From 276b8374284be9a04feb52e3c9613c90bc6d3cb9 Mon Sep 17 00:00:00 2001
From: pompolic <pompolic@special-circumstanc.es>
Date: Sun, 14 May 2023 19:46:45 +0200
Subject: [PATCH] (WIP) Reorganize pending allocations data structure
---
gdb-port/parser.py | 11 +++++++----
gdb-port/top-level-parse.py | 18 +++++++++---------
2 files changed, 16 insertions(+), 13 deletions(-)
diff --git a/gdb-port/parser.py b/gdb-port/parser.py
index 7a907b4..cbc97e6 100644
--- a/gdb-port/parser.py
+++ b/gdb-port/parser.py
@@ -301,7 +301,8 @@ class ParserStack:
#current_event[2].add_mem_use(int(self.arena), current_frame_alloc)
self.committed = len(self.stack_events)-1
# print("commit_at_pop: pop_allocs:", { current_event[2].address: current_frame_alloc }) # DEBUG
- pop_allocs[current_event[2].address] = pop_allocs.get(current_event[2].address, 0) + current_frame_alloc
+ # This assumes the value of self.arena and self.arena_int won't change while there are pending allocations, otherwise it might drop allocations belonging to the old arena
+ pop_allocs[current_event[2].address] = { self.arena_int : pop_allocs.get(self.arena_int, {}).get(current_event[2].address, 0) + current_frame_alloc }
parser_cache[current_event[2].address] = current_event[2]
current_event[2].add_mem_use(self.arena_int, current_frame_alloc)
#profiler.disable() # DEBUG
@@ -318,7 +319,7 @@ class ParserStack:
for index, event in enumerate(ev_list[:-1]):
parser_cache[int(event[2].address)] = event[2]
alloc = bigger[index][1] - smaller[index][1]
- pop_allocs[event[2].address] = pop_allocs.get(event[2].address, 0) + alloc
+ pop_allocs[event[2].address] = { self.arena_int : pop_allocs.get(self.arena_int, {}).get(event[2].address, 0) + alloc }
#alloc = next(differences)
#print("commit_at_pop: alloc:", alloc) # DEBUG
#print("commit_at_pop: adding to parser", event[2].name, hex(event[2].address))
@@ -356,7 +357,7 @@ class ParserStack:
for index, event in enumerate(ev_list[:-1]):
alloc = next(differences)
#event[2].add_mem_use(self.arena_int, alloc)
- push_allocs[event[2].address] = push_allocs.get(event[2].address, 0) + alloc
+ push_allocs[event[2].address] = { self.arena_int : push_allocs.get(self.arena_int, {}).get(event[2].address, 0) + alloc }
#print("commit_at_push: push_allocs:", push_allocs) # DEBUG
# Most recent event is a push event, and we have no information on its allocations yet. So we only increment self.committed up to the index of the previous event
@@ -372,4 +373,6 @@ class ParserStack:
return self.pending_allocs
def clear_pending_allocations(self):
- self.pending_allocs = { self.arena_int : {} }
+ for k, v in self.pending_allocs.items():
+ v.clear()
+ self.pending_allocs = {}
diff --git a/gdb-port/top-level-parse.py b/gdb-port/top-level-parse.py
index ff2af3b..77e553f 100644
--- a/gdb-port/top-level-parse.py
+++ b/gdb-port/top-level-parse.py
@@ -274,12 +274,12 @@ class TopLevelParse:
def gather_allocations_in_parser_stacks(self):
for ps in self.parser_stacks:
- new_allocs_by_arena = ps.get_pending_allocations()
- new_allocs = new_allocs_by_arena.get(ps.arena_int, {})
- pending_allocs = self.pending_allocs.get(ps.arena_int, {})
- for addr, alloc in new_allocs.items():
- pending_allocs[addr] = pending_allocs.get(addr, 0) + alloc
- self.pending_allocs[ps.arena_int] = pending_allocs
+ new_allocs_by_parser = ps.get_pending_allocations()
+ for parser, alloc in new_allocs_by_parser.items():
+ pending_allocs = self.pending_allocs.get(parser, {})
+ pending_allocs[ps.arena_int] = pending_allocs.get(ps.arena_int, 0) + alloc.get(ps.arena_int, 0)
+ self.pending_allocs[parser] = pending_allocs
+ ps.clear_pending_allocations()
# TODO: allocations are grouped by arena, then parser address. we get the arena address from the parser stack, which may be shared between parser stacks
@@ -287,9 +287,9 @@ class TopLevelParse:
# However, the problem is that doing this naively would require two nested for loops minimum
# TODO: check if grouping { parser : { arena : alloc } } would work better
def commit_pending_allocations(self):
- for arena, allocs in self.pending_allocs.items():
- for addr, alloc in allocs.items():
- self.parser_by_address(addr).add_mem_use(arena, alloc)
+ for parser, allocs in self.pending_allocs.items():
+ for arena, alloc in allocs.items():
+ self.parser_by_address(parser).add_mem_use(arena, alloc)
self.pending_allocs = {}
top_level_parse = TopLevelParse()
--
GitLab