This is probably related to bug 51739 and bug 53393. On my Lenovo X220s (Sandy bridge, 00:02.0 VGA compatible controller [0300]: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller [8086:0126] (rev 09)), I often end up in full power mode, even when idle at the gnome 3 desktop. I've seen this on Fedora 17 with various 3.6 and 3.7 kernels. i915_enable_rc6=7 has no effect. When the problem is happening, i915_drpc_info looks like: RC information accurate: yes Video Turbo Mode: yes HW control enabled: yes SW control enabled: no RC1e Enabled: no RC6 Enabled: yes Deep RC6 Enabled: no Deepest RC6 Enabled: no Current RC state: on Core Power Down: no RC6 "Locked to RPn" residency since boot: 0 RC6 residency since boot: 121924 RC6+ residency since boot: 0 RC6++ residency since boot: 0 sysfs_rc6_residency fails. The RC6 residency since boot: 121924 number doesn't change at all, so presumably the chip is just stuck at full power instead of flipping in and out of rc6. The only strange kernel options I have are pcie_aspm=force pcie_aspm.policy=powersave, which are mostly a relic from when the kernel misinterpreted the BIOS lack of support for OS ASPM control.
Can you please retest with latest drm-intel-fixes?
Sorry for the absurd delay. 3.8.1 has the same problem. Is there a git tree I should try?
3.9-rc1 has pretty much all drm/i915 patches, so retesting that might be worth a shot. But tbh we haven't really gained any clue at all as to why rc6 doesn't work too well on snb sometimes :(
Is it worth trying to detect the problem and resetting something (the RC6 registers? toggling between SW and HW control?) when it happens?
*** Bug 57677 has been marked as a duplicate of this bug. ***
When this happens, can you try: $ intel_reg_write 0xa094 0xf0000 and then check the RC6 residency counters again? If idle, that's *supposed* to force us into RC6.
[luto@antithesis tools]$ sudo cat /sys/kernel/debug/dri/0/i915_drpc_info RC information accurate: yes Video Turbo Mode: yes HW control enabled: yes SW control enabled: no RC1e Enabled: no RC6 Enabled: yes Deep RC6 Enabled: no Deepest RC6 Enabled: no Current RC state: on Core Power Down: no RC6 "Locked to RPn" residency since boot: 0 RC6 residency since boot: 115522 RC6+ residency since boot: 0 RC6++ residency since boot: 0 RC6 voltage: 450mV RC6+ voltage: 245mV RC6++ voltage: 245mV [luto@antithesis tools]$ sudo intel_reg_read 0xa094 Couldn't map MMIO region: No such file or directory [luto@antithesis tools]$ ^Cdo intel_reg_read 0xa094 (reverse-i-search)`': ^C [luto@antithesis tools]$ sudo intel_reg_write 0xa094 0xf0000 Couldn't map MMIO region: No such file or directory [luto@antithesis tools]$ sudo intel_reg_write 0xa094 0xf000^C $ sudo ./intel_reg_read 0xa094 0xA094 : 0x0 $ sudo ./intel_reg_write 0xa094 0xf0000 Value before: 0x0 Value after: 0x70000 $ sudo ./intel_reg_read 0xa094 0xA094 : 0x70000 $ sudo cat /sys/kernel/debug/dri/0/i915_drpc_info RC information accurate: yes Video Turbo Mode: yes HW control enabled: yes SW control enabled: no RC1e Enabled: no RC6 Enabled: yes Deep RC6 Enabled: no Deepest RC6 Enabled: no Current RC state: on Core Power Down: no RC6 "Locked to RPn" residency since boot: 0 RC6 residency since boot: 115522 RC6+ residency since boot: 0 RC6++ residency since boot: 0 RC6 voltage: 450mV RC6+ voltage: 245mV RC6++ voltage: 245mV Are you sure that's the right register? (Can you point me to the right section of the programming manual? I tried awhile ago but I couldn't find it.)
Can you please try with this patch: https://patchwork.kernel.org/patch/2707341/ as it claims to fix some instability with rc6 on SandyBridge?
(In reply to comment #8) > Can you please try with this patch: > https://patchwork.kernel.org/patch/2707341/ as it claims to fix some > instability with rc6 on SandyBridge? Did a local build with this patch on top of fedora's kernel-3.9.5-201.fc18.x86_64 and if anything, this makes it more likely to happen on my laptop, a thinkpad X220. A possibly interesting data point is that this seems more likely to happen when I do a suspend cycle with the laptop docked.
We have a lead for rc6 issues after suspend... Hopefully that patch will be ready for stable@ shortly.
(In reply to comment #10) > We have a lead for rc6 issues after suspend... Hopefully that patch will be > ready for stable@ shortly. I.e. please test this patch and report what happens: https://patchwork.kernel.org/patch/2827634/ Thanks, Daniel
Can you please test with drm-intel-fixes, in particular commit 7dcd2677ea912573d9ed4bcd629b0023b2d11505 Author: Konstantin Khlebnikov <khlebnikov@openvz.org> Date: Wed Jul 17 10:22:58 2013 +0400 drm/i915: fix long-standing SNB regression in power consumption after resume
I ran https://patchwork.kernel.org/patch/2707341/ for a day without triggering the problem. I'm building the newer version now.
I'll declare this resolved now, please reopen if the latest patch doesn't work for you.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.