xref: /openbmc/linux/Documentation/fb/matroxfb.rst (revision d9fd5a71)
1=================
2What is matroxfb?
3=================
4
5.. [This file is cloned from VesaFB. Thanks go to Gerd Knorr]
6
7
8This is a driver for a graphic framebuffer for Matrox devices on
9Alpha, Intel and PPC boxes.
10
11Advantages:
12
13 * It provides a nice large console (128 cols + 48 lines with 1024x768)
14   without using tiny, unreadable fonts.
15 * You can run XF{68,86}_FBDev or XFree86 fbdev driver on top of /dev/fb0
16 * Most important: boot logo :-)
17
18Disadvantages:
19
20 * graphic mode is slower than text mode... but you should not notice
21   if you use same resolution as you used in textmode.
22
23
24How to use it?
25==============
26
27Switching modes is done using the video=matroxfb:vesa:... boot parameter
28or using `fbset` program.
29
30If you want, for example, enable a resolution of 1280x1024x24bpp you should
31pass to the kernel this command line: "video=matroxfb:vesa:0x1BB".
32
33You should compile in both vgacon (to boot if you remove you Matrox from
34box) and matroxfb (for graphics mode). You should not compile-in vesafb
35unless you have primary display on non-Matrox VBE2.0 device (see
36Documentation/fb/vesafb.rst for details).
37
38Currently supported video modes are (through vesa:... interface, PowerMac
39has [as addon] compatibility code):
40
41
42Graphic modes
43-------------
44
45===  =======  =======  =======  =======  =======
46bpp  640x400  640x480  768x576  800x600  960x720
47===  =======  =======  =======  =======  =======
48  4             0x12             0x102
49  8   0x100    0x101    0x180    0x103    0x188
50 15            0x110    0x181    0x113    0x189
51 16            0x111    0x182    0x114    0x18A
52 24            0x1B2    0x184    0x1B5    0x18C
53 32            0x112    0x183    0x115    0x18B
54===  =======  =======  =======  =======  =======
55
56
57Graphic modes (continued)
58-------------------------
59
60===  ======== ======== ========= ========= =========
61bpp  1024x768 1152x864 1280x1024 1408x1056 1600x1200
62===  ======== ======== ========= ========= =========
63  4    0x104             0x106
64  8    0x105    0x190    0x107     0x198     0x11C
65 15    0x116    0x191    0x119     0x199     0x11D
66 16    0x117    0x192    0x11A     0x19A     0x11E
67 24    0x1B8    0x194    0x1BB     0x19C     0x1BF
68 32    0x118    0x193    0x11B     0x19B
69===  ======== ======== ========= ========= =========
70
71
72Text modes
73----------
74
75==== =======  =======  ========  ========  ========
76text 640x400  640x480  1056x344  1056x400  1056x480
77==== =======  =======  ========  ========  ========
78 8x8   0x1C0    0x108     0x10A     0x10B     0x10C
798x16 2, 3, 7                        0x109
80==== =======  =======  ========  ========  ========
81
82You can enter these number either hexadecimal (leading `0x`) or decimal
83(0x100 = 256). You can also use value + 512 to achieve compatibility
84with your old number passed to vesafb.
85
86Non-listed number can be achieved by more complicated command-line, for
87example 1600x1200x32bpp can be specified by `video=matroxfb:vesa:0x11C,depth:32`.
88
89
90X11
91===
92
93XF{68,86}_FBDev should work just fine, but it is non-accelerated. On non-intel
94architectures there are some glitches for 24bpp videomodes. 8, 16 and 32bpp
95works fine.
96
97Running another (accelerated) X-Server like XF86_SVGA works too. But (at least)
98XFree servers have big troubles in multihead configurations (even on first
99head, not even talking about second). Running XFree86 4.x accelerated mga
100driver is possible, but you must not enable DRI - if you do, resolution and
101color depth of your X desktop must match resolution and color depths of your
102virtual consoles, otherwise X will corrupt accelerator settings.
103
104
105SVGALib
106=======
107
108Driver contains SVGALib compatibility code. It is turned on by choosing textual
109mode for console. You can do it at boot time by using videomode
1102,3,7,0x108-0x10C or 0x1C0. At runtime, `fbset -depth 0` does this work.
111Unfortunately, after SVGALib application exits, screen contents is corrupted.
112Switching to another console and back fixes it. I hope that it is SVGALib's
113problem and not mine, but I'm not sure.
114
115
116Configuration
117=============
118
119You can pass kernel command line options to matroxfb with
120`video=matroxfb:option1,option2:value2,option3` (multiple options should be
121separated by comma, values are separated from options by `:`).
122Accepted options:
123
124============ ===================================================================
125mem:X        size of memory (X can be in megabytes, kilobytes or bytes)
126	     You can only decrease value determined by driver because of
127	     it always probe for memory. Default is to use whole detected
128	     memory usable for on-screen display (i.e. max. 8 MB).
129disabled     do not load driver; you can use also `off`, but `disabled`
130	     is here too.
131enabled      load driver, if you have `video=matroxfb:disabled` in LILO
132	     configuration, you can override it by this (you cannot override
133	     `off`). It is default.
134noaccel      do not use acceleration engine. It does not work on Alphas.
135accel        use acceleration engine. It is default.
136nopan        create initial consoles with vyres = yres, thus disabling virtual
137	     scrolling.
138pan          create initial consoles as tall as possible (vyres = memory/vxres).
139	     It is default.
140nopciretry   disable PCI retries. It is needed for some broken chipsets,
141	     it is autodetected for intel's 82437. In this case device does
142	     not comply to PCI 2.1 specs (it will not guarantee that every
143	     transaction terminate with success or retry in 32 PCLK).
144pciretry     enable PCI retries. It is default, except for intel's 82437.
145novga        disables VGA I/O ports. It is default if BIOS did not enable
146	     device. You should not use this option, some boards then do not
147	     restart without power off.
148vga          preserve state of VGA I/O ports. It is default. Driver does not
149	     enable VGA I/O if BIOS did not it (it is not safe to enable it in
150	     most cases).
151nobios       disables BIOS ROM. It is default if BIOS did not enable BIOS
152	     itself. You should not use this option, some boards then do not
153	     restart without power off.
154bios         preserve state of BIOS ROM. It is default. Driver does not enable
155	     BIOS if BIOS was not enabled before.
156noinit       tells driver, that devices were already initialized. You should use
157	     it if you have G100 and/or if driver cannot detect memory, you see
158	     strange pattern on screen and so on. Devices not enabled by BIOS
159	     are still initialized. It is default.
160init         driver initializes every device it knows about.
161memtype      specifies memory type, implies 'init'. This is valid only for G200
162	     and G400 and has following meaning:
163
164	       G200:
165		 -  0 -> 2x128Kx32 chips, 2MB onboard, probably sgram
166		 -  1 -> 2x128Kx32 chips, 4MB onboard, probably sgram
167		 -  2 -> 2x256Kx32 chips, 4MB onboard, probably sgram
168		 -  3 -> 2x256Kx32 chips, 8MB onboard, probably sgram
169		 -  4 -> 2x512Kx16 chips, 8/16MB onboard, probably sdram only
170		 -  5 -> same as above
171		 -  6 -> 4x128Kx32 chips, 4MB onboard, probably sgram
172		 -  7 -> 4x128Kx32 chips, 8MB onboard, probably sgram
173	       G400:
174		 -  0 -> 2x512Kx16 SDRAM, 16/32MB
175		 -	 2x512Kx32 SGRAM, 16/32MB
176		 -  1 -> 2x256Kx32 SGRAM, 8/16MB
177		 -  2 -> 4x128Kx32 SGRAM, 8/16MB
178		 -  3 -> 4x512Kx32 SDRAM, 32MB
179		 -  4 -> 4x256Kx32 SGRAM, 16/32MB
180		 -  5 -> 2x1Mx32 SDRAM, 32MB
181		 -  6 -> reserved
182		 -  7 -> reserved
183
184	     You should use sdram or sgram parameter in addition to memtype
185	     parameter.
186nomtrr       disables write combining on frame buffer. This slows down driver
187	     but there is reported minor incompatibility between GUS DMA and
188	     XFree under high loads if write combining is enabled (sound
189	     dropouts).
190mtrr         enables write combining on frame buffer. It speeds up video
191	     accesses much. It is default. You must have MTRR support enabled
192	     in kernel and your CPU must have MTRR (f.e. Pentium II have them).
193sgram        tells to driver that you have Gxx0 with SGRAM memory. It has no
194	     effect without `init`.
195sdram        tells to driver that you have Gxx0 with SDRAM memory.
196	     It is a default.
197inv24        change timings parameters for 24bpp modes on Millennium and
198	     Millennium II. Specify this if you see strange color shadows
199	     around  characters.
200noinv24      use standard timings. It is the default.
201inverse      invert colors on screen (for LCD displays)
202noinverse    show true colors on screen. It is default.
203dev:X        bind driver to device X. Driver numbers device from 0 up to N,
204	     where device 0 is first `known` device found, 1 second and so on.
205	     lspci lists devices in this order.
206	     Default is `every` known device.
207nohwcursor   disables hardware cursor (use software cursor instead).
208hwcursor     enables hardware cursor. It is default. If you are using
209	     non-accelerated mode (`noaccel` or `fbset -accel false`), software
210	     cursor is used (except for text mode).
211noblink      disables cursor blinking. Cursor in text mode always blinks (hw
212	     limitation).
213blink        enables cursor blinking. It is default.
214nofastfont   disables fastfont feature. It is default.
215fastfont:X   enables fastfont feature. X specifies size of memory reserved for
216	     font data, it must be >= (fontwidth*fontheight*chars_in_font)/8.
217	     It is faster on Gx00 series, but slower on older cards.
218grayscale    enable grayscale summing. It works in PSEUDOCOLOR modes (text,
219	     4bpp, 8bpp). In DIRECTCOLOR modes it is limited to characters
220	     displayed through putc/putcs. Direct accesses to framebuffer
221	     can paint colors.
222nograyscale  disable grayscale summing. It is default.
223cross4MB     enables that pixel line can cross 4MB boundary. It is default for
224	     non-Millennium.
225nocross4MB   pixel line must not cross 4MB boundary. It is default for
226	     Millennium I or II, because of these devices have hardware
227	     limitations which do not allow this. But this option is
228	     incompatible with some (if not all yet released) versions of
229	     XF86_FBDev.
230dfp          enables digital flat panel interface. This option is incompatible
231	     with secondary (TV) output - if DFP is active, TV output must be
232	     inactive and vice versa. DFP always uses same timing as primary
233	     (monitor) output.
234dfp:X        use settings X for digital flat panel interface. X is number from
235	     0 to 0xFF, and meaning of each individual bit is described in
236	     G400 manual, in description of DAC register 0x1F. For normal
237	     operation you should set all bits to zero, except lowest bit. This
238	     lowest bit selects who is source of display clocks, whether G400,
239	     or panel. Default value is now read back from hardware - so you
240	     should specify this value only if you are also using `init`
241	     parameter.
242outputs:XYZ  set mapping between CRTC and outputs. Each letter can have value
243	     of 0 (for no CRTC), 1 (CRTC1) or 2 (CRTC2), and first letter
244	     corresponds to primary analog output, second letter to the
245	     secondary analog output and third letter to the DVI output.
246	     Default setting is 100 for cards below G400 or G400 without DFP,
247	     101 for G400 with DFP, and 111 for G450 and G550. You can set
248	     mapping only on first card, use matroxset for setting up other
249	     devices.
250vesa:X       selects startup videomode. X is number from 0 to 0x1FF, see table
251	     above for detailed explanation. Default is 640x480x8bpp if driver
252	     has 8bpp support. Otherwise first available of 640x350x4bpp,
253	     640x480x15bpp, 640x480x24bpp, 640x480x32bpp or 80x25 text
254	     (80x25 text is always available).
255============ ===================================================================
256
257If you are not satisfied with videomode selected by `vesa` option, you
258can modify it with these options:
259
260============ ===================================================================
261xres:X       horizontal resolution, in pixels. Default is derived from `vesa`
262	     option.
263yres:X       vertical resolution, in pixel lines. Default is derived from `vesa`
264	     option.
265upper:X      top boundary: lines between end of VSYNC pulse and start of first
266	     pixel line of picture. Default is derived from `vesa` option.
267lower:X      bottom boundary: lines between end of picture and start of VSYNC
268	     pulse. Default is derived from `vesa` option.
269vslen:X      length of VSYNC pulse, in lines. Default is derived from `vesa`
270	     option.
271left:X       left boundary: pixels between end of HSYNC pulse and first pixel.
272	     Default is derived from `vesa` option.
273right:X      right boundary: pixels between end of picture and start of HSYNC
274	     pulse. Default is derived from `vesa` option.
275hslen:X      length of HSYNC pulse, in pixels. Default is derived from `vesa`
276	     option.
277pixclock:X   dotclocks, in ps (picoseconds). Default is derived from `vesa`
278	     option and from `fh` and `fv` options.
279sync:X       sync. pulse - bit 0 inverts HSYNC polarity, bit 1 VSYNC polarity.
280	     If bit 3 (value 0x08) is set, composite sync instead of HSYNC is
281	     generated. If bit 5 (value 0x20) is set, sync on green is turned
282	     on. Do not forget that if you want sync on green, you also probably
283	     want composite sync.
284	     Default depends on `vesa`.
285depth:X      Bits per pixel: 0=text, 4,8,15,16,24 or 32. Default depends on
286	     `vesa`.
287============ ===================================================================
288
289If you know capabilities of your monitor, you can specify some (or all) of
290`maxclk`, `fh` and `fv`. In this case, `pixclock` is computed so that
291pixclock <= maxclk, real_fh <= fh and real_fv <= fv.
292
293============ ==================================================================
294maxclk:X     maximum dotclock. X can be specified in MHz, kHz or Hz. Default is
295	     `don`t care`.
296fh:X         maximum horizontal synchronization frequency. X can be specified
297	     in kHz or Hz. Default is `don't care`.
298fv:X         maximum vertical frequency. X must be specified in Hz. Default is
299	     70 for modes derived from `vesa` with yres <= 400, 60Hz for
300	     yres > 400.
301============ ==================================================================
302
303
304Limitations
305===========
306
307There are known and unknown bugs, features and misfeatures.
308Currently there are following known bugs:
309
310 - SVGALib does not restore screen on exit
311 - generic fbcon-cfbX procedures do not work on Alphas. Due to this,
312   `noaccel` (and cfb4 accel) driver does not work on Alpha. So everyone
313   with access to `/dev/fb*` on Alpha can hang machine (you should restrict
314   access to `/dev/fb*` - everyone with access to this device can destroy
315   your monitor, believe me...).
316 - 24bpp does not support correctly XF-FBDev on big-endian architectures.
317 - interlaced text mode is not supported; it looks like hardware limitation,
318   but I'm not sure.
319 - Gxx0 SGRAM/SDRAM is not autodetected.
320 - maybe more...
321
322And following misfeatures:
323
324 - SVGALib does not restore screen on exit.
325 - pixclock for text modes is limited by hardware to
326
327    - 83 MHz on G200
328    - 66 MHz on Millennium I
329    - 60 MHz on Millennium II
330
331   Because I have no access to other devices, I do not know specific
332   frequencies for them. So driver does not check this and allows you to
333   set frequency higher that this. It causes sparks, black holes and other
334   pretty effects on screen. Device was not destroyed during tests. :-)
335 - my Millennium G200 oscillator has frequency range from 35 MHz to 380 MHz
336   (and it works with 8bpp on about 320 MHz dotclocks (and changed mclk)).
337   But Matrox says on product sheet that VCO limit is 50-250 MHz, so I believe
338   them (maybe that chip overheats, but it has a very big cooler (G100 has
339   none), so it should work).
340 - special mixed video/graphics videomodes of Mystique and Gx00 - 2G8V16 and
341   G16V16 are not supported
342 - color keying is not supported
343 - feature connector of Mystique and Gx00 is set to VGA mode (it is disabled
344   by BIOS)
345 - DDC (monitor detection) is supported through dualhead driver
346 - some check for input values are not so strict how it should be (you can
347   specify vslen=4000 and so on).
348 - maybe more...
349
350And following features:
351
352 - 4bpp is available only on Millennium I and Millennium II. It is hardware
353   limitation.
354 - selection between 1:5:5:5 and 5:6:5 16bpp videomode is done by -rgba
355   option of fbset: "fbset -depth 16 -rgba 5,5,5" selects 1:5:5:5, anything
356   else selects 5:6:5 mode.
357 - text mode uses 6 bit VGA palette instead of 8 bit (one of 262144 colors
358   instead of one of 16M colors). It is due to hardware limitation of
359   Millennium I/II and SVGALib compatibility.
360
361
362Benchmarks
363==========
364It is time to redraw whole screen 1000 times in 1024x768, 60Hz. It is
365time for draw 6144000 characters on screen through /dev/vcsa
366(for 32bpp it is about 3GB of data (exactly 3000 MB); for 8x16 font in
36716 seconds, i.e. 187 MBps).
368Times were obtained from one older version of driver, now they are about 3%
369faster, it is kernel-space only time on P-II/350 MHz, Millennium I in 33 MHz
370PCI slot, G200 in AGP 2x slot. I did not test vgacon::
371
372  NOACCEL
373	8x16                 12x22
374	Millennium I  G200   Millennium I  G200
375  8bpp    16.42         9.54   12.33         9.13
376  16bpp   21.00        15.70   19.11        15.02
377  24bpp   36.66        36.66   35.00        35.00
378  32bpp   35.00        30.00   33.85        28.66
379
380  ACCEL, nofastfont
381	8x16                 12x22                6x11
382	Millennium I  G200   Millennium I  G200   Millennium I  G200
383  8bpp     7.79         7.24   13.55         7.78   30.00        21.01
384  16bpp    9.13         7.78   16.16         7.78   30.00        21.01
385  24bpp   14.17        10.72   18.69        10.24   34.99        21.01
386  32bpp   16.15	     16.16   18.73        13.09   34.99        21.01
387
388  ACCEL, fastfont
389	8x16                 12x22                6x11
390	Millennium I  G200   Millennium I  G200   Millennium I  G200
391  8bpp     8.41         6.01    6.54         4.37   16.00        10.51
392  16bpp    9.54         9.12    8.76         6.17   17.52        14.01
393  24bpp   15.00        12.36   11.67        10.00   22.01        18.32
394  32bpp   16.18        18.29*  12.71        12.74   24.44        21.00
395
396  TEXT
397	8x16
398	Millennium I  G200
399  TEXT     3.29         1.50
400
401  * Yes, it is slower than Millennium I.
402
403
404Dualhead G400
405=============
406Driver supports dualhead G400 with some limitations:
407 + secondary head shares videomemory with primary head. It is not problem
408   if you have 32MB of videoram, but if you have only 16MB, you may have
409   to think twice before choosing videomode (for example twice 1880x1440x32bpp
410   is not possible).
411 + due to hardware limitation, secondary head can use only 16 and 32bpp
412   videomodes.
413 + secondary head is not accelerated. There were bad problems with accelerated
414   XFree when secondary head used to use acceleration.
415 + secondary head always powerups in 640x480@60-32 videomode. You have to use
416   fbset to change this mode.
417 + secondary head always powerups in monitor mode. You have to use fbmatroxset
418   to change it to TV mode. Also, you must select at least 525 lines for
419   NTSC output and 625 lines for PAL output.
420 + kernel is not fully multihead ready. So some things are impossible to do.
421 + if you compiled it as module, you must insert i2c-matroxfb, matroxfb_maven
422   and matroxfb_crtc2 into kernel.
423
424
425Dualhead G450
426=============
427Driver supports dualhead G450 with some limitations:
428 + secondary head shares videomemory with primary head. It is not problem
429   if you have 32MB of videoram, but if you have only 16MB, you may have
430   to think twice before choosing videomode.
431 + due to hardware limitation, secondary head can use only 16 and 32bpp
432   videomodes.
433 + secondary head is not accelerated.
434 + secondary head always powerups in 640x480@60-32 videomode. You have to use
435   fbset to change this mode.
436 + TV output is not supported
437 + kernel is not fully multihead ready, so some things are impossible to do.
438 + if you compiled it as module, you must insert matroxfb_g450 and matroxfb_crtc2
439   into kernel.
440
441Petr Vandrovec <vandrove@vc.cvut.cz>
442