From d62b568caac667c42f3dca6e0335aaf4a2e9c083 Mon Sep 17 00:00:00 2001 From: P. J. McDermott Date: Sun, 20 Sep 2015 07:47:18 -0400 Subject: docs/hcl/gm45_lcd.html: Describe solution to LVDS channels problem --- diff --git a/docs/hcl/gm45_lcd.html b/docs/hcl/gm45_lcd.html index 137bc68..29b31ec 100644 --- a/docs/hcl/gm45_lcd.html +++ b/docs/hcl/gm45_lcd.html @@ -64,6 +64,43 @@ lvds_channel_mode:Specify LVDS channel mode (0=probe BIOS [default], 1=single-channel, 2=dual-channel) (int) - from /sbin/modinfo i915

+

+ Proper solution to LVDS second channel enablement problem +

+

+ The definition of register "gfx.lvds_dual_channel" = "0" in + coreboot/src/mainboard/lenovo/t400/devicetree.cb in libreboot affects the + int lvds_dual_channel member of a struct i915_gpu_controller_info + (defined in coreboot/src/drivers/intel/gma/i915.h) named gfx in a + struct northbridge_intel_gm45_config (defined in + coreboot/src/northbridge/intel/gm45/chip.h). That + gfx.lvds_dual_channel in that struct northbridge_intel_gm45_config + affects various registers set in intel_gma_init() in + coreboot/src/northbridge/intel/gm45/gma.c. +

+

+ A proper solution would likely be as follows: +

+ +

+ Note that this doesn't completely solve the problem for those dual-channel + displays and that GRUB still won't have graphics output. More work is still + needed to investigate why coreboot's graphics initialization code for GM45 + doesn't properly handle dual-channel displays (but the code for Sandy/Ivy Bridge + apparently does, as does Linux's i915 driver) and resolve that. +

-- cgit v0.9.1