Replace hard coded reservations in x86-64 early boot code with dynamic table

Index: linux/arch/x86/kernel/e820_64.c
===================================================================
@@ -47,56 +47,65 @@ unsigned long end_pfn_map;
  */
 static unsigned long __initdata end_user_pfn = MAXMEM>>PAGE_SHIFT;
 
-/* Check for some hardcoded bad areas that early boot is not allowed to touch */
-static inline int bad_addr(unsigned long *addrp, unsigned long size)
-{
-	unsigned long addr = *addrp, last = addr + size;
+/*
+ * Early reserved memory areas.
+ */
+#define MAX_EARLY_RES 20
 
-	/* various gunk below that needed for SMP startup */
-	if (addr < 0x8000) {
-		*addrp = PAGE_ALIGN(0x8000);
-		return 1;
-	}
+struct early_res {
+	unsigned long start, end;
+};
+static struct early_res early_res[MAX_EARLY_RES] __initdata = {
+	{ 0, PAGE_SIZE },			/* BIOS data page */
+#ifdef CONFIG_SMP
+	{ SMP_TRAMPOLINE_BASE, SMP_TRAMPOLINE_BASE + 2*PAGE_SIZE },
+#endif
+	{}
+};
 
-	/* direct mapping tables of the kernel */
-	if (last >= table_start<<PAGE_SHIFT && addr < table_end<<PAGE_SHIFT) {
-		*addrp = PAGE_ALIGN(table_end << PAGE_SHIFT);
-		return 1;
+void __init reserve_early(unsigned long start, unsigned long end)
+{
+	int i;
+	struct early_res *r;
+	for (i = 0; i < MAX_EARLY_RES && early_res[i].end; i++) {
+		r = &early_res[i];
+		if (end > r->start && start < r->end)
+			panic("Duplicated early reservation %lx-%lx\n",
+			      start, end);
 	}
+	if (i >= MAX_EARLY_RES)
+		panic("Too many early reservations");
+	r = &early_res[i];
+	r->start = start;
+	r->end = end;
+}
 
-	/* initrd */
-#ifdef CONFIG_BLK_DEV_INITRD
-	if (boot_params.hdr.type_of_loader && boot_params.hdr.ramdisk_image) {
-		unsigned long ramdisk_image = boot_params.hdr.ramdisk_image;
-		unsigned long ramdisk_size  = boot_params.hdr.ramdisk_size;
-		unsigned long ramdisk_end   = ramdisk_image+ramdisk_size;
+void __init early_res_to_bootmem(void)
+{
+	int i;
+	for (i = 0; i < MAX_EARLY_RES && early_res[i].end; i++) {
+	struct early_res *r = &early_res[i];
+		reserve_bootmem_generic(r->start, r->end - r->start);
+ 	}
+}
 
-		if (last >= ramdisk_image && addr < ramdisk_end) {
-			*addrp = PAGE_ALIGN(ramdisk_end);
-			return 1;
+/* Check for already reserved areas */
+static inline int bad_addr(unsigned long *addrp, unsigned long size)
+{
+	int i;
+	unsigned long addr = *addrp, last;
+	int changed = 0;
+again:
+	last = addr + size;
+	for (i = 0; i < MAX_EARLY_RES && early_res[i].end; i++) {
+		struct early_res *r = &early_res[i];
+		if (last >= r->start && addr < r->end) {
+			*addrp = addr = r->end;
+			changed = 1;
+			goto again;
 		}
-	}
-#endif
-	/* kernel code */
-	if (last >= __pa_symbol(&_text) && addr < __pa_symbol(&_end)) {
-		*addrp = PAGE_ALIGN(__pa_symbol(&_end));
-		return 1;
-	}
-
-	if (last >= ebda_addr && addr < ebda_addr + ebda_size) {
-		*addrp = PAGE_ALIGN(ebda_addr + ebda_size);
-		return 1;
-	}
-
-#ifdef CONFIG_NUMA
-	/* NUMA memory to node map */
-	if (last >= nodemap_addr && addr < nodemap_addr + nodemap_size) {
-		*addrp = nodemap_addr + nodemap_size;
-		return 1;
-	}
-#endif
-	/* XXX ramdisk image here? */
-	return 0;
+ 	}
+	return changed;
 }
 
 /*

65 lines not on this page · raw message

91 lines not on this page · raw message

Index: linux/arch/x86/mm/numa_64.c
===================================================================
@@ -99,6 +99,7 @@ static int __init allocate_cachealigned_
 	}
 	pad_addr = (nodemap_addr + pad) & ~pad;
 	memnodemap = phys_to_virt(pad_addr);
+	reserve_early(nodemap_addr, nodemap_addr + nodemap_size);
 
 	printk(KERN_DEBUG "NUMA: Allocated memnodemap from %lx - %lx\n",
 	       nodemap_addr, nodemap_addr + nodemap_size);

Index: linux/include/asm-x86/e820_64.h
===================================================================
@@ -36,8 +36,9 @@ extern void finish_e820_parsing(void);
 
 extern struct e820map e820;
 
-extern unsigned ebda_addr, ebda_size;
-extern unsigned long nodemap_addr, nodemap_size;
+extern void reserve_early(unsigned long start, unsigned long end);
+extern void early_res_to_bootmem(void);
+
 #endif/*!__ASSEMBLY__*/
 
 #endif/*__E820_HEADER*/

Index: linux/arch/x86/mm/init_64.c
===================================================================
@@ -176,7 +176,8 @@ __set_fixmap (enum fixed_addresses idx, 
 	set_pte_phys(address, phys, prot);
 }
 
-unsigned long __meminitdata table_start, table_end;
+static unsigned long __initdata table_start;
+static unsigned long __meminitdata table_end;
 
 static __meminit void *alloc_low_page(unsigned long *phys)
 { 

@@ -387,6 +388,8 @@ void __init_refok init_memory_mapping(un
 	if (!after_bootmem)
 		mmu_cr4_features = read_cr4();
 	__flush_tlb_all();
+
+	reserve_early(table_start << PAGE_SHIFT, table_end << PAGE_SHIFT);
 }
 
 #ifndef CONFIG_NUMA

Index: linux/include/asm-x86/proto.h
===================================================================
@@ -22,8 +22,6 @@ extern void syscall32_cpu_init(void);
 
 extern void check_efer(void);
 
-extern unsigned long table_start, table_end;
-
 extern int reboot_force;
 
 long do_arch_prctl(struct task_struct *task, int code, unsigned long addr);

Index: linux/arch/x86/kernel/e820_64.c
===================================================================
@@ -819,3 +819,17 @@ int __init arch_get_ram_range(int slot, 
 		max_pfn << PAGE_SHIFT) - *addr;
 	return i + 1;
 }
+
+#define EARLY_ALLOC_START (128<<20)
+__init void *arch_early_alloc(unsigned long size)
+{
+	unsigned long p = find_e820_area(EARLY_ALLOC_START, -1UL, size);
+	if (p == -1ULL) {
+		/* Risk filling the DMA zone */
+		p = find_e820_area(0, -1UL, size);
+		if (p == -1ULL)
+			panic("arch_early_alloc %lu failed", size);
+	}
+	reserve_early(p, p + size);
+	return __va(p);
+}

Index: linux/include/asm-x86/pgtable_64.h
===================================================================
@@ -436,6 +436,9 @@ pte_t *lookup_address(unsigned long addr
 #define	kc_offset_to_vaddr(o) \
    (((o) & (1UL << (__VIRTUAL_MASK_SHIFT-1))) ? ((o) | (~__VIRTUAL_MASK)) : (o))
 
+#define ARCH_HAS_EARLY_ALLOC
+extern void *arch_early_alloc(unsigned long size);
+
 #define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
 #define __HAVE_ARCH_PTEP_GET_AND_CLEAR
 #define __HAVE_ARCH_PTEP_GET_AND_CLEAR_FULL

Index: linux/kernel/lockdep.c
===================================================================
@@ -109,7 +109,7 @@ static inline int debug_locks_off_graph_
 static int lockdep_initialized;
 
 unsigned long nr_list_entries;
-static struct lock_list list_entries[MAX_LOCKDEP_ENTRIES];
+static struct lock_list *list_entries;
 
 /*
  * All data structures here are protected by the global debug_lock.

@@ -118,7 +118,7 @@ static struct lock_list list_entries[MAX
  * get freed - this significantly simplifies the debugging code.
  */
 unsigned long nr_lock_classes;
-static struct lock_class lock_classes[MAX_LOCKDEP_KEYS];
+static struct lock_class *lock_classes;
 
 #ifdef CONFIG_LOCK_STAT
 static DEFINE_PER_CPU(struct lock_class_stats[MAX_LOCKDEP_KEYS], lock_stats);

@@ -257,7 +257,7 @@ static struct list_head classhash_table[
 #define __chainhashfn(chain)	hash_long(chain, CHAINHASH_BITS)
 #define chainhashentry(chain)	(chainhash_table + __chainhashfn((chain)))
 
-static struct list_head chainhash_table[CHAINHASH_SIZE];
+static struct list_head *chainhash_table;
 
 /*
  * The hash key of the lock dependency chains is a hash itself too:

@@ -332,7 +332,7 @@ static int verbose(struct lock_class *cl
  * addresses. Protected by the graph_lock.
  */
 unsigned long nr_stack_trace_entries;
-static unsigned long stack_trace[MAX_STACK_TRACE_ENTRIES];
+static unsigned long *stack_trace;
 
 static int save_trace(struct stack_trace *trace)
 {

@@ -1454,7 +1454,7 @@ out_bug:
 }
 
 unsigned long nr_lock_chains;
-static struct lock_chain lock_chains[MAX_LOCKDEP_CHAINS];
+static struct lock_chain *lock_chains;
 
 /*
  * Look up a dependency chain. If the key is not present yet then

@@ -3010,6 +3010,27 @@ out_restore:
 	raw_local_irq_restore(flags);
 }
 
+/*
+ * The large arrays can bloat the kernel image too much causing problems
+ * because it needs an continuous area in memory. Allocate them
+ * using a special allocator if possible. This is before bootmem and only
+ * works on some architectures.
+ */
+
+#ifndef ARCH_HAS_EARLY_ALLOC
+#define LARGEVAR(x,y) { static typeof(*x) __ ## x[y];  x = __ ## x; }
+#else
+#define LARGEVAR(x,y) x = arch_early_alloc(sizeof(*x) * y)
+#endif
+
+void lockdep_init_mem(void)
+{
+	LARGEVAR(stack_trace, MAX_STACK_TRACE_ENTRIES);
+	LARGEVAR(list_entries, MAX_LOCKDEP_ENTRIES);
+	LARGEVAR(lock_chains, MAX_LOCKDEP_CHAINS);
+	LARGEVAR(lock_classes, MAX_LOCKDEP_KEYS);
+	LARGEVAR(chainhash_table, CHAINHASH_SIZE);
+}
 void lockdep_init(void)
 {
 	int i;

@@ -3023,6 +3044,8 @@ void lockdep_init(void)
 	if (lockdep_initialized)
 		return;
 
+	lockdep_init_mem();
+
 	for (i = 0; i < CLASSHASH_SIZE; i++)
 		INIT_LIST_HEAD(classhash_table + i);