RE: [PATCH] xhci: wait CNR when doing xhci resume

[Rick Tseng <rtseng@xxxxxxxxxx>]

Hi Mathias,

Thanks for suggestion.
The reason I do not use xhci_handshake() is we get build fail when configuring below as module:
USB_XHCI_HCD = m
USB_XHCI_PCI = m

Fail message as below:
ERROR: "xhci_handshake" [drivers/usb/host/xhci-pci.ko] undefined!

So I write my own function to check CNR.

Thanks,
Rick

--
nvpublic

  
寄件者: Mathias Nyman <mathias.nyman@xxxxxxxxxxxxxxx>
 寄件日期: 2019年8月13日 上午 03:39
 收件者: Oliver Neukum <oneukum@xxxxxxxx>; Rick Tseng <rtseng@xxxxxxxxxx>; mathias.nyman@xxxxxxxxx <mathias.nyman@xxxxxxxxx>; gregkh@xxxxxxxxxxxxxxxxxxx <gregkh@xxxxxxxxxxxxxxxxxxx>
 副本: linux-usb@xxxxxxxxxxxxxxx <linux-usb@xxxxxxxxxxxxxxx>
 主旨: Re: [PATCH] xhci: wait CNR when doing xhci resume 
   
 
On 12.8.2019 11.19, Oliver Neukum wrote:
 > Am Montag, den 12.08.2019, 15:24 +0800 schrieb Rick Tseng:
 >> From: Rick <rtseng@xxxxxxxxxx>
 >>
 >> NVIDIA 3.1 xHCI card would lose power when moving power state into D3Cold.
 >> Thus we need to wait CNR bit to clear when xhci resmue as xhci init.
 > 
 > Should any controller have CNR set? Why is this specific to a vendor?
 > 
 
 Good point, always checking CNR in resume sounds like a good idea.
 
 And use xhci_handshake() as Felipe pointed out, just like in xhci_reset():
 
 ret = xhci_handshake(&xhci->op_regs->status, STS_CNR, 0, 10 * 1000 * 1000);
 
 Rick, would you like to write a patch for this?
 
 -Mathias