[Bug 110303] [regression, bisected] vt console display doesn't update with psr

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=110303

--- Comment #4 from Dhinakaran Pandiyan <dhinakaran.pandiyan at intel.com> ---
(In reply to Carl Michal from comment #0)
> On a Dell XPS13 9343, with kernels newer than 4.17, a vt switch
> (CTRL-ALT-Fn) to a console vt has become unusable if psr is enabled. Any
> console output only appears after a long delay (or in more recent kernels
> only after a vt switch away and back).
> 
> I bisected to this commit:
> 75cbec033c08f6d41c4775784f66ab860d02a6b5
> 
> In current drm-tip the problem is still there. If I comment out setting
> EDP_PSR_DEBUG_MASK_DISP_REG_WRITE in intel_psr.c
> then the vt behaves as expected.
> 
The reason we made that change was to not cause PSR exit on any display
register write, we want PSR exits to be predictable. If this issue only occurs
on BDW, it looks like the workaround we have in intel_psr_flush() isn't working
on BDW or for some reason console updates aren't calling intel_psr_flush().


> This is a Broadwell machine:
> #lspci:
> 00:02.0 VGA compatible controller: Intel Corporation HD Graphics 5500 (rev
> 09)
> 
> I realize that psr isn't exactly supported on BDW - though I've been using
> it for years and have never seen the lockups that people complained about
> that led to its being turned off by default.

I have seen a couple of other bug reports on BDW too (flip timeouts). So, we
won't be able to enable PSR on BDW by default until we fix all known issues.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20190412/32847278/attachment.html>