On Tue, 26 Aug 2014, Jassi Brar wrote:
> Cool... we can avoid runtime calculations by maybe picking the
> pre-defined 'length pattern' at module load time to match the rate
> selected. And have those many usb requests allocated and their
> 'length' initialized to the pattern. Then the rest of code would
> remain unchanged. Though I won't be surprised if you have some better
> idea.
The normal approach is to perform a simple runtime calculation (no
pre-allocated pattern). It's not complex.
Let S be the number of samples per second at the nominal transfer rate
(for example, S = 44100). Let R be the number of packets per second
(1000 because you transfer one packet every millisecond).
Precompute M = DIV_ROUND_UP(S,R); that is the maximum number of samples
you ever have to send. So the maxpacket value is M * (number of bytes
per sample), for example, M * 4 for 16-bit stereo. That value should
be stored in the endpoint descriptor.
Also precompute n = S / R and p = S % R (the truncated quotient and the
remainder). Each packet will contain either n or n + 1 samples.
Here's how you decide which:
When the data stream begins, set x = 0. When preparing each packet,
do:
if (x < R) {
put n samples in the next packet;
} else {
x = x - R;
put n + 1 samples in the next packet;
}
x = x + p;
That always gives exactly what you want.
Alan Stern
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
