x86-64: Stack protector and percpu improvements
Subsystems (8)
v4 cover x86-64: Stack protector and percpu improvements
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
v4:
- Updated to current tip tree
- Added two new cleanups made possible by the removal of IA-64.
- Small improvements to the objtool conversion code.
Brian Gerst (16):
x86/stackprotector/32: Remove stack protector test script x86/stackprotector/64: Remove stack protector test script x86/boot: Disable stack protector for early boot code x86/pvh: Use fixed_percpu_data for early boot GSBASE x86/relocs: Handle R_X86_64_REX_GOTPCRELX relocations
objtool: Allow adding relocations to an existing section
objtool: Convert fixed location stack protector accesses x86/stackprotector/64: Convert to normal percpu variable x86/percpu/64: Use relative percpu offsets x86/percpu/64: Remove fixed_percpu_data x86/boot/64: Remove inverse relocations x86/percpu/64: Remove INIT_PER_CPU macros
percpu: Remove PER_CPU_FIRST_SECTION
percpu: Remove PERCPU_VADDR()
percpu: Remove __per_cpu_load
kallsyms: Remove KALLSYMS_ABSOLUTE_PERCPU
arch/x86/Kconfig | 16 +-- arch/x86/Makefile | 21 ++-- arch/x86/boot/compressed/misc.c | 14 +-- arch/x86/entry/entry_64.S | 2 +- arch/x86/include/asm/desc.h | 1 - arch/x86/include/asm/percpu.h | 22 ---- arch/x86/include/asm/processor.h | 28 +---- arch/x86/include/asm/stackprotector.h | 36 +----- arch/x86/kernel/Makefile | 2 + arch/x86/kernel/asm-offsets_64.c | 6 - arch/x86/kernel/cpu/common.c | 9 +- arch/x86/kernel/head64.c | 2 +- arch/x86/kernel/head_64.S | 20 ++- arch/x86/kernel/irq_64.c | 1 - arch/x86/kernel/setup_percpu.c | 12 +- arch/x86/kernel/vmlinux.lds.S | 35 ------ arch/x86/platform/pvh/head.S | 10 +- arch/x86/tools/relocs.c | 143 ++-------------------- arch/x86/xen/xen-head.S | 10 +- include/asm-generic/sections.h | 2 +- include/asm-generic/vmlinux.lds.h | 43 +------ include/linux/percpu-defs.h | 12 -- init/Kconfig | 11 +- kernel/kallsyms.c | 12 +- mm/percpu.c | 4 +- scripts/Makefile.lib | 2 + scripts/gcc-x86_32-has-stack-protector.sh | 8 -- scripts/gcc-x86_64-has-stack-protector.sh | 4 - scripts/kallsyms.c | 80 +++--------- scripts/link-vmlinux.sh | 4 - tools/objtool/arch/x86/decode.c | 46 +++++++ tools/objtool/arch/x86/special.c | 91 ++++++++++++++ tools/objtool/builtin-check.c | 9 +- tools/objtool/check.c | 14 ++- tools/objtool/elf.c | 133 ++++++++++++++++---- tools/objtool/include/objtool/arch.h | 3 + tools/objtool/include/objtool/builtin.h | 2 + tools/objtool/include/objtool/elf.h | 90 +++++++++++--- 38 files changed, 442 insertions(+), 518 deletions(-) delete mode 100755 scripts/gcc-x86_32-has-stack-protector.sh delete mode 100755 scripts/gcc-x86_64-has-stack-protector.sh
base-commit: 30052fd948a3b43506c83590eaaada12d1f2dd09
Signature · 1 line
2.44.0
PATCH v4 1/16 x86/stackprotector/32: Remove stack protector test script
Test for compiler support directly in Kconfig.
--- arch/x86/Kconfig | 2 +- scripts/gcc-x86_32-has-stack-protector.sh | 8 -------- 2 files changed, 1 insertion(+), 9 deletions(-) delete mode 100755 scripts/gcc-x86_32-has-stack-protector.sh
arch/x86/Kconfig +1 -1
@@ -412,7 +412,7 @@ config PGTABLE_LEVELS config CC_HAS_SANE_STACKPROTECTOR bool default $(success,$(srctree)/scripts/gcc-x86_64-has-stack-protector.sh $(CC) $(CLANG_FLAGS)) if 64BIT - default $(success,$(srctree)/scripts/gcc-x86_32-has-stack-protector.sh $(CC) $(CLANG_FLAGS)) + default $(cc-option,-mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard) help We have to make sure stack protector is unconditionally disabled if the compiler produces broken code or if it does not let us control
scripts/gcc-x86_32-has-stack-protector.sh deleted +0 -8
@@ -1,8 +0,0 @@ -#!/bin/sh -# SPDX-License-Identifier: GPL-2.0 - -# This requires GCC 8.1 or better. Specifically, we require -# -mstack-protector-guard-reg, added by -# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81708 - -echo "int foo(void) { char X[200]; return 3; }" | $* -S -x c -c -m32 -O0 -fstack-protector -mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard - -o - 2> /dev/null | grep -q "%fs"
Signature · 1 line
2.44.0
PATCH v4 2/16 x86/stackprotector/64: Remove stack protector test script
In reply to Brian GerstThis test for the stack protector was added in 2006 to make sure the compiler had the PR28281 patch applied. With GCC 5.1 being the minimum supported compiler now, it is no longer necessary.
--- arch/x86/Kconfig | 5 ++--- scripts/gcc-x86_64-has-stack-protector.sh | 4 ---- 2 files changed, 2 insertions(+), 7 deletions(-) delete mode 100755 scripts/gcc-x86_64-has-stack-protector.sh
arch/x86/Kconfig +2 -3
@@ -411,12 +411,11 @@ config PGTABLE_LEVELS config CC_HAS_SANE_STACKPROTECTOR bool - default $(success,$(srctree)/scripts/gcc-x86_64-has-stack-protector.sh $(CC) $(CLANG_FLAGS)) if 64BIT + default y if 64BIT default $(cc-option,-mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard) help We have to make sure stack protector is unconditionally disabled if - the compiler produces broken code or if it does not let us control - the segment on 32-bit kernels. + the compiler does not allow control of the segment and symbol. menu "Processor type and features"
scripts/gcc-x86_64-has-stack-protector.sh deleted +0 -4
@@ -1,4 +0,0 @@ -#!/bin/sh -# SPDX-License-Identifier: GPL-2.0 - -echo "int foo(void) { char X[200]; return 3; }" | $* -S -x c -c -m64 -O0 -mcmodel=kernel -fno-PIE -fstack-protector - -o - 2> /dev/null | grep -q "%gs"
Signature · 1 line
2.44.0
PATCH v4 3/16 x86/boot: Disable stack protector for early boot code
In reply to Brian GerstOn 64-bit, this will prevent crashes when the canary access is changed from %gs:40 to %gs:__stack_chk_guard(%rip). RIP-relative addresses from the identity-mapped early boot code will target the wrong address with zero-based percpu. KASLR could then shift that address to an unmapped page causing a crash on boot.
This early boot code runs well before userspace is active and does not need stack protector enabled.
--- arch/x86/kernel/Makefile | 2 ++ 1 file changed, 2 insertions(+)
arch/x86/kernel/Makefile +2 -0
@@ -40,6 +40,8 @@ KMSAN_SANITIZE_sev.o := n KCOV_INSTRUMENT_head$(BITS).o := n KCOV_INSTRUMENT_sev.o := n +CFLAGS_head32.o := -fno-stack-protector +CFLAGS_head64.o := -fno-stack-protector CFLAGS_irq.o := -I $(srctree)/$(src)/../include/asm/trace obj-y += head_$(BITS).o
Signature · 1 line
2.44.0
PATCH v4 4/16 x86/pvh: Use fixed_percpu_data for early boot GSBASE
In reply to Brian GerstInstead of having a private area for the stack canary, use fixed_percpu_data for GSBASE like the native kernel.
--- arch/x86/platform/pvh/head.S | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-)
arch/x86/platform/pvh/head.S +8 -5
@@ -95,10 +95,15 @@ SYM_CODE_START_LOCAL(pvh_start_xen) /* 64-bit entry point. */ .code64 1: - /* Set base address in stack canary descriptor. */ + /* + * Set up GSBASE. + * Note that, on SMP, the boot cpu uses init data section until + * the per cpu areas are set up. + */ mov $MSR_GS_BASE,%ecx - mov $_pa(canary), %eax - xor %edx, %edx + lea INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx + mov %edx, %eax + shr $32, %rdx wrmsr call xen_prepare_pvh
@@ -157,8 +162,6 @@ SYM_DATA_START_LOCAL(gdt_start) SYM_DATA_END_LABEL(gdt_start, SYM_L_LOCAL, gdt_end) .balign 16 -SYM_DATA_LOCAL(canary, .fill 48, 1, 0) - SYM_DATA_START_LOCAL(early_stack) .fill BOOT_STACK_SIZE, 1, 0 SYM_DATA_END_LABEL(early_stack, SYM_L_LOCAL, early_stack_end)
Signature · 1 line
2.44.0
PATCH v4 5/16 x86/relocs: Handle R_X86_64_REX_GOTPCRELX relocations
In reply to Brian GerstClang may produce R_X86_64_REX_GOTPCRELX relocations when redefining the stack protector location. Treat them as another type of PC-relative relocation.
--- arch/x86/tools/relocs.c | 7 +++++++ 1 file changed, 7 insertions(+)
arch/x86/tools/relocs.c +7 -0
@@ -31,6 +31,11 @@ static struct relocs relocs32;static struct relocs relocs32neg;static struct relocs relocs64;#define FMT PRIu64 + +#ifndef R_X86_64_REX_GOTPCRELX +#define R_X86_64_REX_GOTPCRELX 42 +#endif +#else #define FMT PRIu32#endif
@@ -224,6 +229,7 @@ static const char *rel_type(unsigned type)REL_TYPE (R_X86_64_PC16 ),REL_TYPE (R_X86_64_8 ),REL_TYPE (R_X86_64_PC8 ), +REL_TYPE (R_X86_64_REX_GOTPCRELX ), #elseREL_TYPE (R_386_NONE ), REL_TYPE(R_386_32),
@@ -865,6 +871,7 @@ static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,case R_X86_64_PC32 :case R_X86_64_PLT32 : +case R_X86_64_REX_GOTPCRELX :/* * PC relative relocations don't need to be adjusted unless * referencing a percpu symbol.
Signature · 1 line
2.44.0
PATCH v4 6/16 objtool: Allow adding relocations to an existing section
In reply to Brian GerstIn order to add relocations to existing sections (e.g. ".rela.text"), encapsulate the reloc array in a block header to allow chaining blocks to add more relocs without moving and relinking the existing ones. This adds minimal memory overhead, while still being able to easily access the arrays by index.
--- tools/objtool/check.c | 2 +- tools/objtool/elf.c | 99 +++++++++++++++++++++++------ tools/objtool/include/objtool/elf.h | 84 +++++++++++++++++++----- 3 files changed, 148 insertions(+), 37 deletions(-)
tools/objtool/check.c +1 -1
@@ -4421,7 +4421,7 @@ static int validate_ibt_data_reloc(struct objtool_file *file,return 0 ;WARN_FUNC ("data relocation to !ENDBR: %s" , - reloc->sec->base,reloc_offset (reloc), + reloc->block->sec->base,reloc_offset (reloc),offstr (dest->sec, dest->offset));return 1 ;
tools/objtool/elf.c +80 -19
178 lines not on this page · raw message
tools/objtool/include/objtool/elf.h +67 -17
162 lines not on this page · raw message
Signature · 1 line
2.44.0
PATCH v4 7/16 objtool: Convert fixed location stack protector accesses
In reply to Brian GerstOlder versions of GCC fixed the location of the stack protector canary at %gs:40. Use objtool to convert these accesses to normal percpu accesses to __stack_chk_guard.
--- arch/x86/Kconfig | 4 ++ scripts/Makefile.lib | 2 + tools/objtool/arch/x86/decode.c | 46 +++++++++++++ tools/objtool/arch/x86/special.c | 91 +++++++++++++++++++++++++ tools/objtool/builtin-check.c | 9 ++- tools/objtool/check.c | 12 ++++ tools/objtool/elf.c | 34 +++++++-- tools/objtool/include/objtool/arch.h | 3 + tools/objtool/include/objtool/builtin.h | 2 + tools/objtool/include/objtool/elf.h | 6 ++ 10 files changed, 204 insertions(+), 5 deletions(-)
arch/x86/Kconfig +4 -0
@@ -417,6 +417,10 @@ config CC_HAS_SANE_STACKPROTECTOR We have to make sure stack protector is unconditionally disabled if the compiler does not allow control of the segment and symbol. +config STACKPROTECTOR_OBJTOOL + bool + default n + menu "Processor type and features" config SMP
scripts/Makefile.lib +2 -0
@@ -258,6 +258,8 @@ objtool := $(objtree)/tools/objtool/objtool objtool-args-$(CONFIG_HAVE_JUMP_LABEL_HACK) += --hacks=jump_label objtool-args-$(CONFIG_HAVE_NOINSTR_HACK) += --hacks=noinstr objtool-args-$(CONFIG_MITIGATION_CALL_DEPTH_TRACKING) += --hacks=skylake +objtool-args-$(CONFIG_STACKPROTECTOR_OBJTOOL) += --hacks=stackprotector +objtool-args-$(CONFIG_SMP) += --smp objtool-args-$(CONFIG_X86_KERNEL_IBT) += --ibt objtool-args-$(CONFIG_FINEIBT) += --cfi objtool-args-$(CONFIG_FTRACE_MCOUNT_USE_OBJTOOL) += --mcount
tools/objtool/arch/x86/decode.c +46 -0
62 lines not on this page · raw message
tools/objtool/arch/x86/special.c +91 -0
100 lines not on this page · raw message
tools/objtool/builtin-check.c +8 -1
@@ -62,12 +62,17 @@ static int parse_hacks(const struct option *opt, const char *str, int unset) found = true; } +if (!str ||strstr (str,"stackprotector" )) { + opts.hack_stackprotector = true; + found = true; + } +return found ?0 :-1 ; }static const struct option check_options[] = {OPT_GROUP ("Actions:" ), -OPT_CALLBACK_OPTARG ('h' ,"hacks" ,NULL ,NULL ,"jump_label,noinstr,skylake" ,"patch toolchain bugs/limitations" , parse_hacks), +OPT_CALLBACK_OPTARG ('h' ,"hacks" ,NULL ,NULL ,"jump_label,noinstr,skylake,stackprotector" ,"patch toolchain bugs/limitations" , parse_hacks),OPT_BOOLEAN ('i' ,"ibt" , &opts.ibt,"validate and annotate IBT" ),OPT_BOOLEAN ('m' ,"mcount" , &opts.mcount,"annotate mcount/fentry calls for ftrace" ),OPT_BOOLEAN ('n' ,"noinstr" , &opts.noinstr,"validate noinstr rules" ),
@@ -94,6 +99,7 @@ static const struct option check_options[] = {OPT_BOOLEAN (0 ,"sec-address" , &opts.sec_address,"print section addresses in warnings" ),OPT_BOOLEAN (0 ,"stats" , &opts.stats,"print statistics" ),OPT_BOOLEAN ('v' ,"verbose" , &opts.verbose,"verbose warnings" ), +OPT_BOOLEAN (0 ,"smp" , &opts.smp,"building an SMP kernel" ), OPT_END(), };
@@ -133,6 +139,7 @@ static bool opts_valid(void) { if (opts.hack_jump_label || opts.hack_noinstr || + opts.hack_stackprotector || opts.ibt || opts.mcount || opts.noinstr ||
tools/objtool/check.c +12 -0
@@ -1315,6 +1315,11 @@ __weak bool arch_is_embedded_insn(struct symbol *sym)return false; } +__weak int arch_hack_stackprotector (struct objtool_file *file) +{ +return 0 ; +} +static struct reloc *insn_reloc (struct objtool_file *file,struct instruction *insn) {struct reloc *reloc;
@@ -4824,6 +4829,13 @@ int check(struct objtool_file *file) warnings += ret; } +if (opts.hack_stackprotector) { + ret =arch_hack_stackprotector (file); +if (ret <0 ) + goto out; + warnings += ret; + } +free_insns (file);if (opts.verbose)
tools/objtool/elf.c +30 -4
@@ -846,6 +846,32 @@ elf_create_prefix_symbol(struct elf *elf, struct symbol *orig, long size)return sym; } +struct symbol * +elf_create_undef_symbol (struct elf *elf,const char *sym_name) +{ +struct symbol *sym =calloc (1 ,sizeof (*sym)); +char *name =strdup (sym_name); + +if (!sym || !name) { +perror ("malloc" ); +return NULL ; + } + + sym->name = name; + sym->sec =find_section_by_index (elf,0 ); + + sym->sym.st_name =elf_add_string (elf,NULL , name); + sym->sym.st_info =GELF_ST_INFO (STB_GLOBAL ,STT_NOTYPE ); + sym->sym.st_value =0 ; + sym->sym.st_size =0 ; + + sym =__elf_create_symbol (elf, sym); +if (sym) +elf_add_symbol (elf, sym); + +return sym; +} + static struct reloc *elf_init_reloc(struct elf *elf, struct section *rsec, unsigned int reloc_idx, unsigned long offset, struct symbol *sym,
@@ -924,7 +950,7 @@ struct reloc *elf_init_reloc_data_sym(struct elf *elf, struct section *sec, struct symbol *sym, s64 addend) { -if (sym->sec && (sec->sh.sh_flags &SHF_EXECINSTR )) { +if (sym->sec && (sym->sec->sh.sh_flags &SHF_EXECINSTR )) {WARN ("bad call to %s() for text symbol %s" , __func__, sym->name);return NULL ;
@@ -1196,9 +1222,9 @@ struct section *elf_create_section(struct elf *elf, const char *name,return sec; } -static struct section *elf_create_rela_section (struct elf *elf, -struct section *sec, -unsigned int reloc_nr) +struct section *elf_create_rela_section (struct elf *elf, +struct section *sec, +unsigned int reloc_nr) {struct section *rsec;struct reloc_block *block;
tools/objtool/include/objtool/arch.h +3 -0
@@ -28,6 +28,7 @@ enum insn_type {INSN_CLD ,INSN_TRAP ,INSN_ENDBR , +INSN_STACKPROTECTOR , INSN_OTHER, };
@@ -96,4 +97,6 @@ int arch_rewrite_retpolines(struct objtool_file *file);bool arch_pc_relative_reloc (struct reloc *reloc); +int arch_hack_stackprotector (struct objtool_file *file); +#endif /* _ARCH_H */
tools/objtool/include/objtool/builtin.h +2 -0
@@ -13,6 +13,7 @@ struct opts {bool hack_jump_label;bool hack_noinstr;bool hack_skylake; +bool hack_stackprotector;bool ibt;bool mcount;bool noinstr;
@@ -38,6 +39,7 @@ struct opts {bool sec_address;bool stats;bool verbose; +bool smp; };extern struct opts opts;
tools/objtool/include/objtool/elf.h +6 -0
@@ -120,6 +120,10 @@ struct elf *elf_open_read(const char *name, int flags);struct section *elf_create_section (struct elf *elf,const char *name,size_t entsize,unsigned int nr); +struct section *elf_create_rela_section (struct elf *elf, +struct section *sec, +unsigned int reloc_nr); +int elf_extend_rela_section (struct elf *elf,struct section *rsec,int add_relocs);
@@ -130,6 +134,8 @@ struct section *elf_create_section_pair(struct elf *elf, const char *name,struct symbol *elf_create_prefix_symbol (struct elf *elf,struct symbol *orig,long size); +struct symbol *elf_create_undef_symbol (struct elf *elf,const char *sym_name); + struct reloc *elf_init_reloc_text_sym(struct elf *elf, struct section *sec, unsigned long offset, unsigned int reloc_idx,
Signature · 1 line
2.44.0
PATCH v4 8/16 x86/stackprotector/64: Convert to normal percpu variable
In reply to Brian GerstOlder versions of GCC fixed the location of the stack protector canary at %gs:40. This constraint forced the percpu section to be linked at virtual address 0 so that the canary could be the first data object in the percpu section. Supporting the zero-based percpu section requires additional code to handle relocations for RIP-relative references to percpu data, extra complexity to kallsyms, and workarounds for linker bugs due to the use of absolute symbols.
Use compiler options to redefine the stack protector location if supported, otherwise use objtool. This will remove the contraint that the percpu section must be zero-based.
--- arch/x86/Kconfig | 11 ++++---- arch/x86/Makefile | 21 ++++++++++------ arch/x86/entry/entry_64.S | 2 +- arch/x86/include/asm/processor.h | 16 ++---------- arch/x86/include/asm/stackprotector.h | 36 ++++----------------------- arch/x86/kernel/asm-offsets_64.c | 6 ----- arch/x86/kernel/cpu/common.c | 5 +--- arch/x86/kernel/head_64.S | 3 +-- arch/x86/xen/xen-head.S | 3 +-- 9 files changed, 30 insertions(+), 73 deletions(-)
arch/x86/Kconfig +5 -6
@@ -271,7 +271,7 @@ config X86 select HAVE_FUNCTION_ARG_ACCESS_API select HAVE_SETUP_PER_CPU_AREA select HAVE_SOFTIRQ_ON_OWN_STACK - select HAVE_STACKPROTECTOR if CC_HAS_SANE_STACKPROTECTOR + select HAVE_STACKPROTECTOR if X86_64 || CC_HAS_SANE_STACKPROTECTOR select HAVE_STACK_VALIDATION if HAVE_OBJTOOL select HAVE_STATIC_CALL select HAVE_STATIC_CALL_INLINE if HAVE_OBJTOOL
@@ -411,15 +411,14 @@ config PGTABLE_LEVELS config CC_HAS_SANE_STACKPROTECTOR bool - default y if 64BIT + default $(cc-option,-mstack-protector-guard-reg=gs -mstack-protector-guard-symbol=__stack_chk_guard) if 64BIT default $(cc-option,-mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard) - help - We have to make sure stack protector is unconditionally disabled if - the compiler does not allow control of the segment and symbol. config STACKPROTECTOR_OBJTOOL bool - default n + depends on X86_64 && STACKPROTECTOR + default !CC_HAS_SANE_STACKPROTECTOR + prompt "Debug objtool stack protector conversion" if CC_HAS_SANE_STACKPROTECTOR && DEBUG_KERNEL menu "Processor type and features"
arch/x86/Makefile +14 -7
@@ -116,13 +116,7 @@ ifeq ($(CONFIG_X86_32),y) # temporary until string.h is fixed KBUILD_CFLAGS += -ffreestanding - ifeq ($(CONFIG_STACKPROTECTOR),y) - ifeq ($(CONFIG_SMP),y) - KBUILD_CFLAGS += -mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard - else - KBUILD_CFLAGS += -mstack-protector-guard=global - endif - endif + percpu_seg := fs else BITS := 64 UTS_MACHINE := x86_64
@@ -172,6 +166,19 @@ else KBUILD_CFLAGS += -mcmodel=kernel KBUILD_RUSTFLAGS += -Cno-redzone=y KBUILD_RUSTFLAGS += -Ccode-model=kernel + + percpu_seg := gs +endif + +ifeq ($(CONFIG_STACKPROTECTOR),y) + ifneq ($(CONFIG_STACKPROTECTOR_OBJTOOL),y) + ifeq ($(CONFIG_SMP),y) + KBUILD_CFLAGS += -mstack-protector-guard-reg=$(percpu_seg) + KBUILD_CFLAGS += -mstack-protector-guard-symbol=__stack_chk_guard + else + KBUILD_CFLAGS += -mstack-protector-guard=global + endif + endif endif #
arch/x86/entry/entry_64.S +1 -1
@@ -191,7 +191,7 @@ SYM_FUNC_START(__switch_to_asm) #ifdef CONFIG_STACKPROTECTOR movq TASK_stack_canary(%rsi), %rbx - movq %rbx, PER_CPU_VAR(fixed_percpu_data + FIXED_stack_canary) + movq %rbx, PER_CPU_VAR(__stack_chk_guard) #endif /*
arch/x86/include/asm/processor.h +2 -14
@@ -387,16 +387,8 @@ struct irq_stack {#ifdef CONFIG_X86_64 struct fixed_percpu_data { -/* -* GCC hardcodes the stack canary as %gs:40. Since the -* irq_stack is the object at %gs:0, we reserve the bottom -* 48 bytes of the irq stack for the canary. -* -* Once we are willing to require -mstack-protector-guard-symbol= -* support for x86_64 stackprotector, we can get rid of this. -*/ char gs_base[40 ]; -unsigned long stack_canary; +unsigned long reserved; }; DECLARE_PER_CPU_FIRST(struct fixed_percpu_data, fixed_percpu_data) __visible;
@@ -411,11 +403,7 @@ extern asmlinkage void entry_SYSCALL32_ignore(void);/* Save actual FS/GS selectors and bases to current->thread */ void current_save_fsgs (void ); -#else /* X86_64 */ -#ifdef CONFIG_STACKPROTECTOR -DECLARE_PER_CPU(unsigned long, __stack_chk_guard); -#endif -#endif /* !X86_64 */ +#endif /* X86_64 */ struct perf_event ;
arch/x86/include/asm/stackprotector.h +5 -31
@@ -2,26 +2,10 @@/* * GCC stack protector support. * -* Stack protector works by putting predefined pattern at the start of +* Stack protector works by putting a predefined pattern at the start of * the stack frame and verifying that it hasn't been overwritten when -* returning from the function. The pattern is called stack canary -* and unfortunately gcc historically required it to be at a fixed offset -* from the percpu segment base. On x86_64, the offset is 40 bytes. -* -* The same segment is shared by percpu area and stack canary. On -* x86_64, percpu symbols are zero based and %gs (64-bit) points to the -* base of percpu area. The first occupant of the percpu area is always -* fixed_percpu_data which contains stack_canary at the appropriate -* offset. On x86_32, the stack canary is just a regular percpu -* variable. -* -* Putting percpu data in %fs on 32-bit is a minor optimization compared to -* using %gs. Since 32-bit userspace normally has %fs == 0, we are likely -* to load 0 into %fs on exit to usermode, whereas with percpu data in -* %gs, we are likely to load a non-null %gs on return to user mode. -* -* Once we are willing to require GCC 8.1 or better for 64-bit stackprotector -* support, we can remove some of this complexity. +* returning from the function. The pattern is called the stack canary +* and is a unique value for each task. */ #ifndef _ASM_STACKPROTECTOR_H
@@ -36,6 +20,8 @@#include <linux/sched.h> +DECLARE_PER_CPU(unsigned long, __stack_chk_guard); +/* * Initialize the stackprotector canary value. *
@@ -51,25 +37,13 @@ static __always_inline void boot_init_stack_canary(void) {unsigned long canary =get_random_canary (); -#ifdef CONFIG_X86_64 -BUILD_BUG_ON (offsetof(struct fixed_percpu_data , stack_canary) !=40 ); -#endif - current->stack_canary = canary; -#ifdef CONFIG_X86_64 -this_cpu_write (fixed_percpu_data.stack_canary, canary); -#else this_cpu_write (__stack_chk_guard, canary); -#endif }static inline void cpu_init_stack_canary (int cpu,struct task_struct *idle) { -#ifdef CONFIG_X86_64 -per_cpu (fixed_percpu_data.stack_canary, cpu) = idle->stack_canary; -#else per_cpu (__stack_chk_guard, cpu) = idle->stack_canary; -#endif }#else /* STACKPROTECTOR */
arch/x86/kernel/asm-offsets_64.c +0 -6
@@ -54,11 +54,5 @@ int main(void)BLANK ();#undef ENTRY -BLANK (); - -#ifdef CONFIG_STACKPROTECTOR -OFFSET (FIXED_stack_canary, fixed_percpu_data, stack_canary); -BLANK (); -#endif return 0 ; }
arch/x86/kernel/cpu/common.c +1 -4
@@ -2063,16 +2063,13 @@ void syscall_init(void)if (!cpu_feature_enabled (X86_FEATURE_FRED ))idt_syscall_init (); } - -#else /* CONFIG_X86_64 */ +#endif /* CONFIG_X86_64 */ #ifdef CONFIG_STACKPROTECTOR DEFINE_PER_CPU(unsigned long, __stack_chk_guard);EXPORT_PER_CPU_SYMBOL (__stack_chk_guard);#endif -#endif /* CONFIG_X86_64 */ -/* * Clear all 6 debug registers: */
arch/x86/kernel/head_64.S +1 -2
@@ -361,8 +361,7 @@ SYM_INNER_LABEL(common_startup_64, SYM_L_LOCAL) /* Set up %gs. * - * The base of %gs always points to fixed_percpu_data. If the - * stack protector canary is enabled, it is located at %gs:40. + * The base of %gs always points to fixed_percpu_data. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */
arch/x86/xen/xen-head.S +1 -2
@@ -53,8 +53,7 @@ SYM_CODE_START(startup_xen) /* Set up %gs. * - * The base of %gs always points to fixed_percpu_data. If the - * stack protector canary is enabled, it is located at %gs:40. + * The base of %gs always points to fixed_percpu_data. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */
Signature · 1 line
2.44.0
PATCH v4 9/16 x86/percpu/64: Use relative percpu offsets
In reply to Brian GerstThe percpu section is currently linked at virtual address 0, because older compilers hardcoded the stack protector canary value at a fixed offset from the start of the GS segment. Now that the canary is a normal percpu variable, the percpu section does not need to be linked at a specific virtual address. This means that x86-64 will calculate percpu offsets like most other architectures, as the delta between the initial percpu address and the dynamically allocated memory.
--- arch/x86/include/asm/processor.h | 6 +++++- arch/x86/kernel/head_64.S | 19 +++++++++---------- arch/x86/kernel/setup_percpu.c | 12 ++---------- arch/x86/kernel/vmlinux.lds.S | 29 +---------------------------- arch/x86/platform/pvh/head.S | 5 ++--- arch/x86/tools/relocs.c | 10 +++------- arch/x86/xen/xen-head.S | 9 ++++----- init/Kconfig | 2 +- 8 files changed, 27 insertions(+), 65 deletions(-)
arch/x86/include/asm/processor.h +5 -1
@@ -396,7 +396,11 @@ DECLARE_INIT_PER_CPU(fixed_percpu_data);static inline unsigned long cpu_kernelmode_gs_base (int cpu) { -return (unsigned long )per_cpu (fixed_percpu_data.gs_base, cpu); +#ifdef CONFIG_SMP +return per_cpu_offset (cpu); +#else +return 0 ; +#endif } extern asmlinkage void entry_SYSCALL32_ignore(void);
arch/x86/kernel/head_64.S +9 -10
@@ -68,11 +68,14 @@ SYM_CODE_START_NOALIGN(startup_64) /* Set up the stack for verify_cpu() */ leaq __top_init_kernel_stack(%rip), %rsp - /* Setup GSBASE to allow stack canary access for C code */ + /* + * Set up GSBASE. + * Note that, on SMP, the boot cpu uses init data section until + * the per cpu areas are set up. + */ movl $MSR_GS_BASE, %ecx - leaq INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx - movl %edx, %eax - shrq $32, %rdx + xorl %eax, %eax + xorl %edx, %edx wrmsr call startup_64_setup_gdt_idt
@@ -359,16 +362,12 @@ SYM_INNER_LABEL(common_startup_64, SYM_L_LOCAL) movl %eax,%fs movl %eax,%gs - /* Set up %gs. - * - * The base of %gs always points to fixed_percpu_data. + /* + * Set up GSBASE. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */ movl $MSR_GS_BASE,%ecx -#ifndef CONFIG_SMP - leaq INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx -#endif movl %edx, %eax shrq $32, %rdx wrmsr
arch/x86/kernel/setup_percpu.c +2 -10
@@ -23,18 +23,10 @@#include <asm/cpumask.h> #include <asm/cpu.h> -#ifdef CONFIG_X86_64 -#define BOOT_PERCPU_OFFSET ((unsigned long)__per_cpu_load) -#else -#define BOOT_PERCPU_OFFSET 0 -#endif - -DEFINE_PER_CPU_READ_MOSTLY(unsigned long, this_cpu_off) = BOOT_PERCPU_OFFSET; +DEFINE_PER_CPU_READ_MOSTLY(unsigned long, this_cpu_off);EXPORT_PER_CPU_SYMBOL (this_cpu_off); -unsigned long __per_cpu_offset[NR_CPUS ]__ro_after_init = { - [0 ...NR_CPUS -1 ] =BOOT_PERCPU_OFFSET , -}; +unsigned long __per_cpu_offset[NR_CPUS ]__ro_after_init ;EXPORT_SYMBOL (__per_cpu_offset);/*
arch/x86/kernel/vmlinux.lds.S +1 -28
@@ -99,12 +99,6 @@ const_pcpu_hot = pcpu_hot; PHDRS { text PT_LOAD FLAGS(5); /* R_E */ data PT_LOAD FLAGS(6); /* RW_ */ -#ifdef CONFIG_X86_64 -#ifdef CONFIG_SMP - percpu PT_LOAD FLAGS(6); /* RW_ */ -#endif - init PT_LOAD FLAGS(7); /* RWE */ -#endif note PT_NOTE FLAGS(0); /* ___ */ }
@@ -222,21 +216,7 @@ SECTIONS __init_begin = .; /* paired with __init_end */ } -#if defined(CONFIG_X86_64) && defined(CONFIG_SMP) - /* - * percpu offsets are zero-based on SMP. PERCPU_VADDR() changes the - * output PHDR, so the next output section - .init.text - should - * start another segment - init. - */ - PERCPU_VADDR(INTERNODE_CACHE_BYTES, 0, :percpu) - ASSERT(SIZEOF(.data..percpu) < CONFIG_PHYSICAL_START, - "per-CPU data too large - increase CONFIG_PHYSICAL_START") -#endif - INIT_TEXT_SECTION(PAGE_SIZE) -#ifdef CONFIG_X86_64 - :init -#endif /* * Section for code used exclusively before alternatives are run. All
@@ -353,9 +333,7 @@ SECTIONS EXIT_DATA } -#if !defined(CONFIG_X86_64) || !defined(CONFIG_SMP) PERCPU_SECTION(INTERNODE_CACHE_BYTES) -#endif . = ALIGN(PAGE_SIZE);
@@ -493,16 +471,11 @@ SECTIONS * Per-cpu symbols which need to be offset from __per_cpu_load * for the boot processor. */ -#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) + __per_cpu_load +#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) INIT_PER_CPU(gdt_page); INIT_PER_CPU(fixed_percpu_data); INIT_PER_CPU(irq_stack_backing_store); -#ifdef CONFIG_SMP -. = ASSERT((fixed_percpu_data == 0), - "fixed_percpu_data is not at start of per-cpu area"); -#endif - #ifdef CONFIG_MITIGATION_UNRET_ENTRY . = ASSERT((retbleed_return_thunk & 0x3f) == 0, "retbleed_return_thunk not cacheline-aligned"); #endif
arch/x86/platform/pvh/head.S +2 -3
@@ -101,9 +101,8 @@ SYM_CODE_START_LOCAL(pvh_start_xen) * the per cpu areas are set up. */ mov $MSR_GS_BASE,%ecx - lea INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx - mov %edx, %eax - shr $32, %rdx + xor %eax, %eax + xor %edx, %edx wrmsr call xen_prepare_pvh
arch/x86/tools/relocs.c +3 -7
@@ -839,12 +839,7 @@ static void percpu_init(void)*/ static int is_percpu_sym (ElfW (Sym) *sym,const char *symname) { -int shndx =sym_index (sym); - -return (shndx == per_cpu_shndx) && -strcmp (symname,"__init_begin" ) && -strcmp (symname,"__per_cpu_load" ) && -strncmp (symname,"init_per_cpu_" ,13 ); +return 0 ; }
@@ -1068,7 +1063,8 @@ static int cmp_relocs(const void *va, const void *vb)static void sort_relocs (struct relocs *r) { -qsort (r->offset, r->count,sizeof (r->offset[0 ]), cmp_relocs); +if (r->count) +qsort (r->offset, r->count,sizeof (r->offset[0 ]), cmp_relocs); }static int write32 (uint32_t v,FILE *f)
arch/x86/xen/xen-head.S +4 -5
@@ -51,15 +51,14 @@ SYM_CODE_START(startup_xen) leaq __top_init_kernel_stack(%rip), %rsp - /* Set up %gs. - * - * The base of %gs always points to fixed_percpu_data. + /* + * Set up GSBASE. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */ movl $MSR_GS_BASE,%ecx - movq $INIT_PER_CPU_VAR(fixed_percpu_data),%rax - cdq + xorl %eax, %eax + xorl %edx, %edx wrmsr mov %rsi, %rdi
init/Kconfig +1 -1
@@ -1784,7 +1784,7 @@ config KALLSYMS_ALL config KALLSYMS_ABSOLUTE_PERCPU bool depends on KALLSYMS - default X86_64 && SMP + default n config KALLSYMS_BASE_RELATIVE bool
Signature · 1 line
2.44.0
PATCH v4 10/16 x86/percpu/64: Remove fixed_percpu_data
In reply to Brian GerstNow that the stack protector canary value is a normal percpu variable, fixed_percpu_data is unused and can be removed.
--- arch/x86/include/asm/processor.h | 8 -------- arch/x86/kernel/cpu/common.c | 4 ---- arch/x86/kernel/vmlinux.lds.S | 1 - arch/x86/tools/relocs.c | 1 - 4 files changed, 14 deletions(-)
arch/x86/include/asm/processor.h +0 -8
@@ -386,14 +386,6 @@ struct irq_stack { }__aligned(IRQ_STACK_SIZE) ;#ifdef CONFIG_X86_64 -struct fixed_percpu_data { -char gs_base[40 ]; -unsigned long reserved; -}; - -DECLARE_PER_CPU_FIRST(struct fixed_percpu_data, fixed_percpu_data) __visible; -DECLARE_INIT_PER_CPU (fixed_percpu_data); -static inline unsigned long cpu_kernelmode_gs_base (int cpu) {#ifdef CONFIG_SMP
arch/x86/kernel/cpu/common.c +0 -4
@@ -1997,10 +1997,6 @@ EXPORT_PER_CPU_SYMBOL(pcpu_hot);EXPORT_PER_CPU_SYMBOL (const_pcpu_hot);#ifdef CONFIG_X86_64 -DEFINE_PER_CPU_FIRST(struct fixed_percpu_data, - fixed_percpu_data)__aligned(PAGE_SIZE) __visible; -EXPORT_PER_CPU_SYMBOL_GPL (fixed_percpu_data); -static void wrmsrl_cstar (unsigned long val) {/*
arch/x86/kernel/vmlinux.lds.S +0 -1
@@ -473,7 +473,6 @@ SECTIONS */ #define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) INIT_PER_CPU(gdt_page); -INIT_PER_CPU(fixed_percpu_data); INIT_PER_CPU(irq_stack_backing_store); #ifdef CONFIG_MITIGATION_UNRET_ENTRY
arch/x86/tools/relocs.c +0 -1
@@ -834,7 +834,6 @@ static void percpu_init(void)* __per_cpu_load * * The "gold" linker incorrectly associates: -* init_per_cpu__fixed_percpu_data * init_per_cpu__gdt_page */ static int is_percpu_sym (ElfW (Sym) *sym,const char *symname)
Signature · 1 line
2.44.0
PATCH v4 11/16 x86/boot/64: Remove inverse relocations
In reply to Brian GerstInverse relocations were needed to offset the effects of relocation for RIP-relative accesses to zero-based percpu data. Now that the percpu section is linked normally as part of the kernel image, they are no longer needed.
--- arch/x86/boot/compressed/misc.c | 14 +--- arch/x86/tools/relocs.c | 126 +------------------------------- 2 files changed, 2 insertions(+), 138 deletions(-)
arch/x86/boot/compressed/misc.c +1 -13
@@ -235,7 +235,7 @@ static void handle_relocations(void *output, unsigned long output_len,/* * Process relocations: 32 bit relocations first then 64 bit after. -* Three sets of binary relocations are added to the end of the kernel +* Two sets of binary relocations are added to the end of the kernel * before compression. Each relocation table entry is the kernel * address of the location which needs to be updated stored as a * 32-bit value which is sign extended to 64 bits.
@@ -245,8 +245,6 @@ static void handle_relocations(void *output, unsigned long output_len,* kernel bits... * 0 - zero terminator for 64 bit relocations * 64 bit relocation repeated -* 0 - zero terminator for inverse 32 bit relocations -* 32 bit inverse relocation repeated * 0 - zero terminator for 32 bit relocations * 32 bit relocation repeated *
@@ -263,16 +261,6 @@ static void handle_relocations(void *output, unsigned long output_len, *(uint32_t *)ptr += delta; }#ifdef CONFIG_X86_64 -while (*--reloc) { -long extended = *reloc; - extended += map; - - ptr = (unsigned long )extended; -if (ptr < min_addr || ptr > max_addr) -error ("inverse 32-bit relocation outside of kernel!\n" ); - - *(int32_t *)ptr -= delta; - }for (reloc--; *reloc; reloc--) {long extended = *reloc; extended += map;
arch/x86/tools/relocs.c +1 -125
196 lines not on this page · raw message
Signature · 1 line
2.44.0
PATCH v4 12/16 x86/percpu/64: Remove INIT_PER_CPU macros
In reply to Brian GerstNow that the load and link addresses of percpu variables are the same, these macros are no longer necessary.
--- arch/x86/include/asm/desc.h | 1 - arch/x86/include/asm/percpu.h | 22 ---------------------- arch/x86/kernel/head64.c | 2 +- arch/x86/kernel/irq_64.c | 1 - arch/x86/kernel/vmlinux.lds.S | 7 ------- arch/x86/tools/relocs.c | 1 - 6 files changed, 1 insertion(+), 33 deletions(-)
arch/x86/include/asm/desc.h +0 -1
@@ -46,7 +46,6 @@ struct gdt_page { } __attribute__((aligned(PAGE_SIZE))); DECLARE_PER_CPU_PAGE_ALIGNED(struct gdt_page, gdt_page); -DECLARE_INIT_PER_CPU (gdt_page);/* Provide the original GDT */ static inline struct desc_struct *get_cpu_gdt_rw (unsigned int cpu)
arch/x86/include/asm/percpu.h +0 -22
@@ -20,12 +20,6 @@#define PER_CPU_VAR (var) __percpu(var)__percpu_rel -#ifdef CONFIG_X86_64_SMP -#define INIT_PER_CPU_VAR (var) init_per_cpu__##var -#else -#define INIT_PER_CPU_VAR (var) var -#endif -#else /* ...!ASSEMBLY */ #include <linux/build_bug.h>
@@ -101,22 +95,6 @@#define __percpu_arg (x) __percpu_prefix "%" #x#define __force_percpu_arg (x) __force_percpu_prefix "%" #x -/* -* Initialized pointers to per-cpu variables needed for the boot -* processor need to use these macros to get the proper address -* offset from __per_cpu_load on SMP. -* -* There also must be an entry in vmlinux_64.lds.S -*/ -#define DECLARE_INIT_PER_CPU (var) \ - extern typeof(var) init_per_cpu_var(var) - -#ifdef CONFIG_X86_64_SMP -#define init_per_cpu_var (var) init_per_cpu__##var -#else -#define init_per_cpu_var (var) var -#endif -/* For arch-specific code, we can use direct single-insn ops (they * don't give an lvalue though). */
arch/x86/kernel/head64.c +1 -1
@@ -564,7 +564,7 @@ void __head startup_64_setup_gdt_idt(void)void *handler =NULL ;struct desc_ptr startup_gdt_descr = { - .address = (unsigned long )&RIP_REL_REF (init_per_cpu_var (gdt_page.gdt)), + .address = (unsigned long )&RIP_REL_REF (gdt_page.gdt), .size =GDT_SIZE -1 , };
arch/x86/kernel/irq_64.c +0 -1
@@ -26,7 +26,6 @@#include <asm/apic.h> DEFINE_PER_CPU_PAGE_ALIGNED(struct irq_stack, irq_stack_backing_store) __visible; -DECLARE_INIT_PER_CPU (irq_stack_backing_store);#ifdef CONFIG_VMAP_STACK /*
arch/x86/kernel/vmlinux.lds.S +0 -7
@@ -467,13 +467,6 @@ SECTIONS "kernel image bigger than KERNEL_IMAGE_SIZE"); #ifdef CONFIG_X86_64 -/* - * Per-cpu symbols which need to be offset from __per_cpu_load - * for the boot processor. - */ -#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) -INIT_PER_CPU(gdt_page); -INIT_PER_CPU(irq_stack_backing_store); #ifdef CONFIG_MITIGATION_UNRET_ENTRY . = ASSERT((retbleed_return_thunk & 0x3f) == 0, "retbleed_return_thunk not cacheline-aligned");
arch/x86/tools/relocs.c +0 -1
@@ -88,7 +88,6 @@ static const char * const sym_regex_kernel[S_NSYMTYPES] = {"__initramfs_start|" "(jiffies|jiffies_64)|"#if ELF_BITS ==64 -"init_per_cpu__.*|" "__end_rodata_hpage_align|"#endif "__vvar_page|"
Signature · 1 line
2.44.0
PATCH v4 13/16 percpu: Remove PER_CPU_FIRST_SECTION
In reply to Brian Gerstx86-64 was the last user.
--- include/asm-generic/vmlinux.lds.h | 1 - include/linux/percpu-defs.h | 12 ------------ 2 files changed, 13 deletions(-)
include/asm-generic/vmlinux.lds.h +0 -1
@@ -1033,7 +1033,6 @@*/ #define PERCPU_INPUT (cacheline) \ __per_cpu_start = .; \ - *(.data..percpu..first) \ . = ALIGN(PAGE_SIZE); \ *(.data..percpu..page_aligned) \ . = ALIGN(cacheline); \
include/linux/percpu-defs.h +0 -12
@@ -26,13 +26,11 @@#define PER_CPU_SHARED_ALIGNED_SECTION "..shared_aligned"#define PER_CPU_ALIGNED_SECTION "..shared_aligned"#endif -#define PER_CPU_FIRST_SECTION "..first"#else #define PER_CPU_SHARED_ALIGNED_SECTION ""#define PER_CPU_ALIGNED_SECTION "..shared_aligned" -#define PER_CPU_FIRST_SECTION ""#endif
@@ -114,16 +112,6 @@#define DEFINE_PER_CPU (type, name) \ DEFINE_PER_CPU_SECTION(type, name, "") -/* -* Declaration/definition used for per-CPU variables that must come first in -* the set of variables. -*/ -#define DECLARE_PER_CPU_FIRST (type, name) \ - DECLARE_PER_CPU_SECTION(type, name, PER_CPU_FIRST_SECTION) - -#define DEFINE_PER_CPU_FIRST (type, name) \ - DEFINE_PER_CPU_SECTION(type, name, PER_CPU_FIRST_SECTION) -/* * Declaration/definition used for per-CPU variables that must be cacheline * aligned under SMP conditions so that, whilst a particular instance of the
Signature · 1 line
2.44.0
PATCH v4 14/16 percpu: Remove PERCPU_VADDR()
In reply to Brian Gerstx86-64 was the last user.
--- include/asm-generic/vmlinux.lds.h | 41 ++----------------------------- 1 file changed, 2 insertions(+), 39 deletions(-)
include/asm-generic/vmlinux.lds.h +2 -39
@@ -1044,47 +1044,10 @@ __per_cpu_end = .;/** -* PERCPU_VADDR - define output section for percpu area +* PERCPU_SECTION - define output section for percpu area * @cacheline: cacheline size -* @vaddr: explicit base address (optional) -* @phdr: destination PHDR (optional) * -* Macro which expands to output section for percpu area. -* -* @cacheline is used to align subsections to avoid false cacheline -* sharing between subsections for different purposes. -* -* If @vaddr is not blank, it specifies explicit base address and all -* percpu symbols will be offset from the given address. If blank, -* @vaddr always equals @laddr + LOAD_OFFSET. -* -* @phdr defines the output PHDR to use if not blank. Be warned that -* output PHDR is sticky. If @phdr is specified, the next output -* section in the linker script will go there too. @phdr should have -* a leading colon. -* -* Note that this macros defines __per_cpu_load as an absolute symbol. -* If there is no need to put the percpu section at a predetermined -* address, use PERCPU_SECTION. -*/ -#define PERCPU_VADDR (cacheline, vaddr, phdr) \ - __per_cpu_load = .; \ - .data..percpu vaddr : AT(__per_cpu_load - LOAD_OFFSET) { \ - PERCPU_INPUT(cacheline) \ - } phdr \ - . = __per_cpu_load + SIZEOF(.data..percpu); - -/** -* PERCPU_SECTION - define output section for percpu area, simple version -* @cacheline: cacheline size -* -* Align to PAGE_SIZE and outputs output section for percpu area. This -* macro doesn't manipulate @vaddr or @phdr and __per_cpu_load and -* __per_cpu_start will be identical. -* -* This macro is equivalent to ALIGN(PAGE_SIZE); PERCPU_VADDR(@cacheline,,) -* except that __per_cpu_load is defined as a relative symbol against -* .data..percpu which is required for relocatable x86_32 configuration. +* Align to PAGE_SIZE and outputs output section for percpu area. */ #define PERCPU_SECTION (cacheline) \ . = ALIGN(PAGE_SIZE); \
Signature · 1 line
2.44.0
PATCH v4 15/16 percpu: Remove __per_cpu_load
In reply to Brian Gerst__per_cpu_load is now always equal to __per_cpu_start.
--- include/asm-generic/sections.h | 2 +- include/asm-generic/vmlinux.lds.h | 1 - mm/percpu.c | 4 ++-- 3 files changed, 3 insertions(+), 4 deletions(-)
include/asm-generic/sections.h +1 -1
@@ -39,7 +39,7 @@ extern char __init_begin[], __init_end[];extern char _sinittext[], _einittext[];extern char __start_ro_after_init[], __end_ro_after_init[];extern char _end[]; -extern char __per_cpu_load[], __per_cpu_start[], __per_cpu_end[]; +extern char __per_cpu_start[], __per_cpu_end[];extern char __kprobes_text_start[], __kprobes_text_end[];extern char __entry_text_start[], __entry_text_end[];extern char __start_rodata[], __end_rodata[];
include/asm-generic/vmlinux.lds.h +0 -1
@@ -1052,7 +1052,6 @@#define PERCPU_SECTION (cacheline) \ . = ALIGN(PAGE_SIZE); \ .data..percpu : AT(ADDR(.data..percpu) - LOAD_OFFSET) { \ - __per_cpu_load = .; \ PERCPU_INPUT(cacheline) \ }
mm/percpu.c +2 -2
@@ -3153,7 +3153,7 @@ int __init pcpu_embed_first_chunk(size_t reserved_size, size_t dyn_size,continue ; }/* copy and return the unused part */ -memcpy (ptr, __per_cpu_load, ai->static_size); +memcpy (ptr, __per_cpu_start, ai->static_size);pcpu_fc_free (ptr + size_sum, ai->unit_size - size_sum); } }
@@ -3336,7 +3336,7 @@ int __init pcpu_page_first_chunk(size_t reserved_size, pcpu_fc_cpu_to_node_fn_tflush_cache_vmap_early (unit_addr, unit_addr + ai->unit_size);/* copy static data */ -memcpy ((void *)unit_addr, __per_cpu_load, ai->static_size); +memcpy ((void *)unit_addr, __per_cpu_start, ai->static_size); }/* we're ready, commit */
Signature · 1 line
2.44.0
PATCH v4 16/16 kallsyms: Remove KALLSYMS_ABSOLUTE_PERCPU
In reply to Brian Gerstx86-64 was the only user.
--- init/Kconfig | 11 +----- kernel/kallsyms.c | 12 ++----- scripts/kallsyms.c | 80 ++++++++--------------------------------- scripts/link-vmlinux.sh | 4 --- 4 files changed, 18 insertions(+), 89 deletions(-)
init/Kconfig +1 -10
@@ -1781,11 +1781,6 @@ config KALLSYMS_ALL Say N unless you really need all symbols, or kernel live patching. -config KALLSYMS_ABSOLUTE_PERCPU - bool - depends on KALLSYMS - default n - config KALLSYMS_BASE_RELATIVE bool depends on KALLSYMS
@@ -1793,11 +1788,7 @@ config KALLSYMS_BASE_RELATIVE help Instead of emitting them as absolute values in the native word size, emit the symbol references in the kallsyms table as 32-bit entries, - each containing a relative value in the range [base, base + U32_MAX] - or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either - an absolute value in the range [0, S32_MAX] or a relative value in the - range [base, base + S32_MAX], where base is the lowest relative symbol - address encountered in the image. + each containing a relative value in the range [base, base + U32_MAX]. On 64-bit builds, this reduces the size of the address table by 50%, but more importantly, it results in entries whose values are build
kernel/kallsyms.c +2 -10
@@ -151,16 +151,8 @@ unsigned long kallsyms_sym_address(int idx)if (!IS_ENABLED (CONFIG_KALLSYMS_BASE_RELATIVE ))return kallsyms_addresses[idx]; -/* values are unsigned offsets if --absolute-percpu is not in effect */ -if (!IS_ENABLED (CONFIG_KALLSYMS_ABSOLUTE_PERCPU )) -return kallsyms_relative_base + (u32 )kallsyms_offsets[idx]; - -/* ...otherwise, positive offsets are absolute values */ -if (kallsyms_offsets[idx] >=0 ) -return kallsyms_offsets[idx]; - -/* ...and negative offsets are relative to kallsyms_relative_base - 1 */ -return kallsyms_relative_base -1 - kallsyms_offsets[idx]; +/* values are unsigned offsets */ +return kallsyms_relative_base + (u32 )kallsyms_offsets[idx]; }static void cleanup_symbol_name (char *s)
scripts/kallsyms.c +15 -65
157 lines not on this page · raw message
scripts/link-vmlinux.sh +0 -4
@@ -153,10 +153,6 @@ kallsyms() kallsymopt="${kallsymopt} --all-symbols" fi - if is_enabled CONFIG_KALLSYMS_ABSOLUTE_PERCPU; then - kallsymopt="${kallsymopt} --absolute-percpu" - fi - if is_enabled CONFIG_KALLSYMS_BASE_RELATIVE; then kallsymopt="${kallsymopt} --base-relative" fi
Signature · 1 line
2.44.0
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago x86_32 simply dropped stack protector support with earlier compilers and IMO, we should follow this example also with x86_64, because:
a) There are currently 5 (soon 6) GCC major releases that support configurable location for stack protector value. GCC 10 is already out of active maintenance, and GCC 7 is considered an ancient release at this time. For x86_32, it was advised to drop the support for stack protector entirely with too old compilers to somehow force users to upgrade the compiler.
b) Stack protector is not a core feature - the kernel will still boot without stack protector. So, if someone really has the urge to use ancient compilers with the bleeding edge kernel, it is still possible to create a bootable image. I do not think using ancient compilers to compile bleeding edge kernels makes any sense at all.
c) Maintenance burden - an objtool feature will have to be maintained until gcc-8.1 is the minimum required compiler version. This feature will IMO be seldom used and thus prone to bitrot.
d) Discrepancy between x86_32 and x86_64 - either both targets should use objtool fixups for stack protector, or none at all. As shown by x86_32 approach in the past, removing stack protector support with ancient compilers was not problematic at all.
That said, the whole series is heartily Acked-by: Uros Bizjak <[email protected]>
[1] https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=e1769bdd4cef522ada32aec863feba41116b183a
[2] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81708
Thanks,
Uros.
Quoted 71 lines
v4:
- Updated to current tip tree
- Added two new cleanups made possible by the removal of IA-64.
- Small improvements to the objtool conversion code.Brian Gerst (16):
x86/stackprotector/32: Remove stack protector test script x86/stackprotector/64: Remove stack protector test script x86/boot: Disable stack protector for early boot code x86/pvh: Usefixed_percpu_datafor early boot GSBASE x86/relocs: HandleR_X86_64_REX_GOTPCRELXrelocations
objtool: Allow adding relocations to an existing section
objtool: Convert fixed location stack protector accesses x86/stackprotector/64: Convert to normal percpu variable x86/percpu/64: Use relative percpu offsets x86/percpu/64: Removefixed_percpu_datax86/boot/64: Remove inverse relocations x86/percpu/64: RemoveINIT_PER_CPUmacros
percpu: RemovePER_CPU_FIRST_SECTION
percpu: RemovePERCPU_VADDR()
percpu: Remove__per_cpu_load
kallsyms: RemoveKALLSYMS_ABSOLUTE_PERCPUarch/x86/Kconfig | 16 +-- arch/x86/Makefile | 21 ++-- arch/x86/boot/compressed/misc.c | 14 +-- arch/x86/entry/entry_64.S | 2 +- arch/x86/include/asm/desc.h | 1 - arch/x86/include/asm/percpu.h | 22 ---- arch/x86/include/asm/processor.h | 28 +---- arch/x86/include/asm/stackprotector.h | 36 +----- arch/x86/kernel/Makefile | 2 + arch/x86/kernel/asm-offsets_64.c | 6 - arch/x86/kernel/cpu/common.c | 9 +- arch/x86/kernel/head64.c | 2 +- arch/x86/kernel/head_64.S | 20 ++- arch/x86/kernel/irq_64.c | 1 - arch/x86/kernel/setup_percpu.c | 12 +- arch/x86/kernel/vmlinux.lds.S | 35 ------ arch/x86/platform/pvh/head.S | 10 +- arch/x86/tools/relocs.c | 143 ++-------------------- arch/x86/xen/xen-head.S | 10 +- include/asm-generic/sections.h | 2 +- include/asm-generic/vmlinux.lds.h | 43 +------ include/linux/percpu-defs.h | 12 -- init/Kconfig | 11 +- kernel/kallsyms.c | 12 +- mm/percpu.c | 4 +- scripts/Makefile.lib | 2 + scripts/gcc-x86_32-has-stack-protector.sh | 8 -- scripts/gcc-x86_64-has-stack-protector.sh | 4 - scripts/kallsyms.c | 80 +++--------- scripts/link-vmlinux.sh | 4 - tools/objtool/arch/x86/decode.c | 46 +++++++ tools/objtool/arch/x86/special.c | 91 ++++++++++++++ tools/objtool/builtin-check.c | 9 +- tools/objtool/check.c | 14 ++- tools/objtool/elf.c | 133 ++++++++++++++++---- tools/objtool/include/objtool/arch.h | 3 + tools/objtool/include/objtool/builtin.h | 2 + tools/objtool/include/objtool/elf.h | 90 +++++++++++--- 38 files changed, 442 insertions(+), 518 deletions(-) delete mode 100755 scripts/gcc-x86_32-has-stack-protector.sh delete mode 100755 scripts/gcc-x86_64-has-stack-protector.shbase-commit:
30052fd948a3b43506c83590eaaada12d1f2dd09--
2.44.0
Quoted 30 lines · Uros Bizjak
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago
x86_32simply dropped stack protector support with earlier compilers and IMO, we should follow this example also withx86_64, because:
Quoted 6 lines
a) There are currently 5 (soon 6) GCC major releases that support configurable location for stack protector value. GCC 10 is already out of active maintenance, and GCC 7 is considered an ancient release at this time. For
x86_32, it was advised to drop the support for stack protector entirely with too old compilers to somehow force users to upgrade the compiler.
At least one developer claimed to be using an older compiler. I asked for more details on why, but got no response.
Quoted 5 lines
b) Stack protector is not a core feature - the kernel will still boot without stack protector. So, if someone really has the urge to use ancient compilers with the bleeding edge kernel, it is still possible to create a bootable image. I do not think using ancient compilers to compile bleeding edge kernels makes any sense at all.
One small issue is that Kconfig would silently disable istackprotector if the compiler doesn't support the new options. That said, the number of people that this would affect is very small, as just about any modern distribution ships a compiler newer than 8.1.
I'm all in favor of only supporting compilers that are supported upstream.
Quoted 3 lines
c) Maintenance burden - an objtool feature will have to be maintained until gcc-8.1 is the minimum required compiler version. This feature will IMO be seldom used and thus prone to bitrot.
That's the reason I added a config option to allow testing objtool even if the compiler has support.
Quoted 4 lines
d) Discrepancy between
x86_32andx86_64- either both targets should use objtool fixups for stack protector, or none at all. As shown byx86_32approach in the past, removing stack protector support with ancient compilers was not problematic at all.
Objtool doesn't support x86-32, mostly because working with REL type relocations is a pain.
Quoted 8 lines
That said, the whole series is heartily Acked-by: Uros Bizjak <[email protected]>
[1] https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=e1769bdd4cef522ada32aec863feba41116b183a
[2] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81708Thanks,
Uros.
Brian Gerst
Quoted 5 lines · Brian Gerst
One small issue is that
Kconfigwould silently disable istackprotector if the compiler doesn't support the new options. That said, the number of people that this would affect is very small, as just about any modern distribution ships a compiler newer than 8.1.
Yes, let's make the rule be that you can still compile the kernel with gcc-5.1+, but you can't get stackprotector support unless you have gcc-8.1+.
I'd hate to add the objtool support for an old compiler - this is a hardening feature, not a core feature, and anybody who insists on old compilers just won't get it.
And we have other cases like this where various debug features depend on the gcc version, eg
config CC_HAS_WORKING_NOSANITIZE_ADDRESS
def_bool !CC_IS_GCC || GCC_VERSION >= 80300so we could easily do the same for stack protector support.
And we might as well also do the semi-yearly compiler version review. We raised the minimum to 4.9 almost four years ago, and then the jump to 5.1 was first for arm64 due to a serious gcc code generation bug and then globally in Sept 2021.
So it's probably time to think about that anyway,
That said, we don't actually have all that many gcc version checks around any more, so I think the jump to 5.1 got rid of the worst of the horrors. Most of the GCC_VERSION checks are either in gcc-plugins (which we should just remove, imnsho - not the version checks, the plugins entirely), or for various random minor details (warnign enablement and the asm goto workaround).
So there doesn't seem to be a major reason to up the versioning, since the stack protector thing can just be disabled for older versions.
But maybe even enterprise distros have upgraded anyway, and we should be proactive.
Cc'ing Arnd, who has historically been one of the people pushing this. He may no longer care because we haven't had huge issues.
Linus
Re: [PATCH v4 01/16] x86/stackprotector/32: Remove stack protector test script
In reply to Brian GerstQuoted 4 lines · Brian Gerst
Test for compiler support directly in
Kconfig.Signed-off-by Brian Gerst <[email protected]>
Quoted 36 lines
--- arch/x86/Kconfig | 2 +- scripts/gcc-x86_32-has-stack-protector.sh | 8 -------- 2 files changed, 1 insertion(+), 9 deletions(-) delete mode 100755 scripts/gcc-x86_32-has-stack-protector.sharch/x86/Kconfig +1
-1@@ -412,7 +412,7 @@ config PGTABLE_LEVELS config CC_HAS_SANE_STACKPROTECTOR bool default $(success,$(srctree)/scripts/gcc-x86_64-has-stack-protector.sh $(CC) $(CLANG_FLAGS)) if 64BIT - default $(success,$(srctree)/scripts/gcc-x86_32-has-stack-protector.sh $(CC) $(CLANG_FLAGS)) + default $(cc-option,-mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard) help We have to make sure stack protector is unconditionally disabled if the compiler produces broken code or if it does not let us controlscripts/gcc-x86_32-has-stack-protector.sh deleted +0
-9@@ -1,8 +0,0 @@ -#!/bin/sh -# SPDX-License-Identifier: GPL-2.0 - -# This requires GCC 8.1 or better. Specifically, we require -# -mstack-protector-guard-reg, added by -# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81708 - -echo "int foo(void) { char X[200]; return 3; }" | $* -S -x c -c -m32 -O0 -fstack-protector -mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard - -o - 2> /dev/null | grep -q "%fs" --2.44.0
Re: [PATCH v4 02/16] x86/stackprotector/64: Remove stack protector test script
In reply to Brian GerstQuoted 6 lines · Brian Gerst
This test for the stack protector was added in 2006 to make sure the compiler had the PR28281 patch applied. With GCC 5.1 being the minimum supported compiler now, it is no longer necessary.
Signed-off-by Brian Gerst <[email protected]>
Quoted 38 lines
--- arch/x86/Kconfig | 5 ++--- scripts/gcc-x86_64-has-stack-protector.sh | 4 ---- 2 files changed, 2 insertions(+), 7 deletions(-) delete mode 100755 scripts/gcc-x86_64-has-stack-protector.sharch/x86/Kconfig +2
-3@@ -411,12 +411,11 @@ config PGTABLE_LEVELS config CC_HAS_SANE_STACKPROTECTOR bool - default $(success,$(srctree)/scripts/gcc-x86_64-has-stack-protector.sh $(CC) $(CLANG_FLAGS)) if 64BIT + default y if 64BIT default $(cc-option,-mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard) help We have to make sure stack protector is unconditionally disabled if - the compiler produces broken code or if it does not let us control - the segment on 32-bit kernels. + the compiler does not allow control of the segment and symbol. menu "Processor type and features"scripts/gcc-x86_64-has-stack-protector.sh deleted +0
-5@@ -1,4 +0,0 @@ -#!/bin/sh -# SPDX-License-Identifier: GPL-2.0 - -echo "int foo(void) { char X[200]; return 3; }" | $* -S -x c -c -m64 -O0 -mcmodel=kernel -fno-PIE -fstack-protector - -o - 2> /dev/null | grep -q "%gs" --2.44.0
Quoted 5 lines · Linus Torvalds
And we might as well also do the semi-yearly compiler version review. We raised the minimum to 4.9 almost four years ago, and then the jump to 5.1 was first for arm64 due to a serious gcc code generation bug and then globally in Sept 2021.
Looking at RHEL, I find a page that claims
RHEL9 : gcc 11.x in app stream
RHEL8 : gcc 8.x or gcc 9.x in app stream.
RHEL7 : gcc 4.8.x
so RHEL7 is already immaterial from a kernel compiler standpoint, and so it looks like at least as far as RHEL is concerned, we could just jump to gcc 8.1 as a minimum.
RHEL also has a "Developer Toolset" that allows you to pick a compiler upgrade, so it's not *quite* as black-and-white as that, but it does seem like we could at some point just pick gcc-8 as a new minimum with very little pain on that front.
The SLES situation seems somewhat similar, with SLES12 being 4.8.x and SLES15 being 7.3. But again with a "Development Tools Module" setup. So that *might* argue for 7.3.
I can't make sense of Debian releases. There's "stable" (bookworm) that comes with gcc-12.2, but there's oldstable, oldoldstable, and various "archived" releases still under LTS. I can't even begin to guess what may be relevant.
I don't think we care that deeply on the kernel side, other than a "maybe we should be a bit more proactive about raising gcc version requirements". I don't think we have any huge issues right now with old gcc versions.
Linus
Re: [PATCH v4 08/16] x86/stackprotector/64: Convert to normal percpu variable
In reply to Brian GerstQuoted 9 lines · Brian Gerst
Older versions of GCC fixed the location of the stack protector canary at %gs:40. This constraint forced the percpu section to be linked at virtual address 0 so that the canary could be the first data object in the percpu section. Supporting the zero-based percpu section requires additional code to handle relocations for RIP-relative references to percpu data, extra complexity to kallsyms, and workarounds for linker bugs due to the use of absolute symbols.
Quoted 5 lines
Use compiler options to redefine the stack protector location if supported, otherwise use objtool. This will remove the contraint that the percpu section must be zero-based.
Signed-off-by Brian Gerst <[email protected]>
HAVE_STACKPROTECTOR change is not needed without STACKPROTECTOR_OBJTOOL support.
With the above remark:
Quoted 270 lines
--- arch/x86/Kconfig | 11 ++++---- arch/x86/Makefile | 21 ++++++++++------ arch/x86/entry/entry_64.S | 2 +- arch/x86/include/asm/processor.h | 16 ++---------- arch/x86/include/asm/stackprotector.h | 36 ++++----------------------- arch/x86/kernel/asm-offsets_64.c | 6 ----- arch/x86/kernel/cpu/common.c | 5 +--- arch/x86/kernel/head_64.S | 3 +-- arch/x86/xen/xen-head.S | 3 +-- 9 files changed, 30 insertions(+), 73 deletions(-)arch/x86/Kconfig +5
-6@@ -271,7 +271,7 @@ config X86 select HAVE_FUNCTION_ARG_ACCESS_API select HAVE_SETUP_PER_CPU_AREA select HAVE_SOFTIRQ_ON_OWN_STACK - select HAVE_STACKPROTECTOR if CC_HAS_SANE_STACKPROTECTOR + select HAVE_STACKPROTECTOR if X86_64 || CC_HAS_SANE_STACKPROTECTOR select HAVE_STACK_VALIDATION if HAVE_OBJTOOL select HAVE_STATIC_CALL select HAVE_STATIC_CALL_INLINE if HAVE_OBJTOOL@@ -411,15 +411,14 @@ config PGTABLE_LEVELS config CC_HAS_SANE_STACKPROTECTOR bool - default y if 64BIT + default $(cc-option,-mstack-protector-guard-reg=gs -mstack-protector-guard-symbol=__stack_chk_guard) if 64BIT default $(cc-option,-mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard) - help - We have to make sure stack protector is unconditionally disabled if - the compiler does not allow control of the segment and symbol. config STACKPROTECTOR_OBJTOOL bool - default n + depends on X86_64 && STACKPROTECTOR + default !CC_HAS_SANE_STACKPROTECTOR + prompt "Debug objtool stack protector conversion" if CC_HAS_SANE_STACKPROTECTOR && DEBUG_KERNEL menu "Processor type and features"arch/x86/Makefile +14
-7@@ -116,13 +116,7 @@ ifeq ($(CONFIG_X86_32),y) # temporary until string.h is fixed KBUILD_CFLAGS += -ffreestanding - ifeq ($(CONFIG_STACKPROTECTOR),y) - ifeq ($(CONFIG_SMP),y) - KBUILD_CFLAGS += -mstack-protector-guard-reg=fs -mstack-protector-guard-symbol=__stack_chk_guard - else - KBUILD_CFLAGS += -mstack-protector-guard=global - endif - endif + percpu_seg := fs else BITS := 64 UTS_MACHINE := x86_64@@ -172,6 +166,19 @@ else KBUILD_CFLAGS += -mcmodel=kernel KBUILD_RUSTFLAGS += -Cno-redzone=y KBUILD_RUSTFLAGS += -Ccode-model=kernel + + percpu_seg := gs +endif + +ifeq ($(CONFIG_STACKPROTECTOR),y) + ifneq ($(CONFIG_STACKPROTECTOR_OBJTOOL),y) + ifeq ($(CONFIG_SMP),y) + KBUILD_CFLAGS += -mstack-protector-guard-reg=$(percpu_seg) + KBUILD_CFLAGS += -mstack-protector-guard-symbol=__stack_chk_guard + else + KBUILD_CFLAGS += -mstack-protector-guard=global + endif + endif endif #arch/x86/entry/entry_64.S +1
-1@@ -191,7 +191,7 @@ SYM_FUNC_START(__switch_to_asm) #ifdef CONFIG_STACKPROTECTOR movq TASK_stack_canary(%rsi), %rbx - movq %rbx, PER_CPU_VAR(fixed_percpu_data + FIXED_stack_canary) + movq %rbx, PER_CPU_VAR(__stack_chk_guard) #endif /*arch/x86/include/asm/processor.h +2
-14@@ -387,16 +387,8 @@ struct irq_stack {#ifdef CONFIG_X86_64 struct fixed_percpu_data { -/* -* GCC hardcodes the stack canary as %gs:40. Since the -* irq_stack is the object at %gs:0, we reserve the bottom -* 48 bytes of the irq stack for the canary. -* -* Once we are willing to require -mstack-protector-guard-symbol= -* support for x86_64 stackprotector, we can get rid of this. -*/ char gs_base[40 ]; -unsigned long stack_canary; +unsigned long reserved; }; DECLARE_PER_CPU_FIRST(struct fixed_percpu_data, fixed_percpu_data) __visible;@@ -411,11 +403,7 @@ extern asmlinkage void entry_SYSCALL32_ignore(void);/* Save actual FS/GS selectors and bases to current->thread */ void current_save_fsgs (void ); -#else /* X86_64 */ -#ifdef CONFIG_STACKPROTECTOR -DECLARE_PER_CPU(unsigned long, __stack_chk_guard); -#endif -#endif /* !X86_64 */ +#endif /* X86_64 */ struct perf_event ;arch/x86/include/asm/stackprotector.h +5
-31@@ -2,26 +2,10 @@/* * GCC stack protector support. * -* Stack protector works by putting predefined pattern at the start of +* Stack protector works by putting a predefined pattern at the start of * the stack frame and verifying that it hasn't been overwritten when -* returning from the function. The pattern is called stack canary -* and unfortunately gcc historically required it to be at a fixed offset -* from the percpu segment base. On x86_64, the offset is 40 bytes. -* -* The same segment is shared by percpu area and stack canary. On -* x86_64, percpu symbols are zero based and %gs (64-bit) points to the -* base of percpu area. The first occupant of the percpu area is always -* fixed_percpu_data which contains stack_canary at the appropriate -* offset. On x86_32, the stack canary is just a regular percpu -* variable. -* -* Putting percpu data in %fs on 32-bit is a minor optimization compared to -* using %gs. Since 32-bit userspace normally has %fs == 0, we are likely -* to load 0 into %fs on exit to usermode, whereas with percpu data in -* %gs, we are likely to load a non-null %gs on return to user mode. -* -* Once we are willing to require GCC 8.1 or better for 64-bit stackprotector -* support, we can remove some of this complexity. +* returning from the function. The pattern is called the stack canary +* and is a unique value for each task. */ #ifndef _ASM_STACKPROTECTOR_H@@ -36,6 +20,8 @@#include <linux/sched.h> +DECLARE_PER_CPU(unsigned long, __stack_chk_guard); +/* * Initialize the stackprotector canary value. * @@ -51,25 +37,13 @@ static __always_inline void boot_init_stack_canary(void) {unsigned long canary =get_random_canary (); -#ifdef CONFIG_X86_64 -BUILD_BUG_ON (offsetof(struct fixed_percpu_data , stack_canary) !=40 ); -#endif - current->stack_canary = canary; -#ifdef CONFIG_X86_64 -this_cpu_write (fixed_percpu_data.stack_canary, canary); -#else this_cpu_write (__stack_chk_guard, canary); -#endif }static inline void cpu_init_stack_canary (int cpu,struct task_struct *idle) { -#ifdef CONFIG_X86_64 -per_cpu (fixed_percpu_data.stack_canary, cpu) = idle->stack_canary; -#else per_cpu (__stack_chk_guard, cpu) = idle->stack_canary; -#endif }#else /* STACKPROTECTOR */ arch/x86/kernel/asm-offsets_64.c +0
-6@@ -54,11 +54,5 @@ int main(void)BLANK ();#undef ENTRY -BLANK (); - -#ifdef CONFIG_STACKPROTECTOR -OFFSET (FIXED_stack_canary, fixed_percpu_data, stack_canary); -BLANK (); -#endif return 0 ; }arch/x86/kernel/cpu/common.c +1
-4@@ -2063,16 +2063,13 @@ void syscall_init(void)if (!cpu_feature_enabled (X86_FEATURE_FRED ))idt_syscall_init (); } - -#else /* CONFIG_X86_64 */ +#endif /* CONFIG_X86_64 */ #ifdef CONFIG_STACKPROTECTOR DEFINE_PER_CPU(unsigned long, __stack_chk_guard);EXPORT_PER_CPU_SYMBOL (__stack_chk_guard);#endif -#endif /* CONFIG_X86_64 */ -/* * Clear all 6 debug registers: */ arch/x86/kernel/head_64.S +1
-2@@ -361,8 +361,7 @@ SYM_INNER_LABEL(common_startup_64, SYM_L_LOCAL) /* Set up %gs. * - * The base of %gs always points to fixed_percpu_data. If the - * stack protector canary is enabled, it is located at %gs:40. + * The base of %gs always points to fixed_percpu_data. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */arch/x86/xen/xen-head.S +1
-3@@ -53,8 +53,7 @@ SYM_CODE_START(startup_xen) /* Set up %gs. * - * The base of %gs always points to fixed_percpu_data. If the - * stack protector canary is enabled, it is located at %gs:40. + * The base of %gs always points to fixed_percpu_data. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */ --2.44.0
Re: [PATCH v4 09/16] x86/percpu/64: Use relative percpu offsets
In reply to Brian GerstQuoted 10 lines · Brian Gerst
The percpu section is currently linked at virtual address 0, because older compilers hardcoded the stack protector canary value at a fixed offset from the start of the GS segment. Now that the canary is a normal percpu variable, the percpu section does not need to be linked at a specific virtual address. This means that x86-64 will calculate percpu offsets like most other architectures, as the delta between the initial percpu address and the dynamically allocated memory.
Signed-off-by Brian Gerst <[email protected]>
Quoted 246 lines
--- arch/x86/include/asm/processor.h | 6 +++++- arch/x86/kernel/head_64.S | 19 +++++++++---------- arch/x86/kernel/setup_percpu.c | 12 ++---------- arch/x86/kernel/vmlinux.lds.S | 29 +---------------------------- arch/x86/platform/pvh/head.S | 5 ++--- arch/x86/tools/relocs.c | 10 +++------- arch/x86/xen/xen-head.S | 9 ++++----- init/Kconfig | 2 +- 8 files changed, 27 insertions(+), 65 deletions(-)arch/x86/include/asm/processor.h +5
-1@@ -396,7 +396,11 @@ DECLARE_INIT_PER_CPU(fixed_percpu_data);static inline unsigned long cpu_kernelmode_gs_base (int cpu) { -return (unsigned long )per_cpu (fixed_percpu_data.gs_base, cpu); +#ifdef CONFIG_SMP +return per_cpu_offset (cpu); +#else +return 0 ; +#endif } extern asmlinkage void entry_SYSCALL32_ignore(void);arch/x86/kernel/head_64.S +9
-10@@ -68,11 +68,14 @@ SYM_CODE_START_NOALIGN(startup_64) /* Set up the stack for verify_cpu() */ leaq __top_init_kernel_stack(%rip), %rsp - /* Setup GSBASE to allow stack canary access for C code */ + /* + * Set up GSBASE. + * Note that, on SMP, the boot cpu uses init data section until + * the per cpu areas are set up. + */ movl $MSR_GS_BASE, %ecx - leaq INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx - movl %edx, %eax - shrq $32, %rdx + xorl %eax, %eax + xorl %edx, %edx wrmsr call startup_64_setup_gdt_idt@@ -359,16 +362,12 @@ SYM_INNER_LABEL(common_startup_64, SYM_L_LOCAL) movl %eax,%fs movl %eax,%gs - /* Set up %gs. - * - * The base of %gs always points to fixed_percpu_data. + /* + * Set up GSBASE. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */ movl $MSR_GS_BASE,%ecx -#ifndef CONFIG_SMP - leaq INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx -#endif movl %edx, %eax shrq $32, %rdx wrmsrarch/x86/kernel/setup_percpu.c +2
-10@@ -23,18 +23,10 @@#include <asm/cpumask.h> #include <asm/cpu.h> -#ifdef CONFIG_X86_64 -#define BOOT_PERCPU_OFFSET ((unsigned long)__per_cpu_load) -#else -#define BOOT_PERCPU_OFFSET 0 -#endif - -DEFINE_PER_CPU_READ_MOSTLY(unsigned long, this_cpu_off) = BOOT_PERCPU_OFFSET; +DEFINE_PER_CPU_READ_MOSTLY(unsigned long, this_cpu_off);EXPORT_PER_CPU_SYMBOL (this_cpu_off); -unsigned long __per_cpu_offset[NR_CPUS ]__ro_after_init = { - [0 ...NR_CPUS -1 ] =BOOT_PERCPU_OFFSET , -}; +unsigned long __per_cpu_offset[NR_CPUS ]__ro_after_init ;EXPORT_SYMBOL (__per_cpu_offset);/* arch/x86/kernel/vmlinux.lds.S +1
-28@@ -99,12 +99,6 @@ const_pcpu_hot = pcpu_hot; PHDRS { text PT_LOAD FLAGS(5); /* R_E */ data PT_LOAD FLAGS(6); /* RW_ */ -#ifdef CONFIG_X86_64 -#ifdef CONFIG_SMP - percpu PT_LOAD FLAGS(6); /* RW_ */ -#endif - init PT_LOAD FLAGS(7); /* RWE */ -#endif note PT_NOTE FLAGS(0); /* ___ */ }@@ -222,21 +216,7 @@ SECTIONS __init_begin = .; /* paired with __init_end */ } -#if defined(CONFIG_X86_64) && defined(CONFIG_SMP) - /* - * percpu offsets are zero-based on SMP. PERCPU_VADDR() changes the - * output PHDR, so the next output section - .init.text - should - * start another segment - init. - */ - PERCPU_VADDR(INTERNODE_CACHE_BYTES, 0, :percpu) - ASSERT(SIZEOF(.data..percpu) < CONFIG_PHYSICAL_START, - "per-CPU data too large - increase CONFIG_PHYSICAL_START") -#endif - INIT_TEXT_SECTION(PAGE_SIZE) -#ifdef CONFIG_X86_64 - :init -#endif /* * Section for code used exclusively before alternatives are run. All@@ -353,9 +333,7 @@ SECTIONS EXIT_DATA } -#if !defined(CONFIG_X86_64) || !defined(CONFIG_SMP) PERCPU_SECTION(INTERNODE_CACHE_BYTES) -#endif . = ALIGN(PAGE_SIZE);@@ -493,16 +471,11 @@ SECTIONS * Per-cpu symbols which need to be offset from __per_cpu_load * for the boot processor. */ -#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) + __per_cpu_load +#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) INIT_PER_CPU(gdt_page); INIT_PER_CPU(fixed_percpu_data); INIT_PER_CPU(irq_stack_backing_store); -#ifdef CONFIG_SMP -. = ASSERT((fixed_percpu_data == 0), - "fixed_percpu_data is not at start of per-cpu area"); -#endif - #ifdef CONFIG_MITIGATION_UNRET_ENTRY . = ASSERT((retbleed_return_thunk & 0x3f) == 0, "retbleed_return_thunk not cacheline-aligned"); #endifarch/x86/platform/pvh/head.S +2
-3@@ -101,9 +101,8 @@ SYM_CODE_START_LOCAL(pvh_start_xen) * the per cpu areas are set up. */ mov $MSR_GS_BASE,%ecx - lea INIT_PER_CPU_VAR(fixed_percpu_data)(%rip), %rdx - mov %edx, %eax - shr $32, %rdx + xor %eax, %eax + xor %edx, %edx wrmsr call xen_prepare_pvharch/x86/tools/relocs.c +3
-7@@ -839,12 +839,7 @@ static void percpu_init(void)*/ static int is_percpu_sym (ElfW (Sym) *sym,const char *symname) { -int shndx =sym_index (sym); - -return (shndx == per_cpu_shndx) && -strcmp (symname,"__init_begin" ) && -strcmp (symname,"__per_cpu_load" ) && -strncmp (symname,"init_per_cpu_" ,13 ); +return 0 ; }@@ -1068,7 +1063,8 @@ static int cmp_relocs(const void *va, const void *vb)static void sort_relocs (struct relocs *r) { -qsort (r->offset, r->count,sizeof (r->offset[0 ]), cmp_relocs); +if (r->count) +qsort (r->offset, r->count,sizeof (r->offset[0 ]), cmp_relocs); }static int write32 (uint32_t v,FILE *f)arch/x86/xen/xen-head.S +4
-5@@ -51,15 +51,14 @@ SYM_CODE_START(startup_xen) leaq __top_init_kernel_stack(%rip), %rsp - /* Set up %gs. - * - * The base of %gs always points to fixed_percpu_data. + /* + * Set up GSBASE. * Note that, on SMP, the boot cpu uses init data section until * the per cpu areas are set up. */ movl $MSR_GS_BASE,%ecx - movq $INIT_PER_CPU_VAR(fixed_percpu_data),%rax - cdq + xorl %eax, %eax + xorl %edx, %edx wrmsr mov %rsi, %rdiinit/Kconfig +1
-2@@ -1784,7 +1784,7 @@ config KALLSYMS_ALL config KALLSYMS_ABSOLUTE_PERCPU bool depends on KALLSYMS - default X86_64 && SMP + default n config KALLSYMS_BASE_RELATIVE bool --2.44.0
Re: [PATCH v4 10/16] x86/percpu/64: Remove fixed_percpu_data
In reply to Brian GerstQuoted 5 lines · Brian Gerst
Now that the stack protector canary value is a normal percpu variable,
fixed_percpu_datais unused and can be removed.Signed-off-by Brian Gerst <[email protected]>
Quoted 68 lines
--- arch/x86/include/asm/processor.h | 8 -------- arch/x86/kernel/cpu/common.c | 4 ---- arch/x86/kernel/vmlinux.lds.S | 1 - arch/x86/tools/relocs.c | 1 - 4 files changed, 14 deletions(-)arch/x86/include/asm/processor.h +0
-8@@ -386,14 +386,6 @@ struct irq_stack { }__aligned(IRQ_STACK_SIZE) ;#ifdef CONFIG_X86_64 -struct fixed_percpu_data { -char gs_base[40 ]; -unsigned long reserved; -}; - -DECLARE_PER_CPU_FIRST(struct fixed_percpu_data, fixed_percpu_data) __visible; -DECLARE_INIT_PER_CPU (fixed_percpu_data); -static inline unsigned long cpu_kernelmode_gs_base (int cpu) {#ifdef CONFIG_SMP arch/x86/kernel/cpu/common.c +0
-4@@ -1997,10 +1997,6 @@ EXPORT_PER_CPU_SYMBOL(pcpu_hot);EXPORT_PER_CPU_SYMBOL (const_pcpu_hot);#ifdef CONFIG_X86_64 -DEFINE_PER_CPU_FIRST(struct fixed_percpu_data, - fixed_percpu_data)__aligned(PAGE_SIZE) __visible; -EXPORT_PER_CPU_SYMBOL_GPL (fixed_percpu_data); -static void wrmsrl_cstar (unsigned long val) {/* arch/x86/kernel/vmlinux.lds.S +0
-1@@ -473,7 +473,6 @@ SECTIONS */ #define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) INIT_PER_CPU(gdt_page); -INIT_PER_CPU(fixed_percpu_data); INIT_PER_CPU(irq_stack_backing_store); #ifdef CONFIG_MITIGATION_UNRET_ENTRYarch/x86/tools/relocs.c +0
-2@@ -834,7 +834,6 @@ static void percpu_init(void)* __per_cpu_load * * The "gold" linker incorrectly associates: -* init_per_cpu__fixed_percpu_data * init_per_cpu__gdt_page */ static int is_percpu_sym (ElfW (Sym) *sym,const char *symname) --2.44.0
Re: [PATCH v4 12/16] x86/percpu/64: Remove INIT_PER_CPU macros
In reply to Brian GerstQuoted 5 lines · Brian Gerst
Now that the load and link addresses of percpu variables are the same, these macros are no longer necessary.
Signed-off-by Brian Gerst <[email protected]>
Quoted 119 lines
--- arch/x86/include/asm/desc.h | 1 - arch/x86/include/asm/percpu.h | 22 ---------------------- arch/x86/kernel/head64.c | 2 +- arch/x86/kernel/irq_64.c | 1 - arch/x86/kernel/vmlinux.lds.S | 7 ------- arch/x86/tools/relocs.c | 1 - 6 files changed, 1 insertion(+), 33 deletions(-)arch/x86/include/asm/desc.h +0
-1@@ -46,7 +46,6 @@ struct gdt_page { } __attribute__((aligned(PAGE_SIZE))); DECLARE_PER_CPU_PAGE_ALIGNED(struct gdt_page, gdt_page); -DECLARE_INIT_PER_CPU (gdt_page);/* Provide the original GDT */ static inline struct desc_struct *get_cpu_gdt_rw (unsigned int cpu)arch/x86/include/asm/percpu.h +0
-22@@ -20,12 +20,6 @@#define PER_CPU_VAR (var) __percpu(var)__percpu_rel -#ifdef CONFIG_X86_64_SMP -#define INIT_PER_CPU_VAR (var) init_per_cpu__##var -#else -#define INIT_PER_CPU_VAR (var) var -#endif -#else /* ...!ASSEMBLY */ #include <linux/build_bug.h> @@ -101,22 +95,6 @@#define __percpu_arg (x) __percpu_prefix "%" #x#define __force_percpu_arg (x) __force_percpu_prefix "%" #x -/* -* Initialized pointers to per-cpu variables needed for the boot -* processor need to use these macros to get the proper address -* offset from __per_cpu_load on SMP. -* -* There also must be an entry in vmlinux_64.lds.S -*/ -#define DECLARE_INIT_PER_CPU (var) \ - extern typeof(var) init_per_cpu_var(var) - -#ifdef CONFIG_X86_64_SMP -#define init_per_cpu_var (var) init_per_cpu__##var -#else -#define init_per_cpu_var (var) var -#endif -/* For arch-specific code, we can use direct single-insn ops (they * don't give an lvalue though). */ arch/x86/kernel/head64.c +1
-1@@ -564,7 +564,7 @@ void __head startup_64_setup_gdt_idt(void)void *handler =NULL ;struct desc_ptr startup_gdt_descr = { - .address = (unsigned long )&RIP_REL_REF (init_per_cpu_var (gdt_page.gdt)), + .address = (unsigned long )&RIP_REL_REF (gdt_page.gdt), .size =GDT_SIZE -1 , };arch/x86/kernel/irq_64.c +0
-1@@ -26,7 +26,6 @@#include <asm/apic.h> DEFINE_PER_CPU_PAGE_ALIGNED(struct irq_stack, irq_stack_backing_store) __visible; -DECLARE_INIT_PER_CPU (irq_stack_backing_store);#ifdef CONFIG_VMAP_STACK /* arch/x86/kernel/vmlinux.lds.S +0
-7@@ -467,13 +467,6 @@ SECTIONS "kernel image bigger than KERNEL_IMAGE_SIZE"); #ifdef CONFIG_X86_64 -/* - * Per-cpu symbols which need to be offset from __per_cpu_load - * for the boot processor. - */ -#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) -INIT_PER_CPU(gdt_page); -INIT_PER_CPU(irq_stack_backing_store); #ifdef CONFIG_MITIGATION_UNRET_ENTRY . = ASSERT((retbleed_return_thunk & 0x3f) == 0, "retbleed_return_thunk not cacheline-aligned");arch/x86/tools/relocs.c +0
-2@@ -88,7 +88,6 @@ static const char * const sym_regex_kernel[S_NSYMTYPES] = {"__initramfs_start|" "(jiffies|jiffies_64)|"#if ELF_BITS ==64 -"init_per_cpu__.*|" "__end_rodata_hpage_align|"#endif "__vvar_page|" --2.44.0
Re: [PATCH v4 13/16] percpu: Remove PER_CPU_FIRST_SECTION
In reply to Brian GerstQuoted 4 lines · Brian Gerst
x86-64 was the last user.
Signed-off-by Brian Gerst <[email protected]>
Quoted 55 lines
--- include/asm-generic/vmlinux.lds.h | 1 - include/linux/percpu-defs.h | 12 ------------ 2 files changed, 13 deletions(-)include/asm-generic/vmlinux.lds.h +0
-1@@ -1033,7 +1033,6 @@*/ #define PERCPU_INPUT (cacheline) \ __per_cpu_start = .; \ - *(.data..percpu..first) \ . = ALIGN(PAGE_SIZE); \ *(.data..percpu..page_aligned) \ . = ALIGN(cacheline); \include/linux/percpu-defs.h +0
-13@@ -26,13 +26,11 @@#define PER_CPU_SHARED_ALIGNED_SECTION "..shared_aligned"#define PER_CPU_ALIGNED_SECTION "..shared_aligned"#endif -#define PER_CPU_FIRST_SECTION "..first"#else #define PER_CPU_SHARED_ALIGNED_SECTION ""#define PER_CPU_ALIGNED_SECTION "..shared_aligned" -#define PER_CPU_FIRST_SECTION ""#endif @@ -114,16 +112,6 @@#define DEFINE_PER_CPU (type, name) \ DEFINE_PER_CPU_SECTION(type, name, "") -/* -* Declaration/definition used for per-CPU variables that must come first in -* the set of variables. -*/ -#define DECLARE_PER_CPU_FIRST (type, name) \ - DECLARE_PER_CPU_SECTION(type, name, PER_CPU_FIRST_SECTION) - -#define DEFINE_PER_CPU_FIRST (type, name) \ - DEFINE_PER_CPU_SECTION(type, name, PER_CPU_FIRST_SECTION) -/* * Declaration/definition used for per-CPU variables that must be cacheline * aligned under SMP conditions so that, whilst a particular instance of the -- 2.44.0
Arnd Bergmann
maintainerQuoted 7 lines · Linus Torvalds
That said, we don't actually have all that many gcc version checks around any more, so I think the jump to 5.1 got rid of the worst of the horrors. Most of the
GCC_VERSIONchecks are either in gcc-plugins (which we should just remove, imnsho - not the version checks, the plugins entirely), or for various random minor details (warnign enablement and the asm goto workaround).
There are also a bunch of feature checks in Kconfig that don't check the version number but instead try if something works.
At least three of the six plugins can be removed with sufficiently new compilers: GCC_PLUGIN_SANCOV after gcc-6, and both GCC_PLUGIN_STRUCTLEAK and GCC_PLUGIN_ARM_SSP_PER_TASK after gcc-12.
Quoted 8 lines
So there doesn't seem to be a major reason to up the versioning, since the stack protector thing can just be disabled for older versions.
But maybe even enterprise distros have upgraded anyway, and we should be proactive.
Cc'ing Arnd, who has historically been one of the people pushing this. He may no longer care because we haven't had huge issues.
I'm not aware of any major issues, but it keeps coming up and a number of people have asked me about it because of smaller ones. Unfortunately I didn't write down what the problems were.
I think based on what compiler versions are shipped by LTS distros, gcc-8.1 is the next logical step when we do it, covering Debian 10, RHEL 8 and Ubuntu 20.04, which are probably the oldest we still need to support.
RHEL 7 and SLES 12 are still technically supported distros as well, but those shipped with gcc-4.8, so we dropped them in 2020 with the move to gcc-4.9.
So in short, not a lot to gain or lose from raising the minimum to 8.1, but it would be nice to reduce the possible test matrix from 10 gcc versions back to the 7 we had in the past, as well as clean up a few version dependencies. Similarly we can probably raise the oldest binutils version to 2.30, as that seems to be the earliest version that was paired with gcc-8 (in RHEL-8).
Arnd
Quoted 21 lines · Uros Bizjak
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
Quoted 7 lines
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago
x86_32simply dropped stack protector support with earlier compilers and IMO, we should follow this example also withx86_64, because:
Ack on raising the minimum version requirement for x86-64 stackprotector to 8.1 or so - this causes no real pain on the distro
side: when *this* new kernel of ours is picked by a distro, it almost always goes hand in hand with a compiler version upgrade.
We should be careful with fixes marked for -stable backport, but other than that, new improvements like Brian's series are a fair game to tweak compiler version requirements.
But please emit a (single) prominent build-time warning if a feature is disabled though, even if there are no functional side-effects, such as for hardening features.
In general distro kernel developers & maintainers like seeing the performance (and other) effects of their compiler version choices, but we are not very transparent about this: our fallbacks are way too opaque right now.
Thanks,
Ingo
Quoted 38 lines · Ingo Molnar
Quoted 21 lines · Uros Bizjak
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
Quoted 7 lines
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago
x86_32simply dropped stack protector support with earlier compilers and IMO, we should follow this example also withx86_64, because:Ack on raising the minimum version requirement for x86-64 stackprotector to 8.1 or so - this causes no real pain on the distro
side: when *this* new kernel of ours is picked by a distro, it almost always goes hand in hand with a compiler version upgrade.
Quoted 7 lines
We should be careful with fixes marked for -stable backport, but other than that, new improvements like Brian's series are a fair game to tweak compiler version requirements.
But please emit a (single) prominent build-time warning if a feature is disabled though, even if there are no functional side-effects, such as for hardening features.
Disabled for any reason or only if the compiler lacks support?
Brian Gerst
Quoted 43 lines · Brian Gerst
Quoted 42 lines · Ingo Molnar
Quoted 21 lines · Uros Bizjak
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
Quoted 7 lines
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago
x86_32simply dropped stack protector support with earlier compilers and IMO, we should follow this example also withx86_64, because:Ack on raising the minimum version requirement for x86-64 stackprotector to 8.1 or so - this causes no real pain on the distro
side: when *this* new kernel of ours is picked by a distro, it almost always goes hand in hand with a compiler version upgrade.We should be careful with fixes marked for -stable backport, but other than that, new improvements like Brian's series are a fair game to tweak compiler version requirements.
Quoted 5 lines
Quoted 3 lines
But please emit a (single) prominent build-time warning if a feature is disabled though, even if there are no functional side-effects, such as for hardening features.
Disabled for any reason or only if the compiler lacks support?
Only if the user desires to have it enabled, but it's not possible due to compiler (or other build environment) reasons. Ie. if something unexpected happens from the user's perspective.
The .config option is preserved even if the compiler doesn't support it, right?
I suspect this should also cover features that get select-ed by a feature that the user enables. (Not sure about architecture level select-ed options.)
Thanks,
Ingo
Quoted 60 lines · Ingo Molnar
Quoted 43 lines · Brian Gerst
Quoted 42 lines · Ingo Molnar
Quoted 21 lines · Uros Bizjak
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
Quoted 7 lines
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago
x86_32simply dropped stack protector support with earlier compilers and IMO, we should follow this example also withx86_64, because:Ack on raising the minimum version requirement for x86-64 stackprotector to 8.1 or so - this causes no real pain on the distro
side: when *this* new kernel of ours is picked by a distro, it almost always goes hand in hand with a compiler version upgrade.We should be careful with fixes marked for -stable backport, but other than that, new improvements like Brian's series are a fair game to tweak compiler version requirements.
Quoted 5 lines
Quoted 3 lines
But please emit a (single) prominent build-time warning if a feature is disabled though, even if there are no functional side-effects, such as for hardening features.
Disabled for any reason or only if the compiler lacks support?
Only if the user desires to have it enabled, but it's not possible due to compiler (or other build environment) reasons. Ie. if something unexpected happens from the user's perspective.
The .config option is preserved even if the compiler doesn't support it, right?
Quoted 7 lines
I suspect this should also cover features that get select-ed by a feature that the user enables. (Not sure about architecture level select-ed options.)
Thanks,
Ingo
I could add something like:
comment "Stack protector is not supported by the architecture or compiler"
depends on !HAVE_STACKPROTECTOR
But, "make oldconfig" will still silently disable stack protector if the compiler doesn't support the new options. It does put the comment into the .config file though, so that may be enough.
Brian Gerst
Quoted 69 lines · Brian Gerst
Quoted 60 lines · Ingo Molnar
Quoted 43 lines · Brian Gerst
Quoted 42 lines · Ingo Molnar
Quoted 21 lines · Uros Bizjak
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
Quoted 7 lines
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago
x86_32simply dropped stack protector support with earlier compilers and IMO, we should follow this example also withx86_64, because:Ack on raising the minimum version requirement for x86-64 stackprotector to 8.1 or so - this causes no real pain on the distro
side: when *this* new kernel of ours is picked by a distro, it almost always goes hand in hand with a compiler version upgrade.We should be careful with fixes marked for -stable backport, but other than that, new improvements like Brian's series are a fair game to tweak compiler version requirements.
Quoted 5 lines
Quoted 3 lines
But please emit a (single) prominent build-time warning if a feature is disabled though, even if there are no functional side-effects, such as for hardening features.
Disabled for any reason or only if the compiler lacks support?
Only if the user desires to have it enabled, but it's not possible due to compiler (or other build environment) reasons. Ie. if something unexpected happens from the user's perspective.
The .config option is preserved even if the compiler doesn't support it, right?
Quoted 7 lines
I suspect this should also cover features that get select-ed by a feature that the user enables. (Not sure about architecture level select-ed options.)
Thanks,
Ingo
Quoted 8 lines
I could add something like:
comment "Stack protector is not supported by the architecture or compiler"
depends on !HAVE_STACKPROTECTORBut, "make oldconfig" will still silently disable stack protector if the compiler doesn't support the new options. It does put the comment into the .config file though, so that may be enough.
So I was thinking more along the lines of emitting an actual warning to the build log, every time the compiler check is executed and fails to detect [certain] essential or good-to-have compiler features.
A bit like the red '[ OFF ]' build lines during the perf build:
Auto-detecting system features:
... dwarf: [ on ] ... dwarf_getlocations: [ on ] ... glibc: [ on ] ... libbfd: [ on ] ... libbfd-buildid: [ on ] ... libcap: [ on ] ... libelf: [ on ] ... libnuma: [ on ] ... numa_num_possible_cpus: [ on ] ... libperl: [ on ] ... libpython: [ on ] ... libcrypto: [ on ] ... libunwind: [ on ] ... libdw-dwarf-unwind: [ on ] ... libcapstone: [ OFF ] <======== ... zlib: [ on ] ... lzma: [ on ] ... get_cpuid: [ on ] ... bpf: [ on ] ... libaio: [ on ] ... libzstd: [ on ]
... or something like that.
Thanks,
Ingo
Quoted 112 lines · Ingo Molnar
Quoted 69 lines · Brian Gerst
Quoted 60 lines · Ingo Molnar
Quoted 43 lines · Brian Gerst
Quoted 42 lines · Ingo Molnar
Quoted 21 lines · Uros Bizjak
Quoted 7 lines · Brian Gerst
Currently, x86-64 uses an unusual percpu layout, where the percpu section is linked at absolute address 0. The reason behind this is that older GCC versions placed the stack protector (if enabled) at a fixed offset from the GS segment base. Since the GS segement is also used for percpu variables, this forced the current layout.
Quoted 5 lines
GCC since version 8.1 supports a configurable location for the stack protector value, which allows removal of the restriction on how the percpu section is linked. This allows the percpu section to be linked normally, like other architectures. In turn, this allows removal of code that was needed to support the zero-based percpu section.
The number of simplifications throughout the code, enabled by this patch set, is really impressive, and it reflects the number of workarounds to enable the feature that was originally not designed for the kernel usage. As noted above, this issue was recognized in the GCC compiler and the stack protector support was generalized by adding configurable location for the stack protector value [1,2].
Quoted 7 lines
The improved stack protector support was implemented in gcc-8.1, released on May 2, 2018, when linux 4.17 was in development. In light of this fact, and 5 (soon 6) GCC major releases later, I'd like to ask if the objtool support to fixup earlier compilers is really necessary. Please note that years ago
x86_32simply dropped stack protector support with earlier compilers and IMO, we should follow this example also withx86_64, because:Ack on raising the minimum version requirement for x86-64 stackprotector to 8.1 or so - this causes no real pain on the distro
side: when *this* new kernel of ours is picked by a distro, it almost always goes hand in hand with a compiler version upgrade.We should be careful with fixes marked for -stable backport, but other than that, new improvements like Brian's series are a fair game to tweak compiler version requirements.
Quoted 5 lines
Quoted 3 lines
But please emit a (single) prominent build-time warning if a feature is disabled though, even if there are no functional side-effects, such as for hardening features.
Disabled for any reason or only if the compiler lacks support?
Only if the user desires to have it enabled, but it's not possible due to compiler (or other build environment) reasons. Ie. if something unexpected happens from the user's perspective.
The .config option is preserved even if the compiler doesn't support it, right?
Quoted 7 lines
I suspect this should also cover features that get select-ed by a feature that the user enables. (Not sure about architecture level select-ed options.)
Thanks,
Ingo
Quoted 8 lines
I could add something like:
comment "Stack protector is not supported by the architecture or compiler"
depends on !HAVE_STACKPROTECTORBut, "make oldconfig" will still silently disable stack protector if the compiler doesn't support the new options. It does put the comment into the .config file though, so that may be enough.
So I was thinking more along the lines of emitting an actual warning to the build log, every time the compiler check is executed and fails to detect [certain] essential or good-to-have compiler features.
A bit like the red '[ OFF ]' build lines during the perf build:
Auto-detecting system features:
... dwarf: [ on ] ... dwarf_getlocations: [ on ] ... glibc: [ on ] ... libbfd: [ on ] ... libbfd-buildid: [ on ] ... libcap: [ on ] ... libelf: [ on ] ... libnuma: [ on ] ... numa_num_possible_cpus: [ on ] ... libperl: [ on ] ... libpython: [ on ] ... libcrypto: [ on ] ... libunwind: [ on ] ... libdw-dwarf-unwind: [ on ] ... libcapstone: [ OFF ] <======== ... zlib: [ on ] ... lzma: [ on ] ... get_cpuid: [ on ] ... bpf: [ on ] ... libaio: [ on ] ... libzstd: [ on ]
Quoted 5 lines
... or something like that.
Thanks,
Ingo
That list comes from the perf tool itself (tools/perf/builtin-version.c), not the kernel config or build system. Something like that could be added to the main kernel build. But it should be a separate patch series as it will likely need a lot of design iteration.
Brian Gerst
Quoted 34 lines · Brian Gerst
Quoted 34 lines
Quoted 3 lines
But, "make oldconfig" will still silently disable stack protector if the compiler doesn't support the new options. It does put the comment into the .config file though, so that may be enough.
So I was thinking more along the lines of emitting an actual warning to the build log, every time the compiler check is executed and fails to detect [certain] essential or good-to-have compiler features.
A bit like the red '[ OFF ]' build lines during the perf build:
Auto-detecting system features:
... dwarf: [ on ] ... dwarf_getlocations: [ on ] ... glibc: [ on ] ... libbfd: [ on ] ... libbfd-buildid: [ on ] ... libcap: [ on ] ... libelf: [ on ] ... libnuma: [ on ] ... numa_num_possible_cpus: [ on ] ... libperl: [ on ] ... libpython: [ on ] ... libcrypto: [ on ] ... libunwind: [ on ] ... libdw-dwarf-unwind: [ on ] ... libcapstone: [ OFF ] <======== ... zlib: [ on ] ... lzma: [ on ] ... get_cpuid: [ on ] ... bpf: [ on ] ... libaio: [ on ] ... libzstd: [ on ]
Quoted 8 lines
Quoted 5 lines
... or something like that.
Thanks,
Ingo
That list comes from the perf tool itself (
tools/perf/builtin-version.c), not the kernel config or build system.
Yeah, I know, I wrote the initial version. ;-)
( See upstream commits b6aa9979416e~1..4cc9117a35b2 )
Quoted 3 lines
Something like that could be added to the main kernel build. But it should be a separate patch series as it will likely need a lot of design iteration.
Doesn't have to be complicated really, but obviously not a requirement for this series.
Thanks,
Ingo
From: Linus Torvalds
Quoted 10 lines
Sent: 23 March 2024 17:07
Quoted 5 lines · Linus Torvalds
And we might as well also do the semi-yearly compiler version review. We raised the minimum to 4.9 almost four years ago, and then the jump to 5.1 was first for arm64 due to a serious gcc code generation bug and then globally in Sept 2021.
Quoted 5 lines
Looking at RHEL, I find a page that claims
RHEL9 : gcc 11.x in app stream
RHEL8 : gcc 8.x or gcc 9.x in app stream.
RHEL7 : gcc 4.8.x
I'm running Ubuntu 18.04 LTS (and still getting updates) and that has gcc 7.3.0.
Clearly it might be time to update/reinstall that system :-)
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Arnd Bergmann
maintainerQuoted 4 lines · Linus Torvalds
On Sat, 23 Mar 2024 at 09:16, Linus Torvalds <[email protected]> wrote:
The SLES situation seems somewhat similar, with SLES12 being 4.8.x and SLES15 being 7.3. But again with a "Development Tools Module" setup. So that *might* argue for 7.3.
According to https://distrowatch.com/table.php?distribution=sle, they also provide gcc-12.2.1 with the sp5 update, so we're probably fine.
On the other hand, I can see that OpenSUSE Leap 15.6 contains a fairly modern kernel (6.4.x) built with the gcc-7.3 system compiler, and I think this is the same one as in SLES.
Not sure if they plan to update the kernel release beyond that, or how inconvenient it would be for them to require using the other compiler for future updates, so I've added the developers that last touched the OpenSUSE kernel RPM package to Cc here.
Arnd
Quoted 3 lines · Arnd Bergmann
Quoted 1 line · Linus Torvalds
...
Quoted 23 lines
Quoted 8 lines
So there doesn't seem to be a major reason to up the versioning, since the stack protector thing can just be disabled for older versions.
But maybe even enterprise distros have upgraded anyway, and we should be proactive.
Cc'ing Arnd, who has historically been one of the people pushing this. He may no longer care because we haven't had huge issues.
I'm not aware of any major issues, but it keeps coming up and a number of people have asked me about it because of smaller ones. Unfortunately I didn't write down what the problems were.
I think based on what compiler versions are shipped by LTS distros, gcc-8.1 is the next logical step when we do it, covering Debian 10, RHEL 8 and Ubuntu 20.04, which are probably the oldest we still need to support.
RHEL 7 and SLES 12 are still technically supported distros as well, but those shipped with gcc-4.8, so we dropped them in 2020 with the move to gcc-4.9.
Quoted 8 lines
So in short, not a lot to gain or lose from raising the minimum to 8.1, but it would be nice to reduce the possible test matrix from 10 gcc versions back to the 7 we had in the past, as well as clean up a few version dependencies. Similarly we can probably raise the oldest binutils version to 2.30, as that seems to be the earliest version that was paired with gcc-8 (in RHEL-8).
x86_64/SMP uses a pile of hacks to create a runtime relocatable kernel, one of which is a workaround for the offset based addressing of per-CPU variables. This requires RIP-relative per-CPU references, e.g.,
leal %gs:foo(%rip), %reg
to be fixed up in the opposite direction (displacement subtracted rather than added) in the decompressor. This scheme is used because older GCCs can only access the stack protector cookie via a fixed offset of GS+40, and so GS must carry the address of the start of the per-CPU region rather than an arbitrary relative offset between the per-CPU region in vmlinux and the one belonging to a CPU.
GCC 8.1 and later allow the cookie to be specified using a symbol, and this would allow us to revert to the ordinary per-CPU addressing, where the base is the vmlinux copy of a symbol, and each CPU carries a different offset in GS that produces the address of its respective private copy. [0]
With that out of the way, we could get rid of the weird relocation format and just use the linker to link vmlinux in PIE mode (like other architectures), using the condensed RELR format which only takes a fraction of the space. Using PIC codegen and PIE linking also brings us closer to what toolchains expect, and so fewer quirks/surprises when moving to newer versions. (Currently on x86, we do a position dependent link of vmlinux, and rely on the static relocations produced by --emit-relocs to create the metadata we need to perform the relocation fixups. Static relocations cannot describe all the transformations and relaxations that the linker might apply, and so static relocations might go out of sync with the actual code.)
Another minor cleanup is __GCC_ASM_FLAG_OUTPUTS__, which would always be supported if we require 8.1 or newer.
None of this is high priority, though, so not a reason in itself to deprecate GCC 7 and older, more of a nice bonus when we do get there.
Signature · 5 lines
Ard. [0] https://git.kernel.org/pub/scm/linux/kernel/git/ardb/linux.git/log/?h=x86-remove-absolute-percpu
Quoted 14 lines · Arnd Bergmann
Quoted 4 lines · Linus Torvalds
On Sat, 23 Mar 2024 at 09:16, Linus Torvalds <[email protected]> wrote:
The SLES situation seems somewhat similar, with SLES12 being 4.8.x and SLES15 being 7.3. But again with a "Development Tools Module" setup. So that *might* argue for 7.3.According to https://distrowatch.com/table.php?distribution=sle, they also provide gcc-12.2.1 with the sp5 update, so we're probably fine.
On the other hand, I can see that OpenSUSE Leap 15.6 contains a fairly modern kernel (6.4.x) built with the gcc-7.3 system compiler, and I think this is the same one as in SLES.
Quoted 5 lines
Not sure if they plan to update the kernel release beyond that, or how inconvenient it would be for them to require using the other compiler for future updates, so I've added the developers that last touched the OpenSUSE kernel RPM package to Cc here.
SLE15-SP6 kernel (based on 6.4.x) is still built with gcc7, currently gcc 7.5, indeed. openSUSE Leap shares the very same kernel, so it's with gcc 7.5, too. Even though gcc-13 is provided as additional compiler package, it's not used for the kernel package build.
AFAIK, it's not decided yet about SP7 kernel. But since we take a conservative approach for SLE, I guess SLE15-SP7 will be likely sticking with the old gcc, unless forced to change by some reason.
SLE12 is built with the old gcc 4.8, and SLE12-SP5 (based on 4.12) is still actively maintained, but only for a few months until October 2024.
The next generation of SLE is built with the latest gcc (gcc-13 for now). So SLE16 will be a totally different story.
openSUSE Tumbleweed always uses a bleeding edge compiler (gcc-13), too.
Takashi
Arnd Bergmann
maintainerQuoted 8 lines · Takashi Iwai
Quoted 5 lines · Arnd Bergmann
Not sure if they plan to update the kernel release beyond that, or how inconvenient it would be for them to require using the other compiler for future updates, so I've added the developers that last touched the OpenSUSE kernel RPM package to Cc here.
Quoted 4 lines
SLE15-SP6 kernel (based on 6.4.x) is still built with gcc7, currently gcc 7.5, indeed. openSUSE Leap shares the very same kernel, so it's with gcc 7.5, too. Even though gcc-13 is provided as additional compiler package, it's not used for the kernel package build.
Ok, so for SP6 there won't be a problem because it won't update the kernel.
Quoted 3 lines
AFAIK, it's not decided yet about SP7 kernel. But since we take a conservative approach for SLE, I guess SLE15-SP7 will be likely sticking with the old gcc, unless forced to change by some reason.
From https://www.suse.com/support/kb/doc/?id=000019587, it looks like kernel versions are historically only updated for even numbered updates (SP2, SP4, SP6), so I guess there is a good chance that SP7 won't upgrade either, even if that decision is still to be made.
I would assume that if SP7 does get a new kernel, it would not be any later than the next LTS kernel (6.12 according to phb-crystal-ball), and we could just decide to move to gcc-8 only after that is out.
Quoted 3 lines
SLE12 is built with the old gcc 4.8, and SLE12-SP5 (based on 4.12) is still actively maintained, but only for a few months until October 2024.
Right, and it also never updated beyond linux-4.12, which still supported gcc-4.8. gcc-4.9 only became a requirement in linux-5.9.
Quoted 5 lines
The next generation of SLE is built with the latest gcc (gcc-13 for now). So SLE16 will be a totally different story.
openSUSE Tumbleweed always uses a bleeding edge compiler (gcc-13), too.
Thanks for the detailed reply!
Arnd
Quoted 28 lines · Arnd Bergmann
Quoted 8 lines · Takashi Iwai
Quoted 5 lines · Arnd Bergmann
Not sure if they plan to update the kernel release beyond that, or how inconvenient it would be for them to require using the other compiler for future updates, so I've added the developers that last touched the OpenSUSE kernel RPM package to Cc here.
Quoted 4 lines
SLE15-SP6 kernel (based on 6.4.x) is still built with gcc7, currently gcc 7.5, indeed. openSUSE Leap shares the very same kernel, so it's with gcc 7.5, too. Even though gcc-13 is provided as additional compiler package, it's not used for the kernel package build.
Ok, so for SP6 there won't be a problem because it won't update the kernel.
Quoted 3 lines
AFAIK, it's not decided yet about SP7 kernel. But since we take a conservative approach for SLE, I guess SLE15-SP7 will be likely sticking with the old gcc, unless forced to change by some reason.
From https://www.suse.com/support/kb/doc/?id=000019587, it looks like kernel versions are historically only updated for even numbered updates (SP2, SP4, SP6), so I guess there is a good chance that SP7 won't upgrade either, even if that decision is still to be made.
Quoted 4 lines
I would assume that if SP7 does get a new kernel, it would not be any later than the next LTS kernel (6.12 according to phb-crystal-ball), and we could just decide to move to gcc-8 only after that is out.
I think that upgrading the compiler to 8.1 after 6.12 LTS is a good plan.
Following that plan also means that the original stack protector improvement series that includes objtool changes is the way to go forward. The stack protector specific part can be removed from objtool after the compiler is upgraded.
Thanks,
Uros.