From: Steinar H. Gunderson <sgunderson@bigfoot.com>
Date: Sat, 19 Sep 2015 22:42:34 +0000 (+0200)
Subject: Some musings about firmware versions.
X-Git-Tag: 0.4~72
X-Git-Url: https://git.sesse.net/?a=commitdiff_plain;h=67bbba40bacadfaa53ed6aa0a58e56c4153e46d5;p=bmusb

Some musings about firmware versions.
---

diff --git a/bmusb.cpp b/bmusb.cpp
index 7f565f0..c3914ac 100644
--- a/bmusb.cpp
+++ b/bmusb.cpp
@@ -675,6 +675,8 @@ void BMUSBCapture::start_bm_capture()
 	//
 	//    so only first 16 bits count, and 0x0100 is a mask for ok/stable signal?
 	//
+	//    Bottom 16 bits of this register seem to be firmware version number (possibly not all all of them).
+	//
 	//    28 and 32 seems to be analog audio input levels (one byte for each of the eight channels).
 	//    however, if setting 32 with HDMI embedded audio, it is immediately overwritten back (to 0xe137002a).
 	//
@@ -684,6 +686,10 @@ void BMUSBCapture::start_bm_capture()
 	//    36 can be set to 0 with no apparent effect (all of this tested on both video and audio),
 	//    but the driver sets it to 0x8036802a at some point.
 	//
+	//    0 is written to during firmware upgrade. Probably best to stay out of it unless you know
+	//    what you're doing.
+	//
+	//
 	// register 16:
  	// first byte is 0x39 for a stable 576p60 signal, 0x2d for a stable 720p60 signal, 0x20 for no signal
 	//