Hi Adam, (CC'ing Jyri and Tomi) I'm sorry for the very late reply, your e-mail got burried in my inbox. On Wed, Aug 21, 2019 at 09:39:27PM -0500, Adam Ford wrote: > On Wed, Aug 21, 2019 at 9:08 PM Laurent Pinchart wrote: > > On Wed, Aug 21, 2019 at 08:14:43PM -0500, Adam Ford wrote: > > > I know it's been nearly 9 months this this was removed, but for those > > > of us who still define our displays in the device tree expecting the > > > dpi-panel, we're not getting video. > > > > > > The commit message only states: > > > > > > Panels are now supported through the drm_panel infrastructure, remove > > > the omapdrm-specific driver. > > > > > > It does not give examples of how to do this, and I feel like we should > > > have been given some warning or indication. Is there an example I can > > > follow for linking a dpi panel into the omap DSS? > > > > Sorry to have left you with non-working systems :-( > > I have really only been following the omap-linux mailing list and only > really focus on the LTS kernels because my employer uses the LTS > kernels as the basis for the their linux distributions. If there is a > different mailing list I should follow, let me know. I just wish it > would have been marked as deprecated or something before just being > killed. > > > If the panel is supported by a mainline DRM panel driver the change > > should be transparent (provided of course that the driver is compiled in > > the kernel or as a module). Most panels are supported by the > > panel-simple driver (CONFIG_DRM_PANEL_SIMPLE), with a few dozen of other > > panels supported by dedicated drivers (in drivers/gpu/drm/panel/) > > > > Could you point me to the DT sources of one (or all) of the affected > > systems ? > > Sure, > The same panel is used on these these two boards: > logicpd-som-lv-baseboard.dtsi > logicpd-torpedo-37xx-devkit-28.dts > > A second panel is used on: > logicpd-torpedo-37xx-devkit.dts which has the LCD timings defined in > logicpd-torpedo-baseboard.dtsi > The am3517-evm also uses the same timings, but the gpio enables are different. I see. The omap3-ha-lcd.dts is also affected. There are only two ways to fix this as far as I can tell: - Add a panel driver matching against the panel-dpi compatible string, and parsing the panel timings from DT there. Thierry Reding, the DRM panel maintainer, has rejected this option multiple times (but it seems people are still trying). - Add the exact panel model to the above DT files (see omap3-thunder.dts for an example), and make sure we have a kernel driver for those panels (possibly extending the panel-simple driver). It seems that we are also missing support in panel drivers for innolux,at070tn83, osddisplays,osd057T0559-34ts, samsung,lte430wq-f0c and startek,startek-kd050c. Tomi, which approach do you think is best at this point ? > The da850-evm uses the same panel as the am3517-evm, but it's not > using the same video driver, and I haven't had a chance to see if that > driver still exists or not. The tilcdc driver is still present in the kernel (I'm not aware of a plan to remove it), and uses a similar panel support hack as the omapdrm driver. I think it would make sense to fix it at some point, but I have no plan to do so myself at the moment. > Thanks for your quick response. > :-) Looks like I did a way worse job this time. Sorry again. > Sorry if my e-mail came across angrily, it wasn't my intention. No worries, and in any case you had reasons to be annoyed. -- Regards, Laurent Pinchart
