refactor: promote new architecture to repository root

Archive the deprecated implementation under lume-legacy, make the new kernel the default root-level build, and update repository entry points without changing kernel behavior.

1 files changed, 235 insertions, 0 deletions

diff --git a/mm/vmm.cc b/mm/vmm.cc
new file mode 100644
index 0000000..51617e3
--- /dev/null
+++ b/mm/vmm.cc
@@ -0,0 +1,235 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/* Copyright (C) 2026 Richard Qin */
+
+/*
+ * vmm.cc — Virtual Memory Manager (Phase 1: Kernel Page Table)
+ *
+ * Builds a fine-grained kernel page table, activates it, and retires
+ * the early_pgdir identity mapping from entry.S.
+ *
+ * With memblock integration, the initialization order is:
+ *   memblock_init → vmm_init (allocates page_map from memblock,
+ *   maps vmemmap, activates page table) → pmm_init (buddy system
+ *   starts with page_map already at vmemmap VA — no relocation).
+ *
+ * Reference: docs/specs/vmm.md
+ */
+
+#include <arch/config.h>
+#include <arch/mmu.h>
+#include <arch/pmap.h>
+#include <lume/tlb.h>
+#include <lume/addr.h>
+#include <lume/console.h>
+#include <lume/fdt.h>
+#include <lume/page.h>
+#include <lume/klog.h>
+#include <lume/kprintf.h>
+#include <lume/memblock.h>
+#include <lume/pmm.h>
+#include <lume/types.h>
+#include <lume/vmm.h>
+#include <lume/config.h>
+
+/* Linker-exported section boundaries (virtual addresses) */
+extern "C" char _stext[], _etext[];
+extern "C" char _srodata[], _erodata[];
+extern "C" char _sdata[], _edata[];
+extern "C" char _sbss[], _ebss[];
+extern "C" char _kernel_end[];
+
+/* Global kernel page table root PA — external linkage for task_exit */
+uint64 g_kernel_pgtbl = 0;
+
+namespace {
+
+/* VMM exports for PMM: page_map location after vmemmap mapping. */
+uint64 g_vmemmap_page_map_pa = 0;
+uint64 g_vmemmap_page_map_size = 0;
+uint64 g_vmemmap_num_pages = 0;
+
+/* Convert a linker symbol to PA (kernel is executing in higher-half VA). */
+inline uint64 sym_to_pa(const void *sym) {
+  return boot_va_to_pa(virt_addr(reinterpret_cast<uint64>(sym))).raw;
+}
+
+/* Map a PA range into the kernel page table via the direct map. */
+void map_range(uint64 root, uint64 pa_start, uint64 pa_end,
+                      uint64 perm) {
+  for (uint64 pa = pa_start; pa < pa_end; pa += kPageSize) {
+    uint64 va = pa_to_va(phys_addr(pa)).raw;
+    int ret = pmap::map(root, va, pa, perm | pmap::PTE_G);
+    if (ret != 0) {
+      kernel_panic("vmm_init: pmap::map failed");
+    }
+  }
+}
+
+/* Map a PA range into the kernel page table at the fixed linked VA. */
+void map_range_linked(uint64 root, uint64 pa_start, uint64 pa_end,
+                             uint64 perm) {
+  for (uint64 pa = pa_start; pa < pa_end; pa += kPageSize) {
+    uint64 va = boot_pa_to_va(phys_addr(pa)).raw;
+    int ret = pmap::map(root, va, pa, perm | pmap::PTE_G);
+    if (ret != 0) {
+      kernel_panic("vmm_init: pmap::map (linked) failed");
+    }
+  }
+}
+
+/* Map a PA range to an explicit VA base, using 2MB pages where possible. */
+void map_range_at_huge(uint64 root, uint64 va_base, uint64 pa_start,
+                              uint64 pa_end, uint64 perm) {
+  constexpr uint64 k2MB = 2ULL * 1024 * 1024;
+  uint64 pa = pa_start;
+  uint64 va = va_base;
+
+  /* Map 2MB-aligned chunks with mega pages */
+  while (pa + k2MB <= pa_end) {
+    if ((va & (k2MB - 1)) == 0 && (pa & (k2MB - 1)) == 0) {
+      int ret = pmap::map_2mb(root, va, pa, perm | pmap::PTE_G);
+      if (ret != 0) {
+        kernel_panic("vmm_init: pmap::map_2mb failed");
+      }
+      pa += k2MB;
+      va += k2MB;
+    } else {
+      break;
+    }
+  }
+
+  /* Map remaining 4KB tail */
+  while (pa < pa_end) {
+    int ret = pmap::map(root, va, pa, perm | pmap::PTE_G);
+    if (ret != 0) {
+      kernel_panic("vmm_init: vmemmap 4KB pmap::map failed");
+    }
+    pa += kPageSize;
+    va += kPageSize;
+  }
+}
+
+} // namespace
+
+uint64 vmm_get_page_map_pa() { return g_vmemmap_page_map_pa; }
+uint64 vmm_get_page_map_size() { return g_vmemmap_page_map_size; }
+uint64 vmm_get_num_pages() { return g_vmemmap_num_pages; }
+
+void vmm_init() {
+  /* 1. Allocate new root page table (from memblock) */
+  uint64 root = pmap::create();
+  if (!root) {
+    kernel_panic("vmm_init: failed to create root page table");
+  }
+
+  /* 2. Compute PA boundaries of each kernel section */
+  uint64 text_start = page_align_down(sym_to_pa(_stext));
+  uint64 text_end = page_align_up(sym_to_pa(_etext));
+  uint64 ro_start = page_align_down(sym_to_pa(_srodata));
+  uint64 ro_end = page_align_up(sym_to_pa(_erodata));
+  uint64 data_start = page_align_down(sym_to_pa(_sdata));
+  uint64 bss_end = page_align_up(sym_to_pa(_ebss));
+
+  /* Memory extent from memblock */
+  uint64 mem_base = memblock_get_mem_base();
+  uint64 mem_size = memblock_get_mem_size();
+  uint64 mem_end = mem_base + mem_size;
+
+  /* 3. Map kernel image at FIXED linked VA.
+   *
+   * Kernel text/data are linked at the bootstrap higher-half addresses
+   * (kDirectMapBaseDefault + PA). They must remain executable/readable at
+   * those addresses until full kernel-image relocation is implemented.
+   */
+  map_range_linked(root, text_start, text_end, pmap::PTE_R | pmap::PTE_X);
+
+  /* 4. Map .rodata (R) at fixed linked VA.
+   * We map [_srodata, _sdata) so that the .lume.reloc section (which
+   * sits in its own padded output section between _erodata and _sdata)
+   * is also covered.  Relocator only reads it during early boot, but
+   * mapping it keeps post-vmm accesses to __lume_reloc_* symbols sane. */
+  map_range_linked(root, ro_start, page_align_down(sym_to_pa(_sdata)),
+                   pmap::PTE_R);
+
+  /* 5. Map .data + .bss (RW) at fixed linked VA */
+  map_range_linked(root, data_start, bss_end, pmap::PTE_R | pmap::PTE_W);
+
+  /* 6. Map usable physical RAM into the KASLR-slid direct map.
+   *
+   * The direct map provides pa_to_va() / va_to_pa() for the physical
+   * address space.  We map from kPhysStart (the kernel's linked load
+   * address), skipping [mem_base, kPhysStart) because that region is
+   * PMP-protected firmware (OpenSBI) — S-mode cannot access it.
+   *
+   * Linux does the same on RISC-V: the linear map covers only the
+   * usable RAM range, not firmware-reserved regions.
+   *
+   * Uses 2MB huge pages for efficiency (64 entries for 128MB vs 32K).
+   */
+  uint64 dm_start = kPhysStart;
+  if (dm_start < mem_base) {
+    dm_start = mem_base;
+  }
+  uint64 dm_va_base = pa_to_va(phys_addr(dm_start)).raw;
+  map_range_at_huge(root, dm_va_base, dm_start, mem_end,
+                    pmap::PTE_R | pmap::PTE_W);
+
+  /* 7. Allocate page_map from memblock (top of RAM) and map at VMEMMAP.
+   *
+   * Since memblock is active (PMM not yet initialized), pmap::map
+   * allocates page table pages from memblock too. page_map starts
+   * life at the vmemmap VA — no relocation ever needed.
+   */
+  uint64 num_pages = mem_size / kPageSize;
+  uint64 page_map_bytes = num_pages * sizeof(Page);
+  uint64 page_map_span = page_align_up(page_map_bytes);
+  uint64 page_map_pa = memblock_alloc_top(page_map_span, kPageSize);
+
+  if (!page_map_pa) {
+    kernel_panic("vmm_init: memblock failed to allocate page_map");
+  }
+
+  kprintf("[vmm] vmemmap: %llu pages, %llu KB at PA 0x%llx\n", num_pages,
+          page_map_span / 1024, page_map_pa);
+
+  /* Map page_map at vmemmap_base from KaslrLayout (randomised by KASLR), using 2MB pages where possible */
+  map_range_at_huge(root, arch::kaslr_layout_get().vmemmap_base, page_map_pa,
+                    page_map_pa + page_map_span, pmap::PTE_R | pmap::PTE_W);
+
+  /* Export for pmm_init() */
+  g_vmemmap_page_map_pa = page_map_pa;
+  g_vmemmap_page_map_size = page_map_bytes;
+  g_vmemmap_num_pages = num_pages;
+
+  /* 8. Pre-map console MMIO before page table switch */
+  g_kernel_pgtbl = root;
+  console_init();
+
+  /* 9. Activate the new page table */
+  pmap::activate(root);
+
+  /* Console MMIO access must now use runtime (KASLR) direct-map base. */
+  console_use_runtime_mapping();
+
+  kprintf("[vmm] vmm_init complete (vmemmap at 0x%llx)\n", arch::kaslr_layout_get().vmemmap_base);
+}
+
+void vmm_init_ap() {
+  if (!g_kernel_pgtbl) {
+    kernel_panic("vmm_init_ap: kernel page table not ready");
+  }
+  pmap::activate(g_kernel_pgtbl);
+}
+
+void vmm_map_kernel_mmio(uint64 pa, uint64 size) {
+  if (!g_kernel_pgtbl) {
+    kernel_panic("vmm_map_kernel_mmio: kernel page table not ready");
+  }
+
+  uint64 aligned_pa = page_align_down(pa);
+  uint64 aligned_size = page_align_up(pa + size) - aligned_pa;
+
+  map_range(g_kernel_pgtbl, aligned_pa, aligned_pa + aligned_size,
+            pmap::PTE_R | pmap::PTE_W | pmap::PTE_G);
+  tlb_shootdown_all();
+}