dvd/dvd_dump.ixx (1)

854-859: Buffer allocation precedes sectors_at_once adjustment.

file_data and file_state are allocated at lines 830–831 using the original sectors_at_once value, but the adjustment to 16 happens here. The buffers end up oversized but never undersized, so there's no correctness issue — just a minor inefficiency.

Consider moving the slim drive adjustment before the buffer allocation for clarity and to avoid the over-allocation.

♻️ Suggested reordering

+    // TODO: can be implemented later
+    if(raw)
+    {
+        rom_update = false;
+        fs_ctx.search = false;
+
+        if(is_omnidrive_slim(ctx.drive_config) && sectors_at_once >= 32)
+        {
+            LOG("warning: setting dump read size to 16 for raw dumping");
+            sectors_at_once = 16;
+        }
+    }
+
     std::vector<uint8_t> file_data(sectors_at_once * cfg.sector_size);
     std::vector<State> file_state(sectors_at_once);
-    ...
-    // TODO: can be implemented later
-    if(raw)
-    {
-        rom_update = false;
-        fs_ctx.search = false;
-
-        if(is_omnidrive_slim(ctx.drive_config) && sectors_at_once >= 32)
-        {
-            LOG("warning: setting dump read size to 16 for raw dumping");
-            sectors_at_once = 16;
-        }
-    }

dvd/dvd_dump.ixx (1)

936-938: Signed/unsigned comparison is correct in practice but fragile.

lba is int32_t while sectors_count and sectors_to_read are uint32_t. When lba is negative (raw DVD pre-data area), implicit conversion to unsigned makes lba < sectors_count evaluate to false, which happens to be the desired behavior — but only by accident of two's complement wrapping.

An explicit cast or guard makes the intent clearer and future-proof:

Suggested clarification

-        // ensure requested sectors don't cross lead-out boundary
-        if(lba < sectors_count && lba + sectors_to_read > sectors_count)
-            sectors_to_read = sectors_count - lba;
+        // ensure requested sectors don't cross lead-out boundary
+        if(lba >= 0 && static_cast<uint32_t>(lba) < sectors_count && static_cast<uint32_t>(lba) + sectors_to_read > sectors_count)
+            sectors_to_read = sectors_count - static_cast<uint32_t>(lba);