[07:42] <RAOF> mlankhorst: Hey, have you got drm_kms_helper.edid=$FOO:edid/<thing>.bin to work with i915? I can't seem to get it to override the broken EDID on my monitor that's causing i915 to pretend there's nothing there.
[07:44] <RAOF> Specifically, throwing “drm_kms_helper.edid_firmware=HDMI2:edid/1280x1024.bin” on the kernel cmd line (or HDMIA2, or HDMI-A-2, or whatever) doesn't result in i915 doing anything different. If I drop the HDMI2: bit i915 applies that EDID to my eDP and DP outputs, as expected, but _still_ not to HDMI2.
[07:46] <mlankhorst> hm funstuff :P
[07:46] <mlankhorst> I just overrode all my EDIDs, I have 2 identical monitors ;-)
[07:46] <mlankhorst> lets see..
[07:48] <RAOF> For those playing at home, http://paste.ubuntu.com/7371321/ is what i915 says about my poor old LCD. :)
[07:53] <mlankhorst> should be HDMI-A-2 then
[07:54] <mlankhorst> that's the connector name afaict
[08:09] <RAOF> Tried HDMI-A-2 as well :(
[08:09]  * RAOF → EOD though.
[08:11] <mlankhorst> what happens then?
[10:22] <RAOF> mlankhorst: Same EDID checksum spew from i915, no monitor for RAOF.
[10:23] <mlankhorst> odd, that should be the magic invocation if dmesg can be trusted :P
[10:23] <RAOF> It's also possible that i915 won't load the EDID over a present, but broken one :/
[10:24] <RAOF> This worked for radeon, though.
[10:24] <mlankhorst> might be
[10:24] <mlankhorst> but since it uses strncmp, shouldn't HDMI:edid/1280x1024.bin work?
[10:26] <mlankhorst> though that'd probably make it apply to all hdmi :P
[10:26] <RAOF> Well, I've only got one HDMI port, so...
[10:26] <RAOF> I've no idea why i915 thinks I've got two.
[10:27] <mlankhorst> hm that should fix my 'optimus' config too :P
[14:19] <pkern> RAOF: You're not using a DP adapter or the like, right? :)
[23:15] <RAOF> pkern: I *am* using a HDMI→DVI adaptor.
[23:15] <RAOF> pkern: But that's obviously working, because the BIOS can bring up the display and GRUB will draw correctly to it.