Announcement

Collapse
No announcement yet.

Radeon PM change artifacts?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Radeon PM change artifacts?

    Hi,

    Is anyone else experiencing flickering during PM change? Running 2.6.34-rc3-git5 kernel and X Server 1.7.6 with DRM, Mesa, and xf86-video-ati from git. Desktop XFCE & Compiz. Hardware is Macbook Pro with RV530/X1600. Seeing the following messages:
    Code:
    [drm] Requested: e: 31300 m: 30000 p: 16
    [drm] Setting: e: 31300 m: 30000 p: 16
    [drm] not in vbl for pm change 00020002 00000000 at entry
    [drm] not in vbl for pm change 00010002 00000000 at exit
    Current kernel code cannot determine when display is in VBLANK? And is PM actually doing anything with my chipset? dmesg has the same entry over and over again.

  • #2
    Originally posted by amphigory View Post
    Hi,
    Is anyone else experiencing flickering during PM change?
    I see this also - it's a known issue AFAIK and there is newer code waiting to get into kernel.

    Comment


    • #3
      I'm seeing this too with 2.6.34.2 and my RV790.

      These lines are repeated over and over again, once I set the "dynpm=1 dynclks=1" options on the radeon module:

      Code:
      [drm] Requested: e: 80000 m: 97500 p: 16
      [drm] Setting: e: 80000 m: 97500 p: 16
      [drm] not in vbl for pm change 00020002 00000000 at entry
      [drm] Requested: e: 75000 m: 85000 p: 16
      [drm] Setting: e: 75000 m: 85000 p: 16
      [drm] not in vbl for pm change 00020002 00000000 at entry
      Dynamic power management does seem to be enabled:
      Code:
      [drm] initializing kernel modesetting (RV770 0x1002:0x9460).
      [drm] register mmio base: 0xFBCE0000
      [drm] register mmio size: 65536
      ATOM BIOS: RV790
      [drm] Clocks initialized !
      [drm] Internal thermal controller with fan control
      [drm] 5 Power State(s)
      [drm] State 0 Default (default)
      [drm]   16 PCIE Lanes
      [drm]   3 Clock Mode(s)
      [drm]           0 engine/memory: 750000/850000
      [drm]           1 engine/memory: 750000/850000
      [drm]           2 engine/memory: 750000/850000
      [drm] State 1 Performance 
      [drm]   16 PCIE Lanes
      [drm]   3 Clock Mode(s)
      [drm]           0 engine/memory: 500000/975000
      [drm]           1 engine/memory: 800000/975000
      [drm]           2 engine/memory: 850000/975000
      [drm] State 2 Default 
      [drm]   16 PCIE Lanes
      [drm]   3 Clock Mode(s)
      [drm]           0 engine/memory: 800000/975000
      [drm]           1 engine/memory: 800000/975000
      [drm]           2 engine/memory: 800000/975000
      [drm] State 3 Performance 
      [drm]   16 PCIE Lanes
      [drm]   3 Clock Mode(s)
      [drm]           0 engine/memory: 600000/975000
      [drm]           1 engine/memory: 800000/975000
      [drm]           2 engine/memory: 850000/975000
      [drm] State 4 Default 
      [drm]   16 PCIE Lanes
      [drm]   3 Clock Mode(s)
      [drm]           0 engine/memory: 240000/975000
      [drm]           1 engine/memory: 240000/975000
      [drm]           2 engine/memory: 240000/975000
      [drm] radeon: dynamic power management enabled
      [drm] radeon: power management initialized
      There is no obvious flickering. But the dmesg spam is annoying.

      Comment


      • #4
        Try a newer kernel. The pm code has changed significantly in 2.6.35.

        Comment


        • #5
          Awaiting the start of the 2.6.35.x series, then.

          Originally posted by agd5f View Post
          Try a newer kernel. The pm code has changed significantly in 2.6.35.
          Hopefully, the 2.6.35.1 kernel will materialise sooner than 2.6.34.1 did!

          On a side note, I also upgraded my old AGPx8 RV350 machine from 2.6.33.6 to 2.6.34.2 last night, and was presented with only a single Power State in the dmesg log. From which I would infer that such old hardware has no "power management" potential at all. Would this be a correct assumption please? Or are other power savings still available?

          Comment

          Working...
          X