areusch commented on a change in pull request #7785:
URL: https://github.com/apache/tvm/pull/7785#discussion_r619380889
##########
File path: src/runtime/crt/memory/stack_allocator.c
##########
@@ -16,17 +16,22 @@
* specific language governing permissions and limitations
* under the License.
*/
-
// LINT_C_FILE
-
#include <tvm/runtime/crt/stack_allocator.h>
+#ifdef TVM_CRT_DEBUG
+#include <tvm/runtime/crt/logging.h>
+#endif
void* StackMemoryManager_Allocate(tvm_workspace_t* tvm_runtime_workspace,
int32_t nbytes) {
uint32_t offset_bytes = (~nbytes + 1) & (TVM_RUNTIME_ALLOC_ALIGNMENT_BYTES -
1);
uint8_t* current_alloc = tvm_runtime_workspace->next_alloc;
uint8_t* next_alloc = tvm_runtime_workspace->next_alloc + nbytes +
offset_bytes;
uint8_t* workspace_end = tvm_runtime_workspace->workspace +
tvm_runtime_workspace->workspace_size;
-
+#ifdef TVM_CRT_DEBUG
Review comment:
the case i'm thinking about is where other code running on the user
device overwrites e.g. `next_alloc` or the `ptr` passed to Free(). in this
case, the error is likely in the user application rather than in TVM codegen,
and this allocator would be a bit of a black box (of course, they could inspect
the source code, but it would be helpful to provide an indication of memory
corruption at the earliest point it could be detected).
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
