[0.100.3] CVS Commit

[Rui Nuno Capela]

* Make channels 3+4 of US428 work.

  This is done by hacking driver->capture_interleave_skip
  in usx2y_driver_start(). Other changes in usx2y.c improve
  rawusb mode debugging.

  An "if (unlikely(x))" optimization commonly found in kernel
  code, applied to alsa_driver_run_cycle(), as proposed by
  Lee Revell.

  Signed-off-by: Karsten Wiese <[hidden email]>
  Signed-off-by: Rui Nuno Capela <[hidden email]>

Index: configure.ac
===================================================================
RCS file: /cvsroot/jackit/jack/configure.ac,v
retrieving revision 1.64
diff -u -r1.64 configure.ac
--- configure.ac 16 Aug 2005 10:53:06 -0000 1.64
+++ configure.ac 22 Aug 2005 20:36:33 -0000
@@ -15,7 +15,7 @@
 dnl ---
 JACK_MAJOR_VERSION=0
 JACK_MINOR_VERSION=100
-JACK_MICRO_VERSION=2
+JACK_MICRO_VERSION=3

 dnl ---
 dnl HOWTO: updating the jack protocol version
Index: drivers/alsa/alsa_driver.c
===================================================================
RCS file: /cvsroot/jackit/jack/drivers/alsa/alsa_driver.c,v
retrieving revision 1.69
diff -u -r1.69 alsa_driver.c
--- drivers/alsa/alsa_driver.c 15 Jun 2005 09:21:31 -0000 1.69
+++ drivers/alsa/alsa_driver.c 22 Aug 2005 20:36:34 -0000
@@ -1681,10 +1681,10 @@

  DEBUG ("alsaback from wait, nframes = %lu", nframes);

- if (wait_status < 0)
+ if (unlikely(wait_status < 0))
  return -1; /* driver failed */

- if (nframes == 0) {
+ if (unlikely(nframes == 0)) {

  /* we detected an xrun and restarted: notify
  * clients about the delay.
Index: drivers/alsa/usx2y.c
===================================================================
RCS file: /cvsroot/jackit/jack/drivers/alsa/usx2y.c,v
retrieving revision 1.3
diff -u -r1.3 usx2y.c
--- drivers/alsa/usx2y.c 9 Jun 2005 22:10:35 -0000 1.3
+++ drivers/alsa/usx2y.c 22 Aug 2005 20:36:34 -0000
@@ -31,7 +31,12 @@
 #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
 #endif

+//#define DBGHWDEP

+#ifdef DBGHWDEP
+int dbg_offset;
+char dbg_buffer[8096];
+#endif
 static
 int usx2y_set_input_monitor_mask (jack_hardware_t *hw, unsigned long mask)
 {
@@ -118,7 +123,7 @@
  if (dbg_offset < (sizeof(dbg_buffer) - 256))
  dbg_offset += sprintf(dbg_buffer + dbg_offset, "avail %li@%p\n",
*playback_avail, driver->playback_addr[0]);
  else {
- jack_error(dbg_buffer);
+ printf(dbg_buffer);
  return -1;
  }
 #endif
@@ -143,7 +148,7 @@
  return 0; /* FIXME: return -1; */
  h->capture_iso_bytes_done = 0;
 #ifdef DBGHWDEP
- dbg_offset = sprintf(dbg_buffer, "first iso = %i %i@%p:%i\n",
+ dbg_offset = sprintf(dbg_buffer, "cfirst iso = %i %i@%p:%i\n",
  iso, h->hwdep_pcm_shm->captured_iso[iso].length,
  h->hwdep_pcm_shm->capture0x8,
  h->hwdep_pcm_shm->captured_iso[iso].offset);
@@ -152,7 +157,7 @@
  iso = h->capture_iso_start;
  }
 #ifdef DBGHWDEP
- dbg_offset += sprintf(dbg_buffer + dbg_offset, "iso = %i(%i;%i); ", iso,
+ dbg_offset += sprintf(dbg_buffer + dbg_offset, "ciso = %i(%i;%i); ", iso,
  h->hwdep_pcm_shm->captured_iso[iso].offset,
  h->hwdep_pcm_shm->captured_iso[iso].frame);
 #endif
@@ -178,10 +183,21 @@
  ((chn & 1) ? driver->capture_sample_bytes : 0);
  }
 #ifdef DBGHWDEP
+ {
+ int f = 0;
+ unsigned *u = driver->capture_addr[0];
+ static unsigned last;
+ dbg_offset += sprintf(dbg_buffer + dbg_offset, "\nvon %6u  bis %6u\n",
last, u[0]);
+ while (f < *capture_avail && dbg_offset < (sizeof(dbg_buffer) - 256)) {
+ if (u[f] != last + 1)
+ dbg_offset += sprintf(dbg_buffer + dbg_offset, "\nooops %6u  %6u\n",
last, u[f]);
+ last = u[f++];
+ }
+ }
  if (dbg_offset < (sizeof(dbg_buffer) - 256))
  dbg_offset += sprintf(dbg_buffer + dbg_offset, "avail %li@%p\n",
*capture_avail, driver->capture_addr[0]);
  else {
- jack_error(dbg_buffer);
+ printf(dbg_buffer);
  return -1;
  }
 #endif
@@ -197,6 +213,12 @@

  usx2y_t *h = (usx2y_t *) driver->hw->private;

+ if (driver->capture_nchannels == 4) {
+ // US428 channels 3+4 are on a seperate 2 channel stream.
+ // ALSA thinks its 1 stream with 4 channels, so we have to hack here.
+ driver->capture_interleave_skip = 2 * driver->capture_sample_bytes;
+ }
+
  driver->poll_last = 0;
  driver->poll_next = 0;

Index: jack/internal.h
===================================================================
RCS file: /cvsroot/jackit/jack/jack/internal.h,v
retrieving revision 1.81
diff -u -r1.81 internal.h
--- jack/internal.h 11 May 2005 02:06:46 -0000 1.81
+++ jack/internal.h 22 Aug 2005 20:36:34 -0000
@@ -445,5 +445,13 @@
 jack_port_t *jack_port_by_name_int (jack_client_t *client,
     const char *port_name);

+#ifdef __GNUC__
+#  define likely(x) __builtin_expect((x),1)
+#  define unlikely(x) __builtin_expect((x),0)
+#else
+#  define likely(x) (x)
+#  define unlikely(x) (x)
+#endif
+
 #endif /* __jack_internal_h__ */





-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel

[Thomas Charbonnel]

Hi all,

Do we really want this ?

> - jack_error(dbg_buffer);
> + printf(dbg_buffer);

Thomas




-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel

[Rui Nuno Capela]

> Do we really want this ?
>
>> - jack_error(dbg_buffer);
>> + printf(dbg_buffer);
>

No, we don't.

However, I don't think we should mind at all. That's only effective on a
very special debug mode (#ifdef DBGHWDEP) which is not compiled in by
default. Ultimately, it's only of use to Karsten, not us mere mortals ;)
As he's the one to credit for the original usx2y code, I find it fair to
make like so. After all he's the one who shall debug the stuff, isn't he?
Or so I believe :)

Cheers.
--
rncbc aka Rui Nuno Capela
[hidden email]



-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel