Skip to contentSkip to conversation

hdsp fixes

PATCH thomas charbonnel 4 participants 7 messages Jan 26, 2005 – Jan 28, 2005

lore
Share & tools
lore
#0 lore raw

Hi all,

Sorry for my long absence. Here's a patch against 1.0.8 that should solve the firmware issues people are experiencing.

Thomas

ChangeLog:

* init sequence cleanup and firmware upload related bugfixes
* more robust revision detection scheme
(should transparently handle new revisions)
* allow hdsploader and the kernel fw loader to coexist this is useful for cardbus user who compiled the driver in-kernel (userspace may not be ready to upload the firmware when the card is probed)
* removed confusing and obsolete passthru option (was interfering with the mixer when opening the device for capture or playback) this change requires a recompile of the userspace tools against the patched hdsp.h
* removed confusing and obsolete line_outs_monitor module param
* made precise_ptr the default behaviour, and runtime tweakable (removed corresponding module param)
* small endianness fix
* ALSA-Configuration.txt HDSP entry update
* error messages cleanup

ALSA-Configuration.txt renamed from ALSA-Configuration.txt.old +0 -7
@@ -604,13 +604,6 @@
 
     Module for RME Hammerfall DSP audio interface(s)
 
-    precise_ptr	      - Enable precise pointer (doesn't work reliably).
-			(default = 0)
-    line_outs_monitor - Send all input and playback streams to line outs
-			by default. (default = 0)
-    force_firmware    - Force a reload of the I/O box firmware
-			(default = 0)
-
     Module supports up to 8 cards.
 
     Note: The firmware data can be automatically loaded via hotplug
hdsp.c renamed from hdsp.c.old +154 -258

840 lines not on this page · raw message

hdsp.h renamed from hdsp.h.old +0 -1
@@ -63,7 +63,6 @@
 	unsigned char clock_source;
 	unsigned char autosync_ref;
 	unsigned char line_out;
-	unsigned char passthru; 
 	unsigned char da_gain;
 	unsigned char ad_gain;
 	unsigned char phone_gain;
#1 lore raw
Quoted 2 lines · thomas charbonnel

Sorry for my long absence. Here's a patch against 1.0.8 that should solve the firmware issues people are experiencing.

Wow! Thank you!!
I just tested a build of it on top of 1.0.8 and it loaded the firmware on the first try with hotplug, no problems so far in simple tests (9652
+ Digiface on top of Fedora Core 3).

Takashi-san: sorry, I did not try your smaller patch, I went with what Thomas posted.

Planet CCRMA dwellers... packages coming up soon :-) Beautiful!
-- Fernando

Quoted 18 lines

ChangeLog:
* init sequence cleanup and firmware upload related bugfixes
* more robust revision detection scheme
(should transparently handle new revisions)
* allow hdsploader and the kernel fw loader to coexist this is useful for cardbus user who compiled the driver in-kernel (userspace may not be ready to upload the firmware when the card is probed)
* removed confusing and obsolete passthru option (was interfering with the mixer when opening the device for capture or playback) this change requires a recompile of the userspace tools against the patched hdsp.h
* removed confusing and obsolete line_outs_monitor module param
* made precise_ptr the default behaviour, and runtime tweakable (removed corresponding module param)
* small endianness fix
* ALSA-Configuration.txt HDSP entry update
* error messages cleanup

#2 lore raw

|--==> Fernando Lopez-Lezcano writes:

Hi Nando,

I think I missed the first part of this thread.. :/

Is the patch available online? Has it been committed in ALSA CVS?

Thanks,

Free

Quoted 2 lines · thomas charbonnel
Quoted 2 lines

Sorry for my long absence. Here's a patch against 1.0.8 that should solve the firmware issues people are experiencing.

FL> Wow! Thank you!!
FL> I just tested a build of it on top of 1.0.8 and it loaded the firmware FL> on the first try with hotplug, no problems so far in simple tests (9652 FL> + Digiface on top of Fedora Core 3).

FL> Takashi-san: sorry, I did not try your smaller patch, I went with what FL> Thomas posted.

FL> Planet CCRMA dwellers... packages coming up soon :-) FL> Beautiful!
FL> -- Fernando

Quoted 10 lines
Quoted 10 lines

ChangeLog:
* init sequence cleanup and firmware upload related bugfixes
* more robust revision detection scheme
(should transparently handle new revisions)
* allow hdsploader and the kernel fw loader to coexist this is useful for cardbus user who compiled the driver in-kernel (userspace may not be ready to upload the firmware when the card is probed)
* removed confusing and obsolete passthru option (was interfering with the mixer when opening the device for capture or playback)

Quoted 8 lines
Quoted 8 lines

this change requires a recompile of the userspace tools against the patched hdsp.h
* removed confusing and obsolete line_outs_monitor module param
* made precise_ptr the default behaviour, and runtime tweakable (removed corresponding module param)
* small endianness fix
* ALSA-Configuration.txt HDSP entry update
* error messages cleanup

FL> _______________________________________________ FL> PlanetCCRMA mailing list
FL> [email protected]
FL> http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma

#3 lore raw
In reply to thomas charbonnel

At Wed, 26 Jan 2005 18:42:11 +0100,

Quoted 21 lines · thomas charbonnel

[1 <text/plain (7bit)>]
Hi all,

Sorry for my long absence. Here's a patch against 1.0.8 that should solve the firmware issues people are experiencing.

Thomas

ChangeLog:

* init sequence cleanup and firmware upload related bugfixes
* more robust revision detection scheme
(should transparently handle new revisions)
* allow hdsploader and the kernel fw loader to coexist this is useful for cardbus user who compiled the driver in-kernel (userspace may not be ready to upload the firmware when the card is probed)
* removed confusing and obsolete passthru option (was interfering with the mixer when opening the device for capture or playback)

Quoted 8 lines

this change requires a recompile of the userspace tools against the patched hdsp.h
* removed confusing and obsolete line_outs_monitor module param
* made precise_ptr the default behaviour, and runtime tweakable (removed corresponding module param)
* small endianness fix
* ALSA-Configuration.txt HDSP entry update
* error messages cleanup

Could you give the signed-off-by line to apply your patch? Just pass
Signed-off-by: Your Name <address>
so that the patch to the linux kernel tree is signed properly.

thanks,

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl

#4 lore raw
In reply to thomas charbonnel

Hi,

the following are some things I noticed in the patch:

hdsp.c line 2731

Quoted 10 lines
hdsp.c renamed from hdsp.c.old +4 -3
+#define HDSP_PRECISE_POINTER(xname, xindex) \
 { .iface = SNDRV_CTL_ELEM_IFACE_HWDEP, \
   .name = xname, \
   .index = xindex, \
-  .info = snd_hdsp_info_line_out, \
-  .get = snd_hdsp_get_line_out, \
-  .put = snd_hdsp_put_line_out \
+  .info = snd_hdsp_info_precise_pointer, \
+  .get = snd_hdsp_get_precise_pointer, \
+  .put = snd_hdsp_put_precise_pointer \

Do we need this as a control?
Isn't preceise_ptr reliable?

hdsp.c line 4979

Quoted 8 lines
hdsp.c renamed from hdsp.c.old +8 -0
+	if (!is_9652 && !is_9632) {
+		/* we wait 2 seconds to let freshly inserted cardbus cards do their hardware init */
+ 		if ((1000 / HZ) < 2000) {
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			schedule_timeout((2000 * HZ + 999) / 1000);
+		} else {
+			mdelay(2000);
+		}

You can use ssleep(2) here.

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl

#5 lore raw
Quoted 16 lines · Takashi Iwai

Hi,

the following are some things I noticed in the patch:

hdsp.c line 2731

Quoted 10 lines
hdsp.c renamed from hdsp.c.old +4 -3
+#define HDSP_PRECISE_POINTER(xname, xindex) \
 { .iface = SNDRV_CTL_ELEM_IFACE_HWDEP, \
   .name = xname, \
   .index = xindex, \
-  .info = snd_hdsp_info_line_out, \
-  .get = snd_hdsp_get_line_out, \
-  .put = snd_hdsp_put_line_out \
+  .info = snd_hdsp_info_precise_pointer, \
+  .get = snd_hdsp_get_precise_pointer, \
+  .put = snd_hdsp_put_precise_pointer \
Quoted 3 lines

Do we need this as a control?
Isn't preceise_ptr reliable?

It should be reliable now. It was not for the rme9652, and the hdsp code forked from there. It was badly broken until recently (Ed Wildgoose's patch, IIRC), and I think few people tried it (it defaulted to off). We don't really need it as a control, but I thought it would be easier for people to test this way. If it proves reliable, I'll eventually remove the alternative code.
Incidentally, I just sent a new patch without seeing your mail... I'm affraid you won't be fond of the new "use midi tasklet" ctl :)

Quoted 11 lines

hdsp.c line 4979

Quoted 8 lines
hdsp.c renamed from hdsp.c.old +8 -0
+	if (!is_9652 && !is_9632) {
+		/* we wait 2 seconds to let freshly inserted cardbus cards do their hardware init */
+ 		if ((1000 / HZ) < 2000) {
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			schedule_timeout((2000 * HZ + 999) / 1000);
+		} else {
+			mdelay(2000);
+		}

You can use ssleep(2) here.

Sure, I didn't know about it.
There is another instance in the code that could use ssleep also.

Thanks,
Thomas

-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl

#6 lore raw

At Thu, 27 Jan 2005 22:42:28 +0100,

Quoted 22 lines · thomas charbonnel
Quoted 16 lines · Takashi Iwai

Hi,

the following are some things I noticed in the patch:

hdsp.c line 2731

Quoted 10 lines
hdsp.c renamed from hdsp.c.old +4 -3
+#define HDSP_PRECISE_POINTER(xname, xindex) \
 { .iface = SNDRV_CTL_ELEM_IFACE_HWDEP, \
   .name = xname, \
   .index = xindex, \
-  .info = snd_hdsp_info_line_out, \
-  .get = snd_hdsp_get_line_out, \
-  .put = snd_hdsp_put_line_out \
+  .info = snd_hdsp_info_precise_pointer, \
+  .get = snd_hdsp_get_precise_pointer, \
+  .put = snd_hdsp_put_precise_pointer \
Quoted 3 lines

Do we need this as a control?
Isn't preceise_ptr reliable?

Quoted 6 lines

It should be reliable now. It was not for the rme9652, and the hdsp code forked from there. It was badly broken until recently (Ed Wildgoose's patch, IIRC), and I think few people tried it (it defaulted to off). We don't really need it as a control, but I thought it would be easier for people to test this way. If it proves reliable, I'll eventually remove the alternative code.

OK.

Quoted 2 lines

Incidentally, I just sent a new patch without seeing your mail... I'm affraid you won't be fond of the new "use midi tasklet" ctl :)

No problem :)

Takashi

-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl