Re: [PATCH 0/5] thunderbolt: Add support for runtime PM

[Mika Westerberg]

On Mon, Jun 18, 2018 at 02:07:26PM +0300, Mika Westerberg wrote:
> Hi all,
> 
> In recent PCs such as Lenovo X1 Carbon 6th generation the Thunderbolt
> controller is in RTD3 mode (Runtime D3). This is different from the
> previous modes because now the controller is present most of the time (it
> still will be hot-removed/hot-added during NVM firmware upgrade). Because
> of that we need to dynamically power it down whenever possible to save some
> power.
> 
> This patch series adds Linux runtime PM support for the Thunderbolt host
> controller driver using ICM firmware but it should be generic enough for
> future additions to allow similar functionality for the older Apple
> hardware as well (even though those system do not support full RTD3, it
> still makes it possible for the host controller to go to low power state if
> cable is not connected).
> 
> With these patches the driver automatically runtime suspends the host
> controller after being idle for 15s. The connected Thunderbolt devices (if
> any) need to support RTD3 mode as well. Typically all 3rd generation
> devices (Alpine Ridge, Titan Ridge) support this.
> 
> However, while this provides some power savings, there is more work to do
> in order to allow powering down the PCIe root port leading to the
> Thunderbolt PCIe hierarchy. This work is still in progress.
> 
> Mika Westerberg (5):
>   thunderbolt: Use 64-bit DMA mask if supported by the platform
>   thunderbolt: Do not unnecessarily call ICM get route
>   thunderbolt: No need to take tb->lock in domain suspend/complete
>   thunderbolt: Use correct ICM commands in system suspend
>   thunderbolt: Add support for runtime PM

All applied to thunderbolt.git/next.