Commit Graph

81 Commits

Author SHA1 Message Date
Henrik Rydgård
28ab4518ed Add a workaround and an assert, trying to fix #18982
Not sure what it is if this doesn't fix #18982.

We already seem to be doing things correctly.
2024-04-29 13:32:26 +02:00
Henrik Rydgård
d28e9dc006 Minor cleanup 2024-04-06 17:03:40 +02:00
Henrik Rydgård
ae6e0d182e Vulkan: Fix function loading regression affecting MSAA
When loading function that have been merged into core, use the core
names.
2024-04-05 17:28:37 +02:00
Henrik Rydgård
9f9881dfe3 Move some code, minor cleanup.
See #18987
2024-04-04 12:30:41 +02:00
Henrik Rydgård
32fe47d761 Fix or ignore some various Vulkan validation errors 2024-04-03 13:37:40 +02:00
Henrik Rydgård
86ea5665f0 Show some more Vulkan extensions in sysinfo 2024-02-01 17:08:54 +01:00
Henrik Rydgård
9253bf9cb5 Add a new screen for managing installed Adreno drivers 2024-01-17 10:51:08 +01:00
Henrik Rydgård
e3177ac870 Make some global string pointers const, not just the strings.
Minor cleanup.
2023-12-29 14:09:45 +01:00
Henrik Rydgård
7634eba083 Merge pull request #18532 from dima-xd/adrenotools
Android: Implement custom driver loading for ARM64 Android devices
2023-12-14 15:54:24 +01:00
Herman Semenov
5a0d54be83 [Common/Data/GPU/Input/System/UI] Added const reference for function params 2023-12-14 14:21:11 +03:00
Dzmitry Dubrova
a461800973 Implement custom driver loading for ARM64 Android devices 2023-12-14 08:53:03 +03:00
Henrik Rydgård
2c751d39f8 Vulkan: Add simple delete count to GPU profiler, to make sure we don't have more resource churn
Also fix the render pass type name lookup table.
2023-10-12 09:02:45 +02:00
Henrik Rydgård
d9e9bde596 Vulkan: Fix disabling VSync on SDL platforms that support IMMEDIATE but not MAILBOX
Fixes #18084
2023-09-06 18:46:19 +02:00
Henrik Rydgård
13cfd9c3d6 Add Mesa as a known GPU driver "vendor". 2023-08-17 22:06:03 +02:00
Henrik Rydgård
aadb700505 Fix turning off VSync on Android, logic fix 2023-08-14 11:02:29 +02:00
Henrik Rydgård
8e74907710 Don't try to use VK_GOOGLE_display_timing on Mac/iOS. MoltenVK has an implementation that seems broken. 2023-08-08 11:46:36 +02:00
Henrik Rydgård
a6b5340d97 Don't crash if VK_KHR_present_id / VK_KHR_present_wait are not exposed (but feature still reported available)
Fixes the new crash in #17858
2023-08-07 11:36:49 +02:00
Henrik Rydgård
e16cac6548 Initial work on supporting VK_GOOGLE_display_timing. Not working yet. 2023-08-03 11:11:16 +02:00
Henrik Rydgård
14377259b5 Use VK_KHR_present_wait to get numbers on how much latency we have to screen
This extension is not available on Android, there they have
VK_GOOGLE_display_timing, which they also have an abstraction library
for, so will look at that later.

Early part of work on #17685
2023-08-02 20:34:22 +02:00
Henrik Rydgård
1c83c4a1fe Vulkan: Enable the present-id/present-wait/display-timing extensions
Doesn't actually use any of them yet, just making sure enabling them
doesn't cause any weird issues. They will be used for the future display
timing rewrite.
2023-07-14 10:43:47 +02:00
Henrik Rydgård
38178d34b5 Sanity checks in InitSwapchain 2023-05-05 00:03:20 +02:00
Henrik Rydgård
c81ec8c74d List Vulkan present modes in system info, show the current one 2023-04-20 00:21:06 +02:00
Henrik Rydgård
f376f3cb5f Another piece of the fix for #17250 (opensuse build error) 2023-04-09 16:57:39 +02:00
Henrik Rydgård
42efc08e64 Couple buildfixes after the g_display refactoring
Should help #17250
2023-04-08 09:25:16 +02:00
Henrik Rydgård
b13c5c2d1b Add compat setting to force using maximum depth resolution
Fixes #17014

Even if our depth-testing-equal heuristic believes that the game needs
lower depth resolution.

This removes some depth-fighting artifacts (that are present on the real
PSP, but nice to avoid) in Outrun, Split/Second and Cars: Race-o-Rama -
essentially reverting these to the behavior we had before the heuristic.

(The heuristic is good though - it means less compat.ini hacks going in
the other direction).

In the case of Outrun, this relies on two passes that pass exactly the
same vertex coordinates twice resulting in the exact same final
geometry. This is actually guaranteed by the spec if the vertex math is
exactly the same and "invariant" is set on the position output, though
I guess you never know.. Haven't seen any issues at least.

Also sneak in disabling some validation messages from using extra Vulkan
validation layers other than the default.
2023-03-26 00:27:58 +01:00