On Thu, 2021-02-18 at 17:31 +0200, Ville Syrjälä wrote:
On Thu, Feb 18, 2021 at 10:35:05AM +0200, Jani Nikula wrote:
On Fri, 12 Feb 2021, Lyude Paul lyude@redhat.com wrote:
I think it wouldn't be a bad idea to just address this with a followup series instead and use the old DRM_DEBUG_* macros in the mean time.
aux->dev is there, could also use dev_dbg et al. in the mean time. They handle NULL dev gracefully too if the driver didn't set that.
Last I looked aux->dev was random. Some drivers point it at the connector vs. some at the the pci/platform device.
That's correct-for most SoCs the AUX channel is actually a standalone platform device that isn't associated with the DRM device by default. /But/ I went through the tree yesterday and the day before and did a bunch of cleanup around DP aux registration, added a drm_dev field and hooked it up in every driver with an aux channel, and then converted all of the DP helpers (including dual mode and MST) over to using drm_dbg_*() variants. Once I've gotten through reading all my email for today I'm going to do a quick sanity check on it and then post the series to dri-devel.