From b986f49e1a97c7ce9d6686dcb8a3124dd2b990a2 Mon Sep 17 00:00:00 2001
From: Kaj Schittecat <kaj@schittecat.com>
Date: Fri, 26 Jun 2026 19:11:13 +0200
Subject: [PATCH] =?UTF-8?q?core:=20move=20the=5Fmesh=20(~42=20KB)=20to=20P?=
 =?UTF-8?q?SRAM=20=E2=80=94=20reclaim=20internal=20DRAM?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

the_mesh was the single biggest static internal-DRAM consumer at 42 KB (dominated
by the MAX_CONTACTS=100 ContactInfo array inside it). Internal DRAM (~320 KB) is the
scarce resource on the ESP32-S3 — Wi-Fi/BLE DMA + heap fight over it, and the V4 was
hitting ~99% on map-tile load (issue #47b). PSRAM (8 MB) is plentiful.

Place the whole the_mesh object in PSRAM via placement-new into a heap_caps_malloc
(MALLOC_CAP_SPIRAM) block and bind a reference, so every the_mesh.foo() call site is
unchanged. The constructor still runs at the SAME static-init point as the old direct
global (PSRAM is already up then), so timing and behaviour are identical; only the
address moves off internal DRAM. Falls back to internal RAM if PSRAM is absent.
ESP32-only (STM32/RP2040 keep the plain global). The contacts array rides along
inside the object — no vendored-core change.

Measured on heltec_v4 (.dram0.bss): 124160 -> 82000 B = ~41 KB internal DRAM freed.
Verified on the T-Deck: clean boot, [BOOT] mesh ok, reason=0, zero panic markers.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---
 src/MyMesh.h |  4 ++++
 src/main.cpp | 24 ++++++++++++++++++++++++
 2 files changed, 28 insertions(+)

diff --git a/src/MyMesh.h b/src/MyMesh.h
index 74502f0..44b6752 100644
--- a/src/MyMesh.h
+++ b/src/MyMesh.h
@@ -951,4 +951,8 @@ private:
   uint32_t _ui_trace_ping_tag = 0;
 };
 
+#if defined(ESP32_PLATFORM)
+extern MyMesh& the_mesh;   // PSRAM-resident (placement-new'd in main.cpp); reference keeps call sites unchanged
+#else
 extern MyMesh the_mesh;
+#endif
diff --git a/src/main.cpp b/src/main.cpp
index 02e342d..335efa1 100644
--- a/src/main.cpp
+++ b/src/main.cpp
@@ -1,6 +1,10 @@
 #include <Arduino.h>   // needed for PlatformIO
 #include <Mesh.h>
 #include "MyMesh.h"
+#if defined(ESP32_PLATFORM)
+  #include <new>               // placement-new for the PSRAM-resident the_mesh
+  #include "esp_heap_caps.h"   // heap_caps_malloc(MALLOC_CAP_SPIRAM)
+#endif
 #if defined(ESP32_PLATFORM) && defined(HAS_TOUCH_UI)
 #include <Preferences.h>
 #include <esp_system.h>
@@ -127,11 +131,31 @@ static uint32_t _atoi(const char* sp) {
 
 StdRNG fast_rng;
 SimpleMeshTables tables;
+#if defined(ESP32_PLATFORM)
+// the_mesh is ~42 KB (dominated by the MAX_CONTACTS ContactInfo array) and was the
+// single biggest static internal-DRAM consumer. Place the whole object in PSRAM —
+// the contacts array rides along inside it — and bind a reference so every
+// `the_mesh.foo()` call site is unchanged. The constructor still runs HERE at
+// static-init (PSRAM is already up; the UITask psAlloc statics rely on the same),
+// so timing/behaviour are identical to the old direct global — only the address
+// moves off internal DRAM. heap_caps falls back to internal RAM if PSRAM is absent.
+static MyMesh& makeTheMesh() {
+  void* mem = heap_caps_malloc(sizeof(MyMesh), MALLOC_CAP_SPIRAM | MALLOC_CAP_8BIT);
+  if (!mem) mem = malloc(sizeof(MyMesh));   // no-PSRAM fallback: behaves as before
+  return *new (mem) MyMesh(radio_driver, fast_rng, rtc_clock, tables, store
+   #ifdef DISPLAY_CLASS
+      , &ui_task
+   #endif
+  );
+}
+MyMesh& the_mesh = makeTheMesh();
+#else
 MyMesh the_mesh(radio_driver, fast_rng, rtc_clock, tables, store
    #ifdef DISPLAY_CLASS
       , &ui_task
    #endif
 );
+#endif
 
 /* END GLOBAL OBJECTS */