[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Patch-gnuradio] [PATCH 0/4] usrp: libusb-1.0
From: |
Eric Blossom |
Subject: |
Re: [Patch-gnuradio] [PATCH 0/4] usrp: libusb-1.0 |
Date: |
Mon, 17 Aug 2009 18:08:32 -0700 |
User-agent: |
Mutt/1.5.18 (2008-05-17) |
On Mon, Aug 17, 2009 at 06:08:28PM -0400, Thomas Tsou wrote:
> This patch set updates the usrp to support libusb-1.0. Asynchronous I/O
> through libusb is added with fusb_libusb1.*, which is heavily based on
> fusb_linux.*. In short, URB's and ioctl calls are replaced with
> libusb_transfer structs and native calls. Transfer reaping is handled by
> libusb and associated callbacks. I get 32Mb/s on all of my machines using
> test_usrp_standard_rx or tx. Due to the API rewrite in 1.0 from 0.12, there
> are alot of changes, many are simply name changes.
>
> Known Issues:
>
> Transmit and receive both work, but not at same time (e.g.
> usrp_benchmark_usb.py). libusb does not create any internal threads, so for
> a single session fusb_libusb1 works in the same manner as fusb_linux with
> the callback called at controlled times. With multiple libusb sessions the
> callback may occur at any time and threading issues come into play causing
> behavior to become undefined. The use of separate libusb_contexts _might_
> solve this issue; I have not had the time to look into it.
>
> Comments welcome.
>
> Thanks.
>
> --
> Thomas Tsou
> address@hidden
Thanks Thomas!
In the places where you removed the usb_strerror(), is there any way
to get additional information about why the calls failed?
Did you happen to do any cpu utilization comparison between the 0.12
and 1.0 versions? oprofile?
Eric