Comment 22 for bug 1774845

Zakhar (alainb06) wrote :

Dear Christopher,

as I posted "upstream" (https://bugs.freedesktop.org/show_bug.cgi?id=93782), both

- Latest ubuntu drm-tip kernel (as of June 3rd)
- applying the patch suggested C#12 of the upstream post (on the source I already had of 4.17 rc3 drm-tip)

were NOT successful fixing the bug.

If you look at C#11 of my initial post upstream (https://bugs.freedesktop.org/show_bug.cgi?id=106422#c11), Jim Rees implies that bissecting has already been done and that it boils down to commit: ea0000f0 "Roll out the helper nonblock tracking"

This commit is supposed to be "too difficult" to revert starting from kernel 4.14 (probably due to major rewrites of parts of the drm stack).

You are right, things can get worse... and they did: indeed the "bug" was only spitting some messages in the system log prior to the change, and is now triggering several time outs.
But for the moment, I'll have to take my chances that this tv-out issue doesn't get worse, since I really don't have time to do a lot more tests.

Also, I don't have permanent access to this hardware (it's not mine), and that doesn't help!