|
1
|
+commit b467d2569a003da05ad222b0dc095bee5eec450a
|
|
2
|
+Author: Michel Dänzer <mdaenzer@redhat.com>
|
|
3
|
+Date: Fri Oct 11 17:10:10 2019 +0200
|
|
4
|
+
|
|
5
|
+ Bump version for the 19.1.0 release
|
|
6
|
+
|
|
7
|
+commit a1b7263277c033e109629829c370c0e95978e061
|
|
8
|
+Author: Michel Dänzer <mdaenzer@redhat.com>
|
|
9
|
+Date: Thu Sep 26 15:56:59 2019 +0200
|
|
10
|
+
|
|
11
|
+ Don't unreference FBs of pixmaps from different screens in LeaveVT
|
|
12
|
+
|
|
13
|
+ FindClientResourcesByType finds pixmaps from all screens, but trying to
|
|
14
|
+ process ones from other screens here makes no sense and likely results
|
|
15
|
+ in a crash or memory corruption.
|
|
16
|
+
|
|
17
|
+ Fixes: c16ff42f927d ("Make all active CRTCs scan out an all-black
|
|
18
|
+ framebuffer in LeaveVT")
|
|
19
|
+ (Ported from radeon commit 2faaecc69b127248718e759c6c98c84d56dd1b6b)
|
|
20
|
+
|
|
21
|
+commit 5b8bc9fc505c551dcd9b0ed5ab835a49fa4f9fda
|
|
22
|
+Author: Michel Dänzer <mdaenzer@redhat.com>
|
|
23
|
+Date: Wed Sep 18 12:55:45 2019 +0200
|
|
24
|
+
|
|
25
|
+ Don't set up black scanout buffer if LeaveVT is called from CloseScreen
|
|
26
|
+
|
|
27
|
+ Avoids a crash described in
|
|
28
|
+ https://gitlab.freedesktop.org/xorg/driver/xf86-video-amdgpu/merge_requests/43#note_223718
|
|
29
|
+
|
|
30
|
+ Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
|
|
31
|
+
|
|
32
|
+commit e6fce59a071220967fcd4e2c9e4a262c72870761
|
|
33
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
34
|
+Date: Wed Jul 24 16:05:05 2019 +0200
|
|
35
|
+
|
|
36
|
+ present: Don't check pixmap pitch in check_flip with non-DC >= 3.34
|
|
37
|
+
|
|
38
|
+ The current non-DC kernel driver also handles flipping between different
|
|
39
|
+ pitches correctly.
|
|
40
|
+
|
|
41
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
42
|
+
|
|
43
|
+commit 5bb2580b266468f87843b5585ae64e056b63bb88
|
|
44
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
45
|
+Date: Wed Jul 24 15:55:19 2019 +0200
|
|
46
|
+
|
|
47
|
+ present: Don't check pixmap pitch in check_flip with current DC
|
|
48
|
+
|
|
49
|
+ Current DC handles flipping between different pitches correctly.
|
|
50
|
+
|
|
51
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
52
|
+
|
|
53
|
+commit ac66086613cbd0974b421cd5eda872adc15242ed
|
|
54
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
55
|
+Date: Wed Jul 24 15:45:21 2019 +0200
|
|
56
|
+
|
|
57
|
+ present: Also check pixmap pitch in check_flip with current xserver
|
|
58
|
+
|
|
59
|
+ The corresponding check in the xserver Present code was removed again,
|
|
60
|
+ because flipping between different pitches can work in some cases.
|
|
61
|
+
|
|
62
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
63
|
+
|
|
64
|
+commit 98f172eb2d2353e19edd8167f22215ce596811f8
|
|
65
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
66
|
+Date: Mon Jul 29 18:54:24 2019 +0200
|
|
67
|
+
|
|
68
|
+ gitlab-ci: Use templates from wayland/ci-templates
|
|
69
|
+
|
|
70
|
+ These are already used by xserver, Mesa and some other projects.
|
|
71
|
+
|
|
72
|
+ Current Debian testing brings e.g. GCC 8.3.0 and clang 7.0.1.
|
|
73
|
+
|
|
74
|
+commit 87f41ace4920fd2069794211683659eb25b025a6
|
|
75
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
76
|
+Date: Fri Jul 5 12:43:53 2019 +0200
|
|
77
|
+
|
|
78
|
+ Don't disable page flipping completely with SW cursor
|
|
79
|
+
|
|
80
|
+ Even with SW cursor, page flipping can be used while no X cursor is
|
|
81
|
+ visible.
|
|
82
|
+
|
|
83
|
+ Occurred to me in the context of xorg/xserver#828.
|
|
84
|
+
|
|
85
|
+commit 7d3fef72e0c871e1677e9e544f4cae5e238b5c52
|
|
86
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
87
|
+Date: Thu May 9 17:39:49 2019 +0200
|
|
88
|
+
|
|
89
|
+ present: Check that we can get a KMS FB for flipping
|
|
90
|
+
|
|
91
|
+ This can legitimately fail if the pixmap's storage is shared from
|
|
92
|
+ another device, e.g. when using PRIME render offloading.
|
|
93
|
+
|
|
94
|
+commit ea19a5207054bb159fc7fb6d88e0ceb10c3da010
|
|
95
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
96
|
+Date: Thu Jun 6 11:02:15 2019 +0200
|
|
97
|
+
|
|
98
|
+ Remove dri2_drawable_crtc parameter consider_disabled
|
|
99
|
+
|
|
100
|
+ All callers were passing TRUE.
|
|
101
|
+
|
|
102
|
+ Reviewed-and-tested-by: Flora Cui <flora.cui@amd.com>
|
|
103
|
+
|
|
104
|
+commit 3109f088fdbd89c2ee8078625d4f073852492656
|
|
105
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
106
|
+Date: Thu Jun 6 11:22:09 2019 +0200
|
|
107
|
+
|
|
108
|
+ dri2: Re-use previous CRTC when possible if pick_best_crtc returns NULL
|
|
109
|
+
|
|
110
|
+ This way, the MSC will continue ticking at the rate of (the last mode
|
|
111
|
+ which was enabled for) that CRTC, instead of the client running
|
|
112
|
+ unthrottled.
|
|
113
|
+
|
|
114
|
+ Reviewed-and-tested-by: Flora Cui <flora.cui@amd.com>
|
|
115
|
+
|
|
116
|
+commit fb06fb814700a47464abd756e1111dcc76d0d776
|
|
117
|
+Author: Flora Cui <flora.cui@amd.com>
|
|
118
|
+Date: Wed May 29 14:18:50 2019 +0800
|
|
119
|
+
|
|
120
|
+ dri2: reply to client for WaitMSC request in any case
|
|
121
|
+
|
|
122
|
+ otherwise client would wait for reply forever and desktop appears hang.
|
|
123
|
+
|
|
124
|
+ Signed-off-by: Flora Cui <flora.cui@amd.com>
|
|
125
|
+ Acked-by: Feifei Xu <Feifei.Xu@amd.com>
|
|
126
|
+ Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
127
|
+
|
|
128
|
+commit 4b17533fcb30842caf0035ba593b7d986520cc85
|
|
129
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
130
|
+Date: Tue Apr 30 17:50:15 2019 +0200
|
|
131
|
+
|
|
132
|
+ dri3: Always flush glamor before sharing pixmap storage with clients
|
|
133
|
+
|
|
134
|
+ Even if glamor_gbm_bo_from_pixmap / glamor_fd_from_pixmap themselves
|
|
135
|
+ don't trigger any drawing, there could already be unflushed drawing to
|
|
136
|
+ the pixmap whose storage we share with a client.
|
|
137
|
+
|
|
138
|
+commit bf61e6d7ac1a5754b1026d7f80acf25ef622c491
|
|
139
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
140
|
+Date: Thu Apr 18 19:21:40 2019 +0200
|
|
141
|
+
|
|
142
|
+ Retry get_fb_ptr in get_fb
|
|
143
|
+
|
|
144
|
+ If get_fb_ptr returns NULL, try again after pixmap_get_handle, it should
|
|
145
|
+ work then.
|
|
146
|
+
|
|
147
|
+ Fixes spurious Present page flipping failures using "normal" pixmaps
|
|
148
|
+ which aren't shared with direct rendering clients, e.g. with a
|
|
149
|
+ compositor using the RENDER extension.
|
|
150
|
+
|
|
151
|
+ Bugzilla: https://bugs.freedesktop.org/110417
|
|
152
|
+ Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
|
|
153
|
+
|
|
154
|
+commit bd4ffd4ebbdf1c43ab9e1ef9ba8b812fd2dde4a4
|
|
155
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
156
|
+Date: Tue Mar 19 18:44:31 2019 +0100
|
|
157
|
+
|
|
158
|
+ Bump version for the 19.0.1 release
|
|
159
|
+
|
|
160
|
+commit 6ee857726166f495abcd68e4ff60e3a09593d079
|
|
161
|
+Author: Dave Airlie <airlied@redhat.com>
|
|
162
|
+Date: Mon Mar 23 11:33:23 2015 +1000
|
|
163
|
+
|
|
164
|
+ modesetting: add tile property support
|
|
165
|
+
|
|
166
|
+ This adds tiling support to the driver, it retrieves the tile info from
|
|
167
|
+ the kernel and translates it into the server format and exposes the
|
|
168
|
+ property.
|
|
169
|
+
|
|
170
|
+ (Ported from xserver commits 8fb8bbb3062f1a06621ab7030a9e89d5e8367b35
|
|
171
|
+ and 6abdb54a11dac4e8854ff94ecdcb90a14321ab31)
|
|
172
|
+ Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
|
|
173
|
+
|
|
174
|
+commit 9534bf3bb33d14cd3a5af08e36ef42b309647fc7
|
|
175
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
176
|
+Date: Wed Mar 6 12:05:14 2019 +0100
|
|
177
|
+
|
|
178
|
+ Bump version for the 19.0.0 release
|
|
179
|
+
|
|
180
|
+commit a2b32e72fdaff3007a79b84929997d8176c2d512
|
|
181
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
182
|
+Date: Fri Mar 1 17:42:08 2019 +0100
|
|
183
|
+
|
|
184
|
+ present: Don't check tiling parameters with DC & DRM minor version >= 31
|
|
185
|
+
|
|
186
|
+ Current DC handles any changes of tiling parameters for flips.
|
|
187
|
+
|
|
188
|
+ v2:
|
|
189
|
+ * Just check all tiling bits if DRM minor < 31 or DC is disabled.
|
|
190
|
+
|
|
191
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
192
|
+
|
|
193
|
+commit 2798244be78df3ef3a7841597577506bfbe50156
|
|
194
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
195
|
+Date: Fri Mar 1 17:47:24 2019 +0100
|
|
196
|
+
|
|
197
|
+ Make drmmode_cm_enabled an inline function
|
|
198
|
+
|
|
199
|
+ So that it can be used outside of drmmode_display.c as well.
|
|
200
|
+
|
|
201
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
202
|
+
|
|
203
|
+commit 72653455e4f652ca6c7c290c7f1e8a889b77f5ce
|
|
204
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
205
|
+Date: Fri Mar 1 17:35:48 2019 +0100
|
|
206
|
+
|
|
207
|
+ Revert "Remove set but unused amdgpu_dri2::pKernelDRMVersion"
|
|
208
|
+
|
|
209
|
+ This reverts commit 720a61000aeb139005bd8125908cec66a6e69554.
|
|
210
|
+
|
|
211
|
+ We're going to make use of it now.
|
|
212
|
+
|
|
213
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
214
|
+
|
|
215
|
+commit 28cd209ebf20561e65d14fa2e8bbfaedf6965948
|
|
216
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
217
|
+Date: Wed Feb 27 17:35:26 2019 +0100
|
|
218
|
+
|
|
219
|
+ Revert "gitlab-ci: Only run docker-image stage if relevant source files change"
|
|
220
|
+
|
|
221
|
+ This reverts commit 9c23076b9e81c36ac2408c491f9b2d546829ee8e.
|
|
222
|
+
|
|
223
|
+ Some scenarios have come to light where this failed to ensure the docker
|
|
224
|
+ image exists:
|
|
225
|
+
|
|
226
|
+ * If the master branch of a forked repository is used for an MR which
|
|
227
|
+ doesn't modify .gitlab-ci.yml, the docker-image job may not run.
|
|
228
|
+ * If the docker-image job of the first pipeline in a forked repository
|
|
229
|
+ is cancelled or fails for any reason, and .gitlab-ci.yml isn't
|
|
230
|
+ modified for the next pipeline run.
|
|
231
|
+
|
|
232
|
+commit 09be74a3d1dd9604336d9a27f98d132b262dcbaf
|
|
233
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
234
|
+Date: Thu Feb 28 17:31:55 2019 +0100
|
|
235
|
+
|
|
236
|
+ dri2: Call drm_queue_handle_deferred in dri2_deferred_event
|
|
237
|
+
|
|
238
|
+ drm_queue_handler just puts the event on the signalled list; without
|
|
239
|
+ calling drm_queue_handle_deferred, actual processing of the event may be
|
|
240
|
+ delayed indefinitely, e.g. until another event arrives from the kernel.
|
|
241
|
+
|
|
242
|
+ This could result in DRI2 clients hanging during DPMS off.
|
|
243
|
+
|
|
244
|
+ Fixes: 739181c8d3334 "Add amdgpu_drm_handle_event wrapper for
|
|
245
|
+ drmHandleEvent"
|
|
246
|
+ Reviewed-by: Aaron Liu <aaron.liu@amd.com>
|
|
247
|
+ Tested-by: Aaron Liu <aaron.liu@amd.com>
|
|
248
|
+
|
|
249
|
+commit a636f42b496b0604ca00a144690ece61d1a88a27
|
|
250
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
251
|
+Date: Wed Feb 27 18:43:27 2019 +0100
|
|
252
|
+
|
|
253
|
+ present: Check that flip and screen pixmap pitches match
|
|
254
|
+
|
|
255
|
+ If they don't, flipping will result in corrupted display.
|
|
256
|
+
|
|
257
|
+ Test case:
|
|
258
|
+
|
|
259
|
+ * Run Xorg at 1920x1080 with no window manager
|
|
260
|
+ * glxgears -geometry 2048x1080
|
|
261
|
+
|
|
262
|
+ The Present extension code in xserver 1.21 will check for this.
|
|
263
|
+
|
|
264
|
+ Tested-by: Jax Lin <jax.lin@amd.com>
|
|
265
|
+
|
|
266
|
+commit bd090f389f19f1f4a3f662ffdd891345a3899539
|
|
267
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
268
|
+Date: Tue Feb 12 17:57:17 2019 +0100
|
|
269
|
+
|
|
270
|
+ Call amdgpu_present_set_screen_vrr from amdgpu_vrr_property_update
|
|
271
|
+
|
|
272
|
+ If the window is currently flipping.
|
|
273
|
+
|
|
274
|
+ This might make a difference when the property gets disabled: Variable
|
|
275
|
+ refresh will now be disabled immediately in that case, instead of only
|
|
276
|
+ when the window can no longer use page flipping at all.
|
|
277
|
+
|
|
278
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
279
|
+
|
|
280
|
+commit d9be5d712d469595e1e610f7294bc670ca3b1985
|
|
281
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
282
|
+Date: Tue Feb 12 12:26:25 2019 +0100
|
|
283
|
+
|
|
284
|
+ Make use of property request wrappers for VRR property
|
|
285
|
+
|
|
286
|
+ Instead of scanning for PropertyNotify events. Reasons:
|
|
287
|
+
|
|
288
|
+ * Works even if no client listens to PropertyNotify events for the
|
|
289
|
+ window.
|
|
290
|
+ * No overhead on delivery of unrelated events, and no overhead at all
|
|
291
|
+ if Option "VariableRefresh" is disabled.
|
|
292
|
+
|
|
293
|
+ v2:
|
|
294
|
+ * Use shorter variable name amdgpu_vrr_atom.
|
|
295
|
+ * Call MakeAtom regardless of info->instance_id, for robustness vs VRR
|
|
296
|
+ being enabled in some but not all AMDGPU screens.
|
|
297
|
+
|
|
298
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
299
|
+
|
|
300
|
+commit ef8fbe33b7d97f7fb5518db9c0e4d2dcbf2fab6f
|
|
301
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
302
|
+Date: Thu Jan 17 18:41:11 2019 +0100
|
|
303
|
+
|
|
304
|
+ Wrap change/delete window property request handlers
|
|
305
|
+
|
|
306
|
+ Preparation for the following change.
|
|
307
|
+
|
|
308
|
+ v2:
|
|
309
|
+ * Add comments explaining what the wrappers are wrapping.
|
|
310
|
+ * Use global amdgpu_property_vectors_wrapped to keep track of whether
|
|
311
|
+ the vectors need to be (un)wrapped, for robustness against VRR being
|
|
312
|
+ enabled in some but not all AMDGPU screens.
|
|
313
|
+
|
|
314
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
315
|
+
|
|
316
|
+commit 09a45ff8fe3ac07bafa3a0822b1598c41f9ca200
|
|
317
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
318
|
+Date: Tue Feb 12 13:01:04 2019 +0100
|
|
319
|
+
|
|
320
|
+ Don't enable the VRR support code for GPU screens
|
|
321
|
+
|
|
322
|
+ Windows aren't associated with GPU screens, and amdgpu_present_flip is
|
|
323
|
+ never called for them, so VRR can never actually be enabled for them.
|
|
324
|
+
|
|
325
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
326
|
+
|
|
327
|
+commit 2a3d00dc7ed2b4fca698e2d699e1b94da6d0ddb8
|
|
328
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
329
|
+Date: Tue Feb 12 18:12:23 2019 +0100
|
|
330
|
+
|
|
331
|
+ Don't register a window private if VRR is disabled
|
|
332
|
+
|
|
333
|
+ It's not used in that case.
|
|
334
|
+
|
|
335
|
+ Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
|
|
336
|
+
|
|
337
|
+commit 5f91be77e059d0c4a4268ec10cbd9aa1052f53eb
|
|
338
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
339
|
+Date: Mon Feb 11 18:32:07 2019 +0100
|
|
340
|
+
|
|
341
|
+ gitlab-ci: Don't rely on $CI_PROJECT_NAME
|
|
342
|
+
|
|
343
|
+ The name of a forked repository can be changed later, in which case this
|
|
344
|
+ would fail to refer to the main repository.
|
|
345
|
+
|
|
346
|
+ Pointed out by Eric Engestrom in
|
|
347
|
+ https://gitlab.freedesktop.org/mesa/mesa/merge_requests/224 .
|
|
348
|
+
|
|
349
|
+commit 9c23076b9e81c36ac2408c491f9b2d546829ee8e
|
|
350
|
+Author: Michel Dänzer <michel.daenzer@amd.com>
|
|
351
|
+Date: Thu Feb 7 17:35:13 2019 +0100
|
|
352
|
+
|
|
353
|
+ gitlab-ci: Only run docker-image stage if relevant source files change
|
|
354
|
+
|
|
355
|
+ Otherwise there's normally no need to run it. It will also run when a
|
|
356
|
+ new branch is created, which ensures that the docker image always exists
|
|
357
|
+ (e.g. in a newly forked repository).
|
|
358
|
+
|
|
359
|
+ Inspired by https://gitlab.freedesktop.org/mesa/mesa/merge_requests/143
|
|
360
|
+
|
1
|
361
|
commit 9045fb310f88780e250e60b80431ca153330e61b
|
2
|
362
|
Author: Michel Dänzer <michel.daenzer@amd.com>
|
3
|
363
|
Date: Thu Jan 24 18:31:40 2019 +0100
|
... |
... |
@@ -5876,7 +6236,7 @@ Date: Tue Jun 2 17:04:21 2015 +0900 |
5876
|
6236
|
|
5877
|
6237
|
Must be enabled with
|
5878
|
6238
|
|
5879
|
|
- Option "DRI3"
|
|
6239
|
+ Option "DRI3"
|
5880
|
6240
|
|
5881
|
6241
|
in xorg.conf.
|
5882
|
6242
|
|