On Mon, Jul 13, 2020 at 08:26:14PM +0200, Yves-Alexis Perez wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA256 > > On Mon, 2020-07-13 at 10:33 +0800, Aaron Lu wrote: > > On Tue, May 12, 2020 at 11:13:52AM +0200, Yves-Alexis Perez wrote: > > > I've also re-pinged upstream Linux and the patch should be included in the > > > next Linux point release (hopefully 4.19.123), and which will then be > > > included > > > in Debian buster for the next Debian point release (hopefully 10.5). It's > > > not > > > tomorrow but still it should help at one point. > > > > Just want to report that I have installed 5.6.0-0.bpo.2-amd64 from > > buster-backport and now the display can be powered back on when I > > pressed my keyboard or moved my mouse. > > Could you try linux-image-4.19.0-10-amd64 from testing-proposed-updates and > report back? It should include the fix.
It doesn't seem easy to install packages from testing-proposed-updates. https://wiki.debian.org/TestingProposedUpdates doesn't tell me how to do that and after adding: deb http://some_debian_mirror/debian/ testing-proposed-updates main 'apt-cache search linux-image' doesn't show me linux-image-4.19.0-10-amd64.. Anyway, I downloaded this package directly from the mirror's pool: linux-image-4.19.0-10-amd64-unsigned_4.19.131-2_amd64.deb Is the above package correct? And the good news is, this kernel also works fine :-) > > I do not see any other problems right now, everything seems to work > > fine, except one error message in dmesg everytime the display goes the > > off-on cycle: > > broken atomic modeset userspace detected, disabling atomic > > But I guess it's a different problem and it doesn't seem to cause any > > problem. > > Actually the problem lies in the way Xorg uses atomic, and that's why the fix > is to disable them. So if you see the message that means the fix is correctly > applied. I see, thanks for the explanation.