* Sarah Sharp | 2012-01-16 12:35:49 [-0800]: >Hi Sebastian, Hi Sarah, >Here are the patches in my queue for 3.4 for the host-side UAS driver. >I'm going to send what we have off to Greg, so we can get the ball >rolling. They see us rollin, they hatin.. >I don't have anything new on top of the pull request you sent me on Dec >22nd and the three patches you've sent out to the USB mailing list since >then. okay. >I started reworking my RCU cancellation code on top of your patches, and >then I realized the major flaw in the plan was that the UAS workqueue is >global, and trying to guarantee that the workqueue was stopped when we >needed to cancel a command for a particular device wouldn't work if the >SCSI layer started a new command for a completely different UAS device, >and the UAS driver needed to add that new command to the workqueue. So >I started just ripping out the workqueue, which meant I needed to have a >plan/way to cancel status and data URBs when the setup URB failed to be >queued, and then I had to put aside the work for Christmas and more >pressing USB 3.0 power management patches. No worries. Yes, please get rid of the workqueue. Allocate all resources with GFP_KERNEL at command enqueue time. If something fails (either due to -ENOMEM or something else) clean up and tell the SCSI layer to retry or something. This should make things easier for you. >Anyway, let's get what we have into Greg's queue for 3.4 and hopefully >queue more before the merge window. Can you please test these patches >against your UAS target device and verify they don't cause any >additional issues? Otherwise they will have to wait until Wednesday, >when I can get back into the office and test with the UAS prototypes I >have. I merged a bunch of trees including your series and the result passes my test case in high and super speed against dummy_hcd. >Sarah Sharp Sebastian -- 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
