Change Malloc to request an aligned memory size.
This fixes a problem on Linux where allocated memory would get overun in some use cases (ie. in TestExprs2.py).
llvm-svn: 180614
diff --git a/lldb/source/Expression/IRMemoryMap.cpp b/lldb/source/Expression/IRMemoryMap.cpp
index db5b96f..be8a589 100644
--- a/lldb/source/Expression/IRMemoryMap.cpp
+++ b/lldb/source/Expression/IRMemoryMap.cpp
@@ -235,8 +235,14 @@
lldb::ProcessSP process_sp;
lldb::addr_t allocation_address = LLDB_INVALID_ADDRESS;
lldb::addr_t aligned_address = LLDB_INVALID_ADDRESS;
-
- size_t allocation_size = (size ? size : 1) + alignment - 1;
+
+ size_t alignment_mask = alignment - 1;
+ size_t allocation_size;
+
+ if (size == 0)
+ allocation_size = alignment;
+ else
+ allocation_size = (size & alignment_mask) ? ((size + alignment) & (~alignment_mask)) : size;
switch (policy)
{
@@ -305,7 +311,7 @@
m_allocations[aligned_address] = Allocation(allocation_address,
aligned_address,
- size,
+ allocation_size,
permissions,
alignment,
policy);
@@ -331,7 +337,7 @@
}
log->Printf("IRMemoryMap::Malloc (%llu, 0x%llx, 0x%llx, %s) -> 0x%llx",
- (uint64_t)size,
+ (uint64_t)allocation_size,
(uint64_t)alignment,
(uint64_t)permissions,
policy_string,