At Mon, 24 Jan 2011 13:10:58 +0100, Torsten Schenk wrote:
Hello,
I had to review the code since an error came up: when sound was playing and I disconnected the device, the whole system crashed. This bug is now solved (pcm.c: check if a isoc-packet has invalid status on retire, if so, abort streaming).
diff -Nur a/sound/usb/6fire/pcm.c b/sound/usb/6fire/pcm.c --- a/sound/usb/6fire/pcm.c 1970-01-01 01:00:00.000000000 +0100 +++ b/sound/usb/6fire/pcm.c 2011-01-20 23:07:24.000000000 +0100
...
+/* keep next two synced with
- FW_EP_W_MAX_PACKET_SIZE[] and RATES_MAX_PACKET_SIZE */
+static const int RATES_IN_PACKET_SIZE[] = { 228, 228, 420, 420, 404, 404 }; +static const int RATES_OUT_PACKET_SIZE[] = { 228, 228, 420, 420, 604, 604 }; +static const int RATES[] = { 44100, 48000, 88200, 96000, 176400, 192000 };
Any reason to use capital letters for these? Because they are const?
That was exactly the idea.
Hm, but it's not so common. Not too annoying as well, though.
Ok, I thought it would be that way... No matter, I changed it to small letters.
Why x86-dependent?
I thought so because of the firmware stuff. If bit- or byte-order are changed, the firmware uploading might not work. If I figured out the #ifdef stuff for bit and byte order, I will remove this dependency.
But you are decoding in bytes, not in words or so. Thus the CPU byte-order doesn't matter.
I reviewed the code and found exactly one endian-dependend thing: a const array I defined with type u16. I changed that to u8 and removed the x86 dependency.
Thanks. I committed your patch yesterday to sound git tree.
For further works, please base on the latest sound git tree.
Takashi