"drm/i915/gen8+: Add RC6 CTX corruption WA" (d4360736a7c0a6326e3bbdf7d41181f6ed03d9a6) in 5.3 stable broke RC6 on my Cofee Lake ThinkPad P52. This prevents the CPU package from entering package C-states. Happens every time, no suspend/resume needed. Since this is not going to be fixed until 5.5 (see https://www.phoronix.com/scan.php?page=news_item&px=Intel-Patches-Fast-Soft-RC6) may we have the 5.2 back in gentoo tree? The Cofeelake laptops are crappy on pre-5.0 kernels.
I'm not keen on restoring 5.2 to the tree. As it's EOL'd and unsupported upsteam. It will have none of the newest security patches and these types of patches are coming out more and more these days. Does the latest 4.19 have the same issue? You do have the option of restoring your ebuild yourself. Have you thought about doing that?
4.19 doesn't have this particular issue, however it has tons of other issues fixed in 5.0-5.2 (especially in i915 driver). I have to stay with modsetting driver in X to avoid screen corruption. It has further limitations on using the hdmi/dp connectors that are wired to the nvidia quadro card when the graphics is hybrid. I still have several 5.2 kernels installed: PF16W6Y2 ~ # eselect kernel list Available kernel symlink targets: [1] linux-4.19.10-rt8 [2] linux-4.19.66-gentoo [3] linux-4.19.67-gentoo [4] linux-4.19.68-gentoo [5] linux-4.19.84-gentoo [6] linux-5.1.21-gentoo [7] linux-5.2.14-rt7 [8] linux-5.2.15-gentoo [9] linux-5.2.16-gentoo [10] linux-5.2.17-gentoo [11] linux-5.2.20-gentoo * [12] linux-5.3.11-gentoo
These patches were supposed to land in 5.5, as indicated. Can you test the latest 5.5.X, which is 5.5.10 as of this writing
| GPU | | | | Powered On 10,2% | | RC6 89,8% | | RC6p 0,0% | | RC6pp 0,0% | I think this bug can be closed as resolved. anton@PF16W6Y2 ~ $ uname -a Linux PF16W6Y2 5.5.10-gentoo #1 SMP PREEMPT Thu Mar 19 12:37:42 MSK 2020 x86_64 Intel(R) Core(TM) i7-8850H CPU @ 2.60GHz GenuineIntel GNU/Linux