Changeset 31043 in vbox

Timestamp:

Jul 23, 2010 9:26:17 AM (15 years ago)

Author:

vboxsync

svn:sync-xref-src-repo-rev:

63965

Message:

HDA: More coded review comments.

File:

• trunk/src/VBox/Devices/Audio/DevCodec.cpp (modified) (3 diffs)

trunk/src/VBox/Devices/Audio/DevCodec.cpp

@@ -59 +59 @@
 #define CODEC_B_INDEX(cmd)      ((cmd) & CODEC_VERB_B_INDEX)
 
+//** @todo r=michaln: Please document why this is bit 36 (it's not the same as the link format)
 #define CODEC_RESPONSE_UNSOLICITED RT_BIT_64(36)
 
@@ -89 +90 @@
     Log(("codecUnimplemented: cmd(raw:%x: cad:%x, d:%c, nid:%x, verb:%x)\n", cmd,
         CODEC_CAD(cmd), CODEC_DIRECT(cmd) ? 'N' : 'Y', CODEC_NID(cmd), CODEC_VERBDATA(cmd)));
+    //** @todo r=michaln: The zero used as a codec ID shouldn't be hardcoded!
     if (CODEC_CAD(cmd) != 0)
+        //** @todo r=michaln: Are we really supposed to respond to commands sent to nonexistent codecs??
+        //** @todo r=michaln: Where in the specs is this response format defined?
         *pResp = ((uint64_t)CODEC_CAD(cmd) << 4)| 0xFF;
     else
@@ -485 +489 @@
     {
         *pfn = CODEC_CAD(cmd) != 0 ? codecUnimplemented : codecBreak;
+        //** @todo r=michaln: Why "intelHD" and not e.g. "HDAcodec"?
+        //** @todo r=michaln: There needs to be a counter to avoid log flooding (see e.g. DevRTC.cpp)
         LogRel(("intelHD: cmd %x was ignored\n", cmd));
         return VINF_SUCCESS;