From ac3b43283923440900b4f36ca5f9f0b1ca43b70e Mon Sep 17 00:00:00 2001 From: Song Liu Date: Mon, 6 Feb 2023 16:28:02 -0800 Subject: 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 Cc: Luis Chamberlain Cc: Thomas Gleixner Cc: Peter Zijlstra Cc: Guenter Roeck Cc: Christophe Leroy Reviewed-by: Thomas Gleixner Reviewed-by: Christophe Leroy Reviewed-by: Luis Chamberlain Signed-off-by: Luis Chamberlain --- kernel/module/internal.h | 40 ++--- kernel/module/kallsyms.c | 58 ++++--- kernel/module/kdb.c | 17 +- kernel/module/main.c | 375 ++++++++++++++++++++++---------------------- kernel/module/procfs.c | 16 +- kernel/module/strict_rwx.c | 99 +++--------- kernel/module/tree_lookup.c | 39 ++--- 7 files changed, 294 insertions(+), 350 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/internal.h b/kernel/module/internal.h index 2e2bf236f558..c6c44ceca07a 100644 --- a/kernel/module/internal.h +++ b/kernel/module/internal.h @@ -17,27 +17,19 @@ #define ARCH_SHF_SMALL 0 #endif -/* If this is set, the section belongs in the init part of the module */ -#define INIT_OFFSET_MASK (1UL << (BITS_PER_LONG - 1)) -/* Maximum number of characters written by module_flags() */ -#define MODULE_FLAGS_BUF_SIZE (TAINT_FLAGS_COUNT + 4) - -#ifndef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC -#define data_layout core_layout -#endif - /* - * Modules' sections will be aligned on page boundaries - * to ensure complete separation of code and data, but - * only when CONFIG_STRICT_MODULE_RWX=y + * Use highest 4 bits of sh_entsize to store the mod_mem_type of this + * section. This leaves 28 bits for offset on 32-bit systems, which is + * about 256 MiB (WARN_ON_ONCE if we exceed that). */ -static inline unsigned int strict_align(unsigned int size) -{ - if (IS_ENABLED(CONFIG_STRICT_MODULE_RWX)) - return PAGE_ALIGN(size); - else - return size; -} + +#define SH_ENTSIZE_TYPE_BITS 4 +#define SH_ENTSIZE_TYPE_SHIFT (BITS_PER_LONG - SH_ENTSIZE_TYPE_BITS) +#define SH_ENTSIZE_TYPE_MASK ((1UL << SH_ENTSIZE_TYPE_BITS) - 1) +#define SH_ENTSIZE_OFFSET_MASK ((1UL << (BITS_PER_LONG - SH_ENTSIZE_TYPE_BITS)) - 1) + +/* Maximum number of characters written by module_flags() */ +#define MODULE_FLAGS_BUF_SIZE (TAINT_FLAGS_COUNT + 4) extern struct mutex module_mutex; extern struct list_head modules; @@ -101,8 +93,8 @@ int try_to_force_load(struct module *mod, const char *reason); bool find_symbol(struct find_symbol_arg *fsa); struct module *find_module_all(const char *name, size_t len, bool even_unformed); int cmp_name(const void *name, const void *sym); -long module_get_offset(struct module *mod, unsigned int *size, Elf_Shdr *sechdr, - unsigned int section); +long module_get_offset_and_type(struct module *mod, enum mod_mem_type type, + Elf_Shdr *sechdr, unsigned int section); char *module_flags(struct module *mod, char *buf, bool show_state); size_t module_flags_taint(unsigned long taints, char *buf); @@ -190,10 +182,13 @@ struct mod_tree_root { #endif unsigned long addr_min; unsigned long addr_max; +#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC + unsigned long data_addr_min; + unsigned long data_addr_max; +#endif }; extern struct mod_tree_root mod_tree; -extern struct mod_tree_root mod_data_tree; #ifdef CONFIG_MODULES_TREE_LOOKUP void mod_tree_insert(struct module *mod); @@ -224,7 +219,6 @@ void module_enable_nx(const struct module *mod); void module_enable_x(const struct module *mod); int module_enforce_rwx_sections(Elf_Ehdr *hdr, Elf_Shdr *sechdrs, char *secstrings, struct module *mod); -bool module_check_misalignment(const struct module *mod); #ifdef CONFIG_MODULE_SIG int module_sig_check(struct load_info *info, int flags); diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c index ab2376a1be88..8e89f459ffdd 100644 --- a/kernel/module/kallsyms.c +++ b/kernel/module/kallsyms.c @@ -78,6 +78,7 @@ static bool is_core_symbol(const Elf_Sym *src, const Elf_Shdr *sechdrs, unsigned int shnum, unsigned int pcpundx) { const Elf_Shdr *sec; + enum mod_mem_type type; if (src->st_shndx == SHN_UNDEF || src->st_shndx >= shnum || @@ -90,11 +91,12 @@ static bool is_core_symbol(const Elf_Sym *src, const Elf_Shdr *sechdrs, #endif sec = sechdrs + src->st_shndx; + type = sec->sh_entsize >> SH_ENTSIZE_TYPE_SHIFT; if (!(sec->sh_flags & SHF_ALLOC) #ifndef CONFIG_KALLSYMS_ALL || !(sec->sh_flags & SHF_EXECINSTR) #endif - || (sec->sh_entsize & INIT_OFFSET_MASK)) + || mod_mem_type_is_init(type)) return false; return true; @@ -113,11 +115,13 @@ void layout_symtab(struct module *mod, struct load_info *info) Elf_Shdr *strsect = info->sechdrs + info->index.str; const Elf_Sym *src; unsigned int i, nsrc, ndst, strtab_size = 0; + struct module_memory *mod_mem_data = &mod->mem[MOD_DATA]; + struct module_memory *mod_mem_init_data = &mod->mem[MOD_INIT_DATA]; /* Put symbol section at end of init part of module. */ symsect->sh_flags |= SHF_ALLOC; - symsect->sh_entsize = module_get_offset(mod, &mod->init_layout.size, symsect, - info->index.sym) | INIT_OFFSET_MASK; + symsect->sh_entsize = module_get_offset_and_type(mod, MOD_INIT_DATA, + symsect, info->index.sym); pr_debug("\t%s\n", info->secstrings + symsect->sh_name); src = (void *)info->hdr + symsect->sh_offset; @@ -134,28 +138,27 @@ void layout_symtab(struct module *mod, struct load_info *info) } /* Append room for core symbols at end of core part. */ - info->symoffs = ALIGN(mod->data_layout.size, symsect->sh_addralign ?: 1); - info->stroffs = mod->data_layout.size = info->symoffs + ndst * sizeof(Elf_Sym); - mod->data_layout.size += strtab_size; + info->symoffs = ALIGN(mod_mem_data->size, symsect->sh_addralign ?: 1); + info->stroffs = mod_mem_data->size = info->symoffs + ndst * sizeof(Elf_Sym); + mod_mem_data->size += strtab_size; /* Note add_kallsyms() computes strtab_size as core_typeoffs - stroffs */ - info->core_typeoffs = mod->data_layout.size; - mod->data_layout.size += ndst * sizeof(char); - mod->data_layout.size = strict_align(mod->data_layout.size); + info->core_typeoffs = mod_mem_data->size; + mod_mem_data->size += ndst * sizeof(char); /* Put string table section at end of init part of module. */ strsect->sh_flags |= SHF_ALLOC; - strsect->sh_entsize = module_get_offset(mod, &mod->init_layout.size, strsect, - info->index.str) | INIT_OFFSET_MASK; + strsect->sh_entsize = module_get_offset_and_type(mod, MOD_INIT_DATA, + strsect, info->index.str); pr_debug("\t%s\n", info->secstrings + strsect->sh_name); /* We'll tack temporary mod_kallsyms on the end. */ - mod->init_layout.size = ALIGN(mod->init_layout.size, - __alignof__(struct mod_kallsyms)); - info->mod_kallsyms_init_off = mod->init_layout.size; - mod->init_layout.size += sizeof(struct mod_kallsyms); - info->init_typeoffs = mod->init_layout.size; - mod->init_layout.size += nsrc * sizeof(char); - mod->init_layout.size = strict_align(mod->init_layout.size); + mod_mem_init_data->size = ALIGN(mod_mem_init_data->size, + __alignof__(struct mod_kallsyms)); + info->mod_kallsyms_init_off = mod_mem_init_data->size; + + mod_mem_init_data->size += sizeof(struct mod_kallsyms); + info->init_typeoffs = mod_mem_init_data->size; + mod_mem_init_data->size += nsrc * sizeof(char); } /* @@ -171,9 +174,11 @@ void add_kallsyms(struct module *mod, const struct load_info *info) char *s; Elf_Shdr *symsec = &info->sechdrs[info->index.sym]; unsigned long strtab_size; + void *data_base = mod->mem[MOD_DATA].base; + void *init_data_base = mod->mem[MOD_INIT_DATA].base; /* Set up to point into init section. */ - mod->kallsyms = (void __rcu *)mod->init_layout.base + + mod->kallsyms = (void __rcu *)init_data_base + info->mod_kallsyms_init_off; rcu_read_lock(); @@ -183,15 +188,15 @@ void add_kallsyms(struct module *mod, const struct load_info *info) /* Make sure we get permanent strtab: don't use info->strtab. */ rcu_dereference(mod->kallsyms)->strtab = (void *)info->sechdrs[info->index.str].sh_addr; - rcu_dereference(mod->kallsyms)->typetab = mod->init_layout.base + info->init_typeoffs; + rcu_dereference(mod->kallsyms)->typetab = init_data_base + info->init_typeoffs; /* * Now populate the cut down core kallsyms for after init * and set types up while we still have access to sections. */ - mod->core_kallsyms.symtab = dst = mod->data_layout.base + info->symoffs; - mod->core_kallsyms.strtab = s = mod->data_layout.base + info->stroffs; - mod->core_kallsyms.typetab = mod->data_layout.base + info->core_typeoffs; + mod->core_kallsyms.symtab = dst = data_base + info->symoffs; + mod->core_kallsyms.strtab = s = data_base + info->stroffs; + mod->core_kallsyms.typetab = data_base + info->core_typeoffs; strtab_size = info->core_typeoffs - info->stroffs; src = rcu_dereference(mod->kallsyms)->symtab; for (ndst = i = 0; i < rcu_dereference(mod->kallsyms)->num_symtab; i++) { @@ -267,12 +272,15 @@ static const char *find_kallsyms_symbol(struct module *mod, unsigned int i, best = 0; unsigned long nextval, bestval; struct mod_kallsyms *kallsyms = rcu_dereference_sched(mod->kallsyms); + struct module_memory *mod_mem; /* At worse, next value is at end of module */ if (within_module_init(addr, mod)) - nextval = (unsigned long)mod->init_layout.base + mod->init_layout.text_size; + mod_mem = &mod->mem[MOD_INIT_TEXT]; else - nextval = (unsigned long)mod->core_layout.base + mod->core_layout.text_size; + mod_mem = &mod->mem[MOD_TEXT]; + + nextval = (unsigned long)mod_mem->base + mod_mem->size; bestval = kallsyms_symbol_value(&kallsyms->symtab[best]); diff --git a/kernel/module/kdb.c b/kernel/module/kdb.c index f4317f92e189..995c32d3698f 100644 --- a/kernel/module/kdb.c +++ b/kernel/module/kdb.c @@ -26,10 +26,11 @@ int kdb_lsmod(int argc, const char **argv) if (mod->state == MODULE_STATE_UNFORMED) continue; - kdb_printf("%-20s%8u", mod->name, mod->core_layout.size); -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - kdb_printf("/%8u", mod->data_layout.size); -#endif + kdb_printf("%-20s%8u", mod->name, mod->mem[MOD_TEXT].size); + kdb_printf("/%8u", mod->mem[MOD_RODATA].size); + kdb_printf("/%8u", mod->mem[MOD_RO_AFTER_INIT].size); + kdb_printf("/%8u", mod->mem[MOD_DATA].size); + kdb_printf(" 0x%px ", (void *)mod); #ifdef CONFIG_MODULE_UNLOAD kdb_printf("%4d ", module_refcount(mod)); @@ -40,10 +41,10 @@ int kdb_lsmod(int argc, const char **argv) kdb_printf(" (Loading)"); else kdb_printf(" (Live)"); - kdb_printf(" 0x%px", mod->core_layout.base); -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - kdb_printf("/0x%px", mod->data_layout.base); -#endif + kdb_printf(" 0x%px", mod->mem[MOD_TEXT].base); + kdb_printf("/0x%px", mod->mem[MOD_RODATA].base); + kdb_printf("/0x%px", mod->mem[MOD_RO_AFTER_INIT].base); + kdb_printf("/0x%px", mod->mem[MOD_DATA].base); #ifdef CONFIG_MODULE_UNLOAD { diff --git a/kernel/module/main.c b/kernel/module/main.c index d3be89de706d..2724bc1b9a90 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -80,12 +80,6 @@ struct mod_tree_root mod_tree __cacheline_aligned = { .addr_min = -1UL, }; -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC -struct mod_tree_root mod_data_tree __cacheline_aligned = { - .addr_min = -1UL, -}; -#endif - struct symsearch { const struct kernel_symbol *start, *stop; const s32 *crcs; @@ -93,14 +87,24 @@ struct symsearch { }; /* - * Bounds of module text, for speeding up __module_address. + * Bounds of module memory, for speeding up __module_address. * Protected by module_mutex. */ -static void __mod_update_bounds(void *base, unsigned int size, struct mod_tree_root *tree) +static void __mod_update_bounds(enum mod_mem_type type __maybe_unused, void *base, + unsigned int size, struct mod_tree_root *tree) { unsigned long min = (unsigned long)base; unsigned long max = min + size; +#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC + if (mod_mem_type_is_core_data(type)) { + if (min < tree->data_addr_min) + tree->data_addr_min = min; + if (max > tree->data_addr_max) + tree->data_addr_max = max; + return; + } +#endif if (min < tree->addr_min) tree->addr_min = min; if (max > tree->addr_max) @@ -109,12 +113,12 @@ static void __mod_update_bounds(void *base, unsigned int size, struct mod_tree_r static void mod_update_bounds(struct module *mod) { - __mod_update_bounds(mod->core_layout.base, mod->core_layout.size, &mod_tree); - if (mod->init_layout.size) - __mod_update_bounds(mod->init_layout.base, mod->init_layout.size, &mod_tree); -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - __mod_update_bounds(mod->data_layout.base, mod->data_layout.size, &mod_data_tree); -#endif + for_each_mod_mem_type(type) { + struct module_memory *mod_mem = &mod->mem[type]; + + if (mod_mem->size) + __mod_update_bounds(type, mod_mem->base, mod_mem->size, &mod_tree); + } } /* Block module loading/unloading? */ @@ -926,7 +930,13 @@ struct module_attribute module_uevent = static ssize_t show_coresize(struct module_attribute *mattr, struct module_kobject *mk, char *buffer) { - return sprintf(buffer, "%u\n", mk->mod->core_layout.size); + unsigned int size = mk->mod->mem[MOD_TEXT].size; + + if (!IS_ENABLED(CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC)) { + for_class_mod_mem_type(type, core_data) + size += mk->mod->mem[type].size; + } + return sprintf(buffer, "%u\n", size); } static struct module_attribute modinfo_coresize = @@ -936,7 +946,11 @@ static struct module_attribute modinfo_coresize = static ssize_t show_datasize(struct module_attribute *mattr, struct module_kobject *mk, char *buffer) { - return sprintf(buffer, "%u\n", mk->mod->data_layout.size); + unsigned int size = 0; + + for_class_mod_mem_type(type, core_data) + size += mk->mod->mem[type].size; + return sprintf(buffer, "%u\n", size); } static struct module_attribute modinfo_datasize = @@ -946,7 +960,11 @@ static struct module_attribute modinfo_datasize = static ssize_t show_initsize(struct module_attribute *mattr, struct module_kobject *mk, char *buffer) { - return sprintf(buffer, "%u\n", mk->mod->init_layout.size); + unsigned int size = 0; + + for_class_mod_mem_type(type, init) + size += mk->mod->mem[type].size; + return sprintf(buffer, "%u\n", size); } static struct module_attribute modinfo_initsize = @@ -1143,6 +1161,46 @@ void __weak module_arch_freeing_init(struct module *mod) { } +static bool mod_mem_use_vmalloc(enum mod_mem_type type) +{ + return IS_ENABLED(CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC) && + mod_mem_type_is_core_data(type); +} + +static void *module_memory_alloc(unsigned int size, enum mod_mem_type type) +{ + if (mod_mem_use_vmalloc(type)) + return vzalloc(size); + return module_alloc(size); +} + +static void module_memory_free(void *ptr, enum mod_mem_type type) +{ + if (mod_mem_use_vmalloc(type)) + vfree(ptr); + else + module_memfree(ptr); +} + +static void free_mod_mem(struct module *mod) +{ + for_each_mod_mem_type(type) { + struct module_memory *mod_mem = &mod->mem[type]; + + if (type == MOD_DATA) + continue; + + /* Free lock-classes; relies on the preceding sync_rcu(). */ + lockdep_free_key_range(mod_mem->base, mod_mem->size); + if (mod_mem->size) + module_memory_free(mod_mem->base, type); + } + + /* MOD_DATA hosts mod, so free it at last */ + lockdep_free_key_range(mod->mem[MOD_DATA].base, mod->mem[MOD_DATA].size); + module_memory_free(mod->mem[MOD_DATA].base, MOD_DATA); +} + /* Free a module, remove from lists, etc. */ static void free_module(struct module *mod) { @@ -1189,18 +1247,10 @@ static void free_module(struct module *mod) /* This may be empty, but that's OK */ module_arch_freeing_init(mod); - module_memfree(mod->init_layout.base); kfree(mod->args); percpu_modfree(mod); - /* Free lock-classes; relies on the preceding sync_rcu(). */ - lockdep_free_key_range(mod->data_layout.base, mod->data_layout.size); - - /* Finally, free the core (containing the module structure) */ - module_memfree(mod->core_layout.base); -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - vfree(mod->data_layout.base); -#endif + free_mod_mem(mod); } void *__symbol_get(const char *symbol) @@ -1387,16 +1437,18 @@ unsigned int __weak arch_mod_section_prepend(struct module *mod, return 0; } -/* Update size with this section: return offset. */ -long module_get_offset(struct module *mod, unsigned int *size, - Elf_Shdr *sechdr, unsigned int section) +long module_get_offset_and_type(struct module *mod, enum mod_mem_type type, + Elf_Shdr *sechdr, unsigned int section) { - long ret; + long offset; + long mask = ((unsigned long)(type) & SH_ENTSIZE_TYPE_MASK) << SH_ENTSIZE_TYPE_SHIFT; - *size += arch_mod_section_prepend(mod, section); - ret = ALIGN(*size, sechdr->sh_addralign ?: 1); - *size = ret + sechdr->sh_size; - return ret; + mod->mem[type].size += arch_mod_section_prepend(mod, section); + offset = ALIGN(mod->mem[type].size, sechdr->sh_addralign ?: 1); + mod->mem[type].size = offset + sechdr->sh_size; + + WARN_ON_ONCE(offset & mask); + return offset | mask; } static bool module_init_layout_section(const char *sname) @@ -1408,15 +1460,11 @@ static bool module_init_layout_section(const char *sname) return module_init_section(sname); } -/* - * Lay out the SHF_ALLOC sections in a way not dissimilar to how ld - * might -- code, read-only data, read-write data, small data. Tally - * sizes, and place the offsets into sh_entsize fields: high bit means it - * belongs in init. - */ -static void layout_sections(struct module *mod, struct load_info *info) +static void __layout_sections(struct module *mod, struct load_info *info, bool is_init) { - static unsigned long const masks[][2] = { + unsigned int m, i; + + static const unsigned long masks[][2] = { /* * NOTE: all executable code must be the first section * in this array; otherwise modify the text_size @@ -1428,84 +1476,63 @@ static void layout_sections(struct module *mod, struct load_info *info) { SHF_WRITE | SHF_ALLOC, ARCH_SHF_SMALL }, { ARCH_SHF_SMALL | SHF_ALLOC, 0 } }; - unsigned int m, i; - - for (i = 0; i < info->hdr->e_shnum; i++) - info->sechdrs[i].sh_entsize = ~0UL; + static const int core_m_to_mem_type[] = { + MOD_TEXT, + MOD_RODATA, + MOD_RO_AFTER_INIT, + MOD_DATA, + MOD_INVALID, /* This is needed to match the masks array */ + }; + static const int init_m_to_mem_type[] = { + MOD_INIT_TEXT, + MOD_INIT_RODATA, + MOD_INVALID, + MOD_INIT_DATA, + MOD_INVALID, /* This is needed to match the masks array */ + }; - pr_debug("Core section allocation order:\n"); for (m = 0; m < ARRAY_SIZE(masks); ++m) { + enum mod_mem_type type = is_init ? init_m_to_mem_type[m] : core_m_to_mem_type[m]; + for (i = 0; i < info->hdr->e_shnum; ++i) { Elf_Shdr *s = &info->sechdrs[i]; const char *sname = info->secstrings + s->sh_name; - unsigned int *sizep; if ((s->sh_flags & masks[m][0]) != masks[m][0] || (s->sh_flags & masks[m][1]) || s->sh_entsize != ~0UL - || module_init_layout_section(sname)) + || is_init != module_init_layout_section(sname)) continue; - sizep = m ? &mod->data_layout.size : &mod->core_layout.size; - s->sh_entsize = module_get_offset(mod, sizep, s, i); - pr_debug("\t%s\n", sname); - } - switch (m) { - case 0: /* executable */ - mod->core_layout.size = strict_align(mod->core_layout.size); - mod->core_layout.text_size = mod->core_layout.size; - break; - case 1: /* RO: text and ro-data */ - mod->data_layout.size = strict_align(mod->data_layout.size); - mod->data_layout.ro_size = mod->data_layout.size; - break; - case 2: /* RO after init */ - mod->data_layout.size = strict_align(mod->data_layout.size); - mod->data_layout.ro_after_init_size = mod->data_layout.size; - break; - case 4: /* whole core */ - mod->data_layout.size = strict_align(mod->data_layout.size); - break; - } - } - - pr_debug("Init section allocation order:\n"); - for (m = 0; m < ARRAY_SIZE(masks); ++m) { - for (i = 0; i < info->hdr->e_shnum; ++i) { - Elf_Shdr *s = &info->sechdrs[i]; - const char *sname = info->secstrings + s->sh_name; - if ((s->sh_flags & masks[m][0]) != masks[m][0] - || (s->sh_flags & masks[m][1]) - || s->sh_entsize != ~0UL - || !module_init_layout_section(sname)) + if (WARN_ON_ONCE(type == MOD_INVALID)) continue; - s->sh_entsize = (module_get_offset(mod, &mod->init_layout.size, s, i) - | INIT_OFFSET_MASK); + + s->sh_entsize = module_get_offset_and_type(mod, type, s, i); pr_debug("\t%s\n", sname); } - switch (m) { - case 0: /* executable */ - mod->init_layout.size = strict_align(mod->init_layout.size); - mod->init_layout.text_size = mod->init_layout.size; - break; - case 1: /* RO: text and ro-data */ - mod->init_layout.size = strict_align(mod->init_layout.size); - mod->init_layout.ro_size = mod->init_layout.size; - break; - case 2: - /* - * RO after init doesn't apply to init_layout (only - * core_layout), so it just takes the value of ro_size. - */ - mod->init_layout.ro_after_init_size = mod->init_layout.ro_size; - break; - case 4: /* whole init */ - mod->init_layout.size = strict_align(mod->init_layout.size); - break; - } } } +/* + * Lay out the SHF_ALLOC sections in a way not dissimilar to how ld + * might -- code, read-only data, read-write data, small data. Tally + * sizes, and place the offsets into sh_entsize fields: high bit means it + * belongs in init. + */ +static void layout_sections(struct module *mod, struct load_info *info) +{ + unsigned int i; + + for (i = 0; i < info->hdr->e_shnum; i++) + info->sechdrs[i].sh_entsize = ~0UL; + + pr_debug("Core section allocation order:\n"); + __layout_sections(mod, info, false); + + pr_debug("Init section allocation order:\n"); + __layout_sections(mod, info, true); +} + static void set_license(struct module *mod, const char *license) { if (!license) @@ -2122,72 +2149,41 @@ static int move_module(struct module *mod, struct load_info *info) { int i; void *ptr; + enum mod_mem_type t; - /* Do the allocs. */ - ptr = module_alloc(mod->core_layout.size); - /* - * The pointer to this block is stored in the module structure - * which is inside the block. Just mark it as not being a - * leak. - */ - kmemleak_not_leak(ptr); - if (!ptr) - return -ENOMEM; - - memset(ptr, 0, mod->core_layout.size); - mod->core_layout.base = ptr; + for_each_mod_mem_type(type) { + if (!mod->mem[type].size) { + mod->mem[type].base = NULL; + continue; + } + mod->mem[type].size = PAGE_ALIGN(mod->mem[type].size); + ptr = module_memory_alloc(mod->mem[type].size, type); - if (mod->init_layout.size) { - ptr = module_alloc(mod->init_layout.size); /* * The pointer to this block is stored in the module structure - * which is inside the block. This block doesn't need to be - * scanned as it contains data and code that will be freed - * after the module is initialized. + * which is inside the block. Just mark it as not being a + * leak. */ kmemleak_ignore(ptr); if (!ptr) { - module_memfree(mod->core_layout.base); - return -ENOMEM; + t = type; + goto out_enomem; } - memset(ptr, 0, mod->init_layout.size); - mod->init_layout.base = ptr; - } else - mod->init_layout.base = NULL; - -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - /* Do the allocs. */ - ptr = vzalloc(mod->data_layout.size); - /* - * The pointer to this block is stored in the module structure - * which is inside the block. Just mark it as not being a - * leak. - */ - kmemleak_not_leak(ptr); - if (!ptr) { - module_memfree(mod->core_layout.base); - module_memfree(mod->init_layout.base); - return -ENOMEM; + memset(ptr, 0, mod->mem[type].size); + mod->mem[type].base = ptr; } - mod->data_layout.base = ptr; -#endif /* Transfer each section which specifies SHF_ALLOC */ pr_debug("final section addresses:\n"); for (i = 0; i < info->hdr->e_shnum; i++) { void *dest; Elf_Shdr *shdr = &info->sechdrs[i]; + enum mod_mem_type type = shdr->sh_entsize >> SH_ENTSIZE_TYPE_SHIFT; if (!(shdr->sh_flags & SHF_ALLOC)) continue; - if (shdr->sh_entsize & INIT_OFFSET_MASK) - dest = mod->init_layout.base - + (shdr->sh_entsize & ~INIT_OFFSET_MASK); - else if (!(shdr->sh_flags & SHF_EXECINSTR)) - dest = mod->data_layout.base + shdr->sh_entsize; - else - dest = mod->core_layout.base + shdr->sh_entsize; + dest = mod->mem[type].base + (shdr->sh_entsize & SH_ENTSIZE_OFFSET_MASK); if (shdr->sh_type != SHT_NOBITS) memcpy(dest, (void *)shdr->sh_addr, shdr->sh_size); @@ -2198,6 +2194,10 @@ static int move_module(struct module *mod, struct load_info *info) } return 0; +out_enomem: + for (t--; t >= 0; t--) + module_memory_free(mod->mem[t].base, t); + return -ENOMEM; } static int check_module_license_and_versions(struct module *mod) @@ -2242,12 +2242,14 @@ static void flush_module_icache(const struct module *mod) * Do it before processing of module parameters, so the module * can provide parameter accessor functions of its own. */ - if (mod->init_layout.base) - flush_icache_range((unsigned long)mod->init_layout.base, - (unsigned long)mod->init_layout.base - + mod->init_layout.size); - flush_icache_range((unsigned long)mod->core_layout.base, - (unsigned long)mod->core_layout.base + mod->core_layout.size); + for_each_mod_mem_type(type) { + const struct module_memory *mod_mem = &mod->mem[type]; + + if (mod_mem->size) { + flush_icache_range((unsigned long)mod_mem->base, + (unsigned long)mod_mem->base + mod_mem->size); + } + } } bool __weak module_elf_check_arch(Elf_Ehdr *hdr) @@ -2350,11 +2352,8 @@ static void module_deallocate(struct module *mod, struct load_info *info) { percpu_modfree(mod); module_arch_freeing_init(mod); - module_memfree(mod->init_layout.base); - module_memfree(mod->core_layout.base); -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - vfree(mod->data_layout.base); -#endif + + free_mod_mem(mod); } int __weak module_finalize(const Elf_Ehdr *hdr, @@ -2415,7 +2414,9 @@ static void do_mod_ctors(struct module *mod) /* For freeing module_init on success, in case kallsyms traversing */ struct mod_initfree { struct llist_node node; - void *module_init; + void *init_text; + void *init_data; + void *init_rodata; }; static void do_free_init(struct work_struct *w) @@ -2429,7 +2430,9 @@ static void do_free_init(struct work_struct *w) llist_for_each_safe(pos, n, list) { initfree = container_of(pos, struct mod_initfree, node); - module_memfree(initfree->module_init); + module_memfree(initfree->init_text); + module_memfree(initfree->init_data); + module_memfree(initfree->init_rodata); kfree(initfree); } } @@ -2456,7 +2459,9 @@ static noinline int do_init_module(struct module *mod) ret = -ENOMEM; goto fail; } - freeinit->module_init = mod->init_layout.base; + freeinit->init_text = mod->mem[MOD_INIT_TEXT].base; + freeinit->init_data = mod->mem[MOD_INIT_DATA].base; + freeinit->init_rodata = mod->mem[MOD_INIT_RODATA].base; do_mod_ctors(mod); /* Start the module */ @@ -2492,8 +2497,8 @@ static noinline int do_init_module(struct module *mod) if (!mod->async_probe_requested) async_synchronize_full(); - ftrace_free_mem(mod, mod->init_layout.base, mod->init_layout.base + - mod->init_layout.size); + ftrace_free_mem(mod, mod->mem[MOD_INIT_TEXT].base, + mod->mem[MOD_INIT_TEXT].base + mod->mem[MOD_INIT_TEXT].size); mutex_lock(&module_mutex); /* Drop initial reference. */ module_put(mod); @@ -2505,11 +2510,10 @@ static noinline int do_init_module(struct module *mod) module_enable_ro(mod, true); mod_tree_remove_init(mod); module_arch_freeing_init(mod); - mod->init_layout.base = NULL; - mod->init_layout.size = 0; - mod->init_layout.ro_size = 0; - mod->init_layout.ro_after_init_size = 0; - mod->init_layout.text_size = 0; + for_class_mod_mem_type(type, init) { + mod->mem[type].base = NULL; + mod->mem[type].size = 0; + } #ifdef CONFIG_DEBUG_INFO_BTF_MODULES /* .BTF is not SHF_ALLOC and will get removed, so sanitize pointer */ mod->btf_data = NULL; @@ -2628,9 +2632,6 @@ static int complete_formation(struct module *mod, struct load_info *info) module_bug_finalize(info->hdr, info->sechdrs, mod); module_cfi_finalize(info->hdr, info->sechdrs, mod); - if (module_check_misalignment(mod)) - goto out_misaligned; - module_enable_ro(mod, false); module_enable_nx(mod); module_enable_x(mod); @@ -2644,8 +2645,6 @@ static int complete_formation(struct module *mod, struct load_info *info) return 0; -out_misaligned: - err = -EINVAL; out: mutex_unlock(&module_mutex); return err; @@ -2909,7 +2908,10 @@ static int load_module(struct load_info *info, const char __user *uargs, mutex_unlock(&module_mutex); free_module: /* Free lock-classes; relies on the preceding sync_rcu() */ - lockdep_free_key_range(mod->data_layout.base, mod->data_layout.size); + for_class_mod_mem_type(type, core_data) { + lockdep_free_key_range(mod->mem[type].base, + mod->mem[type].size); + } module_deallocate(mod, info); free_copy: @@ -3060,20 +3062,21 @@ bool is_module_address(unsigned long addr) struct module *__module_address(unsigned long addr) { struct module *mod; - struct mod_tree_root *tree; if (addr >= mod_tree.addr_min && addr <= mod_tree.addr_max) - tree = &mod_tree; + goto lookup; + #ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - else if (addr >= mod_data_tree.addr_min && addr <= mod_data_tree.addr_max) - tree = &mod_data_tree; + if (addr >= mod_tree.data_addr_min && addr <= mod_tree.data_addr_max) + goto lookup; #endif - else - return NULL; + return NULL; + +lookup: module_assert_mutex_or_preempt(); - mod = mod_find(addr, tree); + mod = mod_find(addr, &mod_tree); if (mod) { BUG_ON(!within_module(addr, mod)); if (mod->state == MODULE_STATE_UNFORMED) @@ -3113,8 +3116,8 @@ struct module *__module_text_address(unsigned long addr) struct module *mod = __module_address(addr); if (mod) { /* Make sure it's within the text section. */ - if (!within(addr, mod->init_layout.base, mod->init_layout.text_size) - && !within(addr, mod->core_layout.base, mod->core_layout.text_size)) + if (!within_module_mem_type(addr, mod, MOD_TEXT) && + !within_module_mem_type(addr, mod, MOD_INIT_TEXT)) mod = NULL; } return mod; diff --git a/kernel/module/procfs.c b/kernel/module/procfs.c index cf5b9f1e6ec4..0a4841e88adb 100644 --- a/kernel/module/procfs.c +++ b/kernel/module/procfs.c @@ -62,6 +62,15 @@ static void m_stop(struct seq_file *m, void *p) mutex_unlock(&module_mutex); } +static unsigned int module_total_size(struct module *mod) +{ + int size = 0; + + for_each_mod_mem_type(type) + size += mod->mem[type].size; + return size; +} + static int m_show(struct seq_file *m, void *p) { struct module *mod = list_entry(p, struct module, list); @@ -73,10 +82,7 @@ static int m_show(struct seq_file *m, void *p) if (mod->state == MODULE_STATE_UNFORMED) return 0; - size = mod->init_layout.size + mod->core_layout.size; -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - size += mod->data_layout.size; -#endif + size = module_total_size(mod); seq_printf(m, "%s %u", mod->name, size); print_unload_info(m, mod); @@ -86,7 +92,7 @@ static int m_show(struct seq_file *m, void *p) mod->state == MODULE_STATE_COMING ? "Loading" : "Live"); /* Used by oprofile and other similar tools. */ - value = m->private ? NULL : mod->core_layout.base; + value = m->private ? NULL : mod->mem[MOD_TEXT].base; seq_printf(m, " 0x%px", value); /* Taints info */ diff --git a/kernel/module/strict_rwx.c b/kernel/module/strict_rwx.c index 14fbea66f12f..a2b656b4e3d2 100644 --- a/kernel/module/strict_rwx.c +++ b/kernel/module/strict_rwx.c @@ -11,82 +11,25 @@ #include #include "internal.h" -/* - * LKM RO/NX protection: protect module's text/ro-data - * from modification and any data from execution. - * - * General layout of module is: - * [text] [read-only-data] [ro-after-init] [writable data] - * text_size -----^ ^ ^ ^ - * ro_size ------------------------| | | - * ro_after_init_size -----------------------------| | - * size -----------------------------------------------------------| - * - * These values are always page-aligned (as is base) when - * CONFIG_STRICT_MODULE_RWX is set. - */ +static void module_set_memory(const struct module *mod, enum mod_mem_type type, + int (*set_memory)(unsigned long start, int num_pages)) +{ + const struct module_memory *mod_mem = &mod->mem[type]; + + set_vm_flush_reset_perms(mod_mem->base); + set_memory((unsigned long)mod_mem->base, mod_mem->size >> PAGE_SHIFT); +} /* * Since some arches are moving towards PAGE_KERNEL module allocations instead - * of PAGE_KERNEL_EXEC, keep frob_text() and module_enable_x() independent of + * of PAGE_KERNEL_EXEC, keep module_enable_x() independent of * CONFIG_STRICT_MODULE_RWX because they are needed regardless of whether we * are strict. */ -static void frob_text(const struct module_layout *layout, - int (*set_memory)(unsigned long start, int num_pages)) -{ - set_memory((unsigned long)layout->base, - PAGE_ALIGN(layout->text_size) >> PAGE_SHIFT); -} - -static void frob_rodata(const struct module_layout *layout, - int (*set_memory)(unsigned long start, int num_pages)) -{ - set_memory((unsigned long)layout->base + layout->text_size, - (layout->ro_size - layout->text_size) >> PAGE_SHIFT); -} - -static void frob_ro_after_init(const struct module_layout *layout, - int (*set_memory)(unsigned long start, int num_pages)) -{ - set_memory((unsigned long)layout->base + layout->ro_size, - (layout->ro_after_init_size - layout->ro_size) >> PAGE_SHIFT); -} - -static void frob_writable_data(const struct module_layout *layout, - int (*set_memory)(unsigned long start, int num_pages)) -{ - set_memory((unsigned long)layout->base + layout->ro_after_init_size, - (layout->size - layout->ro_after_init_size) >> PAGE_SHIFT); -} - -static bool layout_check_misalignment(const struct module_layout *layout) -{ - return WARN_ON(!PAGE_ALIGNED(layout->base)) || - WARN_ON(!PAGE_ALIGNED(layout->text_size)) || - WARN_ON(!PAGE_ALIGNED(layout->ro_size)) || - WARN_ON(!PAGE_ALIGNED(layout->ro_after_init_size)) || - WARN_ON(!PAGE_ALIGNED(layout->size)); -} - -bool module_check_misalignment(const struct module *mod) -{ - if (!IS_ENABLED(CONFIG_STRICT_MODULE_RWX)) - return false; - - return layout_check_misalignment(&mod->core_layout) || - layout_check_misalignment(&mod->data_layout) || - layout_check_misalignment(&mod->init_layout); -} - void module_enable_x(const struct module *mod) { - if (!PAGE_ALIGNED(mod->core_layout.base) || - !PAGE_ALIGNED(mod->init_layout.base)) - return; - - frob_text(&mod->core_layout, set_memory_x); - frob_text(&mod->init_layout, set_memory_x); + for_class_mod_mem_type(type, text) + module_set_memory(mod, type, set_memory_x); } void module_enable_ro(const struct module *mod, bool after_init) @@ -98,16 +41,13 @@ void module_enable_ro(const struct module *mod, bool after_init) return; #endif - set_vm_flush_reset_perms(mod->core_layout.base); - set_vm_flush_reset_perms(mod->init_layout.base); - frob_text(&mod->core_layout, set_memory_ro); - - frob_rodata(&mod->data_layout, set_memory_ro); - frob_text(&mod->init_layout, set_memory_ro); - frob_rodata(&mod->init_layout, set_memory_ro); + module_set_memory(mod, MOD_TEXT, set_memory_ro); + module_set_memory(mod, MOD_INIT_TEXT, set_memory_ro); + module_set_memory(mod, MOD_RODATA, set_memory_ro); + module_set_memory(mod, MOD_INIT_RODATA, set_memory_ro); if (after_init) - frob_ro_after_init(&mod->data_layout, set_memory_ro); + module_set_memory(mod, MOD_RO_AFTER_INIT, set_memory_ro); } void module_enable_nx(const struct module *mod) @@ -115,11 +55,8 @@ void module_enable_nx(const struct module *mod) if (!IS_ENABLED(CONFIG_STRICT_MODULE_RWX)) return; - frob_rodata(&mod->data_layout, set_memory_nx); - frob_ro_after_init(&mod->data_layout, set_memory_nx); - frob_writable_data(&mod->data_layout, set_memory_nx); - frob_rodata(&mod->init_layout, set_memory_nx); - frob_writable_data(&mod->init_layout, set_memory_nx); + for_class_mod_mem_type(type, data) + module_set_memory(mod, type, set_memory_nx); } int module_enforce_rwx_sections(Elf_Ehdr *hdr, Elf_Shdr *sechdrs, diff --git a/kernel/module/tree_lookup.c b/kernel/module/tree_lookup.c index 8ec5cfd60496..277197977d43 100644 --- a/kernel/module/tree_lookup.c +++ b/kernel/module/tree_lookup.c @@ -21,16 +21,16 @@ static __always_inline unsigned long __mod_tree_val(struct latch_tree_node *n) { - struct module_layout *layout = container_of(n, struct module_layout, mtn.node); + struct module_memory *mod_mem = container_of(n, struct module_memory, mtn.node); - return (unsigned long)layout->base; + return (unsigned long)mod_mem->base; } static __always_inline unsigned long __mod_tree_size(struct latch_tree_node *n) { - struct module_layout *layout = container_of(n, struct module_layout, mtn.node); + struct module_memory *mod_mem = container_of(n, struct module_memory, mtn.node); - return (unsigned long)layout->size; + return (unsigned long)mod_mem->size; } static __always_inline bool @@ -77,32 +77,27 @@ static void __mod_tree_remove(struct mod_tree_node *node, struct mod_tree_root * */ void mod_tree_insert(struct module *mod) { - mod->core_layout.mtn.mod = mod; - mod->init_layout.mtn.mod = mod; - - __mod_tree_insert(&mod->core_layout.mtn, &mod_tree); - if (mod->init_layout.size) - __mod_tree_insert(&mod->init_layout.mtn, &mod_tree); - -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - mod->data_layout.mtn.mod = mod; - __mod_tree_insert(&mod->data_layout.mtn, &mod_data_tree); -#endif + for_each_mod_mem_type(type) { + mod->mem[type].mtn.mod = mod; + if (mod->mem[type].size) + __mod_tree_insert(&mod->mem[type].mtn, &mod_tree); + } } void mod_tree_remove_init(struct module *mod) { - if (mod->init_layout.size) - __mod_tree_remove(&mod->init_layout.mtn, &mod_tree); + for_class_mod_mem_type(type, init) { + if (mod->mem[type].size) + __mod_tree_remove(&mod->mem[type].mtn, &mod_tree); + } } void mod_tree_remove(struct module *mod) { - __mod_tree_remove(&mod->core_layout.mtn, &mod_tree); - mod_tree_remove_init(mod); -#ifdef CONFIG_ARCH_WANTS_MODULES_DATA_IN_VMALLOC - __mod_tree_remove(&mod->data_layout.mtn, &mod_data_tree); -#endif + for_each_mod_mem_type(type) { + if (mod->mem[type].size) + __mod_tree_remove(&mod->mem[type].mtn, &mod_tree); + } } struct module *mod_find(unsigned long addr, struct mod_tree_root *tree) -- cgit From 9e07f161717ab8e8ac1206bf82546511e24cbb7b Mon Sep 17 00:00:00 2001 From: Jiapeng Chong Date: Fri, 10 Feb 2023 14:42:43 +0800 Subject: module: Remove the unused function within The function within is defined in the main.c file, but not called elsewhere, so remove this unused function. This routine became no longer used after commit ("module: replace module_layout with module_memory"). kernel/module/main.c:3007:19: warning: unused function 'within'. Reported-by: Abaci Robot Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=4035 Signed-off-by: Jiapeng Chong [mcgrof: adjust commit log to explain why this change is needed] Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 2724bc1b9a90..1fa529668a45 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2975,11 +2975,6 @@ SYSCALL_DEFINE3(finit_module, int, fd, const char __user *, uargs, int, flags) return load_module(&info, uargs, flags); } -static inline int within(unsigned long addr, void *start, unsigned long size) -{ - return ((void *)addr >= start && (void *)addr < start + size); -} - /* Keep in sync with MODULE_FLAGS_BUF_SIZE !!! */ char *module_flags(struct module *mod, char *buf, bool show_state) { -- cgit From 7deabd67498869640c937c9bd83472574b7dea0b Mon Sep 17 00:00:00 2001 From: Jason Baron Date: Fri, 3 Mar 2023 11:50:56 -0500 Subject: dyndbg: use the module notifier callbacks Bring dynamic debug in line with other subsystems by using the module notifier callbacks. This results in a net decrease in core module code. Additionally, Jim Cromie has a new dynamic debug classmap feature, which requires that jump labels be initialized prior to dynamic debug. Specifically, the new feature toggles a jump label from the existing dynamic_debug_setup() function. However, this does not currently work properly, because jump labels are initialized via the 'module_notify_list' notifier chain, which is invoked after the current call to dynamic_debug_setup(). Thus, this patch ensures that jump labels are initialized prior to dynamic debug by setting the dynamic debug notifier priority to 0, while jump labels have the higher priority of 1. Tested by Jim using his new test case, and I've verfied the correct printing via: # modprobe test_dynamic_debug dyndbg. Link: https://lore.kernel.org/lkml/20230113193016.749791-21-jim.cromie@gmail.com/ Reported-by: kernel test robot Link: https://lore.kernel.org/oe-kbuild-all/202302190427.9iIK2NfJ-lkp@intel.com/ Tested-by: Jim Cromie Reviewed-by: Vincenzo Palazzo Cc: Peter Zijlstra CC: Jim Cromie Cc: Luis Chamberlain Cc: Greg Kroah-Hartman Signed-off-by: Jason Baron Signed-off-by: Luis Chamberlain --- kernel/module/internal.h | 2 -- kernel/module/main.c | 30 ++++++++---------------------- 2 files changed, 8 insertions(+), 24 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/internal.h b/kernel/module/internal.h index c6c44ceca07a..e3883b7d4840 100644 --- a/kernel/module/internal.h +++ b/kernel/module/internal.h @@ -45,7 +45,6 @@ extern const struct kernel_symbol __stop___ksymtab_gpl[]; extern const s32 __start___kcrctab[]; extern const s32 __start___kcrctab_gpl[]; -#include struct load_info { const char *name; /* pointer to module in temporary copy, freed at end of load_module() */ @@ -55,7 +54,6 @@ struct load_info { Elf_Shdr *sechdrs; char *secstrings, *strtab; unsigned long symoffs, stroffs, init_typeoffs, core_typeoffs; - struct _ddebug_info dyndbg; bool sig_ok; #ifdef CONFIG_KALLSYMS unsigned long mod_kallsyms_init_off; diff --git a/kernel/module/main.c b/kernel/module/main.c index 1fa529668a45..b4759f1695b7 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1216,9 +1216,6 @@ static void free_module(struct module *mod) mod->state = MODULE_STATE_UNFORMED; mutex_unlock(&module_mutex); - /* Remove dynamic debug info */ - ddebug_remove_module(mod->name); - /* Arch-specific cleanup. */ module_arch_cleanup(mod); @@ -1619,19 +1616,6 @@ static void free_modinfo(struct module *mod) } } -static void dynamic_debug_setup(struct module *mod, struct _ddebug_info *dyndbg) -{ - if (!dyndbg->num_descs) - return; - ddebug_add_module(dyndbg, mod->name); -} - -static void dynamic_debug_remove(struct module *mod, struct _ddebug_info *dyndbg) -{ - if (dyndbg->num_descs) - ddebug_remove_module(mod->name); -} - void * __weak module_alloc(unsigned long size) { return __vmalloc_node_range(size, 1, VMALLOC_START, VMALLOC_END, @@ -2137,10 +2121,14 @@ static int find_module_sections(struct module *mod, struct load_info *info) if (section_addr(info, "__obsparm")) pr_warn("%s: Ignoring obsolete parameters\n", mod->name); - info->dyndbg.descs = section_objs(info, "__dyndbg", - sizeof(*info->dyndbg.descs), &info->dyndbg.num_descs); - info->dyndbg.classes = section_objs(info, "__dyndbg_classes", - sizeof(*info->dyndbg.classes), &info->dyndbg.num_classes); +#ifdef CONFIG_DYNAMIC_DEBUG_CORE + mod->dyndbg_info.descs = section_objs(info, "__dyndbg", + sizeof(*mod->dyndbg_info.descs), + &mod->dyndbg_info.num_descs); + mod->dyndbg_info.classes = section_objs(info, "__dyndbg_classes", + sizeof(*mod->dyndbg_info.classes), + &mod->dyndbg_info.num_classes); +#endif return 0; } @@ -2824,7 +2812,6 @@ static int load_module(struct load_info *info, const char __user *uargs, } init_build_id(mod, info); - dynamic_debug_setup(mod, &info->dyndbg); /* Ftrace init must be called in the MODULE_STATE_UNFORMED state */ ftrace_module_init(mod); @@ -2888,7 +2875,6 @@ static int load_module(struct load_info *info, const char __user *uargs, ddebug_cleanup: ftrace_release_mod(mod); - dynamic_debug_remove(mod, &info->dyndbg); synchronize_rcu(); kfree(mod->args); free_arch_cleanup: -- cgit From 3703bd54cd37e7875f51ece8df8c85c184e40bba Mon Sep 17 00:00:00 2001 From: Zhen Lei Date: Wed, 8 Mar 2023 15:38:46 +0800 Subject: kallsyms: Delete an unused parameter related to {module_}kallsyms_on_each_symbol() The parameter 'struct module *' in the hook function associated with {module_}kallsyms_on_each_symbol() is no longer used. Delete it. Suggested-by: Petr Mladek Signed-off-by: Zhen Lei Reviewed-by: Vincenzo Palazzo Acked-by: Jiri Olsa Acked-by: Steven Rostedt (Google) Signed-off-by: Luis Chamberlain --- kernel/module/kallsyms.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c index 8e89f459ffdd..3320586584f1 100644 --- a/kernel/module/kallsyms.c +++ b/kernel/module/kallsyms.c @@ -503,8 +503,7 @@ unsigned long module_kallsyms_lookup_name(const char *name) } int module_kallsyms_on_each_symbol(const char *modname, - int (*fn)(void *, const char *, - struct module *, unsigned long), + int (*fn)(void *, const char *, unsigned long), void *data) { struct module *mod; @@ -533,7 +532,7 @@ int module_kallsyms_on_each_symbol(const char *modname, continue; ret = fn(data, kallsyms_symbol_name(kallsyms, i), - mod, kallsyms_symbol_value(sym)); + kallsyms_symbol_value(sym)); if (ret != 0) goto out; } -- cgit From 3c17655ab13704582fe25e8ea3200a9b2f8bf20a Mon Sep 17 00:00:00 2001 From: "Fabio M. De Francesco" Date: Wed, 15 Mar 2023 13:52:56 +0100 Subject: module/decompress: Never use kunmap() for local un-mappings Use kunmap_local() to unmap pages locally mapped with kmap_local_page(). kunmap_local() must be called on the kernel virtual address returned by kmap_local_page(), differently from how we use kunmap() which instead expects the mapped page as its argument. In module_zstd_decompress() we currently map with kmap_local_page() and unmap with kunmap(). This breaks the code and so it should be fixed. Cc: Piotr Gorski Cc: Dmitry Torokhov Cc: Luis Chamberlain Cc: Stephen Boyd Cc: Ira Weiny Fixes: 169a58ad824d ("module/decompress: Support zstd in-kernel decompression") Signed-off-by: Fabio M. De Francesco Reviewed-by: Stephen Boyd Reviewed-by: Ira Weiny Reviewed-by: Piotr Gorski Signed-off-by: Luis Chamberlain --- kernel/module/decompress.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/module') diff --git a/kernel/module/decompress.c b/kernel/module/decompress.c index bb79ac1a6d8f..7ddc87bee274 100644 --- a/kernel/module/decompress.c +++ b/kernel/module/decompress.c @@ -267,7 +267,7 @@ static ssize_t module_zstd_decompress(struct load_info *info, zstd_dec.size = PAGE_SIZE; ret = zstd_decompress_stream(dstream, &zstd_dec, &zstd_buf); - kunmap(page); + kunmap_local(zstd_dec.dst); retval = zstd_get_error_code(ret); if (retval) break; -- cgit From b66973b82d4426b318f78a20c6b39ddd977a508a Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:35 -0700 Subject: module: move get_modinfo() helpers all above Instead of forward declaring routines for get_modinfo() just move everything up. This makes no functional changes. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 100 +++++++++++++++++++++++++-------------------------- 1 file changed, 48 insertions(+), 52 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index b4759f1695b7..1e739f534100 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1016,9 +1016,55 @@ int try_to_force_load(struct module *mod, const char *reason) #endif } -static char *get_modinfo(const struct load_info *info, const char *tag); +/* Parse tag=value strings from .modinfo section */ +static char *next_string(char *string, unsigned long *secsize) +{ + /* Skip non-zero chars */ + while (string[0]) { + string++; + if ((*secsize)-- <= 1) + return NULL; + } + + /* Skip any zero padding. */ + while (!string[0]) { + string++; + if ((*secsize)-- <= 1) + return NULL; + } + return string; +} + static char *get_next_modinfo(const struct load_info *info, const char *tag, - char *prev); + char *prev) +{ + char *p; + unsigned int taglen = strlen(tag); + Elf_Shdr *infosec = &info->sechdrs[info->index.info]; + unsigned long size = infosec->sh_size; + + /* + * get_modinfo() calls made before rewrite_section_headers() + * must use sh_offset, as sh_addr isn't set! + */ + char *modinfo = (char *)info->hdr + infosec->sh_offset; + + if (prev) { + size -= prev - modinfo; + modinfo = next_string(prev, &size); + } + + for (p = modinfo; p; p = next_string(p, &size)) { + if (strncmp(p, tag, taglen) == 0 && p[taglen] == '=') + return p + taglen + 1; + } + return NULL; +} + +static char *get_modinfo(const struct load_info *info, const char *tag) +{ + return get_next_modinfo(info, tag, NULL); +} static int verify_namespace_is_imported(const struct load_info *info, const struct kernel_symbol *sym, @@ -1544,56 +1590,6 @@ static void set_license(struct module *mod, const char *license) } } -/* Parse tag=value strings from .modinfo section */ -static char *next_string(char *string, unsigned long *secsize) -{ - /* Skip non-zero chars */ - while (string[0]) { - string++; - if ((*secsize)-- <= 1) - return NULL; - } - - /* Skip any zero padding. */ - while (!string[0]) { - string++; - if ((*secsize)-- <= 1) - return NULL; - } - return string; -} - -static char *get_next_modinfo(const struct load_info *info, const char *tag, - char *prev) -{ - char *p; - unsigned int taglen = strlen(tag); - Elf_Shdr *infosec = &info->sechdrs[info->index.info]; - unsigned long size = infosec->sh_size; - - /* - * get_modinfo() calls made before rewrite_section_headers() - * must use sh_offset, as sh_addr isn't set! - */ - char *modinfo = (char *)info->hdr + infosec->sh_offset; - - if (prev) { - size -= prev - modinfo; - modinfo = next_string(prev, &size); - } - - for (p = modinfo; p; p = next_string(p, &size)) { - if (strncmp(p, tag, taglen) == 0 && p[taglen] == '=') - return p + taglen + 1; - } - return NULL; -} - -static char *get_modinfo(const struct load_info *info, const char *tag) -{ - return get_next_modinfo(info, tag, NULL); -} - static void setup_modinfo(struct module *mod, struct load_info *info) { struct module_attribute *attr; -- cgit From feb5b784a26363b690f618213450faf244c1c58e Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:36 -0700 Subject: module: rename next_string() to module_next_tag_pair() This makes it clearer what it is doing. While at it, make it available to other code other than main.c. This will be used in the subsequent patch and make the changes easier to read. Signed-off-by: Luis Chamberlain --- kernel/module/internal.h | 2 ++ kernel/module/main.c | 6 +++--- 2 files changed, 5 insertions(+), 3 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/internal.h b/kernel/module/internal.h index e3883b7d4840..1fa2328636ec 100644 --- a/kernel/module/internal.h +++ b/kernel/module/internal.h @@ -96,6 +96,8 @@ long module_get_offset_and_type(struct module *mod, enum mod_mem_type type, char *module_flags(struct module *mod, char *buf, bool show_state); size_t module_flags_taint(unsigned long taints, char *buf); +char *module_next_tag_pair(char *string, unsigned long *secsize); + static inline void module_assert_mutex_or_preempt(void) { #ifdef CONFIG_LOCKDEP diff --git a/kernel/module/main.c b/kernel/module/main.c index 1e739f534100..ebb5e6b92a48 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1017,7 +1017,7 @@ int try_to_force_load(struct module *mod, const char *reason) } /* Parse tag=value strings from .modinfo section */ -static char *next_string(char *string, unsigned long *secsize) +char *module_next_tag_pair(char *string, unsigned long *secsize) { /* Skip non-zero chars */ while (string[0]) { @@ -1051,10 +1051,10 @@ static char *get_next_modinfo(const struct load_info *info, const char *tag, if (prev) { size -= prev - modinfo; - modinfo = next_string(prev, &size); + modinfo = module_next_tag_pair(prev, &size); } - for (p = modinfo; p; p = next_string(p, &size)) { + for (p = modinfo; p; p = module_next_tag_pair(p, &size)) { if (strncmp(p, tag, taglen) == 0 && p[taglen] == '=') return p + taglen + 1; } -- cgit From 1e684172358453df1cb783d7c101a09ff08ceee1 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:37 -0700 Subject: module: add a for_each_modinfo_entry() Add a for_each_modinfo_entry() to make it easier to read and use. This produces no functional changes but makes this code easiert to read as we are used to with loops in the kernel and trims more lines of code. Signed-off-by: Luis Chamberlain --- kernel/module/internal.h | 3 +++ kernel/module/main.c | 5 +---- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/internal.h b/kernel/module/internal.h index 1fa2328636ec..6ae29bb8836f 100644 --- a/kernel/module/internal.h +++ b/kernel/module/internal.h @@ -98,6 +98,9 @@ size_t module_flags_taint(unsigned long taints, char *buf); char *module_next_tag_pair(char *string, unsigned long *secsize); +#define for_each_modinfo_entry(entry, info, name) \ + for (entry = get_modinfo(info, name); entry; entry = get_next_modinfo(info, name, entry)) + static inline void module_assert_mutex_or_preempt(void) { #ifdef CONFIG_LOCKDEP diff --git a/kernel/module/main.c b/kernel/module/main.c index ebb5e6b92a48..427284ab31f1 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1075,12 +1075,9 @@ static int verify_namespace_is_imported(const struct load_info *info, namespace = kernel_symbol_namespace(sym); if (namespace && namespace[0]) { - imported_namespace = get_modinfo(info, "import_ns"); - while (imported_namespace) { + for_each_modinfo_entry(imported_namespace, info, "import_ns") { if (strcmp(namespace, imported_namespace) == 0) return 0; - imported_namespace = get_next_modinfo( - info, "import_ns", imported_namespace); } #ifdef CONFIG_MODULE_ALLOW_MISSING_NAMESPACE_IMPORTS pr_warn( -- cgit From 85e6f61c134f111232d27d3f63667c1bccbbc12d Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:38 -0700 Subject: module: move early sanity checks into a helper Move early sanity checkers for the module into a helper. This let's us make it clear when we are working with the local copy of the module prior to allocation. This produces no functional changes, it just makes subsequent changes easier to read. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 43 ++++++++++++++++++++++++++----------------- 1 file changed, 26 insertions(+), 17 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 427284ab31f1..22596cb46dc8 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2668,6 +2668,31 @@ static int unknown_module_param_cb(char *param, char *val, const char *modname, return 0; } +/* Module within temporary copy, this doesn't do any allocation */ +static int early_mod_check(struct load_info *info, int flags) +{ + int err; + + /* + * Now that we know we have the correct module name, check + * if it's blacklisted. + */ + if (blacklisted(info->name)) { + pr_err("Module %s is blacklisted\n", info->name); + return -EPERM; + } + + err = rewrite_section_headers(info, flags); + if (err) + return err; + + /* Check module struct version now, before we try to use module. */ + if (!check_modstruct_version(info, info->mod)) + return -ENOEXEC; + + return 0; +} + /* * Allocate and load the module: note that size of section 0 is always * zero, and we rely on this for optional sections. @@ -2711,26 +2736,10 @@ static int load_module(struct load_info *info, const char __user *uargs, if (err) goto free_copy; - /* - * Now that we know we have the correct module name, check - * if it's blacklisted. - */ - if (blacklisted(info->name)) { - err = -EPERM; - pr_err("Module %s is blacklisted\n", info->name); - goto free_copy; - } - - err = rewrite_section_headers(info, flags); + err = early_mod_check(info, flags); if (err) goto free_copy; - /* Check module struct version now, before we try to use module. */ - if (!check_modstruct_version(info, info->mod)) { - err = -ENOEXEC; - goto free_copy; - } - /* Figure out module layout, and allocate all the memory. */ mod = layout_and_allocate(info, flags); if (IS_ERR(mod)) { -- cgit From 02da2cbab452a236fa67abc9fc9e47430934e652 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:39 -0700 Subject: module: move check_modinfo() early to early_mod_check() This moves check_modinfo() to early_mod_check(). This doesn't make any functional changes either, as check_modinfo() was the first call on layout_and_allocate(), so we're just moving it back one routine and at the end. This let's us keep separate the checkers from the allocator. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 22596cb46dc8..95fd705328ac 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2273,10 +2273,6 @@ static struct module *layout_and_allocate(struct load_info *info, int flags) unsigned int ndx; int err; - err = check_modinfo(info->mod, info, flags); - if (err) - return ERR_PTR(err); - /* Allow arches to frob section contents and sizes. */ err = module_frob_arch_sections(info->hdr, info->sechdrs, info->secstrings, info->mod); @@ -2690,6 +2686,10 @@ static int early_mod_check(struct load_info *info, int flags) if (!check_modstruct_version(info, info->mod)) return -ENOEXEC; + err = check_modinfo(info->mod, info, flags); + if (err) + return err; + return 0; } -- cgit From ad8d3a36e981064f8a646531cddca30516894457 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:40 -0700 Subject: module: rename set_license() to module_license_taint_check() The set_license() routine would seem to a reader to do some sort of setting, but it does not. It just adds a taint if the license is not set or proprietary. This makes what the code is doing clearer, so much we can remove the comment about it. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 95fd705328ac..5e64485ac05a 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1573,7 +1573,7 @@ static void layout_sections(struct module *mod, struct load_info *info) __layout_sections(mod, info, true); } -static void set_license(struct module *mod, const char *license) +static void module_license_taint_check(struct module *mod, const char *license) { if (!license) license = "unspecified"; @@ -1993,8 +1993,7 @@ static int check_modinfo(struct module *mod, struct load_info *info, int flags) if (err) return err; - /* Set up license info based on the info section */ - set_license(mod, get_modinfo(info, "license")); + module_license_taint_check(mod, get_modinfo(info, "license")); if (get_modinfo(info, "test")) { if (!test_taint(TAINT_TEST)) -- cgit From ed52cabecb7a7e4242cee9c370c2622a47177d5d Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:41 -0700 Subject: module: split taint work out of check_modinfo_livepatch() The work to taint the kernel due to a module should be split up eventually. To aid with this, split up the tainting on check_modinfo_livepatch(). This let's us bring more early checks together which do return a value, and makes changes easier to read later where we stuff all the work to do the taints in one single routine. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 5e64485ac05a..cfb2ff5185fe 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1808,12 +1808,8 @@ static int check_modinfo_livepatch(struct module *mod, struct load_info *info) /* Nothing more to do */ return 0; - if (set_livepatch_module(mod)) { - add_taint_module(mod, TAINT_LIVEPATCH, LOCKDEP_STILL_OK); - pr_notice_once("%s: tainting kernel with TAINT_LIVEPATCH\n", - mod->name); + if (set_livepatch_module(mod)) return 0; - } pr_err("%s: module is marked as livepatch module, but livepatch support is disabled", mod->name); @@ -1993,6 +1989,11 @@ static int check_modinfo(struct module *mod, struct load_info *info, int flags) if (err) return err; + if (is_livepatch_module(mod)) { + add_taint_module(mod, TAINT_LIVEPATCH, LOCKDEP_STILL_OK); + pr_notice_once("%s: tainting kernel with TAINT_LIVEPATCH\n", + mod->name); + } module_license_taint_check(mod, get_modinfo(info, "license")); if (get_modinfo(info, "test")) { -- cgit From 437c1f9cc61fd37829eaf12d8ae2f7dcc5dddce0 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:42 -0700 Subject: module: split taint adding with info checking check_modinfo() actually does two things: a) sanity checks, some of which are fatal, and so we prevent the user from completing trying to load a module b) taints the kernel The taints are pretty heavy handed because we're tainting the kernel *before* we ever even get to load the module into the modules linked list. That is, it it can fail for other reasons later as we review the module's structure. But this commit makes no functional changes, it just makes the intent clearer and splits the code up where needed to make that happen. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 62 +++++++++++++++++++++++++++++++++------------------- 1 file changed, 40 insertions(+), 22 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index cfb2ff5185fe..a3953ca18090 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1951,25 +1951,10 @@ static int setup_load_info(struct load_info *info, int flags) return 0; } -static int check_modinfo(struct module *mod, struct load_info *info, int flags) +/* + * These calls taint the kernel depending certain module circumstances */ +static void module_augment_kernel_taints(struct module *mod, struct load_info *info) { - const char *modmagic = get_modinfo(info, "vermagic"); - int err; - - if (flags & MODULE_INIT_IGNORE_VERMAGIC) - modmagic = NULL; - - /* This is allowed: modprobe --force will invalidate it. */ - if (!modmagic) { - err = try_to_force_load(mod, "bad vermagic"); - if (err) - return err; - } else if (!same_magic(modmagic, vermagic, info->index.vers)) { - pr_err("%s: version magic '%s' should be '%s'\n", - info->name, modmagic, vermagic); - return -ENOEXEC; - } - if (!get_modinfo(info, "intree")) { if (!test_taint(TAINT_OOT_MODULE)) pr_warn("%s: loading out-of-tree module taints kernel.\n", @@ -1985,15 +1970,12 @@ static int check_modinfo(struct module *mod, struct load_info *info, int flags) "is unknown, you have been warned.\n", mod->name); } - err = check_modinfo_livepatch(mod, info); - if (err) - return err; - if (is_livepatch_module(mod)) { add_taint_module(mod, TAINT_LIVEPATCH, LOCKDEP_STILL_OK); pr_notice_once("%s: tainting kernel with TAINT_LIVEPATCH\n", mod->name); } + module_license_taint_check(mod, get_modinfo(info, "license")); if (get_modinfo(info, "test")) { @@ -2002,6 +1984,42 @@ static int check_modinfo(struct module *mod, struct load_info *info, int flags) mod->name); add_taint_module(mod, TAINT_TEST, LOCKDEP_STILL_OK); } +} + +static int check_modinfo(struct module *mod, struct load_info *info, int flags) +{ + const char *modmagic = get_modinfo(info, "vermagic"); + int err; + + if (flags & MODULE_INIT_IGNORE_VERMAGIC) + modmagic = NULL; + + /* This is allowed: modprobe --force will invalidate it. */ + if (!modmagic) { + err = try_to_force_load(mod, "bad vermagic"); + if (err) + return err; + } else if (!same_magic(modmagic, vermagic, info->index.vers)) { + pr_err("%s: version magic '%s' should be '%s'\n", + info->name, modmagic, vermagic); + return -ENOEXEC; + } + + err = check_modinfo_livepatch(mod, info); + if (err) + return err; + + /* + * We are tainting your kernel *even* if you try to load + * modules with possible taints and we fail to load these + * modules for other reasons. + * + * We have a descrepancy though, see the other taints for + * signature and those in check_module_license_and_versions(). + * + * We should compromise and converge. + */ + module_augment_kernel_taints(mod, info); return 0; } -- cgit From a12b94511cf36855cd731c16005bd535e2007552 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:43 -0700 Subject: module: move tainting until after a module hits our linked list It is silly to have taints spread out all over, we can just compromise and add them if the module ever hit our linked list. Our sanity checkers should just prevent crappy drivers / bogus ELF modules / etc and kconfig options should be enough to let you *not* load things you don't want. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 22 ++++++++++------------ 1 file changed, 10 insertions(+), 12 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index a3953ca18090..1aa71f82aca2 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2009,18 +2009,6 @@ static int check_modinfo(struct module *mod, struct load_info *info, int flags) if (err) return err; - /* - * We are tainting your kernel *even* if you try to load - * modules with possible taints and we fail to load these - * modules for other reasons. - * - * We have a descrepancy though, see the other taints for - * signature and those in check_module_license_and_versions(). - * - * We should compromise and converge. - */ - module_augment_kernel_taints(mod, info); - return 0; } @@ -2772,6 +2760,16 @@ static int load_module(struct load_info *info, const char __user *uargs, if (err) goto free_module; + /* + * We are tainting your kernel if your module gets into + * the modules linked list somehow. + * + * We have a descrepancy though, see the other taints for + * signature and those in check_module_license_and_versions(). + * + * We should compromise and converge. + */ + module_augment_kernel_taints(mod, info); #ifdef CONFIG_MODULE_SIG mod->sig_ok = info->sig_ok; if (!mod->sig_ok) { -- cgit From c3bbf62ebf8c9e87cea875cfa146f44f46af4145 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:44 -0700 Subject: module: move signature taint to module_augment_kernel_taints() Just move the signature taint into the helper: module_augment_kernel_taints() Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 1aa71f82aca2..2f1988137965 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1984,6 +1984,15 @@ static void module_augment_kernel_taints(struct module *mod, struct load_info *i mod->name); add_taint_module(mod, TAINT_TEST, LOCKDEP_STILL_OK); } +#ifdef CONFIG_MODULE_SIG + mod->sig_ok = info->sig_ok; + if (!mod->sig_ok) { + pr_notice_once("%s: module verification failed: signature " + "and/or required key missing - tainting " + "kernel\n", mod->name); + add_taint_module(mod, TAINT_UNSIGNED_MODULE, LOCKDEP_STILL_OK); + } +#endif } static int check_modinfo(struct module *mod, struct load_info *info, int flags) @@ -2770,15 +2779,6 @@ static int load_module(struct load_info *info, const char __user *uargs, * We should compromise and converge. */ module_augment_kernel_taints(mod, info); -#ifdef CONFIG_MODULE_SIG - mod->sig_ok = info->sig_ok; - if (!mod->sig_ok) { - pr_notice_once("%s: module verification failed: signature " - "and/or required key missing - tainting " - "kernel\n", mod->name); - add_taint_module(mod, TAINT_UNSIGNED_MODULE, LOCKDEP_STILL_OK); - } -#endif /* To avoid stressing percpu allocator, do this once we're unique. */ err = percpu_modalloc(mod, info); -- cgit From 72f08b3cc631f4ebcaa9f373d18fc0b877fb6458 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:45 -0700 Subject: module: converge taint work together Converge on a compromise: so long as we have a module hit our linked list of modules we taint. That is, the module was about to become live. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 52 ++++++++++++++++++++++++---------------------------- 1 file changed, 24 insertions(+), 28 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 2f1988137965..f165d93a4ef9 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1955,6 +1955,8 @@ static int setup_load_info(struct load_info *info, int flags) * These calls taint the kernel depending certain module circumstances */ static void module_augment_kernel_taints(struct module *mod, struct load_info *info) { + int prev_taint = test_taint(TAINT_PROPRIETARY_MODULE); + if (!get_modinfo(info, "intree")) { if (!test_taint(TAINT_OOT_MODULE)) pr_warn("%s: loading out-of-tree module taints kernel.\n", @@ -1993,6 +1995,28 @@ static void module_augment_kernel_taints(struct module *mod, struct load_info *i add_taint_module(mod, TAINT_UNSIGNED_MODULE, LOCKDEP_STILL_OK); } #endif + + /* + * ndiswrapper is under GPL by itself, but loads proprietary modules. + * Don't use add_taint_module(), as it would prevent ndiswrapper from + * using GPL-only symbols it needs. + */ + if (strcmp(mod->name, "ndiswrapper") == 0) + add_taint(TAINT_PROPRIETARY_MODULE, LOCKDEP_NOW_UNRELIABLE); + + /* driverloader was caught wrongly pretending to be under GPL */ + if (strcmp(mod->name, "driverloader") == 0) + add_taint_module(mod, TAINT_PROPRIETARY_MODULE, + LOCKDEP_NOW_UNRELIABLE); + + /* lve claims to be GPL but upstream won't provide source */ + if (strcmp(mod->name, "lve") == 0) + add_taint_module(mod, TAINT_PROPRIETARY_MODULE, + LOCKDEP_NOW_UNRELIABLE); + + if (!prev_taint && test_taint(TAINT_PROPRIETARY_MODULE)) + pr_warn("%s: module license taints kernel.\n", mod->name); + } static int check_modinfo(struct module *mod, struct load_info *info, int flags) @@ -2198,29 +2222,6 @@ out_enomem: static int check_module_license_and_versions(struct module *mod) { - int prev_taint = test_taint(TAINT_PROPRIETARY_MODULE); - - /* - * ndiswrapper is under GPL by itself, but loads proprietary modules. - * Don't use add_taint_module(), as it would prevent ndiswrapper from - * using GPL-only symbols it needs. - */ - if (strcmp(mod->name, "ndiswrapper") == 0) - add_taint(TAINT_PROPRIETARY_MODULE, LOCKDEP_NOW_UNRELIABLE); - - /* driverloader was caught wrongly pretending to be under GPL */ - if (strcmp(mod->name, "driverloader") == 0) - add_taint_module(mod, TAINT_PROPRIETARY_MODULE, - LOCKDEP_NOW_UNRELIABLE); - - /* lve claims to be GPL but upstream won't provide source */ - if (strcmp(mod->name, "lve") == 0) - add_taint_module(mod, TAINT_PROPRIETARY_MODULE, - LOCKDEP_NOW_UNRELIABLE); - - if (!prev_taint && test_taint(TAINT_PROPRIETARY_MODULE)) - pr_warn("%s: module license taints kernel.\n", mod->name); - #ifdef CONFIG_MODVERSIONS if ((mod->num_syms && !mod->crcs) || (mod->num_gpl_syms && !mod->gpl_crcs)) { @@ -2772,11 +2773,6 @@ static int load_module(struct load_info *info, const char __user *uargs, /* * We are tainting your kernel if your module gets into * the modules linked list somehow. - * - * We have a descrepancy though, see the other taints for - * signature and those in check_module_license_and_versions(). - * - * We should compromise and converge. */ module_augment_kernel_taints(mod, info); -- cgit From 419e1a20f7bdef5380fde5ed73f05c98c28a598b Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:27:46 -0700 Subject: module: rename check_module_license_and_versions() to check_export_symbol_versions() This makes the routine easier to understand what the check its checking for. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index f165d93a4ef9..cf097ffe6a4a 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2220,7 +2220,7 @@ out_enomem: return -ENOMEM; } -static int check_module_license_and_versions(struct module *mod) +static int check_export_symbol_versions(struct module *mod) { #ifdef CONFIG_MODVERSIONS if ((mod->num_syms && !mod->crcs) || @@ -2796,7 +2796,7 @@ static int load_module(struct load_info *info, const char __user *uargs, if (err) goto free_unload; - err = check_module_license_and_versions(mod); + err = check_export_symbol_versions(mod); if (err) goto free_unload; -- cgit From 46752820f9abc013b6bd8172562b642376723313 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:35:38 -0700 Subject: module: add sanity check for ELF module section The ELF ".gnu.linkonce.this_module" section is special, it is what we use to construct the struct module __this_module, which THIS_MODULE points to. When userspace loads a module we always deal first with a copy of the userspace buffer, and twiddle with the userspace copy's version of the struct module. Eventually we allocate memory to do a memcpy() of that struct module, under the assumption that the module size is right. But we have no validity checks against the size or the requirements for the section. Add some validity checks for the special module section early and while at it, cache the module section index early, so we don't have to do that later. While at it, just move over the assigment of the info->mod to make the code clearer. The validity checker also adds an explicit size check to ensure the module section size matches the kernel's run time size for sizeof(struct module). This should prevent sloppy loads of modules which are built today *without* actually increasing the size of the struct module. A developer today can for example expand the size of struct module, rebuild a directoroy 'make fs/xfs/' for example and then try to insmode the driver there. That module would in effect have an incorrect size. This new size check would put a stop gap against such mistakes. This also makes the entire goal of ".gnu.linkonce.this_module" pretty clear. Before this patch verification of the goal / intent required some Indian Jones whips, torches and cleaning up big old spider webs. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 62 ++++++++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 53 insertions(+), 9 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index cf097ffe6a4a..e1a9dd51c036 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2,6 +2,7 @@ /* * Copyright (C) 2002 Richard Henderson * Copyright (C) 2001 Rusty Russell, 2002, 2010 Rusty Russell IBM. + * Copyright (C) 2023 Luis Chamberlain */ #define INCLUDE_VERMAGIC @@ -1656,6 +1657,7 @@ static int elf_validity_check(struct load_info *info) unsigned int i; Elf_Shdr *shdr, *strhdr; int err; + unsigned int num_mod_secs = 0, mod_idx; if (info->len < sizeof(*(info->hdr))) { pr_err("Invalid ELF header len %lu\n", info->len); @@ -1767,6 +1769,11 @@ static int elf_validity_check(struct load_info *info) i, shdr->sh_type); return err; } + if (strcmp(info->secstrings + shdr->sh_name, + ".gnu.linkonce.this_module") == 0) { + num_mod_secs++; + mod_idx = i; + } if (shdr->sh_flags & SHF_ALLOC) { if (shdr->sh_name >= strhdr->sh_size) { @@ -1779,6 +1786,52 @@ static int elf_validity_check(struct load_info *info) } } + /* + * The ".gnu.linkonce.this_module" ELF section is special. It is + * what modpost uses to refer to __this_module and let's use rely + * on THIS_MODULE to point to &__this_module properly. The kernel's + * modpost declares it on each modules's *.mod.c file. If the struct + * module of the kernel changes a full kernel rebuild is required. + * + * We have a few expectaions for this special section, the following + * code validates all this for us: + * + * o Only one section must exist + * o We expect the kernel to always have to allocate it: SHF_ALLOC + * o The section size must match the kernel's run time's struct module + * size + */ + if (num_mod_secs != 1) { + pr_err("Only one .gnu.linkonce.this_module section must exist.\n"); + goto no_exec; + } + + shdr = &info->sechdrs[mod_idx]; + + /* + * This is already implied on the switch above, however let's be + * pedantic about it. + */ + if (shdr->sh_type == SHT_NOBITS) { + pr_err(".gnu.linkonce.this_module section must have a size set\n"); + goto no_exec; + } + + if (!(shdr->sh_flags & SHF_ALLOC)) { + pr_err(".gnu.linkonce.this_module must occupy memory during process execution\n"); + goto no_exec; + } + + if (shdr->sh_size != sizeof(struct module)) { + pr_err(".gnu.linkonce.this_module section size must match the kernel's built struct module size at run time\n"); + goto no_exec; + } + + info->index.mod = mod_idx; + + /* This is temporary: point mod into copy of data. */ + info->mod = (void *)info->hdr + shdr->sh_offset; + return 0; no_exec: @@ -1925,15 +1978,6 @@ static int setup_load_info(struct load_info *info, int flags) return -ENOEXEC; } - info->index.mod = find_sec(info, ".gnu.linkonce.this_module"); - if (!info->index.mod) { - pr_warn("%s: No module found in object\n", - info->name ?: "(missing .modinfo section or name field)"); - return -ENOEXEC; - } - /* This is temporary: point mod into copy of data. */ - info->mod = (void *)info->hdr + info->sechdrs[info->index.mod].sh_offset; - /* * If we didn't load the .modinfo 'name' field earlier, fall back to * on-disk struct mod 'name' field. -- cgit From c7ee8aebf6c0588c0aab76538aff395c3abf811c Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:35:39 -0700 Subject: module: add stop-grap sanity check on module memcpy() The integrity of the struct module we load is important, and although our ELF validator already checks that the module section must match struct module, add a stop-gap check before we memcpy() the final minted module. This also makes those inspecting the code what the goal is. While at it, clarify the goal behind updating the sh_addr address. The current comment is pretty misleading. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 26 ++++++++++++++++++++++---- 1 file changed, 22 insertions(+), 4 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index e1a9dd51c036..fbe62d1625bc 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2213,7 +2213,8 @@ static int move_module(struct module *mod, struct load_info *info) { int i; void *ptr; - enum mod_mem_type t; + enum mod_mem_type t = 0; + int ret = -ENOMEM; for_each_mod_mem_type(type) { if (!mod->mem[type].size) { @@ -2249,9 +2250,26 @@ static int move_module(struct module *mod, struct load_info *info) dest = mod->mem[type].base + (shdr->sh_entsize & SH_ENTSIZE_OFFSET_MASK); - if (shdr->sh_type != SHT_NOBITS) + if (shdr->sh_type != SHT_NOBITS) { + /* + * Our ELF checker already validated this, but let's + * be pedantic and make the goal clearer. We actually + * end up copying over all modifications made to the + * userspace copy of the entire struct module. + */ + if (i == info->index.mod && + (WARN_ON_ONCE(shdr->sh_size != sizeof(struct module)))) { + ret = -ENOEXEC; + goto out_enomem; + } memcpy(dest, (void *)shdr->sh_addr, shdr->sh_size); - /* Update sh_addr to point to copy in image. */ + } + /* + * Update the userspace copy's ELF section address to point to + * our newly allocated memory as a pure convenience so that + * users of info can keep taking advantage and using the newly + * minted official memory area. + */ shdr->sh_addr = (unsigned long)dest; pr_debug("\t0x%lx %s\n", (long)shdr->sh_addr, info->secstrings + shdr->sh_name); @@ -2261,7 +2279,7 @@ static int move_module(struct module *mod, struct load_info *info) out_enomem: for (t--; t >= 0; t--) module_memory_free(mod->mem[t].base, t); - return -ENOMEM; + return ret; } static int check_export_symbol_versions(struct module *mod) -- cgit From 1bb49db9919a4d4186cba288930e7026d8f7ec96 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:35:40 -0700 Subject: module: move more elf validity checks to elf_validity_check() The symbol and strings section validation currently happen in setup_load_info() but since they are also doing validity checks move this to elf_validity_check(). Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 79 +++++++++++++++++++++++++++++----------------------- 1 file changed, 44 insertions(+), 35 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index fbe62d1625bc..84a7f96cf35a 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1658,6 +1658,8 @@ static int elf_validity_check(struct load_info *info) Elf_Shdr *shdr, *strhdr; int err; unsigned int num_mod_secs = 0, mod_idx; + unsigned int num_info_secs = 0, info_idx; + unsigned int num_sym_secs = 0, sym_idx; if (info->len < sizeof(*(info->hdr))) { pr_err("Invalid ELF header len %lu\n", info->len); @@ -1761,6 +1763,8 @@ static int elf_validity_check(struct load_info *info) info->hdr->e_shnum); goto no_exec; } + num_sym_secs++; + sym_idx = i; fallthrough; default: err = validate_section_offset(info, shdr); @@ -1773,6 +1777,10 @@ static int elf_validity_check(struct load_info *info) ".gnu.linkonce.this_module") == 0) { num_mod_secs++; mod_idx = i; + } else if (strcmp(info->secstrings + shdr->sh_name, + ".modinfo") == 0) { + num_info_secs++; + info_idx = i; } if (shdr->sh_flags & SHF_ALLOC) { @@ -1786,6 +1794,27 @@ static int elf_validity_check(struct load_info *info) } } + if (num_info_secs > 1) { + pr_err("Only one .modinfo section must exist.\n"); + goto no_exec; + } else if (num_info_secs == 1) { + /* Try to find a name early so we can log errors with a module name */ + info->index.info = info_idx; + info->name = get_modinfo(info, "name"); + } + + if (num_sym_secs != 1) { + pr_warn("%s: module has no symbols (stripped?)\n", + info->name ?: "(missing .modinfo section or name field)"); + goto no_exec; + } + + /* Sets internal symbols and strings. */ + info->index.sym = sym_idx; + shdr = &info->sechdrs[sym_idx]; + info->index.str = shdr->sh_link; + info->strtab = (char *)info->hdr + info->sechdrs[info->index.str].sh_offset; + /* * The ".gnu.linkonce.this_module" ELF section is special. It is * what modpost uses to refer to __this_module and let's use rely @@ -1802,7 +1831,8 @@ static int elf_validity_check(struct load_info *info) * size */ if (num_mod_secs != 1) { - pr_err("Only one .gnu.linkonce.this_module section must exist.\n"); + pr_err("module %s: Only one .gnu.linkonce.this_module section must exist.\n", + info->name ?: "(missing .modinfo section or name field)"); goto no_exec; } @@ -1813,17 +1843,20 @@ static int elf_validity_check(struct load_info *info) * pedantic about it. */ if (shdr->sh_type == SHT_NOBITS) { - pr_err(".gnu.linkonce.this_module section must have a size set\n"); + pr_err("module %s: .gnu.linkonce.this_module section must have a size set\n", + info->name ?: "(missing .modinfo section or name field)"); goto no_exec; } if (!(shdr->sh_flags & SHF_ALLOC)) { - pr_err(".gnu.linkonce.this_module must occupy memory during process execution\n"); + pr_err("module %s: .gnu.linkonce.this_module must occupy memory during process execution\n", + info->name ?: "(missing .modinfo section or name field)"); goto no_exec; } if (shdr->sh_size != sizeof(struct module)) { - pr_err(".gnu.linkonce.this_module section size must match the kernel's built struct module size at run time\n"); + pr_err("module %s: .gnu.linkonce.this_module section size must match the kernel's built struct module size at run time\n", + info->name ?: "(missing .modinfo section or name field)"); goto no_exec; } @@ -1832,6 +1865,13 @@ static int elf_validity_check(struct load_info *info) /* This is temporary: point mod into copy of data. */ info->mod = (void *)info->hdr + shdr->sh_offset; + /* + * If we didn't load the .modinfo 'name' field earlier, fall back to + * on-disk struct mod 'name' field. + */ + if (!info->name) + info->name = info->mod->name; + return 0; no_exec: @@ -1954,37 +1994,6 @@ static int rewrite_section_headers(struct load_info *info, int flags) */ static int setup_load_info(struct load_info *info, int flags) { - unsigned int i; - - /* Try to find a name early so we can log errors with a module name */ - info->index.info = find_sec(info, ".modinfo"); - if (info->index.info) - info->name = get_modinfo(info, "name"); - - /* Find internal symbols and strings. */ - for (i = 1; i < info->hdr->e_shnum; i++) { - if (info->sechdrs[i].sh_type == SHT_SYMTAB) { - info->index.sym = i; - info->index.str = info->sechdrs[i].sh_link; - info->strtab = (char *)info->hdr - + info->sechdrs[info->index.str].sh_offset; - break; - } - } - - if (info->index.sym == 0) { - pr_warn("%s: module has no symbols (stripped?)\n", - info->name ?: "(missing .modinfo section or name field)"); - return -ENOEXEC; - } - - /* - * If we didn't load the .modinfo 'name' field earlier, fall back to - * on-disk struct mod 'name' field. - */ - if (!info->name) - info->name = info->mod->name; - if (flags & MODULE_INIT_IGNORE_MODVERSIONS) info->index.vers = 0; /* Pretend no __versions section! */ else -- cgit From 3d40bb903ed1f654707d34bdd61ee2c332000e4b Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:35:41 -0700 Subject: module: merge remnants of setup_load_info() to elf validation The setup_load_info() was actually had ELF validation checks of its own. To later cache useful variables as an secondary step just means looping again over the ELF sections we just validated. We can simply keep tabs of the key sections of interest as we validate the module ELF section in one swoop, so do that and merge the two routines together. Expand a bit on the documentation / intent / goals. Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 60 +++++++++++++++++++++++----------------------------- 1 file changed, 27 insertions(+), 33 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 84a7f96cf35a..929644d79d38 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1647,12 +1647,26 @@ static int validate_section_offset(struct load_info *info, Elf_Shdr *shdr) } /* - * Sanity checks against invalid binaries, wrong arch, weird elf version. + * Check userspace passed ELF module against our expectations, and cache + * useful variables for further processing as we go. * - * Also do basic validity checks against section offsets and sizes, the + * This does basic validity checks against section offsets and sizes, the * section name string table, and the indices used for it (sh_name). + * + * As a last step, since we're already checking the ELF sections we cache + * useful variables which will be used later for our convenience: + * + * o pointers to section headers + * o cache the modinfo symbol section + * o cache the string symbol section + * o cache the module section + * + * As a last step we set info->mod to the temporary copy of the module in + * info->hdr. The final one will be allocated in move_module(). Any + * modifications we make to our copy of the module will be carried over + * to the final minted module. */ -static int elf_validity_check(struct load_info *info) +static int elf_validity_cache_copy(struct load_info *info, int flags) { unsigned int i; Elf_Shdr *shdr, *strhdr; @@ -1872,6 +1886,13 @@ static int elf_validity_check(struct load_info *info) if (!info->name) info->name = info->mod->name; + if (flags & MODULE_INIT_IGNORE_MODVERSIONS) + info->index.vers = 0; /* Pretend no __versions section! */ + else + info->index.vers = find_sec(info, "__versions"); + + info->index.pcpu = find_pcpusec(info); + return 0; no_exec: @@ -1984,26 +2005,6 @@ static int rewrite_section_headers(struct load_info *info, int flags) return 0; } -/* - * Set up our basic convenience variables (pointers to section headers, - * search for module section index etc), and do some basic section - * verification. - * - * Set info->mod to the temporary copy of the module in info->hdr. The final one - * will be allocated in move_module(). - */ -static int setup_load_info(struct load_info *info, int flags) -{ - if (flags & MODULE_INIT_IGNORE_MODVERSIONS) - info->index.vers = 0; /* Pretend no __versions section! */ - else - info->index.vers = find_sec(info, "__versions"); - - info->index.pcpu = find_pcpusec(info); - - return 0; -} - /* * These calls taint the kernel depending certain module circumstances */ static void module_augment_kernel_taints(struct module *mod, struct load_info *info) @@ -2809,17 +2810,10 @@ static int load_module(struct load_info *info, const char __user *uargs, /* * Do basic sanity checks against the ELF header and - * sections. - */ - err = elf_validity_check(info); - if (err) - goto free_copy; - - /* - * Everything checks out, so set up the section info - * in the info structure. + * sections. Cache useful sections and set the + * info->mod to the userspace passed struct module. */ - err = setup_load_info(info, flags); + err = elf_validity_cache_copy(info, flags); if (err) goto free_copy; -- cgit From 25be451aa4c0e9a96c59a626ab0e93d5cb7f6f48 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Sun, 19 Mar 2023 14:35:42 -0700 Subject: module: fold usermode helper kmod into modules directory The kernel/kmod.c is already only built if we enabled modules, so just stuff it under kernel/module/kmod.c and unify the MAINTAINERS file for it. Signed-off-by: Luis Chamberlain --- kernel/module/Makefile | 4 +- kernel/module/kmod.c | 177 +++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 180 insertions(+), 1 deletion(-) create mode 100644 kernel/module/kmod.c (limited to 'kernel/module') diff --git a/kernel/module/Makefile b/kernel/module/Makefile index 948efea81e85..5b1d26b53b8d 100644 --- a/kernel/module/Makefile +++ b/kernel/module/Makefile @@ -7,7 +7,9 @@ # and produce insane amounts of uninteresting coverage. KCOV_INSTRUMENT_module.o := n -obj-y += main.o strict_rwx.o +obj-y += main.o +obj-y += strict_rwx.o +obj-y += kmod.o obj-$(CONFIG_MODULE_DECOMPRESS) += decompress.o obj-$(CONFIG_MODULE_SIG) += signing.o obj-$(CONFIG_LIVEPATCH) += livepatch.o diff --git a/kernel/module/kmod.c b/kernel/module/kmod.c new file mode 100644 index 000000000000..b717134ebe17 --- /dev/null +++ b/kernel/module/kmod.c @@ -0,0 +1,177 @@ +/* + * kmod - the kernel module loader + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +/* + * Assuming: + * + * threads = div64_u64((u64) totalram_pages * (u64) PAGE_SIZE, + * (u64) THREAD_SIZE * 8UL); + * + * If you need less than 50 threads would mean we're dealing with systems + * smaller than 3200 pages. This assumes you are capable of having ~13M memory, + * and this would only be an upper limit, after which the OOM killer would take + * effect. Systems like these are very unlikely if modules are enabled. + */ +#define MAX_KMOD_CONCURRENT 50 +static atomic_t kmod_concurrent_max = ATOMIC_INIT(MAX_KMOD_CONCURRENT); +static DECLARE_WAIT_QUEUE_HEAD(kmod_wq); + +/* + * This is a restriction on having *all* MAX_KMOD_CONCURRENT threads + * running at the same time without returning. When this happens we + * believe you've somehow ended up with a recursive module dependency + * creating a loop. + * + * We have no option but to fail. + * + * Userspace should proactively try to detect and prevent these. + */ +#define MAX_KMOD_ALL_BUSY_TIMEOUT 5 + +/* + modprobe_path is set via /proc/sys. +*/ +char modprobe_path[KMOD_PATH_LEN] = CONFIG_MODPROBE_PATH; + +static void free_modprobe_argv(struct subprocess_info *info) +{ + kfree(info->argv[3]); /* check call_modprobe() */ + kfree(info->argv); +} + +static int call_modprobe(char *module_name, int wait) +{ + struct subprocess_info *info; + static char *envp[] = { + "HOME=/", + "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", + NULL + }; + + char **argv = kmalloc(sizeof(char *[5]), GFP_KERNEL); + if (!argv) + goto out; + + module_name = kstrdup(module_name, GFP_KERNEL); + if (!module_name) + goto free_argv; + + argv[0] = modprobe_path; + argv[1] = "-q"; + argv[2] = "--"; + argv[3] = module_name; /* check free_modprobe_argv() */ + argv[4] = NULL; + + info = call_usermodehelper_setup(modprobe_path, argv, envp, GFP_KERNEL, + NULL, free_modprobe_argv, NULL); + if (!info) + goto free_module_name; + + return call_usermodehelper_exec(info, wait | UMH_KILLABLE); + +free_module_name: + kfree(module_name); +free_argv: + kfree(argv); +out: + return -ENOMEM; +} + +/** + * __request_module - try to load a kernel module + * @wait: wait (or not) for the operation to complete + * @fmt: printf style format string for the name of the module + * @...: arguments as specified in the format string + * + * Load a module using the user mode module loader. The function returns + * zero on success or a negative errno code or positive exit code from + * "modprobe" on failure. Note that a successful module load does not mean + * the module did not then unload and exit on an error of its own. Callers + * must check that the service they requested is now available not blindly + * invoke it. + * + * If module auto-loading support is disabled then this function + * simply returns -ENOENT. + */ +int __request_module(bool wait, const char *fmt, ...) +{ + va_list args; + char module_name[MODULE_NAME_LEN]; + int ret; + + /* + * We don't allow synchronous module loading from async. Module + * init may invoke async_synchronize_full() which will end up + * waiting for this task which already is waiting for the module + * loading to complete, leading to a deadlock. + */ + WARN_ON_ONCE(wait && current_is_async()); + + if (!modprobe_path[0]) + return -ENOENT; + + va_start(args, fmt); + ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args); + va_end(args); + if (ret >= MODULE_NAME_LEN) + return -ENAMETOOLONG; + + ret = security_kernel_module_request(module_name); + if (ret) + return ret; + + if (atomic_dec_if_positive(&kmod_concurrent_max) < 0) { + pr_warn_ratelimited("request_module: kmod_concurrent_max (%u) close to 0 (max_modprobes: %u), for module %s, throttling...", + atomic_read(&kmod_concurrent_max), + MAX_KMOD_CONCURRENT, module_name); + ret = wait_event_killable_timeout(kmod_wq, + atomic_dec_if_positive(&kmod_concurrent_max) >= 0, + MAX_KMOD_ALL_BUSY_TIMEOUT * HZ); + if (!ret) { + pr_warn_ratelimited("request_module: modprobe %s cannot be processed, kmod busy with %d threads for more than %d seconds now", + module_name, MAX_KMOD_CONCURRENT, MAX_KMOD_ALL_BUSY_TIMEOUT); + return -ETIME; + } else if (ret == -ERESTARTSYS) { + pr_warn_ratelimited("request_module: sigkill sent for modprobe %s, giving up", module_name); + return ret; + } + } + + trace_module_request(module_name, wait, _RET_IP_); + + ret = call_modprobe(module_name, wait ? UMH_WAIT_PROC : UMH_WAIT_EXEC); + + atomic_inc(&kmod_concurrent_max); + wake_up(&kmod_wq); + + return ret; +} +EXPORT_SYMBOL(__request_module); -- cgit From 6ed81802d4d1b037ad2d1657511ff0c2e9aeda14 Mon Sep 17 00:00:00 2001 From: Jim Cromie Date: Tue, 21 Mar 2023 19:36:20 -0600 Subject: module: in layout_sections, move_module: add the modname layout_sections() and move_module() each issue ~50 messages for each module loaded. Add mod-name into their 2 header lines, to help the reader find his module. no functional changes. Signed-off-by: Jim Cromie Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 929644d79d38..3bea679837e0 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1567,10 +1567,10 @@ static void layout_sections(struct module *mod, struct load_info *info) for (i = 0; i < info->hdr->e_shnum; i++) info->sechdrs[i].sh_entsize = ~0UL; - pr_debug("Core section allocation order:\n"); + pr_debug("Core section allocation order for %s:\n", mod->name); __layout_sections(mod, info, false); - pr_debug("Init section allocation order:\n"); + pr_debug("Init section allocation order for %s:\n", mod->name); __layout_sections(mod, info, true); } @@ -2249,7 +2249,7 @@ static int move_module(struct module *mod, struct load_info *info) } /* Transfer each section which specifies SHF_ALLOC */ - pr_debug("final section addresses:\n"); + pr_debug("Final section addresses for %s:\n", mod->name); for (i = 0; i < info->hdr->e_shnum; i++) { void *dest; Elf_Shdr *shdr = &info->sechdrs[i]; -- cgit From b10addf37bbcaee66672eb54c15532266c8daea6 Mon Sep 17 00:00:00 2001 From: Jim Cromie Date: Tue, 21 Mar 2023 19:36:21 -0600 Subject: module: add symbol-name to pr_debug Absolute symbol The pr_debug("Absolute symbol" ..) reports value, (which is usually 0), but not the name, which is more informative. So add it. no functional changes Signed-off-by: Jim Cromie Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 3bea679837e0..6567d5a156ae 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -1394,8 +1394,8 @@ static int simplify_symbols(struct module *mod, const struct load_info *info) case SHN_ABS: /* Don't need to do anything */ - pr_debug("Absolute symbol: 0x%08lx\n", - (long)sym[i].st_value); + pr_debug("Absolute symbol: 0x%08lx %s\n", + (long)sym[i].st_value, name); break; case SHN_LIVEPATCH: -- cgit From 66a2301edf313d630c2ece4f3721c5b3402653ee Mon Sep 17 00:00:00 2001 From: Jim Cromie Date: Tue, 21 Mar 2023 19:36:22 -0600 Subject: module: add section-size to move_module pr_debug move_module() pr_debug's "Final section addresses for $modname". Add section addresses to the message, for anyone looking at these. no functional changes. Signed-off-by: Jim Cromie Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 6567d5a156ae..1d587e5e9cc6 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2281,8 +2281,8 @@ static int move_module(struct module *mod, struct load_info *info) * minted official memory area. */ shdr->sh_addr = (unsigned long)dest; - pr_debug("\t0x%lx %s\n", - (long)shdr->sh_addr, info->secstrings + shdr->sh_name); + pr_debug("\t0x%lx 0x%.8lx %s\n", (long)shdr->sh_addr, + (long)shdr->sh_size, info->secstrings + shdr->sh_name); } return 0; -- cgit From 33c951f62920d144ca89daa0560180a49afb6f1e Mon Sep 17 00:00:00 2001 From: Jim Cromie Date: Tue, 21 Mar 2023 19:36:23 -0600 Subject: module: already_uses() - reduce pr_debug output volume already_uses() is unnecessarily chatty. `modprobe i915` yields 491 messages like: [ 64.108744] i915 uses drm! This is a normal situation, and isn't worth all the log entries. NOTE: I've preserved the "does not use %s" messages, which happens less often, but does happen. Its not clear to me what it tells a reader, or what info might improve the pr_debug's utility. [ 6847.584999] main:already_uses:569: amdgpu does not use ttm! [ 6847.585001] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585014] main:already_uses:569: amdgpu does not use drm! [ 6847.585016] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585024] main:already_uses:569: amdgpu does not use drm_display_helper! [ 6847.585025] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585084] main:already_uses:569: amdgpu does not use drm_kms_helper! [ 6847.585086] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585175] main:already_uses:569: amdgpu does not use drm_buddy! [ 6847.585176] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585202] main:already_uses:569: amdgpu does not use i2c_algo_bit! [ 6847.585204] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585249] main:already_uses:569: amdgpu does not use gpu_sched! [ 6847.585250] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585314] main:already_uses:569: amdgpu does not use video! [ 6847.585315] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585409] main:already_uses:569: amdgpu does not use iommu_v2! [ 6847.585410] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6847.585816] main:already_uses:569: amdgpu does not use drm_ttm_helper! [ 6847.585818] main:add_module_usage:584: Allocating new usage for amdgpu. [ 6848.762268] dyndbg: add-module: amdgpu.2533 sites no functional changes. Signed-off-by: Jim Cromie Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 1d587e5e9cc6..5cc21083af04 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -564,10 +564,8 @@ static int already_uses(struct module *a, struct module *b) struct module_use *use; list_for_each_entry(use, &b->source_list, source_list) { - if (use->source == a) { - pr_debug("%s uses %s!\n", a->name, b->name); + if (use->source == a) return 1; - } } pr_debug("%s does not use %s!\n", a->name, b->name); return 0; -- cgit From 87e5b1e8f257023ac5c4d2b8f07716a7f3dcc8ea Mon Sep 17 00:00:00 2001 From: Tiezhu Yang Date: Fri, 31 Mar 2023 17:15:51 +0800 Subject: module: Sync code of is_arm_mapping_symbol() After commit 2e3a10a1551d ("ARM: avoid ARM binutils leaking ELF local symbols") and commit d6b732666a1b ("modpost: fix undefined behavior of is_arm_mapping_symbol()"), many differences of is_arm_mapping_symbol() exist in kernel/module/kallsyms.c and scripts/mod/modpost.c, just sync the code to keep consistent. Signed-off-by: Tiezhu Yang Signed-off-by: Luis Chamberlain --- kernel/module/kallsyms.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c index 3320586584f1..a9045fe55bcd 100644 --- a/kernel/module/kallsyms.c +++ b/kernel/module/kallsyms.c @@ -251,8 +251,9 @@ static inline int is_arm_mapping_symbol(const char *str) { if (str[0] == '.' && str[1] == 'L') return true; - return str[0] == '$' && strchr("axtd", str[1]) && - (str[2] == '\0' || str[2] == '.'); + return str[0] == '$' && + (str[1] == 'a' || str[1] == 'd' || str[1] == 't' || str[1] == 'x') + && (str[2] == '\0' || str[2] == '.'); } static const char *kallsyms_symbol_name(struct mod_kallsyms *kallsyms, unsigned int symnum) -- cgit From 987d2e0aaa55de40938435be760aa96428470fd6 Mon Sep 17 00:00:00 2001 From: Tiezhu Yang Date: Fri, 31 Mar 2023 17:15:52 +0800 Subject: module: Move is_arm_mapping_symbol() to module_symbol.h In order to avoid duplicated code, move is_arm_mapping_symbol() to include/linux/module_symbol.h, then remove is_arm_mapping_symbol() in the other places. Signed-off-by: Tiezhu Yang Signed-off-by: Luis Chamberlain --- kernel/module/kallsyms.c | 14 +------------- 1 file changed, 1 insertion(+), 13 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c index a9045fe55bcd..5de3207f18af 100644 --- a/kernel/module/kallsyms.c +++ b/kernel/module/kallsyms.c @@ -6,6 +6,7 @@ */ #include +#include #include #include #include @@ -243,19 +244,6 @@ void init_build_id(struct module *mod, const struct load_info *info) } #endif -/* - * This ignores the intensely annoying "mapping symbols" found - * in ARM ELF files: $a, $t and $d. - */ -static inline int is_arm_mapping_symbol(const char *str) -{ - if (str[0] == '.' && str[1] == 'L') - return true; - return str[0] == '$' && - (str[1] == 'a' || str[1] == 'd' || str[1] == 't' || str[1] == 'x') - && (str[2] == '\0' || str[2] == '.'); -} - static const char *kallsyms_symbol_name(struct mod_kallsyms *kallsyms, unsigned int symnum) { return kallsyms->strtab + kallsyms->symtab[symnum].st_name; -- cgit From 0a3bf86092c38f7b72c56c6901c78dd302411307 Mon Sep 17 00:00:00 2001 From: Tiezhu Yang Date: Fri, 31 Mar 2023 17:15:53 +0800 Subject: module: Ignore L0 and rename is_arm_mapping_symbol() The L0 symbol is generated when build module on LoongArch, ignore it in modpost and when looking at module symbols, otherwise we can not see the expected call trace. Now is_arm_mapping_symbol() is not only for ARM, in order to reflect the reality, rename is_arm_mapping_symbol() to is_mapping_symbol(). This is related with commit c17a2538704f ("mksysmap: Fix the mismatch of 'L0' symbols in System.map"). (1) Simple test case [loongson@linux hello]$ cat hello.c #include #include #include static void test_func(void) { pr_info("This is a test\n"); dump_stack(); } static int __init hello_init(void) { pr_warn("Hello, world\n"); test_func(); return 0; } static void __exit hello_exit(void) { pr_warn("Goodbye\n"); } module_init(hello_init); module_exit(hello_exit); MODULE_LICENSE("GPL"); [loongson@linux hello]$ cat Makefile obj-m:=hello.o ccflags-y += -g -Og all: make -C /lib/modules/$(shell uname -r)/build/ M=$(PWD) modules clean: make -C /lib/modules/$(shell uname -r)/build/ M=$(PWD) clean (2) Test environment system: LoongArch CLFS 5.5 https://github.com/sunhaiyong1978/CLFS-for-LoongArch/releases/tag/5.0 It needs to update grub to avoid booting error "invalid magic number". kernel: 6.3-rc1 with loongson3_defconfig + CONFIG_DYNAMIC_FTRACE=y (3) Test result Without this patch: [root@linux hello]# insmod hello.ko [root@linux hello]# dmesg ... Hello, world This is a test ... Call Trace: [<9000000000223728>] show_stack+0x68/0x18c [<90000000013374cc>] dump_stack_lvl+0x60/0x88 [] L0\x01+0x20/0x2c [hello] [] L0\x01+0x20/0x30 [hello] [<900000000022097c>] do_one_initcall+0x88/0x288 [<90000000002df890>] do_init_module+0x54/0x200 [<90000000002e1e18>] __do_sys_finit_module+0xc4/0x114 [<90000000013382e8>] do_syscall+0x7c/0x94 [<9000000000221e3c>] handle_syscall+0xbc/0x158 With this patch: [root@linux hello]# insmod hello.ko [root@linux hello]# dmesg ... Hello, world This is a test ... Call Trace: [<9000000000223728>] show_stack+0x68/0x18c [<90000000013374cc>] dump_stack_lvl+0x60/0x88 [] test_func+0x28/0x34 [hello] [] hello_init+0x28/0x38 [hello] [<900000000022097c>] do_one_initcall+0x88/0x288 [<90000000002df890>] do_init_module+0x54/0x200 [<90000000002e1e18>] __do_sys_finit_module+0xc4/0x114 [<90000000013382e8>] do_syscall+0x7c/0x94 [<9000000000221e3c>] handle_syscall+0xbc/0x158 Signed-off-by: Tiezhu Yang Tested-by: Youling Tang # for LoongArch Signed-off-by: Luis Chamberlain --- kernel/module/kallsyms.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/module') diff --git a/kernel/module/kallsyms.c b/kernel/module/kallsyms.c index 5de3207f18af..d8e426a9a0cd 100644 --- a/kernel/module/kallsyms.c +++ b/kernel/module/kallsyms.c @@ -289,7 +289,7 @@ static const char *find_kallsyms_symbol(struct module *mod, * and inserted at a whim. */ if (*kallsyms_symbol_name(kallsyms, i) == '\0' || - is_arm_mapping_symbol(kallsyms_symbol_name(kallsyms, i))) + is_mapping_symbol(kallsyms_symbol_name(kallsyms, i))) continue; if (thisval <= addr && thisval > bestval) { -- cgit From 430bb0d1c3376c988982f14bcbe71f917c89e1ab Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Tue, 4 Apr 2023 18:52:47 -0700 Subject: module: fix kmemleak annotations for non init ELF sections Commit ac3b43283923 ("module: replace module_layout with module_memory") reworked the way to handle memory allocations to make it clearer. But it lost in translation how we handled kmemleak_ignore() or kmemleak_not_leak() for different ELF sections. Fix this and clarify the comments a bit more. Contrary to the old way of using kmemleak_ignore() for init.* ELF sections we stick now only to kmemleak_not_leak() as per suggestion by Catalin Marinas so to avoid any false positives and simplify the code. Fixes: ac3b43283923 ("module: replace module_layout with module_memory") Reported-by: Jim Cromie Acked-by: Song Liu Suggested-by: Catalin Marinas Reviewed-by: Catalin Marinas Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 5cc21083af04..32554d8a5791 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2231,13 +2231,18 @@ static int move_module(struct module *mod, struct load_info *info) } mod->mem[type].size = PAGE_ALIGN(mod->mem[type].size); ptr = module_memory_alloc(mod->mem[type].size, type); - /* - * The pointer to this block is stored in the module structure - * which is inside the block. Just mark it as not being a - * leak. + * The pointer to these blocks of memory are stored on the module + * structure and we keep that around so long as the module is + * around. We only free that memory when we unload the module. + * Just mark them as not being a leak then. The .init* ELF + * sections *do* get freed after boot so we *could* treat them + * slightly differently with kmemleak_ignore() and only grey + * them out as they work as typical memory allocations which + * *do* eventually get freed, but let's just keep things simple + * and avoid *any* false positives. */ - kmemleak_ignore(ptr); + kmemleak_not_leak(ptr); if (!ptr) { t = type; goto out_enomem; -- cgit From 25a1b5b518f4336bff934ac8348da6c57158363a Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Fri, 24 Mar 2023 18:38:00 -0700 Subject: modules/kmod: replace implementation with a semaphore Simplify the concurrency delimiter we use for kmod with the semaphore. I had used the kmod strategy to try to implement a similar concurrency delimiter for the kernel_read*() calls from the finit_module() path so to reduce vmalloc() memory pressure. That effort didn't provide yet conclusive results, but one thing that became clear is we can use the suggested alternative solution with semaphores which Linus hinted at instead of using the atomic / wait strategy. I've stress tested this with kmod test 0008: time /data/linux-next/tools/testing/selftests/kmod/kmod.sh -t 0008 And I get only a *slight* delay. That delay however is small, a few seconds for a full test loop run that runs 150 times, for about ~30-40 seconds. The small delay is worth the simplfication IMHO. Reviewed-by: Davidlohr Bueso Reviewed-by: Miroslav Benes Reviewed-by: David Hildenbrand Signed-off-by: Luis Chamberlain --- kernel/module/kmod.c | 26 +++++++------------------- 1 file changed, 7 insertions(+), 19 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/kmod.c b/kernel/module/kmod.c index b717134ebe17..5899083436a3 100644 --- a/kernel/module/kmod.c +++ b/kernel/module/kmod.c @@ -40,8 +40,7 @@ * effect. Systems like these are very unlikely if modules are enabled. */ #define MAX_KMOD_CONCURRENT 50 -static atomic_t kmod_concurrent_max = ATOMIC_INIT(MAX_KMOD_CONCURRENT); -static DECLARE_WAIT_QUEUE_HEAD(kmod_wq); +static DEFINE_SEMAPHORE(kmod_concurrent_max, MAX_KMOD_CONCURRENT); /* * This is a restriction on having *all* MAX_KMOD_CONCURRENT threads @@ -148,29 +147,18 @@ int __request_module(bool wait, const char *fmt, ...) if (ret) return ret; - if (atomic_dec_if_positive(&kmod_concurrent_max) < 0) { - pr_warn_ratelimited("request_module: kmod_concurrent_max (%u) close to 0 (max_modprobes: %u), for module %s, throttling...", - atomic_read(&kmod_concurrent_max), - MAX_KMOD_CONCURRENT, module_name); - ret = wait_event_killable_timeout(kmod_wq, - atomic_dec_if_positive(&kmod_concurrent_max) >= 0, - MAX_KMOD_ALL_BUSY_TIMEOUT * HZ); - if (!ret) { - pr_warn_ratelimited("request_module: modprobe %s cannot be processed, kmod busy with %d threads for more than %d seconds now", - module_name, MAX_KMOD_CONCURRENT, MAX_KMOD_ALL_BUSY_TIMEOUT); - return -ETIME; - } else if (ret == -ERESTARTSYS) { - pr_warn_ratelimited("request_module: sigkill sent for modprobe %s, giving up", module_name); - return ret; - } + ret = down_timeout(&kmod_concurrent_max, MAX_KMOD_ALL_BUSY_TIMEOUT * HZ); + if (ret) { + pr_warn_ratelimited("request_module: modprobe %s cannot be processed, kmod busy with %d threads for more than %d seconds now", + module_name, MAX_KMOD_CONCURRENT, MAX_KMOD_ALL_BUSY_TIMEOUT); + return ret; } trace_module_request(module_name, wait, _RET_IP_); ret = call_modprobe(module_name, wait ? UMH_WAIT_PROC : UMH_WAIT_EXEC); - atomic_inc(&kmod_concurrent_max); - wake_up(&kmod_wq); + up(&kmod_concurrent_max); return ret; } -- cgit From f71afa6a420111da90657fe999a8e32c42d5c7d6 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Fri, 10 Mar 2023 20:05:52 -0800 Subject: module: extract patient module check into helper The patient module check inside add_unformed_module() is large enough as we need it. It is a bit hard to read too, so just move it to a helper and do the inverse checks first to help shift the code and make it easier to read. The new helper then is module_patient_check_exists(). To make this work we need to mvoe the finished_loading() up, we do that without making any functional changes to that routine. Reviewed-by: David Hildenbrand Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 112 +++++++++++++++++++++++++++------------------------ 1 file changed, 60 insertions(+), 52 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 32554d8a5791..75b23257128d 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2447,27 +2447,6 @@ static int post_relocation(struct module *mod, const struct load_info *info) return module_finalize(info->hdr, info->sechdrs, mod); } -/* Is this module of this name done loading? No locks held. */ -static bool finished_loading(const char *name) -{ - struct module *mod; - bool ret; - - /* - * The module_mutex should not be a heavily contended lock; - * if we get the occasional sleep here, we'll go an extra iteration - * in the wait_event_interruptible(), which is harmless. - */ - sched_annotate_sleep(); - mutex_lock(&module_mutex); - mod = find_module_all(name, strlen(name), true); - ret = !mod || mod->state == MODULE_STATE_LIVE - || mod->state == MODULE_STATE_GOING; - mutex_unlock(&module_mutex); - - return ret; -} - /* Call module constructors. */ static void do_mod_ctors(struct module *mod) { @@ -2631,6 +2610,63 @@ static int may_init_module(void) return 0; } +/* Is this module of this name done loading? No locks held. */ +static bool finished_loading(const char *name) +{ + struct module *mod; + bool ret; + + /* + * The module_mutex should not be a heavily contended lock; + * if we get the occasional sleep here, we'll go an extra iteration + * in the wait_event_interruptible(), which is harmless. + */ + sched_annotate_sleep(); + mutex_lock(&module_mutex); + mod = find_module_all(name, strlen(name), true); + ret = !mod || mod->state == MODULE_STATE_LIVE + || mod->state == MODULE_STATE_GOING; + mutex_unlock(&module_mutex); + + return ret; +} + +/* Must be called with module_mutex held */ +static int module_patient_check_exists(const char *name) +{ + struct module *old; + int err = 0; + + old = find_module_all(name, strlen(name), true); + if (old == NULL) + return 0; + + if (old->state == MODULE_STATE_COMING || + old->state == MODULE_STATE_UNFORMED) { + /* Wait in case it fails to load. */ + mutex_unlock(&module_mutex); + err = wait_event_interruptible(module_wq, + finished_loading(name)); + mutex_lock(&module_mutex); + if (err) + return err; + + /* The module might have gone in the meantime. */ + old = find_module_all(name, strlen(name), true); + } + + /* + * We are here only when the same module was being loaded. Do + * not try to load it again right now. It prevents long delays + * caused by serialized module load failures. It might happen + * when more devices of the same type trigger load of + * a particular module. + */ + if (old && old->state == MODULE_STATE_LIVE) + return -EEXIST; + return -EBUSY; +} + /* * We try to place it in the list now to make sure it's unique before * we dedicate too many resources. In particular, temporary percpu @@ -2639,41 +2675,14 @@ static int may_init_module(void) static int add_unformed_module(struct module *mod) { int err; - struct module *old; mod->state = MODULE_STATE_UNFORMED; mutex_lock(&module_mutex); - old = find_module_all(mod->name, strlen(mod->name), true); - if (old != NULL) { - if (old->state == MODULE_STATE_COMING - || old->state == MODULE_STATE_UNFORMED) { - /* Wait in case it fails to load. */ - mutex_unlock(&module_mutex); - err = wait_event_interruptible(module_wq, - finished_loading(mod->name)); - if (err) - goto out_unlocked; - - /* The module might have gone in the meantime. */ - mutex_lock(&module_mutex); - old = find_module_all(mod->name, strlen(mod->name), - true); - } - - /* - * We are here only when the same module was being loaded. Do - * not try to load it again right now. It prevents long delays - * caused by serialized module load failures. It might happen - * when more devices of the same type trigger load of - * a particular module. - */ - if (old && old->state == MODULE_STATE_LIVE) - err = -EEXIST; - else - err = -EBUSY; + err = module_patient_check_exists(mod->name); + if (err) goto out; - } + mod_update_bounds(mod); list_add_rcu(&mod->list, &modules); mod_tree_insert(mod); @@ -2681,7 +2690,6 @@ static int add_unformed_module(struct module *mod) out: mutex_unlock(&module_mutex); -out_unlocked: return err; } -- cgit From df3e764d8e5cd416efee29e0de3c93917dff5d33 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Tue, 28 Mar 2023 20:03:19 -0700 Subject: module: add debug stats to help identify memory pressure Loading modules with finit_module() can end up using vmalloc(), vmap() and vmalloc() again, for a total of up to 3 separate allocations in the worst case for a single module. We always kernel_read*() the module, that's a vmalloc(). Then vmap() is used for the module decompression, and if so the last read buffer is freed as we use the now decompressed module buffer to stuff data into our copy module. The last allocation is specific to each architectures but pretty much that's generally a series of vmalloc() calls or a variation of vmalloc to handle ELF sections with special permissions. Evaluation with new stress-ng module support [1] with just 100 ops is proving that you can end up using GiBs of data easily even with all care we have in the kernel and userspace today in trying to not load modules which are already loaded. 100 ops seems to resemble the sort of pressure a system with about 400 CPUs can create on module loading. Although issues relating to duplicate module requests due to each CPU inucurring a new module reuest is silly and some of these are being fixed, we currently lack proper tooling to help diagnose easily what happened, when it happened and who likely is to blame -- userspace or kernel module autoloading. Provide an initial set of stats which use debugfs to let us easily scrape post-boot information about failed loads. This sort of information can be used on production worklaods to try to optimize *avoiding* redundant memory pressure using finit_module(). There's a few examples that can be provided: A 255 vCPU system without the next patch in this series applied: Startup finished in 19.143s (kernel) + 7.078s (userspace) = 26.221s graphical.target reached after 6.988s in userspace And 13.58 GiB of virtual memory space lost due to failed module loading: root@big ~ # cat /sys/kernel/debug/modules/stats Mods ever loaded 67 Mods failed on kread 0 Mods failed on decompress 0 Mods failed on becoming 0 Mods failed on load 1411 Total module size 11464704 Total mod text size 4194304 Failed kread bytes 0 Failed decompress bytes 0 Failed becoming bytes 0 Failed kmod bytes 14588526272 Virtual mem wasted bytes 14588526272 Average mod size 171115 Average mod text size 62602 Average fail load bytes 10339140 Duplicate failed modules: module-name How-many-times Reason kvm_intel 249 Load kvm 249 Load irqbypass 8 Load crct10dif_pclmul 128 Load ghash_clmulni_intel 27 Load sha512_ssse3 50 Load sha512_generic 200 Load aesni_intel 249 Load crypto_simd 41 Load cryptd 131 Load evdev 2 Load serio_raw 1 Load virtio_pci 3 Load nvme 3 Load nvme_core 3 Load virtio_pci_legacy_dev 3 Load virtio_pci_modern_dev 3 Load t10_pi 3 Load virtio 3 Load crc32_pclmul 6 Load crc64_rocksoft 3 Load crc32c_intel 40 Load virtio_ring 3 Load crc64 3 Load The following screen shot, of a simple 8vcpu 8 GiB KVM guest with the next patch in this series applied, shows 226.53 MiB are wasted in virtual memory allocations which due to duplicate module requests during boot. It also shows an average module memory size of 167.10 KiB and an an average module .text + .init.text size of 61.13 KiB. The end shows all modules which were detected as duplicate requests and whether or not they failed early after just the first kernel_read*() call or late after we've already allocated the private space for the module in layout_and_allocate(). A system with module decompression would reveal more wasted virtual memory space. We should put effort now into identifying the source of these duplicate module requests and trimming these down as much possible. Larger systems will obviously show much more wasted virtual memory allocations. root@kmod ~ # cat /sys/kernel/debug/modules/stats Mods ever loaded 67 Mods failed on kread 0 Mods failed on decompress 0 Mods failed on becoming 83 Mods failed on load 16 Total module size 11464704 Total mod text size 4194304 Failed kread bytes 0 Failed decompress bytes 0 Failed becoming bytes 228959096 Failed kmod bytes 8578080 Virtual mem wasted bytes 237537176 Average mod size 171115 Average mod text size 62602 Avg fail becoming bytes 2758544 Average fail load bytes 536130 Duplicate failed modules: module-name How-many-times Reason kvm_intel 7 Becoming kvm 7 Becoming irqbypass 6 Becoming & Load crct10dif_pclmul 7 Becoming & Load ghash_clmulni_intel 7 Becoming & Load sha512_ssse3 6 Becoming & Load sha512_generic 7 Becoming & Load aesni_intel 7 Becoming crypto_simd 7 Becoming & Load cryptd 3 Becoming & Load evdev 1 Becoming serio_raw 1 Becoming nvme 3 Becoming nvme_core 3 Becoming t10_pi 3 Becoming virtio_pci 3 Becoming crc32_pclmul 6 Becoming & Load crc64_rocksoft 3 Becoming crc32c_intel 3 Becoming virtio_pci_modern_dev 2 Becoming virtio_pci_legacy_dev 1 Becoming crc64 2 Becoming virtio 2 Becoming virtio_ring 2 Becoming [0] https://github.com/ColinIanKing/stress-ng.git [1] echo 0 > /proc/sys/vm/oom_dump_tasks ./stress-ng --module 100 --module-name xfs Signed-off-by: Luis Chamberlain --- kernel/module/Kconfig | 41 ++++- kernel/module/Makefile | 1 + kernel/module/decompress.c | 4 + kernel/module/internal.h | 78 ++++++++ kernel/module/main.c | 65 ++++++- kernel/module/stats.c | 430 +++++++++++++++++++++++++++++++++++++++++++++ kernel/module/tracking.c | 7 +- 7 files changed, 615 insertions(+), 11 deletions(-) create mode 100644 kernel/module/stats.c (limited to 'kernel/module') diff --git a/kernel/module/Kconfig b/kernel/module/Kconfig index 424b3bc58f3f..e6df183e2c80 100644 --- a/kernel/module/Kconfig +++ b/kernel/module/Kconfig @@ -22,6 +22,45 @@ menuconfig MODULES if MODULES +config MODULE_DEBUGFS + bool + +config MODULE_DEBUG + bool "Module debugging" + depends on DEBUG_FS + help + Allows you to enable / disable features which can help you debug + modules. You don't need these options on production systems. + +if MODULE_DEBUG + +config MODULE_STATS + bool "Module statistics" + depends on DEBUG_FS + select MODULE_DEBUGFS + help + This option allows you to maintain a record of module statistics. + For example, size of all modules, average size, text size, a list + of failed modules and the size for each of those. For failed + modules we keep track of modules which failed due to either the + existing module taking too long to load or that module was already + loaded. + + You should enable this if you are debugging production loads + and want to see if userspace or the kernel is doing stupid things + with loading modules when it shouldn't or if you want to help + optimize userspace / kernel space module autoloading schemes. + You might want to do this because failed modules tend to use + up significant amount of memory, and so you'd be doing everyone a + favor in avoiding these failures proactively. + + This functionality is also useful for those experimenting with + module .text ELF section optimization. + + If unsure, say N. + +endif # MODULE_DEBUG + config MODULE_FORCE_LOAD bool "Forced module loading" default n @@ -51,7 +90,7 @@ config MODULE_FORCE_UNLOAD config MODULE_UNLOAD_TAINT_TRACKING bool "Tainted module unload tracking" depends on MODULE_UNLOAD - default n + select MODULE_DEBUGFS help This option allows you to maintain a record of each unloaded module that tainted the kernel. In addition to displaying a diff --git a/kernel/module/Makefile b/kernel/module/Makefile index 5b1d26b53b8d..52340bce497e 100644 --- a/kernel/module/Makefile +++ b/kernel/module/Makefile @@ -21,3 +21,4 @@ obj-$(CONFIG_SYSFS) += sysfs.o obj-$(CONFIG_KGDB_KDB) += kdb.o obj-$(CONFIG_MODVERSIONS) += version.o obj-$(CONFIG_MODULE_UNLOAD_TAINT_TRACKING) += tracking.o +obj-$(CONFIG_MODULE_STATS) += stats.o diff --git a/kernel/module/decompress.c b/kernel/module/decompress.c index 7ddc87bee274..e97232b125eb 100644 --- a/kernel/module/decompress.c +++ b/kernel/module/decompress.c @@ -297,6 +297,10 @@ int module_decompress(struct load_info *info, const void *buf, size_t size) ssize_t data_size; int error; +#if defined(CONFIG_MODULE_STATS) + info->compressed_len = size; +#endif + /* * Start with number of pages twice as big as needed for * compressed data. diff --git a/kernel/module/internal.h b/kernel/module/internal.h index 6ae29bb8836f..1fd75dd346dc 100644 --- a/kernel/module/internal.h +++ b/kernel/module/internal.h @@ -59,6 +59,9 @@ struct load_info { unsigned long mod_kallsyms_init_off; #endif #ifdef CONFIG_MODULE_DECOMPRESS +#ifdef CONFIG_MODULE_STATS + unsigned long compressed_len; +#endif struct page **pages; unsigned int max_pages; unsigned int used_pages; @@ -143,6 +146,81 @@ static inline bool set_livepatch_module(struct module *mod) #endif } +/** + * enum fail_dup_mod_reason - state at which a duplicate module was detected + * + * @FAIL_DUP_MOD_BECOMING: the module is read properly, passes all checks but + * we've determined that another module with the same name is already loaded + * or being processed on our &modules list. This happens on early_mod_check() + * right before layout_and_allocate(). The kernel would have already + * vmalloc()'d space for the entire module through finit_module(). If + * decompression was used two vmap() spaces were used. These failures can + * happen when userspace has not seen the module present on the kernel and + * tries to load the module multiple times at same time. + * @FAIL_DUP_MOD_LOAD: the module has been read properly, passes all validation + * checks and the kernel determines that the module was unique and because + * of this allocated yet another private kernel copy of the module space in + * layout_and_allocate() but after this determined in add_unformed_module() + * that another module with the same name is already loaded or being processed. + * These failures should be mitigated as much as possible and are indicative + * of really fast races in loading modules. Without module decompression + * they waste twice as much vmap space. With module decompression three + * times the module's size vmap space is wasted. + */ +enum fail_dup_mod_reason { + FAIL_DUP_MOD_BECOMING = 0, + FAIL_DUP_MOD_LOAD, +}; + +#ifdef CONFIG_MODULE_DEBUGFS +extern struct dentry *mod_debugfs_root; +#endif + +#ifdef CONFIG_MODULE_STATS + +#define mod_stat_add_long(count, var) atomic_long_add(count, var) +#define mod_stat_inc(name) atomic_inc(name) + +extern atomic_long_t total_mod_size; +extern atomic_long_t total_text_size; +extern atomic_long_t invalid_kread_bytes; +extern atomic_long_t invalid_decompress_bytes; + +extern atomic_t modcount; +extern atomic_t failed_kreads; +extern atomic_t failed_decompress; +struct mod_fail_load { + struct list_head list; + char name[MODULE_NAME_LEN]; + atomic_long_t count; + unsigned long dup_fail_mask; +}; + +int try_add_failed_module(const char *name, enum fail_dup_mod_reason reason); +void mod_stat_bump_invalid(struct load_info *info, int flags); +void mod_stat_bump_becoming(struct load_info *info, int flags); + +#else + +#define mod_stat_add_long(name, var) +#define mod_stat_inc(name) + +static inline int try_add_failed_module(const char *name, + enum fail_dup_mod_reason reason) +{ + return 0; +} + +static inline void mod_stat_bump_invalid(struct load_info *info, int flags) +{ +} + +static inline void mod_stat_bump_becoming(struct load_info *info, int flags) +{ +} + +#endif /* CONFIG_MODULE_STATS */ + #ifdef CONFIG_MODULE_UNLOAD_TAINT_TRACKING struct mod_unload_taint { struct list_head list; diff --git a/kernel/module/main.c b/kernel/module/main.c index 75b23257128d..01fffa8afef2 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -56,6 +56,7 @@ #include #include #include +#include #include #include "internal.h" @@ -2500,6 +2501,18 @@ static noinline int do_init_module(struct module *mod) { int ret = 0; struct mod_initfree *freeinit; +#if defined(CONFIG_MODULE_STATS) + unsigned int text_size = 0, total_size = 0; + + for_each_mod_mem_type(type) { + const struct module_memory *mod_mem = &mod->mem[type]; + if (mod_mem->size) { + total_size += mod_mem->size; + if (type == MOD_TEXT || type == MOD_INIT_TEXT) + text_size += mod_mem->size; + } + } +#endif freeinit = kmalloc(sizeof(*freeinit), GFP_KERNEL); if (!freeinit) { @@ -2561,6 +2574,7 @@ static noinline int do_init_module(struct module *mod) mod->mem[type].base = NULL; mod->mem[type].size = 0; } + #ifdef CONFIG_DEBUG_INFO_BTF_MODULES /* .BTF is not SHF_ALLOC and will get removed, so sanitize pointer */ mod->btf_data = NULL; @@ -2584,6 +2598,11 @@ static noinline int do_init_module(struct module *mod) mutex_unlock(&module_mutex); wake_up_all(&module_wq); + mod_stat_add_long(text_size, &total_text_size); + mod_stat_add_long(total_size, &total_mod_size); + + mod_stat_inc(&modcount); + return 0; fail_free_freeinit: @@ -2599,6 +2618,7 @@ fail: ftrace_release_mod(mod); free_module(mod); wake_up_all(&module_wq); + return ret; } @@ -2632,7 +2652,8 @@ static bool finished_loading(const char *name) } /* Must be called with module_mutex held */ -static int module_patient_check_exists(const char *name) +static int module_patient_check_exists(const char *name, + enum fail_dup_mod_reason reason) { struct module *old; int err = 0; @@ -2655,6 +2676,9 @@ static int module_patient_check_exists(const char *name) old = find_module_all(name, strlen(name), true); } + if (try_add_failed_module(name, reason)) + pr_warn("Could not add fail-tracking for module: %s\n", name); + /* * We are here only when the same module was being loaded. Do * not try to load it again right now. It prevents long delays @@ -2679,7 +2703,7 @@ static int add_unformed_module(struct module *mod) mod->state = MODULE_STATE_UNFORMED; mutex_lock(&module_mutex); - err = module_patient_check_exists(mod->name); + err = module_patient_check_exists(mod->name, FAIL_DUP_MOD_LOAD); if (err) goto out; @@ -2800,6 +2824,7 @@ static int load_module(struct load_info *info, const char __user *uargs, int flags) { struct module *mod; + bool module_allocated = false; long err = 0; char *after_dashes; @@ -2839,6 +2864,8 @@ static int load_module(struct load_info *info, const char __user *uargs, goto free_copy; } + module_allocated = true; + audit_log_kern_module(mod->name); /* Reserve our place in the list. */ @@ -2983,6 +3010,7 @@ static int load_module(struct load_info *info, const char __user *uargs, synchronize_rcu(); mutex_unlock(&module_mutex); free_module: + mod_stat_bump_invalid(info, flags); /* Free lock-classes; relies on the preceding sync_rcu() */ for_class_mod_mem_type(type, core_data) { lockdep_free_key_range(mod->mem[type].base, @@ -2991,6 +3019,13 @@ static int load_module(struct load_info *info, const char __user *uargs, module_deallocate(mod, info); free_copy: + /* + * The info->len is always set. We distinguish between + * failures once the proper module was allocated and + * before that. + */ + if (!module_allocated) + mod_stat_bump_becoming(info, flags); free_copy(info, flags); return err; } @@ -3009,8 +3044,11 @@ SYSCALL_DEFINE3(init_module, void __user *, umod, umod, len, uargs); err = copy_module_from_user(umod, len, &info); - if (err) + if (err) { + mod_stat_inc(&failed_kreads); + mod_stat_add_long(len, &invalid_kread_bytes); return err; + } return load_module(&info, uargs, 0); } @@ -3035,14 +3073,20 @@ SYSCALL_DEFINE3(finit_module, int, fd, const char __user *, uargs, int, flags) len = kernel_read_file_from_fd(fd, 0, &buf, INT_MAX, NULL, READING_MODULE); - if (len < 0) + if (len < 0) { + mod_stat_inc(&failed_kreads); + mod_stat_add_long(len, &invalid_kread_bytes); return len; + } if (flags & MODULE_INIT_COMPRESSED_FILE) { err = module_decompress(&info, buf, len); vfree(buf); /* compressed data is no longer needed */ - if (err) + if (err) { + mod_stat_inc(&failed_decompress); + mod_stat_add_long(len, &invalid_decompress_bytes); return err; + } } else { info.hdr = buf; info.len = len; @@ -3216,3 +3260,14 @@ void print_modules(void) last_unloaded_module.taints); pr_cont("\n"); } + +#ifdef CONFIG_MODULE_DEBUGFS +struct dentry *mod_debugfs_root; + +static int module_debugfs_init(void) +{ + mod_debugfs_root = debugfs_create_dir("modules", NULL); + return 0; +} +module_init(module_debugfs_init); +#endif diff --git a/kernel/module/stats.c b/kernel/module/stats.c new file mode 100644 index 000000000000..3d45744b3920 --- /dev/null +++ b/kernel/module/stats.c @@ -0,0 +1,430 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Debugging module statistics. + * + * Copyright (C) 2023 Luis Chamberlain + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#include "internal.h" + +/** + * DOC: module debugging statistics overview + * + * Enabling CONFIG_MODULE_STATS enables module debugging statistics which + * are useful to monitor and root cause memory pressure issues with module + * loading. These statistics are useful to allow us to improve production + * workloads. + * + * The current module debugging statistics supported help keep track of module + * loading failures to enable improvements either for kernel module auto-loading + * usage (request_module()) or interactions with userspace. Statistics are + * provided to track all possible failures in the finit_module() path and memory + * wasted in this process space. Each of the failure counters are associated + * to a type of module loading failure which is known to incur a certain amount + * of memory allocation loss. In the worst case loading a module will fail after + * a 3 step memory allocation process: + * + * a) memory allocated with kernel_read_file_from_fd() + * b) module decompression processes the file read from + * kernel_read_file_from_fd(), and vmap() is used to map + * the decompressed module to a new local buffer which represents + * a copy of the decompressed module passed from userspace. The buffer + * from kernel_read_file_from_fd() is freed right away. + * c) layout_and_allocate() allocates space for the final resting + * place where we would keep the module if it were to be processed + * successfully. + * + * If a failure occurs after these three different allocations only one + * counter will be incremented with the summation of the allocated bytes freed + * incurred during this failure. Likewise, if module loading failed only after + * step b) a separate counter is used and incremented for the bytes freed and + * not used during both of those allocations. + * + * Virtual memory space can be limited, for example on x86 virtual memory size + * defaults to 128 MiB. We should strive to limit and avoid wasting virtual + * memory allocations when possible. These module debugging statistics help + * to evaluate how much memory is being wasted on bootup due to module loading + * failures. + * + * All counters are designed to be incremental. Atomic counters are used so to + * remain simple and avoid delays and deadlocks. + */ + +/** + * DOC: dup_failed_modules - tracks duplicate failed modules + * + * Linked list of modules which failed to be loaded because an already existing + * module with the same name was already being processed or already loaded. + * The finit_module() system call incurs heavy virtual memory allocations. In + * the worst case an finit_module() system call can end up allocating virtual + * memory 3 times: + * + * 1) kernel_read_file_from_fd() call uses vmalloc() + * 2) optional module decompression uses vmap() + * 3) layout_and allocate() can use vzalloc() or an arch specific variation of + * vmalloc to deal with ELF sections requiring special permissions + * + * In practice on a typical boot today most finit_module() calls fail due to + * the module with the same name already being loaded or about to be processed. + * All virtual memory allocated to these failed modules will be freed with + * no functional use. + * + * To help with this the dup_failed_modules allows us to track modules which + * failed to load due to the fact that a module was already loaded or being + * processed. There are only two points at which we can fail such calls, + * we list them below along with the number of virtual memory allocation + * calls: + * + * a) FAIL_DUP_MOD_BECOMING: at the end of early_mod_check() before + * layout_and_allocate(). This does not yet happen. + * - with module decompression: 2 virtual memory allocation calls + * - without module decompression: 1 virtual memory allocation calls + * b) FAIL_DUP_MOD_LOAD: after layout_and_allocate() on add_unformed_module() + * - with module decompression 3 virtual memory allocation calls + * - without module decompression 2 virtual memory allocation calls + * + * We should strive to get this list to be as small as possible. If this list + * is not empty it is a reflection of possible work or optimizations possible + * either in-kernel or in userspace. + */ +static LIST_HEAD(dup_failed_modules); + +/** + * DOC: module statistics debugfs counters + * + * The total amount of wasted virtual memory allocation space during module + * loading can be computed by adding the total from the summation: + * + * * @invalid_kread_bytes + + * @invalid_decompress_bytes + + * @invalid_becoming_bytes + + * @invalid_mod_bytes + * + * The following debugfs counters are available to inspect module loading + * failures: + * + * * total_mod_size: total bytes ever used by all modules we've dealt with on + * this system + * * total_text_size: total bytes of the .text and .init.text ELF section + * sizes we've dealt with on this system + * * invalid_kread_bytes: bytes allocated and then freed on failures which + * happen due to the initial kernel_read_file_from_fd(). kernel_read_file_from_fd() + * uses vmalloc(). These should typically not happen unless your system is + * under memory pressure. + * * invalid_decompress_bytes: number of bytes allocated and freed due to + * memory allocations in the module decompression path that use vmap(). + * These typically should not happen unless your system is under memory + * pressure. + * * invalid_becoming_bytes: total number of bytes allocated and freed used + * used to read the kernel module userspace wants us to read before we + * promote it to be processed to be added to our @modules linked list. + * These failures could in theory happen if we had a check in + * between a successful kernel_read_file_from_fd() + * call and right before we allocate the our private memory for the module + * which would be kept if the module is successfully loaded. The most common + * reason for this failure is when userspace is racing to load a module + * which it does not yet see loaded. The first module to succeed in + * add_unformed_module() will add a module to our &modules list and + * subsequent loads of modules with the same name will error out at the + * end of early_mod_check(). A check for module_patient_check_exists() + * at the end of early_mod_check() could be added to prevent duplicate allocations + * on layout_and_allocate() for modules already being processed. These + * duplicate failed modules are non-fatal, however they typically are + * indicative of userspace not seeing a module in userspace loaded yet and + * unnecessarily trying to load a module before the kernel even has a chance + * to begin to process prior requests. Although duplicate failures can be + * non-fatal, we should try to reduce vmalloc() pressure proactively, so + * ideally after boot this will be close to as 0 as possible. If module + * decompression was used we also add to this counter the cost of the + * initial kernel_read_file_from_fd() of the compressed module. If module + * decompression was not used the value represents the total allocated and + * freed bytes in kernel_read_file_from_fd() calls for these type of + * failures. These failures can occur because: + * + * * module_sig_check() - module signature checks + * * elf_validity_cache_copy() - some ELF validation issue + * * early_mod_check(): + * + * * blacklisting + * * failed to rewrite section headers + * * version magic + * * live patch requirements didn't check out + * * the module was detected as being already present + * + * * invalid_mod_bytes: these are the total number of bytes allocated and + * freed due to failures after we did all the sanity checks of the module + * which userspace passed to us and after our first check that the module + * is unique. A module can still fail to load if we detect the module is + * loaded after we allocate space for it with layout_and_allocate(), we do + * this check right before processing the module as live and run its + * initialization routines. Note that you have a failure of this type it + * also means the respective kernel_read_file_from_fd() memory space was + * also freed and not used, and so we increment this counter with twice + * the size of the module. Additionally if you used module decompression + * the size of the compressed module is also added to this counter. + * + * * modcount: how many modules we've loaded in our kernel life time + * * failed_kreads: how many modules failed due to failed kernel_read_file_from_fd() + * * failed_decompress: how many failed module decompression attempts we've had. + * These really should not happen unless your compression / decompression + * might be broken. + * * failed_becoming: how many modules failed after we kernel_read_file_from_fd() + * it and before we allocate memory for it with layout_and_allocate(). This + * counter is never incremented if you manage to validate the module and + * call layout_and_allocate() for it. + * * failed_load_modules: how many modules failed once we've allocated our + * private space for our module using layout_and_allocate(). These failures + * should hopefully mostly be dealt with already. Races in theory could + * still exist here, but it would just mean the kernel had started processing + * two threads concurrently up to early_mod_check() and one thread won. + * These failures are good signs the kernel or userspace is doing something + * seriously stupid or that could be improved. We should strive to fix these, + * but it is perhaps not easy to fix them. A recent example are the modules + * requests incurred for frequency modules, a separate module request was + * being issued for each CPU on a system. + */ + +atomic_long_t total_mod_size; +atomic_long_t total_text_size; +atomic_long_t invalid_kread_bytes; +atomic_long_t invalid_decompress_bytes; +static atomic_long_t invalid_becoming_bytes; +static atomic_long_t invalid_mod_bytes; +atomic_t modcount; +atomic_t failed_kreads; +atomic_t failed_decompress; +static atomic_t failed_becoming; +static atomic_t failed_load_modules; + +static const char *mod_fail_to_str(struct mod_fail_load *mod_fail) +{ + if (test_bit(FAIL_DUP_MOD_BECOMING, &mod_fail->dup_fail_mask) && + test_bit(FAIL_DUP_MOD_LOAD, &mod_fail->dup_fail_mask)) + return "Becoming & Load"; + if (test_bit(FAIL_DUP_MOD_BECOMING, &mod_fail->dup_fail_mask)) + return "Becoming"; + if (test_bit(FAIL_DUP_MOD_LOAD, &mod_fail->dup_fail_mask)) + return "Load"; + return "Bug-on-stats"; +} + +void mod_stat_bump_invalid(struct load_info *info, int flags) +{ + atomic_long_add(info->len * 2, &invalid_mod_bytes); + atomic_inc(&failed_load_modules); +#if defined(CONFIG_MODULE_DECOMPRESS) + if (flags & MODULE_INIT_COMPRESSED_FILE) + atomic_long_add(info->compressed_len, &invalid_mod_byte); +#endif +} + +void mod_stat_bump_becoming(struct load_info *info, int flags) +{ + atomic_inc(&failed_becoming); + atomic_long_add(info->len, &invalid_becoming_bytes); +#if defined(CONFIG_MODULE_DECOMPRESS) + if (flags & MODULE_INIT_COMPRESSED_FILE) + atomic_long_add(info->compressed_len, &invalid_becoming_bytes); +#endif +} + +int try_add_failed_module(const char *name, enum fail_dup_mod_reason reason) +{ + struct mod_fail_load *mod_fail; + + list_for_each_entry_rcu(mod_fail, &dup_failed_modules, list, + lockdep_is_held(&module_mutex)) { + if (!strcmp(mod_fail->name, name)) { + atomic_long_inc(&mod_fail->count); + __set_bit(reason, &mod_fail->dup_fail_mask); + goto out; + } + } + + mod_fail = kzalloc(sizeof(*mod_fail), GFP_KERNEL); + if (!mod_fail) + return -ENOMEM; + memcpy(mod_fail->name, name, strlen(name)); + __set_bit(reason, &mod_fail->dup_fail_mask); + atomic_long_inc(&mod_fail->count); + list_add_rcu(&mod_fail->list, &dup_failed_modules); +out: + return 0; +} + +/* + * At 64 bytes per module and assuming a 1024 bytes preamble we can fit the + * 112 module prints within 8k. + * + * 1024 + (64*112) = 8k + */ +#define MAX_PREAMBLE 1024 +#define MAX_FAILED_MOD_PRINT 112 +#define MAX_BYTES_PER_MOD 64 +static ssize_t read_file_mod_stats(struct file *file, char __user *user_buf, + size_t count, loff_t *ppos) +{ + struct mod_fail_load *mod_fail; + unsigned int len, size, count_failed = 0; + char *buf; + u32 live_mod_count, fkreads, fdecompress, fbecoming, floads; + u64 total_size, text_size, ikread_bytes, ibecoming_bytes, idecompress_bytes, imod_bytes, + total_virtual_lost; + + live_mod_count = atomic_read(&modcount); + fkreads = atomic_read(&failed_kreads); + fdecompress = atomic_read(&failed_decompress); + fbecoming = atomic_read(&failed_becoming); + floads = atomic_read(&failed_load_modules); + + total_size = atomic64_read(&total_mod_size); + text_size = atomic64_read(&total_text_size); + ikread_bytes = atomic64_read(&invalid_kread_bytes); + idecompress_bytes = atomic64_read(&invalid_decompress_bytes); + ibecoming_bytes = atomic64_read(&invalid_becoming_bytes); + imod_bytes = atomic64_read(&invalid_mod_bytes); + + total_virtual_lost = ikread_bytes + idecompress_bytes + ibecoming_bytes + imod_bytes; + + size = MAX_PREAMBLE + min((unsigned int)(floads + fbecoming), + (unsigned int)MAX_FAILED_MOD_PRINT) * MAX_BYTES_PER_MOD; + buf = kzalloc(size, GFP_KERNEL); + if (buf == NULL) + return -ENOMEM; + + /* The beginning of our debug preamble */ + len = scnprintf(buf + 0, size - len, "%25s\t%u\n", "Mods ever loaded", live_mod_count); + + len += scnprintf(buf + len, size - len, "%25s\t%u\n", "Mods failed on kread", fkreads); + + len += scnprintf(buf + len, size - len, "%25s\t%u\n", "Mods failed on decompress", + fdecompress); + len += scnprintf(buf + len, size - len, "%25s\t%u\n", "Mods failed on becoming", fbecoming); + + len += scnprintf(buf + len, size - len, "%25s\t%u\n", "Mods failed on load", floads); + + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Total module size", total_size); + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Total mod text size", text_size); + + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed kread bytes", ikread_bytes); + + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed decompress bytes", + idecompress_bytes); + + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed becoming bytes", ibecoming_bytes); + + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed kmod bytes", imod_bytes); + + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Virtual mem wasted bytes", total_virtual_lost); + + if (live_mod_count && total_size) { + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Average mod size", + DIV_ROUND_UP(total_size, live_mod_count)); + } + + if (live_mod_count && text_size) { + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Average mod text size", + DIV_ROUND_UP(text_size, live_mod_count)); + } + + /* + * We use WARN_ON_ONCE() for the counters to ensure we always have parity + * for keeping tabs on a type of failure with one type of byte counter. + * The counters for imod_bytes does not increase for fkreads failures + * for example, and so on. + */ + + WARN_ON_ONCE(ikread_bytes && !fkreads); + if (fkreads && ikread_bytes) { + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Avg fail kread bytes", + DIV_ROUND_UP(ikread_bytes, fkreads)); + } + + WARN_ON_ONCE(ibecoming_bytes && !fbecoming); + if (fbecoming && ibecoming_bytes) { + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Avg fail becoming bytes", + DIV_ROUND_UP(ibecoming_bytes, fbecoming)); + } + + WARN_ON_ONCE(idecompress_bytes && !fdecompress); + if (fdecompress && idecompress_bytes) { + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Avg fail decomp bytes", + DIV_ROUND_UP(idecompress_bytes, fdecompress)); + } + + WARN_ON_ONCE(imod_bytes && !floads); + if (floads && imod_bytes) { + len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Average fail load bytes", + DIV_ROUND_UP(imod_bytes, floads)); + } + + /* End of our debug preamble header. */ + + /* Catch when we've gone beyond our expected preamble */ + WARN_ON_ONCE(len >= MAX_PREAMBLE); + + if (list_empty(&dup_failed_modules)) + goto out; + + len += scnprintf(buf + len, size - len, "Duplicate failed modules:\n"); + len += scnprintf(buf + len, size - len, "%25s\t%15s\t%25s\n", + "Module-name", "How-many-times", "Reason"); + mutex_lock(&module_mutex); + + + list_for_each_entry_rcu(mod_fail, &dup_failed_modules, list) { + if (WARN_ON_ONCE(++count_failed >= MAX_FAILED_MOD_PRINT)) + goto out_unlock; + len += scnprintf(buf + len, size - len, "%25s\t%15llu\t%25s\n", mod_fail->name, + atomic64_read(&mod_fail->count), mod_fail_to_str(mod_fail)); + } +out_unlock: + mutex_unlock(&module_mutex); +out: + kfree(buf); + return simple_read_from_buffer(user_buf, count, ppos, buf, len); +} +#undef MAX_PREAMBLE +#undef MAX_FAILED_MOD_PRINT +#undef MAX_BYTES_PER_MOD + +static const struct file_operations fops_mod_stats = { + .read = read_file_mod_stats, + .open = simple_open, + .owner = THIS_MODULE, + .llseek = default_llseek, +}; + +#define mod_debug_add_ulong(name) debugfs_create_ulong(#name, 0400, mod_debugfs_root, (unsigned long *) &name.counter) +#define mod_debug_add_atomic(name) debugfs_create_atomic_t(#name, 0400, mod_debugfs_root, &name) +static int __init module_stats_init(void) +{ + mod_debug_add_ulong(total_mod_size); + mod_debug_add_ulong(total_text_size); + mod_debug_add_ulong(invalid_kread_bytes); + mod_debug_add_ulong(invalid_decompress_bytes); + mod_debug_add_ulong(invalid_becoming_bytes); + mod_debug_add_ulong(invalid_mod_bytes); + + mod_debug_add_atomic(modcount); + mod_debug_add_atomic(failed_kreads); + mod_debug_add_atomic(failed_decompress); + mod_debug_add_atomic(failed_becoming); + mod_debug_add_atomic(failed_load_modules); + + debugfs_create_file("stats", 0400, mod_debugfs_root, mod_debugfs_root, &fops_mod_stats); + + return 0; +} +#undef mod_debug_add_ulong +#undef mod_debug_add_atomic +module_init(module_stats_init); diff --git a/kernel/module/tracking.c b/kernel/module/tracking.c index 26d812e07615..16742d1c630c 100644 --- a/kernel/module/tracking.c +++ b/kernel/module/tracking.c @@ -15,6 +15,7 @@ #include "internal.h" static LIST_HEAD(unloaded_tainted_modules); +extern struct dentry *mod_debugfs_root; int try_add_tainted_module(struct module *mod) { @@ -120,12 +121,8 @@ static const struct file_operations unloaded_tainted_modules_fops = { static int __init unloaded_tainted_modules_init(void) { - struct dentry *dir; - - dir = debugfs_create_dir("modules", NULL); - debugfs_create_file("unloaded_tainted", 0444, dir, NULL, + debugfs_create_file("unloaded_tainted", 0444, mod_debugfs_root, NULL, &unloaded_tainted_modules_fops); - return 0; } module_init(unloaded_tainted_modules_init); -- cgit From 064f4536d13939b6e8cdb71298ff5d657f4f8caa Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Fri, 10 Mar 2023 20:48:03 -0800 Subject: module: avoid allocation if module is already present and ready The finit_module() system call can create unnecessary virtual memory pressure for duplicate modules. This is because load_module() can in the worse case allocate more than twice the size of a module in virtual memory. This saves at least a full size of the module in wasted vmalloc space memory by trying to avoid duplicates as soon as we can validate the module name in the read module structure. This can only be an issue if a system is getting hammered with userspace loading modules. There are two ways to load modules typically on systems, one is the kernel moduile auto-loading (*request_module*() calls in-kernel) and the other is things like udev. The auto-loading is in-kernel, but that pings back to userspace to just call modprobe. We already have a way to restrict the amount of concurrent kernel auto-loads in a given time, however that still allows multiple requests for the same module to go through and force two threads in userspace racing to call modprobe for the same exact module. Even though libkmod which both modprobe and udev does check if a module is already loaded prior calling finit_module() races are still possible and this is clearly evident today when you have multiple CPUs. To avoid memory pressure for such stupid cases put a stop gap for them. The *earliest* we can detect duplicates from the modules side of things is once we have blessed the module name, sadly after the first vmalloc allocation. We can check for the module being present *before* a secondary vmalloc() allocation. There is a linear relationship between wasted virtual memory bytes and the number of CPU counts. The reason is that udev ends up racing to call tons of the same modules for each of the CPUs. We can see the different linear relationships between wasted virtual memory and CPU count during after boot in the following graph: +----------------------------------------------------------------------------+ 14GB |-+ + + + + *+ +-| | **** | | *** | | ** | 12GB |-+ ** +-| | ** | | ** | | ** | | ** | 10GB |-+ ** +-| | ** | | ** | | ** | 8GB |-+ ** +-| waste | ** ### | | ** #### | | ** ####### | 6GB |-+ **** #### +-| | * #### | | * #### | | ***** #### | 4GB |-+ ** #### +-| | ** #### | | ** #### | | ** #### | 2GB |-+ ** ##### +-| | * #### | | * #### Before ******* | | **## + + + + After ####### | +----------------------------------------------------------------------------+ 0 50 100 150 200 250 300 CPUs count On the y-axis we can see gigabytes of wasted virtual memory during boot due to duplicate module requests which just end up failing. Trying to infer the slope this ends up being about ~463 MiB per CPU lost prior to this patch. After this patch we only loose about ~230 MiB per CPU, for a total savings of about ~233 MiB per CPU. This is all *just on bootup*! On a 8vcpu 8 GiB RAM system using kdevops and testing against selftests kmod.sh -t 0008 I see a saving in the *highest* side of memory consumption of up to ~ 84 MiB with the Linux kernel selftests kmod test 0008. With the new stress-ng module test I see a 145 MiB difference in max memory consumption with 100 ops. The stress-ng module ops tests can be pretty pathalogical -- it is not realistic, however it was used to finally successfully reproduce issues which are only reported to happen on system with over 400 CPUs [0] by just usign 100 ops on a 8vcpu 8 GiB RAM system. Running out of virtual memory space is no surprise given the above graph, since at least on x86_64 we're capped at 128 MiB, eventually we'd hit a series of errors and once can use the above graph to guestimate when. This of course will vary depending on the features you have enabled. So for instance, enabling KASAN seems to make this much worse. The results with kmod and stress-ng can be observed and visualized below. The time it takes to run the test is also not affected. The kmod tests 0008: The gnuplot is set to a range from 400000 KiB (390 Mib) - 580000 (566 Mib) given the tests peak around that range. cat kmod.plot set term dumb set output fileout set yrange [400000:580000] plot filein with linespoints title "Memory usage (KiB)" Before: root@kmod ~ # /data/linux-next/tools/testing/selftests/kmod/kmod.sh -t 0008 root@kmod ~ # free -k -s 1 -c 40 | grep Mem | awk '{print $3}' > log-0008-before.txt ^C root@kmod ~ # sort -n -r log-0008-before.txt | head -1 528732 So ~516.33 MiB After: root@kmod ~ # /data/linux-next/tools/testing/selftests/kmod/kmod.sh -t 0008 root@kmod ~ # free -k -s 1 -c 40 | grep Mem | awk '{print $3}' > log-0008-after.txt ^C root@kmod ~ # sort -n -r log-0008-after.txt | head -1 442516 So ~432.14 MiB That's about 84 ~MiB in savings in the worst case. The graphs: root@kmod ~ # gnuplot -e "filein='log-0008-before.txt'; fileout='graph-0008-before.txt'" kmod.plot root@kmod ~ # gnuplot -e "filein='log-0008-after.txt'; fileout='graph-0008-after.txt'" kmod.plot root@kmod ~ # cat graph-0008-before.txt 580000 +-----------------------------------------------------------------+ | + + + + + + + | 560000 |-+ Memory usage (KiB) ***A***-| | | 540000 |-+ +-| | | | *A *AA*AA*A*AA *A*AA A*A*A *AA*A*AA*A A | 520000 |-+A*A*AA *AA*A *A*AA*A*AA *A*A A *A+-| |*A | 500000 |-+ +-| | | 480000 |-+ +-| | | 460000 |-+ +-| | | | | 440000 |-+ +-| | | 420000 |-+ +-| | + + + + + + + | 400000 +-----------------------------------------------------------------+ 0 5 10 15 20 25 30 35 40 root@kmod ~ # cat graph-0008-after.txt 580000 +-----------------------------------------------------------------+ | + + + + + + + | 560000 |-+ Memory usage (KiB) ***A***-| | | 540000 |-+ +-| | | | | 520000 |-+ +-| | | 500000 |-+ +-| | | 480000 |-+ +-| | | 460000 |-+ +-| | | | *A *A*A | 440000 |-+A*A*AA*A A A*A*AA A*A*AA*A*AA*A*AA*A*AA*AA*A*AA*A*AA-| |*A *A*AA*A | 420000 |-+ +-| | + + + + + + + | 400000 +-----------------------------------------------------------------+ 0 5 10 15 20 25 30 35 40 The stress-ng module tests: This is used to run the test to try to reproduce the vmap issues reported by David: echo 0 > /proc/sys/vm/oom_dump_tasks ./stress-ng --module 100 --module-name xfs Prior to this commit: root@kmod ~ # free -k -s 1 -c 40 | grep Mem | awk '{print $3}' > baseline-stress-ng.txt root@kmod ~ # sort -n -r baseline-stress-ng.txt | head -1 5046456 After this commit: root@kmod ~ # free -k -s 1 -c 40 | grep Mem | awk '{print $3}' > after-stress-ng.txt root@kmod ~ # sort -n -r after-stress-ng.txt | head -1 4896972 5046456 - 4896972 149484 149484/1024 145.98046875000000000000 So this commit using stress-ng reveals saving about 145 MiB in memory using 100 ops from stress-ng which reproduced the vmap issue reported. cat kmod.plot set term dumb set output fileout set yrange [4700000:5070000] plot filein with linespoints title "Memory usage (KiB)" root@kmod ~ # gnuplot -e "filein='baseline-stress-ng.txt'; fileout='graph-stress-ng-before.txt'" kmod-simple-stress-ng.plot root@kmod ~ # gnuplot -e "filein='after-stress-ng.txt'; fileout='graph-stress-ng-after.txt'" kmod-simple-stress-ng.plot root@kmod ~ # cat graph-stress-ng-before.txt +---------------------------------------------------------------+ 5.05e+06 |-+ + A + + + + + + +-| | * Memory usage (KiB) ***A*** | | * A | 5e+06 |-+ ** ** +-| | ** * * A | 4.95e+06 |-+ * * A * A* +-| | * * A A * * * * A | | * * * * * * *A * * * A * | 4.9e+06 |-+ * * * A*A * A*AA*A A *A **A **A*A *+-| | A A*A A * A * * A A * A * ** | | * ** ** * * * * * * * | 4.85e+06 |-+ A A A ** * * ** *-| | * * * * ** * | | * A * * * * | 4.8e+06 |-+ * * * A A-| | * * * | 4.75e+06 |-+ * * * +-| | * ** | | * + + + + + + ** + | 4.7e+06 +---------------------------------------------------------------+ 0 5 10 15 20 25 30 35 40 root@kmod ~ # cat graph-stress-ng-after.txt +---------------------------------------------------------------+ 5.05e+06 |-+ + + + + + + + +-| | Memory usage (KiB) ***A*** | | | 5e+06 |-+ +-| | | 4.95e+06 |-+ +-| | | | | 4.9e+06 |-+ *AA +-| | A*AA*A*A A A*AA*AA*A*AA*A A A A*A *AA*A*A A A*AA*AA | | * * ** * * * ** * *** * | 4.85e+06 |-+* *** * * * * *** A * * +-| | * A * * ** * * A * * | | * * * * ** * * | 4.8e+06 |-+* * * A * * * +-| | * * * A * * | 4.75e+06 |-* * * * * +-| | * * * * * | | * + * *+ + + + + * *+ | 4.7e+06 +---------------------------------------------------------------+ 0 5 10 15 20 25 30 35 40 [0] https://lkml.kernel.org/r/20221013180518.217405-1-david@redhat.com Reported-by: David Hildenbrand Signed-off-by: Luis Chamberlain --- kernel/module/main.c | 6 +++++- kernel/module/stats.c | 11 +++++------ 2 files changed, 10 insertions(+), 7 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/main.c b/kernel/module/main.c index 01fffa8afef2..044aa2c9e3cb 100644 --- a/kernel/module/main.c +++ b/kernel/module/main.c @@ -2813,7 +2813,11 @@ static int early_mod_check(struct load_info *info, int flags) if (err) return err; - return 0; + mutex_lock(&module_mutex); + err = module_patient_check_exists(info->mod->name, FAIL_DUP_MOD_BECOMING); + mutex_unlock(&module_mutex); + + return err; } /* diff --git a/kernel/module/stats.c b/kernel/module/stats.c index 3d45744b3920..1d31f96a5e28 100644 --- a/kernel/module/stats.c +++ b/kernel/module/stats.c @@ -85,7 +85,7 @@ * calls: * * a) FAIL_DUP_MOD_BECOMING: at the end of early_mod_check() before - * layout_and_allocate(). This does not yet happen. + * layout_and_allocate(). * - with module decompression: 2 virtual memory allocation calls * - without module decompression: 1 virtual memory allocation calls * b) FAIL_DUP_MOD_LOAD: after layout_and_allocate() on add_unformed_module() @@ -126,17 +126,16 @@ static LIST_HEAD(dup_failed_modules); * pressure. * * invalid_becoming_bytes: total number of bytes allocated and freed used * used to read the kernel module userspace wants us to read before we - * promote it to be processed to be added to our @modules linked list. - * These failures could in theory happen if we had a check in - * between a successful kernel_read_file_from_fd() + * promote it to be processed to be added to our @modules linked list. These + * failures can happen if we had a check in between a successful kernel_read_file_from_fd() * call and right before we allocate the our private memory for the module * which would be kept if the module is successfully loaded. The most common * reason for this failure is when userspace is racing to load a module * which it does not yet see loaded. The first module to succeed in * add_unformed_module() will add a module to our &modules list and * subsequent loads of modules with the same name will error out at the - * end of early_mod_check(). A check for module_patient_check_exists() - * at the end of early_mod_check() could be added to prevent duplicate allocations + * end of early_mod_check(). The check for module_patient_check_exists() + * at the end of early_mod_check() prevents duplicate allocations * on layout_and_allocate() for modules already being processed. These * duplicate failed modules are non-fatal, however they typically are * indicative of userspace not seeing a module in userspace loaded yet and -- cgit From 635dc38314c75c5727711d896d4c71ec92f6f20b Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 18 Apr 2023 00:02:46 +0200 Subject: module: stats: include uapi/linux/module.h MODULE_INIT_COMPRESSED_FILE is defined in the uapi header, which is not included indirectly from the normal linux/module.h, but has to be pulled in explicitly: kernel/module/stats.c: In function 'mod_stat_bump_invalid': kernel/module/stats.c:227:14: error: 'MODULE_INIT_COMPRESSED_FILE' undeclared (first use in this function) 227 | if (flags & MODULE_INIT_COMPRESSED_FILE) | ^~~~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Arnd Bergmann Signed-off-by: Luis Chamberlain --- kernel/module/stats.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel/module') diff --git a/kernel/module/stats.c b/kernel/module/stats.c index 1d31f96a5e28..c1cccd2c3595 100644 --- a/kernel/module/stats.c +++ b/kernel/module/stats.c @@ -6,6 +6,7 @@ */ #include +#include #include #include #include -- cgit From 719ccd803ed5bd1ad92b0b46fc095b8fe266827e Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 18 Apr 2023 00:48:04 +0200 Subject: module: fix building stats for 32-bit targets The new module statistics code mixes 64-bit types and wordsized 'long' variables, which leads to build failures on 32-bit architectures: kernel/module/stats.c: In function 'read_file_mod_stats': kernel/module/stats.c:291:29: error: passing argument 1 of 'atomic64_read' from incompatible pointer type [-Werror=incompatible-pointer-types] 291 | total_size = atomic64_read(&total_mod_size); x86_64-linux-ld: kernel/module/stats.o: in function `read_file_mod_stats': stats.c:(.text+0x2b2): undefined reference to `__udivdi3' To fix this, the code has to use one of the two types consistently. Change them all to word-size types here. Fixes: df3e764d8e5c ("module: add debug stats to help identify memory pressure") Signed-off-by: Arnd Bergmann Signed-off-by: Luis Chamberlain --- kernel/module/stats.c | 46 +++++++++++++++++++++++----------------------- 1 file changed, 23 insertions(+), 23 deletions(-) (limited to 'kernel/module') diff --git a/kernel/module/stats.c b/kernel/module/stats.c index c1cccd2c3595..a1f545479174 100644 --- a/kernel/module/stats.c +++ b/kernel/module/stats.c @@ -277,8 +277,8 @@ static ssize_t read_file_mod_stats(struct file *file, char __user *user_buf, unsigned int len, size, count_failed = 0; char *buf; u32 live_mod_count, fkreads, fdecompress, fbecoming, floads; - u64 total_size, text_size, ikread_bytes, ibecoming_bytes, idecompress_bytes, imod_bytes, - total_virtual_lost; + unsigned long total_size, text_size, ikread_bytes, ibecoming_bytes, + idecompress_bytes, imod_bytes, total_virtual_lost; live_mod_count = atomic_read(&modcount); fkreads = atomic_read(&failed_kreads); @@ -286,12 +286,12 @@ static ssize_t read_file_mod_stats(struct file *file, char __user *user_buf, fbecoming = atomic_read(&failed_becoming); floads = atomic_read(&failed_load_modules); - total_size = atomic64_read(&total_mod_size); - text_size = atomic64_read(&total_text_size); - ikread_bytes = atomic64_read(&invalid_kread_bytes); - idecompress_bytes = atomic64_read(&invalid_decompress_bytes); - ibecoming_bytes = atomic64_read(&invalid_becoming_bytes); - imod_bytes = atomic64_read(&invalid_mod_bytes); + total_size = atomic_long_read(&total_mod_size); + text_size = atomic_long_read(&total_text_size); + ikread_bytes = atomic_long_read(&invalid_kread_bytes); + idecompress_bytes = atomic_long_read(&invalid_decompress_bytes); + ibecoming_bytes = atomic_long_read(&invalid_becoming_bytes); + imod_bytes = atomic_long_read(&invalid_mod_bytes); total_virtual_lost = ikread_bytes + idecompress_bytes + ibecoming_bytes + imod_bytes; @@ -312,27 +312,27 @@ static ssize_t read_file_mod_stats(struct file *file, char __user *user_buf, len += scnprintf(buf + len, size - len, "%25s\t%u\n", "Mods failed on load", floads); - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Total module size", total_size); - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Total mod text size", text_size); + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Total module size", total_size); + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Total mod text size", text_size); - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed kread bytes", ikread_bytes); + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Failed kread bytes", ikread_bytes); - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed decompress bytes", + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Failed decompress bytes", idecompress_bytes); - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed becoming bytes", ibecoming_bytes); + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Failed becoming bytes", ibecoming_bytes); - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Failed kmod bytes", imod_bytes); + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Failed kmod bytes", imod_bytes); - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Virtual mem wasted bytes", total_virtual_lost); + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Virtual mem wasted bytes", total_virtual_lost); if (live_mod_count && total_size) { - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Average mod size", + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Average mod size", DIV_ROUND_UP(total_size, live_mod_count)); } if (live_mod_count && text_size) { - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Average mod text size", + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Average mod text size", DIV_ROUND_UP(text_size, live_mod_count)); } @@ -345,25 +345,25 @@ static ssize_t read_file_mod_stats(struct file *file, char __user *user_buf, WARN_ON_ONCE(ikread_bytes && !fkreads); if (fkreads && ikread_bytes) { - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Avg fail kread bytes", + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Avg fail kread bytes", DIV_ROUND_UP(ikread_bytes, fkreads)); } WARN_ON_ONCE(ibecoming_bytes && !fbecoming); if (fbecoming && ibecoming_bytes) { - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Avg fail becoming bytes", + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Avg fail becoming bytes", DIV_ROUND_UP(ibecoming_bytes, fbecoming)); } WARN_ON_ONCE(idecompress_bytes && !fdecompress); if (fdecompress && idecompress_bytes) { - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Avg fail decomp bytes", + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Avg fail decomp bytes", DIV_ROUND_UP(idecompress_bytes, fdecompress)); } WARN_ON_ONCE(imod_bytes && !floads); if (floads && imod_bytes) { - len += scnprintf(buf + len, size - len, "%25s\t%llu\n", "Average fail load bytes", + len += scnprintf(buf + len, size - len, "%25s\t%lu\n", "Average fail load bytes", DIV_ROUND_UP(imod_bytes, floads)); } @@ -384,8 +384,8 @@ static ssize_t read_file_mod_stats(struct file *file, char __user *user_buf, list_for_each_entry_rcu(mod_fail, &dup_failed_modules, list) { if (WARN_ON_ONCE(++count_failed >= MAX_FAILED_MOD_PRINT)) goto out_unlock; - len += scnprintf(buf + len, size - len, "%25s\t%15llu\t%25s\n", mod_fail->name, - atomic64_read(&mod_fail->count), mod_fail_to_str(mod_fail)); + len += scnprintf(buf + len, size - len, "%25s\t%15lu\t%25s\n", mod_fail->name, + atomic_long_read(&mod_fail->count), mod_fail_to_str(mod_fail)); } out_unlock: mutex_unlock(&module_mutex); -- cgit From 9f5cab173e19201eebeaca853ff664a9a269fed0 Mon Sep 17 00:00:00 2001 From: Tom Rix Date: Mon, 17 Apr 2023 19:09:57 -0400 Subject: module: remove use of uninitialized variable len clang build reports kernel/module/stats.c:307:34: error: variable 'len' is uninitialized when used here [-Werror,-Wuninitialized] len = scnprintf(buf + 0, size - len, ^~~ At the start of this sequence, neither the '+ 0', nor the '- len' are needed. So remove them and fix using 'len' uninitalized. Fixes: df3e764d8e5c ("module: add debug stats to help identify memory pressure") Signed-off-by: Tom Rix Signed-off-by: Luis Chamberlain --- kernel/module/stats.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/module') diff --git a/kernel/module/stats.c b/kernel/module/stats.c index a1f545479174..6eeb35cc8d9f 100644 --- a/kernel/module/stats.c +++ b/kernel/module/stats.c @@ -302,7 +302,7 @@ static ssize_t read_file_mod_stats(struct file *file, char __user *user_buf, return -ENOMEM; /* The beginning of our debug preamble */ - len = scnprintf(buf + 0, size - len, "%25s\t%u\n", "Mods ever loaded", live_mod_count); + len = scnprintf(buf, size, "%25s\t%u\n", "Mods ever loaded", live_mod_count); len += scnprintf(buf + len, size - len, "%25s\t%u\n", "Mods failed on kread", fkreads); -- cgit From a81b1fc8ea639e03326c1d0dcde041986bc11500 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 18 Apr 2023 09:17:51 +0200 Subject: module: stats: fix invalid_mod_bytes typo This was caught by randconfig builds but does not show up in build testing without CONFIG_MODULE_DECOMPRESS: kernel/module/stats.c: In function 'mod_stat_bump_invalid': kernel/module/stats.c:229:42: error: 'invalid_mod_byte' undeclared (first use in this function); did you mean 'invalid_mod_bytes'? 229 | atomic_long_add(info->compressed_len, &invalid_mod_byte); | ^~~~~~~~~~~~~~~~ | invalid_mod_bytes Fixes: df3e764d8e5c ("module: add debug stats to help identify memory pressure") Signed-off-by: Arnd Bergmann Acked-by: Randy Dunlap Tested-by: Randy Dunlap Signed-off-by: Luis Chamberlain --- kernel/module/stats.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/module') diff --git a/kernel/module/stats.c b/kernel/module/stats.c index 6eeb35cc8d9f..ad7b6ada29f2 100644 --- a/kernel/module/stats.c +++ b/kernel/module/stats.c @@ -223,7 +223,7 @@ void mod_stat_bump_invalid(struct load_info *info, int flags) atomic_inc(&failed_load_modules); #if defined(CONFIG_MODULE_DECOMPRESS) if (flags & MODULE_INIT_COMPRESSED_FILE) - atomic_long_add(info->compressed_len, &invalid_mod_byte); + atomic_long_add(info->compressed_len, &invalid_mod_bytes); #endif } -- cgit From 8660484ed1cf3261e89e0bad94c6395597e87599 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Thu, 13 Apr 2023 22:28:39 -0700 Subject: module: add debugging auto-load duplicate module support The finit_module() system call can in the worst case use up to more than twice of a module's size in virtual memory. Duplicate finit_module() system calls are non fatal, however they unnecessarily strain virtual memory during bootup and in the worst case can cause a system to fail to boot. This is only known to currently be an issue on systems with larger number of CPUs. To help debug this situation we need to consider the different sources for finit_module(). Requests from the kernel that rely on module auto-loading, ie, the kernel's *request_module() API, are one source of calls. Although modprobe checks to see if a module is already loaded prior to calling finit_module() there is a small race possible allowing userspace to trigger multiple modprobe calls racing against modprobe and this not seeing the module yet loaded. This adds debugging support to the kernel module auto-loader (*request_module() calls) to easily detect duplicate module requests. To aid with possible bootup failure issues incurred by this, it will converge duplicates requests to a single request. This avoids any possible strain on virtual memory during bootup which could be incurred by duplicate module autoloading requests. Folks debugging virtual memory abuse on bootup can and should enable this to see what pr_warn()s come on, to see if module auto-loading is to blame for their wores. If they see duplicates they can further debug this by enabling the module.enable_dups_trace kernel parameter or by enabling CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS_TRACE. Current evidence seems to point to only a few duplicates for module auto-loading. And so the source for other duplicates creating heavy virtual memory pressure due to larger number of CPUs should becoming from another place (likely udev). Signed-off-by: Luis Chamberlain --- kernel/module/Kconfig | 59 ++++++++++++ kernel/module/Makefile | 1 + kernel/module/dups.c | 246 +++++++++++++++++++++++++++++++++++++++++++++++ kernel/module/internal.h | 15 +++ kernel/module/kmod.c | 23 ++++- 5 files changed, 340 insertions(+), 4 deletions(-) create mode 100644 kernel/module/dups.c (limited to 'kernel/module') diff --git a/kernel/module/Kconfig b/kernel/module/Kconfig index e6df183e2c80..33a2e991f608 100644 --- a/kernel/module/Kconfig +++ b/kernel/module/Kconfig @@ -59,6 +59,65 @@ config MODULE_STATS If unsure, say N. +config MODULE_DEBUG_AUTOLOAD_DUPS + bool "Debug duplicate modules with auto-loading" + help + Module autoloading allows in-kernel code to request modules through + the *request_module*() API calls. This in turn just calls userspace + modprobe. Although modprobe checks to see if a module is already + loaded before trying to load a module there is a small time window in + which multiple duplicate requests can end up in userspace and multiple + modprobe calls race calling finit_module() around the same time for + duplicate modules. The finit_module() system call can consume in the + worst case more than twice the respective module size in virtual + memory for each duplicate module requests. Although duplicate module + requests are non-fatal virtual memory is a limited resource and each + duplicate module request ends up just unnecessarily straining virtual + memory. + + This debugging facility will create pr_warn() splats for duplicate + module requests to help identify if module auto-loading may be the + culprit to your early boot virtual memory pressure. Since virtual + memory abuse caused by duplicate module requests could render a + system unusable this functionality will also converge races in + requests for the same module to a single request. You can boot with + the module.enable_dups_trace=1 kernel parameter to use WARN_ON() + instead of the pr_warn(). + + If the first module request used request_module_nowait() we cannot + use that as the anchor to wait for duplicate module requests, since + users of request_module() do want a proper return value. If a call + for the same module happened earlier with request_module() though, + then a duplicate request_module_nowait() would be detected. The + non-wait request_module() call is synchronous and waits until modprobe + completes. Subsequent auto-loading requests for the same module do + not trigger a new finit_module() calls and do not strain virtual + memory, and so as soon as modprobe successfully completes we remove + tracking for duplicates for that module. + + Enable this functionality to try to debug virtual memory abuse during + boot on systems which are failing to boot or if you suspect you may be + straining virtual memory during boot, and you want to identify if the + abuse was due to module auto-loading. These issues are currently only + known to occur on systems with many CPUs (over 400) and is likely the + result of udev issuing duplicate module requests for each CPU, and so + module auto-loading is not the culprit. There may very well still be + many duplicate module auto-loading requests which could be optimized + for and this debugging facility can be used to help identify them. + + Only enable this for debugging system functionality, never have it + enabled on real systems. + +config MODULE_DEBUG_AUTOLOAD_DUPS_TRACE + bool "Force full stack trace when duplicates are found" + depends on MODULE_DEBUG_AUTOLOAD_DUPS + help + Enabling this will force a full stack trace for duplicate module + auto-loading requests using WARN_ON() instead of pr_warn(). You + should keep this disabled at all times unless you are a developer + and are doing a manual inspection and want to debug exactly why + these duplicates occur. + endif # MODULE_DEBUG config MODULE_FORCE_LOAD diff --git a/kernel/module/Makefile b/kernel/module/Makefile index 52340bce497e..a10b2b9a6fdf 100644 --- a/kernel/module/Makefile +++ b/kernel/module/Makefile @@ -10,6 +10,7 @@ KCOV_INSTRUMENT_module.o := n obj-y += main.o obj-y += strict_rwx.o obj-y += kmod.o +obj-$(CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS) += dups.o obj-$(CONFIG_MODULE_DECOMPRESS) += decompress.o obj-$(CONFIG_MODULE_SIG) += signing.o obj-$(CONFIG_LIVEPATCH) += livepatch.o diff --git a/kernel/module/dups.c b/kernel/module/dups.c new file mode 100644 index 000000000000..aa8e1361fdb5 --- /dev/null +++ b/kernel/module/dups.c @@ -0,0 +1,246 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * kmod dups - the kernel module autoloader duplicate suppressor + * + * Copyright (C) 2023 Luis Chamberlain + */ + +#define pr_fmt(fmt) "module: " fmt + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#undef MODULE_PARAM_PREFIX +#define MODULE_PARAM_PREFIX "module." +static bool enable_dups_trace = IS_ENABLED(CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS_TRACE); +module_param(enable_dups_trace, bool_enable_only, 0644); + +/* + * Protects dup_kmod_reqs list, adds / removals with RCU. + */ +static DEFINE_MUTEX(kmod_dup_mutex); +static LIST_HEAD(dup_kmod_reqs); + +struct kmod_dup_req { + struct list_head list; + char name[MODULE_NAME_LEN]; + struct completion first_req_done; + struct work_struct complete_work; + struct delayed_work delete_work; + int dup_ret; +}; + +static struct kmod_dup_req *kmod_dup_request_lookup(char *module_name) +{ + struct kmod_dup_req *kmod_req; + + list_for_each_entry_rcu(kmod_req, &dup_kmod_reqs, list, + lockdep_is_held(&kmod_dup_mutex)) { + if (strlen(kmod_req->name) == strlen(module_name) && + !memcmp(kmod_req->name, module_name, strlen(module_name))) { + return kmod_req; + } + } + + return NULL; +} + +static void kmod_dup_request_delete(struct work_struct *work) +{ + struct kmod_dup_req *kmod_req; + kmod_req = container_of(to_delayed_work(work), struct kmod_dup_req, delete_work); + + /* + * The typical situation is a module successully loaded. In that + * situation the module will be present already in userspace. If + * new requests come in after that, userspace will already know the + * module is loaded so will just return 0 right away. There is still + * a small chance right after we delete this entry new request_module() + * calls may happen after that, they can happen. These heuristics + * are to protect finit_module() abuse for auto-loading, if modules + * are still tryign to auto-load even if a module is already loaded, + * that's on them, and those inneficiencies should not be fixed by + * kmod. The inneficies there are a call to modprobe and modprobe + * just returning 0. + */ + mutex_lock(&kmod_dup_mutex); + list_del_rcu(&kmod_req->list); + synchronize_rcu(); + mutex_unlock(&kmod_dup_mutex); + kfree(kmod_req); +} + +static void kmod_dup_request_complete(struct work_struct *work) +{ + struct kmod_dup_req *kmod_req; + + kmod_req = container_of(work, struct kmod_dup_req, complete_work); + + /* + * This will ensure that the kernel will let all the waiters get + * informed its time to check the return value. It's time to + * go home. + */ + complete_all(&kmod_req->first_req_done); + + /* + * Now that we have allowed prior request_module() calls to go on + * with life, let's schedule deleting this entry. We don't have + * to do it right away, but we *eventually* want to do it so to not + * let this linger forever as this is just a boot optimization for + * possible abuses of vmalloc() incurred by finit_module() thrashing. + */ + queue_delayed_work(system_wq, &kmod_req->delete_work, 60 * HZ); +} + +bool kmod_dup_request_exists_wait(char *module_name, bool wait, int *dup_ret) +{ + struct kmod_dup_req *kmod_req, *new_kmod_req; + int ret; + + /* + * Pre-allocate the entry in case we have to use it later + * to avoid contention with the mutex. + */ + new_kmod_req = kzalloc(sizeof(*new_kmod_req), GFP_KERNEL); + if (!new_kmod_req) + return false; + + memcpy(new_kmod_req->name, module_name, strlen(module_name)); + INIT_WORK(&new_kmod_req->complete_work, kmod_dup_request_complete); + INIT_DELAYED_WORK(&new_kmod_req->delete_work, kmod_dup_request_delete); + init_completion(&new_kmod_req->first_req_done); + + mutex_lock(&kmod_dup_mutex); + + kmod_req = kmod_dup_request_lookup(module_name); + if (!kmod_req) { + /* + * If the first request that came through for a module + * was with request_module_nowait() we cannot wait for it + * and share its return value with other users which may + * have used request_module() and need a proper return value + * so just skip using them as an anchor. + * + * If a prior request to this one came through with + * request_module() though, then a request_module_nowait() + * would benefit from duplicate detection. + */ + if (!wait) { + kfree(new_kmod_req); + pr_debug("New request_module_nowait() for %s -- cannot track duplicates for this request\n", module_name); + mutex_unlock(&kmod_dup_mutex); + return false; + } + + /* + * There was no duplicate, just add the request so we can + * keep tab on duplicates later. + */ + pr_debug("New request_module() for %s\n", module_name); + list_add_rcu(&new_kmod_req->list, &dup_kmod_reqs); + mutex_unlock(&kmod_dup_mutex); + return false; + } + mutex_unlock(&kmod_dup_mutex); + + /* We are dealing with a duplicate request now */ + kfree(new_kmod_req); + + /* + * To fix these try to use try_then_request_module() instead as that + * will check if the component you are looking for is present or not. + * You could also just queue a single request to load the module once, + * instead of having each and everything you need try to request for + * the module. + * + * Duplicate request_module() calls can cause quite a bit of wasted + * vmalloc() space when racing with userspace. + */ + if (enable_dups_trace) + WARN(1, "module-autoload: duplicate request for module %s\n", module_name); + else + pr_warn("module-autoload: duplicate request for module %s\n", module_name); + + if (!wait) { + /* + * If request_module_nowait() was used then the user just + * wanted to issue the request and if another module request + * was already its way with the same name we don't care for + * the return value either. Let duplicate request_module_nowait() + * calls bail out right away. + */ + *dup_ret = 0; + return true; + } + + /* + * If a duplicate request_module() was used they *may* care for + * the return value, so we have no other option but to wait for + * the first caller to complete. If the first caller used + * the request_module_nowait() call, subsquent callers will + * deal with the comprmise of getting a successful call with this + * optimization enabled ... + */ + ret = wait_for_completion_state(&kmod_req->first_req_done, + TASK_UNINTERRUPTIBLE | TASK_KILLABLE); + if (ret) { + *dup_ret = ret; + return true; + } + + /* Now the duplicate request has the same exact return value as the first request */ + *dup_ret = kmod_req->dup_ret; + + return true; +} + +void kmod_dup_request_announce(char *module_name, int ret) +{ + struct kmod_dup_req *kmod_req; + + mutex_lock(&kmod_dup_mutex); + + kmod_req = kmod_dup_request_lookup(module_name); + if (!kmod_req) + goto out; + + kmod_req->dup_ret = ret; + + /* + * If we complete() here we may allow duplicate threads + * to continue before the first one that submitted the + * request. We're in no rush also, given that each and + * every bounce back to userspace is slow we avoid that + * with a slight delay here. So queueue up the completion + * and let duplicates suffer, just wait a tad bit longer. + * There is no rush. But we also don't want to hold the + * caller up forever or introduce any boot delays. + */ + queue_work(system_wq, &kmod_req->complete_work); + +out: + mutex_unlock(&kmod_dup_mutex); +} diff --git a/kernel/module/internal.h b/kernel/module/internal.h index 1fd75dd346dc..d67682489d14 100644 --- a/kernel/module/internal.h +++ b/kernel/module/internal.h @@ -3,6 +3,7 @@ * * Copyright (C) 2012 Red Hat, Inc. All Rights Reserved. * Written by David Howells (dhowells@redhat.com) + * Copyright (C) 2023 Luis Chamberlain */ #include @@ -221,6 +222,20 @@ static inline void mod_stat_bump_becoming(struct load_info *info, int flags) #endif /* CONFIG_MODULE_STATS */ +#ifdef CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS +bool kmod_dup_request_exists_wait(char *module_name, bool wait, int *dup_ret); +void kmod_dup_request_announce(char *module_name, int ret); +#else +static inline bool kmod_dup_request_exists_wait(char *module_name, bool wait, int *dup_ret) +{ + return false; +} + +static inline void kmod_dup_request_announce(char *module_name, int ret) +{ +} +#endif + #ifdef CONFIG_MODULE_UNLOAD_TAINT_TRACKING struct mod_unload_taint { struct list_head list; diff --git a/kernel/module/kmod.c b/kernel/module/kmod.c index 5899083436a3..0800d9891692 100644 --- a/kernel/module/kmod.c +++ b/kernel/module/kmod.c @@ -1,6 +1,9 @@ /* * kmod - the kernel module loader + * + * Copyright (C) 2023 Luis Chamberlain */ + #include #include #include @@ -27,6 +30,7 @@ #include #include +#include "internal.h" /* * Assuming: @@ -65,7 +69,7 @@ static void free_modprobe_argv(struct subprocess_info *info) kfree(info->argv); } -static int call_modprobe(char *module_name, int wait) +static int call_modprobe(char *orig_module_name, int wait) { struct subprocess_info *info; static char *envp[] = { @@ -74,12 +78,14 @@ static int call_modprobe(char *module_name, int wait) "PATH=/sbin:/usr/sbin:/bin:/usr/bin", NULL }; + char *module_name; + int ret; char **argv = kmalloc(sizeof(char *[5]), GFP_KERNEL); if (!argv) goto out; - module_name = kstrdup(module_name, GFP_KERNEL); + module_name = kstrdup(orig_module_name, GFP_KERNEL); if (!module_name) goto free_argv; @@ -94,13 +100,16 @@ static int call_modprobe(char *module_name, int wait) if (!info) goto free_module_name; - return call_usermodehelper_exec(info, wait | UMH_KILLABLE); + ret = call_usermodehelper_exec(info, wait | UMH_KILLABLE); + kmod_dup_request_announce(orig_module_name, ret); + return ret; free_module_name: kfree(module_name); free_argv: kfree(argv); out: + kmod_dup_request_announce(orig_module_name, -ENOMEM); return -ENOMEM; } @@ -124,7 +133,7 @@ int __request_module(bool wait, const char *fmt, ...) { va_list args; char module_name[MODULE_NAME_LEN]; - int ret; + int ret, dup_ret; /* * We don't allow synchronous module loading from async. Module @@ -156,8 +165,14 @@ int __request_module(bool wait, const char *fmt, ...) trace_module_request(module_name, wait, _RET_IP_); + if (kmod_dup_request_exists_wait(module_name, wait, &dup_ret)) { + ret = dup_ret; + goto out; + } + ret = call_modprobe(module_name, wait ? UMH_WAIT_PROC : UMH_WAIT_EXEC); +out: up(&kmod_concurrent_max); return ret; -- cgit