module: replace module_layout with module_memory

module_layout manages different types of memory (text, data, rodata, etc.)
in one allocation, which is problematic for some reasons:

1. It is hard to enable CONFIG_STRICT_MODULE_RWX.
2. It is hard to use huge pages in modules (and not break strict rwx).
3. Many archs uses module_layout for arch-specific data, but it is not
   obvious how these data are used (are they RO, RX, or RW?)

Improve the scenario by replacing 2 (or 3) module_layout per module with
up to 7 module_memory per module:

        MOD_TEXT,
        MOD_DATA,
        MOD_RODATA,
        MOD_RO_AFTER_INIT,
        MOD_INIT_TEXT,
        MOD_INIT_DATA,
        MOD_INIT_RODATA,

and allocating them separately. This adds slightly more entries to
mod_tree (from up to 3 entries per module, to up to 7 entries per
module). However, this at most adds a small constant overhead to
__module_address(), which is expected to be fast.

Various archs use module_layout for different data. These data are put
into different module_memory based on their location in module_layout.
IOW, data that used to go with text is allocated with MOD_MEM_TYPE_TEXT;
data that used to go with data is allocated with MOD_MEM_TYPE_DATA, etc.

module_memory simplifies quite some of the module code. For example,
ARCH_WANTS_MODULES_DATA_IN_VMALLOC is a lot cleaner, as it just uses a
different allocator for the data. kernel/module/strict_rwx.c is also
much cleaner with module_memory.

Signed-off-by: Song Liu <song@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

arch/arc/kernel/unwind.c

@@ -369,6 +369,8 @@ void *unwind_add_table(struct module *module, const void *table_start,
 		       unsigned long table_size)
 {
 	struct unwind_table *table;
+	struct module_memory *core_text;
+	struct module_memory *init_text;
 
 	if (table_size <= 0)
 		return NULL;
@@ -377,11 +379,11 @@ void *unwind_add_table(struct module *module, const void *table_start,
 	if (!table)
 		return NULL;
 
-	init_unwind_table(table, module->name,
-			  module->core_layout.base, module->core_layout.size,
-			  module->init_layout.base, module->init_layout.size,
-			  table_start, table_size,
-			  NULL, 0);
+	core_text = &module->mem[MOD_TEXT];
+	init_text = &module->mem[MOD_INIT_TEXT];
+
+	init_unwind_table(table, module->name, core_text->base, core_text->size,
+			  init_text->base, init_text->size, table_start, table_size, NULL, 0);
 
 	init_unwind_hdr(table, unw_hdr_alloc);
 

arch/arm/kernel/module-plts.c

@@ -28,11 +28,6 @@ static const u32 fixed_plts[] = {
 #endif
 };
 
-static bool in_init(const struct module *mod, unsigned long loc)
-{
-	return loc - (u32)mod->init_layout.base < mod->init_layout.size;
-}
-
 static void prealloc_fixed(struct mod_plt_sec *pltsec, struct plt_entries *plt)
 {
 	int i;
@@ -50,8 +45,8 @@ static void prealloc_fixed(struct mod_plt_sec *pltsec, struct plt_entries *plt)
 
 u32 get_module_plt(struct module *mod, unsigned long loc, Elf32_Addr val)
 {
-	struct mod_plt_sec *pltsec = !in_init(mod, loc) ? &mod->arch.core :
-							  &mod->arch.init;
+	struct mod_plt_sec *pltsec = !within_module_init(loc, mod) ?
+						&mod->arch.core : &mod->arch.init;
 	struct plt_entries *plt;
 	int idx;
 

arch/arm64/kernel/module-plts.c

@@ -65,17 +65,12 @@ static bool plt_entries_equal(const struct plt_entry *a,
 	       (q + aarch64_insn_adrp_get_offset(le32_to_cpu(b->adrp)));
 }
 
-static bool in_init(const struct module *mod, void *loc)
-{
-	return (u64)loc - (u64)mod->init_layout.base < mod->init_layout.size;
-}
-
 u64 module_emit_plt_entry(struct module *mod, Elf64_Shdr *sechdrs,
 			  void *loc, const Elf64_Rela *rela,
 			  Elf64_Sym *sym)
 {
-	struct mod_plt_sec *pltsec = !in_init(mod, loc) ? &mod->arch.core :
-							  &mod->arch.init;
+	struct mod_plt_sec *pltsec = !within_module_init((unsigned long)loc, mod) ?
+						&mod->arch.core : &mod->arch.init;
 	struct plt_entry *plt = (struct plt_entry *)sechdrs[pltsec->plt_shndx].sh_addr;
 	int i = pltsec->plt_num_entries;
 	int j = i - 1;
@@ -105,8 +100,8 @@ u64 module_emit_plt_entry(struct module *mod, Elf64_Shdr *sechdrs,
 u64 module_emit_veneer_for_adrp(struct module *mod, Elf64_Shdr *sechdrs,
 				void *loc, u64 val)
 {
-	struct mod_plt_sec *pltsec = !in_init(mod, loc) ? &mod->arch.core :
-							  &mod->arch.init;
+	struct mod_plt_sec *pltsec = !within_module_init((unsigned long)loc, mod) ?
+						&mod->arch.core : &mod->arch.init;
 	struct plt_entry *plt = (struct plt_entry *)sechdrs[pltsec->plt_shndx].sh_addr;
 	int i = pltsec->plt_num_entries++;
 	u32 br;

arch/ia64/kernel/module.c

@@ -485,19 +485,19 @@ module_frob_arch_sections (Elf_Ehdr *ehdr, Elf_Shdr *sechdrs, char *secstrings,
 	return 0;
 }
 
-static inline int
+static inline bool
 in_init (const struct module *mod, uint64_t addr)
 {
-	return addr - (uint64_t) mod->init_layout.base < mod->init_layout.size;
+	return within_module_init(addr, mod);
 }
 
-static inline int
+static inline bool
 in_core (const struct module *mod, uint64_t addr)
 {
-	return addr - (uint64_t) mod->core_layout.base < mod->core_layout.size;
+	return within_module_core(addr, mod);
 }
 
-static inline int
+static inline bool
 is_internal (const struct module *mod, uint64_t value)
 {
 	return in_init(mod, value) || in_core(mod, value);
@@ -677,7 +677,8 @@ do_reloc (struct module *mod, uint8_t r_type, Elf64_Sym *sym, uint64_t addend,
 		break;
 
 	      case RV_BDREL:
-		val -= (uint64_t) (in_init(mod, val) ? mod->init_layout.base : mod->core_layout.base);
+		val -= (uint64_t) (in_init(mod, val) ? mod->mem[MOD_INIT_TEXT].base :
+				   mod->mem[MOD_TEXT].base);
 		break;
 
 	      case RV_LTV:
@@ -812,15 +813,18 @@ apply_relocate_add (Elf64_Shdr *sechdrs, const char *strtab, unsigned int symind
 		 *     addresses have been selected...
 		 */
 		uint64_t gp;
-		if (mod->core_layout.size > MAX_LTOFF)
+		struct module_memory *mod_mem;
+
+		mod_mem = &mod->mem[MOD_DATA];
+		if (mod_mem->size > MAX_LTOFF)
 			/*
 			 * This takes advantage of fact that SHF_ARCH_SMALL gets allocated
 			 * at the end of the module.
 			 */
-			gp = mod->core_layout.size - MAX_LTOFF / 2;
+			gp = mod_mem->size - MAX_LTOFF / 2;
 		else
-			gp = mod->core_layout.size / 2;
-		gp = (uint64_t) mod->core_layout.base + ((gp + 7) & -8);
+			gp = mod_mem->size / 2;
+		gp = (uint64_t) mod_mem->base + ((gp + 7) & -8);
 		mod->arch.gp = gp;
 		DEBUGP("%s: placing gp at 0x%lx\n", __func__, gp);
 	}

arch/mips/kernel/vpe.c

@@ -199,18 +199,17 @@ static void layout_sections(struct module *mod, const Elf_Ehdr *hdr,
 	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
 		for (i = 0; i < hdr->e_shnum; ++i) {
 			Elf_Shdr *s = &sechdrs[i];
+			struct module_memory *mod_mem;
+
+			mod_mem = &mod->mem[MOD_TEXT];
 
 			if ((s->sh_flags & masks[m][0]) != masks[m][0]
 			    || (s->sh_flags & masks[m][1])
 			    || s->sh_entsize != ~0UL)
 				continue;
 			s->sh_entsize =
-				get_offset((unsigned long *)&mod->core_layout.size, s);
+				get_offset((unsigned long *)&mod_mem->size, s);
 		}
-
-		if (m == 0)
-			mod->core_layout.text_size = mod->core_layout.size;
-
 	}
 }
 
@@ -641,7 +640,7 @@ static int vpe_elfload(struct vpe *v)
 		layout_sections(&mod, hdr, sechdrs, secstrings);
 	}
 
-	v->load_addr = alloc_progmem(mod.core_layout.size);
+	v->load_addr = alloc_progmem(mod.mod_mem[MOD_TEXT].size);
 	if (!v->load_addr)
 		return -ENOMEM;
 

arch/parisc/kernel/module.c

@@ -27,9 +27,9 @@
  *      We are not doing SEGREL32 handling correctly. According to the ABI, we
  *      should do a value offset, like this:
  *			if (in_init(me, (void *)val))
- *				val -= (uint32_t)me->init_layout.base;
+ *				val -= (uint32_t)me->mem[MOD_INIT_TEXT].base;
  *			else
- *				val -= (uint32_t)me->core_layout.base;
+ *				val -= (uint32_t)me->mem[MOD_TEXT].base;
  *	However, SEGREL32 is used only for PARISC unwind entries, and we want
  *	those entries to have an absolute address, and not just an offset.
  *
@@ -76,25 +76,6 @@
  * allows us to allocate up to 4095 GOT entries. */
 #define MAX_GOTS	4095
 
-/* three functions to determine where in the module core
- * or init pieces the location is */
-static inline int in_init(struct module *me, void *loc)
-{
-	return (loc >= me->init_layout.base &&
-		loc <= (me->init_layout.base + me->init_layout.size));
-}
-
-static inline int in_core(struct module *me, void *loc)
-{
-	return (loc >= me->core_layout.base &&
-		loc <= (me->core_layout.base + me->core_layout.size));
-}
-
-static inline int in_local(struct module *me, void *loc)
-{
-	return in_init(me, loc) || in_core(me, loc);
-}
-
 #ifndef CONFIG_64BIT
 struct got_entry {
 	Elf32_Addr addr;
@@ -302,6 +283,7 @@ int module_frob_arch_sections(CONST Elf_Ehdr *hdr,
 {
 	unsigned long gots = 0, fdescs = 0, len;
 	unsigned int i;
+	struct module_memory *mod_mem;
 
 	len = hdr->e_shnum * sizeof(me->arch.section[0]);
 	me->arch.section = kzalloc(len, GFP_KERNEL);
@@ -346,14 +328,15 @@ int module_frob_arch_sections(CONST Elf_Ehdr *hdr,
 		me->arch.section[s].stub_entries += count;
 	}
 
+	mod_mem = &me->mem[MOD_TEXT];
 	/* align things a bit */
-	me->core_layout.size = ALIGN(me->core_layout.size, 16);
-	me->arch.got_offset = me->core_layout.size;
-	me->core_layout.size += gots * sizeof(struct got_entry);
+	mod_mem->size = ALIGN(mod_mem->size, 16);
+	me->arch.got_offset = mod_mem->size;
+	mod_mem->size += gots * sizeof(struct got_entry);
 
-	me->core_layout.size = ALIGN(me->core_layout.size, 16);
-	me->arch.fdesc_offset = me->core_layout.size;
-	me->core_layout.size += fdescs * sizeof(Elf_Fdesc);
+	mod_mem->size = ALIGN(mod_mem->size, 16);
+	me->arch.fdesc_offset = mod_mem->size;
+	mod_mem->size += fdescs * sizeof(Elf_Fdesc);
 
 	me->arch.got_max = gots;
 	me->arch.fdesc_max = fdescs;
@@ -371,7 +354,7 @@ static Elf64_Word get_got(struct module *me, unsigned long value, long addend)
 
 	BUG_ON(value == 0);
 
-	got = me->core_layout.base + me->arch.got_offset;
+	got = me->mem[MOD_TEXT].base + me->arch.got_offset;
 	for (i = 0; got[i].addr; i++)
 		if (got[i].addr == value)
 			goto out;
@@ -389,7 +372,7 @@ static Elf64_Word get_got(struct module *me, unsigned long value, long addend)
 #ifdef CONFIG_64BIT
 static Elf_Addr get_fdesc(struct module *me, unsigned long value)
 {
-	Elf_Fdesc *fdesc = me->core_layout.base + me->arch.fdesc_offset;
+	Elf_Fdesc *fdesc = me->mem[MOD_TEXT].base + me->arch.fdesc_offset;
 
 	if (!value) {
 		printk(KERN_ERR "%s: zero OPD requested!\n", me->name);
@@ -407,7 +390,7 @@ static Elf_Addr get_fdesc(struct module *me, unsigned long value)
 
 	/* Create new one */
 	fdesc->addr = value;
-	fdesc->gp = (Elf_Addr)me->core_layout.base + me->arch.got_offset;
+	fdesc->gp = (Elf_Addr)me->mem[MOD_TEXT].base + me->arch.got_offset;
 	return (Elf_Addr)fdesc;
 }
 #endif /* CONFIG_64BIT */
@@ -742,7 +725,7 @@ int apply_relocate_add(Elf_Shdr *sechdrs,
 			       loc, val);
 			val += addend;
 			/* can we reach it locally? */
-			if (in_local(me, (void *)val)) {
+			if (within_module(val, me)) {
 				/* this is the case where the symbol is local
 				 * to the module, but in a different section,
 				 * so stub the jump in case it's more than 22
@@ -801,7 +784,7 @@ int apply_relocate_add(Elf_Shdr *sechdrs,
 			break;
 		case R_PARISC_FPTR64:
 			/* 64-bit function address */
-			if(in_local(me, (void *)(val + addend))) {
+			if (within_module(val + addend, me)) {
 				*loc64 = get_fdesc(me, val+addend);
 				pr_debug("FDESC for %s at %llx points to %llx\n",
 				       strtab + sym->st_name, *loc64,
@@ -839,7 +822,7 @@ register_unwind_table(struct module *me,
 
 	table = (unsigned char *)sechdrs[me->arch.unwind_section].sh_addr;
 	end = table + sechdrs[me->arch.unwind_section].sh_size;
-	gp = (Elf_Addr)me->core_layout.base + me->arch.got_offset;
+	gp = (Elf_Addr)me->mem[MOD_TEXT].base + me->arch.got_offset;
 
 	pr_debug("register_unwind_table(), sect = %d at 0x%p - 0x%p (gp=0x%lx)\n",
 	       me->arch.unwind_section, table, end, gp);
@@ -977,7 +960,7 @@ void module_arch_cleanup(struct module *mod)
 #ifdef CONFIG_64BIT
 void *dereference_module_function_descriptor(struct module *mod, void *ptr)
 {
-	unsigned long start_opd = (Elf64_Addr)mod->core_layout.base +
+	unsigned long start_opd = (Elf64_Addr)mod->mem[MOD_TEXT].base +
 				   mod->arch.fdesc_offset;
 	unsigned long end_opd = start_opd +
 				mod->arch.fdesc_count * sizeof(Elf64_Fdesc);

arch/powerpc/kernel/module_32.c

@@ -163,8 +163,7 @@ static uint32_t do_plt_call(void *location,
 
 	pr_debug("Doing plt for call to 0x%x at 0x%x\n", val, (unsigned int)location);
 	/* Init, or core PLT? */
-	if (location >= mod->core_layout.base
-	    && location < mod->core_layout.base + mod->core_layout.size)
+	if (within_module_core((unsigned long)location, mod))
 		entry = (void *)sechdrs[mod->arch.core_plt_section].sh_addr;
 	else
 		entry = (void *)sechdrs[mod->arch.init_plt_section].sh_addr;
@@ -322,14 +321,14 @@ notrace int module_trampoline_target(struct module *mod, unsigned long addr,
 
 int module_finalize_ftrace(struct module *module, const Elf_Shdr *sechdrs)
 {
-	module->arch.tramp = do_plt_call(module->core_layout.base,
+	module->arch.tramp = do_plt_call(module->mem[MOD_TEXT].base,
 					 (unsigned long)ftrace_caller,
 					 sechdrs, module);
 	if (!module->arch.tramp)
 		return -ENOENT;
 
 #ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS
-	module->arch.tramp_regs = do_plt_call(module->core_layout.base,
+	module->arch.tramp_regs = do_plt_call(module->mem[MOD_TEXT].base,
 					      (unsigned long)ftrace_regs_caller,
 					      sechdrs, module);
 	if (!module->arch.tramp_regs)