Overhaullllllled the way we do streaming. There is an odd bug where

a zero length command (now, no chain) used to stop the streaming.
Now it seems to do the reverse... must investigate.

Made all clock configuration into enums. The strings were painful
and there cant be that many variations that enums cant cover them.
The enums will make more sense to developers than mystery strings.

1 files changed, 10 insertions, 6 deletions

diff --git a/host/lib/usrp/usrp2/fw_common.h b/host/lib/usrp/usrp2/fw_common.h
index 7fcae6fb2..30fee6c32 100644
--- a/host/lib/usrp/usrp2/fw_common.h
+++ b/host/lib/usrp/usrp2/fw_common.h
@@ -32,6 +32,9 @@ extern "C" {
 #define _SINS_
 #endif
 
+//used to differentiate control packets over data port
+#define USRP2_INVALID_VRT_HEADER 0
+
 // size of the vrt header and trailer to the host
 #define USRP2_HOST_RX_VRT_HEADER_WORDS32 5
 #define USRP2_HOST_RX_VRT_TRAILER_WORDS32 1 //FIXME fpga sets wrong header size when no trailer present
@@ -90,8 +93,8 @@ typedef enum{
     USRP2_CTRL_ID_SETUP_THIS_DDC_FOR_ME_BRO,
     USRP2_CTRL_ID_TOTALLY_SETUP_THE_DDC_DUDE,
 
-    USRP2_CTRL_ID_CONFIGURE_STREAMING_FOR_ME_BRO,
-    USRP2_CTRL_ID_CONFIGURED_THAT_STREAMING_DUDE,
+    USRP2_CTRL_ID_SEND_STREAM_COMMAND_FOR_ME_BRO,
+    USRP2_CTRL_ID_GOT_THAT_STREAM_COMMAND_DUDE,
 
     USRP2_CTRL_ID_SETUP_THIS_DUC_FOR_ME_BRO,
     USRP2_CTRL_ID_TOTALLY_SETUP_THE_DUC_DUDE,
@@ -186,12 +189,13 @@ typedef struct{
             _SINS_ uint32_t scale_iq;
         } ddc_args;
         struct {
-            _SINS_ uint8_t enabled;
-            _SINS_ uint8_t _pad[3];
+            _SINS_ uint8_t now; //stream now?
+            _SINS_ uint8_t continuous; //auto-reload commmands?
+            _SINS_ uint8_t _pad[2];
             _SINS_ uint32_t secs;
             _SINS_ uint32_t ticks;
-            _SINS_ uint32_t samples;
-        } streaming;
+            _SINS_ uint32_t num_samps;
+        } stream_cmd;
         struct {
             _SINS_ uint32_t freq_word;
             _SINS_ uint32_t interp;