Hi Andrzej, Am Montag, 20. August 2018, 13:37:24 CEST schrieb Andrzej Hajda: > On 14.08.2018 12:26, Heiko Stuebner wrote: > > >From a specified output port of one dsi controller this function allows to > > iterate over the list of registered dsi controllers trying to find a second > > instance connected to the same display, like it is used in dual-dsi setups. > > As I said earlier it looks for me incorrectly. You create helper to > access/control one DSI controller from another but there is no point in > doing it. If CRTC is able to send video stream via two links it should > be CRTC's responsibility to setup and use DSI controllers attached to > these links. If the panel is able to combine stream from two DSI links > into one image it should be panel's responsibility to inform DSI > controllers about each link's configuration. > > Moreover from usage of this helper (patch 8/8) it looks like you > assumes, that second host must be also synopsys, this is not > true/generic at all. > > I know there could be many obstacles in alternative approach, so you > really insist on this hack, please use it inside the driver, you can use > for example driver_for_each_device/driver_find_device to find 2nd DSI host. Ok, I'll move that inside the rockchip-driver. My rationale still is that both msm (with a Committer-Signed-off by Rob Clark) and Tegra (code from Thierry), do follow the expectation of two controllers of the same type) ... people doing graphics stuff way longer than me. So I'd like to believe the valid issue you are raising will have been discussed at least two times already ;-) . So I really prefer not to open that can of worms myself if I don't have to ;-) . Heiko
