On Fri, 20 Aug 2004, Alan Stern wrote:
> > It's not easy, and it probably requires specialized testing equipment. > The bare minimum would be a decent oscilloscope. Ah, yes. And as the monkey said when he p***ed in the cash register, "This runs into money." > And then you'd have to > take apart a USB connector to get access to the leads. > And I would also have to know what I am doing, and what one is looking for. Perhaps I could get some help from some people I know over in the Electrical and Computer Engineering Department if I asked very politely, but would it be worth the trouble? > > But, as to VIA I have picked up a lot of discussion, and there are those > > who claim there are problems with VIA USB implementation, especially with > > some of their older chips. > > Yes indeed. > (cut) > > > > So if anyone has an idea where the problem might be addressed, I am > > willing to try to tweak something in the USB setup. > > I don't think there's much that can be tweaked. Do you suppose that > timing might be involved here? I mean the time between messages sent by > gphoto2. I guess this is conceivable. But the default port timeout for USB, in gphoto2, is 5000 milliseconds, which I would think means "Gee. Take all day if you need to." And the driver in that other OS takes about 300 milliseconds between sending the "get ready to fork over data" signal and then asking for it, and then getting it (like I said, 300 ms in all, to ask for and receive 0x8000 bytes). If one computer is much slower or faster than the other, Not much difference here. The Athlon-VIA system is 1.1G and runs at either 100 or 133 mhz on the bus (has PC133 SDRAM in it). and the Dell Pentium3 - Intel box has a 1G CPU and will take "PC66/100/133" SDRAM (actually has PC133 in it right now). I suspect that both boxes are running the bus at 100 mhz, though I am not sure how to get this information out of /proc (The Dell has very skimpy BIOS settinga snd IIRC does not give the bus speed explicitly). perhaps > the messages are arriving at the camera too quickly or too slowly. > This cannot be excluded, but I can mention too that this evening I tested the camera on another Athlon or DURON VIA-chipset bo, slightly slower than mine (850mhz), which belongs to my son. Same problem there, too. The driver for the camera gives a very nice debug output, with only one thing missing. Again, most of the data gets dropped. Again, all I did was to add the USB ID to an already-existing sq905 driver, which functions well for the rest of the sq905 cameras around the house. So now the problem is seen to exist on two VIA boxes, and not to exist on one Intel box. I am not yet ready to say that there is a trend, but it does make me wonder even more. Theodore Kilgore ------------------------------------------------------- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
