware PCM volume control (with software fallback) for au88x0 and emu10k1. i'm not too sure, you can better check the alsa-devel mailing list for that... |
Maarten Vanraes <alien999999999> |
Wed 19 Oct 2005 04:00:45 AM UTC, comment #143: If we can run two or more instances of xine , how difficult will be in order to implement a customized gui of xine using the PCM volume control ?
Each instance of xine can change the PCM volume instead of change the volume of mixer in AC97 codec.
Other hardware mixing sound cards (e.g. emu10k1, via82xx, hdsp .. ) have PCM volume controls.
http://sourceforge.net/mailarchive/forum.php?thread_id=7815733&forum_id=1752
|
Anonymous |
Tue 18 Oct 2005 03:01:04 PM UTC, comment #142: try using xine or mplayer to use ALSA once and visit a flash site with
|
Maarten Vanraes <alien999999999> |
Tue 18 Oct 2005 12:04:28 PM UTC, comment #141: au8830 card 0 ALSA 1.0.10rc1 with aoss on my 32bits machine
2 instance of xine 0.99 (oss) , 2 instance of swf (local), 4 instances of mplayer 1.0pre7 (OSS)
mplayer complain /dev/rtc busy, but all applications play music concurrently.
|
Anonymous |
Tue 18 Oct 2005 10:01:05 AM UTC, comment #140: i do not have any SPDIF stuff... |
Maarten Vanraes <alien999999999> |
Tue 18 Oct 2005 09:46:40 AM UTC, comment #139: What is the model of your emu10k1 ?
It is more easy for SB Live! to support SPDIF instead of your au8810.
http://www.dearhoney.idv.tw/SoundCard/SBLive/PC2000-200002/
You can find I2S in CT4620 |
Anonymous |
Tue 18 Oct 2005 05:36:18 AM UTC, comment #138: Diamond Impact S90 with STAC9704 and wavetable daugther card connector
http://www.dearhoney.idv.tw/MUSEUM/Diamond_S90.jpg
The au8820 come with Compaq computer (OEM from Diamond)
There are black,green,red,blue connectors, PHILIPS TDA1517 and WM9701 codec. ( NO wavetable daugther card connector)
|
Anonymous |
Tue 18 Oct 2005 05:14:28 AM UTC, comment #137: au8810 is not supported by OSS.
http://www.opensound.com/osshw.html
Aureal AU8810 PCI **** NOT SUPPORTED ***
The last version of of the linux driver released by Aureal is 1.0.5
The linux driver for kernel 2.2 - 2.4
|
Anonymous |
Tue 18 Oct 2005 04:07:45 AM UTC, comment #136: You will need to trace the PCM_STATE of the application in user space.
http://www.alsa-project.org/alsa-doc/alsa-lib/pcm.html
and the corresponding callback of the driver in kernel.
http://www.alsa-project.org/~iwai/writing-an-alsa-driver/x643.htm |
Anonymous |
Tue 18 Oct 2005 02:51:20 AM UTC, comment #135: static int snd_vortex_pcm_trigger(snd_pcm_substream_t *substream, int cmd) {
printk(KERN_INFO "vortex_pcm_trigger : cmd=%d dma=%d dir=%d type=%d number=%d\n", cmd,dma,substream->stream,
|
Anonymous |
Tue 18 Oct 2005 02:34:48 AM UTC, comment #134: Why ALSA callback snd_vortex_hw_params() so many times before trigger start the DMA transfer (Did your system hang just after that happen) ?
Is it the same application on Oct 17 19:23:17 or a different application ?
Oct 17 19:24:15 queen kernel: direction: 0, type:0, subdevice: 1
|
Anonymous |
Tue 18 Oct 2005 01:12:37 AM UTC, comment #133: Oct 17 19:23:17 queen kernel: vortex: start 0
if ALSA trigger stop, why ALSA trigger start again ?
Oct 17 19:23:18 queen kernel: vortex: start 0
-----------------------------------------------------
Oct 17 19:24:15 queen kernel: vortex: start 1
Did your system hang between Oct 17 19:24:15 , Oct 17 19:24:44 ?
printk(KERN_INFO "Vortex: periods %d, period_bytes %d, channels = %d dir=%d type=%d number=%d\n", params_periods(hw_params),
printk(KERN_INFO "snd_vortex_pcm_close : dir=%d type=%d number=%d\n",
-----------------------------------------------------
#define FIFO_VALID 0x00000010
Oct 17 19:24:44 queen kernel: vortex: stop 0
Let try
nr_capt = 1;
instead of
nr_capt = nr;
|
Anonymous |
Mon 17 Oct 2005 05:34:14 PM UTC, comment #132: Oct 17 19:22:00 queen kernel: ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 17 (level, low) -> IRQ 169
|
Maarten Vanraes <alien999999999> |
Sun 16 Oct 2005 02:31:03 AM UTC, comment #131: http://www.freepatentsonline.com/6167465.html
FIG. 13 is a block diagram of an audio processing peripheral 230, in accordance with one embodiment of the present invention. Audio processing peripheral 230 includes an audio processing device 300. Audio processing device 300 is typically coupled to PCI bus 221. Multiple DMA channels 225 may be established between audio processing device 300 and main memory 212 through PCI bus 221 (see FIG. 9). Audio processing device 300 includes a PCI bus interface 302, a plurality of FIFOs 306, a DMA controller 310, a FIFO controller 320, an internal data bus 330, a register bus 331 and operational components 341-352.
Operational components 341-352 may be typical audio processing blocks that conduct real-time operations on the data retrieved through multiple DMA channels 225 established by device 300. By way of example, operational components may be a parallel port interface 341, a wavetable 342, a sample rate converter 343, a mixer 344, a digital signal processor interface 345, a coder/decoder interface 346, a serial port interface 347, a Sony/Phillips Digital Interface Format (SPDIF) 348, a joystick interface 349, a musical instrument digital interface (MIDI) interface 350, a Phillips I2C protocol interface 351 and a prior art sound card emulator 352. Components 341 and 345-351 may be coupled to the appropriate devices (not shown) which may be included on audio processing peripheral 230. However, any suitable operational component may be utilized in accordance with the present invention. By way of example, A3D, tone controller, crosstalk canceller or any other suitable operational component may be incorporated. |
Anonymous |
Sun 16 Oct 2005 02:18:29 AM UTC, comment #130: I don't know how OSS calculate the number of capture subdevice.
Should define snd_vortex_capture_hw_adb instead of using snd_vortex_playback_hw_adb.
.channel_max in snd_vortex_capture_hw_adb should be 2 for capture device 0 instead of 4 for au8830.
if (idx == VORTEX_PCM_ADB)
|
Anonymous |
Sun 16 Oct 2005 02:00:28 AM UTC, comment #129: au88x0_pcm.c
if ((err =
printk(KERN_INFO "snd_vortex_pcm_open : dir=%d type=%d umber=%d\n",substream->stream,VORTEX_PCM_TYPE(substream->pcm),substream->number);
if (VORTEX_PCM_TYPE(substream->pcm) != VORTEX_PCM_WT) {
static int snd_vortex_pcm_close(snd_pcm_substream_t * substream)
printk(KERN_INFO "snd_vortex_pcm_close : number=%d\n",substream->number);
// the hardware-specific codes will be here
sgbuf = (snd_pcm_sgbuf_t *) substream->runtime->dma_private;
printk(KERN_INFO "Vortex: periods %d, period_bytes %d, channels = %d number=%d\n", params_periods(hw_params),
spin_lock_irq(&chip->lock);
|
Anonymous |
Sun 16 Oct 2005 01:37:15 AM UTC, comment #128: For the FIFO issue.
Before au88x0 migrated from alsa-driver to alsa-kernel
case SNDRV_PCM_TRIGGER_STOP:
After au88x0 migrated to alsa-kernel
http://cvs.sourceforge.net/viewcvs.py/*checkout*/alsa/alsa-kernel/pci/au88x0/au88x0_pcm.c?rev=1.1
case SNDRV_PCM_TRIGGER_STOP:
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/au88x0/au88x0_pcm.c?r1=1.4&r2=1.5
May be you can try to use
vortex_adbdma_stopfifo(chip, dma);
instead of
vortex_adbdma_pausefifo(chip, dma);
|
Anonymous |
Sat 15 Oct 2005 02:38:30 PM UTC, comment #127: also, flash only uses 1 subdevice for input and 1 for output, no matter how many flash stuff you open... , the sound is perfectly mixed inbetween... but only 1 subdevice is used...
i'm guessing software mixing in the flash plugin... |
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 02:36:31 PM UTC, comment #126: playback uses /dev/dsp
if you do 'cat /dev/dsp > /dev/dsp' , you use both...
i'll try and log down the hw:0,0 stuff, but i don't know where those printk's should be put... |
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 02:34:58 PM UTC, comment #125: i've been testing using /dev/dsp (snd-au8810) and /dev/dsp1 (snd-emu10k1), i've left the /dev/adsp and /dev/adsp1 devices alone, haven't touched those...
as far as /dev/dsp1 using the SBLive, i can guarantee it since sound is coming out of headphone... |
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 02:31:51 PM UTC, comment #124: input/output orders are wrong:
with input: the order is 0,7,8,9,10,11,12,13,14,15,1,2
i mean by order, which sub device gets status running in what order... |
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 02:25:23 PM UTC, comment #123: i've been testing a few things...
i've seen this remarkable fact...:
snd-au8810 has 12 !!! inputs and 8 outputs !!! as usable by OSS emulation in /dev/dsp
snd-emu10k1 can only have 1 normal input by OSS /dev/dsp1
I don't know about you, but I think this is quite fishy...
more than that, the /proc/asound/card0/pcm0p/sub*/status shows me that with inputs: 4 sub of input keep being unused (nr 4,5,6,7) and 8 sub of output keep being unused (nr 2,3,4,5,6,7,14,15).
with input: the order is 0,8,9,10,11,12,13,14,15,1,2,3
I don't know if this is normal, but that's how it is... |
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 02:24:51 PM UTC, comment #122: Refer to NOTES ON KERNEL OSS-EMULATION
Unlike the real OSS, ALSA cannot use the device files more than the assigned ones. For example, the first card cannot use /dev/dsp1 or /dev/dsp2, but only /dev/dsp0 and /dev/adsp0.
As seen above, PCM and MIDI may have two devices. Usually, the first PCM device (hw:0,0 in ALSA) is mapped to /dev/dsp0 and the secondary device (hw:0,1) to /dev/adsp0 (if available).
hw:0,1 = au8810 spdif (digital out) which SPDIF hardware is not present on your aureal vortex advantage but the device 1 has been created by the au8810 driver.
Do playback and capture use /dev/dsp0 ?
Please add printk statement to log down any ALSA/OSS application open hw:0,0 for capturing/playback ? |
Anonymous |
Sat 15 Oct 2005 01:40:47 PM UTC, comment #121: i don't think you understand it...
AOSS DOES NOT support more than one application, in fact, the AOSS blocks the ALSA audio device...
BUT, when you stop your OSS app, normal ALSA stuff is back...
i can prove this by using 2 applications on /dev/sound/dsp1 (which is the OSS for the second card). I'll prove that in a bit...
anyway, when i do this with snd-au88x0, after this hang, i can kill all applications using sound until dependencies are 0...
but i can never play any sound again, until i modprobe -r snd-au88x0...
i guess that is the problem, ALSO, the OSS can't be played, when the ALSA is playing, which is happening now...
i know the OSS is limited, but i have a feeling that there is something not quite right with the ALSA driver (something about using X channels, and registering the channel as busy...)
i'll go and test out a few combinations with ALSA and OSS... |
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 01:12:38 PM UTC, comment #120: You will need to ask the ALSA developer whether ALSA support multiple OSS applications running usin AOSS Emulation without blocking.
If yes, please remove all printk statement which you have added.
We only need to know any blocking occur on playback/capture
1) snd_vortex_pcm_open() dir=playback/capture and subdevice number
|
Anonymous |
Sat 15 Oct 2005 12:53:23 PM UTC, comment #119: It the cause of hang is due to 2 or more OSS applications running/blocking, this may be a limitation of AOSS emulation instead of bugs in au88x0 driver.
Just swap the order of
modprobe snd-au8810
to
modprobe snd-emu10k1
If your computer still hang, it is not bug in au88x0 driver.
|
Anonymous |
Sat 15 Oct 2005 10:39:35 AM UTC, comment #118: ok, this time, this is the real source (make install failed last time) and i got it reproduced, allthough this time, it took some more to hung it... (2 flash apps + amarok): (first alsa is started; then played, then playing flash sound, then playing another one; then hanging; then killing all apps -9 with sound, then restarting alsa).
the temp value is the one printed after 'fail' eg: ".... fail:%i!!!\n" ...
this happens 4 times, and it's not -1 or 0xFFFFFFFF , too bad i didn't print it in HEX, it'd be easier to figure out...
Oct 15 12:20:04 queen aiccu: Succesfully retrieved tunnel information for T5247
|
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 09:37:04 AM UTC, comment #117: ok, it was pretty reproducable this time, and your code was not working, allthough i'm not sure (i might have forgotten make install) (i'll try again this time with the correct sources).
however i was able to supply logs from what is happening:
Oct 15 11:20:00 queen CROND[26511]: (root) CMD (/etc/cron.d/rrdupdate.sh)
nevermind the ups, cron, alarms and shorewall logs...
BUT: you should note, that opera gets a segfault (flash plugin in opera, which was visiting the official dido homepage, the second it was loading, sound from amarok hung and stopped; that's when i knew it can be reproduced quite easily.
i'll try to get the temp value when it fails next time, it probably won't be long... |
Maarten Vanraes <alien999999999> |
Sat 15 Oct 2005 01:36:16 AM UTC, comment #116: The matrix should be
http://www.alsa-project.org/alsa-doc/index.php?vendor=vendor-Aureal#matrix
Aureal
Vortex Advantage AU8810 Details (au8810) [MIDIio] (3)
Vortex AU8820 Details (au8820) [MIDIio] (3)
SQ2500 AU8830 Details (au8830) [MIDIio][RCAo] (3)
|
Anonymous |
Fri 14 Oct 2005 04:05:34 PM UTC, comment #115: you may very well be correct, prior to changing anything, i decided to browse a flash site with music, and something did go wrong.
I'll definately do this, and i'll probably be fastly enough to reproduce it... |
Maarten Vanraes <alien999999999> |
Fri 14 Oct 2005 03:06:43 PM UTC, comment #114: don't forget that i'm on x86_64...
are you committing all these changes to alsa CVS?
i'll try and find out the value when this fails...
I don't know when i'm going to get this problem, but when i do, i'll tell you the value of temp... |
Maarten Vanraes <alien999999999> |
Fri 14 Oct 2005 02:56:02 PM UTC, comment #113: The au88x0 driver currently don't provide 4 channels for SQ1500
static snd_pcm_hardware_t snd_vortex_playback_hw_adb = {
Need to call snd_vortex_mixer() to find out the audio codec is stereo or quad before create device 0 using snd_vortex_pcm(chip,VORTEX_PCM_ADB,NR_ADB)
// snd_ac97_mixer and Vortex mixer.
|
Anonymous |
Fri 14 Oct 2005 02:37:30 PM UTC, comment #112: What is the value of temp and fifo when the "Vortex: vortex_fifo_setadbctrl fail" appear ?
Try this to see if this fix your problem.
do {
|
Anonymous |
Fri 14 Oct 2005 12:21:27 PM UTC, comment #111: the 2.1 speaker set, only accepts 2.0 and filters the BASS itself...
I hear plenty of bass...
I never use EQUALIZER, do you mean that there is another way to do this? (internally route TVaudio-out to mixer?)
I don't think i have SPDIF, so that doesn't change much.
allthough my card only accepts 1 pair of audio out, it's advertised to have also a 4 channel mode... (but the commercial paper might be wrong)
this is not something that i have a problem with...
on the second point:
it comes down to this: the OSS has a /dev/sound/dsp and a /dev/sound/dsp1 .
for applications like gnomemeeting, skype, TeamSpeak, I select /dev/sound/dsp1 ...
so browsers would probably use the OSS from the first (snd-au8810) card... |
Maarten Vanraes <alien999999999> |
Fri 14 Oct 2005 06:40:01 AM UTC, comment #110: The Analog EQ channel in Win9x driver has not been implmeneted in ALSA driver.
http://www.dearhoney.idv.tw/SoundCard/Vortex2/2048-5.gif
This mean your Line-in will not go through the Equalizer. It just go through the analog mixer of WM9710A to your 2.1 speakers.
Use alsamixer to change EQ will not affect the BASS/TREBLE of the audio of your TV programme.
Your advanage has only two channels speaker out. Do you hear any bass from bass speaker of your 2.1 speakers ?
If you want to test EQ of au8810, you have to use mplayer,xine,xmms to play music through au8810 ?
|
Anonymous |
Fri 14 Oct 2005 03:42:34 AM UTC, comment #109: SQ1500 http://gamesurf.tiscali.it/hardware/inte5108170808100/immagini_grandi/0.jpg
SQ2500 http://gamesurf.tiscali.it/hardware/inte5108170808100/immagini_grandi/1.jpg
SQ3500 Turbo http://gamesurf.tiscali.it/hardware/inte5108170808100/immagini_grandi/2.jpg
All of them have 1 blue, 1 pink , 2 green jacks and coaxial SPDIF
|
Anonymous |
Fri 14 Oct 2005 01:43:33 AM UTC, comment #108: Please post the content of
/proc/asound/oss/sndstat
Do you have configure modprobes.conf to instruct oss emulation to use second card ?
or
You just
modprobe snd-au8810
and
modprobe -r snd-pcm-oss
Did the sound come out from emu10k1 when using oss application ?
|
Anonymous |
Thu 13 Oct 2005 03:59:16 PM UTC, comment #107: What is the value of temp and fifo when the "Vortex: vortex_fifo_setadbctrl fail" appear ?
Depends on the value of hwread(vortex->mmio, VORTEX_FIFO_ADBCTRL + (fifo << 2)) , there should be 3 conditions which will jump out of the while loop instead of only one condition (temp & FIFO_RDONLY)
do {
|
Anonymous |
Thu 13 Oct 2005 01:01:15 PM UTC, comment #106: ok, i'm getting too confused now:
let's start over:
I have:
TVcards audio-out is connected through a wire to the Line-in of and-au8810, that's all it does.
snd-au8810 speaker out, is connected to a 2.1 speaker set.
a gameport is connected to either the snd-au8810 or snd-emu10k1, but it doesn't work yet (I may have forgotten something in kernel config).
My problem is this:
regularly, the 2.1 speaker set stops playing music, for some odd reason, to solve this problem, i have to do:
modprobe -r snd-pcm-oss
followed by:
modprobe snd-au8810
this solves the problem temporarily, until it happens again.
when i remove the snd-au8810 module and it had the problem of not having sound, than i get an adbctrl fail message, if there was no problem, and i'm just restarting ALSA for fun, i do NOT get this message.
on top of that, i ALWAYS get an (fifo?) IRQ reg error when starting snd-au8810 (ever since i can remember, even when i was still using 32bit)
I only use oss with the second card and my surround headphone.
of course a webbrowser is open, and some webbrowser is playing something through OSS, that could be possible, but that can't really be the problem, cause whenever i close all applications, if i don't restart ALSA, i'll never have any sound, NOT EVEN in the snd-emu10k1...
can you supply me a patch to try out? |
Maarten Vanraes <alien999999999> |
Thu 13 Oct 2005 09:54:42 AM UTC, comment #105: If audio-out of your TVcard connect to line-in of your vortex advantage. (The analog signal only pass through the mixer of the WM9701A codec to the speakers)
However the sound-playing plugin in the web browsers seem still using OSS to play sound using the playback device. I noticed that the capture device was also opened and closed at the end of the playback.
Add printk statement to check whether the capture device has been opened and closed by the browser-plugin abnormally when you are browsing some web site which play sound.(e.g flash)
|
Anonymous |
Thu 13 Oct 2005 07:53:10 AM UTC, comment #104: The error "vortex_fifo_setadbctrl fail" also occur on au8830 in the ALSA 1.0.0rc2
Soundcard: Turtle Beach - Montego II, Chipset Aureal Semiconductor
|
Anonymous |
Thu 13 Oct 2005 06:23:43 AM UTC, comment #103: 4 channels support of au8810
http://www.soundcard-drivers.com/drivers/40/40264.htm
The win9x SQ1500 driver also contains the four 4-channels speaker-test wav files.
OUT1 = front
176446 04-21-99 15:43 OUT1LQ.WAV
These two file are used to test the left/right channels.
88322 04-21-99 15:43 OUTLS.WAV
This means SQ1500 support 4 channels if it use a 4-channels AC97 codec.
It is unlikely that Vortex Advantage support 4-channels since WM9701A codec is a high-quality stereo audio codec compliant with the Intel AC’97 Rev 1.03 specification.
http://www.alsa-project.org/alsa/ftp/manuals/wolfson/WM9701A.pdf
|
Anonymous |
Thu 13 Oct 2005 05:33:59 AM UTC, comment #102: http://lists.nongnu.org/archive/html/openvortex-dev/2005-10/msg00000.html
It seem to me that patch diff11.diff forgot to change the function declaration in au88x0_xtalk.h
static void vortex_XtalkHw_SetSampleRate(vortex_t * vortex, int sr);
|
Anonymous |
Thu 13 Oct 2005 03:49:41 AM UTC, comment #101: |
Anonymous |
Wed 12 Oct 2005 11:40:37 AM UTC, comment #100: Vortex Advantage, Vortex 1, Vortex 2 and Vortex 2 Super Quad Digital PCI Sound Card Installation Guide
http://200.14.207.53/download/AuralogToolsAndUpdates/Drivers/Sound%20cards/aureal_a3d/multi.pdf
I guess some au8830 use STAC9704 stereo codec. |
Anonymous |
Wed 12 Oct 2005 10:15:35 AM UTC, comment #99: i don't think it's used to change the mic-in to headphone.
as far as i remember, it's the line-in to rear speakers...
i remember that there was a diagram in the folder, which showed what the connectors did in each state...
i desparately need to find this folder...
i'll be home within 6hours |
Maarten Vanraes <alien999999999> |
Wed 12 Oct 2005 09:31:15 AM UTC, comment #98: i don't think my card has SPDIF ( the single connector will probably be a CD-IN AUDIO, probably not SPDIF)
but the only folder (commercial) told me that it should have 3D stuff.
furthermore, the folder told me that the line in would be reused as rear (or mic), i'm not entirely sure of this... (since i'm using the line in from the audio-out of my TVcard, i can't be sure, i've never tested something like this)
i'm at work now, when i'm home, i'll try and see if i can still find that alleged folder.
when i'm home, i'll try and fotograph the card again to see what it all is and what not...
i'm beginning to believe that the folder i read was from a vortex2 card or something...
i'm fairly certain that my au8810 will not have a modem codec, since i cannot see if there's anything to connect a modem wire to... (i can be mistaken ofcourse).
i'll try and do all these things tonight, together with patches making |
Maarten Vanraes <alien999999999> |
Wed 12 Oct 2005 09:26:10 AM UTC, comment #97: Are the jumper used to change the Mic-In to/from Headphone ? |
Anonymous |
Wed 12 Oct 2005 09:09:14 AM UTC, comment #96: Hardware EQ should be present in au8810.
If the EQ part of the AMD64 patch is correct.
Listen any music and use alsamixer to change the gain of 10-bands will tell you the presence of the hardware EQ. (e.g Increase 31Hz, 64Hz band - Bass)
SPDIF hardware is not present in your au8810 unless you can construct a D.I.Y. SPDIF circuit board
http://www.dearhoney.idv.tw/MUSEUM/Diamond_MX25.jpg
The Turtle Beach Montego II with STAC9721 stereo codec is most likely using I2S to control the rear out and SPDIF IN/OUT. (This is similar to some SBLive, using AC97 and I2S to control front and rear respectively) I2S suppose to be better than AC97 and Turtle Beach Montego Home studio was the second most expensive au8830 sound card at that time.( SQ3500 with AC3 hardware decode was the most expensive)
http://www.dearhoney.idv.tw/MUSEUM/TurtleBeach_MontegoII_Plus_DigitalIO.jpg
The Aureal Vortex Adantage was used to complete with other sound cards in the low-end market.
Can you point out the location of the jumper ?
http://www.watch.impress.co.jp/akiba/hotline/990724/image/au88101.jpg
Line-in, Mic-in and SPK-Out.
If it has 3D hardware like au8830, it should support A3D 2.0 and A3D 3.0
How do au8810 user manual describe the function of the jumper ?
Some stereo codec can output to two different jacks for headphone and speakers.
/* Enable front left/right slots of AC97-link */
vortex->isquad = ((vortex->codec == NULL) ? 0 : (vortex->codec->ext_id& AC97_EI_SDAC));
if (VORTEX_IS_QUAD(vortex) {
|
Anonymous |
Wed 12 Oct 2005 08:03:17 AM UTC, comment #95: ok, i see it: the sticker where it says advantage is on the back of the card...
are you sure that my vortex advantage doesn't have hardware eq, and 4 channel support? according to the commercial paper that came with the card, the card i have, is able to have A3D, 4Channel support (i have to adjust a jumper on the card) and hardware eq... |
Maarten Vanraes <alien999999999> |
Wed 12 Oct 2005 03:25:49 AM UTC, comment #94: Aureal Vortex FAQ - Card Images
http://www.dayc.vispa.com/faq/vtximg.htm
Aureal Vortex Sound Loss Fix
|
Anonymous |
Wed 12 Oct 2005 03:06:17 AM UTC, comment #93: Aureal Vortex = au8820
However Aureal Vortex Advantage is the cheapest card without SPDIF in the au8810 family.
The hardware EQ hardware and SPDIF is only present in the Vortex2 and SQ1500.
SQ1500 is 4 channels while others au8810 is 2 channels.
Using Aureal Advantage to represent au8810 family in ALSA may not be a good choice and lead to confusion.
Module snd-au8810, snd-au8820, snd-au8830
Module for Aureal Vortex, Vortex2 and Advantage device.
pcifix - Control PCI workarounds
This module supports all ADB PCM channels, ac97 mixer, SPDIF, hardware
Note: Some ALSA mixer applications don't handle the SPDIF sample rate
|
Anonymous |
Wed 12 Oct 2005 12:47:58 AM UTC, comment #92: http://pci-ids.ucw.cz/iii/?i=12eb0001 au8820
SQ2500 has coaxial SPDIF and the connector to Motorola DSP daugther card.
SQ3500 = SQ2500 + Motorola DSP daugther card
My white-box au8830 (Optical SPDIF) has subsystem 12eb:0088 which do not have this Motorola DSP connector, I cannot find the offical name for my au8830 on the internet
104d:8049 AU8830 Vortex 3D Digital Audio Processor
|
Anonymous |
Tue 11 Oct 2005 11:54:26 PM UTC, comment #91: P/N BA88VL10A-02
The part number of your card
http://lists.nongnu.org/archive/html/openvortex-dev/2005-09/pngqq5047afsi.png
is same as this one
|
Anonymous |
Tue 11 Oct 2005 08:17:33 AM UTC, comment #90: 00:0c.0 Class 0401: 12eb:0003 (rev 02) <--- au8810
12eb <--- (vendor id) Aureal
We need a SQ1500 user tell us the subsystem id of his SQ1500.
lspci -vnn |
Anonymous |
Tue 11 Oct 2005 06:24:42 AM UTC, comment #89: None of au8820/au8830 users have reported "IRQ reg error".
Your problem may be related to the "IRQ reg error" which is only appear in au8810.
1) patch the codec init sequence
What is the value of hwread(vortex->mmio, 0x22000) on your au8810 before snd_ac97_mixer() ?
THis may allow us to find the presence of modem codec in au8810 audio/modem combo (e.g ASUS L8400B) |
Anonymous |
Tue 11 Oct 2005 04:00:24 AM UTC, comment #88: There is no need to send any photo.
You have to turn the following photo up-side down.
http://lists.nongnu.org/archive/html/openvortex-dev/2005-09/pngqq5047afsi.png
From the shape of your card and the label.
Your au8810 sound card is definitly Vortex Advantage which don't have SPDIF hardware, you can trace the pin 81 of au8810 chip, it should be open and not connected to anything.
SQ1500 has color jacks (pink,2 green,blue), game port and a coxial digital out.
In win9x, the windows registry contain "spdifenablelist" which is a list of the pci subsystem vendor/device id of au8830 which support SPDIF.
In ALSA au88x0 driver, whenever you play audio through hw:0,0 The alsa au88x0 driver use hardware mixer to copy the signal to both ac97 codec and SPDIF.
device 0
FIFO-> SRC-> MIXER-> EQ -> AC97 -> analog speaker
It seem that we may need to disable this route for all au8810 except SQ1500.
http://lists.nongnu.org/archive/html/openvortex-dev/2003-09/msg00065.html
a) Diamond MX300 (MX25),
http://www.dearhoney.idv.tw/SoundCard/Vortex2/using_cmedia.htm
http://galadriel.mat.utfsm.cl/~mjander/aureal/techdoc/techpaper.html
The codec init sequence in current au88x0 driver is wrong for au8810/au8830.
For AU8810: 0x8068, 0x00e8
For AU8820: 0x00a8, 0x80a8, 0x80e8, 0x80a8, 0x00a8, 0x00e8
For AU8830: 0x8068, 0x00e8
The CODEC_EN is also wrong in current au88x0 code.
PCM0
SDAC0
|
Anonymous |
Mon 10 Oct 2005 04:41:44 PM UTC, comment #87: PS: what's this about my card and changing the pci ID or something...
I've not seen that it is/is not an advantage card... i'm trying to see on the pic you send to find this advantage or similar, but i cannot see it on the pic... if you can tell me where to look, i'll try making photograph again... |
Maarten Vanraes <alien999999999> |
Mon 10 Oct 2005 04:39:00 PM UTC, comment #86: OK, you lost me...
All these posts since my last post, are they all from the same person? who is it directed to? what should i do know? i'm totally lost...
all this info about spdif having or not having, what is it to me, i'm not using it, i may or may not have it, i couldn't care less, my only concern is that the driver works for everyone at 32 or 64 bit, and that it shouldn't suddenly stop playing music, and/or give errors at startup or shutdown...
i uncommented the printk's to give you guys more information...
if anyone has any patch (or latest version) i can try, to see if it suddenly stops playing.
on a second note, i've noticed a post on alsa-devel, which reported similar symptoms, but it was emu10k1 (which is my second card)... I'm not sure anymore what is the problem... |
Maarten Vanraes <alien999999999> |
Sun 09 Oct 2005 03:44:32 AM UTC, comment #85: Do you hear any sound from device 3 WaveTable of your au8820 after applying this patch ?
http://lists.nongnu.org/archive/html/openvortex-dev/2005-09/msg00057.html |
Anonymous |
Sat 08 Oct 2005 08:27:32 AM UTC, comment #84: Refer to your system log, you seem to have many printk statements lie between spin_lock and spin_unlock.
This will also hang the system especially during system shutdown.
You should comment all those unnecessary printk statements.
spin_lock();
// printk(,);
spin_unlock.
|
Anonymous |
Sat 08 Oct 2005 08:08:52 AM UTC, comment #83: Refer to page 40 of the XLerate Pro - Manual (english)
If the 20-pins expansion connector is used to connect to ModemDaugtherCard like ASUS L8400B, we will able to trace
a) 5 AC97-links circuits from 20-pins expansion connector to both primary audio ac97 codec and au8830 chip
|
Anonymous |
Sat 08 Oct 2005 04:55:41 AM UTC, comment #82: The pin 81 (red) of au8830 chip is for SPDIF OUT
http://www.dearhoney.idv.tw/SoundCard/Vortex2/mx300spdifo.htm
1-38 (bottom)
39-64
65-102 (top)
103-128
|
Anonymous |
Sat 08 Oct 2005 04:36:36 AM UTC, comment #81: Those 3D registers on au8830 are most likely not present on au8810.
Equalizer is present on au8810/au8830.
This means the device 2 should not be created for au8810 too.
|
Anonymous |
Sat 08 Oct 2005 04:24:24 AM UTC, comment #80: The value of MMIO 22000 seem to indicate the presence of modem codec , DAA, ..
What is the value of hwread(vortex->mmio, 0x22000) on your au8810 ? |
Anonymous |
Sat 08 Oct 2005 04:04:21 AM UTC, comment #79: Your photo clear indicate that your au8810 is Vortex Advantage
The "Vortex Advantage (..)" is printed on the label
http://lists.nongnu.org/archive/html/openvortex-dev/2005-09/pngqq5047afsi.png
http://pci-ids.ucw.cz/iii/?i=12eb0003:sub=1
Please submit and change the name of subsystem 12eb0003 to Vortex Advantage. |
Anonymous |
Sat 08 Oct 2005 03:43:57 AM UTC, comment #78: The current code in vortex_codec_init() is for au8820 and not for au8810/au8830
It should be
static void vortex_codec_init(vortex_t * vortex)
#ifndef CHIP_AU8820
static int __devinit snd_vortex_mixer(vortex_t * vortex)
...
vortex->isquad = ((vortex->codec == NULL) ? 0 : (vortex->codec->ext_id& AC97_EI_SDAC));
if (VORTEX_IS_QUAD(vortex) {
if ( vortex->enable_modem ) {
|
Anonymous |
Sat 08 Oct 2005 03:21:03 AM UTC, comment #77: The device 1 of au88x0 is SPDIF, however when you play audio through device 0 au88x0 adb, the hardware mixer copy the signal to SPDIF too.
Check the message "IRQ Reg error" will/will not appear by disabling the SPDIF out for your Vortex Advantage ?
|
Anonymous |
Fri 07 Oct 2005 01:33:07 PM UTC, comment #76: Refer to your photo, your au8810 is Aureal Advantage which don't have SPDIF.(digital out)
http://www.dearhoney.idv.tw/MUSEUM/Aureal_VortexV512.jpg
The black connector is CD Analog
The only au8810 sound card which has SPDIF is SQ1500 not Advantage.
Ask Ilya Pinaev <syn@newmail.ru> to find out the PCI subsystem vendor/device ID of his SQ1500
http://lists.gnu.org/archive/html/openvortex-dev/2004-08/msg00004.html
SPDIF output has a copy of the AC97 codec output
au88x0.c
if ((err = pci_read_config_dword(pci, 0x2c,
#ifndef CHIP_AU8820
au88x0_core.c
#ifndef CHIP_AU8820
|
Anonymous |
Sun 25 Sep 2005 04:07:44 PM UTC, comment #75: does anyone have any idea as to why i still get this hang?
the setadbctrl fail thing must have something to do with it, since i don't get these printk's when i just restart alsa, only when i restart alsa when it got hung.
it's just getting really irritating, i'm getting this like 2 or 3 times a week... either when playing music, or watching movies |
Maarten Vanraes <alien999999999> |
Tue 20 Sep 2005 03:18:23 PM UTC, comment #74: i get that ACPI:PCI line each time some alsa module is inserted and/or removed;
since i have a snd-emu10k1 and a snd-au8810:
a restart gives my 4 such lines... |
Maarten Vanraes <alien999999999> |
Tue 20 Sep 2005 02:41:26 PM UTC, comment #73: Why the ACPI: PCI interrupt keep changing ?
Do this "ACPI: PCI interrupt" also occcur in system log if you remove au8810 from you AMD64 ?
Sep 10 14:49:35 queen kernel: ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 17 (level, low) -> IRQ 169
....
Sep 10 14:49:36 queen kernel: vortex: IRQ reg error
Sep 10 14:51:43 queen kernel: ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 17 (level, low) -> IRQ 169 |
Anonymous |
Mon 19 Sep 2005 12:38:38 PM UTC, comment #72: well, the card probably has SPDIF (i'll check when i get home), but i'm just not using it...
BUT, this is not what's causing my sound to hang? or is it? |
Maarten Vanraes <alien999999999> |
Mon 19 Sep 2005 12:11:33 PM UTC, comment #71: http://lists.nongnu.org/archive/html/openvortex-dev/2003-09/msg00049.html
audio played through the adb device are sent over the spdif when the mixin and mixout connected by
vortex_connection_mixin_mix(vortex, en,
This mean that data is also pumping to SPDIF when you play audio through device 0 (adb) but your au8810 don't have any SPDIF to process the data.
I think the driver should enable this mixin mixout connections and device 1 (spdif) only when SPDIF is detected. |
Anonymous |
Mon 19 Sep 2005 10:34:11 AM UTC, comment #70: is this for me to try out?
i do not use spdif, so i can try this... |
Maarten Vanraes <alien999999999> |
Mon 19 Sep 2005 10:05:48 AM UTC, comment #69: I do not have ACPI enabled on my machine.
No idea about the functions of srcAddWTD() and srcdelWTD()
If your au8810 do not have digital out, comment the two routes from ADB to SPDIF.
/* Make playback routes. */
|
Anonymous |
Sun 18 Sep 2005 08:32:19 PM UTC, comment #68: don't mind the sensors, they aren't properly adjusted and set; i can only see that the temparatures are probably correct...
this thing runs 24x7 since more than a year! and on 32bit i never had this issue... THIS IS NOT A HARDWARE ISSUE!
with no fan, this amd64, which frequency scaling is disabled, would not run; with those voltages; the mobo would explode... it's quite clear that these sensors aren't correctly reporting; plus, the sensor for noticing case opening isn't attached...
back to the real world...
the fail happens only when i remove the snd-au8810 module and i only remove the module when sound hangs, i don't know if this is related to the hanging or not, but it's the only noticable thing... and i don't get them if i remove the module when sound did not hang... (as evidenced by previous logs)
I really like to have this problem fixed, it's quite a nuisance; i've only used the sound card for about an hour since i recompiled it to report extra info in the printk...
thanks for all the help so far... is there any other info you need? |
Maarten Vanraes <alien999999999> |
Sun 18 Sep 2005 09:09:24 AM UTC, comment #67: Look like hardware problem
Sep 16 17:02:27 queen sensord: -12V: +0.72 V (min = -13.18 V, max = -10.80 V)
Your -12V and -5V are abnormal.
The fans are not moving , I don't think your AMD64 like that |
Anonymous |
Sun 18 Sep 2005 08:56:51 AM UTC, comment #66: From the log, there are still plenty resources before vortex_fifo_setadbctrl() failt, I think you can remove some of printk statements related to resources (especially those between spin_lock() and spin_unlock()
Sep 16 17:37:22 queen kernel: vortex: stop 1
|
Anonymous |
Sun 18 Sep 2005 08:49:19 AM UTC, comment #65: sensord: Sensor alarm: Chip w83697hf-isa-0290: alarms: Chassis intrusion detection [ALARM]
Did someone try to open your server case or your sensor is malfunction ?
|
Anonymous |
Fri 16 Sep 2005 03:54:09 PM UTC, comment #64: ok, nvm my stupid tryouts, one can evidently see the files below this... |
Maarten Vanraes <alien999999999> |
Fri 16 Sep 2005 03:50:40 PM UTC, comment #63: just #5181 then... (i'm trying to refer to the added attachment) |
Maarten Vanraes <alien999999999> |
Fri 16 Sep 2005 03:49:29 PM UTC, comment #62: ok, the numbering didn't really work... how do i get this done correctly?
can anyone actually see the log?
trying patch #5181 |
Maarten Vanraes <alien999999999> |
Fri 16 Sep 2005 03:47:52 PM UTC, comment #61: i forgot to mention that with bug #5181 , the logs are only AFTER it already happened and it try to modprobe -r snd-au8810 ... |
Maarten Vanraes <alien999999999> |
Fri 16 Sep 2005 03:45:15 PM UTC, comment #60: just now, i've had that function fail and i had printk's in place to show what is passed, i've added it below...
|
Maarten Vanraes <alien999999999> |
Wed 14 Sep 2005 01:15:20 PM UTC, comment #59: added info to the printk with all arguments, we'll see if anything goes wrong again... |
Maarten Vanraes <alien999999999> |
Wed 14 Sep 2005 12:13:13 PM UTC, comment #58: http://lists.nongnu.org/archive/html/openvortex-dev/2003-08/msg00047.html
when the driver die at
printk(KERN_ERR "Vortex: vortex_fifo_setadbctrl fail\n");
dump the parameters of vortex_fifo_setadbctrl(vortex_t * vortex, int fifo, int b, int priority, int empty, int valid, int f)
http://lists.nongnu.org/archive/html/openvortex-dev/2003-04/msg00018.html
|
Anonymous |
Sat 10 Sep 2005 02:07:21 PM UTC, comment #57: btw: http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/au88x0/au88x0_pcm.c?r1=1.9&r2=1.10
was not the issue, i have a printk right next to it, and i don't see the message, i don't think i get there...
maybe it's a similar thing, but i don't know... |
Maarten Vanraes <alien999999999> |
Sat 10 Sep 2005 02:03:05 PM UTC, comment #56: to nr #49:
i finally had the time to recompile the kernel and fetch the latest alsa CVS and put printk's everywhere.
that was yesterday...
today I had me one of the hangs i get lately, nothing destructive, you just have to remove the module and modprobe it again, to have any sound working...
in the middle of a song i was playing and i don't think i was doing anything else besides that with audio stuff; the application frontend hangs, and no sound is played anymore (this sounds like a forgotten spinlock somewhere)
I have gathered some printk's about this in my /var/log/messages:
here is the logs with comments on what is happening:
...
Sep 10 14:37:47 queen kernel: vortex: stop 0
...
Sep 10 14:39:37 queen kernel: vortex: stop 0
...
Sep 10 14:44:45 queen kernel: vortex: stop 0
...
the above is still normal i guess, it looks like when it plays music and next song..., BUT
around here, something goes wrong and it hangs; i kill all apps, but amarokapp is still hanging,
...
Sep 10 14:49:11 queen kernel: vortex: stop 0
I do a /etc/init.d/alsa start, which does modprobe snd-au8810:
Sep 10 14:49:35 queen kernel: ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 17 (level, low) -> IRQ 169
I try a /etc/init.d/alsa restart to see how what it looks like when it works fine, but my other soundcard was in use:
Sep 10 14:50:32 queen kernel: Vortex: shutdown...<3>vortex: IRQ reg error
restarting continued after killing app that uses snd-emu10k1:
Sep 10 14:51:31 queen kernel: ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 17 (level, low) -> IRQ 169
snd-emu10k1 card:
Sep 10 14:51:32 queen kernel: ACPI: PCI interrupt 0000:00:0d.0[A] -> GSI 18 (level, low) -> IRQ 209
donno what the following is: (modprobe snd-pcm-oss ?)
Sep 10 14:51:33 queen kernel: vortex: IRQ reg error
restarting alsa:
Sep 10 14:51:39 queen kernel: Vortex: shutdown...<3>vortex: IRQ reg error
Sep 10 14:51:43 queen kernel: ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 17 (level, low) -> IRQ 169
and it works perfectly now...
NOTES:
|
Maarten Vanraes <alien999999999> |
Mon 05 Sep 2005 06:13:26 AM UTC, comment #55: Review of SQ1500 and SQ2500
http://www.maximum3d.com/article.php?cat=&id=18
Vortex AU8810 Features:
http://www.3dsoundsurge.com/press/pr127.html
Do you think AU8810 has any 3D hardware based on the price of SQ1500 and SQ2500 ?
|
Anonymous |
Mon 05 Sep 2005 01:03:39 AM UTC, comment #54: Is there any problem on AMD64 if call snd_ac97_mixer() (cold reset the ac97) before the snd_vortex_pcm() for the proposed PCM Volume control (per substream)?
diff -Naur ../au88x0_cvs/au88x0.c
|
Anonymous |
Fri 02 Sep 2005 01:18:13 PM UTC, comment #53: After apply diff11.diff, there are compile errors for XTALK
./configure --with-cards=au8810
Do Maarten Vanraes's au8810 patch different from Igor V. Kovalenko's au8820 patch ? |
Anonymous |
Mon 29 Aug 2005 11:40:06 AM UTC, comment #52: actually, in the snd-au8810 driver, i had this error in the past a lot of times, always at init, OR it's a similar one. it was always at the modprobing, but lately i don't have it anymore |
Maarten Vanraes <alien999999999> |
Mon 29 Aug 2005 11:04:33 AM UTC, comment #51: Did any of you find the "vortex: IRQ fifo error" when playing sound in system log, I suspect this error only present in au8830 with quad codec. |
Anonymous |
Thu 25 Aug 2005 06:59:08 AM UTC, comment #50: i'll try this with a printk so i'll effectively know if that's the problem.
I'll try tonight, when i'm done working.
this looks like it really might help, especially since i've had adb failure messages,
thanks alot |
Maarten Vanraes <alien999999999> |
Thu 25 Aug 2005 12:38:20 AM UTC, comment #49: Do this patch fix your problem ?
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/au88x0/au88x0_pcm.c?r1=1.9&r2=1.10 |
Anonymous |
Thu 14 Jul 2005 04:25:47 PM UTC, comment #48:
Do you have abnormal "min=0,max=63" in amixer contents ?
The current au88x0 driver do not support power suspend /resume
http://www.digitalproducer.com/pages/aureal_delivers_new_audio_proces.htm
Aggressive power management techniques as well as support for two separate ACÂ’97 CODECs make the AU8810 well suited for notebook applications.
What is the value of VORTEX_CTRL after cold reset of the au8810 ?
@@ -2663,6 +2641,7 @@
|
Anonymous |
Thu 14 Jul 2005 01:21:24 PM UTC, comment #47: well, i never had any trouble with the emu10k1, (but of course emu10k1 has more subdevices) and i don't think my dmix is default, unless you mean aplay directly.
I'm using ALSA CVS now |
Maarten Vanraes <alien999999999> |
Thu 14 Jul 2005 12:12:32 PM UTC, comment #46: It depends which cards (emu10k1 or au8810) is more stable on your AMD64. you need to specify -D hw:x since the latest version of aplay use dmix by default.
|
Anonymous |
Thu 14 Jul 2005 10:52:28 AM UTC, comment #45: you mean, i try to use gaim like that and see if i ever get the same problem again? |
Maarten Vanraes <alien999999999> |
Thu 14 Jul 2005 03:35:05 AM UTC, comment #44: In order to test it is a gaim bug or not, you can configure gaim ( aplay -D hw:x %s ) where x is the emu10k1 card number since you have au8810 and emu10k1 (both support hardware mixing) on your computer.
|
Anonymous |
Mon 11 Jul 2005 07:00:17 PM UTC, comment #43: RE post #42
|
Igor V. Kovalenko <kovalenko> |
Mon 11 Jul 2005 06:49:48 AM UTC, comment #42: http://sourceforge.net/mailarchive/forum.php?thread_id=7699552&forum_id=33141
Do this patch solve the problem of your STAC9704 ?
|
Anonymous |
Sun 10 Jul 2005 09:02:17 AM UTC, comment #41: Direct memory access in Linux for AMD64 technology requires the use of the dynamic DMA mapping API. This API provides a way to efficiently map 64-bit physical addresses to 32-bit Single Address Cycle PCI addresses, so that hardware that only supports 32-bit accesses can be used in 64-bit systems. A full description of the dynamic DMA mapping API is beyond the scope of this document. Device driver developers who need to use DMA should read DMA-mapping.txt in the /usr/src/linux/Documentation directory.
http://www.alsa-project.org/changes/v1-0-9a--v1-0-9b.txt
alsa-kernel/pci/au88x0/au88x0.c
|
Anonymous |
Thu 16 Jun 2005 03:42:00 PM UTC, comment #40:
How about Joystick(gameport) ?
|
Anonymous |
Tue 07 Jun 2005 12:05:15 PM UTC, comment #39: [alien@queen alsa-driver]$ ./configure --with-cards=au8810,au8820,emu10k1,bt87x --with-sequencer=yes --with-oss=yes
|
Maarten Vanraes <alien999999999> |
Tue 07 Jun 2005 11:31:05 AM UTC, comment #38: Would you post the output of ./configure --with-cards=au8810,au8820
|
Anonymous |
Wed 01 Jun 2005 02:26:58 AM UTC, comment #37: low-high-low-high of au8820
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/ac97/ac97_codec.c?r1=1.163&r2=1.164
Do the volume resoultion return 5 bits or 6 bits on STAC9704 ? |
Anonymous |
Mon 23 May 2005 03:38:22 PM UTC, comment #36: only the ac97_patch.c file was rev 1.82, the other three were already the revisions you noted... isn't it possible for you to try starting 9 instances of aplay on your testing system? i am not too happy to do this, as i'm pretty positive this will make have to reboot this system... unless you don't have an 64bit system... I think i won't be able to cancel an aplay instance when i hit the limit, i think that that somewhere there's a deadlock when i hit the limit of the card... i'd rather try if somethings changed that will have a chance of fixing things. if aplay was to error and stop, i wouldn't have this issue... unless that aplay is stopped, but some kind of stuff in the kernel is still there... or dependency isn't decreased... |
Maarten Vanraes <alien999999999> |
Mon 23 May 2005 02:10:38 PM UTC, comment #35: http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-utils/aplay/aplay.c?rev=1.185&view=markup
static void set_params(void)
err = snd_pcm_hw_params(handle, params);
|
Anonymous |
Mon 23 May 2005 01:57:17 PM UTC, comment #34: http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/ac97/ac97_patch.c?rev=1.84&view=markup
http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/ac97/ac97_codec.c?rev=1.185&view=markup
|
Anonymous |
Mon 23 May 2005 01:08:32 PM UTC, comment #33: what's all this code you keep spouting? I don't even know what's it for, and who it's directed to... maybe it isn't such a bad idea to switch to openvortex-dev. i have subscribed to that just in case... if you are talking to me, i do have alsa CVS so, a patch or whatever or a commit to CVS is sufficient (but you'll have to tell me, since i don't update frequently) |
Maarten Vanraes <alien999999999> |
Mon 23 May 2005 01:03:22 PM UTC, comment #32: err = snd_pcm_hw_params(handle, params);
|
Anonymous |
Mon 23 May 2005 11:44:02 AM UTC, comment #31: static int patch_sigmatel_stac9704_specific(ac97_t *ac97)
static struct snd_ac97_build_ops patch_sigmatel_stac9704_ops = {
int patch_sigmatel_stac9704(ac97_t *ac97)
|
Anonymous |
Mon 23 May 2005 09:07:58 AM UTC, comment #30: gaim might not check an error, but that's not the problem isn't it... when everything is full, and aplay is started again, it should wait, it should error and stop. and the kernel dependencies shouldn't go up and up... it shouldn't deadlock or something, no matter how many aplay's are started at the same time... the penalty for starting too much aplays at the same time shouldn't mean rebooting... |
Maarten Vanraes <alien999999999> |
Sun 22 May 2005 08:37:01 AM UTC, comment #29: lspci -v shows:
|
Igor V. Kovalenko <kovalenko> |
Sun 22 May 2005 08:06:29 AM UTC, comment #28: 16 bit data becomes 18 bits.AC97 codecs
|
Anonymous |
Sun 22 May 2005 07:29:41 AM UTC, comment #27: What model and revision is your Vortex AU8820 card ? You can find it out with "lspci"
|
Anonymous |
Sat 21 May 2005 03:18:33 AM UTC, comment #26: snd_vortex_pcm_hw_params() should return -EBUSY from vortex_adb_allocroute()
I guess gaim just spwan "aplay" and do not check any error ? |
Anonymous |
Sat 21 May 2005 03:04:41 AM UTC, comment #25: asp4mmed.inf
; Turn on/off the Sigmatel STAC9704 attenuation program
|
Anonymous |
Fri 20 May 2005 06:35:08 PM UTC, comment #24: I added the following line before returning from snd_vortex_mixer():
Result is:
I conquer isquad is false.
BTW it's a bit hard to demultiplex savannah posts :)
|
Igor V. Kovalenko <kovalenko> |
Fri 20 May 2005 01:08:48 PM UTC, comment #23: thanks for the info, but shouldn't it return an error instead of a hang? AL13N |
Maarten Vanraes <alien999999999> |
Fri 20 May 2005 12:56:06 PM UTC, comment #22: There are 16 Sample Rate Convertors in au8810, you should able to run 8 stereo streams or 16 mono streams.
You can try "3D" demo (require openal) in vortexcontrol, which allow you to play 16 mono waves (loop) concurrently |
Anonymous |
Fri 20 May 2005 12:32:34 PM UTC, comment #21: I'm sending this to confirm that I still got the audio hang with that last patch... it happens when i start gaim (which is configured to use aplay %s for each message, and the autojoining of lots of channels, causes it to play more than 16 times the same file in a very short time. this effectively to hang all audio, hang all killall command, all ps aux, and similar commands on this machine the dependencies on sound modules like snd-au8810 keep increasing with every aplay while no programs can either be killed and no other modules depend on snd-au8810... soon dependencies are around 210... i'm not sure about the number of aplay's at a very short interval, but it's hanging alright. the only way for sound to come back, is to reboot, which i will do shortly... this really is a major problem... AL13N |
Maarten Vanraes <alien999999999> |
Fri 20 May 2005 09:06:30 AM UTC, comment #20: cat /proc/asound/au8810/codec97#0/ac97#0-0+regs
0:00 = 0140
0:18 = 0707 PCM Out
0:7c = 574d
You can use vortexcontrol to reset the 10 bands equalizer to 0 db (Jazz, Pop, Rock, Classical, Dance )
|
Anonymous |
Fri 20 May 2005 06:39:02 AM UTC, comment #19: cat /proc/asound/au8820/codec97#0/ac97#0-0+regs
0:28 = ffff
Do vortex->isquad return TRUE for STAC9704 stereo codec ?
vortex->isquad should return FALSE for stereo codec |
Anonymous |
Sat 14 May 2005 06:30:21 PM UTC, comment #18: I use alsamixer to control the volume, and alsactl to save/restore.
The codec is:
Capabilities :
Current setup
codec regs at 26% Master and 25% PCM are
$ cat /proc/asound/au8820/codec97#0/ac97#0-0+regs
|
Igor V. Kovalenko <kovalenko> |
Sat 14 May 2005 02:02:49 PM UTC, comment #17: You should use alsamixer, amixer or vortexcontrol to change the volume of playback and capture.
After AC97 is reset, the ac97 registers are reset to default values (all volumes are mutes)
You need to use alsactl to save/restore mixer setting (volume)
/etc/modprobes.conf
post-install snd-au8820 /usr/sbin/alsactl restore 0
Need to know which ac97 codec on your au8820,
please post the result of
more /proc/asound/au8820/codec97#0/ac97#0-0
more /proc/asound/au8820/codec97#0/ac97#0-0+regs
|
Anonymous |
Sat 14 May 2005 01:12:15 PM UTC, comment #16: Well, I've not stated the problem clearly enough.
I looked at old sources (pre-alsa), and the following code may be missing since PCM transferred to alsa AC97 module (it was in write_mixer, just before writing the data to mixer register):
--- CUT ---
|
Igor V. Kovalenko <kovalenko> |
Sat 14 May 2005 12:38:27 PM UTC, comment #15: Are you using alsamixer to change the Master and PCM volume (low-high-low-high) ? |
Anonymous |
Sat 14 May 2005 12:06:11 PM UTC, comment #14: Mine au8820 does not have channels swap problem |
Igor V. Kovalenko <kovalenko> |
Sat 14 May 2005 11:30:24 AM UTC, comment #13: https://bugtrack.alsa-project.org/alsa-bug/view.php?id=59
Do your au8820 has the "left right channels swap problem" ?
alsa-utils
speaker-test -c 2
|
Anonymous |
Sat 14 May 2005 07:11:01 AM UTC, comment #12: yes, it looks horrible, the first 2 seem stereo and the last 2 seem mono... the equalizer itself seems fine, never tested it [root@queen RPM]# cat /proc/asound/au8810/codec97#0/ac97#0-0+regs i have other volume weirdness, i don't have low-high-low-high. my volume changes at seemingly random... |
Maarten Vanraes <alien999999999> |
Sat 14 May 2005 05:04:23 AM UTC, comment #11: Do the graphic equalizer (for au8810/au8830) in vortexcontrol ( support #103017 ) need any modification to work in AMD64 ? |
Anonymous |
Sat 14 May 2005 01:15:45 AM UTC, comment #10: Would you post the content of AC97 registers when Master and PCM control are mute, low, high, low-high-low-high ?
more /proc/asound/au8820/codec97#0/ac97#0-0+regs
Read the manual of your AC97 codec (sigmatel, wolfson or intel/ac97r21.pdf) http://www.alsa-project.org/alsa/ftp/manuals/
|
Anonymous |
Fri 13 May 2005 05:41:51 PM UTC, comment #9: These are not Ooopses, it's like the volume control value is scaled by 2 and high bit of the value is dropped. What data do you need?
|
Igor V. Kovalenko <kovalenko> |
Fri 13 May 2005 03:52:52 PM UTC, comment #8: Would you post the outputs (Master and PCM controls) of amixer on the "low-high-low-high oopses" ? |
Anonymous |
Thu 12 May 2005 06:52:52 PM UTC, comment #7: I am replying just to mention that I got a new 2.6.11 kernel, built ALSA CVS on it with your patch and you forgot to set some function prototype in a .h file. besides that it's working somewhat. I still have the occasional "IRQ reg" error, and the sound volume randomness still exists. I haven't had a hang yet, i suspect that the hang is caused by too much sound sources at once. can someone try to confirm some of these issues? |
Maarten Vanraes <alien999999999> |
Thu 12 May 2005 05:41:10 PM UTC, comment #6: Of course, it's not necessary :) I modified #ifdef'd code for it matched the common integer size pattern, and it may be turned back to life later. I currently do not monitor openvortex development closely. |
Igor V. Kovalenko <kovalenko> |
Wed 11 May 2005 01:47:36 PM UTC, comment #5: There is no need to modify those unused code in diff11.diff
#if 0
// unused code
#end if |
Anonymous |
Mon 09 May 2005 02:57:04 AM UTC, comment #4: The PCM volume is controlled by AC97 (alsa-kernel/pci/ac97/ac97_codec.c) , you need to post your "low-high-low-high oopses on amd64" problem on the alsa-devel mailing list |
Anonymous |
Sat 07 May 2005 08:22:13 AM UTC, comment #3: I've updated patch a bit to somewhat follow porting guide.
|
Igor V. Kovalenko <kovalenko> |
Sat 30 Apr 2005 07:37:32 AM UTC, comment #2: |
Anonymous |
Tue 26 Apr 2005 03:19:54 AM UTC, comment #1: Stereo playback is good, I finally learned to plug extension cable properly :)
|
Igor V. Kovalenko <kovalenko> |
Mon 25 Apr 2005 09:47:27 PM UTC, original submission:
I tried au8820 driver with kernel 2.6.12-rc3 on amd64 and found kernel faults while initializing snd-au8820
Changes summary:
|
Igor V. Kovalenko <kovalenko> |
Depends on the following items: None found
Items that depend on this one: None found
There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.
Follow 6 latest changes.
Date | Changed by | Updated Field | Previous Value | => | Replaced by |
---|---|---|---|---|---|
2005-10-20 | None | Attached File | - | ![]() |
Added au8810_remove_3d.patch, #5347 |
2005-09-16 | alien999999999 | Attached File | - | ![]() |
Added adbfail.txt, #5181 |
2005-06-07 | alien999999999 | Attached File | - | ![]() |
Added config.log, #4701 |
2005-05-07 | kovalenko | Attached File | - | ![]() |
Added diff11.diff, #4575 |
2005-05-01 | alien999999999 | Carbon-Copy | - | ![]() |
Added -email is unavailable- |
2005-04-25 | kovalenko | Attached File | - | ![]() |
Added diff10.diff, #4515 |
Wavetable 342 provides a read.sub.-- fifo, an increment.sub.-- wavetable.sub.-- pointer and a increment.sub.-- count signals to FIFO controller 320 to control the flow of data through the wavetable FIFOs. Wavetable 342 generates a corresponding set of signals for each wavetable FIFO. Wavetable 342 triggers a read.sub.-- fifo signal to FIFO controller 320 when the wavetable 342 requires data. Instead of just providing a single word or sample the wavetable FIFOs make available to the wavetable 342 any number of samples that the wavetable 342 may require in order to perform an interpolation operation.
In one embodiment wavetable 342 receives the first four samples of data from a single wavetable FIFO (e.g. 306(44)) when wavetable 342 requests data from FIFO controller 320. The first four samples corresponds to the four oldest samples stored in the FIFO. In another embodiment, wavetable receives the first four samples of data from two wavetable FIFOs (e.g. 306(44) and 306(45)) that are a stereo pair. The wavetable FIFOs may provide any number of samples or words in a single read operation triggered by a read fifo signal from wavetable 342.
Wavetable 342 may require that the same set of data previously read from FIFO (e.g. 306(44)) be read several more times before discarding the data. Therefore, wavetable 342 controls the increment of a wavetable pointer in an associated control register of wavetable control memory 507 of FIFO controller 320. Wavetable pointer is the wavetable version of data bus pointer 505(e), as described in reference to FIG. 18.
Wavetable 342 controls the increment of the wavetable pointers in the control registers of wavetable control memory 507 rather than the FIFO controller 320 automatically incrementing the wavetable pointers. Wavetable 342 triggers the increment.sub.-- wavetable.sub.-- pointer signal when the wavetable 342 requires the next pieces of data rather than the old pieces of data. FIFO controller 320 appropriately increments wavetable pointer for the specific FIFO after receiving the increment.sub.-- wavetable.sub.-- pointer signal.
Increment.sub.-- count signal determines the amount of increment. Wavetable 342 may need to increment the wavetable pointer by one sample or by a greater number of samples. The type of interpolation and sample rate conversion determines the amount of increment. If FIFO 306(44) typically provides four samples per read and increment.sub.-- count is one, three of the four of the new set of samples provided in the next read are carried over from the last set of samples. That is, only the oldest sample is "popped" from the FIFO.
The wavetable pointer is incremented to point to the beginning of the new set of four samples in the FIFO 306(44). Depending on the sample rate conversion performed by wavetable 342 increment.sub.-- count may be any number from zero to a maximum increment count. As described, FIFO controller 320 compares the wavetable pointer and the DMA pointer to determine if the FIFO has reached the trigger level. FIFO controller 320 makes a request for a DMA transfer when a FIFO triggers.