|Intel GMA3600 - The Grapevine
Date: November 28, 2013, 10:39:04 PM
Here is the URL for a short thread found under the heading of issues for the xf86-video-gma500 github project:
The developer's comments are quite interesting and do suggest that there will soon be more to come.
Also, the developer is maintaining a separate drm-gma500 github project which he apparently uses for testing drm patches/changes. The project can be found at: https://github.com/patjak/drm-gma500. Some new blitter code for the gma500 can be found there. The change dates on the files in this project will give you a sense of how long it takes for changes to make their way into the kernel - approximately half of the changes found here are just now about to appear in linux-3.15, despite having been made several months ago.
A caveat on the linux 3.15 updates: The new blitter code is minimal. In fact, only the idle function will be provided for now. We need to see some serious progress here before the developer can fill in the missing functionality of the xf86-video-gma500 xorg driver.
Latest for May:
2014-04-02 drm: Replace crtc fb with primary plane fb (v3)
The following log entries represent changes for linux-3.15-rc1:
2014-03-28 drm/gma500: Replace DRM_LOG_KMS() by DRM_DEBUG_KMS() Lespiau, Damien 1 -10/+10
2014-03-23 drm/gma500: add locking to fixed panel edid probing Daniel Vetter 3 -0/+11
2014-03-20 ACPI / gpu / drm: Stop selecting VIDEO_OUTPUT_CONTROL Jean Delvare 1 -1/+0
2014-03-17 drm/gma500: remove stub .open/postclose David Herrmann 1 -11/+0
2014-03-17 drm/gma500: Code cleanup - inline documentation Arthur Borsboom 2 -106/+54
2014-03-17 drm/gma500: Code cleanup - style fixes Arthur Borsboom 2 -80/+76
2014-03-17 drm/gma500: Code cleanup - removal of centralized exiting of function Arthur Borsboom 1 -6/+1
2014-03-17 drm/gma500/cdv: Cedarview display cleanups Patrik Jakobsson 1 -52/+19
2014-03-17 drm/gma500: Unify encoder mode fixup Patrik Jakobsson 5 -24/+13
2014-03-17 drm/gma500: Unify _get_core_freq for cdv and psb Patrik Jakobsson 5 -78/+82
2014-03-17 drm/gma500: Move asle interrupt work into a work task Patrik Jakobsson 2 -4/+22
2014-03-17 drm/gma500: Remove dead code Thierry Reding 1 -3/+0
2014-03-17 drm/gma500: Add backing type and base align to psb_gem_create() Patrik Jakobsson 6 -9/+33
2014-03-17 drm/gma500: Remove unused ioctls Patrik Jakobsson 4 -269/+0
2014-03-17 drm/gma500: Always trap MMU page faults Patrik Jakobsson 1 -7/+1
2014-03-17 drm/gma500: Hook up the MMU Patrik Jakobsson 2 -13/+55
2014-03-17 drm/gma500: Add first piece of blitter code Patrik Jakobsson 3 -0/+74
2014-03-17 drm/gma500: Give MMU code it's own header file Patrik Jakobsson 3 -88/+95
2014-03-17 drm/gma500: Add support for SGX interrupts Patrik Jakobsson 1 -12/+69
2014-03-17 drm/gma500: Make SGX MMU driver actually do something Patrik Jakobsson 3 -134/+138
2014-02-27 x86, cpufeature: Rename X86_FEATURE_CLFLSH to X86_FEATURE_CLFLUSH H. Peter Anvin 1 -1/+1
I also noted changes to power.c, made May 7 and obviously not included in the current 3.15-rc1 merge:
2013-04-07 drm/gma500: Add hooks for hibernation Patrik Jakobsson 1 -0/+15
2013-04-07 drm/gma500: Activate the gtt rebuild on suspend/resume Patrik Jakobsson 1 -0/+2
On a separate note, there have still been no changes to xf86-video-gma500 for the past 2 months.
Apparently, work has restarted on the xf86-video-gma500 source tree over the past 30 days. This is a great sign that true gpu-based 2D acceleration may be just around the corner.
Here's the URL to the github for xf86-video-gma500 for those who are interested:
Just some personal observations:
1) during the boot cycle, when X11 is started, gma500 screens had a band of noise (at least 50 lines wide) that appeared at the top of the screen until X11 finished its initialisation. This appears to have been fixed - no more band of noise.
2) there has obviously been progress with LVDS identification, although I want to emphasize "progress" - it's still not entirely out of the woods. I can now boot without the "video-LVDS-1:d" kernel param, BUT, the resulting screen has the same weird problem with vsync that I've seen on so many other distros that claim to support the gma500 family of graphics chipsets. I'm hoping that the gma500 LVDS fixes slated for release along with linux-3.13 go the rest of the way in fixing this issue (and also that they don;t screw up what's working).
Linux 3.13-rc1 first merge. New things coming to the gma500 driver.
Patrik Jakobsson (12):
drm/gma500: Add Minnowboard to the IS_MRST() macro
drm/gma500: Add chip specific sdvo masks
drm/gma500: Add support for aux pci vdc device
drm/gma500: Add aux device support for gmbus
drm/gma500/mrst: Add SDVO clock calculation
drm/gma500/mrst: Add aux register writes when programming pipe
drm/gma500/mrst: Properly route oaktrail hdmi hooks
drm/gma500/mrst: Add aux register writes to SDVO
drm/gma500/mrst: Replace WMs and chickenbits with values from EMGD
drm/gma500/mrst: Setup GMBUS for oaktrail/mrst
drm/gma500/mrst: Don't blindly guess a mode for LVDS - (https://lkml.org/lkml/2013/11/14/262)
drm/gma500/mrst: Add SDVO to output init
It looks like there will be some changes coming to the gma500_gfx kernel driver - merges into v3.12 have already started:
The developer, Patrik Jakobsson, has already inferred that EXA 2D acceleration for the gma500 driver will be introduced in Linux 3.12:
Looking forward to seeing some decent progress on this driver.
Linux 3.12-rc1 has been released, so we can probably expect to see a full release within the next 2 months.
Rating: This article has not been rated yet.