> On Fri, Aug 01, 2014 at 11:00:08AM +0800, Wang, Yu wrote: > > From: "Wang, Yu" <yu.y.wang@xxxxxxxxx> > > > > By current xhci ISR request/release flow, the IRQ request be done by > > xhci driver(xhci_run) whatever it is legacy irq or MSI. But the IRQ > > release flow is a little mess, MSI release be handled by xhci > > driver(xhci_stop), and legacy IRQ release be handled by USB core when > > hcd is removing(usb_remove_hcd). > > > > There have one bug if XHCI_BROKEN_MSI quirk be set and meet STS_SRE > > error during xhci resuming(xhci_resume), it will cause the xhci ISR be > > requested more times. > > > > The reason is xhci driver will try to re-initialize all registers and > > related structures to rescue the restore operation failure. And it > > will call xhci_cleanup_msix to release the ISR and call xhci_run to > > re-request it. In this flow, hcd will not be remove so USB core will > > not to free the legacy xhci ISR. For MSI case, it should be fine due > > to they will be free in xhci_cleanup_msix. But if XHCI_BROKEN_MSI > > quirk be set, the xhci will use legacy irq then reproduce this issue. > > > > This patch will unify the xhci irq request/release flow in xhci driver > > to avoid involve USB core. > > > > Change-Id: I85d589b9c7849c471595da89e5b63347bc3a2939 > > Why is this line here? Please don't do that... [Yu:] Oh. Sorry. It is some intel internal tool add it. I will remove it. Thanks Yu -- 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
