From: Philip.Joslin@xxxxxxxxxxxxx > Knowing that you have experts in USB 3.0 support for Linux, I am writing to your organization hoping > that you may be able to offer some direction on a problem I have encountered. > > I am trying to track down a problem I am having with a USB 3.0 PCIe-x1 card on Linux while trying to > talk to a Cypress FX3 explorer board. I posted the following to the Cypress forums. However, after > further investigation, I think it is the ASM1042 controller on the PCIe card. Other than going to a > different USB 3.0 card, is there anything from the Linux side that I might be able to do to get by the > error condition described below? I am doing asynchronous reads through libusb. Thanks for your help > and insight. There is a bug in one of the USB host controllers (the one on a AMD board I have and might be the asmedia one) that it will only process one entry of the control? ring each time the doorbell is rung. This stops one of the usb ethernet drivers running on that system - they work fine on an Intel i7. This was all a while ago, but I don't remember seeing an xhci fix go through (to re-ring the doorbell if there are further requests outstanding). In the ethernet driver that happened when it did async commands - so two (or more) would get queued. Each time a new command got added an old one would be actions. I didn't try to fix this, and no longer have all the required hardware. But it might all be related to what you are seeing. David -- 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
