When trying to use my external monitor from a MacBook Pro (late 2011 model, 8.2), I have a couple of problems: 1. With a mini-DP -> VGA adapter, my system becomes laggy and I see the following error message in kernel [ 6414.039062] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 6414.618381] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 6415.197134] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 6415.778165] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 6416.358456] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 6416.943138] [drm:intel_dp_start_link_train] *ERROR* failed to get link status [ 6416.943144] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting 2. With a mini-DP -> HDMI adapter, xrandr and GNOME Display settings seem to pick up the EDID information from my monitor correctly, but I have no output on the external display. I'm not sure if this is the right place to report this, so if it's not, please do let me know where this should go. ---- lspci output for the graphics card: 00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09) Software: Kernel: 3.3.2 (with one patch to the i915 driver to make things actually -- https://github.com/fooblahblah/linux-mainline-efi-lvds/blob/master/lvds_dual_channel.patch) X.org server: 1.12.0 xf86-video-intel: 2.18.0 libdrm: 2.4.33
Another dongle: https://bugs.freedesktop.org/show_bug.cgi?id=49402
With the 3.3.5 and 3.4-rc6 kernels, the link train messages are gone, and I can actually see the display in GNOME Display Settings and xrandr (so the EDID information is being picked up). There's still no actual output turning up over HDMI, though (the monitor's not getting any signal at all, afaict).
(In reply to comment #2) > With the 3.3.5 and 3.4-rc6 kernels, the link train messages are gone, and I can > actually see the display in GNOME Display Settings and xrandr (so the EDID > information is being picked up). There's still no actual output turning up over > HDMI, though (the monitor's not getting any signal at all, afaict). Sorry, just noticed that's just repeating the original information.
New kernel (3.7), please test.
Tried just now with 3.8-rc7. With DP->HDMI, same situation - I get EDID, but no output. Could it be due to disabling the Radeon chip completely? With a DP->VGA, I get the link train errors and a backtrace in dmesg when I unplug: [ 580.031796] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 580.547201] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 581.127544] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 581.710798] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 582.294358] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 582.877297] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 583.457772] [drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting [ 584.225601] ------------[ cut here ]------------ [ 584.225614] WARNING: at drivers/gpu/drm/i915/intel_dp.c:1987 intel_dp_link_down+0x1c0/0x1d0() [ 584.225617] Hardware name: MacBookPro8,2 [ 584.225619] Modules linked in: b43 bcma radeon ttm rfcomm bnep uvcvideo videobuf2_vmalloc btusb videobuf2_memops videobuf2_core videodev bluetooth media snd_hda_codec_hdmi snd_hda_codec_cirrus kvm_intel kvm sdhci_pci snd_hda_intel crc32c_intel sdhci mmc_core snd_hda_codec snd_hwdep snd_pcm snd_page_alloc snd_timer snd soundcore fuse xfs exportfs nfs lockd sunrpc reiserfs btrfs libcrc32c zlib_deflate [last unloaded: b43] [ 584.225665] Pid: 4648, comm: Xorg Not tainted 3.8.0-rc7 #2 [ 584.225667] Call Trace: [ 584.225674] [<ffffffff81049cbf>] warn_slowpath_common+0x7f/0xc0 [ 584.225678] [<ffffffff81049d1a>] warn_slowpath_null+0x1a/0x20 [ 584.225681] [<ffffffff813e1300>] intel_dp_link_down+0x1c0/0x1d0 [ 584.225685] [<ffffffff813e3e81>] intel_disable_dp+0x51/0x80 [ 584.225689] [<ffffffff813d085a>] ironlake_crtc_disable+0x6da/0x830 [ 584.225693] [<ffffffff813d5766>] intel_crtc_update_dpms+0x76/0xb0 [ 584.225696] [<ffffffff813d585a>] intel_encoder_dpms+0x1a/0x30 [ 584.225700] [<ffffffff813d75dc>] intel_connector_dpms+0x3c/0x70 [ 584.225704] [<ffffffff8139cdf5>] drm_mode_obj_set_property_ioctl+0x345/0x350 [ 584.225707] [<ffffffff8139ce30>] drm_mode_connector_property_set_ioctl+0x30/0x40 [ 584.225710] [<ffffffff8138ba14>] drm_ioctl+0x474/0x530 [ 584.225714] [<ffffffff8139ce00>] ? drm_mode_obj_set_property_ioctl+0x350/0x350 [ 584.225719] [<ffffffff8116024f>] do_vfs_ioctl+0x8f/0x4f0 [ 584.225723] [<ffffffff81160700>] sys_ioctl+0x50/0x90 [ 584.225727] [<ffffffff812d71be>] ? trace_hardirqs_on_thunk+0x3a/0x3c [ 584.225730] [<ffffffff8139c5a0>] ? drm_mode_detachmode_ioctl+0x110/0x110 [ 584.225734] [<ffffffff816e20c2>] system_call_fastpath+0x16/0x1b [ 584.225736] ---[ end trace e7c6fa97df4c9cd8 ]---
The i915 is physically not connected to the external display port data lines. This is why on OSX the moment you connect an external display, the device swaps to the radeon. Sadly, auto switch doesn't work in linux.
(In reply to comment #6) > The i915 is physically not connected to the external display port data > lines. This is why on OSX the moment you connect an external display, the > device swaps to the radeon. Sadly, auto switch doesn't work in linux. That's interesting. Per comment #0 the aux channel seems to be connected as EDID reading succeeds.
Please try the drm-intel-nightly branch of git://people.freedesktop.org/~danvet/drm - we have some DP training fixes there. Attach dmesg with drm.debug=0xe module parameter. Although if comment #6 is correct, we can't fix this...
Please try a recent kernel and report back.
Timeout. Please reopen if the problem persists with recent kernels.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.