Merge pull request #2107 from pguyot/w08/dump-largest-binaries-on-oom

Dump largest refc binaries on oom crashes

These changes are made under both the "Apache 2.0" and the "GNU Lesser General
Public License 2.1 or later" license terms (dual license).

SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later

Author: bettio

src/libAtomVM/context.c

@@ -1300,7 +1300,7 @@ COLD_FUNC void context_dump(Context *ctx)
         fprintf(stderr, "process_count = %zu\n", process_count);
         fprintf(stderr, "ports_count = %zu\n", ports_count);
         fprintf(stderr, "atoms_count = %zu\n", atom_table_count(glb->atom_table));
-        fprintf(stderr, "refc_binary_total_size = %zu\n", refc_binary_total_size(ctx));
+        refc_binary_dump_info(ctx);
     }
     fprintf(stderr, "\n\n**End Of Crash Report**\n");
 }

src/libAtomVM/refc_binary.c

@@ -146,3 +146,86 @@ size_t refc_binary_total_size(Context *ctx)
     synclist_unlock(&ctx->global->refc_binaries);
     return size;
 }
+
+COLD_FUNC void refc_binary_dump_info(Context *ctx)
+{
+    struct ListHead *item;
+    struct ListHead *refc_binaries = synclist_rdlock(&ctx->global->refc_binaries);
+
+    // Note: This only counts non-const refc binaries (ones that allocate memory).
+    // Const binaries (created by term_from_const_binary) point to existing data
+    // and are never added to the global refc_binaries list, so they don't appear here.
+
+    // First pass: count and calculate total size
+    size_t count = 0;
+    size_t total_size = 0;
+    LIST_FOR_EACH (item, refc_binaries) {
+        struct RefcBinary *refc = GET_LIST_ENTRY(item, struct RefcBinary, head);
+        count++;
+        total_size += refc->size;
+    }
+
+    fprintf(stderr, "refc_binary_count = %d\n", (int) count);
+    fprintf(stderr, "refc_binary_total_size = %d\n", (int) total_size);
+
+    if (count == 0) {
+        synclist_unlock(&ctx->global->refc_binaries);
+        return;
+    }
+
+// Find top 5 largest binaries
+#define TOP_N 5
+    struct RefcBinary *top[TOP_N] = { NULL };
+    size_t top_indices[TOP_N] = { 0 };
+
+    size_t index = 0;
+    LIST_FOR_EACH (item, refc_binaries) {
+        struct RefcBinary *refc = GET_LIST_ENTRY(item, struct RefcBinary, head);
+
+        // Try to insert into top 5
+        for (size_t i = 0; i < TOP_N; i++) {
+            if (top[i] == NULL || refc->size > top[i]->size) {
+                // Shift down
+                for (size_t j = TOP_N - 1; j > i; j--) {
+                    top[j] = top[j - 1];
+                    top_indices[j] = top_indices[j - 1];
+                }
+                top[i] = refc;
+                top_indices[i] = index;
+                break;
+            }
+        }
+        index++;
+    }
+
+    // Display top binaries
+    fprintf(stderr, "\nTop %d largest refc binaries:\n", TOP_N);
+    for (size_t i = 0; i < TOP_N && top[i] != NULL; i++) {
+        struct RefcBinary *refc = top[i];
+        fprintf(stderr, "  [%zu] size=%d bytes (%.1f%%), refcount=%d",
+            top_indices[i],
+            (int) refc->size,
+            (double) refc->size * 100.0 / (double) total_size,
+            (int) refc->ref_count);
+
+        if (refc->resource_type) {
+            fprintf(stderr, " [resource]");
+        }
+
+        // Print first 32 bytes as hex
+        fprintf(stderr, "\n      data: ");
+        size_t print_size = refc->size < 32 ? refc->size : 32;
+        for (size_t j = 0; j < print_size; j++) {
+            fprintf(stderr, "%02x", refc->data[j]);
+            if (j % 4 == 3 && j < print_size - 1) {
+                fprintf(stderr, " ");
+            }
+        }
+        if (refc->size > 32) {
+            fprintf(stderr, "...");
+        }
+        fprintf(stderr, "\n");
+    }
+
+    synclist_unlock(&ctx->global->refc_binaries);
+}

src/libAtomVM/refc_binary.h

@@ -142,6 +142,16 @@ term refc_binary_create_binary_info(Context *ctx);
  */
 size_t refc_binary_total_size(Context *ctx);
 
+/**
+ * @brief Dump detailed information about reference counted binaries
+ *
+ * @details This function prints diagnostic information including the count,
+ * total size, and details about the top 5 largest binaries including
+ * their first bytes. Used for debugging memory issues.
+ * @param ctx the context
+ */
+COLD_FUNC void refc_binary_dump_info(Context *ctx);
+
 #ifdef __cplusplus
 }
 #endif