3.16.80-rc1 review patch. If anyone has any objections, please let me know. ------------------ From: Rick Tseng <rtseng@xxxxxxxxxx> commit a70bcbc322837eda1ab5994d12db941dc9733a7d upstream. NVIDIA 3.1 xHCI card would lose power when moving power state into D3Cold. Thus we need to wait for CNR bit to clear in xhci resume, just as in xhci init. [Minor changes to comment and commit message -Mathias] Signed-off-by: Rick Tseng <rtseng@xxxxxxxxxx> Signed-off-by: Mathias Nyman <mathias.nyman@xxxxxxxxxxxxxxx> Link: https://lore.kernel.org/r/1570190373-30684-6-git-send-email-mathias.nyman@xxxxxxxxxxxxxxx Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> [bwh: Backported to 3.16: Add xhci as parameter to xhci_handshake()] Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx> --- drivers/usb/host/xhci.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) --- a/drivers/usb/host/xhci.c +++ b/drivers/usb/host/xhci.c @@ -1029,6 +1029,18 @@ int xhci_resume(struct xhci_hcd *xhci, b hibernated = true; if (!hibernated) { + /* + * Some controllers might lose power during suspend, so wait + * for controller not ready bit to clear, just as in xHC init. + */ + retval = xhci_handshake(xhci, &xhci->op_regs->status, + STS_CNR, 0, 10 * 1000 * 1000); + if (retval) { + xhci_warn(xhci, "Controller not ready at resume %d\n", + retval); + spin_unlock_irq(&xhci->lock); + return retval; + } /* step 1: restore register */ xhci_restore_registers(xhci); /* step 2: initialize command ring buffer */
