ShuriZma
/
Cxbx-Reloaded
mirror of https://github.com/Cxbx-Reloaded/Cxbx-Reloaded.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

NV2A : Fixed hanging PFIFO (all that was needed, was to lock cache_lock at the start of the while loop in pfifo_puller_thread). 

Alas, this isn't the end of it, as now LLE ends up calling KeWaitForSingleObject on never ending repeat.
This commit is contained in:
PatrickvL 2018-02-16 09:44:01 +01:00
parent 8d8c9db930
commit 53154a84dc
1 changed files with 19 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
src/devices/video/EmuNV2A_PFIFO.cpp
View File

@ -38,7 +38,7 @@ DEVICE_READ32(PFIFO)
SET_MASK(result, NV_PFIFO_CACHE1_PUSH1_MODE, d->pfifo.cache1.mode);
break;
case NV_PFIFO_CACHE1_STATUS: {
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock); // UNTESTED
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock);
if (d->pfifo.cache1.cache.empty()) {
result |= NV_PFIFO_CACHE1_STATUS_LOW_MARK; /* low mark empty */
@ -79,12 +79,12 @@ DEVICE_READ32(PFIFO)
| d->pfifo.cache1.subroutine_active;
break;
case NV_PFIFO_CACHE1_PULL0: {
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock); // UNTESTED
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock);
result = d->pfifo.cache1.pull_enabled;
break;
}
case NV_PFIFO_CACHE1_ENGINE: {
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock); // UNTESTED
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock);
for (int i = 0; i < NV2A_NUM_SUBCHANNELS; i++) {
result |= d->pfifo.cache1.bound_engines[i] << (i * 2);
}
@ -171,7 +171,7 @@ DEVICE_WRITE32(PFIFO)
(value & NV_PFIFO_CACHE1_DMA_SUBROUTINE_STATE);
break;
case NV_PFIFO_CACHE1_PULL0: {
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock); // UNTESTED
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock);
if ((value & NV_PFIFO_CACHE1_PULL0_ACCESS)
&& !d->pfifo.cache1.pull_enabled) {
@ -188,7 +188,7 @@ DEVICE_WRITE32(PFIFO)
break;
}
case NV_PFIFO_CACHE1_ENGINE: {
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock); // UNTESTED
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock);
for (i = 0; i < NV2A_NUM_SUBCHANNELS; i++) {
d->pfifo.cache1.bound_engines[i] = (FIFOEngine)((value >> (i * 2)) & 3);
@ -234,7 +234,8 @@ static void pfifo_run_pusher(NV2AState *d) {
channel_id = state->channel_id;
control = &d->user.channel_control[channel_id];
if (!state->push_enabled) return;
if (!state->push_enabled)
return;
/* only handling DMA for now... */
@ -244,8 +245,10 @@ static void pfifo_run_pusher(NV2AState *d) {
assert(channel_modes & (1 << channel_id));
assert(state->mode == FIFO_DMA);
if (!state->dma_push_enabled) return;
if (state->dma_push_suspended) return;
if (!state->dma_push_enabled)
return;
if (state->dma_push_suspended)
return;
/* We're running so there should be no pending errors... */
assert(state->error == NV_PFIFO_CACHE1_DMA_STATE_ERROR_NONE);
@ -370,8 +373,11 @@ int pfifo_puller_thread(NV2AState *d)
std::unique_lock<std::mutex> cache_unique_lock(d->pfifo.cache1.cache_lock, std::defer_lock);
while (true) {
cache_unique_lock.lock();
while (state->cache.empty() || !state->pull_enabled) {
state->cache_cond.wait(cache_unique_lock); // UNTESTED
state->cache_cond.wait(cache_unique_lock);
if (d->exiting) {
cache_unique_lock.unlock(); // UNTESTED
@ -435,9 +441,9 @@ int pfifo_puller_thread(NV2AState *d)
//qemu_mutex_unlock_iothread();
}
// qemu_mutex_lock(&state->cache_lock);
// state->cache_lock.lock();
enum FIFOEngine engine = state->bound_engines[command->subchannel];
// qemu_mutex_unlock(&state->cache_lock);
// state->cache_lock.unlock();
switch (engine) {
case ENGINE_GRAPHICS:
@ -450,9 +456,9 @@ int pfifo_puller_thread(NV2AState *d)
break;
}
// qemu_mutex_lock(&state->cache_lock);
// state->cache_lock.lock();
state->last_engine = state->bound_engines[command->subchannel];
// qemu_mutex_unlock(&state->cache_lock);
// state->cache_lock.unlock();
}
g_free(command);