On 2013.08.22 19:29, Juan Lang wrote: > Renesas uPD720201/uPD720202 (dev 0194), driver version 2.0.32.0. (I'm > not certain which chip, I haven't looked at the board to identify.)
If you're using driver version 2.0.32.0 then your chip is a uPD720200/uPD720200A. uPD720201/uPD720202 would use driver version 3.x (on my system it's 3.0.23.0) Also, 2.0.32.0 is an old and buggy driver, which should NOT be used by anyone. Renesas released 2.1.16.0 years ago now, so if you see an apparent problem with the driver, at the very least, you should first try to look for an upgrade. The latest driver should be 2.1.39.0. This is actually indicated very prominently at the top of the Windows page for libusbx (http://windows.libusbx.org) and what's more the error you got is the exact error that we reference in the thread linked there. And yeah, right now the link to download the updated drivers doesn't work, but that's because the guys who used to host the driver files are in the process of redesigning their sites (and for some unfathomable reason, Renesas chose not to provide their own drivers on their site), so you'll have to google around to find a copy. For what is worth, I have re-tested both uPD720200/uPD720200A & uPD720201/uPD720202 with the latest drivers, and found no issue (but you want to make sure that you unplug and replug the device before running your test, in case you switch drivers). Now, considering that it's very likely that once you upgrade your Renesas drivers, as you should have done long ago, you will also confirm that your issue disappears, and that I can't help but have some doubts about the freshness of your TI and Intel drivers, I don't think I'm going to want to touch any proposal that aims at working around broken drivers, even more so if you aren't going to contact the manufacturers in the fist place. This is because: 1. Unless you can prove otherwise, it'll really be Intel or TI's job to fix their own bugs (as Renesas did), especially as, unlike us, they do have paid developers whose job it is to do just that. 2. You shouldn't equate the fact that we are easier to reach as an indication that we are going to be more willing to spend time to address a problem you encounter, and that seems to have very little to do with our software. 3. Both the IOCTLs are used by Microsoft's UsbView sample, which is as close to an official test application to validate an USB driver as you're gonna get from MS (and what's more, is an app that comes with the driver development kit). If TI and Intel failed to properly support these IOCTLs, something tells me that they haven't tried very hard when it comes to designing their drivers... > I have a hack that works for me by falling back to WinUsb when these > ioctls fail, but it's pretty hacky since I don't actually know (in > GEN_PASS) whether the device is a WinUsb device or not. I'll work on > postponing the ioctls until DEV_PASS to have something a little more > concrete (and less ugly) to propose. Unless somebody else has any inclination to review and test a workaround, that's likely to break other stuff, since you mentioned that you'd have to remove stuff that you don't really understand the use of, you're going to waste your time. A much more productive use of your time is to first make sure you use the latest drivers always, and, in case you find they are still broken, ask the manufacturer if they are aware of the issue, and what they plan to do about it. Depending on the outcome of the above, I'll be more than happy to update the Windows page and tell our users what hardware or software they should steer away from, as we did with Renesas. Regards, /Pete ------------------------------------------------------------------------------ Introducing Performance Central, a new site from SourceForge and AppDynamics. Performance Central is your source for news, insights, analysis and resources for efficient Application Performance Management. Visit us today! http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk _______________________________________________ libusbx-devel mailing list libusbx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libusbx-devel
