summaryrefslogtreecommitdiff
path: root/xc/programs/Xserver/hw/xfree86/drivers/glint/README.pm3
blob: c4db56f6b4de56075614d63a5d5a31f9e8e5b8b4 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
/* $XFree86: xc/programs/Xserver/hw/xfree86/drivers/glint/README.pm3,v 1.11 2001/05/08 19:31:22 alanh Exp $ */
  
STATUS as of Tue,  8 May 2001 19:01:39 +0200
  
Working :
  * Unaccelerated : Depth 8, 15, 16 and 24 are ok. I tested them upto
    1280x1024 only as my monitor don't support higher modes.
  * Endianess clean, thanks to Romain Dolbeau <dolbeaur@club-internet.fr> for
    helping me debug this.
  * HW cursor.
  * ShadowFb.
  * Full 2D Accels.
    - Sync.
    - Pixmap cache.
    - Offscreen pixmaps.
    - Clipping.
    - Screen to screen copy. 
    - Solid fills.
    - HorVert Solid Lines .
    - 8x8 Mono Pattern Fills.
    - Color Expansion Fills.
    - Images Writes.
    - Bitmap Writes using direct FIFO writes with or without FIFO Disconnect.
    - Pixmap Writes using direct FIFO writes with or without FIFO Disconnect.
  * Appian J2000 second head initialization.
  * Dual head : The console fonts will get trashed, but dual head works.
  * Xv : Hardware video scaler :
    - Needs checking on a big endian machine.
    - Needs acceleration to work.
    - Clipping supported trough the alpha channel in depth 15 and 24
      and with an overlay key color in depth 8 and 16.
    - Support both dual head and single head, trough gamma or permedia3.
    - Attributes are :
      - FILTER : None, Partial (X only) or Full filtering.
      - MIRROR : X and/or Y Axis mirroring.
      - ALPHA :
        - 0 -> FB Only
        - 1 -> 25% Video, 75% FB
        - 2 -> 75% Video, 25% FB
        - 3 -> Video Only
      - [TODO] VIDEOKEY : Overlay Key Color for clipping in depth 8 and 16.
  * DRI : Work is underway.

Not Working :

  * [NOT POSSIBLE] 2D Accel :
    - Solid & Dashed Lines are not possible on glint hardware.
    - 8x8 Color Pattern Fill is almost never used.
    - Clipping needs cleaning up.
  * [TODO] 2D Accel hooks for the Rendering extension :
    - Render extension initialization.
    - CPUToScreenTexture.
    - CPUToScreenAlphaTexture.
  * [NEED FIX] Xv has still some minor problems :
    - depth 8 does not work, but then it may be the app i am using.
    - when using 2D accels there is some unstability in the video display. I
      guess this is because there is then not enough bandwith to do the
      copying of the data in time. After a time of the above, the images
      becomes black. It will come back once stopvideo is called.
      [FIX] i will disable VideoOverlay each 25 frames, this causes a flicker,
      but at least it will bing the image back.
  * [WORK IS UNDERWAY] DRI/OpenGL (only together with a gamma chip) :
    - DRI support should be ok, but accelerated openGL is not yet supported. 
    - The accelerated OpenGL library supposes we are using a gamma together
      with a MX rasterizer, not a Permedia3 one.

Known problems :

  * Console gets broken when using dual headed mode. The culprit seems to be
    the vga graphics index & port (GraphicsIndexReg : 0x63ce, GraphicsPort :
    0x3cf). I had to use IO acces for both these two, because if not, console
    fonts would get trashed even in mono headed mode.
    [FIX] Well, if you really need the console and are running linux, just use
    vesafb, it will be a bit slower, but the fonts will no more become
    corrupt. Maybe i will even try writting a specific fbdev for the pm3,
    which will be much faster and have more functionality.
    [FIX2] try : consolechars -d, it should reload the console fonts.
  * [FIXED] R-B inversion. Sometimes, all of a sudden, it seems as the
    RGB order gets changed. going to console and back solves this.
    Well, this is partly fixed, but still appears under very heavy load.
  * [FIXED] When moving a window around a lot quickly, the video outputs dies.
    Well, this is partly fixed, but still appears under very heavy load.
 => [NOTE] If this two still happens, try disabling the Hardware cursor, with
    the "SWCursor" option to your device XF86Config section.
  * Sometimes there are blue transparent garbage in the form of small
    horizontal bands, few pixels high, and more pixels width, maybe 64pixels ?
    This may be a hardware bug.

Sven Luther <luther@dpt-info.u-strasbg.fr>