> -----Original Message----- > From: Alan Stern [mailto:stern@xxxxxxxxxxxxxxxxxxx] > Sent: Tuesday, August 17, 2010 10:46 PM > To: Xu, Andiry > Cc: sarah.a.sharp@xxxxxxxxxxxxxxx; linux-usb@xxxxxxxxxxxxxxx; > gregkh@xxxxxxx > Subject: Re: [PATCH 1/2] xHCI: change xhci_reset_device() to allocate new > device > > On Tue, 17 Aug 2010, Andiry Xu wrote: > > > >From fb54a4791208d08d42c1bae8f58b2657301a1143 Mon Sep 17 00:00:00 2001 > > From: Andiry Xu <andiry.xu@xxxxxxx> > > Date: Mon, 16 Aug 2010 15:52:18 +0800 > > Subject: [PATCH 1/2] xHCI: change xhci_reset_device() to allocate new > device > > > > Rename xhci_reset_device() to xhci_discover_or_reset_device(). > > If xhci_discover_or_reset_device() is called to reset a device which > does > > not exist, it calls xhci_alloc_dev() to re-allocate the device. > > > > This would prevent the reset device failure, possibly due to the xHC > restore > > error during S3/S4 resume. Note this change would affect > xhci_address_device() > > behavior: udev->config exists, but the device is actually not configured > yet. > > Add a flag in struct xhci_virt_device to indicate its actual state. > > Hmmm. This means that the place where the alloc_dev method gets called > in core/usb.c:usb_alloc_dev() is now redundant. Do you think it should > get removed? > To remove alloc_dev method calling in usb_alloc_dev(), two more places need to be modified: xhci_check_args() in xhci_discover_or_reset_device(), to allow udev with slot_id = 0 to be allocate; and reset super speed device in hub_port_init(). I tried and it seems work fine. But I wonder if there will be any potential issue, especially xhci_discover_or_reset_device() uses udev->slot_id to allocate or reset device. Will there be any udev->slot_id conflict issue in certain cases? Sarah, what's your opinion? Thanks, Andiry -- 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
