the last changes to trident driver

the last changes to trident driver

[Takashi Iwai]

Hi,

if someone has a Trident 4DNX (not DX), could you test the latest CVS
driver?  i hope my last change doesn't break, but i couldn't test the
board atm...


thanks in advance,

Takashi


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[James Tappin]

On Tue, 07 Jan 2003 17:04:47 +0100
Takashi Iwai <tiwai@suse.de> wrote:

> Hi,
> 
> if someone has a Trident 4DNX (not DX), could you test the latest CVS
> driver?  i hope my last change doesn't break, but i couldn't test the
> board atm...
> 
> 
> thanks in advance,
> 
> Takashi

Hi Takashi.

I get a problem with the gameport options not resolving with the following build.

00:09.0 Multimedia audio controller: Trident Microsystems 4DWave NX (rev 02)

./cvscompile --with-cards=trident --with-sequencer --without-isapnp --with-oss

gabrielle:/gsoft/build/ALSA/CVS/alsa-driver# /sbin/depmod -ae 2.4.20
depmod: *** Unresolved symbols in /lib/modules/2.4.20/kernel/sound/pci/trident/snd-trident.o
depmod:         gameport_unregister_port_R70daab68
depmod:         gameport_register_port_R98692e58

Best regards,
	James

-- 
James Tappin,             O__      "I forget the punishment for using
james@tappin.me.uk       --  \/`    Microsoft --- Something lingering
http://www.tappin.me.uk/            with data loss in it I fancy"  


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[James Tappin]

On Tue, 7 Jan 2003 19:33:19 +0000
James Tappin <james@tappin.me.uk> wrote:

> On Tue, 07 Jan 2003 17:04:47 +0100
> Takashi Iwai <tiwai@suse.de> wrote:
> 
> > Hi,
> > 
> > if someone has a Trident 4DNX (not DX), could you test the latest CVS
> > driver?  i hope my last change doesn't break, but i couldn't test the
> > board atm...
> > 
> > 
> > thanks in advance,
> > 
> > Takashi
> 
> Hi Takashi.
> 
> I get a problem with the gameport options not resolving with the
> following build.
> 
> 00:09.0 Multimedia audio controller: Trident Microsystems 4DWave NX (rev
> 02)
> 
> ./cvscompile --with-cards=trident --with-sequencer --without-isapnp
> --with-oss
> 
> gabrielle:/gsoft/build/ALSA/CVS/alsa-driver# /sbin/depmod -ae 2.4.20
> depmod: *** Unresolved symbols in
> /lib/modules/2.4.20/kernel/sound/pci/trident/snd-trident.o depmod:      
>   gameport_unregister_port_R70daab68
> depmod:         gameport_register_port_R98692e58
> 
> Best regards,
> 	James

Ooops dumbcluck error -- /usr/src/linux was pointing to the 2.5 sources,
I'll try again with the right settings.

James

-- 
James Tappin,             O__      "I forget the punishment for using
james@tappin.me.uk       --  \/`    Microsoft --- Something lingering
http://www.tappin.me.uk/            with data loss in it I fancy"  


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[James Tappin]

On Tue, 07 Jan 2003 17:04:47 +0100
Takashi Iwai <tiwai@suse.de> wrote:

> Hi,
> 
> if someone has a Trident 4DNX (not DX), could you test the latest CVS
> driver?  i hope my last change doesn't break, but i couldn't test the
> board atm...
> 
> 

Hi Takashi,
	I've managed to resolve my unresolved references problem by rebuilding
the kernel without gameport support, but now attempting to aplay a wav
file produces an immediate segfault (but it refuses to dump core).
Likewise gwc which uses the OSS interface. Also attempting to restart alsa
now (after the gwc crash) gives a series of device busy messages:

Shutting down sound driver: snd-pcm-oss: Device or resource busy
snd-mixer-oss: Device or resource busy
snd-trident: Device or resource busy
snd-pcm: Device or resource busy
snd-timer: Device or resource busy
snd-ac97-codec: Device or resource busy
snd-mpu401-uart: Device or resource busy
snd-rawmidi: Device or resource busy
snd-util-mem: Device or resource busy
snd-seq-device: Device or resource busy
snd: Device or resource busy
done

So at least for me right now it looks well-and-truly busted. If you can
let me know what further info you need and how to get it I'll try to
provide it though it make take a little while if any major builds are
needed as the box with the trident in it isn't exactly the fastest in the
world (PII 333).

James

-- 
James Tappin,             O__      "I forget the punishment for using
james@tappin.me.uk       --  \/`    Microsoft --- Something lingering
http://www.tappin.me.uk/            with data loss in it I fancy"  


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[Takashi Iwai]

At Tue, 7 Jan 2003 23:20:51 +0000,
James Tappin wrote:
> 
> On Tue, 07 Jan 2003 17:04:47 +0100
> Takashi Iwai <tiwai@suse.de> wrote:
> 
> > Hi,
> > 
> > if someone has a Trident 4DNX (not DX), could you test the latest CVS
> > driver?  i hope my last change doesn't break, but i couldn't test the
> > board atm...
> > 
> > 
> 
> Hi Takashi,
> 	I've managed to resolve my unresolved references problem by rebuilding
> the kernel without gameport support, but now attempting to aplay a wav
> file produces an immediate segfault (but it refuses to dump core).

hmm, perhaps it caused oops in kernel.
could you check the kernel message and get the trace via ksymoops if
possible?


Takashi


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[Takashi Iwai]

[-- Attachment #1: Type: text/plain, Size: 1477 bytes --]

At Wed, 8 Jan 2003 11:06:50 +0000,
James Tappin wrote:
> 
> On Wed, 08 Jan 2003 11:08:39 +0100
> Takashi Iwai <tiwai@suse.de> wrote:
> 
> TI> At Tue, 7 Jan 2003 23:20:51 +0000,
> TI> James Tappin wrote:
> TI> > 
> TI> > On Tue, 07 Jan 2003 17:04:47 +0100
> TI> > Takashi Iwai <tiwai@suse.de> wrote:
> TI> > 
> TI> > > Hi,
> TI> > > 
> TI> > > if someone has a Trident 4DNX (not DX), could you test the latest CVS
> TI> > > driver?  i hope my last change doesn't break, but i couldn't test the
> TI> > > board atm...
> TI> > > 
> TI> > > 
> TI> > 
> TI> > Hi Takashi,
> TI> > 	I've managed to resolve my unresolved references problem by rebuilding
> TI> > the kernel without gameport support, but now attempting to aplay a wav
> TI> > file produces an immediate segfault (but it refuses to dump core).
> TI> 
> TI> hmm, perhaps it caused oops in kernel.
> TI> could you check the kernel message and get the trace via ksymoops if
> TI> possible?
> TI> 
> 
> I've just logged into the machine from work and looked back at the system logs and this 
> is what I see from "ksymoops syslog.0":
> 

hmm, unfortunately the stack wasn't parsed correctly with symbols.
did you build the alsa drivers with debug option (--with-debug=full)?
anyway, please rebuild the drivers with the attached patch (and debug
option, if not yet).  i hope a fatal oops or hang-up can be avoided
now...
oh, also, please update the cvs tree again.  i've done some changes
since yesterday.


thanks,

Takashi

[-- Attachment #2: trident-nx-fix.dif --]
[-- Type: application/octet-stream, Size: 6023 bytes --]

Index: alsa-kernel/pci/trident/trident_main.c
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-kernel/pci/trident/trident_main.c,v
retrieving revision 1.19
diff -u -r1.19 trident_main.c
--- alsa-kernel/pci/trident/trident_main.c	7 Jan 2003 10:51:12 -0000	1.19
+++ alsa-kernel/pci/trident/trident_main.c	8 Jan 2003 11:56:17 -0000
@@ -2041,6 +2041,20 @@
 	.pointer =	snd_trident_playback_pointer,
 };
 
+static snd_pcm_ops_t snd_trident_nx_playback_ops = {
+	.open =		snd_trident_playback_open,
+	.close =	snd_trident_playback_close,
+	.ioctl =	snd_trident_ioctl,
+	.hw_params =	snd_trident_hw_params,
+	.hw_free =	snd_trident_hw_free,
+	.prepare =	snd_trident_playback_prepare,
+	.trigger =	snd_trident_trigger,
+	.pointer =	snd_trident_playback_pointer,
+	.copy =		snd_pcm_sgbuf_ops_copy_playback,
+	.silence =	snd_pcm_sgbuf_ops_silence,
+	.page =		snd_pcm_sgbuf_ops_page,
+};
+
 static snd_pcm_ops_t snd_trident_capture_ops = {
 	.open =		snd_trident_capture_open,
 	.close =	snd_trident_capture_close,
@@ -2052,6 +2066,20 @@
 	.pointer =	snd_trident_capture_pointer,
 };
 
+static snd_pcm_ops_t snd_trident_nx_capture_ops = {
+	.open =		snd_trident_capture_open,
+	.close =	snd_trident_capture_close,
+	.ioctl =	snd_trident_ioctl,
+	.hw_params =	snd_trident_capture_hw_params,
+	.hw_free =	snd_trident_hw_free,
+	.prepare =	snd_trident_capture_prepare,
+	.trigger =	snd_trident_trigger,
+	.pointer =	snd_trident_capture_pointer,
+	.copy =		snd_pcm_sgbuf_ops_copy_capture,
+	.silence =	snd_pcm_sgbuf_ops_silence,
+	.page =		snd_pcm_sgbuf_ops_page,
+};
+
 static snd_pcm_ops_t snd_trident_si7018_capture_ops = {
 	.open =		snd_trident_capture_open,
 	.close =	snd_trident_capture_close,
@@ -2074,6 +2102,20 @@
 	.pointer =	snd_trident_playback_pointer,
 };
 
+static snd_pcm_ops_t snd_trident_nx_foldback_ops = {
+	.open =		snd_trident_foldback_open,
+	.close =	snd_trident_foldback_close,
+	.ioctl =	snd_trident_ioctl,
+	.hw_params =	snd_trident_hw_params,
+	.hw_free =	snd_trident_hw_free,
+	.prepare =	snd_trident_foldback_prepare,
+	.trigger =	snd_trident_trigger,
+	.pointer =	snd_trident_playback_pointer,
+	.copy =		snd_pcm_sgbuf_ops_copy_capture,
+	.silence =	snd_pcm_sgbuf_ops_silence,
+	.page =		snd_pcm_sgbuf_ops_page,
+};
+
 static snd_pcm_ops_t snd_trident_spdif_ops = {
 	.open =		snd_trident_spdif_open,
 	.close =	snd_trident_spdif_close,
@@ -2085,6 +2127,20 @@
 	.pointer =	snd_trident_spdif_pointer,
 };
 
+static snd_pcm_ops_t snd_trident_nx_spdif_ops = {
+	.open =		snd_trident_spdif_open,
+	.close =	snd_trident_spdif_close,
+	.ioctl =	snd_trident_ioctl,
+	.hw_params =	snd_trident_spdif_hw_params,
+	.hw_free =	snd_trident_hw_free,
+	.prepare =	snd_trident_spdif_prepare,
+	.trigger =	snd_trident_trigger,
+	.pointer =	snd_trident_spdif_pointer,
+	.copy =		snd_pcm_sgbuf_ops_copy_playback,
+	.silence =	snd_pcm_sgbuf_ops_silence,
+	.page =		snd_pcm_sgbuf_ops_page,
+};
+
 static snd_pcm_ops_t snd_trident_spdif_7018_ops = {
 	.open =		snd_trident_spdif_open,
 	.close =	snd_trident_spdif_close,
@@ -2154,11 +2210,16 @@
 	pcm->private_data = trident;
 	pcm->private_free = snd_trident_pcm_free;
 
-	snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_PLAYBACK, &snd_trident_playback_ops);
-	snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_CAPTURE,
+	if (trident->tlb.entries) {
+		snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_PLAYBACK, &snd_trident_nx_playback_ops);
+		snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_CAPTURE, &snd_trident_nx_capture_ops);
+	} else {
+		snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_PLAYBACK, &snd_trident_playback_ops);
+		snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_CAPTURE,
 				trident->device != TRIDENT_DEVICE_ID_SI7018 ?
-					&snd_trident_capture_ops :
-					&snd_trident_si7018_capture_ops);
+				&snd_trident_capture_ops :
+				&snd_trident_si7018_capture_ops);
+	}
 
 	pcm->info_flags = 0;
 	pcm->dev_subclass = SNDRV_PCM_SUBCLASS_GENERIC_MIX;
@@ -2200,7 +2261,10 @@
 
 	foldback->private_data = trident;
 	foldback->private_free = snd_trident_foldback_pcm_free;
-	snd_pcm_set_ops(foldback, SNDRV_PCM_STREAM_CAPTURE, &snd_trident_foldback_ops);
+	if (trident->tlb.entries)
+		snd_pcm_set_ops(foldback, SNDRV_PCM_STREAM_CAPTURE, &snd_trident_nx_foldback_ops);
+	else
+		snd_pcm_set_ops(foldback, SNDRV_PCM_STREAM_CAPTURE, &snd_trident_foldback_ops);
 	foldback->info_flags = 0;
 	strcpy(foldback->name, "Trident 4DWave");
 	substream = foldback->streams[SNDRV_PCM_STREAM_CAPTURE].substream;
@@ -2246,7 +2310,9 @@
 
 	spdif->private_data = trident;
 	spdif->private_free = snd_trident_spdif_pcm_free;
-	if (trident->device != TRIDENT_DEVICE_ID_SI7018) {
+	if (trident->tlb.entries) {
+		snd_pcm_set_ops(spdif, SNDRV_PCM_STREAM_PLAYBACK, &snd_trident_nx_spdif_ops);
+	} else if (trident->device != TRIDENT_DEVICE_ID_SI7018) {
 		snd_pcm_set_ops(spdif, SNDRV_PCM_STREAM_PLAYBACK, &snd_trident_spdif_ops);
 	} else {
 		snd_pcm_set_ops(spdif, SNDRV_PCM_STREAM_PLAYBACK, &snd_trident_spdif_7018_ops);
Index: alsa-kernel/pci/trident/trident_memory.c
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-kernel/pci/trident/trident_memory.c,v
retrieving revision 1.6
diff -u -r1.6 trident_memory.c
--- alsa-kernel/pci/trident/trident_memory.c	7 Jan 2003 10:47:57 -0000	1.6
+++ alsa-kernel/pci/trident/trident_memory.c	8 Jan 2003 11:45:21 -0000
@@ -202,12 +202,20 @@
 		up(&hdr->block_mutex);
 		return NULL;
 	}
+	if (lastpg(blk) - firstpg(blk) >= sgbuf->pages) {
+		snd_printk(KERN_ERR "page calculation doesn't match: allocated pages = %d, trident = %d/%d\n", sgbuf->pages, firstpg(blk), lastpg(blk));
+		__snd_util_mem_free(hdr, blk);
+		up(&hdr->block_mutex);
+		return NULL;
+	}
+			   
 	/* set TLB entries */
 	idx = 0;
 	for (page = firstpg(blk); page <= lastpg(blk); page++, idx++) {
 		dma_addr_t addr = sgbuf->table[idx].addr;
 		unsigned long ptr = (unsigned long)sgbuf->table[idx].buf;
 		if (! is_valid_page(addr)) {
+			__snd_util_mem_free(hdr, blk);
 			up(&hdr->block_mutex);
 			return NULL;
 		}

Re: the last changes to trident driver

[James Tappin]

On Wed, 08 Jan 2003 13:00:29 +0100
Takashi Iwai <tiwai@suse.de> wrote:

TI> At Wed, 8 Jan 2003 11:06:50 +0000,
TI> James Tappin wrote:
TI> > 
TI> > On Wed, 08 Jan 2003 11:08:39 +0100
TI> > Takashi Iwai <tiwai@suse.de> wrote:
TI> > 
TI> > TI> At Tue, 7 Jan 2003 23:20:51 +0000,
TI> > TI> James Tappin wrote:
TI> > TI> > 
TI> > TI> > On Tue, 07 Jan 2003 17:04:47 +0100
TI> > TI> > Takashi Iwai <tiwai@suse.de> wrote:
TI> > TI> > 
TI> > TI> > > Hi,
TI> > TI> > > 
TI> > TI> > > if someone has a Trident 4DNX (not DX), could you test the
TI> > TI> > > latest CVS driver?  i hope my last change doesn't break,
TI> > TI> > > but i couldn't test the board atm...
TI> > TI> > > 
TI> > TI> > > 
TI> > TI> > 
TI> > TI> > Hi Takashi,
TI> > TI> > 	I've managed to resolve my unresolved references problem
TI> > TI> > 	by rebuilding
TI> > TI> > the kernel without gameport support, but now attempting to
TI> > TI> > aplay a wav file produces an immediate segfault (but it
TI> > TI> > refuses to dump core).
TI> > TI> 
TI> > TI> hmm, perhaps it caused oops in kernel.
TI> > TI> could you check the kernel message and get the trace via
TI> > TI> ksymoops if possible?
TI> > TI> 
TI> > 
TI> > I've just logged into the machine from work and looked back at the
TI> > system logs and this is what I see from "ksymoops syslog.0":
TI> > 
TI> 
TI> hmm, unfortunately the stack wasn't parsed correctly with symbols.
TI> did you build the alsa drivers with debug option
TI> (--with-debug=full)? anyway, please rebuild the drivers with the
TI> attached patch (and debug option, if not yet).  i hope a fatal oops
TI> or hang-up can be avoided now...
TI> oh, also, please update the cvs tree again.  i've done some changes
TI> since yesterday.

As far as I can tell that works, i.e. the modules load, aplay plays a
wav file and there's nothing in /var/log/syslog -- I can't tell if any
sound's coming out.

Best regards,
	James

P.S. any ideas on the incompatibility with a kernel with gameport
enabled as modules?

-- 
+------------------------+-------------------------------+---------+
| James Tappin           | School of Physics & Astronomy |  O__    |
| sjt@star.sr.bham.ac.uk | University of Birmingham      | --  \/` |
| Ph: 0121-414-6462. Fax: 0121-414-3722                  |         |
+--------------------------------------------------------+---------+


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[Takashi Iwai]

At Wed, 8 Jan 2003 12:52:27 +0000,
James Tappin wrote:
> 
> TI> 
> TI> hmm, unfortunately the stack wasn't parsed correctly with symbols.
> TI> did you build the alsa drivers with debug option
> TI> (--with-debug=full)? anyway, please rebuild the drivers with the
> TI> attached patch (and debug option, if not yet).  i hope a fatal oops
> TI> or hang-up can be avoided now...
> TI> oh, also, please update the cvs tree again.  i've done some changes
> TI> since yesterday.
> 
> As far as I can tell that works, i.e. the modules load, aplay plays a
> wav file and there's nothing in /var/log/syslog -- I can't tell if any
> sound's coming out.

thanks.  then it's better to commit it to cvs, to avoid for other
people to get oops :)

> 
> Best regards,
> 	James
> 
> P.S. any ideas on the incompatibility with a kernel with gameport
> enabled as modules?

ehm, i thought you had compiled the driver with a different kernel
tree?  you build the modules with versions, so it must be
incompatible...


ciao,

Takashi


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[James Tappin]

On Wed, 08 Jan 2003 14:05:35 +0100
Takashi Iwai <tiwai@suse.de> wrote:

TI> > P.S. any ideas on the incompatibility with a kernel with gameport
TI> > enabled as modules?
TI> 
TI> ehm, i thought you had compiled the driver with a different kernel
TI> tree?  you build the modules with versions, so it must be
TI> incompatible...

No, even after I fixed that problem, it still gave unresolved symbols
until I disabled the gameport stuff in the kernel and rebuilt that,
rebooted with the new kernel and did a clean rebuild of ALSA.

James

-- 
+------------------------+-------------------------------+---------+
| James Tappin           | School of Physics & Astronomy |  O__    |
| sjt@star.sr.bham.ac.uk | University of Birmingham      | --  \/` |
| Ph: 0121-414-6462. Fax: 0121-414-3722                  |         |
+--------------------------------------------------------+---------+


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com

Re: the last changes to trident driver

[James Tappin]

On Wed, 08 Jan 2003 14:05:35 +0100
Takashi Iwai <tiwai@suse.de> wrote:

> At Wed, 8 Jan 2003 12:52:27 +0000,
> James Tappin wrote:
> > 
> > TI> 
> > TI> hmm, unfortunately the stack wasn't parsed correctly with symbols.
> > TI> did you build the alsa drivers with debug option
> > TI> (--with-debug=full)? anyway, please rebuild the drivers with the
> > TI> attached patch (and debug option, if not yet).  i hope a fatal
> > TI> oops or hang-up can be avoided now...
> > TI> oh, also, please update the cvs tree again.  i've done some
> > TI> changes since yesterday.
> > 
> > As far as I can tell that works, i.e. the modules load, aplay plays a
> > wav file and there's nothing in /var/log/syslog -- I can't tell if any
> > sound's coming out.
> 
> thanks.  then it's better to commit it to cvs, to avoid for other
> people to get oops :)

Just a quickie to say that the updated version does indeed produce sound
as it should. I can't test anything other than PCM output as thats the
only hardware I have.

Thanks for the good work,
	James


-- 
James Tappin,             O__      "I forget the punishment for using
james@tappin.me.uk       --  \/`    Microsoft --- Something lingering
http://www.tappin.me.uk/            with data loss in it I fancy"  


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com