ChangeLog revision b698ba48
1commit 52f40e7569bbcc9bce8f5d885b0de2868398e8df
2Author: Dave Airlie <airlied@redhat.com>
3Date:   Wed Aug 5 11:09:26 2009 +1000
4
5    fix build after headers moved
6
7commit 9d9bfe0349693b283e5dc1ced6fa882f10619253
8Author: Dave Airlie <airlied@redhat.com>
9Date:   Wed Aug 5 11:08:15 2009 +1000
10
11    update for 1.7.3
12
13commit 115b564f78b038baef2d14c037785850efb4e9ca
14Author: Dave Airlie <airlied@redhat.com>
15Date:   Tue Jul 28 15:22:41 2009 +1000
16
17    siliconmotion: change to using ABI version check
18
19commit 70f2463faa6d21517d97aa460871e1b5f3fa6505
20Author: Dave Airlie <airlied@redhat.com>
21Date:   Tue Jul 28 13:32:36 2009 +1000
22
23    siliconmotion: update for resources/RAC API removal
24
25commit 841df790b5b6700e9a1d6d41ee04b0b913329084
26Author: Peter Hutterer <peter.hutterer@who-t.net>
27Date:   Thu Jul 16 11:55:05 2009 +1000
28
29    Update to xextproto 7.1 support.
30    
31    DPMS header was split into dpms.h (client) and dpmsconst.h (server). Drivers
32    need to include dpmsconst.h if xextproto 7.1 is available.
33    
34    Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
35
36commit d754e9a2954e56a26a94e58d87a450a825c250e6
37Author: Adam Jackson <ajax@redhat.com>
38Date:   Thu Jul 2 11:43:52 2009 -0400
39
40    siliconmotion 1.7.2
41
42commit cd9e9f25294abb739c1772e3a9a7a9c0a921ee82
43Author: Francisco Jerez <currojerez@riseup.net>
44Date:   Tue Jun 23 13:11:02 2009 +0200
45
46    Fix EXA rotation for SM502 at 32bpp (bug 22432).
47
48commit 97498c048c897e5753e61d3b4ab231025974d67c
49Author: Adam Jackson <ajax@redhat.com>
50Date:   Thu May 28 15:17:58 2009 -0400
51
52    Remove useless loader symbol lists.
53
54commit 536e5df957698251206326edc5a13e833f7c72b1
55Author: Francisco Jerez <currojerez@riseup.net>
56Date:   Wed May 27 01:23:36 2009 +0200
57
58    Program MCLK/MXCLK on startup for a secondary SM502.
59    
60    This should allow using a SM502 as secondary display
61    device (bug 21810).
62
63commit 6a370fa2b6b8fcbd556dd6f9bf92872e9bea23e8
64Author: Alan Coopersmith <alan.coopersmith@sun.com>
65Date:   Mon May 11 22:31:30 2009 -0700
66
67    Fill in COPYING file, add SubmittingPatches URL to README
68
69commit 962e0c4f08180ac51994aef8d6d042108d5841c1
70Author: Francisco Jerez <currojerez@riseup.net>
71Date:   Sun May 10 01:53:05 2009 +0200
72
73    Widen the pixel clock readjust interval for SM712.
74    
75    Set it to 3MHz so that the pixel clock frequency is overridden
76    when it's found to be 49MHz, which is reported to be unstable.
77
78commit 3cc1b9d4bcf1b7a756ea4426c93392ff6d82e9a6
79Author: Francisco Jerez <currojerez@riseup.net>
80Date:   Tue May 5 00:45:10 2009 +0200
81
82    Fail gracefully when the specified virtual screen doesn't fit in RAM.
83
84commit 32628d9884d577d7a672c172c12b7097be276700
85Author: Francisco Jerez <currojerez@riseup.net>
86Date:   Tue May 5 00:44:17 2009 +0200
87
88    Probe the amount of installed memory by trial and error on SM712.
89    
90    In some cases the BIOS hasn't filled in the "scratchpad registers"
91    (SR71) with the right amount of memory installed (e.g. MIPS
92    platform). There seems to be no other way to do it than to test it.
93    
94    This should fix bug 21528.
95
96commit 00921b014fa0b5358c22a769cf2450cbd4bdc8a5
97Author: Francisco Jerez <currojerez@riseup.net>
98Date:   Tue Apr 28 22:24:03 2009 +0200
99
100    Bump version to 1.7.1.
101
102commit abf1ba79f2bfe61f24cfa43deb0400d7c5f95bd5
103Author: Francisco Jerez <currojerez@riseup.net>
104Date:   Tue Apr 21 21:37:45 2009 +0200
105
106    Increase the maximum clock value to 200MHz on SM712.
107    
108    The default MCLK setting was higher than the clock limit, and it
109    failed.
110
111commit 41e5c49024d4e27a7be3da02017000a22b59016f
112Author: Francisco Jerez <currojerez@riseup.net>
113Date:   Tue Apr 21 21:36:10 2009 +0200
114
115    Don't attempt monitor detection on SM712.
116    
117    The databook says nothing about it, and it doesn't work.
118
119commit 6691678b5f04d52a1e08f9eaa6d511df9b7eeb4d
120Author: Jamie Lentin <jm@lentin.co.uk>
121Date:   Thu Mar 19 21:54:12 2009 +0000
122
123    Stop clearing of "VESA compliance power down mode" bit
124    
125    Clearing this bit causes an OQO 01+ w/SMI720 to power down the LCD,
126    leave it alone.
127    
128    Signed-off-by: Jamie Lentin <jm@lentin.co.uk>
129    Signed-off-by: Francisco Jerez <currojerez@riseup.net>
130
131commit 05480382c508bd27c1755405518ce9c339f0f80e
132Author: Matthieu Herrb <matthieu.herrb@laas.fr>
133Date:   Fri Mar 13 21:19:16 2009 +0100
134
135    Fix direct access to IO space on chipsets with no IOBase mapping.
136    
137    Signed-off-by: Francisco Jerez <currojerez@riseup.net>
138
139commit e62807140ede74e262ed70c846017dc3506b511e
140Author: Niels de Vos <niels.devos@wincor-nixdorf.com>
141Date:   Wed Feb 11 10:10:13 2009 +0100
142
143    siliconmotion: Fix disabling of debugging if SMI501_CLI_DEBUG is set to 0
144    
145    Due to checking if SMI501_CLI_DEBUG is defined, some debugging will
146    be enabled if SMI501_CLI_DEBUG is set to 0. A single #if should be
147    used instead. Some debugging code already does this.
148    
149    Signed-off-by: Niels de Vos <niels.devos@wincor-nixdorf.com>
150    Signed-off-by: Francisco Jerez <currojerez@riseup.net>
151
152commit 0b34c4a29fcf45ac24361b4691cca0fd99cc3a3f
153Author: Francisco Jerez <currojerez@riseup.net>
154Date:   Wed Feb 4 03:49:08 2009 +0100
155
156    Wait for vertical retrace before writing registers at SMILynx_CrtcDPMS_*
157
158commit 1d46cca27b0d38d5355b5f93dd0ed0fe4f250d3f
159Author: Francisco Jerez <currojerez@riseup.net>
160Date:   Tue Feb 3 22:15:45 2009 +0100
161
162    Drop the outdated configuration options documentation in README.
163
164commit 64f722c3121a8f28cfc5b532ddec0ac83095ac95
165Author: Francisco Jerez <currojerez@riseup.net>
166Date:   Tue Feb 3 22:17:33 2009 +0100
167
168    Minor corrections at the man page.
169
170commit 94d9c659b71c84db46fa279d9d8736e1589e155c
171Author: Francisco Jerez <currojerez@riseup.net>
172Date:   Tue Feb 3 22:18:14 2009 +0100
173
174    Set dualhead to on by default on SM72x chipsets.
175
176commit 4d98e9c6a6e357ed736cd312a957c1e0a7b1d36b
177Author: Francisco Jerez <currojerez@riseup.net>
178Date:   Wed Feb 4 01:56:20 2009 +0100
179
180    Dynamically switch virtual refresh mode.
181    
182    Probably this makes dualhead mode more useful because it makes
183    possible displaying video on the LCD as long as the CRT output is
184    disabled or cloned.
185
186commit 1c57cfa814504c33fcad51115d77f2d166fb7903
187Author: Alan Coopersmith <alan.coopersmith@sun.com>
188Date:   Fri Jan 30 21:00:17 2009 -0800
189
190    Add README with pointers to mailing list, bugzilla & git repos
191
192commit 2aee107d992cb58bfbb2f527cfa73ea9f6957998
193Author: Alan Coopersmith <alan.coopersmith@sun.com>
194Date:   Fri Jan 9 16:33:43 2009 -0800
195
196    Remove xorgconfig & xorgcfg from See Also list in man page
197
198commit fdc2c50682b240e5966e4e5a7b45552ee2043bfc
199Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
200Date:   Mon Dec 29 18:41:05 2008 -0200
201
202    Bump release to 1.7.0.
203    
204      This driver supports smi 50x chipsets, randr 1.2, exa, dual head,
205    etc.
206    
207      Special thanks to Teddy Wang <teddy.wang AT siliconmotion.com.cn>
208    for support and help in responding and/or triaging hardware related
209    questions.
210
211commit 97eb5da1040ca7882e60b86ffaf25b6c021a3df5
212Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
213Date:   Mon Dec 29 18:39:54 2008 -0200
214
215    Use util-macros XORG_CHANGELOG and XORG_CWARNFLAGS.
216    
217      Also correct only compilation warning about possibly
218    uninitialized variable.
219
220commit deb2f845dd370ba819d02cb21b8a481827497af6
221Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
222Date:   Mon Dec 29 18:35:22 2008 -0200
223
224    Simplification and redundancy removal in video code.
225    
226      FOURCC_YV12 and FOURCC_I420 handling also was buggy. First it was
227    doing a noop by swapping offset2 and offset3 values twice, and second,
228    swap is not required when using smi 501/502 CSC video.
229      Changed SMI_DisplayVideo0501_CSC() to not set static values to
230    registers in a possible loop, if there is clipping.
231
232commit 46741589529809c17aa1e9719492a4b623de6ddf
233Author: Francisco Jerez <currojerez@gmail.com>
234Date:   Sat Dec 27 00:52:05 2008 +0100
235
236    Some more quirks for the SM712.
237    
238    * Program the MCLK to 157MHz on startup.
239    * Adjust the requested pixel clock if it's near one of the known
240      stable frequencies.
241    * Prefer the clock alternative with post scalar turned on when the
242      denominator is even.
243
244commit 439adf4455c651926040eeeec52a092f14f3196b
245Author: Francisco Jerez <currojerez@gmail.com>
246Date:   Mon Dec 22 17:28:35 2008 +0100
247
248    Add support for clone mode on Lynx chipsets.
249
250commit 2b3fa385a6da4be5ad6719dd115834b96d1ea3e3
251Author: Francisco Jerez <currojerez@gmail.com>
252Date:   Sun Dec 21 19:37:14 2008 +0100
253
254    Add a CRTC/Output implementation using BIOS for modesetting.
255    
256    After the RandR1.2 implementation the "UseBIOS" option wasn't actually
257    programming the hardware through VESA BIOS, this brings back that
258    functionality.
259
260commit 75b1c309f1418d70e8cf112744ff95f372d4ef75
261Author: Francisco Jerez <currojerez@gmail.com>
262Date:   Sun Dec 21 16:58:29 2008 +0100
263
264    Fix SMI_CrtcShadowAllocate.
265
266commit 068de9c0a302aba21e4d6a27eb6d330375ff2d2f
267Author: Francisco Jerez <currojerez@gmail.com>
268Date:   Sun Dec 21 16:02:57 2008 +0100
269
270    Allocate crtc->funcs and output->funcs in the heap.
271
272commit 373440667636d4e3214028710e33ac89d8f18ff1
273Author: Francisco Jerez <currojerez@gmail.com>
274Date:   Sun Dec 21 13:51:01 2008 +0100
275
276    Cleanup the Lynx register saving/restoring code.
277    
278    Save some registers not previously tracked, and use pSmi->mode instead
279    of continuously reading the hardware state.
280
281commit 67a1e4b090d6647b6c45f1534c67d9ea8a223d44
282Author: Francisco Jerez <currojerez@gmail.com>
283Date:   Sat Dec 20 15:41:13 2008 +0100
284
285    Add some quirks for SM712 modesetting.
286
287commit 7f252345c0c6b260c5b37aff98a22679eb5c6c82
288Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
289Date:   Fri Dec 19 21:15:01 2008 -0200
290
291    Enable the SMI 501/502 command list interpreter in a "debug" build.
292    
293      To enable it, set SMI501_CLI_DEBUG to 1 in smi.h, and use
294    Option "AcellMethod "EXA"
295    in the Device section of /etc/X11/xorg.conf
296    This code is enabled mainly for debug purposes. To make if have an
297    actual performance gain (like when using a sm50x with a "low profile"
298    "main" processor") it should be required to actually do busy loops
299    in kernel mode (and hope the costs of context switch will pay it).
300    In kernel mode it is possible to wait for an interrupt being triggered
301    when the command list is processed, or when the 2d engine is idle.
302    
303      This commit should be functional, but, mainly due to debug messages,
304    should be significantly slower then a build with MI501_CLI_DEBUG
305    defined to 0.
306
307commit 6a91bcc111902c45cc75c865893848b7c6c0a0b1
308Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
309Date:   Thu Dec 18 20:47:24 2008 -0200
310
311    Add definitions for the SMI 501/502 "command list interpreter".
312    
313      This also changes some bit operations to use a "bitfield" equivalent
314    one, with named fields, that should make it easier to understand what
315    is being tested.
316    
317      The enum smi_cli_cmd_code in smi_501.h is code that was added to a
318    experimental smi_drm.h, but the hardware only supports basic 2d accel,
319    and to compensate for the extra overhead for maintaining a command
320    list (assuming it worked correctly) it would be required to have a
321    special handling, like calling an ioctl to do the "busy loop" in the
322    kernel (that is, should wait for an irq or a timeout).
323    
324      The problem is that even if waiting for a idle engine before crafting
325    a command, and waiting again after submitting the command, there would
326    be corruption on screen after some time. So, the "busy loop" in the
327    kernel would only be useful if still using direct writes to mmio
328    registers.
329
330commit 5a07709ae41600bc02205753c64d764f11838240
331Author: Francisco Jerez <currojerez@gmail.com>
332Date:   Tue Dec 16 02:39:49 2008 +0100
333
334    Fix crashes when switching VTs with EXA enabled.
335    
336    Call ModifyPixmapHeader on SMI_EnterVT to update the screen pixmap
337    address tracked by EXA.
338    
339    Set memoryBase in the ExaDriverRec to NULL when switching out: some
340    pixmap could get allocated near the former aperture address and it
341    could be erroneously considered as being in offscreen memory.
342
343commit 47bcfb505886d16ff864f26a9cfa5f43b5da710e
344Author: Francisco Jerez <currojerez@gmail.com>
345Date:   Sat Dec 13 19:13:50 2008 +0100
346
347    Enable linear memory mode on SMI_MapMmio.
348    
349    Set UseBIOS to off by default for SM720. Some improvements on the
350    debugging output.
351
352commit 6f972c9022fc3a315b60246a69399807443d2ca0
353Author: Francisco Jerez <currojerez@gmail.com>
354Date:   Thu Dec 11 16:21:17 2008 +0100
355
356    Fall back to UseBIOS off when VBEInit fails.
357
358commit 1fc340dcf0abd7fa0837fef2c711aaec4fa8bcb8
359Author: Francisco Jerez <currojerez@gmail.com>
360Date:   Thu Dec 11 16:18:50 2008 +0100
361
362    Fix XAA SolidFill with 32 bpp framebuffer.
363
364commit dfc4510e8f4adeed520be47621ad429bc2d18f7e
365Author: Francisco Jerez <currojerez@gmail.com>
366Date:   Thu Dec 4 15:35:22 2008 +0100
367
368    Some corrections on the Lynx modesetting code.
369    
370    This fixes 1280x1024 modes on the VGA output with dual head on.
371    Also, solve a build problem when using -DSMI_DEBUG, as pointed by
372    Richard Schwarting.
373
374commit 6e1d29d19b205eac85b2f83de126ba88501e9b74
375Author: Francisco Jerez <currojerez@gmail.com>
376Date:   Tue Dec 2 23:43:26 2008 +0100
377
378    Disable screen centering on mode initialization.
379
380commit 695a6066cfe7ad0e568edc596b8704bb9bf8754f
381Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
382Date:   Sat Dec 13 04:20:20 2008 -0200
383
384    Janitor/maintenance patches.
385    
386    o Define HAVE_XMODES in config.h.
387    o Correct the LEAVE() macro declaration when SMI_DEBUG is defined.
388    o Remove prototype for SMI_DGAInit() if HAVE_XMODES is not defined.
389    o Properly scale value read from sm501 CURRENT_CLOCK register, to
390      print an appropriate message on startup.
391
392commit 191cbd3f8419cc5b428f262607d7a8594e574f67
393Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
394Date:   Fri Dec 12 15:50:53 2008 -0200
395
396    Unless there is a xorg.conf option, don't modify mclck neither mxclck.
397    
398      The option set_mclck was renamed to just MCLK, and a new MXCLCK option
399    was added for smi501/502 chips.
400      Note that previously, on MSOC chips, the set_mclck option actually
401    changed mxclck, what is now done by the MXCLK option.
402      The manpage was also update to reflect the options usage.
403      A new SMI501_FindMemClock() function was added, as a possible
404    placeholder for choosing alternate clocks when programming MMIO 0x68.
405    But it should be noted that the smi 501 databook says:
406    Miscellaneous Timing
407    	Read/Write MMIO_base + 0x000068
408    	Power-on Default 0x00000000
409    	[...]
410    	5:4 Divider Output Frequency Selection for Second PLL.
411    		00: 336 MHz.
412    		01: 288 MHz.
413    		10: 240 MHz.
414    		11: 192 MHz.
415    while the 502 databook says:
416    	Miscellaneous Timing
417    	Read/Write MMIO_base + 0x000068
418    	Power-on Default 0b0000.00xx.0000.1001.0000.1001.0000.0000
419    	[...]
420    	5:4 Reserved These bits are reserved.
421    but still tells to refer to those bits.
422
423commit a2352ff23787f37530dff68e03028dd78925e4b4
424Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
425Date:   Fri Nov 28 20:23:11 2008 -0200
426
427    Properly restore console on the smi 501/502.
428    
429      The current kernel framebuffer module expects accelerations registers
430    in the default boot state, otherwise, it will display some corruption,
431    or not correctly clear the screen, among other side effects.
432      Now it should be restoring everything the kernel framebuffer relies
433    on not being modified.
434
435commit 5a65559e12452a47f9848c47cfa6e30cfb871f0b
436Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
437Date:   Fri Nov 28 19:49:20 2008 -0200
438
439    DGA is dependent on HAVE_XMODES.
440
441commit 06c78cee718a86b1521c1fc1075d08bb9ad257f1
442Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
443Date:   Tue Nov 25 19:57:05 2008 -0200
444
445    Remove smi_dga.c and polylines code from smi_xaa.c.
446    
447      DGA was also not being properly test to not enable the pSmi fields,
448    but since the dga implementation is very basic, it is better to just
449    remove it (if compiling the sources on pre 1.4 X Server, DGA will be
450    disabled).
451    
452      The polylines code in xmi_xaa.c was already if 0'ed for some time,
453    but it may be reinstantiated to accelerate (back) xaa rotated modes.
454    Currently only randr+exa accelerates rotated modes.
455
456commit 9570cfdc132b0acb1d2b70edc6a3a5a812ef9691
457Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
458Date:   Tue Nov 25 15:07:39 2008 -0200
459
460    Ensure the engine is idle before switching to a virtual terminal.
461    
462      At least in the smi501/502, it may even completely lock in some special
463    cases (like when there is a browser playing some youtube video).
464      The code in SMI_CloseScreen() was almost identical to SMI_LeaveVT(),
465    so it now just call SMI_Leave() if required.
466
467commit 26ce0b163d9cf44ec1bc1d1bd460887aa5b405a7
468Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
469Date:   Mon Nov 24 15:15:43 2008 -0200
470
471    Use a single debug macro for leaving functions.
472    
473      The LEAVE() macro, when debug is disabled, now is defined as:
474    <hash>define LEAVE(...)		return __VA_ARGS__
475    This avoids the requirement of a return statement after the LEAVE()
476    macro call, what can be confusing.
477
478commit 2057d7ae05ef269f75a4d53fe5c4914a10856cb1
479Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
480Date:   Wed Nov 19 16:22:29 2008 -0200
481
482    Declare DGA fields only when not using xf86DiDGAInit.
483
484commit c2e318eaec3ac281457acdca4ea10fa368f646d3
485Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
486Date:   Mon Nov 17 20:08:49 2008 -0200
487
488    Enable XAA and randr rotated modes.
489    
490      Use the same logic for offscreen memory management as used in
491    smi_video.c. This also makes the functions SMI_AllocateMemory and
492    SMI_FreeMemory no longer static in smi_video.c.
493
494commit ed6aba73259a1412e23a954d6a8dba439ad98a8a
495Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
496Date:   Fri Nov 14 22:37:56 2008 -0200
497
498    Correct EXA+Randr+rotated mode.
499    
500      The smi 501/502 cannot rotate-blt more than 32 bytes at a time.
501    Patch based on smi's sample smi_shadow.c.
502      XAA + randr + rotated mode currently not supported.
503
504commit 85d118ee288fe7aa32db003bfa322ddf06944bf6
505Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
506Date:   Fri Nov 14 22:04:17 2008 -0200
507
508    Don't explicitly call the cursor routines.
509    
510      On the MSOC, the hardware cursor was remaining on screen, when using
511    EXA and a randr rotated mode.
512
513commit cd10af22240b9df7c264b15c483d8cc4505ee359
514Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
515Date:   Fri Nov 14 20:36:09 2008 -0200
516
517    Only compile smi_dga.c when xf86DiDGAInit is not available.
518    
519      Besides reducing driver size, this should make DGA+EXA accelerated.
520
521commit 41ef793b0ed6014ed49bef011259f977833aef5d
522Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
523Date:   Fri Nov 14 19:24:45 2008 -0200
524
525    Correct regressions in CSCVideo
526    
527      Commit 0870d46718fe4e01953efd63cec46c54140b20f9 broke the clipping
528    code. That has been corrected in this patch. Anyway, the hardware color
529    space conversion should not be hidden down in the driver code, instead
530    it should be usable by external programs as well, instead of having
531    some applications doing it by software.
532
533commit be226bfaad76e7b48e006fe55ebae09b044ba4b7
534Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
535Date:   Fri Nov 14 16:56:46 2008 -0200
536
537    MSOC: Hide cursor on mode setup when using software cursor.
538    
539      The kernel framebuffer module uses the hw cursor for the console cursor,
540    so the driver must ensure it is hidden when setting a video mode.
541
542commit 5b8583fc2e5626ed98877858158d9687d7c94469
543Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
544Date:   Fri Nov 14 15:47:52 2008 -0200
545
546    Be more liberal about mode width validation.
547    
548      The code was inherited from first import of latest smi sources, but
549    the driver actually programs a mode, so there is no need to check if
550    there will be a table entry with register contents matching the requested
551    mode.
552
553commit ffb4be763ced82158a3a1d2e06fb0e81815e057b
554Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
555Date:   Thu Nov 13 20:01:18 2008 -0200
556
557    Don't restore alpha registers, as they are not being modified.
558    
559      Restoring the registers "without need" caused some instability, with
560    random locks.
561
562commit 6b6da04d566ec5b9d723c9f28791c604f76526dd
563Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
564Date:   Thu Nov 13 19:44:14 2008 -0200
565
566    Changed to use panel plane tl and tr to center modes smaller then panel size.
567    
568      The code is under "<hash>ifdef USE_PANEL_CENTER", as it is buggy, and
569    regardless of value set to right and bottom, it will crop from 0 to
570    mode-width/mode-height, and then display a lot of screen artifacts, due
571    to improper programming. Either way, the CRT in clone mode will display
572    correctly.
573
574commit d2709b1af22a06b24606ec8c01f39f1ca27ca8d3
575Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
576Date:   Thu Nov 13 18:00:46 2008 -0200
577
578    Enable gamma correction on the MSOC.
579
580commit 0870d46718fe4e01953efd63cec46c54140b20f9
581Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
582Date:   Tue Nov 11 19:14:36 2008 -0200
583
584    Correct a problem in CSC video on the smi 501/502.
585    
586      This significantly simplify the function SMI_DisplayVideo0501_CSC(),
587    but the real cause of the problem that this patch corrects is that
588    whenever MMIO 0x1000fc bit 24 (Vertical Linear Filter Control)  is
589    set, the last line of the rendered video will display only noisy,
590    regardless of video being rendered 1x1, or scaled up or down.
591
592commit 09644ad3da4cf4c2576532137fdc81156e9c06e6
593Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
594Date:   Tue Nov 11 15:34:27 2008 -0200
595
596    Disable detection of crt for the smi 501/502.
597    
598      As described by siliconmotion, it does not work on the smi 502.
599    Does it work on older revisions? Keeping <hash>ifdef'ed out for now
600    as it hopefully will be corrected in a newer hardware revision.
601
602commit caac648932df69419b8f86c5793a6b33f5997099
603Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
604Date:   Tue Nov 11 15:06:15 2008 -0200
605
606    Implement an argb cursor using the alpha layer.
607    
608      The code is <hash>ifdef'ed out by <hash>if SMI_CURSOR_ALPHA_PLANE
609    because the smi 502 hardware is buggy, but it appears to have worked
610    for some older hardware revisions (by looking at code available at
611    ftp.siliconmotion.com.tw).
612      Keeping it at least for now, as it serves as a simple way to reproduce
613    the problems described in smi.h.
614
615commit 61dd360a30c8e530185bc71b134c287ab1cdde2c
616Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
617Date:   Tue Nov 11 14:47:58 2008 -0200
618
619    Add support for the reading/writing the alpha layer registers.
620
621commit d018ea2272cc1938cc0236eeb210692c1d1fd1df
622Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
623Date:   Tue Nov 11 14:35:36 2008 -0200
624
625    Correct wrong offset value for secondary hardware cursor.
626    
627      Hardware cursor is still disabled in dual head mode due to some problems,
628    usually when having both panel and crt mapped to the same address, what
629    appears to confuse the "modes" code.
630
631commit f97c57873cd7622c373a394548ed435c1975abbc
632Author: Francisco Jerez <currojerez@gmail.com>
633Date:   Sat Nov 8 21:40:49 2008 +0100
634
635    Remove unused fifo_* options.
636
637commit 418b28028e69d14e74bf673be56123e99103e5bf
638Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
639Date:   Sat Nov 8 16:04:58 2008 -0200
640
641    Remove the ZoomOnLCD option and update manpage for PCIRetry.
642    
643      ZoomOnLCD controlled if Ctrl/Alt/+ or Ctrl/Alt/- could be used to
644    reprogram the mode on the panel. The option was not used, and now, if
645    different resolutions could be used, it should be advertised by randr.
646    
647      PCIBurst and PCIRetry options are now enabled by default.
648    
649      The fifo_xxx options need a review, as they appear to be associated
650    with the older WaitQueue macro, that did ignore its argument. But these
651    options are not used, and may be removed or reimplemented at some time.
652
653commit bbc38b2a14d748bb9d9e153fcdcb259e5b10bdc8
654Author: Francisco Jerez <currojerez@gmail.com>
655Date:   Fri Nov 7 16:57:50 2008 +0100
656
657    Update the man page
658    
659    Reflect changes in the configuration options, additional supported
660    chipsets and RandR1.2 interface.
661
662commit 7f8b7535a9a905c1d6c1d89cd1dd4dd506025429
663Author: Nathael Pajani <nathael.pajani@ed3l.fr>
664Date:   Thu Mar 20 07:25:00 2008 -0800
665
666    Bit twelve on CPR00 bitfield is not bit eleven...
667
668commit 98969b1b96244085ff5b0c76d2a463c1e436bcfa
669Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
670Date:   Thu Nov 6 16:35:15 2008 -0200
671
672    Use the CRT Monitor Detect MSOC register.
673    
674      Probably I misunderstood the usage of this register, as it will always
675    show the same value, regardless of having a crt connected to the "vga"
676    port or not. Only difference is that while the detect bit is set,
677    the crt will be blank. This patch should be more of a placeholder for
678    a possible correction, but it should not cause any side effects, unless
679    the data field can be zero on a valid situation.
680      In my test computer, this field, in base 2 is always 1000000010000100,
681    what doesn't really look like data in rgb 8:8:8...
682
683commit 87a3cc0652666c1796fd56f00810834193f3eca0
684Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
685Date:   Tue Nov 4 16:38:12 2008 -0200
686
687    Allow setting modes larger then panel for secondary output.
688    
689      Sample usage on sm502:
690    $ DISPLAY=:0 xrandr --output VGA --newmode 1280x1024 109.00 1280 1368 1496 1712  1024 1027 1034 1063 -hsync +vsync
691    $ DISPLAY=:0 xrandr --addmode VGA 1280x1024
692    $ DISPLAY=:0 xrandr --output VGA --mode 1280x1024 --pos 768x176
693      Last command assuming using XAA (with ``Virtual 2048 1024''), so that it
694    will have panel in top-left, and crt in bottom-right.
695
696commit 1d7b9547e1d37601f14410857c22f256e5c74f14
697Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
698Date:   Tue Nov 4 15:57:12 2008 -0200
699
700    Add CSCVideo option to smi 501/502.
701    
702      This code is an adaptation of SMI sample implementation.
703      CSCVideo is an alternate way to render video, that should reduce memory
704    bandwidth usage, and uses the color space conversion hardware to render
705    video directly to the framebuffer memory.
706      When using randr dual head mode, currently only XAA is supported. As it
707    has a fixed format framebuffer (when using the Virtual xorg.conf option).
708      This patch also ensures that pScrn->displayWidth, pScrn->virtualX and
709    pScrn->virtualY are only changed by the driver when using EXA.
710
711commit 93308d067c93126dd0aba9b718d9e0ba28dde9f4
712Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
713Date:   Thu Oct 30 19:07:40 2008 -0200
714
715    Make MSOC work with EXA offscreen manager.
716
717commit e5cd9e4640555b9a3d4fab205e4b77faaca89d89
718Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
719Date:   Thu Oct 30 19:05:40 2008 -0200
720
721      Update memory detection for MSOC.
722    
723      Use "local memory" field of mmio 0x010 and not "system memory", as
724    it uses system memory values whenever there is a choice (just sets
725    bits to zero).
726
727commit 5acf7e22c256637ba7139733cdba1c59557dd7ff
728Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
729Date:   Thu Oct 30 19:01:54 2008 -0200
730
731      Disable HW cursor in dualhead mode in smi 501/502
732    
733      Other chipsets are doing the same. But this apparently is to circumvent
734    a problem in randr/cursor implementation, that appears to not work very
735    well with hardware that supports two hw cursors, but not argb cursors.
736
737commit c756ed67cbcf5bdd40c300627802d8bcd3d95123
738Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
739Date:   Wed Oct 29 19:17:54 2008 -0200
740
741      SMI501/502 cursor fixes.
742    
743      Use separate buffers for crt and panel hw cursors.
744      Describe, and workaround a problem with Dualhead and hw cursor enabled.
745    It still has the inconvenient of having a blinking X when exiting the server
746    or switching to a VT; maybe should also save the image currently configured
747    also (and restore on exit), or just disable hw cursor on dual head mode.
748      Also revert change to not initialize video on dualhead. Better to have
749    fully functional video only on panel, that can be resized, then non
750    resizable video..
751
752commit 2671776999a5add36e1aeaad4080bfccedc2b1a5
753Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
754Date:   Tue Oct 28 19:45:31 2008 -0200
755
756    Update msoc to use randr cursor routines
757    
758      pSmi->IsSecondary was not used, and there should not exist two
759    instances of the driver managing the same card.
760      Macro CHECK_SECONDARY() was dependant on a second instance of the driver
761    for the same card, and also removed.
762      Some rework should still be done in the cursor code, as it sometimes,
763    when displaying different data, will leave the hw cursor visible, while
764    using an argb cursor.
765
766commit 91b5613fb07fec0f319623ca7ec1efd67f04379b
767Author: Francisco Jerez <currojerez@gmail.com>
768Date:   Tue Oct 28 00:13:55 2008 +0100
769
770    Remove shadowfb based rotation support.
771    
772    Also done some clean up, like removing unused/redundant members of the
773    screen private structure, and removing now unused configuration file
774    options.
775
776commit 0faeddb208d8fade4737be4bcadb8814696bd7b4
777Author: Francisco Jerez <currojerez@gmail.com>
778Date:   Tue Oct 28 00:06:37 2008 +0100
779
780    Lynx hardware cursor code adapted to the CRTC interfaces.
781
782commit 5d151781515cf2e3af1322ce6ed5f8d871c05791
783Author: Francisco Jerez <currojerez@gmail.com>
784Date:   Mon Oct 27 23:37:30 2008 +0100
785
786    Changes in the video overlay clipping code.
787    
788    Clip the video to the CRTC viewport by using
789    xf86_crtc_clip_video_helper, and allow XV in
790    dual head mode.
791
792commit 486e92399f09b9d765d03184fe29a32b3bc4ecc5
793Author: Francisco Jerez <currojerez@gmail.com>
794Date:   Mon Oct 27 23:12:36 2008 +0100
795
796    Simple EXA Composite implementation.
797    
798    * Implementation of the PictOpSrc render operation with a
799      source coordinate transformation through 2D engine rotate-BITBLTs.
800    * Fix EXA Copy and Solid in 24 bpp packed color mode.
801
802commit c0447d33c82829248e642b3156fd9a3c0d0eb709
803Author: Francisco Jerez <currojerez@gmail.com>
804Date:   Mon Oct 27 22:39:52 2008 +0100
805
806    Some corrections in the CRTC code.
807    
808    * Make screen resizing behave better with XAA.
809    * Move some local variable declarations to the top of
810      the function definition to conform ANSI C.
811    * Make the Lynx panel modesetting code to actually
812      program the panel sync pulse width in dualhead mode.
813    * Also, fix a crash when using libpciaccess
814      (the memory wasn't being unmapped).
815
816commit f2c83671cccc42345bfc9b506365936bbb6b9ef0
817Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
818Date:   Fri Oct 24 17:55:02 2008 -0200
819
820    Make Dualhead option functional.
821    
822      Slightly change clock selection code to start using 501 compatible values,
823    before checking 502 values, if it is a 502.
824      DPMS for the VGA/second output was being set with bits inverted, that is
825    dpms-on was programmed as dpms-off and vice versa. This was one of the
826    reasons of dual head not working.
827      Corrected wrong vsync programming for the crt. Cut&paste/typo caused
828    programming vsync with hsync values, and thus, never getting the crt
829    to accept the mode being programmed.
830      If adding:
831      Option "Dualhead" "True"
832    to xorg.conf, now you should be able to do things like:
833    $ DISPLAY=:0.0 xrandr --output VGA --right-of LVDS
834    $ DISPLAY=:0.0 xrandr --output VGA --below LVDS
835    and so on.
836      *Iff* there is some way to not have any limitations for video playback,
837    this option should be made default or automatically configured.
838
839commit 1efe36ed5cf5b0931daa915ca3ce231b78168d87
840Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
841Date:   Wed Oct 22 19:06:06 2008 -0200
842
843    Make UseFBDev option functional again.
844    
845      It is mean't to be a fallback option, that as long as kernel boots
846    in framebuffer mode, the X Server should also work.
847      Correct incorrect value being checked when printing state of Dualhead
848    option.
849      The M value when programming PLL3 is actually an 8 bits integer,
850    so correct it and comments about it.
851
852commit dd083c26de400cee8c67977f46cf90d048b22ff9
853Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
854Date:   Tue Oct 21 19:41:55 2008 -0200
855
856    Use existing "Dualhead" option in MSOC.
857    
858      If Dualhead is set, it will attempt to use setups that allow different
859    outputs with different contents, instead of crt always cloning panel.
860      Still not fully functional.
861
862commit 197b4231586b9ada1c590a01b8dd7a5fcda936e6
863Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
864Date:   Tue Oct 21 14:22:01 2008 -0200
865
866    Crt interface corrections.
867    
868      Correct wrong clock calculation for the 501 clock setting, that is used
869    for pre 502 panel interface and crt.
870      Adds a few new programmable fields to MSOCRegRec:crt_display_ctl.
871      Call SMI501_CrtcAdjustFrame() after adjusting a mode, instead of
872    before changing registers.
873
874commit 71990e0beece6e60ffbffccfdf56e5a71641a120
875Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
876Date:   Mon Oct 20 15:09:09 2008 -0200
877
878    Don't change M1XCLK unless option specified in xorg.conf.
879    
880      Also add more complete description of programmable clocks.
881
882commit 8b7ce66e26113ec5859566e7f3d0fd15d3e29eaa
883Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
884Date:   Fri Oct 17 18:58:45 2008 -0300
885
886    Update sm502 pll3 programming.
887    
888      Also print the value of sm502 registers (should be a noop on sm501).
889      Added the frequency field to the MSOCRegRec's misc_ctl, as it is
890    a read/write registers, and changing bit 24 may change the logic
891    of pll3 programming (needs clarification with SMI).
892
893commit de1b633f916f4fcaaa95b226a8622d37041c86d4
894Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
895Date:   Thu Oct 16 15:41:44 2008 -0300
896
897    Revert/modify some RandR changes to reenable XAA.
898    
899      XAA was not initialized anymore in RandR patch, but it is still the
900    default accel method. Also reenable offscreen fb manager.
901      At least on the MSOC OEM sample (using X Server 1.4), needs some more
902    work as EXA is crashing in a call to exaPrepareAccess() when switching
903    to a terminal; this seems to be related to disabling exa offscreen
904    pixmaps on the smi501.
905
906commit e020afa359434b4da06ef8509a3e504423f824d2
907Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
908Date:   Thu Oct 16 15:11:36 2008 -0300
909
910    Correct incorrect pll3 calculation.
911    
912      Double value must be truncated when calculating the difference from the
913    possible clock and requested clock.
914
915commit f686d1b4fdd11d3d969ff4db21128d7867c9044d
916Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
917Date:   Tue Oct 14 19:05:31 2008 -0300
918
919    Extra MSOC tweaks for the RandR1.2 changes.
920    
921      The "prototype" computer no longer locks, and only "missing feature"
922    from before the patch is that video and Xaa don't work (only with Exa).
923      For the moment, just force crt as not connected, and set the proper
924    registers to make the secondary output clone the panel.
925
926commit fc722212d297a5141245beb7c8ae396606a8f221
927Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
928Date:   Tue Oct 14 15:46:52 2008 -0300
929
930    Remove dependency on xf86cvt.c.
931    
932      It actually was not added in a previous commit (due to using git reset
933    to remake a commit, and not readding it), but now, changed only call
934    xf86CVTMode if it is available in the X Server, otherwise, for older
935    X Servers, a modeline should be specified in xorg.conf.
936
937commit 5adb3b502ef79827b89fb11926b6198c3a57d2e6
938Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
939Date:   Tue Oct 14 14:48:42 2008 -0300
940
941    Minor corrections for smi501 for the randr1.2 integration.
942    
943      Also correct printing information about MCLK in the smi501/502, it
944    should print M2CLK (aka M1CLK o the 502) as it is the clock that
945    controls video engine/memory.
946
947commit 6e0837305533f940adb7a09f9b013077c0888f2d
948Author: Francisco Jerez <currojerez@gmail.com>
949Date:   Sat Oct 11 14:24:48 2008 +0200
950
951    RandR1.2 initial implementation (WIP)
952    
953    Moved most of the Lynx-specific code out of the main
954    functions at smi_driver.c to some new files: smilynx_hw.c
955    (With CRTC-independent code like global hardware initialization and
956    mode saving/restoring), smilynx_crtc.c (With the CRTC-local
957    procedures) and smilynx_output.c (Output power management, DDC and
958    monitor detection, currently).
959    
960      Done something similar with the SMI501 code: split
961    SMI501_ModeInit in three separate functions: SMI501_HWInit that does
962    the global initialization, and the CRTC mode_set callbacks
963    SMI501_ModeSet_crt and SMI501_ModeSet_lcd at smi501_crtc.c.
964    
965      The SMI501_ModeSet code is divided into SMI501_WriteMode_common,
966    SMI501_WriteMode_lcd and SMI501_WriteMode_crt, each one updates a
967    different register set in the hardware.
968    
969      Inside smi_crtc.c, there is mainly hardware independent
970    code... Initial CRT controller allocation, shadows, rotation and
971    framebuffer resizing code.
972    
973      The shadow code currently relies on EXA... I'm not sure what are the
974    problems of the EXA implementation in MSOC. Does it work?
975    
976      I think it would be a good thing to get rid of XAA soon: The next
977    thing I'll be working on will be EXA Composite... I hope it's possible
978    to get EXA acceleration working at least as fast as the current
979    XAA. (Although Teddy Wang confirmed me that the SMI720 DMA engine is
980    broken...)
981    
982      So, the patch adds some features like new Lynx dualhead modesetting
983    code, but it probably breaks some other things:
984    * Video Overlay: I suppose it does work with EXA activated. It seems
985    it is also possible to have simultaneous overlays in both CRTCs with the
986    Lynx hardware (I have some code for this, but still work in progress).
987    * Hardware Cursor (It needs more integration in the CRTC interfaces).
988    * The old Shadow FB / rotation code, which we should probably drop.
989    
990      These are in my TODO list... Along with some more cleaning and other
991    minor issues in the modesetting code.
992    
993      About the SMI501 RandR1.2 implementation... I suppose it's specially
994    lacking per-output DPMS (The DPMS field in the System Control register
995    only affects the CRT, doesn't it?).
996
997commit 605199b1fa80d5a65e11a38270f6d0974466d1dc
998Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
999Date:   Thu Oct 9 18:51:50 2008 -0300
1000
1001    Add a PanelSize/60Hz CVT mode at driver initialization
1002    
1003      This driver is expected to be used with XServer 1.4 or newer, but a
1004    (probably temporary) copy of xf86cvt.c was also added to the build.
1005      This modeline automatically added implies it would not be required
1006    to specify one in xorg.conf, as it should not be uncommon "non standard"
1007    modes, like 800x480 or 1024x600.
1008
1009commit 55052509d75492f0936273686826c51fcbc831c1
1010Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1011Date:   Thu Oct 9 18:12:02 2008 -0300
1012
1013    Make the input frequency in SMI501_FindPLLClock a variable
1014    
1015      Instead of using a fixed 24Mhz input frequency, as "unofficially"
1016    specified by SMI, a guess of using a 12 multiplier was used, and it
1017    corrects the screen flicker problem as well as make a secondary output
1018    work again.
1019
1020commit 61f814d1c8ad9440bab89eda44e64b01ba0eb703
1021Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1022Date:   Wed Oct 8 18:36:11 2008 -0300
1023
1024    Rewrite WaitQueue and WaitIdle accell macros
1025    
1026      WaitQueue() did receive an unused parameter, that was removed. The
1027    parameter was unused since version 1.0 in XFree86 CVS, so probably
1028    nobody knows for what it was used.
1029      WaitIdle() now also replaces WaitIdleEmpty(), as an idle engine
1030    should also have an empty fifo.
1031      SMI_SubsequentScreenToScreenCopy() was changed to use WaitIdle()
1032    instead of WaitQueue() before submitting the commands to the engine.
1033    This fixes a nasty lockup when using XaaOffscreenPixmaps (what is
1034    desired/wanted to help improve performance), and the lockup in the
1035    MSOC was clearly due to an engine overrun, when sending commands
1036    faster then the engine could process.
1037
1038commit 1656fb6de5308ff586676e8d6a8aa8d4579ebf4c
1039Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1040Date:   Fri Oct 3 18:55:14 2008 -0300
1041
1042    Don't use the 1 multiplier on older chipsets.
1043    
1044      It is not in the specs, so don't use it on chipsets older then the 502.
1045      Some review should be done in SMI501_FindPLLClock() as when hardcoding it
1046    to use the clock selection code for older chips, it appears to choose a
1047    better refresh rate (no flicking at 1024x600).
1048
1049commit da789f720e2d772dff92adc10aa024a6a921fbbd
1050Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1051Date:   Fri Oct 3 16:30:23 2008 -0300
1052
1053    Simplify regsmi.h by removing most unused SMI501 defines.
1054    
1055      The values that should be used, should also have a more meaningful name,
1056    as something like:
1057    doesn't help much, and the fact that these "ports" can be accessed from
1058    several "bases" may be confusing also, as there is SCRBase, DCRBase,
1059    FPRBase, DPRBase, CPRBase (and IOBase). The 501 code should probably be
1060    changed to just use IOBase, as currently it uses SCRBase, that happens
1061    to be at pSmi->MapBase + 0x0000.
1062    
1063      This patch also completely transforms the WaitQueue macro in a noop
1064    for the MSOC. In SMI sample code it is almost the current code, just
1065    that the macro body is commented out. Hopefully this will be addressed
1066    soon.
1067
1068commit bacddd1d31a5f3de14477914958f27b6f081c125
1069Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1070Date:   Fri Oct 3 15:10:55 2008 -0300
1071
1072    Rename macro "bitfield" to "bits" and correct a wrong division.
1073
1074commit 22a7fd6711e2a40633981cd21fa3808c9246f13a
1075Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1076Date:   Thu Oct 2 16:11:59 2008 -0300
1077
1078    Remove the IN_SEQ and OUT_SEQ macros.
1079    
1080      These macros access vga registers, but don't make it clear in their name.
1081
1082commit 973f68bf0464c4d2aec2acf7c8e6db201782f4ec
1083Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1084Date:   Thu Oct 2 15:48:51 2008 -0300
1085
1086    Rewrite some macros to not have side effects in if/else nesting.
1087
1088commit 82e17c442c3969938435a8b46d2679995af64425
1089Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1090Date:   Thu Oct 2 15:25:41 2008 -0300
1091
1092    Add support for the extra divider in the alternate pixel clock setting.
1093    
1094      Also remove extra includes that were not used.
1095
1096commit b32769305275519f7f4d1e0871a71514636ca026
1097Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1098Date:   Wed Oct 1 18:27:27 2008 -0300
1099
1100    Correct clock programming for the SMI 501/502
1101    
1102      The new code is based on the file SM502Clock.pdf provide by SMI engineers
1103    (many thanks). And now it is expected to work correctly. The few remaining
1104    details that are unclear (for the extended 502 modesetting mode) should be
1105    resolved soon, and those are:
1106    *  MMIO:0x74:bit15 (PLL Output Divided by 2)
1107    	0: Disable.
1108    	1: Enable.
1109       (does this mean it can use a 12MHz clock instead of 24? or anything else?)
1110    *  Should bit 31 of "current clock" always be set when using the alternate
1111       mode setting for the 502? The alternate modesetting allows a very closer
1112       pixel clock programming option, usually with 0 difference from the
1113       xf86 mode.
1114    
1115      I will ask SMI to make SM502Clock.pdf also available in the ftp site, it
1116    is just 3 pages, or possibly even better, an update for the existing
1117    documentation.
1118
1119commit ea34e20852ee6007e2f0fcacd5068c151865be1e
1120Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1121Date:   Tue Sep 30 17:49:25 2008 -0300
1122
1123    Remove the macro field, and rename the detail structure to f.
1124    
1125      Now access to bitfields looks like:
1126    record->register.f.field
1127      instead of:
1128    field(record->register, field)
1129      what should be easier to read and understand.
1130
1131commit a91872f813a6299f32b9f833876c54afd0b53aaa
1132Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1133Date:   Tue Sep 30 17:33:36 2008 -0300
1134
1135    Update to match the SMI 502 chipset specs.
1136    
1137      The 502 is a superset of the 501, also recognized as PCI_CHIP_SMI501.
1138      The driver is only known to work with smi_501 chipset
1139    "SM501 VoyagerGX Rev. AA (rev c0)".
1140
1141commit e830572bca9b8eaf50a36a12d1fe31cdfec6280e
1142Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1143Date:   Fri Sep 26 19:43:45 2008 -0300
1144
1145    Don't always program CRT clock and registers.
1146    
1147      This causes instabilities, and there was also a bug where it would
1148    tell the crt to not select panel data (due to removing the "else").
1149
1150commit 314eecb6628840a942ff2e76d8c7402301feac77
1151Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1152Date:   Fri Sep 26 19:30:17 2008 -0300
1153
1154    Properly check pScrn->driverPrivate before deferencing it.
1155    
1156       Add missing divide by 1000 in smi_501.c, as well as a FIXME comment
1157    about a crash, and possible cause as well as possible ways to program
1158    other clocks.
1159
1160commit f18222ffcff8e78b3cfcd220cad06a1daa3fc920
1161Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1162Date:   Fri Sep 26 17:55:42 2008 -0300
1163
1164      Don't try to find the closest clock, just use highest one.
1165    
1166      Previous patch wasn't fully correct, just that the values were generating
1167    the same register dump as the smi 2.2.5 driver version.
1168      When correcting it to not mix Mhz and Khz in the math, the problems
1169    returned.... Anyway, it just works when selecting the highest clock, and
1170    the attached CRT will display the panel image. This is illogical, so maybe
1171    the specs are outdated... The previous code has been #ifdef'ed out in
1172    "CALC_CLOCK".
1173      Note that it still changes the boot default clock, from the default
1174    288Mhz to 336Mhz, and in this patch, it also always sets both, panel and
1175    crt clocks (what happens here, is that most likely, by doing this, it
1176    forces the hardware to "somehow" autoconfigure itself; will try to get
1177    some information from SMI about this, but this is the boot default,
1178    and what smi sources do...).
1179
1180commit bdba7c1397186cd15d519db55c2345db44990be4
1181Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1182Date:   Wed Sep 24 18:58:35 2008 -0300
1183
1184    Fix incorrect understanding of the pixel clock from specs.
1185    
1186      The clock that should be programmed is the bandwidth, and not the pixel
1187    clock itself. Based on analysis of register contents, after "importing"
1188    a slightly modified version of xf86ModeBandwidth() and properly converting
1189    values, now it properly programs both, the LCD and a "clone" CRT or just
1190    a secondary panel.
1191    
1192      Modes generated by either gtf or cvt correctly works.
1193    
1194      A (hopefully temporary) fallback was also added, and if the option
1195    "UseFBDev" is used, the driver will not attempt to set/modify the
1196    current video mode.
1197    
1198      TODO: Generate a modeline "on the fly", based on monitor configured
1199    specs, or some set of defaults, so that it should not be required to
1200    specify a modeline in xorg.conf.
1201
1202commit e64b3d63be2a4148c4d10ac160e2bd2f3e8c7fd3
1203Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1204Date:   Mon Sep 22 22:01:16 2008 -0300
1205
1206    Correct problems in clock setting.
1207    
1208      One clock must be changed at a time, first setting the pll value,
1209    then waiting for 16ms (one vsync), then setting the divider/shift
1210    values, and again waiting 16ms.
1211      Code was working after split of SMI501_ModeInit(), because some
1212    clocks were already using system boot default, but would most likely
1213    cause a crash when actually changing values (currently only p2_xxx
1214    or v2_xxx is changed, but it would cause problems on some system where
1215    the initial values don't match the ones set by the driver).
1216    
1217      TODO: Either don't change m2clk and mclk or ensure other values
1218    work correctly (don't change because the kernel should already
1219    have set those if a value other then the boot default should be
1220    used).
1221
1222commit 1f0af5ea50e3e3b2c9ed882f7ef2fb8871e780bd
1223Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1224Date:   Mon Sep 22 21:41:12 2008 -0300
1225
1226    Correct logic in sw cursor handling and add missing entries to .gitignore.
1227
1228commit 862cbd8c1845b54a38996b92e8e22d153f38cdd2
1229Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1230Date:   Mon Sep 22 19:09:12 2008 -0300
1231
1232    Split SMI501_ModeInit in two functions.
1233    
1234      This allows using alternate MSOCRegRec structures to set a mode.
1235    This should be useful to restore the initial mode.
1236
1237commit bcbf1854be851b8fa1ec3f384b11c312b0533bf3
1238Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1239Date:   Mon Sep 22 18:29:31 2008 -0300
1240
1241    Remove unused .cvsignore files.
1242
1243commit 1014ffe0a2c55119d6bfbcf632f3d4ebe77af212
1244Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1245Date:   Mon Sep 22 18:15:21 2008 -0300
1246
1247    Simplify hw cursor and sw cursor option handling.
1248    
1249      Also rename SMIRec field hwcursor to HwCursor, as there are few usages,
1250    and this way, match the case usage for other field names.
1251
1252commit 6b5c22d8680573c8a6b259d78ba3c8435514fde9
1253Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1254Date:   Mon Sep 22 17:05:21 2008 -0300
1255
1256    Rework/simplify debug macros.
1257    
1258      Instead of cut&paste of the name of the current function everywhere,
1259    just use cpp's __FUNCTION__ predefined macro.
1260      Create two macros to exit a function, named LEAVE() and RETURN().
1261    Functions returning void should call LEAVE() and then explicitly return
1262    for now. "Logged" function calls are indented, so a review was done to
1263    ensure functions with a ENTER() also have the proper exit macro.
1264      The DEBUG macro was changed to have variadic arguments, and this way
1265    it is no longer required to prefix arguments with VERBLEV, but now it
1266    also is not possible to use another "verbosity value", but it wasn't used
1267    in any of the DEBUG macro calls.
1268
1269commit cf7097c19b31671a53bc8161bd29b9f79f1d3d85
1270Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1271Date:   Mon Sep 22 14:25:04 2008 -0300
1272
1273    Fix a leak and minor cosmetic change.
1274    
1275      Leak was due to only releasing return value of xf86GetEntityInfo
1276    if the chipset was PCI_CHIP_SMI501.
1277      Cosmetic change was to log pci burst and pci retry default values
1278    as X_DEFAULT instead of X_PROBED.
1279
1280commit ee1e72a248b2df6522283c0dd3e5433d7f961c62
1281Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1282Date:   Fri Sep 19 18:06:20 2008 -0300
1283
1284    Kludge to not lock the SMI 501 when running at 8bpp.
1285
1286commit 9ae840a358959615020fa4adb76c1487231927f3
1287Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1288Date:   Fri Sep 19 18:00:16 2008 -0300
1289
1290    Add MSOC palette support to run at 8 bpp.
1291    
1292      Colors should be correctly set, but most output using Render is not
1293    properly drawn, i.e. anti aliased fonts, etc. Need to investigate
1294    further, but it should be a general problem, not exclusive to the
1295    driver.
1296
1297commit f4339a6a7aa077cb355dd23b77ea568ad2a481e3
1298Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1299Date:   Fri Sep 19 16:52:17 2008 -0300
1300
1301    Enable pci retry and pci burst by default.
1302    
1303      Rename fields pci_burst and NoPCIRetry to PCIBurst and PCIRetry,
1304    to match field names case usage, and updated logic for the removal
1305    of "No" from the pci retry option.
1306      Pci retry and pci burst greatly increase video playback speed,
1307    allowing smooth video play on the SMI 501.
1308
1309commit 782ef31a9273e33100ae6b4bd20a4d1d3f493aa4
1310Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1311Date:   Fri Sep 19 15:48:34 2008 -0300
1312
1313    Add initial exa support for SMI501.
1314    
1315      This just allows exa to run, with a minimal ammount of features avaiable.
1316
1317commit 5d9ab8195e692f78dd7e57e7d17d8b46f7f59221
1318Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1319Date:   Thu Sep 18 18:27:51 2008 -0300
1320
1321    Add a missing CHECK_SECONDARY macro call.
1322
1323commit 2876b55c3ed82b9be023acbfc4ddf993494c3c9b
1324Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1325Date:   Thu Sep 18 18:23:37 2008 -0300
1326
1327    Correct video offscreen memory allocation routines.
1328    
1329      It was always failing in the first allocation, if using XAA, what caused
1330    some players to not work correct. This is a major problem for the MSOC as
1331    it still only supports XAA.
1332      Also add a workaround for a hardware problem, as described in
1333    "SM501 Rev.AA engineering addendum v0.1.pdf"
1334    "A-27 PIXELS TOO BRIGHT ON VIDEO SHRINK".
1335
1336commit 79297e797de3497773539eb9f89e8594797ee512
1337Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1338Date:   Thu Sep 18 18:16:21 2008 -0300
1339
1340    Update for new smi_501 interfaces.
1341    
1342      This also adds some simplification for the driver code, by adding
1343    some function pointers to avoid the need to check the chipset everywhere.
1344      The new SMI_DetectMCLK function doesn't actually change pSmi->MCLK to
1345    avoid chances of it stoping working on non MSOC chipsets, as it
1346    previously not did set it, and for the MSOC, for the moment, default to
1347    whatever value is already set, i.e. doesn't reprogram it.
1348      Added new memory detection code for the MSOC, that better matches the
1349    documentation.
1350
1351commit 95e312b712a1e4a476ef31c5302faf77c22915a9
1352Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1353Date:   Thu Sep 18 18:10:54 2008 -0300
1354
1355      Complete rewrite of smi_501.c and smi_501.h.
1356    
1357      The previous version was dependant on kernel framebuffer, as it was
1358    just failing a test if not having a modeline defined, and in the test
1359    case, it is running at 1024x600.
1360      Now it properly programs the video hardware, and the procedure is
1361    expected to be very well documented.
1362
1363commit cf5132907e4709c872089fe2d79837b2de4e45b7
1364Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1365Date:   Tue Sep 2 18:16:41 2008 -0300
1366
1367    Rename global smi501 functions to have SMI501 prefix.
1368    
1369      Functions renamed were setDPMS -> SMI501_SetDPMS,
1370    regRead32 -> SMI501_Read32 and regWrite32 -> SMI501_Write32.
1371      Also moved SMI_SetDPMS to start of smi_501.c to have extern
1372    functions at start of the file.
1373
1374commit a2a090e10f031aed6dbf6d143d6ce8abdf7a2392
1375Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1376Date:   Tue Sep 2 17:55:16 2008 -0300
1377
1378    Correct all compiler warning messages.
1379    
1380      Code compiled with:
1381    gcc -Wall -Wbad-function-cast -Wdeclaration-after-statement \
1382        -Wmissing-prototypes -Wmissing-declarations -Wnested-externs \
1383        -fno-strict-aliasing -Wold-style-definition -Wpointer-arith \
1384        -Wstrict-prototypes
1385    
1386      This patch also makes most smi_501.c functions static, as they are not
1387    called from anywhere else.
1388      Also removed prototypes for non existent functions in smi_501.h.
1389
1390commit 2d5ba5f7310fa8e11f7349d116aa871a239ac35f
1391Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1392Date:   Tue Sep 2 16:32:31 2008 -0300
1393
1394    Correct xv video problems on MSOC.
1395    
1396      Set VIDEO_CLIP_TO_VIEWPORT as it is used in MSI sources, and also add
1397    proper code to detect panel size, as this information is used to manage
1398    offscreen memory (but shouldn't it be using pScrn->virtualY, etc?)
1399
1400commit e111c85d6cca3f2206b6bce119b284edccf0cd4f
1401Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1402Date:   Mon Sep 1 18:41:43 2008 -0300
1403
1404    Don't pretend this driver compiles on XFree86.
1405    
1406      Either remove check for XF86_VERSION_CURRENT or at least add the
1407    implementation of RegionsEqual() to smi_video.c.
1408
1409commit b23c5c5e434aaba2128abcae61a8564a62d2e7a1
1410Author: Francisco Jerez <currojerez@gmail.com>
1411Date:   Mon Sep 1 21:28:13 2008 +0200
1412
1413    Fix XAA, ShadowFB and VT switching for non-sm501 chipsets
1414    
1415    Signed-off-by: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1416
1417commit 6a01c87b7c75f698ff2b783f8d341526159ebfe2
1418Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1419Date:   Mon Sep 1 18:00:03 2008 -0300
1420
1421    Split SMI_MapMem in SMI_MapMem and SMI_MapMmio
1422    
1423      This is required because MSOC needs mmio to detect ammount of memory.
1424      The patch also creates a SMI_DetectMem() function, that should slightly
1425    simplify SMI_PreInit().
1426
1427commit c45ef4676812f72dcff53ba85e9fa95836704c75
1428Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1429Date:   Fri Aug 29 19:33:31 2008 -0300
1430
1431    Bump version to 1.6.1.
1432
1433commit 4eee032811d1487df9dcf94708b1bcafc7e63392
1434Author: Arnaud Patard <apatard@mandriva.com>
1435Date:   Fri Aug 29 19:33:31 2008 -0300
1436
1437    Correct a problem when handling i420 format.
1438    
1439    The problem was that the driver was converting i420->YV12 2 times
1440    instead of only once.
1441
1442commit 76c931e149917563485ae6ad2950aa28a531b0d1
1443Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1444Date:   Fri Aug 29 19:31:57 2008 -0300
1445
1446    Update MSOC video interface.
1447    
1448      This synchronizes smi_video.c with smi 2.2.5 sources.
1449
1450commit 0eb35f2065314807a604a44e571bbb898db23543
1451Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1452Date:   Fri Aug 29 19:20:54 2008 -0300
1453
1454    Update xaa and generic acceleration code for the MSOC.
1455    
1456      Currently, the support for EXA is not functional, neither is
1457    ShadowFB or bpp/depth other than 16.
1458
1459commit 375e35fc7df4a21c2e0bc933774cc2f5f29b89d1
1460Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1461Date:   Fri Aug 29 19:36:35 2008 -0300
1462
1463    MSOC doesn't access VGA registers or VBE/INT10
1464    
1465      This should be the biggest patch in the "merge". Done in a single chunk
1466    to have it compilable at all stages.
1467
1468commit 164d921465ae7b3a0fa2551358f7c0b5a18a425a
1469Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1470Date:   Fri Aug 29 18:47:14 2008 -0300
1471
1472    Add code to probe and recognize the SMI501 chipset
1473
1474commit db330497c3b757361ac9c3602b3a4ece0a401dd3
1475Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1476Date:   Fri Aug 29 15:58:16 2008 -0300
1477
1478    Change SILICONMOTION_NAME value
1479    
1480      xf86VDrvMsgVerb prints up to 14 characters prefix, where prefix has the
1481    format "%s(%d): " so, use name "SMI" instead of "Silicon Motion"
1482
1483commit 36309132aef067822a8901ce24fda817be58546d
1484Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1485Date:   Fri Aug 29 15:52:36 2008 -0300
1486
1487    Add initial support and macros for the MSOC.
1488    
1489    This patch add the new files and basic required definitions.
1490
1491commit 96166eafc2b0e5db627aa4f8b0422f953837772b
1492Author: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1493Date:   Fri Aug 29 14:48:44 2008 -0300
1494
1495    Fix build for removal of xf86Version.h
1496    
1497      This is the first patch of series to merge code from the version 2.2.5
1498    from SiliconMotion and code from
1499    http://cgit.freedesktop.org/~agd5f/xf86-video-smi501
1500    
1501      After the merge, work will continue, to correct some extra problems,
1502    and make it work with EXA at least (as only the Xorg version has
1503    EXA support).
1504
1505commit 682b9a1b97a82c7d084c63860a8c753cc45d9faf
1506Author: Francisco Jerez <currojerez@gmail.com>
1507Date:   Sat Aug 16 13:24:27 2008 +0200
1508
1509    Allow using XV and RandR rotation simultaneously.
1510    
1511     As it doesn't seem feasible to rotate the video, I did some modifications
1512    at SMI_PutImage and SMI_ClipVideo to make PutImage work (incorrectly) with
1513    rotation enabled. The image is displayed unrotated.
1514    
1515    Signed-off-by: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1516
1517commit f6110b1ef5934b45812ee95e70b1e64f44386dda
1518Author: Francisco Jerez <currojerez@gmail.com>
1519Date:   Sat Aug 16 13:08:25 2008 +0200
1520
1521    Some fixes in the EXA UTS/DTS code.
1522    
1523     In DTS:
1524      * It uses the screen Bpp inestead of the pixmap Bpp...
1525        this gives some problems when using pixmaps with different depth
1526        to the screen.
1527     In UTS:
1528      * aligned_pitch was computed from src_pitch inestead of the
1529        pixmap width.
1530      * When writing the target coordinates to the DE registers, it does
1531        y*0xFFFF inestead of y & 0xFFFF.
1532      * I renamed source_pitch to src_pixelpitch as it is very confusing
1533        to have src_pitch and source_pitch.
1534      * It isn't necessary to call WaitQueue before copying each scanline,
1535        but it seems it almost doesn't affect performance.
1536    
1537    Signed-off-by: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1538
1539commit 209097ba5b44a0ce0da7f1ea52150dcace2b5244
1540Author: Francisco Jerez <currojerez@gmail.com>
1541Date:   Sat Aug 16 13:00:31 2008 +0200
1542
1543    RandR rotation implemented.
1544    
1545    * I added the configuration file option "RandRRotation".
1546    * I replaced pSmi->ShadowPitch with pSmi->screenStride, it seems it
1547       makes more sense because the lower word of ShadowPitch may change
1548       independently.
1549    * I moved the SMI_DEDataFormat to smi_accel.c because it seems it is
1550      a piece of code repeated many times in the driver.
1551    * At some places, it is assumed the framebuffer is at FBOffset:
1552      when using a shadow framebuffer, FBOffset is the location of the
1553      on-screen framebuffer (0 should be used). This made e.g. EXA
1554      completly useless with ShadowFB enabled (it crashed).
1555    * In the FBManager initialization, I have replaced xf86InitFBManager
1556      with xf86InitFBManagerRegion to reserve some additional space as
1557      screen fb: it's unlikely to be the case, but a less efficient
1558      alignment in the rotated mode could make the rotated mode need more
1559      memory than the unrotated one.
1560      This is not a problem with EXA as the offscreen memory parameters
1561      can be easily modified when doing the rotation.
1562    * In SMI_RefreshArea it's assumed that some DE registers are already
1563      in some state, this is specially not true when using EXA.
1564    * SMI_ValidMode rejects a rotated mode with different dimensions
1565      than the panel. This seems to work now.
1566    
1567    Signed-off-by: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1568
1569commit 45c6aedd6fe4e4d6abe58d1bb39ec01049404f09
1570Author: Francisco Jerez <currojerez@gmail.com>
1571Date:   Sat Aug 16 12:50:26 2008 +0200
1572
1573    Updates in SMI_EnterVT when remapping memory.
1574    
1575      Update the screen pixmap header with the new aperture address
1576    when remapping framebuffer memory in SMI_EnterVT.
1577      Update pSmi->EXADriverPtr->memoryBase too, in case EXA is
1578    being used.
1579    
1580    Signed-off-by: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1581
1582commit 8feca790a6e92799019237ac69a4ef618cacfaae
1583Author: Francisco Jerez <currojerez@gmail.com>
1584Date:   Sat Aug 16 12:40:52 2008 +0200
1585
1586    Make the int10/VBE initialization depend on the UseBIOS configuration option.
1587    
1588    Signed-off-by: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1589
1590commit 772e899e71efab208b844b09c75f9f49ff453386
1591Author: Francisco Jerez <currojerez@gmail.com>
1592Date:   Sat Aug 16 12:25:18 2008 +0200
1593
1594    Memory detection moved before memory mapping in SMI_PreInit.
1595    
1596    Take into account the VideoRam configuration option.
1597    
1598    Signed-off-by: Paulo Cesar Pereira de Andrade <pcpa@mandriva.com.br>
1599
1600commit d28ed6aa0d28fdcf8f555b2bb8a58fe28d1b464b
1601Author: Adam Jackson <ajax@redhat.com>
1602Date:   Fri Aug 15 14:24:31 2008 -0400
1603
1604    Dead code removal.
1605
1606commit 7cfecd313aaea74ad5675c53b4d0099d6e9dec39
1607Author: Adam Jackson <ajax@redhat.com>
1608Date:   Wed Mar 19 15:26:14 2008 -0400
1609
1610    siliconmotion 1.6.0
1611
1612commit c0e80b14a8af7c9df44919b2df58030e59ecbf88
1613Author: Adam Jackson <ajax@redhat.com>
1614Date:   Wed Mar 19 15:25:23 2008 -0400
1615
1616    Fix distcheck
1617
1618commit e53d07525952e1281170db01c48d7a14a6983049
1619Author: Adam Jackson <ajax@redhat.com>
1620Date:   Wed Mar 19 15:24:11 2008 -0400
1621
1622    Death to RCS tags.
1623
1624commit 46e68bf899754f6f70a0f5a7b0d42c5e208aaa5a
1625Author: Dave Airlie <airlied@redhat.com>
1626Date:   Mon Mar 10 13:53:00 2008 +1000
1627
1628    smi: add pciaccess support
1629
1630commit 75182ca115cba8fcf31304aea88ca3d4b9aa66ac
1631Author: Matthieu Herrb <matthieu.herrb@laas.fr>
1632Date:   Sat Mar 8 23:49:58 2008 +0100
1633
1634    Makefile.am: nuke RCS Id
1635
1636commit ae556406ab6c4de410b68b3c8bcebb434b3efc9d
1637Author: Brice Goglin <bgoglin@debian.org>
1638Date:   Tue Aug 7 10:00:25 2007 +0200
1639
1640    Define SILICONMOTION_VERSION* using PACKAGE_VERSION*
1641
1642commit 1a803a8f91a931c00106f9d3d41cfa5d74c19f55
1643Author: Alex Deucher <alex@botch2.com>
1644Date:   Tue Mar 20 21:08:38 2007 -0400
1645
1646    minor fixup
1647
1648commit 503027d38dab759277c0a67746ac205a1dfa9277
1649Author: Dennis De Winter <dennis.de.winter@gmail.com>
1650Date:   Fri Mar 16 21:29:19 2007 -0400
1651
1652    Fix UTS issues with flash
1653    
1654    Also remove safety padding on memory manager calculation
1655
1656commit fe4114033f2853e2df8d37c490deba26bf0366e8
1657Author: Alex Deucher <alex@botch2.com>
1658Date:   Thu Mar 15 02:11:24 2007 -0400
1659
1660    hostdata blit UTS works, but SW is faster
1661    
1662    UTS is working, but SW is faster so leave it disabled
1663
1664commit 62442ce9e62b484e4f2067e3c52c153cd356be2c
1665Author: Alex Deucher <alex@botch2.com>
1666Date:   Wed Mar 14 20:13:28 2007 -0400
1667
1668    more clock fixes
1669
1670commit 6ff1241636e25d2c0f20bd14e92c8585f209aeb0
1671Author: Alex Deucher <alex@botch2.com>
1672Date:   Tue Mar 13 02:32:52 2007 -0400
1673
1674    bump to 1.5.1
1675
1676commit b53db3bc5ba7648d392dc630ef47e57fb8a0afb4
1677Author: Alex Deucher <alex@botch2.com>
1678Date:   Tue Mar 13 02:31:30 2007 -0400
1679
1680    fix clk calculation on older lynx chips
1681
1682commit 7d1392be291e99fd3dde2292b82423b5f78d3a21
1683Author: Alex Deucher <alex@botch2.com>
1684Date:   Tue Mar 13 00:44:29 2007 -0400
1685
1686    bump to 1.5.0
1687
1688commit 449fedec0e38289d10d75799bbf995b0e7189911
1689Author: Alex Deucher <alex@botch2.com>
1690Date:   Tue Mar 13 00:13:35 2007 -0400
1691
1692    fix Solid()
1693    
1694    Unforunately, for 32 bpp solid fills the HW ignores alpha.
1695
1696commit 6d40f87d6cddde2d88f25be686b6392449cdecf1
1697Author: Alex Deucher <alex@botch2.com>
1698Date:   Mon Mar 12 23:58:41 2007 -0400
1699
1700    remove spurious waitqueue() line
1701
1702commit 452cabd6e3870b832cb19c8ad8a1fce27eef9c48
1703Author: Alex Deucher <alex@botch2.com>
1704Date:   Mon Mar 12 01:11:16 2007 -0400
1705
1706    Update copyright.
1707
1708commit 55266778ac693582131acae80def32211e09f18f
1709Author: Alex Deucher <alex@botch2.com>
1710Date:   Mon Mar 12 01:02:53 2007 -0400
1711
1712    switch the engine to quick start mode
1713    
1714    Rather than writing to the engine command register for every
1715    operation, set quick start mode, and the operation executes
1716    automatically when the width register is written.
1717
1718commit 4b37b514907da4a567261569eaebcc5a35aff6db
1719Author: Alex Deucher <alex@botch2.com>
1720Date:   Mon Mar 12 00:12:23 2007 -0400
1721
1722    fixup waitqueue slots
1723
1724commit 493c93de288437501fd898800bce45300379ebc3
1725Author: Alex Deucher <alex@botch2.com>
1726Date:   Sun Mar 11 18:49:48 2007 -0400
1727
1728    more exa fixups, disable UTS for now.
1729
1730commit c1d6373624428c8c52e4d1d295b1d77ed75391a0
1731Author: Alex Deucher <alex@botch2.com>
1732Date:   Sun Mar 11 16:22:05 2007 -0400
1733
1734    engine likes to use blit command for solid fills
1735
1736commit 9db380c99decdb9ab7fbf481788a5b63a84cd6e0
1737Author: Alex Deucher <alex@botch2.com>
1738Date:   Sun Mar 11 15:53:29 2007 -0400
1739
1740    fix postscalar shift
1741
1742commit b6082e628289221864e96ebc41ce93d8efe2d2ce
1743Author: Alex Deucher <alex@botch2.com>
1744Date:   Sun Mar 11 14:09:45 2007 -0400
1745
1746    support 32 bpp
1747
1748commit 3ebbc2e55a1584e01949ec760501ac4a3c086b04
1749Author: Alex Deucher <alex@botch2.com>
1750Date:   Sun Mar 11 13:40:06 2007 -0400
1751
1752    fix up vclk1 setup
1753
1754commit 3e62ec8736ea9ea64a2c0c5c41df9c6d742397ca
1755Author: Alex Deucher <alex@botch2.com>
1756Date:   Tue Mar 6 02:42:52 2007 -0500
1757
1758    revert the last commit.
1759    
1760    FBBase already take the register offset into account.
1761
1762commit 7da75a99159f80f215ee409377128f7dcbc4064d
1763Author: Alex Deucher <alex@botch2.com>
1764Date:   Tue Mar 6 02:38:02 2007 -0500
1765
1766    fix memoryBase
1767    
1768    pSmi->FBOffset is the offset from the start of the framebuffer
1769    pScrn->fbOffset is the aperture offset where the framebuffer
1770    starts
1771
1772commit ba0b2492269ea81e91f7ce4fb8da618187a3997f
1773Author: Alex Deucher <alex@botch2.com>
1774Date:   Tue Mar 6 02:24:08 2007 -0500
1775
1776    first pass at UTS
1777
1778commit e650ca36da5cf7614e4f1a0a524eadb46b1cc21d
1779Author: Alex Deucher <alex@botch2.com>
1780Date:   Mon Mar 5 23:58:25 2007 -0500
1781
1782    fix up EXA code
1783
1784commit c5b23ba581a3a4395b528d4ed824d5df986527e4
1785Author: Dennis De Winter <de.winter.dennis@gmail.com>
1786Date:   Sun Mar 4 23:17:57 2007 -0500
1787
1788    Initial import of EXA support
1789    
1790    Based on Dennis' code from bug 8721 with some minor
1791    changes by me.
1792
1793commit cf83ef7685609cd9daea193b4c842d232edcf506
1794Author: Alex Deucher <alex@botch2.com>
1795Date:   Sun Mar 4 17:18:10 2007 -0500
1796
1797    actually calculate and use vclk1
1798
1799commit 3030e5a6afda5d05ddee5b586295a9e608b7a0de
1800Author: Alex Deucher <alex@botch2.com>
1801Date:   Sun Mar 4 02:45:55 2007 -0500
1802
1803    Whitespace and formatting changes
1804
1805commit 606971fc8a323b0b33bb43524941b4bae6a0c334
1806Author: Alan Coopersmith <alan.coopersmith@sun.com>
1807Date:   Tue Feb 27 13:04:31 2007 -0800
1808
1809    renamed: .cvsignore -> .gitignore
1810
1811commit 4b3d28435d2f046d48a38694d89166fe38b00597
1812Author: Alan Coopersmith <alan.coopersmith@sun.com>
1813Date:   Tue Feb 27 13:04:01 2007 -0800
1814
1815    Replace XFree86 references in man page
1816
1817commit 3ff1d1aaf93df90f3a96b87bd8447db44b1722f5
1818Author: Alex Deucher <alex@samba.(none)>
1819Date:   Wed Jan 10 21:58:47 2007 -0500
1820
1821    only idle the engine on mode switches
1822    
1823    on the initial mode set the engine doesn't idle
1824    for some users.
1825
1826commit 72c1d44182a49dbd502bbdc0b48f69a43ad5aed0
1827Author: Adam Jackson <ajax@benzedrine.nwnk.net>
1828Date:   Fri Oct 13 18:41:44 2006 -0400
1829
1830    Bump to 1.4.2
1831
1832commit 36bbaaf3f44321fc7bb8ff1b71ff1d7f87f3b19b
1833Author: Alex Deucher <alex@samba.(none)>
1834Date:   Sat Sep 16 15:48:54 2006 -0400
1835
1836    make siliconmotion build again.  somehow a bunch of defines in
1837    regsmi.h got lost?  commit 9a58bc9adbc982680dd8df27caeb9a5819f2adad looks
1838    like the culprit, but siliconmotion has built fine subsequent to that commit.
1839    Not sure exactly what's going on.
1840
1841commit c090b2d97dbedfc278a4691b70cf3f73f9799aa6
1842Author:  <alex@samba.(none)>
1843Date:   Thu Jul 20 23:35:30 2006 -0400
1844
1845    - add dualhead support for lynx chips (pretty limited due to severe hardware
1846    limitations -- I'll improve it a bit at some point)
1847    - small driver clean ups
1848
1849commit 845d45cd2066be53f6e2e53d5af6375045dac6e2
1850Author: Adam Jackson <ajax@nwnk.net>
1851Date:   Sat Apr 8 01:43:52 2006 +0000
1852
1853    Bump to 1.4.1 for Xv changes.
1854
1855commit 2951e947c56909c9d51bf3f937502d0f682b1d76
1856Author: Aaron Plattner <aplattner@nvidia.com>
1857Date:   Fri Apr 7 23:13:59 2006 +0000
1858
1859    Add a DrawablePtr argument to the XV functions to pave the way for
1860        redirected video.
1861
1862commit 9a316eaea700e3a356f922506bf2c79e00fd9d83
1863Author: Adam Jackson <ajax@nwnk.net>
1864Date:   Fri Apr 7 20:58:57 2006 +0000
1865
1866    Unlibcwrap. Bump server version requirement. Bump to 1.4.0.
1867
1868commit c394f38c9a998eebec134a9a77ced46b9c005b9d
1869Author: Luc Verhaegen <libv@skynet.be>
1870Date:   Wed Feb 8 16:52:34 2006 +0000
1871
1872    SiliconMotion: Clean up insane pScrn->clock usage.
1873
1874commit 2453cfe259dec826fceab3cff1b2c3803d3d8042
1875Author: Kevin E Martin <kem@kem.org>
1876Date:   Wed Dec 21 02:30:02 2005 +0000
1877
1878    Update package version for X11R7 release.
1879
1880commit e398882fba13cd081916d2ff04ead4f9023c9bfb
1881Author: Adam Jackson <ajax@nwnk.net>
1882Date:   Mon Dec 19 16:25:54 2005 +0000
1883
1884    Stub COPYING files
1885
1886commit 818baab410deef2b3a3b6653dd2134dea29e98de
1887Author: Kevin E Martin <kem@kem.org>
1888Date:   Thu Dec 15 00:24:19 2005 +0000
1889
1890    Update package version number for final X11R7 release candidate.
1891
1892commit 6c95cd7b0ac0c113c25bf985365f3ccb1d30d7c9
1893Author: Kevin E Martin <kem@kem.org>
1894Date:   Tue Dec 6 22:48:37 2005 +0000
1895
1896    Change *man_SOURCES ==> *man_PRE to fix autotools warnings.
1897
1898commit 10d5542f95e86463b96cc1fd31378473c2394bc0
1899Author: Kevin E Martin <kem@kem.org>
1900Date:   Sat Dec 3 05:49:35 2005 +0000
1901
1902    Update package version number for X11R7 RC3 release.
1903
1904commit 8180ea0cb0df275831b900836d9387bb2a730de4
1905Author: Kevin E Martin <kem@kem.org>
1906Date:   Fri Dec 2 02:16:10 2005 +0000
1907
1908    Remove extraneous AC_MSG_RESULT.
1909
1910commit 73e22f89e205a61c315ffaca63cc7f5ab1f2df5f
1911Author: Adam Jackson <ajax@nwnk.net>
1912Date:   Tue Nov 29 23:30:03 2005 +0000
1913
1914    Only build dlloader modules by default.
1915
1916commit 8b156b75c37fb642d6c680cc9fb5361fb0a8ff74
1917Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
1918Date:   Mon Nov 28 22:04:10 2005 +0000
1919
1920    Change *mandir targets to use new *_MAN_DIR variables set by xorg-macros.m4
1921        update to fix bug #5167 (Linux prefers *.1x man pages in man1 subdir)
1922
1923commit 8ad5ff0aa74eabc5b4dca2e0422d36713415b1b1
1924Author: Eric Anholt <anholt@freebsd.org>
1925Date:   Mon Nov 21 10:49:18 2005 +0000
1926
1927    Add .cvsignores for drivers.
1928
1929commit 0d4f340f197b14dacc5c33d82f0094ee1f9bb2a8
1930Author: Kevin E Martin <kem@kem.org>
1931Date:   Wed Nov 9 21:15:20 2005 +0000
1932
1933    Update package version number for X11R7 RC2 release.
1934
1935commit 280644786854eea8e8dbe7873e6c79d18f8048f0
1936Author: Kevin E Martin <kem@kem.org>
1937Date:   Tue Nov 1 15:08:54 2005 +0000
1938
1939    Update pkgcheck depedencies to work with separate build roots.
1940
1941commit bcc4f2555d4e4fa7ba2cf022165d3dd8ef331cd9
1942Author: Kevin E Martin <kem@kem.org>
1943Date:   Wed Oct 19 02:48:04 2005 +0000
1944
1945    Update package version number for RC1 release.
1946
1947commit 2da8c99697874a012dccfc582561a0b89e111ecd
1948Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
1949Date:   Tue Oct 18 00:01:54 2005 +0000
1950
1951    Use @DRIVER_MAN_SUFFIX@ instead of $(DRIVER_MAN_SUFFIX) in macro
1952        substitutions to work better with BSD make
1953
1954commit 9f9d3fcd3dca04b7d960d8777e8f8ba612fa1510
1955Author: Adam Jackson <ajax@nwnk.net>
1956Date:   Mon Oct 17 22:57:35 2005 +0000
1957
1958    More 1.7 braindamage: define EXTRA_DIST in terms of @DRIVER_NAME@ instead
1959        of indirectly
1960
1961commit 286bf9de5baebd75ca129e50af9dc5f66dbf4e4f
1962Author: Alan Coopersmith <Alan.Coopersmith@sun.com>
1963Date:   Mon Oct 17 00:09:05 2005 +0000
1964
1965    Use sed & cpp to substitute variables in driver man pages
1966
1967commit 0db428f930db99d70897ca90b9ec0e130f4e6635
1968Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk>
1969Date:   Tue Sep 27 22:04:08 2005 +0000
1970
1971    Add some more documentation files for siliconmotion driver
1972
1973commit bdc896858ed819dcb79bb7443bfa57d63c8e0fad
1974Author: Daniel Stone <daniel@fooishbar.org>
1975Date:   Thu Aug 18 09:03:48 2005 +0000
1976
1977    Update autogen.sh to one that does objdir != srcdir.
1978
1979commit a3add2e05412767c49ca1c0a1069cd5805ea09e9
1980Author: Kevin E Martin <kem@kem.org>
1981Date:   Fri Jul 29 21:22:44 2005 +0000
1982
1983    Various changes preparing packages for RC0:
1984    - Verify and update package version numbers as needed
1985    - Implement versioning scheme
1986    - Change bug address to point to bugzilla bug entry form
1987    - Disable loadable i18n in libX11 by default (use --enable-loadable-i18n to
1988        reenable it)
1989    - Fix makedepend to use pkgconfig and pass distcheck
1990    - Update build script to build macros first
1991    - Update modular Xorg version
1992
1993commit 5754df51d9b0f8bb8dac7b8c0072da929f1241f0
1994Author: Søren Sandmann Pedersen <sandmann@daimi.au.dk>
1995Date:   Tue Jul 26 22:12:58 2005 +0000
1996
1997    Build system for siliconmotion
1998
1999commit 4b98eb4b30d694139d5e597c445d65d637ab7472
2000Author: Adam Jackson <ajax@nwnk.net>
2001Date:   Mon Jul 11 02:29:59 2005 +0000
2002
2003    Prep for modular builds by adding guarded #include "config.h" everywhere.
2004
2005commit 6389dd3dc1259865a4b5fcc6b6c7b3f4429593cf
2006Author: Daniel Stone <daniel@fooishbar.org>
2007Date:   Fri Jul 1 22:43:12 2005 +0000
2008
2009    Change all misc.h and os.h references to <X11/foo.h>.
2010
2011commit 1dc135becf5898950e04c84cc4d87870840140c1
2012Author: Alex Deucher <agd5f@yahoo.com>
2013Date:   Thu Jun 30 00:14:26 2005 +0000
2014
2015    - add 800x480 mode to smi driver (bug 3639)
2016
2017commit 0670b675c2133dd7c29f79bdf17d24d64181537d
2018Author: Adam Jackson <ajax@nwnk.net>
2019Date:   Sat Jun 25 21:16:56 2005 +0000
2020
2021    Bug #3626: _X_EXPORT tags for video and input drivers.
2022
2023commit 39112b558fdccd2b0a15b8e41d6215e3b0e6e7bd
2024Author: Daniel Stone <daniel@fooishbar.org>
2025Date:   Wed Apr 20 12:25:29 2005 +0000
2026
2027    Fix includes right throughout the Xserver tree:
2028    change "foo.h" to <X11/foo.h> for core headers, e.g. X.h, Xpoll.h;
2029    change "foo.h", "extensions/foo.h" and "X11/foo.h" to
2030        <X11/extensions/foo.h> for extension headers, e.g. Xv.h;
2031    change "foo.[ch]" to <X11/Xtrans/foo.[ch]> for Xtrans files.
2032
2033commit c34e24d83a0f84ad90b7cbaba78d54cc779f74d5
2034Author: Adam Jackson <ajax@nwnk.net>
2035Date:   Fri Sep 17 03:04:54 2004 +0000
2036
2037    Bug #1192: Remove cfb support from drivers where its use is an option.
2038        Delete xf24_32bpp, as s3virge was the last user. Fix up some comments
2039        to refer to fb rather than cfb.
2040
2041commit 61457616aea5d5d7d98e869d3b24e989c2404d6b
2042Author: Adam Jackson <ajax@nwnk.net>
2043Date:   Mon Aug 16 09:13:15 2004 +0000
2044
2045    Bug #1087: Make sure all the *Weak symbols are documented in the import
2046        list, so the loader knows not to complain about unresolved symbols when
2047        an optional module isn't loaded. This typically manifests as warnings
2048        about fbdevHW symbols when the user isn't using the framebuffer
2049        console.
2050
2051commit 74176c460a4f5206cbc7cd7b9f48d1b6cedae7cb
2052Author: Eric Anholt <anholt@freebsd.org>
2053Date:   Sat Jul 31 01:21:51 2004 +0000
2054
2055    Rename some COMPOSITE macros and enums to avoid conflicting with the
2056        COMPOSITE extension's new macro. These appear to only be used
2057        internally.
2058
2059commit 0f239a7f12e1c0a0b7ede424b0f1eb0fe189e9d6
2060Author: Adam Jackson <ajax@nwnk.net>
2061Date:   Sat Jul 31 01:21:19 2004 +0000
2062
2063    Change several LoaderSymbol calls introduced by the bug #400 patch to
2064    *Weak() resolver functions.
2065
2066commit 865ba2a4c1107bb82d22ed648f8a1d63c833d6fc
2067Author: Adam Jackson <ajax@nwnk.net>
2068Date:   Fri Jul 30 20:30:54 2004 +0000
2069
2070    Bug #400 (partial): Driver fixes for the dlloader. When using dlloader, all
2071        framebuffer formats except cfb and the overlay modes should work, and
2072        r128 and radeon need to be loaded from the ati driver (both issues to
2073        be fixed soon). Tested on i740, s3virge, mach64, tdfx, vesa, and vga
2074        drivers. elfloader users shouldn't be affected.
2075
2076commit 9a58bc9adbc982680dd8df27caeb9a5819f2adad
2077Author: Eric Anholt <anholt@freebsd.org>
2078Date:   Wed Jun 16 09:23:38 2004 +0000
2079
2080    DRI XFree86-4_3_99_12-merge import
2081
2082commit 1f957fb7e44762197b0fc5a3b7def25c7888929a
2083Author: Egbert Eich <eich@suse.de>
2084Date:   Wed May 26 16:24:09 2004 +0000
2085
2086    Fixing setting of physical framebuffer base for several drivers. C&T
2087        driver: Fixed setting of minimal clocks for HiQV chipsets. Neomagic
2088        driver: improved support for lowres double scan modes.
2089
2090commit 445708ff1249e96628dd6b11e20f83e568fb4ad1
2091Author: Egbert Eich <eich@suse.de>
2092Date:   Fri Apr 23 19:45:06 2004 +0000
2093
2094    Merging XORG-CURRENT into trunk
2095
2096commit 90141c8b6da8e6b57f2e4bae486a3d39c4069a8b
2097Author: Egbert Eich <eich@suse.de>
2098Date:   Sun Mar 14 08:33:31 2004 +0000
2099
2100    Importing vendor version xf86-4_4_99_1 on Sun Mar 14 00:26:39 PST 2004
2101
2102commit 1082b4e0760359234cb5f51b953bbffc65d594ec
2103Author: Egbert Eich <eich@suse.de>
2104Date:   Wed Mar 3 12:12:24 2004 +0000
2105
2106    Importing vendor version xf86-4_4_0 on Wed Mar 3 04:09:24 PST 2004
2107
2108commit b1c51d42211f715043a3252550ef2fde1a3065e5
2109Author: Egbert Eich <eich@suse.de>
2110Date:   Thu Feb 26 13:35:54 2004 +0000
2111
2112    readding XFree86's cvs IDs
2113
2114commit 2ae35696e96079b192c2cac4e092f6deeb300325
2115Author: Egbert Eich <eich@suse.de>
2116Date:   Thu Feb 26 09:23:23 2004 +0000
2117
2118    Importing vendor version xf86-4_3_99_903 on Wed Feb 26 01:21:00 PST 2004
2119
2120commit ac5b2461e620df7026eefda6a389127e7d41f67d
2121Author: Kaleb Keithley <kaleb@freedesktop.org>
2122Date:   Fri Dec 19 20:55:14 2003 +0000
2123
2124    XFree86 4.3.99.902 (RC 2)
2125
2126commit 14cd4af100e1deb1c7b63e9f19587e38ce1f08f2
2127Author: Kaleb Keithley <kaleb@freedesktop.org>
2128Date:   Tue Nov 25 19:28:39 2003 +0000
2129
2130    XFree86 4.3.99.16 Bring the tree up to date for the Cygwin folks
2131
2132commit 1684aa7a21a194b6e7dd518f3895c9b2dfce229a
2133Author: Kaleb Keithley <kaleb@freedesktop.org>
2134Date:   Fri Nov 14 16:48:55 2003 +0000
2135
2136    XFree86 4.3.0.1
2137
2138commit 78af703cb019a60cc93843efcd4889ccb15abd13
2139Author: Kaleb Keithley <kaleb@freedesktop.org>
2140Date:   Fri Nov 14 16:48:55 2003 +0000
2141
2142    Initial revision
2143