Hi Andrea, On 11 December 2016 at 22:48, Andrea Venturi <[email protected]> wrote:
> hello, > > this mail just to report an experience about using the DRM driver on sun7i > (Olimex A20 SOM EVB). let me start thanking all the guys doing the > "heavylifting" for this feature/driver. > > *TL;DR* i'll be very synthetic, it starts to work (you can see a picture, > then some minutes later screen shuts off) albeit it's still a hard to gain > feature as lot's of info is scattered here and there or implicit, that's > why i try to recap here the simpler recipe to "see the light".. > > first i've got the* 4.9rc7 kernel* runnng from sunxi-next git tree: > > https://github.com/linux-sunxi/linux-sunxi/commits/sunxi-next > > i compiled with the default defconfig for this Olimex A20 SOM EVB board > and booted. > > i've got *simplefb* working; as the FB was working since uboot phase, > thanx to this "modeline": > > CONFIG_VIDEO_LCD_MODE="x:800,y:480,depth:24,pclk_khz:33000, > le:45,ri:209,up:22,lo:22,hs:1,vs:1,sync:3,vmode:0" > > then i added in DTS the extended lines as from this commit by "net147" > (thanx): > > https://github.com/net147/linux/commit/cd93810d8a565d2348e27756f4bb27 > 62dc6cc79d > > BTW i added in *drivers/gpu/drm/panel/panel-simple.c* the setup of my > panel (a noname 800x480 7inch working in uboot), and put accordingly the > panel's compatible entry in place of the *lcd-olinuxino-43-ts* in DTS: > ... > panel: panel { > compatible = "olimex,lcd-olinuxino-43-ts"; > ... > > then i had to patch kernel for "relaxing" the panel *clock validation > step* with this very recent "hack" for the clock matching: > > http://www.spinics.net/lists/kernel/msg2390747.html > > and this is more or less everything to see the DRM LCD working on A20 > > in *dmesg* the relevant lines are: > > --- > [ 1.082842] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). > > [ 1.082846] [drm] No driver support for vblank timestamp query. > > [ 1.083328] sun4i-drm display-engine: bound 1e60000.display-backend > (ops 0xc063d87c) > [ 1.083841] sun4i-drm display-engine: bound 1c0c000.lcd-controller (ops > 0xc063d410) > [ 1.083886] checking generic (7fe89000 177000) vs hw (0 ffffffff) > > [ 1.083890] fb: switching to sun4i-drm-fb from simple > > [ 1.137201 <13%2072%2001>] Console: switching to colour dummy device > 80x30 > > and this is all fantastic. i could display fancy images with *fbv* (i'm > working on straight framebuffer, no X), thanks to all those involved. > > the only issue is that, after few minutes (10 or so), the display switch > to this kind of rendering: > > https://drive.google.com/open?id=0B9QRW6Oy0GmeOGdvM3YyYUtqZHM > > no messages on syslog or console. i'm wondering which kind of "glitch" > could bring to this (i suppose some kind of "clock power down"?). > Console blank is 600 seconds by default - kernel will turn off the console after 10 minutes of inactivity. Try adding consoleblank=0 to your kernel command line. It is powering down the display engine without powering off the LCD. > maybe setting up some *DRM debug level,* could bring to helpful messages. > i'll check how to enable those, if any has suggestions, are welcome. > > later, i'd like to test the mali binary libraries as net147 tells they > works too (i'll test ES GL on plain framebuffer and not X) . > > i see this kernel open sorce low level driver compiles (adding some setup > on *build.sh*): > https://github.com/mripard/sunxi-mali > but i'm stopping on adding the relevant entry in the DTS. and this is the > topic for another report, eventually. > > Andrea > > Regards, Jonathan -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
