1
0
Fork 0
mirror of https://github.com/Ryujinx/Ryujinx.git synced 2024-12-27 00:36:00 +00:00
Ryujinx/Ryujinx.Memory.Tests/TrackingTests.cs

510 lines
19 KiB
C#
Raw Normal View History

Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
using NUnit.Framework;
using Ryujinx.Memory.Tracking;
using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.Threading;
namespace Ryujinx.Memory.Tests
{
public class TrackingTests
{
private const int RndCnt = 3;
private const ulong MemorySize = 0x8000;
private const int PageSize = 4096;
private MemoryBlock _memoryBlock;
private MemoryTracking _tracking;
private MockVirtualMemoryManager _memoryManager;
[SetUp]
public void Setup()
{
_memoryBlock = new MemoryBlock(MemorySize);
_memoryManager = new MockVirtualMemoryManager(MemorySize, PageSize);
POWER - Performance Optimizations With Extensive Ramifications (#2286) * Refactoring of KMemoryManager class * Replace some trivial uses of DRAM address with VA * Get rid of GetDramAddressFromVa * Abstracting more operations on derived page table class * Run auto-format on KPageTableBase * Managed to make TryConvertVaToPa private, few uses remains now * Implement guest physical pages ref counting, remove manual freeing * Make DoMmuOperation private and call new abstract methods only from the base class * Pass pages count rather than size on Map/UnmapMemory * Change memory managers to take host pointers * Fix a guest memory leak and simplify KPageTable * Expose new methods for host range query and mapping * Some refactoring of MapPagesFromClientProcess to allow proper page ref counting and mapping without KPageLists * Remove more uses of AddVaRangeToPageList, now only one remains (shared memory page checking) * Add a SharedMemoryStorage class, will be useful for host mapping * Sayonara AddVaRangeToPageList, you served us well * Start to implement host memory mapping (WIP) * Support memory tracking through host exception handling * Fix some access violations from HLE service guest memory access and CPU * Fix memory tracking * Fix mapping list bugs, including a race and a error adding mapping ranges * Simple page table for memory tracking * Simple "volatile" region handle mode * Update UBOs directly (experimental, rough) * Fix the overlap check * Only set non-modified buffers as volatile * Fix some memory tracking issues * Fix possible race in MapBufferFromClientProcess (block list updates were not locked) * Write uniform update to memory immediately, only defer the buffer set. * Fix some memory tracking issues * Pass correct pages count on shared memory unmap * Armeilleure Signal Handler v1 + Unix changes Unix currently behaves like windows, rather than remapping physical * Actually check if the host platform is unix * Fix decommit on linux. * Implement windows 10 placeholder shared memory, fix a buffer issue. * Make PTC version something that will never match with master * Remove testing variable for block count * Add reference count for memory manager, fix dispose Can still deadlock with OpenAL * Add address validation, use page table for mapped check, add docs Might clean up the page table traversing routines. * Implement batched mapping/tracking. * Move documentation, fix tests. * Cleanup uniform buffer update stuff. * Remove unnecessary assignment. * Add unsafe host mapped memory switch On by default. Would be good to turn this off for untrusted code (homebrew, exefs mods) and give the user the option to turn it on manually, though that requires some UI work. * Remove C# exception handlers They have issues due to current .NET limitations, so the meilleure one fully replaces them for now. * Fix MapPhysicalMemory on the software MemoryManager. * Null check for GetHostAddress, docs * Add configuration for setting memory manager mode (not in UI yet) * Add config to UI * Fix type mismatch on Unix signal handler code emit * Fix 6GB DRAM mode. The size can be greater than `uint.MaxValue` when the DRAM is >4GB. * Address some feedback. * More detailed error if backing memory cannot be mapped. * SetLastError on all OS functions for consistency * Force pages dirty with UBO update instead of setting them directly. Seems to be much faster across a few games. Need retesting. * Rebase, configuration rework, fix mem tracking regression * Fix race in FreePages * Set memory managers null after decrementing ref count * Remove readonly keyword, as this is now modified. * Use a local variable for the signal handler rather than a register. * Fix bug with buffer resize, and index/uniform buffer binding. Should fix flickering in games. * Add InvalidAccessHandler to MemoryTracking Doesn't do anything yet * Call invalid access handler on unmapped read/write. Same rules as the regular memory manager. * Make unsafe mapped memory its own MemoryManagerType * Move FlushUboDirty into UpdateState. * Buffer dirty cache, rather than ubo cache Much cleaner, may be reusable for Inline2Memory updates. * This doesn't return anything anymore. * Add sigaction remove methods, correct a few function signatures. * Return empty list of physical regions for size 0. * Also on AddressSpaceManager Co-authored-by: gdkchan <gab.dark.100@gmail.com>
2021-05-24 21:52:44 +01:00
_tracking = new MemoryTracking(_memoryManager, PageSize);
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
}
[TearDown]
public void Teardown()
{
_memoryBlock.Dispose();
}
private bool TestSingleWrite(RegionHandle handle, ulong address, ulong size)
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
{
handle.Reprotect();
_tracking.VirtualMemoryEvent(address, size, true);
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
return handle.Dirty;
}
[Test]
public void SingleRegion()
{
RegionHandle handle = _tracking.BeginTracking(0, PageSize);
(ulong address, ulong size)? readTrackingTriggered = null;
handle.RegisterAction((address, size) =>
{
readTrackingTriggered = (address, size);
});
bool dirtyInitial = handle.Dirty;
Assert.True(dirtyInitial); // Handle starts dirty.
handle.Reprotect();
bool dirtyAfterReprotect = handle.Dirty;
Assert.False(dirtyAfterReprotect); // Handle is no longer dirty.
_tracking.VirtualMemoryEvent(PageSize * 2, 4, true);
_tracking.VirtualMemoryEvent(PageSize * 2, 4, false);
bool dirtyAfterUnrelatedReadWrite = handle.Dirty;
Assert.False(dirtyAfterUnrelatedReadWrite); // Not dirtied, as the write was to an unrelated address.
Assert.IsNull(readTrackingTriggered); // Hasn't been triggered yet
_tracking.VirtualMemoryEvent(0, 4, false);
bool dirtyAfterRelatedRead = handle.Dirty;
Assert.False(dirtyAfterRelatedRead); // Only triggers on write.
Assert.AreEqual(readTrackingTriggered, (0UL, 4UL)); // Read action was triggered.
readTrackingTriggered = null;
_tracking.VirtualMemoryEvent(0, 4, true);
bool dirtyAfterRelatedWrite = handle.Dirty;
Assert.True(dirtyAfterRelatedWrite); // Dirty flag should now be set.
_tracking.VirtualMemoryEvent(4, 4, true);
bool dirtyAfterRelatedWrite2 = handle.Dirty;
Assert.True(dirtyAfterRelatedWrite2); // Dirty flag should still be set.
handle.Reprotect();
bool dirtyAfterReprotect2 = handle.Dirty;
Assert.False(dirtyAfterReprotect2); // Handle is no longer dirty.
handle.Dispose();
bool dirtyAfterDispose = TestSingleWrite(handle, 0, 4);
Assert.False(dirtyAfterDispose); // Handle cannot be triggered when disposed
}
[Test]
public void OverlappingRegions()
{
RegionHandle allHandle = _tracking.BeginTracking(0, PageSize * 16);
allHandle.Reprotect();
(ulong address, ulong size)? readTrackingTriggeredAll = null;
Action registerReadAction = () =>
{
readTrackingTriggeredAll = null;
allHandle.RegisterAction((address, size) =>
{
readTrackingTriggeredAll = (address, size);
});
};
registerReadAction();
// Create 16 page sized handles contained within the allHandle.
RegionHandle[] containedHandles = new RegionHandle[16];
for (int i = 0; i < 16; i++)
{
containedHandles[i] = _tracking.BeginTracking((ulong)i * PageSize, PageSize);
containedHandles[i].Reprotect();
}
for (int i = 0; i < 16; i++)
{
// No handles are dirty.
Assert.False(allHandle.Dirty);
Assert.IsNull(readTrackingTriggeredAll);
for (int j = 0; j < 16; j++)
{
Assert.False(containedHandles[j].Dirty);
}
_tracking.VirtualMemoryEvent((ulong)i * PageSize, 1, true);
// Only the handle covering the entire range and the relevant contained handle are dirty.
Assert.True(allHandle.Dirty);
Assert.AreEqual(readTrackingTriggeredAll, ((ulong)i * PageSize, 1UL)); // Triggered read tracking
for (int j = 0; j < 16; j++)
{
if (j == i)
{
Assert.True(containedHandles[j].Dirty);
}
else
{
Assert.False(containedHandles[j].Dirty);
}
}
// Clear flags and reset read action.
registerReadAction();
allHandle.Reprotect();
containedHandles[i].Reprotect();
}
}
[Test]
public void PageAlignment(
[Values(1ul, 512ul, 2048ul, 4096ul, 65536ul)] [Random(1ul, 65536ul, RndCnt)] ulong address,
[Values(1ul, 4ul, 1024ul, 4096ul, 65536ul)] [Random(1ul, 65536ul, RndCnt)] ulong size)
{
ulong alignedStart = (address / PageSize) * PageSize;
ulong alignedEnd = ((address + size + PageSize - 1) / PageSize) * PageSize;
ulong alignedSize = alignedEnd - alignedStart;
RegionHandle handle = _tracking.BeginTracking(address, size);
// Anywhere inside the pages the region is contained on should trigger.
bool originalRangeTriggers = TestSingleWrite(handle, address, size);
Assert.True(originalRangeTriggers);
bool alignedRangeTriggers = TestSingleWrite(handle, alignedStart, alignedSize);
Assert.True(alignedRangeTriggers);
bool alignedStartTriggers = TestSingleWrite(handle, alignedStart, 1);
Assert.True(alignedStartTriggers);
bool alignedEndTriggers = TestSingleWrite(handle, alignedEnd - 1, 1);
Assert.True(alignedEndTriggers);
// Outside the tracked range should not trigger.
bool alignedBeforeTriggers = TestSingleWrite(handle, alignedStart - 1, 1);
Assert.False(alignedBeforeTriggers);
bool alignedAfterTriggers = TestSingleWrite(handle, alignedEnd, 1);
Assert.False(alignedAfterTriggers);
}
[Test, Explicit, Timeout(1000)]
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
public void Multithreading()
{
// Multithreading sanity test
// Multiple threads can easily read/write memory regions from any existing handle.
// Handles can also be owned by different threads, though they should have one owner thread.
// Handles can be created and disposed at any time, by any thread.
// This test should not throw or deadlock due to invalid state.
const int threadCount = 1;
const int handlesPerThread = 16;
long finishedTime = 0;
RegionHandle[] handles = new RegionHandle[threadCount * handlesPerThread];
Random globalRand = new Random();
for (int i = 0; i < handles.Length; i++)
{
handles[i] = _tracking.BeginTracking((ulong)i * PageSize, PageSize);
handles[i].Reprotect();
}
List<Thread> testThreads = new List<Thread>();
// Dirty flag consumer threads
int dirtyFlagReprotects = 0;
for (int i = 0; i < threadCount; i++)
{
int randSeed = i;
testThreads.Add(new Thread(() =>
{
int handleBase = randSeed * handlesPerThread;
while (Stopwatch.GetTimestamp() < finishedTime)
{
Random random = new Random(randSeed);
RegionHandle handle = handles[handleBase + random.Next(handlesPerThread)];
if (handle.Dirty)
{
handle.Reprotect();
Interlocked.Increment(ref dirtyFlagReprotects);
}
}
}));
}
// Write trigger threads
int writeTriggers = 0;
for (int i = 0; i < threadCount; i++)
{
int randSeed = i;
testThreads.Add(new Thread(() =>
{
Random random = new Random(randSeed);
ulong handleBase = (ulong)(randSeed * handlesPerThread * PageSize);
while (Stopwatch.GetTimestamp() < finishedTime)
{
_tracking.VirtualMemoryEvent(handleBase + (ulong)random.Next(PageSize * handlesPerThread), PageSize / 2, true);
Interlocked.Increment(ref writeTriggers);
}
}));
}
// Handle create/delete threads
int handleLifecycles = 0;
for (int i = 0; i < threadCount; i++)
{
int randSeed = i;
testThreads.Add(new Thread(() =>
{
int maxAddress = threadCount * handlesPerThread * PageSize;
Random random = new Random(randSeed + 512);
while (Stopwatch.GetTimestamp() < finishedTime)
{
RegionHandle handle = _tracking.BeginTracking((ulong)random.Next(maxAddress), (ulong)random.Next(65536));
handle.Dispose();
Interlocked.Increment(ref handleLifecycles);
}
}));
}
finishedTime = Stopwatch.GetTimestamp() + Stopwatch.Frequency / 2; // Run for 500ms;
foreach (Thread thread in testThreads)
{
thread.Start();
}
foreach (Thread thread in testThreads)
{
thread.Join();
}
Assert.Greater(dirtyFlagReprotects, 10);
Assert.Greater(writeTriggers, 10);
Assert.Greater(handleLifecycles, 10);
}
[Test]
public void ReadActionThreadConsumption()
{
// Read actions should only be triggered once for each registration.
// The implementation should use an interlocked exchange to make sure other threads can't get the action.
RegionHandle handle = _tracking.BeginTracking(0, PageSize);
int triggeredCount = 0;
int registeredCount = 0;
int signalThreadsDone = 0;
bool isRegistered = false;
Action registerReadAction = () =>
{
registeredCount++;
handle.RegisterAction((address, size) =>
{
isRegistered = false;
Interlocked.Increment(ref triggeredCount);
});
};
const int threadCount = 16;
const int iterationCount = 10000;
Thread[] signalThreads = new Thread[threadCount];
for (int i = 0; i < threadCount; i++)
{
int randSeed = i;
signalThreads[i] = new Thread(() =>
{
Random random = new Random(randSeed);
for (int j = 0; j < iterationCount; j++)
{
_tracking.VirtualMemoryEvent((ulong)random.Next(PageSize), 4, false);
}
Interlocked.Increment(ref signalThreadsDone);
});
}
for (int i = 0; i < threadCount; i++)
{
signalThreads[i].Start();
}
while (signalThreadsDone != -1)
{
if (signalThreadsDone == threadCount)
{
signalThreadsDone = -1;
}
if (!isRegistered)
{
isRegistered = true;
registerReadAction();
}
}
// The action should trigger exactly once for every registration,
// then we register once after all the threads signalling it cease.
Assert.AreEqual(registeredCount, triggeredCount + 1);
}
[Test]
public void DisposeHandles()
{
// Ensure that disposed handles correctly remove their virtual and physical regions.
RegionHandle handle = _tracking.BeginTracking(0, PageSize);
handle.Reprotect();
Assert.AreEqual(1, _tracking.GetRegionCount());
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
handle.Dispose();
Assert.AreEqual(0, _tracking.GetRegionCount());
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
// Two handles, small entirely contains big.
// We expect there to be three regions after creating both, one for the small region and two covering the big one around it.
// Regions are always split to avoid overlapping, which is why there are three instead of two.
RegionHandle handleSmall = _tracking.BeginTracking(PageSize, PageSize);
RegionHandle handleBig = _tracking.BeginTracking(0, PageSize * 4);
Assert.AreEqual(3, _tracking.GetRegionCount());
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
// After disposing the big region, only the small one will remain.
handleBig.Dispose();
Assert.AreEqual(1, _tracking.GetRegionCount());
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
handleSmall.Dispose();
Assert.AreEqual(0, _tracking.GetRegionCount());
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
}
Implement lazy flush-on-read for Buffers (SSBO/Copy) (#1790) * Initial implementation of buffer flush (VERY WIP) * Host shaders need to be rebuilt for the SSBO write flag. * New approach with reserved regions and gl sync * Fix a ton of buffer issues. * Remove unused buffer unmapped behaviour * Revert "Remove unused buffer unmapped behaviour" This reverts commit f1700e52fb8760180ac5e0987a07d409d1e70ece. * Delete modified ranges on unmap Fixes potential crashes in Super Smash Bros, where a previously modified range could lie on either side of an unmap. * Cache some more delegates. * Dispose Sync on Close * Also create host sync for GPFifo syncpoint increment. * Copy buffer optimization, add docs * Fix race condition with OpenGL Sync * Enable read tracking on CommandBuffer, insert syncpoint on WaitForIdle * Performance: Only flush individual pages of SSBO at a time This avoids flushing large amounts of data when only a small amount is actually used. * Signal Modified rather than flushing after clear * Fix some docs and code style. * Introduce a new test for tracking memory protection. Sucessfully demonstrates that the bug causing write protection to be cleared by a read action has been fixed. (these tests fail on master) * Address Comments * Add host sync for SetReference This ensures that any indirect draws will correctly flush any related buffer data written before them. Fixes some flashing and misplaced world geometry in MH rise. * Make PageAlign static * Re-enable read tracking, for reads.
2021-01-17 20:08:06 +00:00
[Test]
public void ReadAndWriteProtection()
{
MemoryPermission protection = MemoryPermission.ReadAndWrite;
_memoryManager.OnProtect += (va, size, newProtection) =>
{
Assert.AreEqual((0, PageSize), (va, size)); // Should protect the exact region all the operations use.
protection = newProtection;
};
RegionHandle handle = _tracking.BeginTracking(0, PageSize);
// After creating the handle, there is no protection yet.
Assert.AreEqual(MemoryPermission.ReadAndWrite, protection);
bool dirtyInitial = handle.Dirty;
Assert.True(dirtyInitial); // Handle starts dirty.
handle.Reprotect();
// After a reprotect, there is write protection, which will set a dirty flag when any write happens.
Assert.AreEqual(MemoryPermission.Read, protection);
(ulong address, ulong size)? readTrackingTriggered = null;
handle.RegisterAction((address, size) =>
{
readTrackingTriggered = (address, size);
});
// Registering an action adds read/write protection.
Assert.AreEqual(MemoryPermission.None, protection);
bool dirtyAfterReprotect = handle.Dirty;
Assert.False(dirtyAfterReprotect); // Handle is no longer dirty.
// First we should read, which will trigger the action. This _should not_ remove write protection on the memory.
_tracking.VirtualMemoryEvent(0, 4, false);
bool dirtyAfterRead = handle.Dirty;
Assert.False(dirtyAfterRead); // Not dirtied, as this was a read.
Assert.AreEqual(readTrackingTriggered, (0UL, 4UL)); // Read action was triggered.
Assert.AreEqual(MemoryPermission.Read, protection); // Write protection is still present.
readTrackingTriggered = null;
// Now, perform a write.
_tracking.VirtualMemoryEvent(0, 4, true);
bool dirtyAfterWriteAfterRead = handle.Dirty;
Assert.True(dirtyAfterWriteAfterRead); // Should be dirty.
Assert.AreEqual(MemoryPermission.ReadAndWrite, protection); // All protection is now be removed from the memory.
Assert.IsNull(readTrackingTriggered); // Read tracking was removed when the action fired, as it can only fire once.
handle.Dispose();
}
Replace CacheResourceWrite with more general "precise" write (#2684) * Replace CacheResourceWrite with more general "precise" write The goal of CacheResourceWrite was to notify GPU resources when they were modified directly, by looking up the modified address/size in a structure and calling a method on each resource. The downside of this is that each resource cache has to be queried individually, they all have to implement their own way to do this, and it can only signal to resources using the same PhysicalMemory instance. This PR adds the ability to signal a write as "precise" on the tracking, which signals a special handler (if present) which can be used to avoid unnecessary flush actions, or maybe even more. For buffers, precise writes specifically do not flush, and instead punch a hole in the modified range list to indicate that the data on GPU has been replaced. The downside is that precise actions must ignore the page protection bits and always signal - as they need to notify the target resource to ignore the sequence number optimization. I had to reintroduce the sequence number increment after I2M, as removing it was causing issues in rabbids kingdom battle. However - all resources modified by I2M are notified directly to lower their sequence number, so the problem is likely that another unrelated resource is not being properly updated. Thankfully, doing this does not affect performance in the games I tested. This should fix regressions from #2624. Test any games that were broken by that. (RF4, rabbids kingdom battle) I've also added a sequence number increment to ThreedClass.IncrementSyncpoint, as it seems to fix buffer corruption in OpenGL homebrew. (this was a regression from removing sequence number increment from constant buffer update - another unrelated resource thing) * Add tests. * Add XML docs for GpuRegionHandle * Skip UpdateProtection if only precise actions were called This allows precise actions to skip reprotection costs.
2021-09-29 01:27:03 +01:00
[Test]
public void PreciseAction()
{
RegionHandle handle = _tracking.BeginTracking(0, PageSize);
(ulong address, ulong size, bool write)? preciseTriggered = null;
handle.RegisterPreciseAction((address, size, write) =>
{
preciseTriggered = (address, size, write);
return true;
});
(ulong address, ulong size)? readTrackingTriggered = null;
handle.RegisterAction((address, size) =>
{
readTrackingTriggered = (address, size);
});
handle.Reprotect();
_tracking.VirtualMemoryEvent(0, 4, false, precise: true);
Assert.IsNull(readTrackingTriggered); // Hasn't been triggered - precise action returned true.
Assert.AreEqual(preciseTriggered, (0UL, 4UL, false)); // Precise action was triggered.
_tracking.VirtualMemoryEvent(0, 4, true, precise: true);
Assert.IsNull(readTrackingTriggered); // Still hasn't been triggered.
bool dirtyAfterPreciseActionTrue = handle.Dirty;
Assert.False(dirtyAfterPreciseActionTrue); // Not dirtied - precise action returned true.
Assert.AreEqual(preciseTriggered, (0UL, 4UL, true)); // Precise action was triggered.
// Handle is now dirty.
handle.Reprotect(true);
preciseTriggered = null;
_tracking.VirtualMemoryEvent(4, 4, true, precise: true);
Assert.AreEqual(preciseTriggered, (4UL, 4UL, true)); // Precise action was triggered even though handle was dirty.
handle.Reprotect();
handle.RegisterPreciseAction((address, size, write) =>
{
preciseTriggered = (address, size, write);
return false; // Now, we return false, which indicates that the regular read/write behaviours should trigger.
});
_tracking.VirtualMemoryEvent(8, 4, true, precise: true);
Assert.AreEqual(readTrackingTriggered, (8UL, 4UL)); // Read action triggered, as precise action returned false.
bool dirtyAfterPreciseActionFalse = handle.Dirty;
Assert.True(dirtyAfterPreciseActionFalse); // Dirtied, as precise action returned false.
Assert.AreEqual(preciseTriggered, (8UL, 4UL, true)); // Precise action was triggered.
}
Memory Read/Write Tracking using Region Handles (#1272) * WIP Range Tracking - Texture invalidation seems to have large problems - Buffer/Pool invalidation may have problems - Mirror memory tracking puts an additional `add` in compiled code, we likely just want to make HLE access slower if this is the final solution. - Native project is in the messiest possible location. - [HACK] JIT memory access always uses native "fast" path - [HACK] Trying some things with texture invalidation and views. It works :) Still a few hacks, messy things, slow things More work in progress stuff (also move to memory project) Quite a bit faster now. - Unmapping GPU VA and CPU VA will now correctly update write tracking regions, and invalidate textures for the former. - The Virtual range list is now non-overlapping like the physical one. - Fixed some bugs where regions could leak. - Introduced a weird bug that I still need to track down (consistent invalid buffer in MK8 ribbon road) Move some stuff. I think we'll eventually just put the dll and so for this in a nuget package. Fix rebase. [WIP] MultiRegionHandle variable size ranges - Avoid reprotecting regions that change often (needs some tweaking) - There's still a bug in buffers, somehow. - Might want different api for minimum granularity Fix rebase issue Commit everything needed for software only tracking. Remove native components. Remove more native stuff. Cleanup Use a separate window for the background context, update opentk. (fixes linux) Some experimental changes Should get things working up to scratch - still need to try some things with flush/modification and res scale. Include address with the region action. Initial work to make range tracking work Still a ton of bugs Fix some issues with the new stuff. * Fix texture flush instability There's still some weird behaviour, but it's much improved without this. (textures with cpu modified data were flushing over it) * Find the destination texture for Buffer->Texture full copy Greatly improves performance for nvdec videos (with range tracking) * Further improve texture tracking * Disable Memory Tracking for view parents This is a temporary approach to better match behaviour on master (where invalidations would be soaked up by views, rather than trigger twice) The assumption is that when views are created to a texture, they will cover all of its data anyways. Of course, this can easily be improved in future. * Introduce some tracking tests. WIP * Complete base tests. * Add more tests for multiregion, fix existing test. * Cleanup Part 1 * Remove unnecessary code from memory tracking * Fix some inconsistencies with 3D texture rule. * Add dispose tests. * Use a background thread for the background context. Rather than setting and unsetting a context as current, doing the work on a dedicated thread with signals seems to be a bit faster. Also nerf the multithreading test a bit. * Copy to texture with matching alignment This extends the copy to work for some videos with unusual size, such as tutorial videos in SMO. It will only occur if the destination texture already exists at XCount size. * Track reads for buffer copies. Synchronize new buffers before copying overlaps. * Remove old texture flushing mechanisms. Range tracking all the way, baby. * Wake the background thread when disposing. Avoids a deadlock when games are closed. * Address Feedback 1 * Separate TextureCopy instance for background thread Also `BackgroundContextWorker.InBackground` for a more sensible idenfifier for if we're in a background thread. * Add missing XML docs. * Address Feedback * Maybe I should start drinking coffee. * Some more feedback. * Remove flush warning, Refocus window after making background context
2020-10-16 21:18:35 +01:00
}
}