Re: [Xen-devel] pvusb performance

[Noboru Iwamatsu <n_iwamatsu@xxxxxxxxxxxxxx>]

James,

> I don't think that unlink by itself would solve it. The failure scenario
> I am think of is when Windows gives me a URB with 64k of data to read or
> write and an error occurs. I need to break that up to send to usbback,
> but if I have 128 x 512 bytes requests on the ring and the first one
> fails, usbback will still continue to execute the remaining 127. I'm not
> sure that that matters for a regular read error which is presumably a
> rare occurance, but what about for a device which could do a 'short
> read'?
We might have to understand the difference of linux and windows USB 
driver's behaviors.
In linux, I think,

- If 128 requests on the ring and the first one fails, usbback will
  still continue to the remaining 127.
  Subsequent requests may succeed or fail or not response. In ether
  case, USB device driver in the frontend can identify the urb failure,
  and re-initialize the device or unlink urbs.

- Reporting short-reads as errors or not is defined by the USB
  device driver itself.
  "USB_SHORT_NOT_OK" bit of the urb->transfer_flags is that.

I doubt host controller driver needs to clean up the error.

> Maybe we could have a sequence number in the req and so once the error
> or short read occurs, usbback throws out all the remaining requests with
> that sequence number (just returns them with a status to indicate that
> they weren't used). The expected next sequence number would be kept by
> usbback in the data structure for each and usbback would discard any
> request with an unexpected sequence number...
>
> Why do you allow up to 10 4K segments to be attached to the request? The
> upper limit for usb packet size seems to be 512 bytes... or do ISOC
> requests allow more (I don't know anything about those yet).
USB packat size and urb->transfer_buffer_length are different.
The first is the actual USB transfer size that is defined in USB Spec,
the second is the data size handled by urb.

usbfront and usbback transfers urb.

Some device drivers that require over 10 * 4K segments can not be
supported by current pvusb.

Regards,
Noboru




_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel