Debian 9 (Kernel

i6700k (Skylake)


I get a strange error in dmesg that reads:

[    3.683213] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up
[    3.725006] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up
[    3.766802] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up
[    3.808607] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up
[    3.850408] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up
[    3.892210] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up
[    3.933994] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up
[    3.934444] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to train DP, aborting

Someone in a previous question I asked mentioned this could be related to the kernel, and suggested I update it. I don't just want to jump to a kernel update as I sort of worry I might introduce additional instabilities into my system (before I already had to add the firmware-misc-nonfree so that I actually had the i915 drivers), so first I want to try to troubleshoot this issue. I would appreciate if anyone with an answer could give me a reason for why I want to do something as I am trying to learn as much as possible.

Additionally, I don't currently see any problems stemming from this other than that it outputs in dmesg, but I don't really know if this is something I should worry about or not. I tend to like to avoid errors when possible but if it's really harmless (I only see it when I check for errors on dmesg), then maybe I should just let it go.

  • Regardless of this particular problem, you really should update your kernel; 4.9 is the supported version for Stretch, and has received tons of bug fixes (including at least 50 security fixes, many for vulnerabilities present in 4.8). May 5, 2017 at 5:01
  • Absolutely agree with Stephen - especially with newer hardware it behooves one to use the latest stable kernel with the latest DRM firmware. There have been many such updates to the 4.9.x kernel.
    – SD Allen
    May 5, 2017 at 14:02
  • Regarding this, should I worry that (current) kernel is unsigned? Other than that I have to have secure boot off in the BIOS?
    – Aaron A.
    May 5, 2017 at 18:50
  • @Aaron No, having an unsigned kernel is fine. (You should be able to install linux-image-4.9.0-2-amd64 though, which is signed. Note that all Debian packages are signed anyway, the signed v. unsigned distinction for the kernel is different.) Debian is unfortunately not quite ready for secure boot yet. May 7, 2017 at 21:31

1 Answer 1


As noted by Stephen and Allen, once I upgraded to kernel the issue no longer appears in my dmesg.

  • Feel free to accept your own answer using the big check mark, that will indicate that your question has been answered. May 7, 2017 at 21:29

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .