ed/raddebugger
1
0
Fork 0
mirror of https://github.com/Ed94/raddebugger.git synced 2026-07-01 15:41:48 -07:00
Code Issues Packages Projects Releases Wiki Activity

first pass at call stack cache, need a lot of clean up before this is working...

Browse Source
This commit is contained in:
Ryan Fleury
2025-05-14 21:44:09 -07:00
parent a84739e1c4
commit e5d6a49055
7 changed files with 132 additions and 69 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/ctrl/ctrl_core.c
+72 -26
View File
@@ -1443,7 +1443,7 @@ ctrl_scope_close(CTRL_Scope *scope)
internal void
ctrl_scope_touch_call_stack_node__stripe_r_guarded(CTRL_Scope *scope, CTRL_CallStackCacheNode *node)
{
ins_atomic_u64_eval(&node->scope_touch_count);
ins_atomic_u64_inc_eval(&node->scope_touch_count);
CTRL_ScopeCallStackTouch *touch = ctrl_tctx->free_call_stack_touch;
if(touch != 0)
{
@@ -3299,13 +3299,21 @@ ctrl_call_stack_from_unwind(Arena *arena, CTRL_Entity *process, CTRL_Unwind *bas
}
//- rjf: package
result.count = frame_count;
result.frames = push_array(arena, CTRL_CallStackFrame, result.count);
result.frames_count = frame_count;
result.frames = push_array(arena, CTRL_CallStackFrame, result.frames_count);
result.concrete_frames_count = base_unwind->frames.count;
result.concrete_frames = push_array(arena, CTRL_CallStackFrame *, result.concrete_frames_count);
{
U64 idx = 0;
U64 concrete_idx = 0;
for(FrameNode *n = first_frame; n != 0; n = n->next, idx += 1)
{
MemoryCopyStruct(&result.frames[idx], &n->v);
if(n->v.inline_depth == 0 && concrete_idx < result.concrete_frames_count)
{
result.concrete_frames[concrete_idx] = &result.frames[idx];
concrete_idx += 1;
}
}
}
}
@@ -3320,7 +3328,7 @@ ctrl_call_stack_frame_from_unwind_and_inline_depth(CTRL_CallStack *call_stack, U
CTRL_CallStackFrame *f = 0;
{
U64 base_frame_idx = 0;
for(U64 idx = 0; idx < call_stack->count; idx += 1)
for(U64 idx = 0; idx < call_stack->frames_count; idx += 1)
{
if(call_stack->frames[idx].inline_depth == 0)
{
@@ -3361,7 +3369,7 @@ ctrl_call_stack_from_thread(CTRL_Scope *scope, CTRL_Entity *thread, U64 endt_us)
//- rjf: loop: try to grab cached call stack; request; wait
U64 reg_gen = ctrl_reg_gen();
U64 mem_gen = ctrl_mem_gen();
OS_MutexScopeW(stripe->rw_mutex) for(;;)
OS_MutexScopeR(stripe->rw_mutex) for(;;)
{
//- rjf: try to grab cached
B32 is_good = 0;
@@ -3385,20 +3393,31 @@ ctrl_call_stack_from_thread(CTRL_Scope *scope, CTRL_Entity *thread, U64 endt_us)
}
//- rjf: create node if needed
if(!is_good)
if(!is_good) OS_MutexScopeRWPromote(stripe->rw_mutex)
{
node = push_array(stripe->arena, CTRL_CallStackCacheNode, 1);
DLLPushBack(slot->first, slot->last, node);
node->thread = thread->handle;
for(CTRL_CallStackCacheNode *n = slot->first; n != 0; n = n->next)
{
if(ctrl_handle_match(n->thread, handle))
{
node = n;
break;
}
}
if(node == 0)
{
node = push_array(stripe->arena, CTRL_CallStackCacheNode, 1);
DLLPushBack(slot->first, slot->last, node);
node->thread = thread->handle;
}
}
//- rjf: request if needed
if(!is_working && (!is_good || !is_stale))
if(!is_working && (!is_good || is_stale))
{
if(ctrl_u2csb_enqueue_req(thread->handle, endt_us) &&
async_push_work(ctrl_call_stack_build_work))
{
node->working_count += 1;
ins_atomic_u64_inc_eval(&node->working_count);
}
}
@@ -3409,7 +3428,7 @@ ctrl_call_stack_from_thread(CTRL_Scope *scope, CTRL_Entity *thread, U64 endt_us)
}
//- rjf: time to wait for new result? -> wait
os_condition_variable_wait_rw_w(stripe->cv, stripe->rw_mutex, endt_us);
os_condition_variable_wait_rw_r(stripe->cv, stripe->rw_mutex, endt_us);
}
}
return call_stack;
@@ -6784,30 +6803,47 @@ ASYNC_WORK_DEF(ctrl_call_stack_build_work)
U64 pre_reg_gen = ctrl_reg_gen();
U64 pre_mem_gen = ctrl_mem_gen();
Arena *arena = arena_alloc();
CTRL_Unwind unwind = ctrl_unwind_from_thread(arena, ctx, thread_handle, os_now_microseconds()+1000000);
CTRL_Unwind unwind = ctrl_unwind_from_thread(arena, ctx, thread_handle, 0);
CTRL_CallStack call_stack = ctrl_call_stack_from_unwind(arena, process, &unwind);
U64 post_reg_gen = ctrl_reg_gen();
U64 post_mem_gen = ctrl_mem_gen();
//- rjf: store in cache
//- rjf: store new results in cache
Arena *last_arena = arena;
OS_MutexScopeW(stripe->rw_mutex)
if(pre_reg_gen == post_reg_gen &&
pre_mem_gen == post_mem_gen)
{
for(CTRL_CallStackCacheNode *n = slot->first; n != 0; n = n->next)
for(B32 done = 0; !done;)
{
if(ctrl_handle_match(n->thread, thread_handle))
B32 found = 0;
OS_MutexScopeW(stripe->rw_mutex)
{
if(pre_reg_gen == post_reg_gen &&
pre_mem_gen == post_mem_gen)
for(CTRL_CallStackCacheNode *n = slot->first; n != 0; n = n->next)
{
last_arena = n->arena;
n->arena = arena;
n->reg_gen = pre_reg_gen;
n->mem_gen = pre_mem_gen;
n->unwind = unwind;
n->call_stack = call_stack;
if(ctrl_handle_match(n->thread, thread_handle))
{
found = 1;
if(n->scope_touch_count == 0)
{
done = 1;
if(unwind.flags == 0 || call_stack.frames_count >= n->call_stack.frames_count)
{
last_arena = n->arena;
n->arena = arena;
n->call_stack = call_stack;
}
if(unwind.flags == 0)
{
n->reg_gen = pre_reg_gen;
n->mem_gen = pre_mem_gen;
}
}
break;
}
}
n->working_count -= 1;
}
if(!found)
{
break;
}
}
@@ -6819,6 +6855,16 @@ ASYNC_WORK_DEF(ctrl_call_stack_build_work)
arena_release(last_arena);
}
//- rjf: mark work as done
OS_MutexScopeW(stripe->rw_mutex) for(CTRL_CallStackCacheNode *n = slot->first; n != 0; n = n->next)
{
if(ctrl_handle_match(n->thread, thread_handle))
{
ins_atomic_u64_dec_eval(&n->working_count);
break;
}
}
scratch_end(scratch);
}