On Fri, Jan 17, 2014 at 01:27:01PM +0000, David Laight wrote: > At least some hardware (eg ASMedia PCIe cards) don't like processing a LINK > TRB and then finding that the 'cycle' bit on the next TRB is 'unset'. As I mentioned, this patch is much too big to be backported to the stable trees. You need to break it up into the very minimal amount of code necessary to fix the ASMedia bug, and then clean up patches on top of that. The clean up patches can be queued for just the current kernel version. So, please try to turn this into a series of patches. I would suggest: 1. A minimal patch to ensure a link TRB isn't given over to hardware until after the TD it points to is. Your original patch seemed like the right size and complexity for that small fix. 2. A patch that removes the unused cleanup code at the end of xhci_queue_isoc_tx(). If it is unused as you claim, it does not need to be part of the minimal bug fix patch, since the code will never be run. 3. A patch to remove the 'flip_cycle' parameter to td_to_noop(), since the only code that used it was in xhci_queue_isoc_tx(). 4. Make any simplifications of the cycle handover code. Do not include simplifications in the minimal patch. Please break the simplifications up into easily-readable chunks. 5. Trace short reads only when they're unexpected. I know you've been trying to push some of these cleanups in previous patches, but it's not generally a good idea to push both cleanups and bug fixes into stable. I've been bitten on that in the past, with a simple bug fix and arithmetic shortcut combined in one patch causing four different stable kernels to fail, and a big mess to fix all of them. Please break up these changes into separate commits. It's too hard to review and catch bugs with this large of a change. > - if (trb_comp_code == COMP_SHORT_TX) > + if (*status == -EREMOTEIO) This is an unrelated change not described in your patch description. Please remove it from the bug fix patch, and create a separate patch describing why you think it's necessary. Sarah Sharp -- 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
