diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2023-04-30 21:43:31 +0300 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2023-04-30 21:43:31 +0300 |
commit | 10de638d8ea57ebab4231ea077bed01d9bade775 (patch) | |
tree | f9de1b131e1a94c1cbe051e55cda1ba9b9418ee9 /arch/s390 | |
parent | d55571c0084465f1f7e1e29f22bd910d366a6e1d (diff) | |
parent | 2a405f6bb3a5b2baaa74dfc5aaa0e1b99145bd1b (diff) | |
download | linux-10de638d8ea57ebab4231ea077bed01d9bade775.tar.xz |
Merge tag 's390-6.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Pull s390 updates from Vasily Gorbik:
- Add support for stackleak feature. Also allow specifying
architecture-specific stackleak poison function to enable faster
implementation. On s390, the mvc-based implementation helps decrease
typical overhead from a factor of 3 to just 25%
- Convert all assembler files to use SYM* style macros, deprecating the
ENTRY() macro and other annotations. Select ARCH_USE_SYM_ANNOTATIONS
- Improve KASLR to also randomize module and special amode31 code base
load addresses
- Rework decompressor memory tracking to support memory holes and
improve error handling
- Add support for protected virtualization AP binding
- Add support for set_direct_map() calls
- Implement set_memory_rox() and noexec module_alloc()
- Remove obsolete overriding of mem*() functions for KASAN
- Rework kexec/kdump to avoid using nodat_stack to call purgatory
- Convert the rest of the s390 code to use flexible-array member
instead of a zero-length array
- Clean up uaccess inline asm
- Enable ARCH_HAS_MEMBARRIER_SYNC_CORE
- Convert to using CONFIG_FUNCTION_ALIGNMENT and enable
DEBUG_FORCE_FUNCTION_ALIGN_64B
- Resolve last_break in userspace fault reports
- Simplify one-level sysctl registration
- Clean up branch prediction handling
- Rework CPU counter facility to retrieve available counter sets just
once
- Other various small fixes and improvements all over the code
* tag 's390-6.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (118 commits)
s390/stackleak: provide fast __stackleak_poison() implementation
stackleak: allow to specify arch specific stackleak poison function
s390: select ARCH_USE_SYM_ANNOTATIONS
s390/mm: use VM_FLUSH_RESET_PERMS in module_alloc()
s390: wire up memfd_secret system call
s390/mm: enable ARCH_HAS_SET_DIRECT_MAP
s390/mm: use BIT macro to generate SET_MEMORY bit masks
s390/relocate_kernel: adjust indentation
s390/relocate_kernel: use SYM* macros instead of ENTRY(), etc.
s390/entry: use SYM* macros instead of ENTRY(), etc.
s390/purgatory: use SYM* macros instead of ENTRY(), etc.
s390/kprobes: use SYM* macros instead of ENTRY(), etc.
s390/reipl: use SYM* macros instead of ENTRY(), etc.
s390/head64: use SYM* macros instead of ENTRY(), etc.
s390/earlypgm: use SYM* macros instead of ENTRY(), etc.
s390/mcount: use SYM* macros instead of ENTRY(), etc.
s390/crc32le: use SYM* macros instead of ENTRY(), etc.
s390/crc32be: use SYM* macros instead of ENTRY(), etc.
s390/crypto,chacha: use SYM* macros instead of ENTRY(), etc.
s390/amode31: use SYM* macros instead of ENTRY(), etc.
...
Diffstat (limited to 'arch/s390')
75 files changed, 1971 insertions, 1887 deletions
diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index 61d778397720..beb62f744c61 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -26,10 +26,6 @@ config GENERIC_BUG config GENERIC_BUG_RELATIVE_POINTERS def_bool y -config GENERIC_CSUM - bool - default y if KASAN - config GENERIC_LOCKBREAK def_bool y if PREEMPTION @@ -76,10 +72,12 @@ config S390 select ARCH_HAS_GCOV_PROFILE_ALL select ARCH_HAS_GIGANTIC_PAGE select ARCH_HAS_KCOV + select ARCH_HAS_MEMBARRIER_SYNC_CORE select ARCH_HAS_MEM_ENCRYPT select ARCH_HAS_NMI_SAFE_THIS_CPU_OPS select ARCH_HAS_PTE_SPECIAL select ARCH_HAS_SCALED_CPUTIME + select ARCH_HAS_SET_DIRECT_MAP select ARCH_HAS_SET_MEMORY select ARCH_HAS_STRICT_KERNEL_RWX select ARCH_HAS_STRICT_MODULE_RWX @@ -123,6 +121,7 @@ config S390 select ARCH_SUPPORTS_PER_VMA_LOCK select ARCH_USE_BUILTIN_BSWAP select ARCH_USE_CMPXCHG_LOCKREF + select ARCH_USE_SYM_ANNOTATIONS select ARCH_WANTS_DYNAMIC_TASK_STRUCT select ARCH_WANTS_NO_INSTR select ARCH_WANT_DEFAULT_BPF_JIT @@ -132,6 +131,8 @@ config S390 select CLONE_BACKWARDS2 select DMA_OPS if PCI select DYNAMIC_FTRACE if FUNCTION_TRACER + select FUNCTION_ALIGNMENT_8B if CC_IS_GCC + select FUNCTION_ALIGNMENT_16B if !CC_IS_GCC select GCC12_NO_ARRAY_BOUNDS select GENERIC_ALLOCATOR select GENERIC_CPU_AUTOPROBE @@ -153,6 +154,7 @@ config S390 select HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET select HAVE_ARCH_SECCOMP_FILTER select HAVE_ARCH_SOFT_DIRTY + select HAVE_ARCH_STACKLEAK select HAVE_ARCH_TRACEHOOK select HAVE_ARCH_TRANSPARENT_HUGEPAGE select HAVE_ARCH_VMAP_STACK diff --git a/arch/s390/appldata/appldata_base.c b/arch/s390/appldata/appldata_base.c index c0fd29133f27..b07b0610950e 100644 --- a/arch/s390/appldata/appldata_base.c +++ b/arch/s390/appldata/appldata_base.c @@ -66,16 +66,6 @@ static struct ctl_table appldata_table[] = { { }, }; -static struct ctl_table appldata_dir_table[] = { - { - .procname = appldata_proc_name, - .maxlen = 0, - .mode = S_IRUGO | S_IXUGO, - .child = appldata_table, - }, - { }, -}; - /* * Timer */ @@ -291,7 +281,7 @@ appldata_generic_handler(struct ctl_table *ctl, int write, mutex_lock(&appldata_ops_mutex); list_for_each(lh, &appldata_ops_list) { tmp_ops = list_entry(lh, struct appldata_ops, list); - if (&tmp_ops->ctl_table[2] == ctl) { + if (&tmp_ops->ctl_table[0] == ctl) { found = 1; } } @@ -361,7 +351,8 @@ int appldata_register_ops(struct appldata_ops *ops) if (ops->size > APPLDATA_MAX_REC_SIZE) return -EINVAL; - ops->ctl_table = kcalloc(4, sizeof(struct ctl_table), GFP_KERNEL); + /* The last entry must be an empty one */ + ops->ctl_table = kcalloc(2, sizeof(struct ctl_table), GFP_KERNEL); if (!ops->ctl_table) return -ENOMEM; @@ -369,17 +360,12 @@ int appldata_register_ops(struct appldata_ops *ops) list_add(&ops->list, &appldata_ops_list); mutex_unlock(&appldata_ops_mutex); - ops->ctl_table[0].procname = appldata_proc_name; - ops->ctl_table[0].maxlen = 0; - ops->ctl_table[0].mode = S_IRUGO | S_IXUGO; - ops->ctl_table[0].child = &ops->ctl_table[2]; - - ops->ctl_table[2].procname = ops->name; - ops->ctl_table[2].mode = S_IRUGO | S_IWUSR; - ops->ctl_table[2].proc_handler = appldata_generic_handler; - ops->ctl_table[2].data = ops; + ops->ctl_table[0].procname = ops->name; + ops->ctl_table[0].mode = S_IRUGO | S_IWUSR; + ops->ctl_table[0].proc_handler = appldata_generic_handler; + ops->ctl_table[0].data = ops; - ops->sysctl_header = register_sysctl_table(ops->ctl_table); + ops->sysctl_header = register_sysctl(appldata_proc_name, ops->ctl_table); if (!ops->sysctl_header) goto out; return 0; @@ -422,7 +408,7 @@ static int __init appldata_init(void) appldata_wq = alloc_ordered_workqueue("appldata", 0); if (!appldata_wq) return -ENOMEM; - appldata_sysctl_header = register_sysctl_table(appldata_dir_table); + appldata_sysctl_header = register_sysctl(appldata_proc_name, appldata_table); return 0; } diff --git a/arch/s390/boot/Makefile b/arch/s390/boot/Makefile index cebd4ca16916..c7c81e5f9218 100644 --- a/arch/s390/boot/Makefile +++ b/arch/s390/boot/Makefile @@ -35,7 +35,7 @@ endif CFLAGS_sclp_early_core.o += -I$(srctree)/drivers/s390/char -obj-y := head.o als.o startup.o mem_detect.o ipl_parm.o ipl_report.o vmem.o +obj-y := head.o als.o startup.o physmem_info.o ipl_parm.o ipl_report.o vmem.o obj-y += string.o ebcdic.o sclp_early_core.o mem.o ipl_vmparm.o cmdline.o obj-y += version.o pgm_check_info.o ctype.o ipl_data.o machine_kexec_reloc.o obj-$(findstring y, $(CONFIG_PROTECTED_VIRTUALIZATION_GUEST) $(CONFIG_PGSTE)) += uv.o diff --git a/arch/s390/boot/boot.h b/arch/s390/boot/boot.h index 58ce701d6110..222c6886acf6 100644 --- a/arch/s390/boot/boot.h +++ b/arch/s390/boot/boot.h @@ -8,6 +8,8 @@ #ifndef __ASSEMBLY__ +#include <asm/physmem_info.h> + struct machine_info { unsigned char has_edat1 : 1; unsigned char has_edat2 : 1; @@ -30,24 +32,46 @@ struct vmlinux_info { unsigned long init_mm_off; unsigned long swapper_pg_dir_off; unsigned long invalid_pg_dir_off; +#ifdef CONFIG_KASAN + unsigned long kasan_early_shadow_page_off; + unsigned long kasan_early_shadow_pte_off; + unsigned long kasan_early_shadow_pmd_off; + unsigned long kasan_early_shadow_pud_off; + unsigned long kasan_early_shadow_p4d_off; +#endif }; void startup_kernel(void); -unsigned long detect_memory(unsigned long *safe_addr); -void mem_detect_set_usable_limit(unsigned long limit); +unsigned long detect_max_physmem_end(void); +void detect_physmem_online_ranges(unsigned long max_physmem_end); +void physmem_set_usable_limit(unsigned long limit); +void physmem_reserve(enum reserved_range_type type, unsigned long addr, unsigned long size); +void physmem_free(enum reserved_range_type type); +/* for continuous/multiple allocations per type */ +unsigned long physmem_alloc_top_down(enum reserved_range_type type, unsigned long size, + unsigned long align); +/* for single allocations, 1 per type */ +unsigned long physmem_alloc_range(enum reserved_range_type type, unsigned long size, + unsigned long align, unsigned long min, unsigned long max, + bool die_on_oom); +unsigned long get_physmem_alloc_pos(void); +bool ipl_report_certs_intersects(unsigned long addr, unsigned long size, + unsigned long *intersection_start); bool is_ipl_block_dump(void); void store_ipl_parmblock(void); -unsigned long read_ipl_report(unsigned long safe_addr); +int read_ipl_report(void); +void save_ipl_cert_comp_list(void); void setup_boot_command_line(void); void parse_boot_command_line(void); void verify_facilities(void); void print_missing_facilities(void); void sclp_early_setup_buffer(void); void print_pgm_check_info(void); -unsigned long get_random_base(unsigned long safe_addr); +unsigned long randomize_within_range(unsigned long size, unsigned long align, + unsigned long min, unsigned long max); void setup_vmem(unsigned long asce_limit); -unsigned long vmem_estimate_memory_needs(unsigned long online_mem_total); void __printf(1, 2) decompressor_printk(const char *fmt, ...); +void print_stacktrace(unsigned long sp); void error(char *m); extern struct machine_info machine; @@ -57,12 +81,11 @@ extern const char kernel_version[]; extern unsigned long memory_limit; extern unsigned long vmalloc_size; extern int vmalloc_size_set; -extern int kaslr_enabled; extern char __boot_data_start[], __boot_data_end[]; extern char __boot_data_preserved_start[], __boot_data_preserved_end[]; extern char _decompressor_syms_start[], _decompressor_syms_end[]; extern char _stack_start[], _stack_end[]; -extern char _end[]; +extern char _end[], _decompressor_end[]; extern unsigned char _compressed_start[]; extern unsigned char _compressed_end[]; extern struct vmlinux_info _vmlinux_info; @@ -70,5 +93,10 @@ extern struct vmlinux_info _vmlinux_info; #define __abs_lowcore_pa(x) (((unsigned long)(x) - __abs_lowcore) % sizeof(struct lowcore)) +static inline bool intersects(unsigned long addr0, unsigned long size0, + unsigned long addr1, unsigned long size1) +{ + return addr0 + size0 > addr1 && addr1 + size1 > addr0; +} #endif /* __ASSEMBLY__ */ #endif /* BOOT_BOOT_H */ diff --git a/arch/s390/boot/install.sh b/arch/s390/boot/install.sh index 616ba1660f08..a13dd2f2aa1c 100755 --- a/arch/s390/boot/install.sh +++ b/arch/s390/boot/install.sh @@ -17,8 +17,8 @@ echo "Warning: '${INSTALLKERNEL}' command not available - additional " \ "bootloader config required" >&2 -if [ -f $4/vmlinuz-$1 ]; then mv $4/vmlinuz-$1 $4/vmlinuz-$1.old; fi -if [ -f $4/System.map-$1 ]; then mv $4/System.map-$1 $4/System.map-$1.old; fi +if [ -f "$4/vmlinuz-$1" ]; then mv -- "$4/vmlinuz-$1" "$4/vmlinuz-$1.old"; fi +if [ -f "$4/System.map-$1" ]; then mv -- "$4/System.map-$1" "$4/System.map-$1.old"; fi -cat $2 > $4/vmlinuz-$1 -cp $3 $4/System.map-$1 +cat -- "$2" > "$4/vmlinuz-$1" +cp -- "$3" "$4/System.map-$1" diff --git a/arch/s390/boot/ipl_parm.c b/arch/s390/boot/ipl_parm.c index c1f8f7999fed..8753cb0339e5 100644 --- a/arch/s390/boot/ipl_parm.c +++ b/arch/s390/boot/ipl_parm.c @@ -24,11 +24,11 @@ int __bootdata(noexec_disabled); unsigned int __bootdata_preserved(zlib_dfltcc_support) = ZLIB_DFLTCC_FULL; struct ipl_parameter_block __bootdata_preserved(ipl_block); int __bootdata_preserved(ipl_block_valid); +int __bootdata_preserved(__kaslr_enabled); unsigned long vmalloc_size = VMALLOC_DEFAULT_SIZE; unsigned long memory_limit; int vmalloc_size_set; -int kaslr_enabled; static inline int __diag308(unsigned long subcode, void *addr) { @@ -264,7 +264,7 @@ void parse_boot_command_line(void) char *args; int rc; - kaslr_enabled = IS_ENABLED(CONFIG_RANDOMIZE_BASE); + __kaslr_enabled = IS_ENABLED(CONFIG_RANDOMIZE_BASE); args = strcpy(command_line_buf, early_command_line); while (*args) { args = next_arg(args, ¶m, &val); @@ -300,7 +300,7 @@ void parse_boot_command_line(void) modify_fac_list(val); if (!strcmp(param, "nokaslr")) - kaslr_enabled = 0; + __kaslr_enabled = 0; #if IS_ENABLED(CONFIG_KVM) if (!strcmp(param, "prot_virt")) { diff --git a/arch/s390/boot/ipl_report.c b/arch/s390/boot/ipl_report.c index 74b5cd264862..1803035e68d2 100644 --- a/arch/s390/boot/ipl_report.c +++ b/arch/s390/boot/ipl_report.c @@ -5,6 +5,7 @@ #include <asm/sclp.h> #include <asm/sections.h> #include <asm/boot_data.h> +#include <asm/physmem_info.h> #include <uapi/asm/ipl.h> #include "boot.h" @@ -16,20 +17,16 @@ unsigned long __bootdata_preserved(ipl_cert_list_size); unsigned long __bootdata(early_ipl_comp_list_addr); unsigned long __bootdata(early_ipl_comp_list_size); +static struct ipl_rb_certificates *certs; +static struct ipl_rb_components *comps; +static bool ipl_report_needs_saving; + #define for_each_rb_entry(entry, rb) \ for (entry = rb->entries; \ (void *) entry + sizeof(*entry) <= (void *) rb + rb->len; \ entry++) -static inline bool intersects(unsigned long addr0, unsigned long size0, - unsigned long addr1, unsigned long size1) -{ - return addr0 + size0 > addr1 && addr1 + size1 > addr0; -} - -static unsigned long find_bootdata_space(struct ipl_rb_components *comps, - struct ipl_rb_certificates *certs, - unsigned long safe_addr) +static unsigned long get_cert_comp_list_size(void) { struct ipl_rb_certificate_entry *cert; struct ipl_rb_component_entry *comp; @@ -44,44 +41,27 @@ static unsigned long find_bootdata_space(struct ipl_rb_components *comps, ipl_cert_list_size = 0; for_each_rb_entry(cert, certs) ipl_cert_list_size += sizeof(unsigned int) + cert->len; - size = ipl_cert_list_size + early_ipl_comp_list_size; + return ipl_cert_list_size + early_ipl_comp_list_size; +} - /* - * Start from safe_addr to find a free memory area large - * enough for the IPL report boot data. This area is used - * for ipl_cert_list_addr/ipl_cert_list_size and - * early_ipl_comp_list_addr/early_ipl_comp_list_size. It must - * not overlap with any component or any certificate. - */ -repeat: - if (IS_ENABLED(CONFIG_BLK_DEV_INITRD) && initrd_data.start && initrd_data.size && - intersects(initrd_data.start, initrd_data.size, safe_addr, size)) - safe_addr = initrd_data.start + initrd_data.size; - if (intersects(safe_addr, size, (unsigned long)comps, comps->len)) { - safe_addr = (unsigned long)comps + comps->len; - goto repeat; - } - for_each_rb_entry(comp, comps) - if (intersects(safe_addr, size, comp->addr, comp->len)) { - safe_addr = comp->addr + comp->len; - goto repeat; +bool ipl_report_certs_intersects(unsigned long addr, unsigned long size, + unsigned long *intersection_start) +{ + struct ipl_rb_certificate_entry *cert; + + if (!ipl_report_needs_saving) + return false; + + for_each_rb_entry(cert, certs) { + if (intersects(addr, size, cert->addr, cert->len)) { + *intersection_start = cert->addr; + return true; } - if (intersects(safe_addr, size, (unsigned long)certs, certs->len)) { - safe_addr = (unsigned long)certs + certs->len; - goto repeat; } - for_each_rb_entry(cert, certs) - if (intersects(safe_addr, size, cert->addr, cert->len)) { - safe_addr = cert->addr + cert->len; - goto repeat; - } - early_ipl_comp_list_addr = safe_addr; - ipl_cert_list_addr = safe_addr + early_ipl_comp_list_size; - - return safe_addr + size; + return false; } -static void copy_components_bootdata(struct ipl_rb_components *comps) +static void copy_components_bootdata(void) { struct ipl_rb_component_entry *comp, *ptr; @@ -90,7 +70,7 @@ static void copy_components_bootdata(struct ipl_rb_components *comps) memcpy(ptr++, comp, sizeof(*ptr)); } -static void copy_certificates_bootdata(struct ipl_rb_certificates *certs) +static void copy_certificates_bootdata(void) { struct ipl_rb_certificate_entry *cert; void *ptr; @@ -104,10 +84,8 @@ static void copy_certificates_bootdata(struct ipl_rb_certificates *certs) } } -unsigned long read_ipl_report(unsigned long safe_addr) +int read_ipl_report(void) { - struct ipl_rb_certificates *certs; - struct ipl_rb_components *comps; struct ipl_pl_hdr *pl_hdr; struct ipl_rl_hdr *rl_hdr; struct ipl_rb_hdr *rb_hdr; @@ -120,7 +98,7 @@ unsigned long read_ipl_report(unsigned long safe_addr) */ if (!ipl_block_valid || !(ipl_block.hdr.flags & IPL_PL_FLAG_IPLSR)) - return safe_addr; + return -1; ipl_secure_flag = !!(ipl_block.hdr.flags & IPL_PL_FLAG_SIPL); /* * There is an IPL report, to find it load the pointer to the @@ -158,16 +136,30 @@ unsigned long read_ipl_report(unsigned long safe_addr) * With either the component list or the certificate list * missing the kernel will stay ignorant of secure IPL. */ - if (!comps || !certs) - return safe_addr; + if (!comps || !certs) { + certs = NULL; + return -1; + } - /* - * Copy component and certificate list to a safe area - * where the decompressed kernel can find them. - */ - safe_addr = find_bootdata_space(comps, certs, safe_addr); - copy_components_bootdata(comps); - copy_certificates_bootdata(certs); + ipl_report_needs_saving = true; + physmem_reserve(RR_IPLREPORT, (unsigned long)pl_hdr, + (unsigned long)rl_end - (unsigned long)pl_hdr); + return 0; +} + +void save_ipl_cert_comp_list(void) +{ + unsigned long size; + + if (!ipl_report_needs_saving) + return; + + size = get_cert_comp_list_size(); + early_ipl_comp_list_addr = physmem_alloc_top_down(RR_CERT_COMP_LIST, size, sizeof(int)); + ipl_cert_list_addr = early_ipl_comp_list_addr + early_ipl_comp_list_size; - return safe_addr; + copy_components_bootdata(); + copy_certificates_bootdata(); + physmem_free(RR_IPLREPORT); + ipl_report_needs_saving = false; } diff --git a/arch/s390/boot/kaslr.c b/arch/s390/boot/kaslr.c index 3e3d846400b4..90602101e2ae 100644 --- a/arch/s390/boot/kaslr.c +++ b/arch/s390/boot/kaslr.c @@ -3,7 +3,7 @@ * Copyright IBM Corp. 2019 */ #include <linux/pgtable.h> -#include <asm/mem_detect.h> +#include <asm/physmem_info.h> #include <asm/cpacf.h> #include <asm/timex.h> #include <asm/sclp.h> @@ -91,113 +91,108 @@ static int get_random(unsigned long limit, unsigned long *value) return 0; } -/* - * To randomize kernel base address we have to consider several facts: - * 1. physical online memory might not be continuous and have holes. mem_detect - * info contains list of online memory ranges we should consider. - * 2. we have several memory regions which are occupied and we should not - * overlap and destroy them. Currently safe_addr tells us the border below - * which all those occupied regions are. We are safe to use anything above - * safe_addr. - * 3. the upper limit might apply as well, even if memory above that limit is - * online. Currently those limitations are: - * 3.1. Limit set by "mem=" kernel command line option - * 3.2. memory reserved at the end for kasan initialization. - * 4. kernel base address must be aligned to THREAD_SIZE (kernel stack size). - * Which is required for CONFIG_CHECK_STACK. Currently THREAD_SIZE is 4 pages - * (16 pages when the kernel is built with kasan enabled) - * Assumptions: - * 1. kernel size (including .bss size) and upper memory limit are page aligned. - * 2. mem_detect memory region start is THREAD_SIZE aligned / end is PAGE_SIZE - * aligned (in practice memory configurations granularity on z/VM and LPAR - * is 1mb). - * - * To guarantee uniform distribution of kernel base address among all suitable - * addresses we generate random value just once. For that we need to build a - * continuous range in which every value would be suitable. We can build this - * range by simply counting all suitable addresses (let's call them positions) - * which would be valid as kernel base address. To count positions we iterate - * over online memory ranges. For each range which is big enough for the - * kernel image we count all suitable addresses we can put the kernel image at - * that is - * (end - start - kernel_size) / THREAD_SIZE + 1 - * Two functions count_valid_kernel_positions and position_to_address help - * to count positions in memory range given and then convert position back - * to address. - */ -static unsigned long count_valid_kernel_positions(unsigned long kernel_size, - unsigned long _min, - unsigned long _max) +static void sort_reserved_ranges(struct reserved_range *res, unsigned long size) { - unsigned long start, end, pos = 0; - int i; - - for_each_mem_detect_usable_block(i, &start, &end) { - if (_min >= end) - continue; - if (start >= _max) - break; - start = max(_min, start); - end = min(_max, end); - if (end - start < kernel_size) - continue; - pos += (end - start - kernel_size) / THREAD_SIZE + 1; + struct reserved_range tmp; + int i, j; + + for (i = 1; i < size; i++) { + tmp = res[i]; + for (j = i - 1; j >= 0 && res[j].start > tmp.start; j--) + res[j + 1] = res[j]; + res[j + 1] = tmp; } - - return pos; } -static unsigned long position_to_address(unsigned long pos, unsigned long kernel_size, - unsigned long _min, unsigned long _max) +static unsigned long iterate_valid_positions(unsigned long size, unsigned long align, + unsigned long _min, unsigned long _max, + struct reserved_range *res, size_t res_count, + bool pos_count, unsigned long find_pos) { - unsigned long start, end; + unsigned long start, end, tmp_end, range_pos, pos = 0; + struct reserved_range *res_end = res + res_count; + struct reserved_range *skip_res; int i; - for_each_mem_detect_usable_block(i, &start, &end) { + align = max(align, 8UL); + _min = round_up(_min, align); + for_each_physmem_usable_range(i, &start, &end) { if (_min >= end) continue; + start = round_up(start, align); if (start >= _max) break; start = max(_min, start); end = min(_max, end); - if (end - start < kernel_size) - continue; - if ((end - start - kernel_size) / THREAD_SIZE + 1 >= pos) - return start + (pos - 1) * THREAD_SIZE; - pos -= (end - start - kernel_size) / THREAD_SIZE + 1; + + while (start + size <= end) { + /* skip reserved ranges below the start */ + while (res && res->end <= start) { + res++; + if (res >= res_end) + res = NULL; + } + skip_res = NULL; + tmp_end = end; + /* has intersecting reserved range */ + if (res && res->start < end) { + skip_res = res; + tmp_end = res->start; + } + if (start + size <= tmp_end) { + range_pos = (tmp_end - start - size) / align + 1; + if (pos_count) { + pos += range_pos; + } else { + if (range_pos >= find_pos) + return start + (find_pos - 1) * align; + find_pos -= range_pos; + } + } + if (!skip_res) + break; + start = round_up(skip_res->end, align); + } } - return 0; + return pos_count ? pos : 0; } -unsigned long get_random_base(unsigned long safe_addr) +/* + * Two types of decompressor memory allocations/reserves are considered + * differently. + * + * "Static" or "single" allocations are done via physmem_alloc_range() and + * physmem_reserve(), and they are listed in physmem_info.reserved[]. Each + * type of "static" allocation can only have one allocation per type and + * cannot have chains. + * + * On the other hand, "dynamic" or "repetitive" allocations are done via + * physmem_alloc_top_down(). These allocations are tightly packed together + * top down from the end of online memory. physmem_alloc_pos represents + * current position where those allocations start. + * + * Functions randomize_within_range() and iterate_valid_positions() + * only consider "dynamic" allocations by never looking above + * physmem_alloc_pos. "Static" allocations, however, are explicitly + * considered by checking the "res" (reserves) array. The first + * reserved_range of a "dynamic" allocation may also be checked along the + * way, but it will always be above the maximum value anyway. + */ +unsigned long randomize_within_range(unsigned long size, unsigned long align, + unsigned long min, unsigned long max) { - unsigned long usable_total = get_mem_detect_usable_total(); - unsigned long memory_limit = get_mem_detect_end(); - unsigned long base_pos, max_pos, kernel_size; - int i; - - /* - * Avoid putting kernel in the end of physical memory - * which vmem and kasan code will use for shadow memory and - * pgtable mapping allocations. - */ - memory_limit -= kasan_estimate_memory_needs(usable_total); - memory_limit -= vmem_estimate_memory_needs(usable_total); + struct reserved_range res[RR_MAX]; + unsigned long max_pos, pos; - safe_addr = ALIGN(safe_addr, THREAD_SIZE); - kernel_size = vmlinux.image_size + vmlinux.bss_size; - if (safe_addr + kernel_size > memory_limit) - return 0; + memcpy(res, physmem_info.reserved, sizeof(res)); + sort_reserved_ranges(res, ARRAY_SIZE(res)); + max = min(max, get_physmem_alloc_pos()); - max_pos = count_valid_kernel_positions(kernel_size, safe_addr, memory_limit); - if (!max_pos) { - sclp_early_printk("KASLR disabled: not enough memory\n"); + max_pos = iterate_valid_positions(size, align, min, max, res, ARRAY_SIZE(res), true, 0); + if (!max_pos) return 0; - } - - /* we need a value in the range [1, base_pos] inclusive */ - if (get_random(max_pos, &base_pos)) + if (get_random(max_pos, &pos)) return 0; - return position_to_address(base_pos + 1, kernel_size, safe_addr, memory_limit); + return iterate_valid_positions(size, align, min, max, res, ARRAY_SIZE(res), false, pos + 1); } diff --git a/arch/s390/boot/mem_detect.c b/arch/s390/boot/mem_detect.c deleted file mode 100644 index 35f4ba11f7fd..000000000000 --- a/arch/s390/boot/mem_detect.c +++ /dev/null @@ -1,191 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -#include <linux/errno.h> -#include <linux/init.h> -#include <asm/setup.h> -#include <asm/processor.h> -#include <asm/sclp.h> -#include <asm/sections.h> -#include <asm/mem_detect.h> -#include <asm/sparsemem.h> -#include "decompressor.h" -#include "boot.h" - -struct mem_detect_info __bootdata(mem_detect); - -/* up to 256 storage elements, 1020 subincrements each */ -#define ENTRIES_EXTENDED_MAX \ - (256 * (1020 / 2) * sizeof(struct mem_detect_block)) - -static struct mem_detect_block *__get_mem_detect_block_ptr(u32 n) -{ - if (n < MEM_INLINED_ENTRIES) - return &mem_detect.entries[n]; - return &mem_detect.entries_extended[n - MEM_INLINED_ENTRIES]; -} - -/* - * sequential calls to add_mem_detect_block with adjacent memory areas - * are merged together into single memory block. - */ -void add_mem_detect_block(u64 start, u64 end) -{ - struct mem_detect_block *block; - - if (mem_detect.count) { - block = __get_mem_detect_block_ptr(mem_detect.count - 1); - if (block->end == start) { - block->end = end; - return; - } - } - - block = __get_mem_detect_block_ptr(mem_detect.count); - block->start = start; - block->end = end; - mem_detect.count++; -} - -static int __diag260(unsigned long rx1, unsigned long rx2) -{ - unsigned long reg1, reg2, ry; - union register_pair rx; - psw_t old; - int rc; - - rx.even = rx1; - rx.odd = rx2; - ry = 0x10; /* storage configuration */ - rc = -1; /* fail */ - asm volatile( - " mvc 0(16,%[psw_old]),0(%[psw_pgm])\n" - " epsw %[reg1],%[reg2]\n" - " st %[reg1],0(%[psw_pgm])\n" - " st %[reg2],4(%[psw_pgm])\n" - " larl %[reg1],1f\n" - " stg %[reg1],8(%[psw_pgm])\n" - " diag %[rx],%[ry],0x260\n" - " ipm %[rc]\n" - " srl %[rc],28\n" - "1: mvc 0(16,%[psw_pgm]),0(%[psw_old])\n" - : [reg1] "=&d" (reg1), - [reg2] "=&a" (reg2), - [rc] "+&d" (rc), - [ry] "+&d" (ry), - "+Q" (S390_lowcore.program_new_psw), - "=Q" (old) - : [rx] "d" (rx.pair), - [psw_old] "a" (&old), - [psw_pgm] "a" (&S390_lowcore.program_new_psw) - : "cc", "memory"); - return rc == 0 ? ry : -1; -} - -static int diag260(void) -{ - int rc, i; - - struct { - unsigned long start; - unsigned long end; - } storage_extents[8] __aligned(16); /* VM supports up to 8 extends */ - - memset(storage_extents, 0, sizeof(storage_extents)); - rc = __diag260((unsigned long)storage_extents, sizeof(storage_extents)); - if (rc == -1) - return -1; - - for (i = 0; i < min_t(int, rc, ARRAY_SIZE(storage_extents)); i++) - add_mem_detect_block(storage_extents[i].start, storage_extents[i].end + 1); - return 0; -} - -static int tprot(unsigned long addr) -{ - unsigned long reg1, reg2; - int rc = -EFAULT; - psw_t old; - - asm volatile( - " mvc 0(16,%[psw_old]),0(%[psw_pgm])\n" - " epsw %[reg1],%[reg2]\n" - " st %[reg1],0(%[psw_pgm])\n" - " st %[reg2],4(%[psw_pgm])\n" - " larl %[reg1],1f\n" - " stg %[reg1],8(%[psw_pgm])\n" - " tprot 0(%[addr]),0\n" - " ipm %[rc]\n" - " srl %[rc],28\n" - "1: mvc 0(16,%[psw_pgm]),0(%[psw_old])\n" - : [reg1] "=&d" (reg1), - [reg2] "=&a" (reg2), - [rc] "+&d" (rc), - "=Q" (S390_lowcore.program_new_psw.addr), - "=Q" (old) - : [psw_old] "a" (&old), - [psw_pgm] "a" (&S390_lowcore.program_new_psw), - [addr] "a" (addr) - : "cc", "memory"); - return rc; -} - -static unsigned long search_mem_end(void) -{ - unsigned long range = 1 << (MAX_PHYSMEM_BITS - 20); /* in 1MB blocks */ - unsigned long offset = 0; - unsigned long pivot; - - while (range > 1) { - range >>= 1; - pivot = offset + range; - if (!tprot(pivot << 20)) - offset = pivot; - } - return (offset + 1) << 20; -} - -unsigned long detect_memory(unsigned long *safe_addr) -{ - unsigned long max_physmem_end = 0; - - sclp_early_get_memsize(&max_physmem_end); - mem_detect.entries_extended = (struct mem_detect_block *)ALIGN(*safe_addr, sizeof(u64)); - - if (!sclp_early_read_storage_info()) { - mem_detect.info_source = MEM_DETECT_SCLP_STOR_INFO; - } else if (!diag260()) { - mem_detect.info_source = MEM_DETECT_DIAG260; - max_physmem_end = max_physmem_end ?: get_mem_detect_end(); - } else if (max_physmem_end) { - add_mem_detect_block(0, max_physmem_end); - mem_detect.info_source = MEM_DETECT_SCLP_READ_INFO; - } else { - max_physmem_end = search_mem_end(); - add_mem_detect_block(0, max_physmem_end); - mem_detect.info_source = MEM_DETECT_BIN_SEARCH; - } - - if (mem_detect.count > MEM_INLINED_ENTRIES) { - *safe_addr += (mem_detect.count - MEM_INLINED_ENTRIES) * - sizeof(struct mem_detect_block); - } - - return max_physmem_end; -} - -void mem_detect_set_usable_limit(unsigned long limit) -{ - struct mem_detect_block *block; - int i; - - /* make sure mem_detect.usable ends up within online memory block */ - for (i = 0; i < mem_detect.count; i++) { - block = __get_mem_detect_block_ptr(i); - if (block->start >= limit) - break; - if (block->end >= limit) { - mem_detect.usable = limit; - break; - } - mem_detect.usable = block->end; - } -} diff --git a/arch/s390/boot/pgm_check_info.c b/arch/s390/boot/pgm_check_info.c index c2a1defc79da..97244cd7a206 100644 --- a/arch/s390/boot/pgm_check_info.c +++ b/arch/s390/boot/pgm_check_info.c @@ -123,11 +123,10 @@ out: sclp_early_printk(buf); } -static noinline void print_stacktrace(void) +void print_stacktrace(unsigned long sp) { struct stack_info boot_stack = { STACK_TYPE_TASK, (unsigned long)_stack_start, (unsigned long)_stack_end }; - unsigned long sp = S390_lowcore.gpregs_save_area[15]; bool first = true; decompressor_printk("Call Trace:\n"); @@ -154,7 +153,7 @@ void print_pgm_check_info(void) decompressor_printk("Kernel command line: %s\n", early_command_line); decompressor_printk("Kernel fault: interruption code %04x ilc:%x\n", S390_lowcore.pgm_code, S390_lowcore.pgm_ilc >> 1); - if (kaslr_enabled) + if (kaslr_enabled()) decompressor_printk("Kernel random base: %lx\n", __kaslr_offset); decompressor_printk("PSW : %016lx %016lx (%pS)\n", S390_lowcore.psw_save_area.mask, @@ -173,7 +172,7 @@ void print_pgm_check_info(void) gpregs[8], gpregs[9], gpregs[10], gpregs[11]); decompressor_printk(" %016lx %016lx %016lx %016lx\n", gpregs[12], gpregs[13], gpregs[14], gpregs[15]); - print_stacktrace(); + print_stacktrace(S390_lowcore.gpregs_save_area[15]); decompressor_printk("Last Breaking-Event-Address:\n"); decompressor_printk(" [<%016lx>] %pS\n", (unsigned long)S390_lowcore.pgm_last_break, (void *)S390_lowcore.pgm_last_break); diff --git a/arch/s390/boot/physmem_info.c b/arch/s390/boot/physmem_info.c new file mode 100644 index 000000000000..0cf79826eef9 --- /dev/null +++ b/arch/s390/boot/physmem_info.c @@ -0,0 +1,328 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/processor.h> +#include <linux/errno.h> +#include <linux/init.h> +#include <asm/physmem_info.h> +#include <asm/stacktrace.h> +#include <asm/boot_data.h> +#include <asm/sparsemem.h> +#include <asm/sections.h> +#include <asm/setup.h> +#include <asm/sclp.h> +#include <asm/uv.h> +#include "decompressor.h" +#include "boot.h" + +struct physmem_info __bootdata(physmem_info); +static unsigned int physmem_alloc_ranges; +static unsigned long physmem_alloc_pos; + +/* up to 256 storage elements, 1020 subincrements each */ +#define ENTRIES_EXTENDED_MAX \ + (256 * (1020 / 2) * sizeof(struct physmem_range)) + +static struct physmem_range *__get_physmem_range_ptr(u32 n) +{ + if (n < MEM_INLINED_ENTRIES) + return &physmem_info.online[n]; + if (unlikely(!physmem_info.online_extended)) { + physmem_info.online_extended = (struct physmem_range *)physmem_alloc_range( + RR_MEM_DETECT_EXTENDED, ENTRIES_EXTENDED_MAX, sizeof(long), 0, + physmem_alloc_pos, true); + } + return &physmem_info.online_extended[n - MEM_INLINED_ENTRIES]; +} + +/* + * sequential calls to add_physmem_online_range with adjacent memory ranges + * are merged together into single memory range. + */ +void add_physmem_online_range(u64 start, u64 end) +{ + struct physmem_range *range; + + if (physmem_info.range_count) { + range = __get_physmem_range_ptr(physmem_info.range_count - 1); + if (range->end == start) { + range->end = end; + return; + } + } + + range = __get_physmem_range_ptr(physmem_info.range_count); + range->start = start; + range->end = end; + physmem_info.range_count++; +} + +static int __diag260(unsigned long rx1, unsigned long rx2) +{ + unsigned long reg1, reg2, ry; + union register_pair rx; + psw_t old; + int rc; + + rx.even = rx1; + rx.odd = rx2; + ry = 0x10; /* storage configuration */ + rc = -1; /* fail */ + asm volatile( + " mvc 0(16,%[psw_old]),0(%[psw_pgm])\n" + " epsw %[reg1],%[reg2]\n" + " st %[reg1],0(%[psw_pgm])\n" + " st %[reg2],4(%[psw_pgm])\n" + " larl %[reg1],1f\n" + " stg %[reg1],8(%[psw_pgm])\n" + " diag %[rx],%[ry],0x260\n" + " ipm %[rc]\n" + " srl %[rc],28\n" + "1: mvc 0(16,%[psw_pgm]),0(%[psw_old])\n" + : [reg1] "=&d" (reg1), + [reg2] "=&a" (reg2), + [rc] "+&d" (rc), + [ry] "+&d" (ry), + "+Q" (S390_lowcore.program_new_psw), + "=Q" (old) + : [rx] "d" (rx.pair), + [psw_old] "a" (&old), + [psw_pgm] "a" (&S390_lowcore.program_new_psw) + : "cc", "memory"); + return rc == 0 ? ry : -1; +} + +static int diag260(void) +{ + int rc, i; + + struct { + unsigned long start; + unsigned long end; + } storage_extents[8] __aligned(16); /* VM supports up to 8 extends */ + + memset(storage_extents, 0, sizeof(storage_extents)); + rc = __diag260((unsigned long)storage_extents, sizeof(storage_extents)); + if (rc == -1) + return -1; + + for (i = 0; i < min_t(int, rc, ARRAY_SIZE(storage_extents)); i++) + add_physmem_online_range(storage_extents[i].start, storage_extents[i].end + 1); + return 0; +} + +static int tprot(unsigned long addr) +{ + unsigned long reg1, reg2; + int rc = -EFAULT; + psw_t old; + + asm volatile( + " mvc 0(16,%[psw_old]),0(%[psw_pgm])\n" + " epsw %[reg1],%[reg2]\n" + " st %[reg1],0(%[psw_pgm])\n" + " st %[reg2],4(%[psw_pgm])\n" + " larl %[reg1],1f\n" + " stg %[reg1],8(%[psw_pgm])\n" + " tprot 0(%[addr]),0\n" + " ipm %[rc]\n" + " srl %[rc],28\n" + "1: mvc 0(16,%[psw_pgm]),0(%[psw_old])\n" + : [reg1] "=&d" (reg1), + [reg2] "=&a" (reg2), + [rc] "+&d" (rc), + "=Q" (S390_lowcore.program_new_psw.addr), + "=Q" (old) + : [psw_old] "a" (&old), + [psw_pgm] "a" (&S390_lowcore.program_new_psw), + [addr] "a" (addr) + : "cc", "memory"); + return rc; +} + +static unsigned long search_mem_end(void) +{ + unsigned long range = 1 << (MAX_PHYSMEM_BITS - 20); /* in 1MB blocks */ + unsigned long offset = 0; + unsigned long pivot; + + while (range > 1) { + range >>= 1; + pivot = offset + range; + if (!tprot(pivot << 20)) + offset = pivot; + } + return (offset + 1) << 20; +} + +unsigned long detect_max_physmem_end(void) +{ + unsigned long max_physmem_end = 0; + + if (!sclp_early_get_memsize(&max_physmem_end)) { + physmem_info.info_source = MEM_DETECT_SCLP_READ_INFO; + } else { + max_physmem_end = search_mem_end(); + physmem_info.info_source = MEM_DETECT_BIN_SEARCH; + } + return max_physmem_end; +} + +void detect_physmem_online_ranges(unsigned long max_physmem_end) +{ + if (!sclp_early_read_storage_info()) { + physmem_info.info_source = MEM_DETECT_SCLP_STOR_INFO; + } else if (!diag260()) { + physmem_info.info_source = MEM_DETECT_DIAG260; + } else if (max_physmem_end) { + add_physmem_online_range(0, max_physmem_end); + } +} + +void physmem_set_usable_limit(unsigned long limit) +{ + physmem_info.usable = limit; + physmem_alloc_pos = limit; +} + +static void die_oom(unsigned long size, unsigned long align, unsigned long min, unsigned long max) +{ + unsigned long start, end, total_mem = 0, total_reserved_mem = 0; + struct reserved_range *range; + enum reserved_range_type t; + int i; + + decompressor_printk("Linux version %s\n", kernel_version); + if (!is_prot_virt_guest() && early_command_line[0]) + decompressor_printk("Kernel command line: %s\n", early_command_line); + decompressor_printk("Out of memory allocating %lx bytes %lx aligned in range %lx:%lx\n", + size, align, min, max); + decompressor_printk("Reserved memory ranges:\n"); + for_each_physmem_reserved_range(t, range, &start, &end) { + decompressor_printk("%016lx %016lx %s\n", start, end, get_rr_type_name(t)); + total_reserved_mem += end - start; + } + decompressor_printk("Usable online memory ranges (info source: %s [%x]):\n", + get_physmem_info_source(), physmem_info.info_source); + for_each_physmem_usable_range(i, &start, &end) { + decompressor_printk("%016lx %016lx\n", start, end); + total_mem += end - start; + } + decompressor_printk("Usable online memory total: %lx Reserved: %lx Free: %lx\n", + total_mem, total_reserved_mem, + total_mem > total_reserved_mem ? total_mem - total_reserved_mem : 0); + print_stacktrace(current_frame_address()); + sclp_early_printk("\n\n -- System halted\n"); + disabled_wait(); +} + +void physmem_reserve(enum reserved_range_type type, unsigned long addr, unsigned long size) +{ + physmem_info.reserved[type].start = addr; + physmem_info.reserved[type].end = addr + size; +} + +void physmem_free(enum reserved_range_type type) +{ + physmem_info.reserved[type].start = 0; + physmem_info.reserved[type].end = 0; +} + +static bool __physmem_alloc_intersects(unsigned long addr, unsigned long size, + unsigned long *intersection_start) +{ + unsigned long res_addr, res_size; + int t; + + for (t = 0; t < RR_MAX; t++) { + if (!get_physmem_reserved(t, &res_addr, &res_size)) + continue; + if (intersects(addr, size, res_addr, res_size)) { + *intersection_start = res_addr; + return true; + } + } + return ipl_report_certs_intersects(addr, size, intersection_start); +} + +static unsigned long __physmem_alloc_range(unsigned long size, unsigned long align, + unsigned long min, unsigned long max, + unsigned int from_ranges, unsigned int *ranges_left, + bool die_on_oom) +{ + unsigned int nranges = from_ranges ?: physmem_info.range_count; + unsigned long range_start, range_end; + unsigned long intersection_start; + unsigned long addr, pos = max; + + align = max(align, 8UL); + while (nranges) { + __get_physmem_range(nranges - 1, &range_start, &range_end, false); + pos = min(range_end, pos); + + if (round_up(min, align) + size > pos) + break; + addr = round_down(pos - size, align); + if (range_start > addr) { + nranges--; + continue; + } + if (__physmem_alloc_intersects(addr, size, &intersection_start)) { + pos = intersection_start; + continue; + } + + if (ranges_left) + *ranges_left = nranges; + return addr; + } + if (die_on_oom) + die_oom(size, align, min, max); + return 0; +} + +unsigned long physmem_alloc_range(enum reserved_range_type type, unsigned long size, + unsigned long align, unsigned long min, unsigned long max, + bool die_on_oom) +{ + unsigned long addr; + + max = min(max, physmem_alloc_pos); + addr = __physmem_alloc_range(size, align, min, max, 0, NULL, die_on_oom); + if (addr) + physmem_reserve(type, addr, size); + return addr; +} + +unsigned long physmem_alloc_top_down(enum reserved_range_type type, unsigned long size, + unsigned long align) +{ + struct reserved_range *range = &physmem_info.reserved[type]; + struct reserved_range *new_range; + unsigned int ranges_left; + unsigned long addr; + + addr = __physmem_alloc_range(size, align, 0, physmem_alloc_pos, physmem_alloc_ranges, + &ranges_left, true); + /* if not a consecutive allocation of the same type or first allocation */ + if (range->start != addr + size) { + if (range->end) { + physmem_alloc_pos = __physmem_alloc_range( + sizeof(struct reserved_range), 0, 0, physmem_alloc_pos, + physmem_alloc_ranges, &ranges_left, true); + new_range = (struct reserved_range *)physmem_alloc_pos; + *new_range = *range; + range->chain = new_range; + addr = __physmem_alloc_range(size, align, 0, physmem_alloc_pos, + ranges_left, &ranges_left, true); + } + range->end = addr + size; + } + range->start = addr; + physmem_alloc_pos = addr; + physmem_alloc_ranges = ranges_left; + return addr; +} + +unsigned long get_physmem_alloc_pos(void) +{ + return physmem_alloc_pos; +} diff --git a/arch/s390/boot/startup.c b/arch/s390/boot/startup.c index 11413f0baabc..64bd7ac3e35d 100644 --- a/arch/s390/boot/startup.c +++ b/arch/s390/boot/startup.c @@ -12,7 +12,7 @@ #include <asm/diag.h> #include <asm/uv.h> #include <asm/abs_lowcore.h> -#include <asm/mem_detect.h> +#include <asm/physmem_info.h> #include "decompressor.h" #include "boot.h" #include "uv.h" @@ -21,7 +21,6 @@ unsigned long __bootdata_preserved(__kaslr_offset); unsigned long __bootdata_preserved(__abs_lowcore); unsigned long __bootdata_preserved(__memcpy_real_area); pte_t *__bootdata_preserved(memcpy_real_ptep); -unsigned long __bootdata(__amode31_base); unsigned long __bootdata_preserved(VMALLOC_START); unsigned long __bootdata_preserved(VMALLOC_END); struct page *__bootdata_preserved(vmemmap); @@ -29,8 +28,6 @@ unsigned long __bootdata_preserved(vmemmap_size); unsigned long __bootdata_preserved(MODULES_VADDR); unsigned long __bootdata_preserved(MODULES_END); unsigned long __bootdata(ident_map_size); -int __bootdata(is_full_image) = 1; -struct initrd_data __bootdata(initrd_data); u64 __bootdata_preserved(stfle_fac_list[16]); u64 __bootdata_preserved(alt_stfle_fac_list[16]); @@ -76,17 +73,20 @@ unsigned long mem_safe_offset(void) } #endif -static unsigned long rescue_initrd(unsigned long safe_addr) +static void rescue_initrd(unsigned long min, unsigned long max) { + unsigned long old_addr, addr, size; + if (!IS_ENABLED(CONFIG_BLK_DEV_INITRD)) - return safe_addr; - if (!initrd_data.start || !initrd_data.size) - return safe_addr; - if (initrd_data.start < safe_addr) { - memmove((void *)safe_addr, (void *)initrd_data.start, initrd_data.size); - initrd_data.start = safe_addr; - } - return initrd_data.start + initrd_data.size; + return; + if (!get_physmem_reserved(RR_INITRD, &addr, &size)) + return; + if (addr >= min && addr + size <= max) + return; + old_addr = addr; + physmem_free(RR_INITRD); + addr = physmem_alloc_top_down(RR_INITRD, size, 0); + memmove((void *)addr, (void *)old_addr, size); } static void copy_bootdata(void) @@ -140,7 +140,7 @@ static void handle_relocs(unsigned long offset) * * Consider the following factors: * 1. max_physmem_end - end of physical memory online or standby. - * Always <= end of the last online memory block (get_mem_detect_end()). + * Always >= end of the last online memory range (get_physmem_online_end()). * 2. CONFIG_MAX_PHYSMEM_BITS - the maximum size of physical memory the * kernel is able to support. * 3. "mem=" kernel command line option which limits physical memory usage. @@ -160,10 +160,10 @@ static void setup_ident_map_size(unsigned long max_physmem_end) #ifdef CONFIG_CRASH_DUMP if (oldmem_data.start) { - kaslr_enabled = 0; + __kaslr_enabled = 0; ident_map_size = min(ident_map_size, oldmem_data.size); } else if (ipl_block_valid && is_ipl_block_dump()) { - kaslr_enabled = 0; + __kaslr_enabled = 0; if (!sclp_early_get_hsa_size(&hsa_size) && hsa_size) ident_map_size = min(ident_map_size, hsa_size); } @@ -235,9 +235,9 @@ static unsigned long setup_kernel_memory_layout(void) /* * This function clears the BSS section of the decompressed Linux kernel and NOT the decompressor's. */ -static void clear_bss_section(void) +static void clear_bss_section(unsigned long vmlinux_lma) { - memset((void *)vmlinux.default_lma + vmlinux.image_size, 0, vmlinux.bss_size); + memset((void *)vmlinux_lma + vmlinux.image_size, 0, vmlinux.bss_size); } /* @@ -256,7 +256,6 @@ static void setup_vmalloc_size(void) static void offset_vmlinux_info(unsigned long offset) { - vmlinux.default_lma += offset; *(unsigned long *)(&vmlinux.entry) += offset; vmlinux.bootdata_off += offset; vmlinux.bootdata_preserved_off += offset; @@ -266,60 +265,83 @@ static void offset_vmlinux_info(unsigned long offset) vmlinux.init_mm_off += offset; vmlinux.swapper_pg_dir_off += offset; vmlinux.invalid_pg_dir_off += offset; -} - -static unsigned long reserve_amode31(unsigned long safe_addr) -{ - __amode31_base = PAGE_ALIGN(safe_addr); - return __amode31_base + vmlinux.amode31_size; +#ifdef CONFIG_KASAN + vmlinux.kasan_early_shadow_page_off += offset; + vmlinux.kasan_early_shadow_pte_off += offset; + vmlinux.kasan_early_shadow_pmd_off += offset; + vmlinux.kasan_early_shadow_pud_off += offset; + vmlinux.kasan_early_shadow_p4d_off += offset; +#endif } void startup_kernel(void) { unsigned long max_physmem_end; - unsigned long random_lma; - unsigned long safe_addr; + unsigned long vmlinux_lma = 0; + unsigned long amode31_lma = 0; unsigned long asce_limit; + unsigned long safe_addr; void *img; psw_t psw; - initrd_data.start = parmarea.initrd_start; - initrd_data.size = parmarea.initrd_size; + setup_lpp(); + safe_addr = mem_safe_offset(); + /* + * reserve decompressor memory together with decompression heap, buffer and + * memory which might be occupied by uncompressed kernel at default 1Mb + * position (if KASLR is off or failed). + */ + physmem_reserve(RR_DECOMPRESSOR, 0, safe_addr); + if (IS_ENABLED(CONFIG_BLK_DEV_INITRD) && parmarea.initrd_size) + physmem_reserve(RR_INITRD, parmarea.initrd_start, parmarea.initrd_size); oldmem_data.start = parmarea.oldmem_base; oldmem_data.size = parmarea.oldmem_size; - setup_lpp(); store_ipl_parmblock(); - safe_addr = mem_safe_offset(); - safe_addr = reserve_amode31(safe_addr); - safe_addr = read_ipl_report(safe_addr); + read_ipl_report(); uv_query_info(); - safe_addr = rescue_initrd(safe_addr); sclp_early_read_info(); setup_boot_command_line(); parse_boot_command_line(); detect_facilities(); sanitize_prot_virt_host(); - max_physmem_end = detect_memory(&safe_addr); + max_physmem_end = detect_max_physmem_end(); setup_ident_map_size(max_physmem_end); setup_vmalloc_size(); asce_limit = setup_kernel_memory_layout(); - mem_detect_set_usable_limit(ident_map_size); - - if (IS_ENABLED(CONFIG_RANDOMIZE_BASE) && kaslr_enabled) { - random_lma = get_random_base(safe_addr); - if (random_lma) { - __kaslr_offset = random_lma - vmlinux.default_lma; - img = (void *)vmlinux.default_lma; + /* got final ident_map_size, physmem allocations could be performed now */ + physmem_set_usable_limit(ident_map_size); + detect_physmem_online_ranges(max_physmem_end); + save_ipl_cert_comp_list(); + rescue_initrd(safe_addr, ident_map_size); + + if (kaslr_enabled()) { + vmlinux_lma = randomize_within_range(vmlinux.image_size + vmlinux.bss_size, + THREAD_SIZE, vmlinux.default_lma, + ident_map_size); + if (vmlinux_lma) { + __kaslr_offset = vmlinux_lma - vmlinux.default_lma; offset_vmlinux_info(__kaslr_offset); } } + vmlinux_lma = vmlinux_lma ?: vmlinux.default_lma; + physmem_reserve(RR_VMLINUX, vmlinux_lma, vmlinux.image_size + vmlinux.bss_size); if (!IS_ENABLED(CONFIG_KERNEL_UNCOMPRESSED)) { img = decompress_kernel(); - memmove((void *)vmlinux.default_lma, img, vmlinux.image_size); - } else if (__kaslr_offset) - memcpy((void *)vmlinux.default_lma, img, vmlinux.image_size); + memmove((void *)vmlinux_lma, img, vmlinux.image_size); + } else if (__kaslr_offset) { + img = (void *)vmlinux.default_lma; + memmove((void *)vmlinux_lma, img, vmlinux.image_size); + memset(img, 0, vmlinux.image_size); + } + + /* vmlinux decompression is done, shrink reserved low memory */ + physmem_reserve(RR_DECOMPRESSOR, 0, (unsigned long)_decompressor_end); + if (kaslr_enabled()) + amode31_lma = randomize_within_range(vmlinux.amode31_size, PAGE_SIZE, 0, SZ_2G); + amode31_lma = amode31_lma ?: vmlinux.default_lma - vmlinux.amode31_size; + physmem_reserve(RR_AMODE31, amode31_lma, vmlinux.amode31_size); /* * The order of the following operations is important: @@ -334,21 +356,16 @@ void startup_kernel(void) * - copy_bootdata() must follow setup_vmem() to propagate changes to * bootdata made by setup_vmem() */ - clear_bss_section(); + clear_bss_section(vmlinux_lma); handle_relocs(__kaslr_offset); setup_vmem(asce_limit); copy_bootdata(); - if (__kaslr_offset) { - /* - * Save KASLR offset for early dumps, before vmcore_info is set. - * Mark as uneven to distinguish from real vmcore_info pointer. - */ - S390_lowcore.vmcore_info = __kaslr_offset | 0x1UL; - /* Clear non-relocated kernel */ - if (IS_ENABLED(CONFIG_KERNEL_UNCOMPRESSED)) - memset(img, 0, vmlinux.image_size); - } + /* + * Save KASLR offset for early dumps, before vmcore_info is set. + * Mark as uneven to distinguish from real vmcore_info pointer. + */ + S390_lowcore.vmcore_info = __kaslr_offset ? __kaslr_offset | 0x1UL : 0; /* * Jump to the decompressed kernel entry point and switch DAT mode on. diff --git a/arch/s390/boot/vmem.c b/arch/s390/boot/vmem.c index 4d1d0d8e99cb..acb1f8b53105 100644 --- a/arch/s390/boot/vmem.c +++ b/arch/s390/boot/vmem.c @@ -1,81 +1,217 @@ // SPDX-License-Identifier: GPL-2.0 #include <linux/sched/task.h> #include <linux/pgtable.h> +#include <linux/kasan.h> #include <asm/pgalloc.h> #include <asm/facility.h> #include <asm/sections.h> -#include <asm/mem_detect.h> +#include <asm/physmem_info.h> #include <asm/maccess.h> #include <asm/abs_lowcore.h> #include "decompressor.h" #include "boot.h" +unsigned long __bootdata_preserved(s390_invalid_asce); + +#ifdef CONFIG_PROC_FS +atomic_long_t __bootdata_preserved(direct_pages_count[PG_DIRECT_MAP_MAX]); +#endif + #define init_mm (*(struct mm_struct *)vmlinux.init_mm_off) #define swapper_pg_dir vmlinux.swapper_pg_dir_off #define invalid_pg_dir vmlinux.invalid_pg_dir_off -/* - * Mimic virt_to_kpte() in lack of init_mm symbol. Skip pmd NULL check though. - */ -static inline pte_t *__virt_to_kpte(unsigned long va) -{ - return pte_offset_kernel(pmd_offset(pud_offset(p4d_offset(pgd_offset_k(va), va), va), va), va); -} - -unsigned long __bootdata_preserved(s390_invalid_asce); -unsigned long __bootdata(pgalloc_pos); -unsigned long __bootdata(pgalloc_end); -unsigned long __bootdata(pgalloc_low); - enum populate_mode { POPULATE_NONE, - POPULATE_ONE2ONE, + POPULATE_DIRECT, POPULATE_ABS_LOWCORE, +#ifdef CONFIG_KASAN + POPULATE_KASAN_MAP_SHADOW, + POPULATE_KASAN_ZERO_SHADOW, + POPULATE_KASAN_SHALLOW +#endif }; -static void boot_check_oom(void) +static void pgtable_populate(unsigned long addr, unsigned long end, enum populate_mode mode); + +#ifdef CONFIG_KASAN + +#define kasan_early_shadow_page vmlinux.kasan_early_shadow_page_off +#define kasan_early_shadow_pte ((pte_t *)vmlinux.kasan_early_shadow_pte_off) +#define kasan_early_shadow_pmd ((pmd_t *)vmlinux.kasan_early_shadow_pmd_off) +#define kasan_early_shadow_pud ((pud_t *)vmlinux.kasan_early_shadow_pud_off) +#define kasan_early_shadow_p4d ((p4d_t *)vmlinux.kasan_early_shadow_p4d_off) +#define __sha(x) ((unsigned long)kasan_mem_to_shadow((void *)x)) + +static pte_t pte_z; + +static void kasan_populate_shadow(void) +{ + pmd_t pmd_z = __pmd(__pa(kasan_early_shadow_pte) | _SEGMENT_ENTRY); + pud_t pud_z = __pud(__pa(kasan_early_shadow_pmd) | _REGION3_ENTRY); + p4d_t p4d_z = __p4d(__pa(kasan_early_shadow_pud) | _REGION2_ENTRY); + unsigned long untracked_end; + unsigned long start, end; + int i; + + pte_z = __pte(__pa(kasan_early_shadow_page) | pgprot_val(PAGE_KERNEL_RO)); + if (!machine.has_nx) + pte_z = clear_pte_bit(pte_z, __pgprot(_PAGE_NOEXEC)); + crst_table_init((unsigned long *)kasan_early_shadow_p4d, p4d_val(p4d_z)); + crst_table_init((unsigned long *)kasan_early_shadow_pud, pud_val(pud_z)); + crst_table_init((unsigned long *)kasan_early_shadow_pmd, pmd_val(pmd_z)); + memset64((u64 *)kasan_early_shadow_pte, pte_val(pte_z), PTRS_PER_PTE); + + /* + * Current memory layout: + * +- 0 -------------+ +- shadow start -+ + * |1:1 ident mapping| /|1/8 of ident map| + * | | / | | + * +-end of ident map+ / +----------------+ + * | ... gap ... | / | kasan | + * | | / | zero page | + * +- vmalloc area -+ / | mapping | + * | vmalloc_size | / | (untracked) | + * +- modules vaddr -+ / +----------------+ + * | 2Gb |/ | unmapped | allocated per module + * +- shadow start -+ +----------------+ + * | 1/8 addr space | | zero pg mapping| (untracked) + * +- shadow end ----+---------+- shadow end ---+ + * + * Current memory layout (KASAN_VMALLOC): + * +- 0 -------------+ +- shadow start -+ + * |1:1 ident mapping| /|1/8 of ident map| + * | | / | | + * +-end of ident map+ / +----------------+ + * | ... gap ... | / | kasan zero page| (untracked) + * | | / | mapping | + * +- vmalloc area -+ / +----------------+ + * | vmalloc_size | / |shallow populate| + * +- modules vaddr -+ / +----------------+ + * | 2Gb |/ |shallow populate| + * +- shadow start -+ +----------------+ + * | 1/8 addr space | | zero pg mapping| (untracked) + * +- shadow end ----+---------+- shadow end ---+ + */ + + for_each_physmem_usable_range(i, &start, &end) + pgtable_populate(__sha(start), __sha(end), POPULATE_KASAN_MAP_SHADOW); + if (IS_ENABLED(CONFIG_KASAN_VMALLOC)) { + untracked_end = VMALLOC_START; + /* shallowly populate kasan shadow for vmalloc and modules */ + pgtable_populate(__sha(VMALLOC_START), __sha(MODULES_END), POPULATE_KASAN_SHALLOW); + } else { + untracked_end = MODULES_VADDR; + } + /* populate kasan shadow for untracked memory */ + pgtable_populate(__sha(ident_map_size), __sha(untracked_end), POPULATE_KASAN_ZERO_SHADOW); + pgtable_populate(__sha(MODULES_END), __sha(_REGION1_SIZE), POPULATE_KASAN_ZERO_SHADOW); +} + +static bool kasan_pgd_populate_zero_shadow(pgd_t *pgd, unsigned long addr, + unsigned long end, enum populate_mode mode) { - if (pgalloc_pos < pgalloc_low) - error("out of memory on boot\n"); + if (mode == POPULATE_KASAN_ZERO_SHADOW && + IS_ALIGNED(addr, PGDIR_SIZE) && end - addr >= PGDIR_SIZE) { + pgd_populate(&init_mm, pgd, kasan_early_shadow_p4d); + return true; + } + return false; } -static void pgtable_populate_init(void) +static bool kasan_p4d_populate_zero_shadow(p4d_t *p4d, unsigned long addr, + unsigned long end, enum populate_mode mode) { - unsigned long initrd_end; - unsigned long kernel_end; - - kernel_end = vmlinux.default_lma + vmlinux.image_size + vmlinux.bss_size; - pgalloc_low = round_up(kernel_end, PAGE_SIZE); - if (IS_ENABLED(CONFIG_BLK_DEV_INITRD)) { - initrd_end = round_up(initrd_data.start + initrd_data.size, _SEGMENT_SIZE); - pgalloc_low = max(pgalloc_low, initrd_end); + if (mode == POPULATE_KASAN_ZERO_SHADOW && + IS_ALIGNED(addr, P4D_SIZE) && end - addr >= P4D_SIZE) { + p4d_populate(&init_mm, p4d, kasan_early_shadow_pud); + return true; } + return false; +} - pgalloc_end = round_down(get_mem_detect_end(), PAGE_SIZE); - pgalloc_pos = pgalloc_end; +static bool kasan_pud_populate_zero_shadow(pud_t *pud, unsigned long addr, + unsigned long end, enum populate_mode mode) +{ + if (mode == POPULATE_KASAN_ZERO_SHADOW && + IS_ALIGNED(addr, PUD_SIZE) && end - addr >= PUD_SIZE) { + pud_populate(&init_mm, pud, kasan_early_shadow_pmd); + return true; + } + return false; +} - boot_check_oom(); +static bool kasan_pmd_populate_zero_shadow(pmd_t *pmd, unsigned long addr, + unsigned long end, enum populate_mode mode) +{ + if (mode == POPULATE_KASAN_ZERO_SHADOW && + IS_ALIGNED(addr, PMD_SIZE) && end - addr >= PMD_SIZE) { + pmd_populate(&init_mm, pmd, kasan_early_shadow_pte); + return true; + } + return false; } -static void *boot_alloc_pages(unsigned int order) +static bool kasan_pte_populate_zero_shadow(pte_t *pte, enum populate_mode mode) { - unsigned long size = PAGE_SIZE << order; + pte_t entry; - pgalloc_pos -= size; - pgalloc_pos = round_down(pgalloc_pos, size); + if (mode == POPULATE_KASAN_ZERO_SHADOW) { + set_pte(pte, pte_z); + return true; + } + return false; +} +#else - boot_check_oom(); +static inline void kasan_populate_shadow(void) {} - return (void *)pgalloc_pos; +static inline bool kasan_pgd_populate_zero_shadow(pgd_t *pgd, unsigned long addr, + unsigned long end, enum populate_mode mode) +{ + return false; +} + +static inline bool kasan_p4d_populate_zero_shadow(p4d_t *p4d, unsigned long addr, + unsigned long end, enum populate_mode mode) +{ + return false; +} + +static inline bool kasan_pud_populate_zero_shadow(pud_t *pud, unsigned long addr, + unsigned long end, enum populate_mode mode) +{ + return false; +} + +static inline bool kasan_pmd_populate_zero_shadow(pmd_t *pmd, unsigned long addr, + unsigned long end, enum populate_mode mode) +{ + return false; +} + +static bool kasan_pte_populate_zero_shadow(pte_t *pte, enum populate_mode mode) +{ + return false; +} + +#endif + +/* + * Mimic virt_to_kpte() in lack of init_mm symbol. Skip pmd NULL check though. + */ +static inline pte_t *__virt_to_kpte(unsigned long va) +{ + return pte_offset_kernel(pmd_offset(pud_offset(p4d_offset(pgd_offset_k(va), va), va), va), va); } static void *boot_crst_alloc(unsigned long val) { + unsigned long size = PAGE_SIZE << CRST_ALLOC_ORDER; unsigned long *table; - table = boot_alloc_pages(CRST_ALLOC_ORDER); - if (table) - crst_table_init(table, val); + table = (unsigned long *)physmem_alloc_top_down(RR_VMEM, size, size); + crst_table_init(table, val); return table; } @@ -84,28 +220,37 @@ static pte_t *boot_pte_alloc(void) static void *pte_leftover; pte_t *pte; - BUILD_BUG_ON(_PAGE_TABLE_SIZE * 2 != PAGE_SIZE); - + /* + * handling pte_leftovers this way helps to avoid memory fragmentation + * during POPULATE_KASAN_MAP_SHADOW when EDAT is off + */ if (!pte_leftover) { - pte_leftover = boot_alloc_pages(0); + pte_leftover = (void *)physmem_alloc_top_down(RR_VMEM, PAGE_SIZE, PAGE_SIZE); pte = pte_leftover + _PAGE_TABLE_SIZE; } else { pte = pte_leftover; pte_leftover = NULL; } + memset64((u64 *)pte, _PAGE_INVALID, PTRS_PER_PTE); return pte; } -static unsigned long _pa(unsigned long addr, enum populate_mode mode) +static unsigned long _pa(unsigned long addr, unsigned long size, enum populate_mode mode) { switch (mode) { case POPULATE_NONE: return -1; - case POPULATE_ONE2ONE: + case POPULATE_DIRECT: return addr; case POPULATE_ABS_LOWCORE: return __abs_lowcore_pa(addr); +#ifdef CONFIG_KASAN + case POPULATE_KASAN_MAP_SHADOW: + addr = physmem_alloc_top_down(RR_VMEM, size, size); + memset((void *)addr, 0, size); + return addr; +#endif default: return -1; } @@ -126,23 +271,28 @@ static bool can_large_pmd(pmd_t *pm_dir, unsigned long addr, unsigned long end) static void pgtable_pte_populate(pmd_t *pmd, unsigned long addr, unsigned long end, enum populate_mode mode) { - unsigned long next; + unsigned long pages = 0; pte_t *pte, entry; pte = pte_offset_kernel(pmd, addr); for (; addr < end; addr += PAGE_SIZE, pte++) { if (pte_none(*pte)) { - entry = __pte(_pa(addr, mode)); + if (kasan_pte_populate_zero_shadow(pte, mode)) + continue; + entry = __pte(_pa(addr, PAGE_SIZE, mode)); entry = set_pte_bit(entry, PAGE_KERNEL_EXEC); set_pte(pte, entry); + pages++; } } + if (mode == POPULATE_DIRECT) + update_page_count(PG_DIRECT_MAP_4K, pages); } static void pgtable_pmd_populate(pud_t *pud, unsigned long addr, unsigned long end, enum populate_mode mode) { - unsigned long next; + unsigned long next, pages = 0; pmd_t *pmd, entry; pte_t *pte; @@ -150,10 +300,13 @@ static void pgtable_pmd_populate(pud_t *pud, unsigned long addr, unsigned long e for (; addr < end; addr = next, pmd++) { next = pmd_addr_end(addr, end); if (pmd_none(*pmd)) { + if (kasan_pmd_populate_zero_shadow(pmd, addr, next, mode)) + continue; if (can_large_pmd(pmd, addr, next)) { - entry = __pmd(_pa(addr, mode)); + entry = __pmd(_pa(addr, _SEGMENT_SIZE, mode)); entry = set_pmd_bit(entry, SEGMENT_KERNEL_EXEC); set_pmd(pmd, entry); + pages++; continue; } pte = boot_pte_alloc(); @@ -163,12 +316,14 @@ static void pgtable_pmd_populate(pud_t *pud, unsigned long addr, unsigned long e } pgtable_pte_populate(pmd, addr, next, mode); } + if (mode == POPULATE_DIRECT) + update_page_count(PG_DIRECT_MAP_1M, pages); } static void pgtable_pud_populate(p4d_t *p4d, unsigned long addr, unsigned long end, enum populate_mode mode) { - unsigned long next; + unsigned long next, pages = 0; pud_t *pud, entry; pmd_t *pmd; @@ -176,10 +331,13 @@ static void pgtable_pud_populate(p4d_t *p4d, unsigned long addr, unsigned long e for (; addr < end; addr = next, pud++) { next = pud_addr_end(addr, end); if (pud_none(*pud)) { + if (kasan_pud_populate_zero_shadow(pud, addr, next, mode)) + continue; if (can_large_pud(pud, addr, next)) { - entry = __pud(_pa(addr, mode)); + entry = __pud(_pa(addr, _REGION3_SIZE, mode)); entry = set_pud_bit(entry, REGION3_KERNEL_EXEC); set_pud(pud, entry); + pages++; continue; } pmd = boot_crst_alloc(_SEGMENT_ENTRY_EMPTY); @@ -189,6 +347,8 @@ static void pgtable_pud_populate(p4d_t *p4d, unsigned long addr, unsigned long e } pgtable_pmd_populate(pud, addr, next, mode); } + if (mode == POPULATE_DIRECT) + update_page_count(PG_DIRECT_MAP_2G, pages); } static void pgtable_p4d_populate(pgd_t *pgd, unsigned long addr, unsigned long end, @@ -202,6 +362,8 @@ static void pgtable_p4d_populate(pgd_t *pgd, unsigned long addr, unsigned long e for (; addr < end; addr = next, p4d++) { next = p4d_addr_end(addr, end); if (p4d_none(*p4d)) { + if (kasan_p4d_populate_zero_shadow(p4d, addr, next, mode)) + continue; pud = boot_crst_alloc(_REGION3_ENTRY_EMPTY); p4d_populate(&init_mm, p4d, pud); } @@ -219,9 +381,15 @@ static void pgtable_populate(unsigned long addr, unsigned long end, enum populat for (; addr < end; addr = next, pgd++) { next = pgd_addr_end(addr, end); if (pgd_none(*pgd)) { + if (kasan_pgd_populate_zero_shadow(pgd, addr, next, mode)) + continue; p4d = boot_crst_alloc(_REGION2_ENTRY_EMPTY); pgd_populate(&init_mm, pgd, p4d); } +#ifdef CONFIG_KASAN + if (mode == POPULATE_KASAN_SHALLOW) + continue; +#endif pgtable_p4d_populate(pgd, addr, next, mode); } } @@ -250,16 +418,17 @@ void setup_vmem(unsigned long asce_limit) * To prevent creation of a large page at address 0 first map * the lowcore and create the identity mapping only afterwards. */ - pgtable_populate_init(); - pgtable_populate(0, sizeof(struct lowcore), POPULATE_ONE2ONE); - for_each_mem_detect_usable_block(i, &start, &end) - pgtable_populate(start, end, POPULATE_ONE2ONE); + pgtable_populate(0, sizeof(struct lowcore), POPULATE_DIRECT); + for_each_physmem_usable_range(i, &start, &end) + pgtable_populate(start, end, POPULATE_DIRECT); pgtable_populate(__abs_lowcore, __abs_lowcore + sizeof(struct lowcore), POPULATE_ABS_LOWCORE); pgtable_populate(__memcpy_real_area, __memcpy_real_area + PAGE_SIZE, POPULATE_NONE); memcpy_real_ptep = __virt_to_kpte(__memcpy_real_area); + kasan_populate_shadow(); + S390_lowcore.kernel_asce = swapper_pg_dir | asce_bits; S390_lowcore.user_asce = s390_invalid_asce; @@ -269,10 +438,3 @@ void setup_vmem(unsigned long asce_limit) init_mm.context.asce = S390_lowcore.kernel_asce; } - -unsigned long vmem_estimate_memory_needs(unsigned long online_mem_total) -{ - unsigned long pages = DIV_ROUND_UP(online_mem_total, PAGE_SIZE); - - return DIV_ROUND_UP(pages, _PAGE_ENTRIES) * _PAGE_TABLE_SIZE * 2; -} diff --git a/arch/s390/boot/vmlinux.lds.S b/arch/s390/boot/vmlinux.lds.S index fa9d33b01b85..389df0e0d9e5 100644 --- a/arch/s390/boot/vmlinux.lds.S +++ b/arch/s390/boot/vmlinux.lds.S @@ -93,6 +93,8 @@ SECTIONS _decompressor_syms_end = .; } + _decompressor_end = .; + #ifdef CONFIG_KERNEL_UNCOMPRESSED . = 0x100000; #else diff --git a/arch/s390/crypto/chacha-s390.S b/arch/s390/crypto/chacha-s390.S index 9b033622191c..37cb63f25b17 100644 --- a/arch/s390/crypto/chacha-s390.S +++ b/arch/s390/crypto/chacha-s390.S @@ -13,27 +13,28 @@ #define SP %r15 #define FRAME (16 * 8 + 4 * 8) -.data -.align 32 - -.Lsigma: -.long 0x61707865,0x3320646e,0x79622d32,0x6b206574 # endian-neutral -.long 1,0,0,0 -.long 2,0,0,0 -.long 3,0,0,0 -.long 0x03020100,0x07060504,0x0b0a0908,0x0f0e0d0c # byte swap - -.long 0,1,2,3 -.long 0x61707865,0x61707865,0x61707865,0x61707865 # smashed sigma -.long 0x3320646e,0x3320646e,0x3320646e,0x3320646e -.long 0x79622d32,0x79622d32,0x79622d32,0x79622d32 -.long 0x6b206574,0x6b206574,0x6b206574,0x6b206574 + .data + .balign 32 -.previous +SYM_DATA_START_LOCAL(sigma) + .long 0x61707865,0x3320646e,0x79622d32,0x6b206574 # endian-neutral + .long 1,0,0,0 + .long 2,0,0,0 + .long 3,0,0,0 + .long 0x03020100,0x07060504,0x0b0a0908,0x0f0e0d0c # byte swap + + .long 0,1,2,3 + .long 0x61707865,0x61707865,0x61707865,0x61707865 # smashed sigma + .long 0x3320646e,0x3320646e,0x3320646e,0x3320646e + .long 0x79622d32,0x79622d32,0x79622d32,0x79622d32 + .long 0x6b206574,0x6b206574,0x6b206574,0x6b206574 +SYM_DATA_END(sigma) + + .previous GEN_BR_THUNK %r14 -.text + .text ############################################################################# # void chacha20_vx_4x(u8 *out, counst u8 *inp, size_t len, @@ -78,10 +79,10 @@ #define XT2 %v29 #define XT3 %v30 -ENTRY(chacha20_vx_4x) +SYM_FUNC_START(chacha20_vx_4x) stmg %r6,%r7,6*8(SP) - larl %r7,.Lsigma + larl %r7,sigma lhi %r0,10 lhi %r1,0 @@ -403,7 +404,7 @@ ENTRY(chacha20_vx_4x) lmg %r6,%r7,6*8(SP) BR_EX %r14 -ENDPROC(chacha20_vx_4x) +SYM_FUNC_END(chacha20_vx_4x) #undef OUT #undef INP @@ -471,7 +472,7 @@ ENDPROC(chacha20_vx_4x) #define T2 %v29 #define T3 %v30 -ENTRY(chacha20_vx) +SYM_FUNC_START(chacha20_vx) clgfi LEN,256 jle chacha20_vx_4x stmg %r6,%r7,6*8(SP) @@ -481,7 +482,7 @@ ENTRY(chacha20_vx) la SP,0(%r1,SP) stg %r0,0(SP) # back-chain - larl %r7,.Lsigma + larl %r7,sigma lhi %r0,10 VLM K1,K2,0,KEY,0 # load key @@ -902,6 +903,6 @@ ENTRY(chacha20_vx) lmg %r6,%r7,FRAME+6*8(SP) la SP,FRAME(SP) BR_EX %r14 -ENDPROC(chacha20_vx) +SYM_FUNC_END(chacha20_vx) .previous diff --git a/arch/s390/crypto/crc32be-vx.S b/arch/s390/crypto/crc32be-vx.S index 6b3d1009c392..6ea17628ea10 100644 --- a/arch/s390/crypto/crc32be-vx.S +++ b/arch/s390/crypto/crc32be-vx.S @@ -24,8 +24,8 @@ #define CONST_RU_POLY %v13 #define CONST_CRC_POLY %v14 -.data -.align 8 + .data + .balign 8 /* * The CRC-32 constant block contains reduction constants to fold and @@ -58,19 +58,20 @@ * P'(x) = 0xEDB88320 */ -.Lconstants_CRC_32_BE: +SYM_DATA_START_LOCAL(constants_CRC_32_BE) .quad 0x08833794c, 0x0e6228b11 # R1, R2 .quad 0x0c5b9cd4c, 0x0e8a45605 # R3, R4 .quad 0x0f200aa66, 1 << 32 # R5, x32 .quad 0x0490d678d, 1 # R6, 1 .quad 0x104d101df, 0 # u .quad 0x104C11DB7, 0 # P(x) +SYM_DATA_END(constants_CRC_32_BE) -.previous + .previous GEN_BR_THUNK %r14 -.text + .text /* * The CRC-32 function(s) use these calling conventions: * @@ -90,9 +91,9 @@ * * V9..V14: CRC-32 constants. */ -ENTRY(crc32_be_vgfm_16) +SYM_FUNC_START(crc32_be_vgfm_16) /* Load CRC-32 constants */ - larl %r5,.Lconstants_CRC_32_BE + larl %r5,constants_CRC_32_BE VLM CONST_R1R2,CONST_CRC_POLY,0,%r5 /* Load the initial CRC value into the leftmost word of V0. */ @@ -207,6 +208,6 @@ ENTRY(crc32_be_vgfm_16) .Ldone: VLGVF %r2,%v2,3 BR_EX %r14 -ENDPROC(crc32_be_vgfm_16) +SYM_FUNC_END(crc32_be_vgfm_16) .previous diff --git a/arch/s390/crypto/crc32le-vx.S b/arch/s390/crypto/crc32le-vx.S index 71caf0f4ec08..5a819ae09a0b 100644 --- a/arch/s390/crypto/crc32le-vx.S +++ b/arch/s390/crypto/crc32le-vx.S @@ -25,8 +25,8 @@ #define CONST_RU_POLY %v13 #define CONST_CRC_POLY %v14 -.data -.align 8 + .data + .balign 8 /* * The CRC-32 constant block contains reduction constants to fold and @@ -59,27 +59,29 @@ * P'(x) = 0x82F63B78 */ -.Lconstants_CRC_32_LE: +SYM_DATA_START_LOCAL(constants_CRC_32_LE) .octa 0x0F0E0D0C0B0A09080706050403020100 # BE->LE mask .quad 0x1c6e41596, 0x154442bd4 # R2, R1 .quad 0x0ccaa009e, 0x1751997d0 # R4, R3 .octa 0x163cd6124 # R5 .octa 0x1F7011641 # u' .octa 0x1DB710641 # P'(x) << 1 +SYM_DATA_END(constants_CRC_32_LE) -.Lconstants_CRC_32C_LE: +SYM_DATA_START_LOCAL(constants_CRC_32C_LE) .octa 0x0F0E0D0C0B0A09080706050403020100 # BE->LE mask .quad 0x09e4addf8, 0x740eef02 # R2, R1 .quad 0x14cd00bd6, 0xf20c0dfe # R4, R3 .octa 0x0dd45aab8 # R5 .octa 0x0dea713f1 # u' .octa 0x105ec76f0 # P'(x) << 1 +SYM_DATA_END(constants_CRC_32C_LE) -.previous + .previous GEN_BR_THUNK %r14 -.text + .text /* * The CRC-32 functions use these calling conventions: @@ -102,17 +104,17 @@ * V10..V14: CRC-32 constants. */ -ENTRY(crc32_le_vgfm_16) - larl %r5,.Lconstants_CRC_32_LE +SYM_FUNC_START(crc32_le_vgfm_16) + larl %r5,constants_CRC_32_LE j crc32_le_vgfm_generic -ENDPROC(crc32_le_vgfm_16) +SYM_FUNC_END(crc32_le_vgfm_16) -ENTRY(crc32c_le_vgfm_16) - larl %r5,.Lconstants_CRC_32C_LE +SYM_FUNC_START(crc32c_le_vgfm_16) + larl %r5,constants_CRC_32C_LE j crc32_le_vgfm_generic -ENDPROC(crc32c_le_vgfm_16) +SYM_FUNC_END(crc32c_le_vgfm_16) -ENTRY(crc32_le_vgfm_generic) +SYM_FUNC_START(crc32_le_vgfm_generic) /* Load CRC-32 constants */ VLM CONST_PERM_LE2BE,CONST_CRC_POLY,0,%r5 @@ -268,6 +270,6 @@ ENTRY(crc32_le_vgfm_generic) .Ldone: VLGVF %r2,%v2,2 BR_EX %r14 -ENDPROC(crc32_le_vgfm_generic) +SYM_FUNC_END(crc32_le_vgfm_generic) .previous diff --git a/arch/s390/include/asm/ap.h b/arch/s390/include/asm/ap.h index c699f251a464..d5d967166bac 100644 --- a/arch/s390/include/asm/ap.h +++ b/arch/s390/include/asm/ap.h @@ -43,10 +43,11 @@ struct ap_queue_status { unsigned int queue_empty : 1; unsigned int replies_waiting : 1; unsigned int queue_full : 1; - unsigned int _pad1 : 4; + unsigned int : 3; + unsigned int async : 1; unsigned int irq_enabled : 1; unsigned int response_code : 8; - unsigned int _pad2 : 16; + unsigned int : 16; }; /* @@ -86,6 +87,42 @@ static inline bool ap_instructions_available(void) return reg1 != 0; } +/* TAPQ register GR2 response struct */ +struct ap_tapq_gr2 { + union { + unsigned long value; + struct { + unsigned int fac : 32; /* facility bits */ + unsigned int apinfo : 32; /* ap type, ... */ + }; + struct { + unsigned int s : 1; /* APSC */ + unsigned int m : 1; /* AP4KM */ + unsigned int c : 1; /* AP4KC */ + unsigned int mode : 3; + unsigned int n : 1; /* APXA */ + unsigned int : 1; + unsigned int class : 8; + unsigned int bs : 2; /* SE bind/assoc */ + unsigned int : 14; + unsigned int at : 8; /* ap type */ + unsigned int nd : 8; /* nr of domains */ + unsigned int : 4; + unsigned int ml : 4; /* apxl ml */ + unsigned int : 4; + unsigned int qd : 4; /* queue depth */ + }; + }; +}; + +/* + * Convenience defines to be used with the bs field from struct ap_tapq_gr2 + */ +#define AP_BS_Q_USABLE 0 +#define AP_BS_Q_USABLE_NO_SECURE_KEY 1 +#define AP_BS_Q_AVAIL_FOR_BINDING 2 +#define AP_BS_Q_UNUSABLE 3 + /** * ap_tapq(): Test adjunct processor queue. * @qid: The AP queue number @@ -93,7 +130,7 @@ static inline bool ap_instructions_available(void) * * Returns AP queue status structure. */ -static inline struct ap_queue_status ap_tapq(ap_qid_t qid, unsigned long *info) +static inline struct ap_queue_status ap_tapq(ap_qid_t qid, struct ap_tapq_gr2 *info) { union ap_queue_status_reg reg1; unsigned long reg2; @@ -108,7 +145,7 @@ static inline struct ap_queue_status ap_tapq(ap_qid_t qid, unsigned long *info) : [qid] "d" (qid) : "cc", "0", "1", "2"); if (info) - *info = reg2; + info->value = reg2; return reg1.status; } @@ -116,13 +153,12 @@ static inline struct ap_queue_status ap_tapq(ap_qid_t qid, unsigned long *info) * ap_test_queue(): Test adjunct processor queue. * @qid: The AP queue number * @tbit: Test facilities bit - * @info: Pointer to queue descriptor + * @info: Ptr to tapq gr2 struct * * Returns AP queue status structure. */ -static inline struct ap_queue_status ap_test_queue(ap_qid_t qid, - int tbit, - unsigned long *info) +static inline struct ap_queue_status ap_test_queue(ap_qid_t qid, int tbit, + struct ap_tapq_gr2 *info) { if (tbit) qid |= 1UL << 23; /* set T bit*/ @@ -132,14 +168,18 @@ static inline struct ap_queue_status ap_test_queue(ap_qid_t qid, /** * ap_pqap_rapq(): Reset adjunct processor queue. * @qid: The AP queue number + * @fbit: if != 0 set F bit * * Returns AP queue status structure. */ -static inline struct ap_queue_status ap_rapq(ap_qid_t qid) +static inline struct ap_queue_status ap_rapq(ap_qid_t qid, int fbit) { unsigned long reg0 = qid | (1UL << 24); /* fc 1UL is RAPQ */ union ap_queue_status_reg reg1; + if (fbit) + reg0 |= 1UL << 22; + asm volatile( " lgr 0,%[reg0]\n" /* qid arg into gr0 */ " .insn rre,0xb2af0000,0,0\n" /* PQAP(RAPQ) */ @@ -153,14 +193,18 @@ static inline struct ap_queue_status ap_rapq(ap_qid_t qid) /** * ap_pqap_zapq(): Reset and zeroize adjunct processor queue. * @qid: The AP queue number + * @fbit: if != 0 set F bit * * Returns AP queue status structure. */ -static inline struct ap_queue_status ap_zapq(ap_qid_t qid) +static inline struct ap_queue_status ap_zapq(ap_qid_t qid, int fbit) { unsigned long reg0 = qid | (2UL << 24); /* fc 2UL is ZAPQ */ union ap_queue_status_reg reg1; + if (fbit) + reg0 |= 1UL << 22; + asm volatile( " lgr 0,%[reg0]\n" /* qid arg into gr0 */ " .insn rre,0xb2af0000,0,0\n" /* PQAP(ZAPQ) */ @@ -180,15 +224,16 @@ struct ap_config_info { unsigned int apxa : 1; /* N bit */ unsigned int qact : 1; /* C bit */ unsigned int rc8a : 1; /* R bit */ - unsigned char _reserved1 : 4; - unsigned char _reserved2[3]; - unsigned char Na; /* max # of APs - 1 */ - unsigned char Nd; /* max # of Domains - 1 */ - unsigned char _reserved3[10]; + unsigned int : 4; + unsigned int apsb : 1; /* B bit */ + unsigned int : 23; + unsigned char na; /* max # of APs - 1 */ + unsigned char nd; /* max # of Domains - 1 */ + unsigned char _reserved0[10]; unsigned int apm[8]; /* AP ID mask */ unsigned int aqm[8]; /* AP (usage) queue mask */ unsigned int adm[8]; /* AP (control) domain mask */ - unsigned char _reserved4[16]; + unsigned char _reserved1[16]; } __aligned(8); /** @@ -318,6 +363,59 @@ static inline struct ap_queue_status ap_qact(ap_qid_t qid, int ifbit, return reg1.status; } +/* + * ap_bapq(): SE bind AP queue. + * @qid: The AP queue number + * + * Returns AP queue status structure. + * + * Invoking this function in a non-SE environment + * may case a specification exception. + */ +static inline struct ap_queue_status ap_bapq(ap_qid_t qid) +{ + unsigned long reg0 = qid | (7UL << 24); /* fc 7 is BAPQ */ + union ap_queue_status_reg reg1; + + asm volatile( + " lgr 0,%[reg0]\n" /* qid arg into gr0 */ + " .insn rre,0xb2af0000,0,0\n" /* PQAP(BAPQ) */ + " lgr %[reg1],1\n" /* gr1 (status) into reg1 */ + : [reg1] "=&d" (reg1.value) + : [reg0] "d" (reg0) + : "cc", "0", "1"); + + return reg1.status; +} + +/* + * ap_aapq(): SE associate AP queue. + * @qid: The AP queue number + * @sec_idx: The secret index + * + * Returns AP queue status structure. + * + * Invoking this function in a non-SE environment + * may case a specification exception. + */ +static inline struct ap_queue_status ap_aapq(ap_qid_t qid, unsigned int sec_idx) +{ + unsigned long reg0 = qid | (8UL << 24); /* fc 8 is AAPQ */ + unsigned long reg2 = sec_idx; + union ap_queue_status_reg reg1; + + asm volatile( + " lgr 0,%[reg0]\n" /* qid arg into gr0 */ + " lgr 2,%[reg2]\n" /* secret index into gr2 */ + " .insn rre,0xb2af0000,0,0\n" /* PQAP(AAPQ) */ + " lgr %[reg1],1\n" /* gr1 (status) into reg1 */ + : [reg1] "=&d" (reg1.value) + : [reg0] "d" (reg0), [reg2] "d" (reg2) + : "cc", "0", "1", "2"); + + return reg1.status; +} + /** * ap_nqap(): Send message to adjunct processor queue. * @qid: The AP queue number @@ -359,10 +457,11 @@ static inline struct ap_queue_status ap_nqap(ap_qid_t qid, * ap_dqap(): Receive message from adjunct processor queue. * @qid: The AP queue number * @psmid: Pointer to program supplied message identifier - * @msg: The message text - * @length: The message length - * @reslength: Resitual length on return - * @resgr0: input: gr0 value (only used if != 0), output: resitual gr0 content + * @msg: Pointer to message buffer + * @msglen: Message buffer size + * @length: Pointer to length of actually written bytes + * @reslength: Residual length on return + * @resgr0: input: gr0 value (only used if != 0), output: residual gr0 content * * Returns AP queue status structure. * Condition code 1 on DQAP means the receive has taken place @@ -386,8 +485,9 @@ static inline struct ap_queue_status ap_nqap(ap_qid_t qid, * *resgr0 is to be used instead of qid to further process this entry. */ static inline struct ap_queue_status ap_dqap(ap_qid_t qid, - unsigned long long *psmid, - void *msg, size_t length, + unsigned long *psmid, + void *msg, size_t msglen, + size_t *length, size_t *reslength, unsigned long *resgr0) { @@ -399,7 +499,7 @@ static inline struct ap_queue_status ap_dqap(ap_qid_t qid, rp1.even = 0UL; rp1.odd = 0UL; rp2.even = (unsigned long)msg; - rp2.odd = (unsigned long)length; + rp2.odd = (unsigned long)msglen; asm volatile( " lgr 0,%[reg0]\n" /* qid param into gr0 */ @@ -429,11 +529,15 @@ static inline struct ap_queue_status ap_dqap(ap_qid_t qid, if (resgr0) *resgr0 = reg0; } else { - *psmid = (((unsigned long long)rp1.even) << 32) + rp1.odd; + *psmid = (rp1.even << 32) + rp1.odd; if (resgr0) *resgr0 = 0; } + /* update *length with the nr of bytes stored into the msg buffer */ + if (length) + *length = msglen - rp2.odd; + return reg1.status; } diff --git a/arch/s390/include/asm/checksum.h b/arch/s390/include/asm/checksum.h index d977a3a2f619..69837eec2ff5 100644 --- a/arch/s390/include/asm/checksum.h +++ b/arch/s390/include/asm/checksum.h @@ -12,13 +12,7 @@ #ifndef _S390_CHECKSUM_H #define _S390_CHECKSUM_H -#ifdef CONFIG_GENERIC_CSUM - -#include <asm-generic/checksum.h> - -#else /* CONFIG_GENERIC_CSUM */ - -#include <linux/uaccess.h> +#include <linux/kasan-checks.h> #include <linux/in6.h> /* @@ -40,6 +34,7 @@ static inline __wsum csum_partial(const void *buff, int len, __wsum sum) .odd = (unsigned long) len, }; + kasan_check_read(buff, len); asm volatile( "0: cksm %[sum],%[rp]\n" " jo 0b\n" @@ -135,5 +130,4 @@ static inline __sum16 csum_ipv6_magic(const struct in6_addr *saddr, return csum_fold((__force __wsum)(sum >> 32)); } -#endif /* CONFIG_GENERIC_CSUM */ #endif /* _S390_CHECKSUM_H */ diff --git a/arch/s390/include/asm/diag.h b/arch/s390/include/asm/diag.h index 674a939f16ee..902e0330dd91 100644 --- a/arch/s390/include/asm/diag.h +++ b/arch/s390/include/asm/diag.h @@ -90,7 +90,7 @@ struct diag8c { u8 num_partitions; u16 width; u16 height; - u8 data[0]; + u8 data[]; } __packed __aligned(4); extern int diag8c(struct diag8c *out, struct ccw_dev_id *devno); diff --git a/arch/s390/include/asm/entry-common.h b/arch/s390/include/asm/entry-common.h index 000de2b1e67a..fdd319a622b0 100644 --- a/arch/s390/include/asm/entry-common.h +++ b/arch/s390/include/asm/entry-common.h @@ -60,9 +60,4 @@ static inline void arch_exit_to_user_mode_prepare(struct pt_regs *regs, #define arch_exit_to_user_mode_prepare arch_exit_to_user_mode_prepare -static inline bool on_thread_stack(void) -{ - return !(((unsigned long)(current->stack) ^ current_stack_pointer) & ~(THREAD_SIZE - 1)); -} - #endif diff --git a/arch/s390/include/asm/fcx.h b/arch/s390/include/asm/fcx.h index b8a028a36173..29784b4b44f6 100644 --- a/arch/s390/include/asm/fcx.h +++ b/arch/s390/include/asm/fcx.h @@ -286,7 +286,7 @@ struct tccb_tcat { */ struct tccb { struct tccb_tcah tcah; - u8 tca[0]; + u8 tca[]; } __attribute__ ((packed, aligned(8))); struct tcw *tcw_get_intrg(struct tcw *tcw); diff --git a/arch/s390/include/asm/kasan.h b/arch/s390/include/asm/kasan.h index e5cfc81d5b61..0cffead0f2f2 100644 --- a/arch/s390/include/asm/kasan.h +++ b/arch/s390/include/asm/kasan.h @@ -2,7 +2,7 @@ #ifndef __ASM_KASAN_H #define __ASM_KASAN_H -#include <asm/pgtable.h> +#include <linux/const.h> #ifdef CONFIG_KASAN @@ -13,35 +13,6 @@ #define KASAN_SHADOW_START KASAN_SHADOW_OFFSET #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE) -extern void kasan_early_init(void); - -/* - * Estimate kasan memory requirements, which it will reserve - * at the very end of available physical memory. To estimate - * that, we take into account that kasan would require - * 1/8 of available physical memory (for shadow memory) + - * creating page tables for the shadow memory region. - * To keep page tables estimates simple take the double of - * combined ptes size. - * - * physmem parameter has to be already adjusted if not entire physical memory - * would be used (e.g. due to effect of "mem=" option). - */ -static inline unsigned long kasan_estimate_memory_needs(unsigned long physmem) -{ - unsigned long kasan_needs; - unsigned long pages; - /* for shadow memory */ - kasan_needs = round_up(physmem / 8, PAGE_SIZE); - /* for paging structures */ - pages = DIV_ROUND_UP(kasan_needs, PAGE_SIZE); - kasan_needs += DIV_ROUND_UP(pages, _PAGE_ENTRIES) * _PAGE_TABLE_SIZE * 2; - - return kasan_needs; -} -#else -static inline void kasan_early_init(void) { } -static inline unsigned long kasan_estimate_memory_needs(unsigned long physmem) { return 0; } #endif #endif diff --git a/arch/s390/include/asm/linkage.h b/arch/s390/include/asm/linkage.h index c76777b15fec..df3fb7d8227b 100644 --- a/arch/s390/include/asm/linkage.h +++ b/arch/s390/include/asm/linkage.h @@ -4,7 +4,7 @@ #include <linux/stringify.h> -#define __ALIGN .align 16, 0x07 +#define __ALIGN .balign CONFIG_FUNCTION_ALIGNMENT, 0x07 #define __ALIGN_STR __stringify(__ALIGN) #endif diff --git a/arch/s390/include/asm/mem_detect.h b/arch/s390/include/asm/mem_detect.h deleted file mode 100644 index f9e7354036d2..000000000000 --- a/arch/s390/include/asm/mem_detect.h +++ /dev/null @@ -1,117 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 */ -#ifndef _ASM_S390_MEM_DETECT_H -#define _ASM_S390_MEM_DETECT_H - -#include <linux/types.h> - -enum mem_info_source { - MEM_DETECT_NONE = 0, - MEM_DETECT_SCLP_STOR_INFO, - MEM_DETECT_DIAG260, - MEM_DETECT_SCLP_READ_INFO, - MEM_DETECT_BIN_SEARCH -}; - -struct mem_detect_block { - u64 start; - u64 end; -}; - -/* - * Storage element id is defined as 1 byte (up to 256 storage elements). - * In practise only storage element id 0 and 1 are used). - * According to architecture one storage element could have as much as - * 1020 subincrements. 255 mem_detect_blocks are embedded in mem_detect_info. - * If more mem_detect_blocks are required, a block of memory from already - * known mem_detect_block is taken (entries_extended points to it). - */ -#define MEM_INLINED_ENTRIES 255 /* (PAGE_SIZE - 16) / 16 */ - -struct mem_detect_info { - u32 count; - u8 info_source; - unsigned long usable; - struct mem_detect_block entries[MEM_INLINED_ENTRIES]; - struct mem_detect_block *entries_extended; -}; -extern struct mem_detect_info mem_detect; - -void add_mem_detect_block(u64 start, u64 end); - -static inline int __get_mem_detect_block(u32 n, unsigned long *start, - unsigned long *end, bool respect_usable_limit) -{ - if (n >= mem_detect.count) { - *start = 0; - *end = 0; - return -1; - } - - if (n < MEM_INLINED_ENTRIES) { - *start = (unsigned long)mem_detect.entries[n].start; - *end = (unsigned long)mem_detect.entries[n].end; - } else { - *start = (unsigned long)mem_detect.entries_extended[n - MEM_INLINED_ENTRIES].start; - *end = (unsigned long)mem_detect.entries_extended[n - MEM_INLINED_ENTRIES].end; - } - - if (respect_usable_limit && mem_detect.usable) { - if (*start >= mem_detect.usable) - return -1; - if (*end > mem_detect.usable) - *end = mem_detect.usable; - } - return 0; -} - -/** - * for_each_mem_detect_usable_block - early online memory range iterator - * @i: an integer used as loop variable - * @p_start: ptr to unsigned long for start address of the range - * @p_end: ptr to unsigned long for end address of the range - * - * Walks over detected online memory ranges below usable limit. - */ -#define for_each_mem_detect_usable_block(i, p_start, p_end) \ - for (i = 0; !__get_mem_detect_block(i, p_start, p_end, true); i++) - -/* Walks over all detected online memory ranges disregarding usable limit. */ -#define for_each_mem_detect_block(i, p_start, p_end) \ - for (i = 0; !__get_mem_detect_block(i, p_start, p_end, false); i++) - -static inline unsigned long get_mem_detect_usable_total(void) -{ - unsigned long start, end, total = 0; - int i; - - for_each_mem_detect_usable_block(i, &start, &end) - total += end - start; - - return total; -} - -static inline void get_mem_detect_reserved(unsigned long *start, - unsigned long *size) -{ - *start = (unsigned long)mem_detect.entries_extended; - if (mem_detect.count > MEM_INLINED_ENTRIES) - *size = (mem_detect.count - MEM_INLINED_ENTRIES) * sizeof(struct mem_detect_block); - else - *size = 0; -} - -static inline unsigned long get_mem_detect_end(void) -{ - unsigned long start; - unsigned long end; - - if (mem_detect.usable) - return mem_detect.usable; - if (mem_detect.count) { - __get_mem_detect_block(mem_detect.count - 1, &start, &end, false); - return end; - } - return 0; -} - -#endif diff --git a/arch/s390/include/asm/nospec-insn.h b/arch/s390/include/asm/nospec-insn.h index 7e9e99523e95..7a946c42ad13 100644 --- a/arch/s390/include/asm/nospec-insn.h +++ b/arch/s390/include/asm/nospec-insn.h @@ -2,6 +2,7 @@ #ifndef _ASM_S390_NOSPEC_ASM_H #define _ASM_S390_NOSPEC_ASM_H +#include <linux/linkage.h> #include <asm/dwarf.h> #ifdef __ASSEMBLY__ @@ -16,7 +17,7 @@ .macro __THUNK_PROLOG_NAME name #ifdef CONFIG_EXPOLINE_EXTERN .pushsection .text,"ax",@progbits - .align 16,0x07 + __ALIGN #else .pushsection .text.\name,"axG",@progbits,\name,comdat #endif diff --git a/arch/s390/include/asm/perf_event.h b/arch/s390/include/asm/perf_event.h index b9da71632827..9917e2717b2b 100644 --- a/arch/s390/include/asm/perf_event.h +++ b/arch/s390/include/asm/perf_event.h @@ -60,7 +60,6 @@ struct perf_sf_sde_regs { #define PERF_CPUM_SF_DIAG_MODE 0x0002 /* Diagnostic-sampling flag */ #define PERF_CPUM_SF_MODE_MASK (PERF_CPUM_SF_BASIC_MODE| \ PERF_CPUM_SF_DIAG_MODE) -#define PERF_CPUM_SF_FULL_BLOCKS 0x0004 /* Process full SDBs only */ #define PERF_CPUM_SF_FREQ_MODE 0x0008 /* Sampling with frequency */ #define REG_NONE 0 @@ -71,7 +70,6 @@ struct perf_sf_sde_regs { #define SAMPL_RATE(hwc) ((hwc)->event_base) #define SAMPL_FLAGS(hwc) ((hwc)->config_base) #define SAMPL_DIAG_MODE(hwc) (SAMPL_FLAGS(hwc) & PERF_CPUM_SF_DIAG_MODE) -#define SDB_FULL_BLOCKS(hwc) (SAMPL_FLAGS(hwc) & PERF_CPUM_SF_FULL_BLOCKS) #define SAMPLE_FREQ_MODE(hwc) (SAMPL_FLAGS(hwc) & PERF_CPUM_SF_FREQ_MODE) #define perf_arch_fetch_caller_regs(regs, __ip) do { \ diff --git a/arch/s390/include/asm/pgtable.h b/arch/s390/include/asm/pgtable.h index c1f6b46ec555..6822a11c2c8a 100644 --- a/arch/s390/include/asm/pgtable.h +++ b/arch/s390/include/asm/pgtable.h @@ -34,7 +34,7 @@ enum { PG_DIRECT_MAP_MAX }; -extern atomic_long_t direct_pages_count[PG_DIRECT_MAP_MAX]; +extern atomic_long_t __bootdata_preserved(direct_pages_count[PG_DIRECT_MAP_MAX]); static inline void update_page_count(int level, long count) { diff --git a/arch/s390/include/asm/physmem_info.h b/arch/s390/include/asm/physmem_info.h new file mode 100644 index 000000000000..8e9c582592b3 --- /dev/null +++ b/arch/s390/include/asm/physmem_info.h @@ -0,0 +1,171 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _ASM_S390_MEM_DETECT_H +#define _ASM_S390_MEM_DETECT_H + +#include <linux/types.h> + +enum physmem_info_source { + MEM_DETECT_NONE = 0, + MEM_DETECT_SCLP_STOR_INFO, + MEM_DETECT_DIAG260, + MEM_DETECT_SCLP_READ_INFO, + MEM_DETECT_BIN_SEARCH +}; + +struct physmem_range { + u64 start; + u64 end; +}; + +enum reserved_range_type { + RR_DECOMPRESSOR, + RR_INITRD, + RR_VMLINUX, + RR_AMODE31, + RR_IPLREPORT, + RR_CERT_COMP_LIST, + RR_MEM_DETECT_EXTENDED, + RR_VMEM, + RR_MAX +}; + +struct reserved_range { + unsigned long start; + unsigned long end; + struct reserved_range *chain; +}; + +/* + * Storage element id is defined as 1 byte (up to 256 storage elements). + * In practise only storage element id 0 and 1 are used). + * According to architecture one storage element could have as much as + * 1020 subincrements. 255 physmem_ranges are embedded in physmem_info. + * If more physmem_ranges are required, a block of memory from already + * known physmem_range is taken (online_extended points to it). + */ +#define MEM_INLINED_ENTRIES 255 /* (PAGE_SIZE - 16) / 16 */ + +struct physmem_info { + u32 range_count; + u8 info_source; + unsigned long usable; + struct reserved_range reserved[RR_MAX]; + struct physmem_range online[MEM_INLINED_ENTRIES]; + struct physmem_range *online_extended; +}; + +extern struct physmem_info physmem_info; + +void add_physmem_online_range(u64 start, u64 end); + +static inline int __get_physmem_range(u32 n, unsigned long *start, + unsigned long *end, bool respect_usable_limit) +{ + if (n >= physmem_info.range_count) { + *start = 0; + *end = 0; + return -1; + } + + if (n < MEM_INLINED_ENTRIES) { + *start = (unsigned long)physmem_info.online[n].start; + *end = (unsigned long)physmem_info.online[n].end; + } else { + *start = (unsigned long)physmem_info.online_extended[n - MEM_INLINED_ENTRIES].start; + *end = (unsigned long)physmem_info.online_extended[n - MEM_INLINED_ENTRIES].end; + } + + if (respect_usable_limit && physmem_info.usable) { + if (*start >= physmem_info.usable) + return -1; + if (*end > physmem_info.usable) + *end = physmem_info.usable; + } + return 0; +} + +/** + * for_each_physmem_usable_range - early online memory range iterator + * @i: an integer used as loop variable + * @p_start: ptr to unsigned long for start address of the range + * @p_end: ptr to unsigned long for end address of the range + * + * Walks over detected online memory ranges below usable limit. + */ +#define for_each_physmem_usable_range(i, p_start, p_end) \ + for (i = 0; !__get_physmem_range(i, p_start, p_end, true); i++) + +/* Walks over all detected online memory ranges disregarding usable limit. */ +#define for_each_physmem_online_range(i, p_start, p_end) \ + for (i = 0; !__get_physmem_range(i, p_start, p_end, false); i++) + +static inline const char *get_physmem_info_source(void) +{ + switch (physmem_info.info_source) { + case MEM_DETECT_SCLP_STOR_INFO: + return "sclp storage info"; + case MEM_DETECT_DIAG260: + return "diag260"; + case MEM_DETECT_SCLP_READ_INFO: + return "sclp read info"; + case MEM_DETECT_BIN_SEARCH: + return "binary search"; + } + return "none"; +} + +#define RR_TYPE_NAME(t) case RR_ ## t: return #t +static inline const char *get_rr_type_name(enum reserved_range_type t) +{ + switch (t) { + RR_TYPE_NAME(DECOMPRESSOR); + RR_TYPE_NAME(INITRD); + RR_TYPE_NAME(VMLINUX); + RR_TYPE_NAME(AMODE31); + RR_TYPE_NAME(IPLREPORT); + RR_TYPE_NAME(CERT_COMP_LIST); + RR_TYPE_NAME(MEM_DETECT_EXTENDED); + RR_TYPE_NAME(VMEM); + default: + return "UNKNOWN"; + } +} + +#define for_each_physmem_reserved_type_range(t, range, p_start, p_end) \ + for (range = &physmem_info.reserved[t], *p_start = range->start, *p_end = range->end; \ + range && range->end; range = range->chain, \ + *p_start = range ? range->start : 0, *p_end = range ? range->end : 0) + +static inline struct reserved_range *__physmem_reserved_next(enum reserved_range_type *t, + struct reserved_range *range) +{ + if (!range) { + range = &physmem_info.reserved[*t]; + if (range->end) + return range; + } + if (range->chain) + return range->chain; + while (++*t < RR_MAX) { + range = &physmem_info.reserved[*t]; + if (range->end) + return range; + } + return NULL; +} + +#define for_each_physmem_reserved_range(t, range, p_start, p_end) \ + for (t = 0, range = __physmem_reserved_next(&t, NULL), \ + *p_start = range ? range->start : 0, *p_end = range ? range->end : 0; \ + range; range = __physmem_reserved_next(&t, range), \ + *p_start = range ? range->start : 0, *p_end = range ? range->end : 0) + +static inline unsigned long get_physmem_reserved(enum reserved_range_type type, + unsigned long *addr, unsigned long *size) +{ + *addr = physmem_info.reserved[type].start; + *size = physmem_info.reserved[type].end - physmem_info.reserved[type].start; + return *size; +} + +#endif diff --git a/arch/s390/include/asm/processor.h b/arch/s390/include/asm/processor.h index e98d9650764b..dc17896a001a 100644 --- a/arch/s390/include/asm/processor.h +++ b/arch/s390/include/asm/processor.h @@ -99,7 +99,6 @@ void cpu_detect_mhz_feature(void); extern const struct seq_operations cpuinfo_op; extern void execve_tail(void); -extern void __bpon(void); unsigned long vdso_size(void); /* @@ -119,6 +118,41 @@ unsigned long vdso_size(void); #define HAVE_ARCH_PICK_MMAP_LAYOUT +#define __stackleak_poison __stackleak_poison +static __always_inline void __stackleak_poison(unsigned long erase_low, + unsigned long erase_high, + unsigned long poison) +{ + unsigned long tmp, count; + + count = erase_high - erase_low; + if (!count) + return; + asm volatile( + " cghi %[count],8\n" + " je 2f\n" + " aghi %[count],-(8+1)\n" + " srlg %[tmp],%[count],8\n" + " ltgr %[tmp],%[tmp]\n" + " jz 1f\n" + "0: stg %[poison],0(%[addr])\n" + " mvc 8(256-8,%[addr]),0(%[addr])\n" + " la %[addr],256(%[addr])\n" + " brctg %[tmp],0b\n" + "1: stg %[poison],0(%[addr])\n" + " larl %[tmp],3f\n" + " ex %[count],0(%[tmp])\n" + " j 4f\n" + "2: stg %[poison],0(%[addr])\n" + " j 4f\n" + "3: mvc 8(1,%[addr]),0(%[addr])\n" + "4:\n" + : [addr] "+&a" (erase_low), [count] "+&d" (count), [tmp] "=&a" (tmp) + : [poison] "d" (poison) + : "memory", "cc" + ); +} + /* * Thread structure */ @@ -227,6 +261,13 @@ static __always_inline unsigned long __current_stack_pointer(void) return sp; } +static __always_inline bool on_thread_stack(void) +{ + unsigned long ksp = S390_lowcore.kernel_stack; + + return !((ksp ^ current_stack_pointer) & ~(THREAD_SIZE - 1)); +} + static __always_inline unsigned short stap(void) { unsigned short cpu_address; @@ -329,9 +370,6 @@ static __always_inline void __noreturn disabled_wait(void) #define ARCH_LOW_ADDRESS_LIMIT 0x7fffffffUL -extern int s390_isolate_bp(void); -extern int s390_isolate_bp_guest(void); - static __always_inline bool regs_irqs_disabled(struct pt_regs *regs) { return arch_irqs_disabled_flags(regs->psw.mask); diff --git a/arch/s390/include/asm/set_memory.h b/arch/s390/include/asm/set_memory.h index 950d87bd997a..7a3eefd7a242 100644 --- a/arch/s390/include/asm/set_memory.h +++ b/arch/s390/include/asm/set_memory.h @@ -6,11 +6,23 @@ extern struct mutex cpa_mutex; -#define SET_MEMORY_RO 1UL -#define SET_MEMORY_RW 2UL -#define SET_MEMORY_NX 4UL -#define SET_MEMORY_X 8UL -#define SET_MEMORY_4K 16UL +enum { + _SET_MEMORY_RO_BIT, + _SET_MEMORY_RW_BIT, + _SET_MEMORY_NX_BIT, + _SET_MEMORY_X_BIT, + _SET_MEMORY_4K_BIT, + _SET_MEMORY_INV_BIT, + _SET_MEMORY_DEF_BIT, +}; + +#define SET_MEMORY_RO BIT(_SET_MEMORY_RO_BIT) +#define SET_MEMORY_RW BIT(_SET_MEMORY_RW_BIT) +#define SET_MEMORY_NX BIT(_SET_MEMORY_NX_BIT) +#define SET_MEMORY_X BIT(_SET_MEMORY_X_BIT) +#define SET_MEMORY_4K BIT(_SET_MEMORY_4K_BIT) +#define SET_MEMORY_INV BIT(_SET_MEMORY_INV_BIT) +#define SET_MEMORY_DEF BIT(_SET_MEMORY_DEF_BIT) int __set_memory(unsigned long addr, int numpages, unsigned long flags); @@ -34,9 +46,23 @@ static inline int set_memory_x(unsigned long addr, int numpages) return __set_memory(addr, numpages, SET_MEMORY_X); } +#define set_memory_rox set_memory_rox +static inline int set_memory_rox(unsigned long addr, int numpages) +{ + return __set_memory(addr, numpages, SET_MEMORY_RO | SET_MEMORY_X); +} + +static inline int set_memory_rwnx(unsigned long addr, int numpages) +{ + return __set_memory(addr, numpages, SET_MEMORY_RW | SET_MEMORY_NX); +} + static inline int set_memory_4k(unsigned long addr, int numpages) { return __set_memory(addr, numpages, SET_MEMORY_4K); } +int set_direct_map_invalid_noflush(struct page *page); +int set_direct_map_default_noflush(struct page *page); + #endif diff --git a/arch/s390/include/asm/setup.h b/arch/s390/include/asm/setup.h index 3a1f8825bc7d..f191255c60db 100644 --- a/arch/s390/include/asm/setup.h +++ b/arch/s390/include/asm/setup.h @@ -74,10 +74,6 @@ extern unsigned int zlib_dfltcc_support; extern int noexec_disabled; extern unsigned long ident_map_size; -extern unsigned long pgalloc_pos; -extern unsigned long pgalloc_end; -extern unsigned long pgalloc_low; -extern unsigned long __amode31_base; /* The Write Back bit position in the physaddr is given by the SLPC PCI */ extern unsigned long mio_wb_bit_mask; @@ -150,13 +146,13 @@ static inline unsigned long kaslr_offset(void) return __kaslr_offset; } -extern int is_full_image; - -struct initrd_data { - unsigned long start; - unsigned long size; -}; -extern struct initrd_data initrd_data; +extern int __kaslr_enabled; +static inline int kaslr_enabled(void) +{ + if (IS_ENABLED(CONFIG_RANDOMIZE_BASE)) + return __kaslr_enabled; + return 0; +} struct oldmem_data { unsigned long start; @@ -164,7 +160,7 @@ struct oldmem_data { }; extern struct oldmem_data oldmem_data; -static inline u32 gen_lpswe(unsigned long addr) +static __always_inline u32 gen_lpswe(unsigned long addr) { BUILD_BUG_ON(addr > 0xfff); return 0xb2b20000 | addr; diff --git a/arch/s390/include/asm/stacktrace.h b/arch/s390/include/asm/stacktrace.h index 1802be5abb5d..78f7b729b65f 100644 --- a/arch/s390/include/asm/stacktrace.h +++ b/arch/s390/include/asm/stacktrace.h @@ -189,17 +189,53 @@ static __always_inline unsigned long get_stack_pointer(struct task_struct *task, (rettype)r2; \ }) -#define call_on_stack_noreturn(fn, stack) \ +/* + * Use call_nodat() to call a function with DAT disabled. + * Proper sign and zero extension of function arguments is done. + * Usage: + * + * rc = call_nodat(nr, rettype, fn, t1, a1, t2, a2, ...) + * + * - nr specifies the number of function arguments of fn. + * - fn is the function to be called, where fn is a physical address. + * - rettype is the return type of fn. + * - t1, a1, ... are pairs, where t1 must match the type of the first + * argument of fn, t2 the second, etc. a1 is the corresponding + * first function argument (not name), etc. + * + * fn() is called with standard C function call ABI, with the exception + * that no useful stackframe or stackpointer is passed via register 15. + * Therefore the called function must not use r15 to access the stack. + */ +#define call_nodat(nr, rettype, fn, ...) \ ({ \ - void (*__fn)(void) = fn; \ + rettype (*__fn)(CALL_PARM_##nr(__VA_ARGS__)) = (fn); \ + /* aligned since psw_leave must not cross page boundary */ \ + psw_t __aligned(16) psw_leave; \ + psw_t psw_enter; \ + CALL_LARGS_##nr(__VA_ARGS__); \ + CALL_REGS_##nr; \ \ + CALL_TYPECHECK_##nr(__VA_ARGS__); \ + psw_enter.mask = PSW_KERNEL_BITS & ~PSW_MASK_DAT; \ + psw_enter.addr = (unsigned long)__fn; \ asm volatile( \ - " la 15,0(%[_stack])\n" \ - " xc %[_bc](8,15),%[_bc](15)\n" \ - " brasl 14,%[_fn]\n" \ - ::[_bc] "i" (offsetof(struct stack_frame, back_chain)), \ - [_stack] "a" (stack), [_fn] "X" (__fn)); \ - BUG(); \ + " epsw 0,1\n" \ + " risbg 1,0,0,31,32\n" \ + " larl 7,1f\n" \ + " stg 1,%[psw_leave]\n" \ + " stg 7,8+%[psw_leave]\n" \ + " la 7,%[psw_leave]\n" \ + " lra 7,0(7)\n" \ + " larl 1,0f\n" \ + " lra 14,0(1)\n" \ + " lpswe %[psw_enter]\n" \ + "0: lpswe 0(7)\n" \ + "1:\n" \ + : CALL_FMT_##nr, [psw_leave] "=Q" (psw_leave) \ + : [psw_enter] "Q" (psw_enter) \ + : "7", CALL_CLOBBER_##nr); \ + (rettype)r2; \ }) #endif /* _ASM_S390_STACKTRACE_H */ diff --git a/arch/s390/include/asm/string.h b/arch/s390/include/asm/string.h index 3fae93ddb322..351685de53d2 100644 --- a/arch/s390/include/asm/string.h +++ b/arch/s390/include/asm/string.h @@ -55,18 +55,6 @@ char *strstr(const char *s1, const char *s2); #if defined(CONFIG_KASAN) && !defined(__SANITIZE_ADDRESS__) -extern void *__memcpy(void *dest, const void *src, size_t n); -extern void *__memset(void *s, int c, size_t n); -extern void *__memmove(void *dest, const void *src, size_t n); - -/* - * For files that are not instrumented (e.g. mm/slub.c) we - * should use not instrumented version of mem* functions. - */ - -#define memcpy(dst, src, len) __memcpy(dst, src, len) -#define memmove(dst, src, len) __memmove(dst, src, len) -#define memset(s, c, n) __memset(s, c, n) #define strlen(s) __strlen(s) #define __no_sanitize_prefix_strfunc(x) __##x @@ -79,6 +67,9 @@ extern void *__memmove(void *dest, const void *src, size_t n); #define __no_sanitize_prefix_strfunc(x) x #endif /* defined(CONFIG_KASAN) && !defined(__SANITIZE_ADDRESS__) */ +void *__memcpy(void *dest, const void *src, size_t n); +void *__memset(void *s, int c, size_t n); +void *__memmove(void *dest, const void *src, size_t n); void *__memset16(uint16_t *s, uint16_t v, size_t count); void *__memset32(uint32_t *s, uint32_t v, size_t count); void *__memset64(uint64_t *s, uint64_t v, size_t count); diff --git a/arch/s390/include/asm/thread_info.h b/arch/s390/include/asm/thread_info.h index b2ffcb4fe000..c7c97921ed8d 100644 --- a/arch/s390/include/asm/thread_info.h +++ b/arch/s390/include/asm/thread_info.h @@ -9,6 +9,9 @@ #define _ASM_THREAD_INFO_H #include <linux/bits.h> +#ifndef ASM_OFFSETS_C +#include <asm/asm-offsets.h> +#endif /* * General size of kernel stacks @@ -21,13 +24,12 @@ #define BOOT_STACK_SIZE (PAGE_SIZE << 2) #define THREAD_SIZE (PAGE_SIZE << THREAD_SIZE_ORDER) +#define STACK_INIT_OFFSET (THREAD_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE) + #ifndef __ASSEMBLY__ #include <asm/lowcore.h> #include <asm/page.h> -#define STACK_INIT_OFFSET \ - (THREAD_SIZE - STACK_FRAME_OVERHEAD - sizeof(struct pt_regs)) - /* * low level task data that entry.S needs immediate access to * - this struct should fit entirely inside of one cache line @@ -70,7 +72,6 @@ void arch_setup_new_exec(void); #define TIF_PATCH_PENDING 5 /* pending live patching update */ #define TIF_PGSTE 6 /* New mm's will use 4K page tables */ #define TIF_NOTIFY_SIGNAL 7 /* signal notifications exist */ -#define TIF_ISOLATE_BP 8 /* Run process with isolated BP */ #define TIF_ISOLATE_BP_GUEST 9 /* Run KVM guests with isolated BP */ #define TIF_PER_TRAP 10 /* Need to handle PER trap on exit to usermode */ @@ -94,7 +95,6 @@ void arch_setup_new_exec(void); #define _TIF_UPROBE BIT(TIF_UPROBE) #define _TIF_GUARDED_STORAGE BIT(TIF_GUARDED_STORAGE) #define _TIF_PATCH_PENDING BIT(TIF_PATCH_PENDING) -#define _TIF_ISOLATE_BP BIT(TIF_ISOLATE_BP) #define _TIF_ISOLATE_BP_GUEST BIT(TIF_ISOLATE_BP_GUEST) #define _TIF_PER_TRAP BIT(TIF_PER_TRAP) diff --git a/arch/s390/kernel/debug.c b/arch/s390/kernel/debug.c index b376f0377a2c..a85e0c3e7027 100644 --- a/arch/s390/kernel/debug.c +++ b/arch/s390/kernel/debug.c @@ -60,7 +60,7 @@ typedef struct { * except of floats, and long long (32 bit) * */ - long args[0]; + long args[]; } debug_sprintf_entry_t; /* internal function prototyes */ @@ -981,16 +981,6 @@ static struct ctl_table s390dbf_table[] = { { } }; -static struct ctl_table s390dbf_dir_table[] = { - { - .procname = "s390dbf", - .maxlen = 0, - .mode = S_IRUGO | S_IXUGO, - .child = s390dbf_table, - }, - { } -}; - static struct ctl_table_header *s390dbf_sysctl_header; /** @@ -1574,7 +1564,7 @@ out: */ static int __init debug_init(void) { - s390dbf_sysctl_header = register_sysctl_table(s390dbf_dir_table); + s390dbf_sysctl_header = register_sysctl("s390dbf", s390dbf_table); mutex_lock(&debug_mutex); debug_debugfs_root_entry = debugfs_create_dir(DEBUG_DIR_ROOT, NULL); initialized = 1; diff --git a/arch/s390/kernel/dumpstack.c b/arch/s390/kernel/dumpstack.c index 1e3233eb510a..d2012635b093 100644 --- a/arch/s390/kernel/dumpstack.c +++ b/arch/s390/kernel/dumpstack.c @@ -41,60 +41,50 @@ const char *stack_type_name(enum stack_type type) EXPORT_SYMBOL_GPL(stack_type_name); static inline bool in_stack(unsigned long sp, struct stack_info *info, - enum stack_type type, unsigned long low, - unsigned long high) + enum stack_type type, unsigned long stack) { - if (sp < low || sp >= high) + if (sp < stack || sp >= stack + THREAD_SIZE) return false; info->type = type; - info->begin = low; - info->end = high; + info->begin = stack; + info->end = stack + THREAD_SIZE; return true; } static bool in_task_stack(unsigned long sp, struct task_struct *task, struct stack_info *info) { - unsigned long stack; + unsigned long stack = (unsigned long)task_stack_page(task); - stack = (unsigned long) task_stack_page(task); - return in_stack(sp, info, STACK_TYPE_TASK, stack, stack + THREAD_SIZE); + return in_stack(sp, info, STACK_TYPE_TASK, stack); } static bool in_irq_stack(unsigned long sp, struct stack_info *info) { - unsigned long frame_size, top; + unsigned long stack = S390_lowcore.async_stack - STACK_INIT_OFFSET; - frame_size = STACK_FRAME_OVERHEAD + sizeof(struct pt_regs); - top = S390_lowcore.async_stack + frame_size; - return in_stack(sp, info, STACK_TYPE_IRQ, top - THREAD_SIZE, top); + return in_stack(sp, info, STACK_TYPE_IRQ, stack); } static bool in_nodat_stack(unsigned long sp, struct stack_info *info) { - unsigned long frame_size, top; + unsigned long stack = S390_lowcore.nodat_stack - STACK_INIT_OFFSET; - frame_size = STACK_FRAME_OVERHEAD + sizeof(struct pt_regs); - top = S390_lowcore.nodat_stack + frame_size; - return in_stack(sp, info, STACK_TYPE_NODAT, top - THREAD_SIZE, top); + return in_stack(sp, info, STACK_TYPE_NODAT, stack); } static bool in_mcck_stack(unsigned long sp, struct stack_info *info) { - unsigned long frame_size, top; + unsigned long stack = S390_lowcore.mcck_stack - STACK_INIT_OFFSET; - frame_size = STACK_FRAME_OVERHEAD + sizeof(struct pt_regs); - top = S390_lowcore.mcck_stack + frame_size; - return in_stack(sp, info, STACK_TYPE_MCCK, top - THREAD_SIZE, top); + return in_stack(sp, info, STACK_TYPE_MCCK, stack); } static bool in_restart_stack(unsigned long sp, struct stack_info *info) { - unsigned long frame_size, top; + unsigned long stack = S390_lowcore.restart_stack - STACK_INIT_OFFSET; - frame_size = STACK_FRAME_OVERHEAD + sizeof(struct pt_regs); - top = S390_lowcore.restart_stack + frame_size; - return in_stack(sp, info, STACK_TYPE_RESTART, top - THREAD_SIZE, top); + return in_stack(sp, info, STACK_TYPE_RESTART, stack); } int get_stack_info(unsigned long sp, struct task_struct *task, @@ -152,7 +142,13 @@ void show_stack(struct task_struct *task, unsigned long *stack, static void show_last_breaking_event(struct pt_regs *regs) { printk("Last Breaking-Event-Address:\n"); - printk(" [<%016lx>] %pSR\n", regs->last_break, (void *)regs->last_break); + printk(" [<%016lx>] ", regs->last_break); + if (user_mode(regs)) { + print_vma_addr(KERN_CONT, regs->last_break); + pr_cont("\n"); + } else { + pr_cont("%pSR\n", (void *)regs->last_break); + } } void show_registers(struct pt_regs *regs) diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c index d26f02495636..2dd5976a55ac 100644 --- a/arch/s390/kernel/early.c +++ b/arch/s390/kernel/early.c @@ -34,8 +34,6 @@ #include <asm/switch_to.h> #include "entry.h" -int __bootdata(is_full_image); - #define decompressor_handled_param(param) \ static int __init ignore_decompressor_param_##param(char *s) \ { \ @@ -53,6 +51,14 @@ decompressor_handled_param(nokaslr); decompressor_handled_param(prot_virt); #endif +static void __init kasan_early_init(void) +{ +#ifdef CONFIG_KASAN + init_task.kasan_depth = 0; + sclp_early_printk("KernelAddressSanitizer initialized\n"); +#endif +} + static void __init reset_tod_clock(void) { union tod_clock clk; @@ -288,17 +294,6 @@ static void __init setup_boot_command_line(void) strscpy(boot_command_line, early_command_line, COMMAND_LINE_SIZE); } -static void __init check_image_bootable(void) -{ - if (is_full_image) - return; - - sclp_early_printk("Linux kernel boot failure: An attempt to boot a vmlinux ELF image failed.\n"); - sclp_early_printk("This image does not contain all parts necessary for starting up. Use\n"); - sclp_early_printk("bzImage or arch/s390/boot/compressed/vmlinux instead.\n"); - disabled_wait(); -} - static void __init sort_amode31_extable(void) { sort_extable(__start_amode31_ex_table, __stop_amode31_ex_table); @@ -306,8 +301,8 @@ static void __init sort_amode31_extable(void) void __init startup_init(void) { + kasan_early_init(); reset_tod_clock(); - check_image_bootable(); time_early_init(); init_kernel_storage_key(); lockdep_off(); diff --git a/arch/s390/kernel/earlypgm.S b/arch/s390/kernel/earlypgm.S index f521c6da37b8..c634871f0d90 100644 --- a/arch/s390/kernel/earlypgm.S +++ b/arch/s390/kernel/earlypgm.S @@ -7,7 +7,7 @@ #include <linux/linkage.h> #include <asm/asm-offsets.h> -ENTRY(early_pgm_check_handler) +SYM_CODE_START(early_pgm_check_handler) stmg %r8,%r15,__LC_SAVE_AREA_SYNC aghi %r15,-(STACK_FRAME_OVERHEAD+__PT_SIZE) la %r11,STACK_FRAME_OVERHEAD(%r15) @@ -20,4 +20,4 @@ ENTRY(early_pgm_check_handler) mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) lmg %r0,%r15,STACK_FRAME_OVERHEAD+__PT_R0(%r15) lpswe __LC_RETURN_PSW -ENDPROC(early_pgm_check_handler) +SYM_CODE_END(early_pgm_check_handler) diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S index 76a06f3d3671..e5b6c1369e8e 100644 --- a/arch/s390/kernel/entry.S +++ b/arch/s390/kernel/entry.S @@ -29,10 +29,6 @@ #include <asm/export.h> #include <asm/nospec-insn.h> -STACK_SHIFT = PAGE_SHIFT + THREAD_SIZE_ORDER -STACK_SIZE = 1 << STACK_SHIFT -STACK_INIT = STACK_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE - _LPP_OFFSET = __LC_LPP .macro STBEAR address @@ -53,7 +49,7 @@ _LPP_OFFSET = __LC_LPP .macro CHECK_STACK savearea #ifdef CONFIG_CHECK_STACK - tml %r15,STACK_SIZE - CONFIG_STACK_GUARD + tml %r15,THREAD_SIZE - CONFIG_STACK_GUARD lghi %r14,\savearea jz stack_overflow #endif @@ -62,8 +58,8 @@ _LPP_OFFSET = __LC_LPP .macro CHECK_VMAP_STACK savearea,oklabel #ifdef CONFIG_VMAP_STACK lgr %r14,%r15 - nill %r14,0x10000 - STACK_SIZE - oill %r14,STACK_INIT + nill %r14,0x10000 - THREAD_SIZE + oill %r14,STACK_INIT_OFFSET clg %r14,__LC_KERNEL_STACK je \oklabel clg %r14,__LC_ASYNC_STACK @@ -154,26 +150,26 @@ _LPP_OFFSET = __LC_LPP .endm #endif + .macro STACKLEAK_ERASE +#ifdef CONFIG_GCC_PLUGIN_STACKLEAK + brasl %r14,stackleak_erase_on_task_stack +#endif + .endm + GEN_BR_THUNK %r14 .section .kprobes.text, "ax" .Ldummy: /* - * This nop exists only in order to avoid that __bpon starts at - * the beginning of the kprobes text section. In that case we would - * have several symbols at the same address. E.g. objdump would take - * an arbitrary symbol name when disassembling this code. - * With the added nop in between the __bpon symbol is unique - * again. + * The following nop exists only in order to avoid that the next + * symbol starts at the beginning of the kprobes text section. + * In that case there would be several symbols at the same address. + * E.g. objdump would take an arbitrary symbol when disassembling + * the code. + * With the added nop in between this cannot happen. */ nop 0 -ENTRY(__bpon) - .globl __bpon - BPON - BR_EX %r14 -ENDPROC(__bpon) - /* * Scheduler resume function, called by switch_to * gpr2 = (task_struct *) prev @@ -181,11 +177,11 @@ ENDPROC(__bpon) * Returns: * gpr2 = prev */ -ENTRY(__switch_to) +SYM_FUNC_START(__switch_to) stmg %r6,%r15,__SF_GPRS(%r15) # store gprs of prev task lghi %r4,__TASK_stack lghi %r1,__TASK_thread - llill %r5,STACK_INIT + llill %r5,STACK_INIT_OFFSET stg %r15,__THREAD_ksp(%r1,%r2) # store kernel stack of prev lg %r15,0(%r4,%r3) # start of kernel stack of next agr %r15,%r5 # end of kernel stack of next @@ -197,7 +193,7 @@ ENTRY(__switch_to) lmg %r6,%r15,__SF_GPRS(%r15) # load gprs of next task ALTERNATIVE "nop", "lpp _LPP_OFFSET", 40 BR_EX %r14 -ENDPROC(__switch_to) +SYM_FUNC_END(__switch_to) #if IS_ENABLED(CONFIG_KVM) /* @@ -206,7 +202,7 @@ ENDPROC(__switch_to) * %r3 pointer to sie control block virt * %r4 guest register save area */ -ENTRY(__sie64a) +SYM_FUNC_START(__sie64a) stmg %r6,%r14,__SF_GPRS(%r15) # save kernel registers lg %r12,__LC_CURRENT stg %r2,__SF_SIE_CONTROL_PHYS(%r15) # save sie block physical.. @@ -227,7 +223,7 @@ ENTRY(__sie64a) TSTMSK __LC_CPU_FLAGS,_CIF_FPU jo .Lsie_skip # exit if fp/vx regs changed lg %r14,__SF_SIE_CONTROL_PHYS(%r15) # get sie block phys addr - BPEXIT __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST) + BPEXIT __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST .Lsie_entry: sie 0(%r14) # Let the next instruction be NOP to avoid triggering a machine check @@ -235,7 +231,7 @@ ENTRY(__sie64a) nopr 7 .Lsie_leave: BPOFF - BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST) + BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST .Lsie_skip: lg %r14,__SF_SIE_CONTROL(%r15) # get control block pointer ni __SIE_PROG0C+3(%r14),0xfe # no longer in SIE @@ -252,8 +248,7 @@ ENTRY(__sie64a) nopr 7 .Lrewind_pad2: nopr 7 - .globl sie_exit -sie_exit: +SYM_INNER_LABEL(sie_exit, SYM_L_GLOBAL) lg %r14,__SF_SIE_SAVEAREA(%r15) # load guest register save area stmg %r0,%r13,0(%r14) # save guest gprs 0-13 xgr %r0,%r0 # clear guest registers to @@ -273,7 +268,7 @@ sie_exit: EX_TABLE(.Lrewind_pad4,.Lsie_fault) EX_TABLE(.Lrewind_pad2,.Lsie_fault) EX_TABLE(sie_exit,.Lsie_fault) -ENDPROC(__sie64a) +SYM_FUNC_END(__sie64a) EXPORT_SYMBOL(__sie64a) EXPORT_SYMBOL(sie_exit) #endif @@ -283,7 +278,7 @@ EXPORT_SYMBOL(sie_exit) * are entered with interrupts disabled. */ -ENTRY(system_call) +SYM_CODE_START(system_call) stpt __LC_SYS_ENTER_TIMER stmg %r8,%r15,__LC_SAVE_AREA_SYNC BPOFF @@ -291,11 +286,9 @@ ENTRY(system_call) .Lsysc_per: STBEAR __LC_LAST_BREAK lctlg %c1,%c1,__LC_KERNEL_ASCE - lg %r12,__LC_CURRENT lg %r15,__LC_KERNEL_STACK xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) stmg %r0,%r7,STACK_FRAME_OVERHEAD+__PT_R0(%r15) - BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP # clear user controlled register to prevent speculative use xgr %r0,%r0 xgr %r1,%r1 @@ -312,39 +305,40 @@ ENTRY(system_call) MBEAR %r2 lgr %r3,%r14 brasl %r14,__do_syscall + STACKLEAK_ERASE lctlg %c1,%c1,__LC_USER_ASCE mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) - BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP + BPON LBEAR STACK_FRAME_OVERHEAD+__PT_LAST_BREAK(%r15) lmg %r0,%r15,STACK_FRAME_OVERHEAD+__PT_R0(%r15) stpt __LC_EXIT_TIMER LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE -ENDPROC(system_call) +SYM_CODE_END(system_call) # # a new process exits the kernel with ret_from_fork # -ENTRY(ret_from_fork) +SYM_CODE_START(ret_from_fork) lgr %r3,%r11 brasl %r14,__ret_from_fork + STACKLEAK_ERASE lctlg %c1,%c1,__LC_USER_ASCE mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) - BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP + BPON LBEAR STACK_FRAME_OVERHEAD+__PT_LAST_BREAK(%r15) lmg %r0,%r15,STACK_FRAME_OVERHEAD+__PT_R0(%r15) stpt __LC_EXIT_TIMER LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE -ENDPROC(ret_from_fork) +SYM_CODE_END(ret_from_fork) /* * Program check handler routine */ -ENTRY(pgm_check_handler) +SYM_CODE_START(pgm_check_handler) stpt __LC_SYS_ENTER_TIMER BPOFF stmg %r8,%r15,__LC_SAVE_AREA_SYNC - lg %r12,__LC_CURRENT lghi %r10,0 lmg %r8,%r9,__LC_PGM_OLD_PSW tmhh %r8,0x0001 # coming from user space? @@ -355,6 +349,7 @@ ENTRY(pgm_check_handler) #if IS_ENABLED(CONFIG_KVM) # cleanup critical section for program checks in __sie64a OUTSIDE %r9,.Lsie_gmap,.Lsie_done,1f + BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST SIEEXIT lghi %r10,_PIF_GUEST_FAULT #endif @@ -366,8 +361,7 @@ ENTRY(pgm_check_handler) aghi %r15,-(STACK_FRAME_OVERHEAD + __PT_SIZE) # CHECK_VMAP_STACK branches to stack_overflow or 4f CHECK_VMAP_STACK __LC_SAVE_AREA_SYNC,4f -3: BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP - lg %r15,__LC_KERNEL_STACK +3: lg %r15,__LC_KERNEL_STACK 4: la %r11,STACK_FRAME_OVERHEAD(%r15) stg %r10,__PT_FLAGS(%r11) xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) @@ -388,8 +382,9 @@ ENTRY(pgm_check_handler) brasl %r14,__do_pgm_check tmhh %r8,0x0001 # returning to user space? jno .Lpgm_exit_kernel + STACKLEAK_ERASE lctlg %c1,%c1,__LC_USER_ASCE - BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP + BPON stpt __LC_EXIT_TIMER .Lpgm_exit_kernel: mvc __LC_RETURN_PSW(16),STACK_FRAME_OVERHEAD+__PT_PSW(%r15) @@ -407,32 +402,30 @@ ENTRY(pgm_check_handler) lghi %r14,1 LBEAR __LC_PGM_LAST_BREAK LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE # branch to .Lsysc_per -ENDPROC(pgm_check_handler) +SYM_CODE_END(pgm_check_handler) /* * Interrupt handler macro used for external and IO interrupts. */ .macro INT_HANDLER name,lc_old_psw,handler -ENTRY(\name) +SYM_CODE_START(\name) stckf __LC_INT_CLOCK stpt __LC_SYS_ENTER_TIMER STBEAR __LC_LAST_BREAK BPOFF stmg %r8,%r15,__LC_SAVE_AREA_ASYNC - lg %r12,__LC_CURRENT lmg %r8,%r9,\lc_old_psw tmhh %r8,0x0001 # interrupting from user ? jnz 1f #if IS_ENABLED(CONFIG_KVM) OUTSIDE %r9,.Lsie_gmap,.Lsie_done,0f - BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST) + BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST SIEEXIT #endif 0: CHECK_STACK __LC_SAVE_AREA_ASYNC aghi %r15,-(STACK_FRAME_OVERHEAD + __PT_SIZE) j 2f -1: BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP - lctlg %c1,%c1,__LC_KERNEL_ASCE +1: lctlg %c1,%c1,__LC_KERNEL_ASCE lg %r15,__LC_KERNEL_STACK 2: xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) la %r11,STACK_FRAME_OVERHEAD(%r15) @@ -455,13 +448,14 @@ ENTRY(\name) mvc __LC_RETURN_PSW(16),__PT_PSW(%r11) tmhh %r8,0x0001 # returning to user ? jno 2f + STACKLEAK_ERASE lctlg %c1,%c1,__LC_USER_ASCE - BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP + BPON stpt __LC_EXIT_TIMER 2: LBEAR __PT_LAST_BREAK(%r11) lmg %r0,%r15,__PT_R0(%r11) LPSWEY __LC_RETURN_PSW,__LC_RETURN_LPSWE -ENDPROC(\name) +SYM_CODE_END(\name) .endm INT_HANDLER ext_int_handler,__LC_EXT_OLD_PSW,do_ext_irq @@ -470,7 +464,7 @@ INT_HANDLER io_int_handler,__LC_IO_OLD_PSW,do_io_irq /* * Load idle PSW. */ -ENTRY(psw_idle) +SYM_FUNC_START(psw_idle) stg %r14,(__SF_GPRS+8*8)(%r15) stg %r3,__SF_EMPTY(%r15) larl %r1,psw_idle_exit @@ -486,29 +480,26 @@ ENTRY(psw_idle) stckf __CLOCK_IDLE_ENTER(%r2) stpt __TIMER_IDLE_ENTER(%r2) lpswe __SF_EMPTY(%r15) -.globl psw_idle_exit -psw_idle_exit: +SYM_INNER_LABEL(psw_idle_exit, SYM_L_GLOBAL) BR_EX %r14 -ENDPROC(psw_idle) +SYM_FUNC_END(psw_idle) /* * Machine check handler routines */ -ENTRY(mcck_int_handler) +SYM_CODE_START(mcck_int_handler) stckf __LC_MCCK_CLOCK BPOFF la %r1,4095 # validate r1 spt __LC_CPU_TIMER_SAVE_AREA-4095(%r1) # validate cpu timer LBEAR __LC_LAST_BREAK_SAVE_AREA-4095(%r1) # validate bear - lmg %r0,%r15,__LC_GPREGS_SAVE_AREA-4095(%r1)# validate gprs - lg %r12,__LC_CURRENT + lmg %r0,%r15,__LC_GPREGS_SAVE_AREA # validate gprs lmg %r8,%r9,__LC_MCK_OLD_PSW TSTMSK __LC_MCCK_CODE,MCCK_CODE_SYSTEM_DAMAGE jo .Lmcck_panic # yes -> rest of mcck code invalid TSTMSK __LC_MCCK_CODE,MCCK_CODE_CR_VALID jno .Lmcck_panic # control registers invalid -> panic - la %r14,4095 - lctlg %c0,%c15,__LC_CREGS_SAVE_AREA-4095(%r14) # validate ctl regs + lctlg %c0,%c15,__LC_CREGS_SAVE_AREA # validate ctl regs ptlb lghi %r14,__LC_CPU_TIMER_SAVE_AREA mvc __LC_MCCK_ENTER_TIMER(8),0(%r14) @@ -530,16 +521,13 @@ ENTRY(mcck_int_handler) TSTMSK __LC_MCCK_CODE,MCCK_CODE_PSW_IA_VALID jno .Lmcck_panic #if IS_ENABLED(CONFIG_KVM) - OUTSIDE %r9,.Lsie_gmap,.Lsie_done,.Lmcck_stack + OUTSIDE %r9,.Lsie_gmap,.Lsie_done,.Lmcck_user OUTSIDE %r9,.Lsie_entry,.Lsie_leave,4f oi __LC_CPU_FLAGS+7, _CIF_MCCK_GUEST -4: BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST) +4: BPENTER __SF_SIE_FLAGS(%r15),_TIF_ISOLATE_BP_GUEST SIEEXIT - j .Lmcck_stack #endif .Lmcck_user: - BPENTER __TI_flags(%r12),_TIF_ISOLATE_BP -.Lmcck_stack: lg %r15,__LC_MCCK_STACK la %r11,STACK_FRAME_OVERHEAD(%r15) stctg %c1,%c1,__PT_CR1(%r11) @@ -567,7 +555,7 @@ ENTRY(mcck_int_handler) mvc __LC_RETURN_MCCK_PSW(16),__PT_PSW(%r11) # move return PSW tm __LC_RETURN_MCCK_PSW+1,0x01 # returning to user ? jno 0f - BPEXIT __TI_flags(%r12),_TIF_ISOLATE_BP + BPON stpt __LC_EXIT_TIMER 0: ALTERNATIVE "nop", __stringify(lghi %r12,__LC_LAST_BREAK_SAVE_AREA),193 LBEAR 0(%r12) @@ -583,10 +571,10 @@ ENTRY(mcck_int_handler) */ lhi %r5,0 lhi %r6,1 - larl %r7,.Lstop_lock + larl %r7,stop_lock cs %r5,%r6,0(%r7) # single CPU-stopper only jnz 4f - larl %r7,.Lthis_cpu + larl %r7,this_cpu stap 0(%r7) # this CPU address lh %r4,0(%r7) nilh %r4,0 @@ -602,16 +590,15 @@ ENTRY(mcck_int_handler) 3: sigp %r1,%r4,SIGP_STOP # stop this CPU brc SIGP_CC_BUSY,3b 4: j 4b -ENDPROC(mcck_int_handler) +SYM_CODE_END(mcck_int_handler) -ENTRY(restart_int_handler) +SYM_CODE_START(restart_int_handler) ALTERNATIVE "nop", "lpp _LPP_OFFSET", 40 stg %r15,__LC_SAVE_AREA_RESTART TSTMSK __LC_RESTART_FLAGS,RESTART_FLAG_CTLREGS,4 jz 0f - la %r15,4095 - lctlg %c0,%c15,__LC_CREGS_SAVE_AREA-4095(%r15) -0: larl %r15,.Lstosm_tmp + lctlg %c0,%c15,__LC_CREGS_SAVE_AREA +0: larl %r15,stosm_tmp stosm 0(%r15),0x04 # turn dat on, keep irqs off lg %r15,__LC_RESTART_STACK xc STACK_FRAME_OVERHEAD(__PT_SIZE,%r15),STACK_FRAME_OVERHEAD(%r15) @@ -632,7 +619,7 @@ ENTRY(restart_int_handler) 2: sigp %r4,%r3,SIGP_STOP # sigp stop to current cpu brc 2,2b 3: j 3b -ENDPROC(restart_int_handler) +SYM_CODE_END(restart_int_handler) .section .kprobes.text, "ax" @@ -642,7 +629,7 @@ ENDPROC(restart_int_handler) * No need to properly save the registers, we are going to panic anyway. * Setup a pt_regs so that show_trace can provide a good call trace. */ -ENTRY(stack_overflow) +SYM_CODE_START(stack_overflow) lg %r15,__LC_NODAT_STACK # change to panic stack la %r11,STACK_FRAME_OVERHEAD(%r15) stmg %r0,%r7,__PT_R0(%r11) @@ -652,26 +639,27 @@ ENTRY(stack_overflow) xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) lgr %r2,%r11 # pass pointer to pt_regs jg kernel_stack_overflow -ENDPROC(stack_overflow) +SYM_CODE_END(stack_overflow) #endif .section .data, "aw" - .align 4 -.Lstop_lock: .long 0 -.Lthis_cpu: .short 0 -.Lstosm_tmp: .byte 0 + .balign 4 +SYM_DATA_LOCAL(stop_lock, .long 0) +SYM_DATA_LOCAL(this_cpu, .short 0) +SYM_DATA_LOCAL(stosm_tmp, .byte 0) + .section .rodata, "a" #define SYSCALL(esame,emu) .quad __s390x_ ## esame - .globl sys_call_table -sys_call_table: +SYM_DATA_START(sys_call_table) #include "asm/syscall_table.h" +SYM_DATA_END(sys_call_table) #undef SYSCALL #ifdef CONFIG_COMPAT #define SYSCALL(esame,emu) .quad __s390_ ## emu - .globl sys_call_table_emu -sys_call_table_emu: +SYM_DATA_START(sys_call_table_emu) #include "asm/syscall_table.h" +SYM_DATA_END(sys_call_table_emu) #undef SYSCALL #endif diff --git a/arch/s390/kernel/ftrace.c b/arch/s390/kernel/ftrace.c index 416b5a94353d..c46381ea04ec 100644 --- a/arch/s390/kernel/ftrace.c +++ b/arch/s390/kernel/ftrace.c @@ -49,26 +49,6 @@ struct ftrace_insn { s32 disp; } __packed; -asm( - " .align 16\n" - "ftrace_shared_hotpatch_trampoline_br:\n" - " lmg %r0,%r1,2(%r1)\n" - " br %r1\n" - "ftrace_shared_hotpatch_trampoline_br_end:\n" -); - -#ifdef CONFIG_EXPOLINE -asm( - " .align 16\n" - "ftrace_shared_hotpatch_trampoline_exrl:\n" - " lmg %r0,%r1,2(%r1)\n" - " exrl %r0,0f\n" - " j .\n" - "0: br %r1\n" - "ftrace_shared_hotpatch_trampoline_exrl_end:\n" -); -#endif /* CONFIG_EXPOLINE */ - #ifdef CONFIG_MODULES static char *ftrace_plt; #endif /* CONFIG_MODULES */ @@ -246,7 +226,7 @@ static int __init ftrace_plt_init(void) start = ftrace_shared_hotpatch_trampoline(&end); memcpy(ftrace_plt, start, end - start); - set_memory_ro((unsigned long)ftrace_plt, 1); + set_memory_rox((unsigned long)ftrace_plt, 1); return 0; } device_initcall(ftrace_plt_init); diff --git a/arch/s390/kernel/head64.S b/arch/s390/kernel/head64.S index 3b3bf8329e6c..df77ba102096 100644 --- a/arch/s390/kernel/head64.S +++ b/arch/s390/kernel/head64.S @@ -16,7 +16,7 @@ #include <asm/ptrace.h> __HEAD -ENTRY(startup_continue) +SYM_CODE_START(startup_continue) larl %r1,tod_clock_base mvc 0(16,%r1),__LC_BOOT_CLOCK # @@ -24,19 +24,17 @@ ENTRY(startup_continue) # larl %r14,init_task stg %r14,__LC_CURRENT - larl %r15,init_thread_union+THREAD_SIZE-STACK_FRAME_OVERHEAD-__PT_SIZE + larl %r15,init_thread_union+STACK_INIT_OFFSET + stg %r15,__LC_KERNEL_STACK brasl %r14,sclp_early_adjust_va # allow sclp_early_printk -#ifdef CONFIG_KASAN - brasl %r14,kasan_early_init -#endif brasl %r14,startup_init # s390 specific early init brasl %r14,start_kernel # common init code # # We returned from start_kernel ?!? PANIK # basr %r13,0 - lpswe .Ldw-.(%r13) # load disabled wait psw + lpswe dw_psw-.(%r13) # load disabled wait psw +SYM_CODE_END(startup_continue) .align 16 -.LPG1: -.Ldw: .quad 0x0002000180000000,0x0000000000000000 +SYM_DATA_LOCAL(dw_psw, .quad 0x0002000180000000,0x0000000000000000) diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c index 5f0f5c86963a..43de939b7af1 100644 --- a/arch/s390/kernel/ipl.c +++ b/arch/s390/kernel/ipl.c @@ -176,11 +176,11 @@ static bool reipl_fcp_clear; static bool reipl_ccw_clear; static bool reipl_eckd_clear; -static inline int __diag308(unsigned long subcode, void *addr) +static inline int __diag308(unsigned long subcode, unsigned long addr) { union register_pair r1; - r1.even = (unsigned long) addr; + r1.even = addr; r1.odd = 0; asm volatile( " diag %[r1],%[subcode],0x308\n" @@ -195,7 +195,7 @@ static inline int __diag308(unsigned long subcode, void *addr) int diag308(unsigned long subcode, void *addr) { diag_stat_inc(DIAG_STAT_X308); - return __diag308(subcode, addr); + return __diag308(subcode, addr ? virt_to_phys(addr) : 0); } EXPORT_SYMBOL_GPL(diag308); @@ -649,7 +649,6 @@ static struct kset *ipl_kset; static void __ipl_run(void *unused) { - __bpon(); diag308(DIAG308_LOAD_CLEAR, NULL); } diff --git a/arch/s390/kernel/kprobes.c b/arch/s390/kernel/kprobes.c index 7b41ceecbb25..d4b863ed0aa7 100644 --- a/arch/s390/kernel/kprobes.c +++ b/arch/s390/kernel/kprobes.c @@ -41,7 +41,7 @@ void *alloc_insn_page(void) page = module_alloc(PAGE_SIZE); if (!page) return NULL; - __set_memory((unsigned long) page, 1, SET_MEMORY_RO | SET_MEMORY_X); + set_memory_rox((unsigned long)page, 1); return page; } diff --git a/arch/s390/kernel/kprobes_insn_page.S b/arch/s390/kernel/kprobes_insn_page.S index f6cb022ef8c8..b6335296dcd8 100644 --- a/arch/s390/kernel/kprobes_insn_page.S +++ b/arch/s390/kernel/kprobes_insn_page.S @@ -14,9 +14,9 @@ */ .section .kprobes.text, "ax" .align 4096 -ENTRY(kprobes_insn_page) +SYM_CODE_START(kprobes_insn_page) .rept 2048 .word 0x07fe .endr -ENDPROC(kprobes_insn_page) +SYM_CODE_END(kprobes_insn_page) .previous diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c index 2a8e73266428..6d9276c096a6 100644 --- a/arch/s390/kernel/machine_kexec.c +++ b/arch/s390/kernel/machine_kexec.c @@ -29,8 +29,8 @@ #include <asm/nmi.h> #include <asm/sclp.h> -typedef void (*relocate_kernel_t)(kimage_entry_t *, unsigned long, - unsigned long); +typedef void (*relocate_kernel_t)(unsigned long, unsigned long, unsigned long); +typedef int (*purgatory_t)(int); extern const unsigned char relocate_kernel[]; extern const unsigned long long relocate_kernel_len; @@ -41,11 +41,14 @@ extern const unsigned long long relocate_kernel_len; * Reset the system, copy boot CPU registers to absolute zero, * and jump to the kdump image */ -static void __do_machine_kdump(void *image) +static void __do_machine_kdump(void *data) { - int (*start_kdump)(int); + struct kimage *image = data; + purgatory_t purgatory; unsigned long prefix; + purgatory = (purgatory_t)image->start; + /* store_status() saved the prefix register to lowcore */ prefix = (unsigned long) S390_lowcore.prefixreg_save_area; @@ -58,13 +61,11 @@ static void __do_machine_kdump(void *image) * prefix register of this CPU to zero */ memcpy(absolute_pointer(__LC_FPREGS_SAVE_AREA), - (void *)(prefix + __LC_FPREGS_SAVE_AREA), 512); + phys_to_virt(prefix + __LC_FPREGS_SAVE_AREA), 512); - __load_psw_mask(PSW_MASK_BASE | PSW_DEFAULT_KEY | PSW_MASK_EA | PSW_MASK_BA); - start_kdump = (void *)((struct kimage *) image)->start; - start_kdump(1); + call_nodat(1, int, purgatory, int, 1); - /* Die if start_kdump returns */ + /* Die if kdump returns */ disabled_wait(); } @@ -111,18 +112,6 @@ static noinline void __machine_kdump(void *image) store_status(__do_machine_kdump, image); } -static unsigned long do_start_kdump(unsigned long addr) -{ - struct kimage *image = (struct kimage *) addr; - int (*start_kdump)(int) = (void *)image->start; - int rc; - - __arch_local_irq_stnsm(0xfb); /* disable DAT */ - rc = start_kdump(0); - __arch_local_irq_stosm(0x04); /* enable DAT */ - return rc; -} - #endif /* CONFIG_CRASH_DUMP */ /* @@ -131,12 +120,10 @@ static unsigned long do_start_kdump(unsigned long addr) static bool kdump_csum_valid(struct kimage *image) { #ifdef CONFIG_CRASH_DUMP + purgatory_t purgatory = (purgatory_t)image->start; int rc; - preempt_disable(); - rc = call_on_stack(1, S390_lowcore.nodat_stack, unsigned long, do_start_kdump, - unsigned long, (unsigned long)image); - preempt_enable(); + rc = call_nodat(1, int, purgatory, int, 0); return rc == 0; #else return false; @@ -210,7 +197,7 @@ int machine_kexec_prepare(struct kimage *image) return -EINVAL; /* Get the destination where the assembler code should be copied to.*/ - reboot_code_buffer = (void *) page_to_phys(image->control_code_page); + reboot_code_buffer = page_to_virt(image->control_code_page); /* Then copy it */ memcpy(reboot_code_buffer, relocate_kernel, relocate_kernel_len); @@ -250,19 +237,20 @@ void machine_crash_shutdown(struct pt_regs *regs) */ static void __do_machine_kexec(void *data) { - unsigned long diag308_subcode; - relocate_kernel_t data_mover; + unsigned long data_mover, entry, diag308_subcode; struct kimage *image = data; - s390_reset_system(); - data_mover = (relocate_kernel_t) page_to_phys(image->control_code_page); - - __arch_local_irq_stnsm(0xfb); /* disable DAT - avoid no-execute */ - /* Call the moving routine */ + data_mover = page_to_phys(image->control_code_page); + entry = virt_to_phys(&image->head); diag308_subcode = DIAG308_CLEAR_RESET; if (sclp.has_iplcc) diag308_subcode |= DIAG308_FLAG_EI; - (*data_mover)(&image->head, image->start, diag308_subcode); + s390_reset_system(); + + call_nodat(3, void, (relocate_kernel_t)data_mover, + unsigned long, entry, + unsigned long, image->start, + unsigned long, diag308_subcode); /* Die if kexec returns */ disabled_wait(); diff --git a/arch/s390/kernel/mcount.S b/arch/s390/kernel/mcount.S index 6c10da43b538..dbece2803c50 100644 --- a/arch/s390/kernel/mcount.S +++ b/arch/s390/kernel/mcount.S @@ -28,9 +28,9 @@ .section .kprobes.text, "ax" -ENTRY(ftrace_stub) +SYM_FUNC_START(ftrace_stub) BR_EX %r14 -ENDPROC(ftrace_stub) +SYM_FUNC_END(ftrace_stub) SYM_CODE_START(ftrace_stub_direct_tramp) lgr %r1, %r0 @@ -140,10 +140,25 @@ SYM_FUNC_END(return_to_handler) #endif #endif /* CONFIG_FUNCTION_TRACER */ -#ifdef CONFIG_RETHOOK +SYM_CODE_START(ftrace_shared_hotpatch_trampoline_br) + lmg %r0,%r1,2(%r1) + br %r1 +SYM_INNER_LABEL(ftrace_shared_hotpatch_trampoline_br_end, SYM_L_GLOBAL) +SYM_CODE_END(ftrace_shared_hotpatch_trampoline_br) + +#ifdef CONFIG_EXPOLINE +SYM_CODE_START(ftrace_shared_hotpatch_trampoline_exrl) + lmg %r0,%r1,2(%r1) + exrl %r0,0f + j . +0: br %r1 +SYM_INNER_LABEL(ftrace_shared_hotpatch_trampoline_exrl_end, SYM_L_GLOBAL) +SYM_CODE_END(ftrace_shared_hotpatch_trampoline_exrl) +#endif /* CONFIG_EXPOLINE */ -SYM_FUNC_START(arch_rethook_trampoline) +#ifdef CONFIG_RETHOOK +SYM_CODE_START(arch_rethook_trampoline) stg %r14,(__SF_GPRS+8*8)(%r15) lay %r15,-STACK_FRAME_SIZE(%r15) stmg %r0,%r14,STACK_PTREGS_GPRS(%r15) @@ -166,7 +181,6 @@ SYM_FUNC_START(arch_rethook_trampoline) mvc __SF_EMPTY(16,%r7),STACK_PTREGS_PSW(%r15) lmg %r0,%r15,STACK_PTREGS_GPRS(%r15) lpswe __SF_EMPTY(%r15) - -SYM_FUNC_END(arch_rethook_trampoline) +SYM_CODE_END(arch_rethook_trampoline) #endif /* CONFIG_RETHOOK */ diff --git a/arch/s390/kernel/module.c b/arch/s390/kernel/module.c index b9dbf0c483ed..f1b35dcdf3eb 100644 --- a/arch/s390/kernel/module.c +++ b/arch/s390/kernel/module.c @@ -26,6 +26,7 @@ #include <asm/facility.h> #include <asm/ftrace.lds.h> #include <asm/set_memory.h> +#include <asm/setup.h> #if 0 #define DEBUGP printk @@ -35,6 +36,24 @@ #define PLT_ENTRY_SIZE 22 +static unsigned long get_module_load_offset(void) +{ + static DEFINE_MUTEX(module_kaslr_mutex); + static unsigned long module_load_offset; + + if (!kaslr_enabled()) + return 0; + /* + * Calculate the module_load_offset the first time this code + * is called. Once calculated it stays the same until reboot. + */ + mutex_lock(&module_kaslr_mutex); + if (!module_load_offset) + module_load_offset = get_random_u32_inclusive(1, 1024) * PAGE_SIZE; + mutex_unlock(&module_kaslr_mutex); + return module_load_offset; +} + void *module_alloc(unsigned long size) { gfp_t gfp_mask = GFP_KERNEL; @@ -42,9 +61,11 @@ void *module_alloc(unsigned long size) if (PAGE_ALIGN(size) > MODULES_LEN) return NULL; - p = __vmalloc_node_range(size, MODULE_ALIGN, MODULES_VADDR, MODULES_END, - gfp_mask, PAGE_KERNEL_EXEC, VM_DEFER_KMEMLEAK, NUMA_NO_NODE, - __builtin_return_address(0)); + p = __vmalloc_node_range(size, MODULE_ALIGN, + MODULES_VADDR + get_module_load_offset(), + MODULES_END, gfp_mask, PAGE_KERNEL, + VM_FLUSH_RESET_PERMS | VM_DEFER_KMEMLEAK, + NUMA_NO_NODE, __builtin_return_address(0)); if (p && (kasan_alloc_module_shadow(p, size, gfp_mask) < 0)) { vfree(p); return NULL; @@ -491,7 +512,7 @@ static int module_alloc_ftrace_hotpatch_trampolines(struct module *me, start = module_alloc(numpages * PAGE_SIZE); if (!start) return -ENOMEM; - set_memory_ro((unsigned long)start, numpages); + set_memory_rox((unsigned long)start, numpages); end = start + size; me->arch.trampolines_start = (struct ftrace_hotpatch_trampoline *)start; diff --git a/arch/s390/kernel/perf_cpum_cf.c b/arch/s390/kernel/perf_cpum_cf.c index c9ab971498d6..cf1b6e8a708d 100644 --- a/arch/s390/kernel/perf_cpum_cf.c +++ b/arch/s390/kernel/perf_cpum_cf.c @@ -76,7 +76,6 @@ static inline int ctr_stcctm(enum cpumf_ctr_set set, u64 range, u64 *dest) } struct cpu_cf_events { - struct cpumf_ctr_info info; atomic_t ctr_set[CPUMF_CTR_SET_MAX]; u64 state; /* For perf_event_open SVC */ u64 dev_state; /* For /dev/hwctr */ @@ -95,6 +94,15 @@ static DEFINE_PER_CPU(struct cpu_cf_events, cpu_cf_events); static unsigned int cfdiag_cpu_speed; /* CPU speed for CF_DIAG trailer */ static debug_info_t *cf_dbg; +/* + * The CPU Measurement query counter information instruction contains + * information which varies per machine generation, but is constant and + * does not change when running on a particular machine, such as counter + * first and second version number. This is needed to determine the size + * of counter sets. Extract this information at device driver initialization. + */ +static struct cpumf_ctr_info cpumf_ctr_info; + #define CF_DIAG_CTRSET_DEF 0xfeef /* Counter set header mark */ /* interval in seconds */ @@ -167,11 +175,10 @@ struct cf_trailer_entry { /* CPU-M CF_DIAG trailer (64 byte) */ /* Create the trailer data at the end of a page. */ static void cfdiag_trailer(struct cf_trailer_entry *te) { - struct cpu_cf_events *cpuhw = this_cpu_ptr(&cpu_cf_events); struct cpuid cpuid; - te->cfvn = cpuhw->info.cfvn; /* Counter version numbers */ - te->csvn = cpuhw->info.csvn; + te->cfvn = cpumf_ctr_info.cfvn; /* Counter version numbers */ + te->csvn = cpumf_ctr_info.csvn; get_cpu_id(&cpuid); /* Machine type */ te->mach_type = cpuid.machine; @@ -184,50 +191,60 @@ static void cfdiag_trailer(struct cf_trailer_entry *te) } /* - * Return the maximum possible counter set size (in number of 8 byte counters) - * depending on type and model number. + * The number of counters per counter set varies between machine generations, + * but is constant when running on a particular machine generation. + * Determine each counter set size at device driver initialization and + * retrieve it later. */ -static size_t cpum_cf_ctrset_size(enum cpumf_ctr_set ctrset, - struct cpumf_ctr_info *info) +static size_t cpumf_ctr_setsizes[CPUMF_CTR_SET_MAX]; +static void cpum_cf_make_setsize(enum cpumf_ctr_set ctrset) { size_t ctrset_size = 0; switch (ctrset) { case CPUMF_CTR_SET_BASIC: - if (info->cfvn >= 1) + if (cpumf_ctr_info.cfvn >= 1) ctrset_size = 6; break; case CPUMF_CTR_SET_USER: - if (info->cfvn == 1) + if (cpumf_ctr_info.cfvn == 1) ctrset_size = 6; - else if (info->cfvn >= 3) + else if (cpumf_ctr_info.cfvn >= 3) ctrset_size = 2; break; case CPUMF_CTR_SET_CRYPTO: - if (info->csvn >= 1 && info->csvn <= 5) + if (cpumf_ctr_info.csvn >= 1 && cpumf_ctr_info.csvn <= 5) ctrset_size = 16; - else if (info->csvn == 6 || info->csvn == 7) + else if (cpumf_ctr_info.csvn == 6 || cpumf_ctr_info.csvn == 7) ctrset_size = 20; break; case CPUMF_CTR_SET_EXT: - if (info->csvn == 1) + if (cpumf_ctr_info.csvn == 1) ctrset_size = 32; - else if (info->csvn == 2) + else if (cpumf_ctr_info.csvn == 2) ctrset_size = 48; - else if (info->csvn >= 3 && info->csvn <= 5) + else if (cpumf_ctr_info.csvn >= 3 && cpumf_ctr_info.csvn <= 5) ctrset_size = 128; - else if (info->csvn == 6 || info->csvn == 7) + else if (cpumf_ctr_info.csvn == 6 || cpumf_ctr_info.csvn == 7) ctrset_size = 160; break; case CPUMF_CTR_SET_MT_DIAG: - if (info->csvn > 3) + if (cpumf_ctr_info.csvn > 3) ctrset_size = 48; break; case CPUMF_CTR_SET_MAX: break; } + cpumf_ctr_setsizes[ctrset] = ctrset_size; +} - return ctrset_size; +/* + * Return the maximum possible counter set size (in number of 8 byte counters) + * depending on type and model number. + */ +static size_t cpum_cf_read_setsize(enum cpumf_ctr_set ctrset) +{ + return cpumf_ctr_setsizes[ctrset]; } /* Read a counter set. The counter set number determines the counter set and @@ -248,14 +265,13 @@ static size_t cpum_cf_ctrset_size(enum cpumf_ctr_set ctrset, static size_t cfdiag_getctrset(struct cf_ctrset_entry *ctrdata, int ctrset, size_t room, bool error_ok) { - struct cpu_cf_events *cpuhw = this_cpu_ptr(&cpu_cf_events); size_t ctrset_size, need = 0; int rc = 3; /* Assume write failure */ ctrdata->def = CF_DIAG_CTRSET_DEF; ctrdata->set = ctrset; ctrdata->res1 = 0; - ctrset_size = cpum_cf_ctrset_size(ctrset, &cpuhw->info); + ctrset_size = cpum_cf_read_setsize(ctrset); if (ctrset_size) { /* Save data */ need = ctrset_size * sizeof(u64) + sizeof(*ctrdata); @@ -269,10 +285,6 @@ static size_t cfdiag_getctrset(struct cf_ctrset_entry *ctrdata, int ctrset, need = 0; } - debug_sprintf_event(cf_dbg, 3, - "%s ctrset %d ctrset_size %zu cfvn %d csvn %d" - " need %zd rc %d\n", __func__, ctrset, ctrset_size, - cpuhw->info.cfvn, cpuhw->info.csvn, need, rc); return need; } @@ -377,40 +389,35 @@ static enum cpumf_ctr_set get_counter_set(u64 event) return set; } -static int validate_ctr_version(const struct hw_perf_event *hwc, - enum cpumf_ctr_set set) +static int validate_ctr_version(const u64 config, enum cpumf_ctr_set set) { - struct cpu_cf_events *cpuhw; - int err = 0; u16 mtdiag_ctl; - - cpuhw = &get_cpu_var(cpu_cf_events); + int err = 0; /* check required version for counter sets */ switch (set) { case CPUMF_CTR_SET_BASIC: case CPUMF_CTR_SET_USER: - if (cpuhw->info.cfvn < 1) + if (cpumf_ctr_info.cfvn < 1) err = -EOPNOTSUPP; break; case CPUMF_CTR_SET_CRYPTO: - if ((cpuhw->info.csvn >= 1 && cpuhw->info.csvn <= 5 && - hwc->config > 79) || - (cpuhw->info.csvn >= 6 && hwc->config > 83)) + if ((cpumf_ctr_info.csvn >= 1 && cpumf_ctr_info.csvn <= 5 && + config > 79) || (cpumf_ctr_info.csvn >= 6 && config > 83)) err = -EOPNOTSUPP; break; case CPUMF_CTR_SET_EXT: - if (cpuhw->info.csvn < 1) + if (cpumf_ctr_info.csvn < 1) err = -EOPNOTSUPP; - if ((cpuhw->info.csvn == 1 && hwc->config > 159) || - (cpuhw->info.csvn == 2 && hwc->config > 175) || - (cpuhw->info.csvn >= 3 && cpuhw->info.csvn <= 5 - && hwc->config > 255) || - (cpuhw->info.csvn >= 6 && hwc->config > 287)) + if ((cpumf_ctr_info.csvn == 1 && config > 159) || + (cpumf_ctr_info.csvn == 2 && config > 175) || + (cpumf_ctr_info.csvn >= 3 && cpumf_ctr_info.csvn <= 5 && + config > 255) || + (cpumf_ctr_info.csvn >= 6 && config > 287)) err = -EOPNOTSUPP; break; case CPUMF_CTR_SET_MT_DIAG: - if (cpuhw->info.csvn <= 3) + if (cpumf_ctr_info.csvn <= 3) err = -EOPNOTSUPP; /* * MT-diagnostic counters are read-only. The counter set @@ -425,35 +432,15 @@ static int validate_ctr_version(const struct hw_perf_event *hwc, * counter set is enabled and active. */ mtdiag_ctl = cpumf_ctr_ctl[CPUMF_CTR_SET_MT_DIAG]; - if (!((cpuhw->info.auth_ctl & mtdiag_ctl) && - (cpuhw->info.enable_ctl & mtdiag_ctl) && - (cpuhw->info.act_ctl & mtdiag_ctl))) + if (!((cpumf_ctr_info.auth_ctl & mtdiag_ctl) && + (cpumf_ctr_info.enable_ctl & mtdiag_ctl) && + (cpumf_ctr_info.act_ctl & mtdiag_ctl))) err = -EOPNOTSUPP; break; case CPUMF_CTR_SET_MAX: err = -EOPNOTSUPP; } - put_cpu_var(cpu_cf_events); - return err; -} - -static int validate_ctr_auth(const struct hw_perf_event *hwc) -{ - struct cpu_cf_events *cpuhw; - int err = 0; - - cpuhw = &get_cpu_var(cpu_cf_events); - - /* Check authorization for cpu counter sets. - * If the particular CPU counter set is not authorized, - * return with -ENOENT in order to fall back to other - * PMUs that might suffice the event request. - */ - if (!(hwc->config_base & cpuhw->info.auth_ctl)) - err = -ENOENT; - - put_cpu_var(cpu_cf_events); return err; } @@ -471,13 +458,10 @@ static void cpumf_pmu_enable(struct pmu *pmu) return; err = lcctl(cpuhw->state | cpuhw->dev_state); - if (err) { - pr_err("Enabling the performance measuring unit " - "failed with rc=%x\n", err); - return; - } - - cpuhw->flags |= PMU_F_ENABLED; + if (err) + pr_err("Enabling the performance measuring unit failed with rc=%x\n", err); + else + cpuhw->flags |= PMU_F_ENABLED; } /* @@ -497,13 +481,10 @@ static void cpumf_pmu_disable(struct pmu *pmu) inactive = cpuhw->state & ~((1 << CPUMF_LCCTL_ENABLE_SHIFT) - 1); inactive |= cpuhw->dev_state; err = lcctl(inactive); - if (err) { - pr_err("Disabling the performance measuring unit " - "failed with rc=%x\n", err); - return; - } - - cpuhw->flags &= ~PMU_F_ENABLED; + if (err) + pr_err("Disabling the performance measuring unit failed with rc=%x\n", err); + else + cpuhw->flags &= ~PMU_F_ENABLED; } #define PMC_INIT 0UL @@ -515,8 +496,6 @@ static void cpum_cf_setup_cpu(void *flags) switch ((unsigned long)flags) { case PMC_INIT: - memset(&cpuhw->info, 0, sizeof(cpuhw->info)); - qctri(&cpuhw->info); cpuhw->flags |= PMU_F_RESERVED; break; @@ -602,7 +581,6 @@ static int __hw_perf_event_init(struct perf_event *event, unsigned int type) struct perf_event_attr *attr = &event->attr; struct hw_perf_event *hwc = &event->hw; enum cpumf_ctr_set set; - int err = 0; u64 ev; switch (type) { @@ -678,12 +656,15 @@ static int __hw_perf_event_init(struct perf_event *event, unsigned int type) cpumf_hw_inuse(); event->destroy = hw_perf_event_destroy; - /* Finally, validate version and authorization of the counter set */ - err = validate_ctr_auth(hwc); - if (!err) - err = validate_ctr_version(hwc, set); - - return err; + /* + * Finally, validate version and authorization of the counter set. + * If the particular CPU counter set is not authorized, + * return with -ENOENT in order to fall back to other + * PMUs that might suffice the event request. + */ + if (!(hwc->config_base & cpumf_ctr_info.auth_ctl)) + return -ENOENT; + return validate_ctr_version(hwc->config, set); } /* Events CPU_CYLCES and INSTRUCTIONS can be submitted with two different @@ -983,7 +964,7 @@ static void cpumf_measurement_alert(struct ext_code ext_code, /* counter authorization change alert */ if (alert & CPU_MF_INT_CF_CACA) - qctri(&cpuhw->info); + qctri(&cpumf_ctr_info); /* loss of counter data alert */ if (alert & CPU_MF_INT_CF_LCDA) @@ -1000,9 +981,14 @@ static int __init cpumf_pmu_init(void) { int rc; - if (!cpum_cf_avail()) + /* Extract counter measurement facility information */ + if (!cpum_cf_avail() || qctri(&cpumf_ctr_info)) return -ENODEV; + /* Determine and store counter set sizes for later reference */ + for (rc = CPUMF_CTR_SET_BASIC; rc < CPUMF_CTR_SET_MAX; ++rc) + cpum_cf_make_setsize(rc); + /* * Clear bit 15 of cr0 to unauthorize problem-state to * extract measurement counters @@ -1269,28 +1255,26 @@ static int cfset_all_start(struct cfset_request *req) */ static size_t cfset_needspace(unsigned int sets) { - struct cpu_cf_events *cpuhw = get_cpu_ptr(&cpu_cf_events); size_t bytes = 0; int i; for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { if (!(sets & cpumf_ctr_ctl[i])) continue; - bytes += cpum_cf_ctrset_size(i, &cpuhw->info) * sizeof(u64) + + bytes += cpum_cf_read_setsize(i) * sizeof(u64) + sizeof(((struct s390_ctrset_setdata *)0)->set) + sizeof(((struct s390_ctrset_setdata *)0)->no_cnts); } bytes = sizeof(((struct s390_ctrset_read *)0)->no_cpus) + nr_cpu_ids * (bytes + sizeof(((struct s390_ctrset_cpudata *)0)->cpu_nr) + sizeof(((struct s390_ctrset_cpudata *)0)->no_sets)); - put_cpu_ptr(&cpu_cf_events); return bytes; } static int cfset_all_copy(unsigned long arg, cpumask_t *mask) { struct s390_ctrset_read __user *ctrset_read; - unsigned int cpu, cpus, rc; + unsigned int cpu, cpus, rc = 0; void __user *uptr; ctrset_read = (struct s390_ctrset_read __user *)arg; @@ -1304,17 +1288,20 @@ static int cfset_all_copy(unsigned long arg, cpumask_t *mask) rc |= put_user(cpuhw->sets, &ctrset_cpudata->no_sets); rc |= copy_to_user(ctrset_cpudata->data, cpuhw->data, cpuhw->used); - if (rc) - return -EFAULT; + if (rc) { + rc = -EFAULT; + goto out; + } uptr += sizeof(struct s390_ctrset_cpudata) + cpuhw->used; cond_resched(); } cpus = cpumask_weight(mask); if (put_user(cpus, &ctrset_read->no_cpus)) - return -EFAULT; - debug_sprintf_event(cf_dbg, 4, "%s copied %ld\n", __func__, + rc = -EFAULT; +out: + debug_sprintf_event(cf_dbg, 4, "%s rc %d copied %ld\n", __func__, rc, uptr - (void __user *)ctrset_read->data); - return 0; + return rc; } static size_t cfset_cpuset_read(struct s390_ctrset_setdata *p, int ctrset, @@ -1354,7 +1341,7 @@ static void cfset_cpu_read(void *parm) if (!(p->sets & cpumf_ctr_ctl[set])) continue; /* Counter set not in list */ - set_size = cpum_cf_ctrset_size(set, &cpuhw->info); + set_size = cpum_cf_read_setsize(set); space = sizeof(cpuhw->data) - cpuhw->used; space = cfset_cpuset_read(sp, set, set_size, space); if (space) { @@ -1385,14 +1372,10 @@ static int cfset_all_read(unsigned long arg, struct cfset_request *req) static long cfset_ioctl_read(unsigned long arg, struct cfset_request *req) { - struct s390_ctrset_read read; int ret = -ENODATA; - if (req && req->ctrset) { - if (copy_from_user(&read, (char __user *)arg, sizeof(read))) - return -EFAULT; + if (req && req->ctrset) ret = cfset_all_read(arg, req); - } return ret; } @@ -1569,16 +1552,13 @@ static void cfdiag_read(struct perf_event *event) static int get_authctrsets(void) { - struct cpu_cf_events *cpuhw; unsigned long auth = 0; enum cpumf_ctr_set i; - cpuhw = &get_cpu_var(cpu_cf_events); for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { - if (cpuhw->info.auth_ctl & cpumf_ctr_ctl[i]) + if (cpumf_ctr_info.auth_ctl & cpumf_ctr_ctl[i]) auth |= cpumf_ctr_ctl[i]; } - put_cpu_var(cpu_cf_events); return auth; } @@ -1716,7 +1696,7 @@ static size_t cfdiag_maxsize(struct cpumf_ctr_info *info) enum cpumf_ctr_set i; for (i = CPUMF_CTR_SET_BASIC; i < CPUMF_CTR_SET_MAX; ++i) { - size_t size = cpum_cf_ctrset_size(i, info); + size_t size = cpum_cf_read_setsize(i); if (size) max_size += size * sizeof(u64) + @@ -1750,16 +1730,12 @@ static void cfdiag_get_cpu_speed(void) static int cfset_init(void) { - struct cpumf_ctr_info info; size_t need; int rc; - if (qctri(&info)) - return -ENODEV; - cfdiag_get_cpu_speed(); /* Make sure the counter set data fits into predefined buffer. */ - need = cfdiag_maxsize(&info); + need = cfdiag_maxsize(&cpumf_ctr_info); if (need > sizeof(((struct cpu_cf_events *)0)->start)) { pr_err("Insufficient memory for PMU(cpum_cf_diag) need=%zu\n", need); diff --git a/arch/s390/kernel/perf_cpum_sf.c b/arch/s390/kernel/perf_cpum_sf.c index e7b867e2f73f..7ef72f5ff52e 100644 --- a/arch/s390/kernel/perf_cpum_sf.c +++ b/arch/s390/kernel/perf_cpum_sf.c @@ -882,10 +882,6 @@ static int __hw_perf_event_init(struct perf_event *event) SAMPL_FLAGS(hwc) |= PERF_CPUM_SF_DIAG_MODE; } - /* Check and set other sampling flags */ - if (attr->config1 & PERF_CPUM_SF_FULL_BLOCKS) - SAMPL_FLAGS(hwc) |= PERF_CPUM_SF_FULL_BLOCKS; - err = __hw_perf_event_init_rate(event, &si); if (err) goto out; @@ -1293,11 +1289,8 @@ static inline __uint128_t __cdsg(__uint128_t *ptr, __uint128_t old, __uint128_t * The sampling buffer position are retrieved and saved in the TEAR_REG * register of the specified perf event. * - * Only full sample-data-blocks are processed. Specify the flash_all flag - * to also walk through partially filled sample-data-blocks. It is ignored - * if PERF_CPUM_SF_FULL_BLOCKS is set. The PERF_CPUM_SF_FULL_BLOCKS flag - * enforces the processing of full sample-data-blocks only (trailer entries - * with the block-full-indicator bit set). + * Only full sample-data-blocks are processed. Specify the flush_all flag + * to also walk through partially filled sample-data-blocks. */ static void hw_perf_event_update(struct perf_event *event, int flush_all) { @@ -1315,9 +1308,6 @@ static void hw_perf_event_update(struct perf_event *event, int flush_all) if (SAMPL_DIAG_MODE(&event->hw)) return; - if (flush_all && SDB_FULL_BLOCKS(hwc)) - flush_all = 0; - sdbt = (unsigned long *) TEAR_REG(hwc); done = event_overflow = sampl_overflow = num_sdb = 0; while (!done) { diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c index 67df64ef4839..87ca3a727604 100644 --- a/arch/s390/kernel/process.c +++ b/arch/s390/kernel/process.c @@ -136,12 +136,12 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args) p->thread.last_break = 1; frame->sf.back_chain = 0; - frame->sf.gprs[5] = (unsigned long)frame + sizeof(struct stack_frame); - frame->sf.gprs[6] = (unsigned long)p; + frame->sf.gprs[11 - 6] = (unsigned long)&frame->childregs; + frame->sf.gprs[12 - 6] = (unsigned long)p; /* new return point is ret_from_fork */ - frame->sf.gprs[8] = (unsigned long)ret_from_fork; + frame->sf.gprs[14 - 6] = (unsigned long)ret_from_fork; /* fake return stack for resume(), don't go back to schedule */ - frame->sf.gprs[9] = (unsigned long)frame; + frame->sf.gprs[15 - 6] = (unsigned long)frame; /* Store access registers to kernel stack of new process. */ if (unlikely(args->fn)) { @@ -149,8 +149,6 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args) memset(&frame->childregs, 0, sizeof(struct pt_regs)); frame->childregs.psw.mask = PSW_KERNEL_BITS | PSW_MASK_IO | PSW_MASK_EXT | PSW_MASK_MCHECK; - frame->childregs.psw.addr = - (unsigned long)__ret_from_fork; frame->childregs.gprs[9] = (unsigned long)args->fn; frame->childregs.gprs[10] = (unsigned long)args->fn_arg; frame->childregs.orig_gpr2 = -1; diff --git a/arch/s390/kernel/processor.c b/arch/s390/kernel/processor.c index a194611ba88c..0a999c8226d7 100644 --- a/arch/s390/kernel/processor.c +++ b/arch/s390/kernel/processor.c @@ -364,21 +364,3 @@ const struct seq_operations cpuinfo_op = { .stop = c_stop, .show = show_cpuinfo, }; - -int s390_isolate_bp(void) -{ - if (!test_facility(82)) - return -EOPNOTSUPP; - set_thread_flag(TIF_ISOLATE_BP); - return 0; -} -EXPORT_SYMBOL(s390_isolate_bp); - -int s390_isolate_bp_guest(void) -{ - if (!test_facility(82)) - return -EOPNOTSUPP; - set_thread_flag(TIF_ISOLATE_BP_GUEST); - return 0; -} -EXPORT_SYMBOL(s390_isolate_bp_guest); diff --git a/arch/s390/kernel/reipl.S b/arch/s390/kernel/reipl.S index 4a22163962eb..88087a32ebc6 100644 --- a/arch/s390/kernel/reipl.S +++ b/arch/s390/kernel/reipl.S @@ -19,7 +19,7 @@ # r2 = Function to be called after store status # r3 = Parameter for function # -ENTRY(store_status) +SYM_CODE_START(store_status) /* Save register one and load save area base */ stg %r1,__LC_SAVE_AREA_RESTART /* General purpose registers */ @@ -61,7 +61,7 @@ ENTRY(store_status) stpx 0(%r1) /* Clock comparator - seven bytes */ lghi %r1,__LC_CLOCK_COMP_SAVE_AREA - larl %r4,.Lclkcmp + larl %r4,clkcmp stckc 0(%r4) mvc 1(7,%r1),1(%r4) /* Program status word */ @@ -73,9 +73,9 @@ ENTRY(store_status) lgr %r9,%r2 lgr %r2,%r3 BR_EX %r9 -ENDPROC(store_status) +SYM_CODE_END(store_status) .section .bss - .align 8 -.Lclkcmp: .quad 0x0000000000000000 + .balign 8 +SYM_DATA_LOCAL(clkcmp, .quad 0x0000000000000000) .previous diff --git a/arch/s390/kernel/relocate_kernel.S b/arch/s390/kernel/relocate_kernel.S index a9a1a6f45375..0ae297c82afd 100644 --- a/arch/s390/kernel/relocate_kernel.S +++ b/arch/s390/kernel/relocate_kernel.S @@ -26,53 +26,51 @@ */ .text -ENTRY(relocate_kernel) - basr %r13,0 # base address - .base: - lghi %r7,PAGE_SIZE # load PAGE_SIZE in r7 - lghi %r9,PAGE_SIZE # load PAGE_SIZE in r9 - lg %r5,0(%r2) # read another word for indirection page - aghi %r2,8 # increment pointer - tml %r5,0x1 # is it a destination page? - je .indir_check # NO, goto "indir_check" - lgr %r6,%r5 # r6 = r5 - nill %r6,0xf000 # mask it out and... - j .base # ...next iteration - .indir_check: - tml %r5,0x2 # is it a indirection page? - je .done_test # NO, goto "done_test" - nill %r5,0xf000 # YES, mask out, - lgr %r2,%r5 # move it into the right register, - j .base # and read next... - .done_test: - tml %r5,0x4 # is it the done indicator? - je .source_test # NO! Well, then it should be the source indicator... - j .done # ok, lets finish it here... - .source_test: - tml %r5,0x8 # it should be a source indicator... - je .base # NO, ignore it... - lgr %r8,%r5 # r8 = r5 - nill %r8,0xf000 # masking - 0: mvcle %r6,%r8,0x0 # copy PAGE_SIZE bytes from r8 to r6 - pad with 0 - jo 0b - j .base - .done: - lgr %r0,%r4 # subcode - cghi %r3,0 - je .diag - la %r4,load_psw-.base(%r13) # load psw-address into the register - o %r3,4(%r4) # or load address into psw - st %r3,4(%r4) - mvc 0(8,%r0),0(%r4) # copy psw to absolute address 0 - .diag: - diag %r0,%r0,0x308 -ENDPROC(relocate_kernel) +SYM_CODE_START(relocate_kernel) + basr %r13,0 # base address +.base: + lghi %r7,PAGE_SIZE # load PAGE_SIZE in r7 + lghi %r9,PAGE_SIZE # load PAGE_SIZE in r9 + lg %r5,0(%r2) # read another word for indirection page + aghi %r2,8 # increment pointer + tml %r5,0x1 # is it a destination page? + je .indir_check # NO, goto "indir_check" + lgr %r6,%r5 # r6 = r5 + nill %r6,0xf000 # mask it out and... + j .base # ...next iteration +.indir_check: + tml %r5,0x2 # is it a indirection page? + je .done_test # NO, goto "done_test" + nill %r5,0xf000 # YES, mask out, + lgr %r2,%r5 # move it into the right register, + j .base # and read next... +.done_test: + tml %r5,0x4 # is it the done indicator? + je .source_test # NO! Well, then it should be the source indicator... + j .done # ok, lets finish it here... +.source_test: + tml %r5,0x8 # it should be a source indicator... + je .base # NO, ignore it... + lgr %r8,%r5 # r8 = r5 + nill %r8,0xf000 # masking +0: mvcle %r6,%r8,0x0 # copy PAGE_SIZE bytes from r8 to r6 - pad with 0 + jo 0b + j .base +.done: + lgr %r0,%r4 # subcode + cghi %r3,0 + je .diag + la %r4,load_psw-.base(%r13) # load psw-address into the register + o %r3,4(%r4) # or load address into psw + st %r3,4(%r4) + mvc 0(8,%r0),0(%r4) # copy psw to absolute address 0 +.diag: + diag %r0,%r0,0x308 +SYM_CODE_END(relocate_kernel) - .align 8 - load_psw: - .long 0x00080000,0x80000000 - relocate_kernel_end: - .align 8 - .globl relocate_kernel_len - relocate_kernel_len: - .quad relocate_kernel_end - relocate_kernel + .balign 8 +SYM_DATA_START_LOCAL(load_psw) + .long 0x00080000,0x80000000 +SYM_DATA_END_LABEL(load_psw, SYM_L_LOCAL, relocate_kernel_end) + .balign 8 +SYM_DATA(relocate_kernel_len, .quad relocate_kernel_end - relocate_kernel) diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c index 4259b6c50516..fe10da1a271e 100644 --- a/arch/s390/kernel/setup.c +++ b/arch/s390/kernel/setup.c @@ -74,7 +74,7 @@ #include <asm/numa.h> #include <asm/alternative.h> #include <asm/nospec-branch.h> -#include <asm/mem_detect.h> +#include <asm/physmem_info.h> #include <asm/maccess.h> #include <asm/uv.h> #include <asm/asm-offsets.h> @@ -147,14 +147,10 @@ static u32 __amode31_ref *__ctl_duct = __ctl_duct_amode31; int __bootdata(noexec_disabled); unsigned long __bootdata(ident_map_size); -struct mem_detect_info __bootdata(mem_detect); -struct initrd_data __bootdata(initrd_data); -unsigned long __bootdata(pgalloc_pos); -unsigned long __bootdata(pgalloc_end); -unsigned long __bootdata(pgalloc_low); +struct physmem_info __bootdata(physmem_info); unsigned long __bootdata_preserved(__kaslr_offset); -unsigned long __bootdata(__amode31_base); +int __bootdata_preserved(__kaslr_enabled); unsigned int __bootdata_preserved(zlib_dfltcc_support); EXPORT_SYMBOL(zlib_dfltcc_support); u64 __bootdata_preserved(stfle_fac_list[16]); @@ -385,39 +381,27 @@ void stack_free(unsigned long stack) #endif } -int __init arch_early_irq_init(void) +void __init __noreturn arch_call_rest_init(void) { - unsigned long stack; - - stack = __get_free_pages(GFP_KERNEL, THREAD_SIZE_ORDER); - if (!stack) - panic("Couldn't allocate async stack"); - S390_lowcore.async_stack = stack + STACK_INIT_OFFSET; - return 0; + smp_reinit_ipl_cpu(); + rest_init(); } -void __init __noreturn arch_call_rest_init(void) +static unsigned long __init stack_alloc_early(void) { unsigned long stack; - smp_reinit_ipl_cpu(); - stack = stack_alloc(); - if (!stack) - panic("Couldn't allocate kernel stack"); - current->stack = (void *) stack; -#ifdef CONFIG_VMAP_STACK - current->stack_vm_area = (void *) stack; -#endif - set_task_stack_end_magic(current); - stack += STACK_INIT_OFFSET; - S390_lowcore.kernel_stack = stack; - call_on_stack_noreturn(rest_init, stack); + stack = (unsigned long)memblock_alloc(THREAD_SIZE, THREAD_SIZE); + if (!stack) { + panic("%s: Failed to allocate %lu bytes align=0x%lx\n", + __func__, THREAD_SIZE, THREAD_SIZE); + } + return stack; } static void __init setup_lowcore(void) { struct lowcore *lc, *abs_lc; - unsigned long mcck_stack; /* * Setup lowcore for boot cpu @@ -441,8 +425,6 @@ static void __init setup_lowcore(void) lc->io_new_psw.mask = PSW_KERNEL_BITS | PSW_MASK_MCHECK; lc->io_new_psw.addr = (unsigned long) io_int_handler; lc->clock_comparator = clock_comparator_max; - lc->nodat_stack = ((unsigned long) &init_thread_union) - + THREAD_SIZE - STACK_FRAME_OVERHEAD - sizeof(struct pt_regs); lc->current_task = (unsigned long)&init_task; lc->lpp = LPP_MAGIC; lc->machine_flags = S390_lowcore.machine_flags; @@ -455,17 +437,15 @@ static void __init setup_lowcore(void) lc->steal_timer = S390_lowcore.steal_timer; lc->last_update_timer = S390_lowcore.last_update_timer; lc->last_update_clock = S390_lowcore.last_update_clock; - /* * Allocate the global restart stack which is the same for - * all CPUs in cast *one* of them does a PSW restart. + * all CPUs in case *one* of them does a PSW restart. */ - restart_stack = memblock_alloc(THREAD_SIZE, THREAD_SIZE); - if (!restart_stack) - panic("%s: Failed to allocate %lu bytes align=0x%lx\n", - __func__, THREAD_SIZE, THREAD_SIZE); - restart_stack += STACK_INIT_OFFSET; - + restart_stack = (void *)(stack_alloc_early() + STACK_INIT_OFFSET); + lc->mcck_stack = stack_alloc_early() + STACK_INIT_OFFSET; + lc->async_stack = stack_alloc_early() + STACK_INIT_OFFSET; + lc->nodat_stack = stack_alloc_early() + STACK_INIT_OFFSET; + lc->kernel_stack = S390_lowcore.kernel_stack; /* * Set up PSW restart to call ipl.c:do_restart(). Copy the relevant * restart data to the absolute zero lowcore. This is necessary if @@ -476,13 +456,6 @@ static void __init setup_lowcore(void) lc->restart_data = 0; lc->restart_source = -1U; __ctl_store(lc->cregs_save_area, 0, 15); - - mcck_stack = (unsigned long)memblock_alloc(THREAD_SIZE, THREAD_SIZE); - if (!mcck_stack) - panic("%s: Failed to allocate %lu bytes align=0x%lx\n", - __func__, THREAD_SIZE, THREAD_SIZE); - lc->mcck_stack = mcck_stack + STACK_INIT_OFFSET; - lc->spinlock_lockval = arch_spin_lockval(0); lc->spinlock_index = 0; arch_spin_lock_setup(0); @@ -635,7 +608,11 @@ static struct notifier_block kdump_mem_nb = { */ static void __init reserve_pgtables(void) { - memblock_reserve(pgalloc_pos, pgalloc_end - pgalloc_pos); + unsigned long start, end; + struct reserved_range *range; + + for_each_physmem_reserved_type_range(RR_VMEM, range, &start, &end) + memblock_reserve(start, end - start); } /* @@ -712,13 +689,13 @@ static void __init reserve_crashkernel(void) */ static void __init reserve_initrd(void) { -#ifdef CONFIG_BLK_DEV_INITRD - if (!initrd_data.start || !initrd_data.size) + unsigned long addr, size; + + if (!IS_ENABLED(CONFIG_BLK_DEV_INITRD) || !get_physmem_reserved(RR_INITRD, &addr, &size)) return; - initrd_start = (unsigned long)__va(initrd_data.start); - initrd_end = initrd_start + initrd_data.size; - memblock_reserve(initrd_data.start, initrd_data.size); -#endif + initrd_start = (unsigned long)__va(addr); + initrd_end = initrd_start + size; + memblock_reserve(addr, size); } /* @@ -730,72 +707,40 @@ static void __init reserve_certificate_list(void) memblock_reserve(ipl_cert_list_addr, ipl_cert_list_size); } -static void __init reserve_mem_detect_info(void) +static void __init reserve_physmem_info(void) { - unsigned long start, size; + unsigned long addr, size; - get_mem_detect_reserved(&start, &size); - if (size) - memblock_reserve(start, size); + if (get_physmem_reserved(RR_MEM_DETECT_EXTENDED, &addr, &size)) + memblock_reserve(addr, size); } -static void __init free_mem_detect_info(void) +static void __init free_physmem_info(void) { - unsigned long start, size; + unsigned long addr, size; - get_mem_detect_reserved(&start, &size); - if (size) - memblock_phys_free(start, size); + if (get_physmem_reserved(RR_MEM_DETECT_EXTENDED, &addr, &size)) + memblock_phys_free(addr, size); } -static const char * __init get_mem_info_source(void) -{ - switch (mem_detect.info_source) { - case MEM_DETECT_SCLP_STOR_INFO: - return "sclp storage info"; - case MEM_DETECT_DIAG260: - return "diag260"; - case MEM_DETECT_SCLP_READ_INFO: - return "sclp read info"; - case MEM_DETECT_BIN_SEARCH: - return "binary search"; - } - return "none"; -} - -static void __init memblock_add_mem_detect_info(void) +static void __init memblock_add_physmem_info(void) { unsigned long start, end; int i; pr_debug("physmem info source: %s (%hhd)\n", - get_mem_info_source(), mem_detect.info_source); + get_physmem_info_source(), physmem_info.info_source); /* keep memblock lists close to the kernel */ memblock_set_bottom_up(true); - for_each_mem_detect_usable_block(i, &start, &end) + for_each_physmem_usable_range(i, &start, &end) memblock_add(start, end - start); - for_each_mem_detect_block(i, &start, &end) + for_each_physmem_online_range(i, &start, &end) memblock_physmem_add(start, end - start); memblock_set_bottom_up(false); memblock_set_node(0, ULONG_MAX, &memblock.memory, 0); } /* - * Check for initrd being in usable memory - */ -static void __init check_initrd(void) -{ -#ifdef CONFIG_BLK_DEV_INITRD - if (initrd_data.start && initrd_data.size && - !memblock_is_region_memory(initrd_data.start, initrd_data.size)) { - pr_err("The initial RAM disk does not fit into the memory\n"); - memblock_phys_free(initrd_data.start, initrd_data.size); - initrd_start = initrd_end = 0; - } -#endif -} - -/* * Reserve memory used for lowcore/command line/kernel image. */ static void __init reserve_kernel(void) @@ -803,7 +748,7 @@ static void __init reserve_kernel(void) memblock_reserve(0, STARTUP_NORMAL_OFFSET); memblock_reserve(OLDMEM_BASE, sizeof(unsigned long)); memblock_reserve(OLDMEM_SIZE, sizeof(unsigned long)); - memblock_reserve(__amode31_base, __eamode31 - __samode31); + memblock_reserve(physmem_info.reserved[RR_AMODE31].start, __eamode31 - __samode31); memblock_reserve(__pa(sclp_early_sccb), EXT_SCCB_READ_SCP); memblock_reserve(__pa(_stext), _end - _stext); } @@ -825,13 +770,13 @@ static void __init setup_memory(void) static void __init relocate_amode31_section(void) { unsigned long amode31_size = __eamode31 - __samode31; - long amode31_offset = __amode31_base - __samode31; + long amode31_offset = physmem_info.reserved[RR_AMODE31].start - __samode31; long *ptr; pr_info("Relocating AMODE31 section of size 0x%08lx\n", amode31_size); /* Move original AMODE31 section to the new one */ - memmove((void *)__amode31_base, (void *)__samode31, amode31_size); + memmove((void *)physmem_info.reserved[RR_AMODE31].start, (void *)__samode31, amode31_size); /* Zero out the old AMODE31 section to catch invalid accesses within it */ memset((void *)__samode31, 0, amode31_size); @@ -997,14 +942,14 @@ void __init setup_arch(char **cmdline_p) reserve_kernel(); reserve_initrd(); reserve_certificate_list(); - reserve_mem_detect_info(); + reserve_physmem_info(); memblock_set_current_limit(ident_map_size); memblock_allow_resize(); /* Get information about *all* installed memory */ - memblock_add_mem_detect_info(); + memblock_add_physmem_info(); - free_mem_detect_info(); + free_physmem_info(); setup_memory_end(); memblock_dump_all(); setup_memory(); @@ -1017,7 +962,6 @@ void __init setup_arch(char **cmdline_p) if (MACHINE_HAS_EDAT2) hugetlb_cma_reserve(PUD_SHIFT - PAGE_SHIFT); - check_initrd(); reserve_crashkernel(); #ifdef CONFIG_CRASH_DUMP /* diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c index 70a84748f806..726de4f4df01 100644 --- a/arch/s390/kernel/smp.c +++ b/arch/s390/kernel/smp.c @@ -280,9 +280,8 @@ static void pcpu_attach_task(struct pcpu *pcpu, struct task_struct *tsk) cpu = pcpu - pcpu_devices; lc = lowcore_ptr[cpu]; - lc->kernel_stack = (unsigned long) task_stack_page(tsk) - + THREAD_SIZE - STACK_FRAME_OVERHEAD - sizeof(struct pt_regs); - lc->current_task = (unsigned long) tsk; + lc->kernel_stack = (unsigned long)task_stack_page(tsk) + STACK_INIT_OFFSET; + lc->current_task = (unsigned long)tsk; lc->lpp = LPP_MAGIC; lc->current_pid = tsk->pid; lc->user_timer = tsk->thread.user_timer; @@ -348,7 +347,6 @@ static void pcpu_delegate(struct pcpu *pcpu, abs_lc->restart_source = source_cpu; put_abs_lowcore(abs_lc); } - __bpon(); asm volatile( "0: sigp 0,%0,%2 # sigp restart to target cpu\n" " brc 2,0b # busy, try again\n" @@ -986,7 +984,6 @@ void __cpu_die(unsigned int cpu) void __noreturn cpu_die(void) { idle_task_exit(); - __bpon(); pcpu_sigp_retry(pcpu_devices + smp_processor_id(), SIGP_STOP, 0); for (;;) ; } @@ -1302,9 +1299,9 @@ int __init smp_reinit_ipl_cpu(void) local_mcck_enable(); local_irq_restore(flags); - free_pages(lc_ipl->async_stack - STACK_INIT_OFFSET, THREAD_SIZE_ORDER); memblock_free_late(__pa(lc_ipl->mcck_stack - STACK_INIT_OFFSET), THREAD_SIZE); + memblock_free_late(__pa(lc_ipl->async_stack - STACK_INIT_OFFSET), THREAD_SIZE); + memblock_free_late(__pa(lc_ipl->nodat_stack - STACK_INIT_OFFSET), THREAD_SIZE); memblock_free_late(__pa(lc_ipl), sizeof(*lc_ipl)); - return 0; } diff --git a/arch/s390/kernel/syscalls/syscall.tbl b/arch/s390/kernel/syscalls/syscall.tbl index 799147658dee..b68f47541169 100644 --- a/arch/s390/kernel/syscalls/syscall.tbl +++ b/arch/s390/kernel/syscalls/syscall.tbl @@ -449,7 +449,7 @@ 444 common landlock_create_ruleset sys_landlock_create_ruleset sys_landlock_create_ruleset 445 common landlock_add_rule sys_landlock_add_rule sys_landlock_add_rule 446 common landlock_restrict_self sys_landlock_restrict_self sys_landlock_restrict_self -# 447 reserved for memfd_secret +447 common memfd_secret sys_memfd_secret sys_memfd_secret 448 common process_mrelease sys_process_mrelease sys_process_mrelease 449 common futex_waitv sys_futex_waitv sys_futex_waitv 450 common set_mempolicy_home_node sys_set_mempolicy_home_node sys_set_mempolicy_home_node diff --git a/arch/s390/kernel/text_amode31.S b/arch/s390/kernel/text_amode31.S index e0f01ce251f5..14c6d25c035f 100644 --- a/arch/s390/kernel/text_amode31.S +++ b/arch/s390/kernel/text_amode31.S @@ -27,7 +27,7 @@ /* * int _diag14_amode31(unsigned long rx, unsigned long ry1, unsigned long subcode) */ -ENTRY(_diag14_amode31) +SYM_FUNC_START(_diag14_amode31) lgr %r1,%r2 lgr %r2,%r3 lgr %r3,%r4 @@ -42,12 +42,12 @@ ENTRY(_diag14_amode31) lgfr %r2,%r5 BR_EX_AMODE31_r14 EX_TABLE_AMODE31(.Ldiag14_ex, .Ldiag14_fault) -ENDPROC(_diag14_amode31) +SYM_FUNC_END(_diag14_amode31) /* * int _diag210_amode31(struct diag210 *addr) */ -ENTRY(_diag210_amode31) +SYM_FUNC_START(_diag210_amode31) lgr %r1,%r2 lhi %r2,-1 sam31 @@ -60,12 +60,12 @@ ENTRY(_diag210_amode31) lgfr %r2,%r2 BR_EX_AMODE31_r14 EX_TABLE_AMODE31(.Ldiag210_ex, .Ldiag210_fault) -ENDPROC(_diag210_amode31) +SYM_FUNC_END(_diag210_amode31) /* * int diag8c(struct diag8c *addr, struct ccw_dev_id *devno, size_t len) */ -ENTRY(_diag8c_amode31) +SYM_FUNC_START(_diag8c_amode31) llgf %r3,0(%r3) sam31 diag %r2,%r4,0x8c @@ -74,11 +74,11 @@ ENTRY(_diag8c_amode31) lgfr %r2,%r3 BR_EX_AMODE31_r14 EX_TABLE_AMODE31(.Ldiag8c_ex, .Ldiag8c_ex) -ENDPROC(_diag8c_amode31) +SYM_FUNC_END(_diag8c_amode31) /* * int _diag26c_amode31(void *req, void *resp, enum diag26c_sc subcode) */ -ENTRY(_diag26c_amode31) +SYM_FUNC_START(_diag26c_amode31) lghi %r5,-EOPNOTSUPP sam31 diag %r2,%r4,0x26c @@ -87,42 +87,42 @@ ENTRY(_diag26c_amode31) lgfr %r2,%r5 BR_EX_AMODE31_r14 EX_TABLE_AMODE31(.Ldiag26c_ex, .Ldiag26c_ex) -ENDPROC(_diag26c_amode31) +SYM_FUNC_END(_diag26c_amode31) /* * void _diag0c_amode31(struct hypfs_diag0c_entry *entry) */ -ENTRY(_diag0c_amode31) +SYM_FUNC_START(_diag0c_amode31) sam31 diag %r2,%r2,0x0c sam64 BR_EX_AMODE31_r14 -ENDPROC(_diag0c_amode31) +SYM_FUNC_END(_diag0c_amode31) /* * void _diag308_reset_amode31(void) * * Calls diag 308 subcode 1 and continues execution */ -ENTRY(_diag308_reset_amode31) - larl %r4,.Lctlregs # Save control registers +SYM_FUNC_START(_diag308_reset_amode31) + larl %r4,ctlregs # Save control registers stctg %c0,%c15,0(%r4) lg %r2,0(%r4) # Disable lowcore protection nilh %r2,0xefff - larl %r4,.Lctlreg0 + larl %r4,ctlreg0 stg %r2,0(%r4) lctlg %c0,%c0,0(%r4) - larl %r4,.Lfpctl # Floating point control register + larl %r4,fpctl # Floating point control register stfpc 0(%r4) - larl %r4,.Lprefix # Save prefix register + larl %r4,prefix # Save prefix register stpx 0(%r4) - larl %r4,.Lprefix_zero # Set prefix register to 0 + larl %r4,prefix_zero # Set prefix register to 0 spx 0(%r4) - larl %r4,.Lcontinue_psw # Save PSW flags + larl %r4,continue_psw # Save PSW flags epsw %r2,%r3 stm %r2,%r3,0(%r4) larl %r4,.Lrestart_part2 # Setup restart PSW at absolute 0 - larl %r3,.Lrestart_diag308_psw + larl %r3,restart_diag308_psw og %r4,0(%r3) # Save PSW lghi %r3,0 sturg %r4,%r3 # Use sturg, because of large pages @@ -134,39 +134,26 @@ ENTRY(_diag308_reset_amode31) lhi %r1,2 # Use mode 2 = ESAME (dump) sigp %r1,%r0,SIGP_SET_ARCHITECTURE # Switch to ESAME mode sam64 # Switch to 64 bit addressing mode - larl %r4,.Lctlregs # Restore control registers + larl %r4,ctlregs # Restore control registers lctlg %c0,%c15,0(%r4) - larl %r4,.Lfpctl # Restore floating point ctl register + larl %r4,fpctl # Restore floating point ctl register lfpc 0(%r4) - larl %r4,.Lprefix # Restore prefix register + larl %r4,prefix # Restore prefix register spx 0(%r4) - larl %r4,.Lcontinue_psw # Restore PSW flags + larl %r4,continue_psw # Restore PSW flags larl %r2,.Lcontinue stg %r2,8(%r4) lpswe 0(%r4) .Lcontinue: BR_EX_AMODE31_r14 -ENDPROC(_diag308_reset_amode31) +SYM_FUNC_END(_diag308_reset_amode31) .section .amode31.data,"aw",@progbits -.align 8 -.Lrestart_diag308_psw: - .long 0x00080000,0x80000000 - -.align 8 -.Lcontinue_psw: - .quad 0,0 - -.align 8 -.Lctlreg0: - .quad 0 -.Lctlregs: - .rept 16 - .quad 0 - .endr -.Lfpctl: - .long 0 -.Lprefix: - .long 0 -.Lprefix_zero: - .long 0 + .balign 8 +SYM_DATA_LOCAL(restart_diag308_psw, .long 0x00080000,0x80000000) +SYM_DATA_LOCAL(continue_psw, .quad 0,0) +SYM_DATA_LOCAL(ctlreg0, .quad 0) +SYM_DATA_LOCAL(ctlregs, .fill 16,8,0) +SYM_DATA_LOCAL(fpctl, .long 0) +SYM_DATA_LOCAL(prefix, .long 0) +SYM_DATA_LOCAL(prefix_zero, .long 0) diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c index 72af753d1bba..9fd19530c9a5 100644 --- a/arch/s390/kernel/topology.c +++ b/arch/s390/kernel/topology.c @@ -637,16 +637,6 @@ static struct ctl_table topology_ctl_table[] = { { }, }; -static struct ctl_table topology_dir_table[] = { - { - .procname = "s390", - .maxlen = 0, - .mode = 0555, - .child = topology_ctl_table, - }, - { }, -}; - static int __init topology_init(void) { struct device *dev_root; @@ -657,7 +647,7 @@ static int __init topology_init(void) set_topology_timer(); else topology_update_polarization_simple(); - register_sysctl_table(topology_dir_table); + register_sysctl("s390", topology_ctl_table); dev_root = bus_get_dev_root(&cpu_subsys); if (dev_root) { diff --git a/arch/s390/kernel/vdso32/vdso_user_wrapper.S b/arch/s390/kernel/vdso32/vdso_user_wrapper.S index 3f42f27f978c..2e645003fdaf 100644 --- a/arch/s390/kernel/vdso32/vdso_user_wrapper.S +++ b/arch/s390/kernel/vdso32/vdso_user_wrapper.S @@ -1,12 +1,13 @@ /* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/linkage.h> #include <asm/unistd.h> #include <asm/dwarf.h> .macro vdso_syscall func,syscall .globl __kernel_compat_\func .type __kernel_compat_\func,@function - .align 8 + __ALIGN __kernel_compat_\func: CFI_STARTPROC svc \syscall diff --git a/arch/s390/kernel/vdso64/vdso_user_wrapper.S b/arch/s390/kernel/vdso64/vdso_user_wrapper.S index 97f0c0a669a5..57f62596e53b 100644 --- a/arch/s390/kernel/vdso64/vdso_user_wrapper.S +++ b/arch/s390/kernel/vdso64/vdso_user_wrapper.S @@ -1,4 +1,5 @@ /* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/linkage.h> #include <asm/vdso.h> #include <asm/unistd.h> #include <asm/asm-offsets.h> @@ -16,7 +17,7 @@ .macro vdso_func func .globl __kernel_\func .type __kernel_\func,@function - .align 8 + __ALIGN __kernel_\func: CFI_STARTPROC aghi %r15,-WRAPPER_FRAME_SIZE @@ -41,7 +42,7 @@ vdso_func getcpu .macro vdso_syscall func,syscall .globl __kernel_\func .type __kernel_\func,@function - .align 8 + __ALIGN __kernel_\func: CFI_STARTPROC svc \syscall diff --git a/arch/s390/kernel/vmlinux.lds.S b/arch/s390/kernel/vmlinux.lds.S index b653ba8d51e6..2ae201ebf90b 100644 --- a/arch/s390/kernel/vmlinux.lds.S +++ b/arch/s390/kernel/vmlinux.lds.S @@ -14,6 +14,8 @@ #define BSS_FIRST_SECTIONS *(.bss..swapper_pg_dir) \ *(.bss..invalid_pg_dir) +#define RO_EXCEPTION_TABLE_ALIGN 16 + /* Handle ro_after_init data on our own. */ #define RO_AFTER_INIT_DATA @@ -66,7 +68,6 @@ SECTIONS *(.data..ro_after_init) JUMP_TABLE_DATA } :data - EXCEPTION_TABLE(16) . = ALIGN(PAGE_SIZE); __end_ro_after_init = .; @@ -219,6 +220,13 @@ SECTIONS QUAD(init_mm) QUAD(swapper_pg_dir) QUAD(invalid_pg_dir) +#ifdef CONFIG_KASAN + QUAD(kasan_early_shadow_page) + QUAD(kasan_early_shadow_pte) + QUAD(kasan_early_shadow_pmd) + QUAD(kasan_early_shadow_pud) + QUAD(kasan_early_shadow_p4d) +#endif } :NONE /* Debugging sections. */ diff --git a/arch/s390/lib/mem.S b/arch/s390/lib/mem.S index dc0874f2e203..5a9a55de2e10 100644 --- a/arch/s390/lib/mem.S +++ b/arch/s390/lib/mem.S @@ -14,8 +14,7 @@ /* * void *memmove(void *dest, const void *src, size_t n) */ -WEAK(memmove) -ENTRY(__memmove) +SYM_FUNC_START(__memmove) ltgr %r4,%r4 lgr %r1,%r2 jz .Lmemmove_exit @@ -48,7 +47,10 @@ ENTRY(__memmove) BR_EX %r14 .Lmemmove_mvc: mvc 0(1,%r1),0(%r3) -ENDPROC(__memmove) +SYM_FUNC_END(__memmove) +EXPORT_SYMBOL(__memmove) + +SYM_FUNC_ALIAS(memmove, __memmove) EXPORT_SYMBOL(memmove) /* @@ -66,8 +68,7 @@ EXPORT_SYMBOL(memmove) * return __builtin_memset(s, c, n); * } */ -WEAK(memset) -ENTRY(__memset) +SYM_FUNC_START(__memset) ltgr %r4,%r4 jz .Lmemset_exit ltgr %r3,%r3 @@ -111,7 +112,10 @@ ENTRY(__memset) xc 0(1,%r1),0(%r1) .Lmemset_mvc: mvc 1(1,%r1),0(%r1) -ENDPROC(__memset) +SYM_FUNC_END(__memset) +EXPORT_SYMBOL(__memset) + +SYM_FUNC_ALIAS(memset, __memset) EXPORT_SYMBOL(memset) /* @@ -119,8 +123,7 @@ EXPORT_SYMBOL(memset) * * void *memcpy(void *dest, const void *src, size_t n) */ -WEAK(memcpy) -ENTRY(__memcpy) +SYM_FUNC_START(__memcpy) ltgr %r4,%r4 jz .Lmemcpy_exit aghi %r4,-1 @@ -141,7 +144,10 @@ ENTRY(__memcpy) j .Lmemcpy_remainder .Lmemcpy_mvc: mvc 0(1,%r1),0(%r3) -ENDPROC(__memcpy) +SYM_FUNC_END(__memcpy) +EXPORT_SYMBOL(__memcpy) + +SYM_FUNC_ALIAS(memcpy, __memcpy) EXPORT_SYMBOL(memcpy) /* @@ -152,7 +158,7 @@ EXPORT_SYMBOL(memcpy) * void *__memset64(uint64_t *s, uint64_t v, size_t count) */ .macro __MEMSET bits,bytes,insn -ENTRY(__memset\bits) +SYM_FUNC_START(__memset\bits) ltgr %r4,%r4 jz .L__memset_exit\bits cghi %r4,\bytes @@ -178,7 +184,7 @@ ENTRY(__memset\bits) BR_EX %r14 .L__memset_mvc\bits: mvc \bytes(1,%r1),0(%r1) -ENDPROC(__memset\bits) +SYM_FUNC_END(__memset\bits) .endm __MEMSET 16,2,sth diff --git a/arch/s390/lib/uaccess.c b/arch/s390/lib/uaccess.c index d44214072779..e4a13d7cab6e 100644 --- a/arch/s390/lib/uaccess.c +++ b/arch/s390/lib/uaccess.c @@ -27,14 +27,13 @@ void debug_user_asce(int exit) "kernel: %016llx user: %016llx\n", exit ? "exit" : "entry", cr1, cr7, S390_lowcore.kernel_asce, S390_lowcore.user_asce); - } #endif /*CONFIG_DEBUG_ENTRY */ static unsigned long raw_copy_from_user_key(void *to, const void __user *from, unsigned long size, unsigned long key) { - unsigned long tmp1, tmp2; + unsigned long rem; union oac spec = { .oac2.key = key, .oac2.as = PSW_BITS_AS_SECONDARY, @@ -42,28 +41,30 @@ static unsigned long raw_copy_from_user_key(void *to, const void __user *from, .oac2.a = 1, }; - tmp1 = -4096UL; asm volatile( - " lr 0,%[spec]\n" - "0: mvcos 0(%2),0(%1),%0\n" - "6: jz 4f\n" - "1: algr %0,%3\n" - " slgr %1,%3\n" - " slgr %2,%3\n" - " j 0b\n" - "2: la %4,4095(%1)\n"/* %4 = ptr + 4095 */ - " nr %4,%3\n" /* %4 = (ptr + 4095) & -4096 */ - " slgr %4,%1\n" - " clgr %0,%4\n" /* copy crosses next page boundary? */ - " jnh 5f\n" - "3: mvcos 0(%2),0(%1),%4\n" - "7: slgr %0,%4\n" - " j 5f\n" - "4: slgr %0,%0\n" - "5:\n" - EX_TABLE(0b,2b) EX_TABLE(3b,5b) EX_TABLE(6b,2b) EX_TABLE(7b,5b) - : "+a" (size), "+a" (from), "+a" (to), "+a" (tmp1), "=a" (tmp2) - : [spec] "d" (spec.val) + " lr 0,%[spec]\n" + "0: mvcos 0(%[to]),0(%[from]),%[size]\n" + "1: jz 5f\n" + " algr %[size],%[val]\n" + " slgr %[from],%[val]\n" + " slgr %[to],%[val]\n" + " j 0b\n" + "2: la %[rem],4095(%[from])\n" /* rem = from + 4095 */ + " nr %[rem],%[val]\n" /* rem = (from + 4095) & -4096 */ + " slgr %[rem],%[from]\n" + " clgr %[size],%[rem]\n" /* copy crosses next page boundary? */ + " jnh 6f\n" + "3: mvcos 0(%[to]),0(%[from]),%[rem]\n" + "4: slgr %[size],%[rem]\n" + " j 6f\n" + "5: slgr %[size],%[size]\n" + "6:\n" + EX_TABLE(0b, 2b) + EX_TABLE(1b, 2b) + EX_TABLE(3b, 6b) + EX_TABLE(4b, 6b) + : [size] "+&a" (size), [from] "+&a" (from), [to] "+&a" (to), [rem] "=&a" (rem) + : [val] "a" (-4096UL), [spec] "d" (spec.val) : "cc", "memory", "0"); return size; } @@ -94,7 +95,7 @@ EXPORT_SYMBOL(_copy_from_user_key); static unsigned long raw_copy_to_user_key(void __user *to, const void *from, unsigned long size, unsigned long key) { - unsigned long tmp1, tmp2; + unsigned long rem; union oac spec = { .oac1.key = key, .oac1.as = PSW_BITS_AS_SECONDARY, @@ -102,28 +103,30 @@ static unsigned long raw_copy_to_user_key(void __user *to, const void *from, .oac1.a = 1, }; - tmp1 = -4096UL; asm volatile( - " lr 0,%[spec]\n" - "0: mvcos 0(%1),0(%2),%0\n" - "6: jz 4f\n" - "1: algr %0,%3\n" - " slgr %1,%3\n" - " slgr %2,%3\n" - " j 0b\n" - "2: la %4,4095(%1)\n"/* %4 = ptr + 4095 */ - " nr %4,%3\n" /* %4 = (ptr + 4095) & -4096 */ - " slgr %4,%1\n" - " clgr %0,%4\n" /* copy crosses next page boundary? */ - " jnh 5f\n" - "3: mvcos 0(%1),0(%2),%4\n" - "7: slgr %0,%4\n" - " j 5f\n" - "4: slgr %0,%0\n" - "5:\n" - EX_TABLE(0b,2b) EX_TABLE(3b,5b) EX_TABLE(6b,2b) EX_TABLE(7b,5b) - : "+a" (size), "+a" (to), "+a" (from), "+a" (tmp1), "=a" (tmp2) - : [spec] "d" (spec.val) + " lr 0,%[spec]\n" + "0: mvcos 0(%[to]),0(%[from]),%[size]\n" + "1: jz 5f\n" + " algr %[size],%[val]\n" + " slgr %[to],%[val]\n" + " slgr %[from],%[val]\n" + " j 0b\n" + "2: la %[rem],4095(%[to])\n" /* rem = to + 4095 */ + " nr %[rem],%[val]\n" /* rem = (to + 4095) & -4096 */ + " slgr %[rem],%[to]\n" + " clgr %[size],%[rem]\n" /* copy crosses next page boundary? */ + " jnh 6f\n" + "3: mvcos 0(%[to]),0(%[from]),%[rem]\n" + "4: slgr %[size],%[rem]\n" + " j 6f\n" + "5: slgr %[size],%[size]\n" + "6:\n" + EX_TABLE(0b, 2b) + EX_TABLE(1b, 2b) + EX_TABLE(3b, 6b) + EX_TABLE(4b, 6b) + : [size] "+&a" (size), [to] "+&a" (to), [from] "+&a" (from), [rem] "=&a" (rem) + : [val] "a" (-4096UL), [spec] "d" (spec.val) : "cc", "memory", "0"); return size; } @@ -147,33 +150,35 @@ EXPORT_SYMBOL(_copy_to_user_key); unsigned long __clear_user(void __user *to, unsigned long size) { - unsigned long tmp1, tmp2; + unsigned long rem; union oac spec = { .oac1.as = PSW_BITS_AS_SECONDARY, .oac1.a = 1, }; - tmp1 = -4096UL; asm volatile( - " lr 0,%[spec]\n" - "0: mvcos 0(%1),0(%4),%0\n" - "6: jz 4f\n" - "1: algr %0,%2\n" - " slgr %1,%2\n" - " j 0b\n" - "2: la %3,4095(%1)\n"/* %4 = to + 4095 */ - " nr %3,%2\n" /* %4 = (to + 4095) & -4096 */ - " slgr %3,%1\n" - " clgr %0,%3\n" /* copy crosses next page boundary? */ - " jnh 5f\n" - "3: mvcos 0(%1),0(%4),%3\n" - "7: slgr %0,%3\n" - " j 5f\n" - "4: slgr %0,%0\n" - "5:\n" - EX_TABLE(0b,2b) EX_TABLE(6b,2b) EX_TABLE(3b,5b) EX_TABLE(7b,5b) - : "+&a" (size), "+&a" (to), "+a" (tmp1), "=&a" (tmp2) - : "a" (empty_zero_page), [spec] "d" (spec.val) + " lr 0,%[spec]\n" + "0: mvcos 0(%[to]),0(%[zeropg]),%[size]\n" + "1: jz 5f\n" + " algr %[size],%[val]\n" + " slgr %[to],%[val]\n" + " j 0b\n" + "2: la %[rem],4095(%[to])\n" /* rem = to + 4095 */ + " nr %[rem],%[val]\n" /* rem = (to + 4095) & -4096 */ + " slgr %[rem],%[to]\n" + " clgr %[size],%[rem]\n" /* copy crosses next page boundary? */ + " jnh 6f\n" + "3: mvcos 0(%[to]),0(%[zeropg]),%[rem]\n" + "4: slgr %[size],%[rem]\n" + " j 6f\n" + "5: slgr %[size],%[size]\n" + "6:\n" + EX_TABLE(0b, 2b) + EX_TABLE(1b, 2b) + EX_TABLE(3b, 6b) + EX_TABLE(4b, 6b) + : [size] "+&a" (size), [to] "+&a" (to), [rem] "=&a" (rem) + : [val] "a" (-4096UL), [zeropg] "a" (empty_zero_page), [spec] "d" (spec.val) : "cc", "memory", "0"); return size; } diff --git a/arch/s390/mm/Makefile b/arch/s390/mm/Makefile index 57e4f3a24829..d90db06a8af5 100644 --- a/arch/s390/mm/Makefile +++ b/arch/s390/mm/Makefile @@ -10,6 +10,3 @@ obj-$(CONFIG_CMM) += cmm.o obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o obj-$(CONFIG_PTDUMP_CORE) += dump_pagetables.o obj-$(CONFIG_PGSTE) += gmap.o - -KASAN_SANITIZE_kasan_init.o := n -obj-$(CONFIG_KASAN) += kasan_init.o diff --git a/arch/s390/mm/cmm.c b/arch/s390/mm/cmm.c index 9141ed4c52e9..5300c6867d5e 100644 --- a/arch/s390/mm/cmm.c +++ b/arch/s390/mm/cmm.c @@ -335,16 +335,6 @@ static struct ctl_table cmm_table[] = { { } }; -static struct ctl_table cmm_dir_table[] = { - { - .procname = "vm", - .maxlen = 0, - .mode = 0555, - .child = cmm_table, - }, - { } -}; - #ifdef CONFIG_CMM_IUCV #define SMSG_PREFIX "CMM" static void cmm_smsg_target(const char *from, char *msg) @@ -389,7 +379,7 @@ static int __init cmm_init(void) { int rc = -ENOMEM; - cmm_sysctl_header = register_sysctl_table(cmm_dir_table); + cmm_sysctl_header = register_sysctl("vm", cmm_table); if (!cmm_sysctl_header) goto out_sysctl; #ifdef CONFIG_CMM_IUCV diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c index 144447d5cb4c..8d94e29adcdb 100644 --- a/arch/s390/mm/init.c +++ b/arch/s390/mm/init.c @@ -176,9 +176,8 @@ void __init mem_init(void) void free_initmem(void) { - __set_memory((unsigned long)_sinittext, - (unsigned long)(_einittext - _sinittext) >> PAGE_SHIFT, - SET_MEMORY_RW | SET_MEMORY_NX); + set_memory_rwnx((unsigned long)_sinittext, + (unsigned long)(_einittext - _sinittext) >> PAGE_SHIFT); free_initmem_default(POISON_FREE_INITMEM); } diff --git a/arch/s390/mm/kasan_init.c b/arch/s390/mm/kasan_init.c deleted file mode 100644 index ef89a5f26853..000000000000 --- a/arch/s390/mm/kasan_init.c +++ /dev/null @@ -1,301 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -#include <linux/kasan.h> -#include <linux/sched/task.h> -#include <linux/pgtable.h> -#include <asm/pgalloc.h> -#include <asm/kasan.h> -#include <asm/mem_detect.h> -#include <asm/processor.h> -#include <asm/sclp.h> -#include <asm/facility.h> -#include <asm/sections.h> -#include <asm/setup.h> -#include <asm/uv.h> - -static unsigned long segment_pos __initdata; -static unsigned long segment_low __initdata; -static bool has_edat __initdata; -static bool has_nx __initdata; - -#define __sha(x) ((unsigned long)kasan_mem_to_shadow((void *)x)) - -static void __init kasan_early_panic(const char *reason) -{ - sclp_early_printk("The Linux kernel failed to boot with the KernelAddressSanitizer:\n"); - sclp_early_printk(reason); - disabled_wait(); -} - -static void * __init kasan_early_alloc_segment(void) -{ - segment_pos -= _SEGMENT_SIZE; - - if (segment_pos < segment_low) - kasan_early_panic("out of memory during initialisation\n"); - - return __va(segment_pos); -} - -static void * __init kasan_early_alloc_pages(unsigned int order) -{ - pgalloc_pos -= (PAGE_SIZE << order); - - if (pgalloc_pos < pgalloc_low) - kasan_early_panic("out of memory during initialisation\n"); - - return __va(pgalloc_pos); -} - -static void * __init kasan_early_crst_alloc(unsigned long val) -{ - unsigned long *table; - - table = kasan_early_alloc_pages(CRST_ALLOC_ORDER); - if (table) - crst_table_init(table, val); - return table; -} - -static pte_t * __init kasan_early_pte_alloc(void) -{ - static void *pte_leftover; - pte_t *pte; - - BUILD_BUG_ON(_PAGE_TABLE_SIZE * 2 != PAGE_SIZE); - - if (!pte_leftover) { - pte_leftover = kasan_early_alloc_pages(0); - pte = pte_leftover + _PAGE_TABLE_SIZE; - } else { - pte = pte_leftover; - pte_leftover = NULL; - } - memset64((u64 *)pte, _PAGE_INVALID, PTRS_PER_PTE); - return pte; -} - -enum populate_mode { - POPULATE_MAP, - POPULATE_ZERO_SHADOW, - POPULATE_SHALLOW -}; - -static inline pgprot_t pgprot_clear_bit(pgprot_t pgprot, unsigned long bit) -{ - return __pgprot(pgprot_val(pgprot) & ~bit); -} - -static void __init kasan_early_pgtable_populate(unsigned long address, - unsigned long end, - enum populate_mode mode) -{ - pgprot_t pgt_prot_zero = PAGE_KERNEL_RO; - pgprot_t pgt_prot = PAGE_KERNEL; - pgprot_t sgt_prot = SEGMENT_KERNEL; - pgd_t *pg_dir; - p4d_t *p4_dir; - pud_t *pu_dir; - pmd_t *pm_dir; - pte_t *pt_dir; - pmd_t pmd; - pte_t pte; - - if (!has_nx) { - pgt_prot_zero = pgprot_clear_bit(pgt_prot_zero, _PAGE_NOEXEC); - pgt_prot = pgprot_clear_bit(pgt_prot, _PAGE_NOEXEC); - sgt_prot = pgprot_clear_bit(sgt_prot, _SEGMENT_ENTRY_NOEXEC); - } - - while (address < end) { - pg_dir = pgd_offset_k(address); - if (pgd_none(*pg_dir)) { - if (mode == POPULATE_ZERO_SHADOW && - IS_ALIGNED(address, PGDIR_SIZE) && - end - address >= PGDIR_SIZE) { - pgd_populate(&init_mm, pg_dir, - kasan_early_shadow_p4d); - address = (address + PGDIR_SIZE) & PGDIR_MASK; - continue; - } - p4_dir = kasan_early_crst_alloc(_REGION2_ENTRY_EMPTY); - pgd_populate(&init_mm, pg_dir, p4_dir); - } - - if (mode == POPULATE_SHALLOW) { - address = (address + P4D_SIZE) & P4D_MASK; - continue; - } - - p4_dir = p4d_offset(pg_dir, address); - if (p4d_none(*p4_dir)) { - if (mode == POPULATE_ZERO_SHADOW && - IS_ALIGNED(address, P4D_SIZE) && - end - address >= P4D_SIZE) { - p4d_populate(&init_mm, p4_dir, - kasan_early_shadow_pud); - address = (address + P4D_SIZE) & P4D_MASK; - continue; - } - pu_dir = kasan_early_crst_alloc(_REGION3_ENTRY_EMPTY); - p4d_populate(&init_mm, p4_dir, pu_dir); - } - - pu_dir = pud_offset(p4_dir, address); - if (pud_none(*pu_dir)) { - if (mode == POPULATE_ZERO_SHADOW && - IS_ALIGNED(address, PUD_SIZE) && - end - address >= PUD_SIZE) { - pud_populate(&init_mm, pu_dir, - kasan_early_shadow_pmd); - address = (address + PUD_SIZE) & PUD_MASK; - continue; - } - pm_dir = kasan_early_crst_alloc(_SEGMENT_ENTRY_EMPTY); - pud_populate(&init_mm, pu_dir, pm_dir); - } - - pm_dir = pmd_offset(pu_dir, address); - if (pmd_none(*pm_dir)) { - if (IS_ALIGNED(address, PMD_SIZE) && - end - address >= PMD_SIZE) { - if (mode == POPULATE_ZERO_SHADOW) { - pmd_populate(&init_mm, pm_dir, kasan_early_shadow_pte); - address = (address + PMD_SIZE) & PMD_MASK; - continue; - } else if (has_edat) { - void *page = kasan_early_alloc_segment(); - - memset(page, 0, _SEGMENT_SIZE); - pmd = __pmd(__pa(page)); - pmd = set_pmd_bit(pmd, sgt_prot); - set_pmd(pm_dir, pmd); - address = (address + PMD_SIZE) & PMD_MASK; - continue; - } - } - pt_dir = kasan_early_pte_alloc(); - pmd_populate(&init_mm, pm_dir, pt_dir); - } else if (pmd_large(*pm_dir)) { - address = (address + PMD_SIZE) & PMD_MASK; - continue; - } - - pt_dir = pte_offset_kernel(pm_dir, address); - if (pte_none(*pt_dir)) { - void *page; - - switch (mode) { - case POPULATE_MAP: - page = kasan_early_alloc_pages(0); - memset(page, 0, PAGE_SIZE); - pte = __pte(__pa(page)); - pte = set_pte_bit(pte, pgt_prot); - set_pte(pt_dir, pte); - break; - case POPULATE_ZERO_SHADOW: - page = kasan_early_shadow_page; - pte = __pte(__pa(page)); - pte = set_pte_bit(pte, pgt_prot_zero); - set_pte(pt_dir, pte); - break; - case POPULATE_SHALLOW: - /* should never happen */ - break; - } - } - address += PAGE_SIZE; - } -} - -static void __init kasan_early_detect_facilities(void) -{ - if (test_facility(8)) { - has_edat = true; - __ctl_set_bit(0, 23); - } - if (!noexec_disabled && test_facility(130)) { - has_nx = true; - __ctl_set_bit(0, 20); - } -} - -void __init kasan_early_init(void) -{ - pte_t pte_z = __pte(__pa(kasan_early_shadow_page) | pgprot_val(PAGE_KERNEL_RO)); - pmd_t pmd_z = __pmd(__pa(kasan_early_shadow_pte) | _SEGMENT_ENTRY); - pud_t pud_z = __pud(__pa(kasan_early_shadow_pmd) | _REGION3_ENTRY); - p4d_t p4d_z = __p4d(__pa(kasan_early_shadow_pud) | _REGION2_ENTRY); - unsigned long untracked_end = MODULES_VADDR; - unsigned long shadow_alloc_size; - unsigned long start, end; - int i; - - kasan_early_detect_facilities(); - if (!has_nx) - pte_z = clear_pte_bit(pte_z, __pgprot(_PAGE_NOEXEC)); - - BUILD_BUG_ON(!IS_ALIGNED(KASAN_SHADOW_START, P4D_SIZE)); - BUILD_BUG_ON(!IS_ALIGNED(KASAN_SHADOW_END, P4D_SIZE)); - - /* init kasan zero shadow */ - crst_table_init((unsigned long *)kasan_early_shadow_p4d, p4d_val(p4d_z)); - crst_table_init((unsigned long *)kasan_early_shadow_pud, pud_val(pud_z)); - crst_table_init((unsigned long *)kasan_early_shadow_pmd, pmd_val(pmd_z)); - memset64((u64 *)kasan_early_shadow_pte, pte_val(pte_z), PTRS_PER_PTE); - - if (has_edat) { - shadow_alloc_size = get_mem_detect_usable_total() >> KASAN_SHADOW_SCALE_SHIFT; - segment_pos = round_down(pgalloc_pos, _SEGMENT_SIZE); - segment_low = segment_pos - shadow_alloc_size; - segment_low = round_down(segment_low, _SEGMENT_SIZE); - pgalloc_pos = segment_low; - } - /* - * Current memory layout: - * +- 0 -------------+ +- shadow start -+ - * |1:1 ident mapping| /|1/8 of ident map| - * | | / | | - * +-end of ident map+ / +----------------+ - * | ... gap ... | / | kasan | - * | | / | zero page | - * +- vmalloc area -+ / | mapping | - * | vmalloc_size | / | (untracked) | - * +- modules vaddr -+ / +----------------+ - * | 2Gb |/ | unmapped | allocated per module - * +- shadow start -+ +----------------+ - * | 1/8 addr space | | zero pg mapping| (untracked) - * +- shadow end ----+---------+- shadow end ---+ - * - * Current memory layout (KASAN_VMALLOC): - * +- 0 -------------+ +- shadow start -+ - * |1:1 ident mapping| /|1/8 of ident map| - * | | / | | - * +-end of ident map+ / +----------------+ - * | ... gap ... | / | kasan zero page| (untracked) - * | | / | mapping | - * +- vmalloc area -+ / +----------------+ - * | vmalloc_size | / |shallow populate| - * +- modules vaddr -+ / +----------------+ - * | 2Gb |/ |shallow populate| - * +- shadow start -+ +----------------+ - * | 1/8 addr space | | zero pg mapping| (untracked) - * +- shadow end ----+---------+- shadow end ---+ - */ - /* populate kasan shadow (for identity mapping and zero page mapping) */ - for_each_mem_detect_usable_block(i, &start, &end) - kasan_early_pgtable_populate(__sha(start), __sha(end), POPULATE_MAP); - if (IS_ENABLED(CONFIG_KASAN_VMALLOC)) { - untracked_end = VMALLOC_START; - /* shallowly populate kasan shadow for vmalloc and modules */ - kasan_early_pgtable_populate(__sha(VMALLOC_START), __sha(MODULES_END), - POPULATE_SHALLOW); - } - /* populate kasan shadow for untracked memory */ - kasan_early_pgtable_populate(__sha(ident_map_size), __sha(untracked_end), - POPULATE_ZERO_SHADOW); - kasan_early_pgtable_populate(__sha(MODULES_END), __sha(_REGION1_SIZE), - POPULATE_ZERO_SHADOW); - /* enable kasan */ - init_task.kasan_depth = 0; - sclp_early_printk("KernelAddressSanitizer initialized\n"); -} diff --git a/arch/s390/mm/pageattr.c b/arch/s390/mm/pageattr.c index 85195c18b2e8..5ba3bd8a7b12 100644 --- a/arch/s390/mm/pageattr.c +++ b/arch/s390/mm/pageattr.c @@ -4,6 +4,7 @@ * Author(s): Jan Glauber <jang@linux.vnet.ibm.com> */ #include <linux/hugetlb.h> +#include <linux/vmalloc.h> #include <linux/mm.h> #include <asm/cacheflush.h> #include <asm/facility.h> @@ -41,7 +42,7 @@ void __storage_key_init_range(unsigned long start, unsigned long end) } #ifdef CONFIG_PROC_FS -atomic_long_t direct_pages_count[PG_DIRECT_MAP_MAX]; +atomic_long_t __bootdata_preserved(direct_pages_count[PG_DIRECT_MAP_MAX]); void arch_report_meminfo(struct seq_file *m) { @@ -101,6 +102,14 @@ static int walk_pte_level(pmd_t *pmdp, unsigned long addr, unsigned long end, new = set_pte_bit(new, __pgprot(_PAGE_NOEXEC)); else if (flags & SET_MEMORY_X) new = clear_pte_bit(new, __pgprot(_PAGE_NOEXEC)); + if (flags & SET_MEMORY_INV) { + new = set_pte_bit(new, __pgprot(_PAGE_INVALID)); + } else if (flags & SET_MEMORY_DEF) { + new = __pte(pte_val(new) & PAGE_MASK); + new = set_pte_bit(new, PAGE_KERNEL); + if (!MACHINE_HAS_NX) + new = clear_pte_bit(new, __pgprot(_PAGE_NOEXEC)); + } pgt_set((unsigned long *)ptep, pte_val(new), addr, CRDTE_DTT_PAGE); ptep++; addr += PAGE_SIZE; @@ -151,6 +160,14 @@ static void modify_pmd_page(pmd_t *pmdp, unsigned long addr, new = set_pmd_bit(new, __pgprot(_SEGMENT_ENTRY_NOEXEC)); else if (flags & SET_MEMORY_X) new = clear_pmd_bit(new, __pgprot(_SEGMENT_ENTRY_NOEXEC)); + if (flags & SET_MEMORY_INV) { + new = set_pmd_bit(new, __pgprot(_SEGMENT_ENTRY_INVALID)); + } else if (flags & SET_MEMORY_DEF) { + new = __pmd(pmd_val(new) & PMD_MASK); + new = set_pmd_bit(new, SEGMENT_KERNEL); + if (!MACHINE_HAS_NX) + new = clear_pmd_bit(new, __pgprot(_SEGMENT_ENTRY_NOEXEC)); + } pgt_set((unsigned long *)pmdp, pmd_val(new), addr, CRDTE_DTT_SEGMENT); } @@ -232,6 +249,14 @@ static void modify_pud_page(pud_t *pudp, unsigned long addr, new = set_pud_bit(new, __pgprot(_REGION_ENTRY_NOEXEC)); else if (flags & SET_MEMORY_X) new = clear_pud_bit(new, __pgprot(_REGION_ENTRY_NOEXEC)); + if (flags & SET_MEMORY_INV) { + new = set_pud_bit(new, __pgprot(_REGION_ENTRY_INVALID)); + } else if (flags & SET_MEMORY_DEF) { + new = __pud(pud_val(new) & PUD_MASK); + new = set_pud_bit(new, REGION3_KERNEL); + if (!MACHINE_HAS_NX) + new = clear_pud_bit(new, __pgprot(_REGION_ENTRY_NOEXEC)); + } pgt_set((unsigned long *)pudp, pud_val(new), addr, CRDTE_DTT_REGION3); } @@ -298,11 +323,6 @@ static int change_page_attr(unsigned long addr, unsigned long end, int rc = -EINVAL; pgd_t *pgdp; - if (addr == end) - return 0; - if (end >= MODULES_END) - return -EINVAL; - mutex_lock(&cpa_mutex); pgdp = pgd_offset_k(addr); do { if (pgd_none(*pgdp)) @@ -313,18 +333,76 @@ static int change_page_attr(unsigned long addr, unsigned long end, break; cond_resched(); } while (pgdp++, addr = next, addr < end && !rc); - mutex_unlock(&cpa_mutex); + return rc; +} + +static int change_page_attr_alias(unsigned long addr, unsigned long end, + unsigned long flags) +{ + unsigned long alias, offset, va_start, va_end; + struct vm_struct *area; + int rc = 0; + + /* + * Changes to read-only permissions on kernel VA mappings are also + * applied to the kernel direct mapping. Execute permissions are + * intentionally not transferred to keep all allocated pages within + * the direct mapping non-executable. + */ + flags &= SET_MEMORY_RO | SET_MEMORY_RW; + if (!flags) + return 0; + area = NULL; + while (addr < end) { + if (!area) + area = find_vm_area((void *)addr); + if (!area || !(area->flags & VM_ALLOC)) + return 0; + va_start = (unsigned long)area->addr; + va_end = va_start + area->nr_pages * PAGE_SIZE; + offset = (addr - va_start) >> PAGE_SHIFT; + alias = (unsigned long)page_address(area->pages[offset]); + rc = change_page_attr(alias, alias + PAGE_SIZE, flags); + if (rc) + break; + addr += PAGE_SIZE; + if (addr >= va_end) + area = NULL; + } return rc; } int __set_memory(unsigned long addr, int numpages, unsigned long flags) { + unsigned long end; + int rc; + if (!MACHINE_HAS_NX) flags &= ~(SET_MEMORY_NX | SET_MEMORY_X); if (!flags) return 0; + if (!numpages) + return 0; addr &= PAGE_MASK; - return change_page_attr(addr, addr + numpages * PAGE_SIZE, flags); + end = addr + numpages * PAGE_SIZE; + mutex_lock(&cpa_mutex); + rc = change_page_attr(addr, end, flags); + if (rc) + goto out; + rc = change_page_attr_alias(addr, end, flags); +out: + mutex_unlock(&cpa_mutex); + return rc; +} + +int set_direct_map_invalid_noflush(struct page *page) +{ + return __set_memory((unsigned long)page_to_virt(page), 1, SET_MEMORY_INV); +} + +int set_direct_map_default_noflush(struct page *page) +{ + return __set_memory((unsigned long)page_to_virt(page), 1, SET_MEMORY_DEF); } #if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KFENCE) diff --git a/arch/s390/mm/pgalloc.c b/arch/s390/mm/pgalloc.c index 2de48b2c1b04..66ab68db9842 100644 --- a/arch/s390/mm/pgalloc.c +++ b/arch/s390/mm/pgalloc.c @@ -33,19 +33,9 @@ static struct ctl_table page_table_sysctl[] = { { } }; -static struct ctl_table page_table_sysctl_dir[] = { - { - .procname = "vm", - .maxlen = 0, - .mode = 0555, - .child = page_table_sysctl, - }, - { } -}; - static int __init page_table_register_sysctl(void) { - return register_sysctl_table(page_table_sysctl_dir) ? 0 : -ENOMEM; + return register_sysctl("vm", page_table_sysctl) ? 0 : -ENOMEM; } __initcall(page_table_register_sysctl); @@ -143,13 +133,7 @@ err_p4d: static inline unsigned int atomic_xor_bits(atomic_t *v, unsigned int bits) { - unsigned int old, new; - - do { - old = atomic_read(v); - new = old ^ bits; - } while (atomic_cmpxchg(v, old, new) != old); - return new; + return atomic_fetch_xor(bits, v) ^ bits; } #ifdef CONFIG_PGSTE diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c index 4113a7ffa149..5b22c6e24528 100644 --- a/arch/s390/mm/vmem.c +++ b/arch/s390/mm/vmem.c @@ -5,6 +5,7 @@ #include <linux/memory_hotplug.h> #include <linux/memblock.h> +#include <linux/kasan.h> #include <linux/pfn.h> #include <linux/mm.h> #include <linux/init.h> @@ -664,6 +665,9 @@ static void __init memblock_region_swap(void *a, void *b, int size) swap(*(struct memblock_region *)a, *(struct memblock_region *)b); } +#ifdef CONFIG_KASAN +#define __sha(x) ((unsigned long)kasan_mem_to_shadow((void *)x)) +#endif /* * map whole physical memory to virtual memory (identity mapping) * we reserve enough space in the vmalloc area for vmemmap to hotplug @@ -728,23 +732,24 @@ void __init vmem_map_init(void) memblock_region_cmp, memblock_region_swap); __for_each_mem_range(i, &memblock.memory, &memory_rwx, NUMA_NO_NODE, MEMBLOCK_NONE, &base, &end, NULL) { - __set_memory((unsigned long)__va(base), - (end - base) >> PAGE_SHIFT, - SET_MEMORY_RW | SET_MEMORY_NX); + set_memory_rwnx((unsigned long)__va(base), + (end - base) >> PAGE_SHIFT); } - __set_memory((unsigned long)_stext, - (unsigned long)(_etext - _stext) >> PAGE_SHIFT, - SET_MEMORY_RO | SET_MEMORY_X); - __set_memory((unsigned long)_etext, - (unsigned long)(__end_rodata - _etext) >> PAGE_SHIFT, - SET_MEMORY_RO); - __set_memory((unsigned long)_sinittext, - (unsigned long)(_einittext - _sinittext) >> PAGE_SHIFT, - SET_MEMORY_RO | SET_MEMORY_X); - __set_memory(__stext_amode31, - (__etext_amode31 - __stext_amode31) >> PAGE_SHIFT, - SET_MEMORY_RO | SET_MEMORY_X); +#ifdef CONFIG_KASAN + for_each_mem_range(i, &base, &end) { + set_memory_rwnx(__sha(base), + (__sha(end) - __sha(base)) >> PAGE_SHIFT); + } +#endif + set_memory_rox((unsigned long)_stext, + (unsigned long)(_etext - _stext) >> PAGE_SHIFT); + set_memory_ro((unsigned long)_etext, + (unsigned long)(__end_rodata - _etext) >> PAGE_SHIFT); + set_memory_rox((unsigned long)_sinittext, + (unsigned long)(_einittext - _sinittext) >> PAGE_SHIFT); + set_memory_rox(__stext_amode31, + (__etext_amode31 - __stext_amode31) >> PAGE_SHIFT); /* lowcore must be executable for LPSWE */ if (static_key_enabled(&cpu_has_bear)) diff --git a/arch/s390/pci/pci.c b/arch/s390/pci/pci.c index e16afacc8fd1..afc3f33788da 100644 --- a/arch/s390/pci/pci.c +++ b/arch/s390/pci/pci.c @@ -874,32 +874,15 @@ bool zpci_is_device_configured(struct zpci_dev *zdev) * @fh: The general function handle supplied by the platform * * Given a device in the configuration state Configured, enables, scans and - * adds it to the common code PCI subsystem if possible. If the PCI device is - * parked because we can not yet create a PCI bus because we have not seen - * function 0, it is ignored but will be scanned once function 0 appears. - * If any failure occurs, the zpci_dev is left disabled. + * adds it to the common code PCI subsystem if possible. If any failure occurs, + * the zpci_dev is left disabled. * * Return: 0 on success, or an error code otherwise */ int zpci_scan_configured_device(struct zpci_dev *zdev, u32 fh) { - int rc; - zpci_update_fh(zdev, fh); - /* the PCI function will be scanned once function 0 appears */ - if (!zdev->zbus->bus) - return 0; - - /* For function 0 on a multi-function bus scan whole bus as we might - * have to pick up existing functions waiting for it to allow creating - * the PCI bus - */ - if (zdev->devfn == 0 && zdev->zbus->multifunction) - rc = zpci_bus_scan_bus(zdev->zbus); - else - rc = zpci_bus_scan_device(zdev); - - return rc; + return zpci_bus_scan_device(zdev); } /** diff --git a/arch/s390/pci/pci_bus.c b/arch/s390/pci/pci_bus.c index a99926af2b69..32245b970a0c 100644 --- a/arch/s390/pci/pci_bus.c +++ b/arch/s390/pci/pci_bus.c @@ -85,9 +85,8 @@ int zpci_bus_scan_device(struct zpci_dev *zdev) if (!pdev) return -ENODEV; - pci_bus_add_device(pdev); pci_lock_rescan_remove(); - pci_bus_add_devices(zdev->zbus->bus); + pci_bus_add_device(pdev); pci_unlock_rescan_remove(); return 0; @@ -130,11 +129,8 @@ void zpci_bus_remove_device(struct zpci_dev *zdev, bool set_error) * @zbus: the zbus to be scanned * * Enables and scans all PCI functions on the bus making them available to the - * common PCI code. If there is no function 0 on the zbus nothing is scanned. If - * a function does not have a slot yet because it was added to the zbus before - * function 0 the slot is created. If a PCI function fails to be initialized - * an error will be returned but attempts will still be made for all other - * functions on the bus. + * common PCI code. If a PCI function fails to be initialized an error will be + * returned but attempts will still be made for all other functions on the bus. * * Return: 0 on success, an error value otherwise */ @@ -211,7 +207,6 @@ static int zpci_bus_create_pci_bus(struct zpci_bus *zbus, struct zpci_dev *fr, s } zbus->bus = bus; - pci_bus_add_devices(bus); return 0; } diff --git a/arch/s390/purgatory/head.S b/arch/s390/purgatory/head.S index 6f835124ee82..e5bd1a503528 100644 --- a/arch/s390/purgatory/head.S +++ b/arch/s390/purgatory/head.S @@ -76,9 +76,9 @@ diag %r0,%r1,0x308 .endm -.text -.align PAGE_SIZE -ENTRY(purgatory_start) + .text + .balign PAGE_SIZE +SYM_CODE_START(purgatory_start) /* The purgatory might be called after a diag308 so better set * architecture and addressing mode. */ @@ -245,45 +245,21 @@ ENTRY(purgatory_start) /* start crash kernel */ START_NEXT_KERNEL .base_dst 1 - - -load_psw_mask: - .long 0x00080000,0x80000000 - - .align 8 -disabled_wait_psw: - .quad 0x0002000180000000 - .quad 0x0000000000000000 + .do_checksum_verification - -gprregs: - .rept 10 - .quad 0 - .endr - -/* Macro to define a global variable with name and size (in bytes) to be - * shared with C code. - * - * Add the .size and .type attribute to satisfy checks on the Elf_Sym during - * purgatory load. - */ -.macro GLOBAL_VARIABLE name,size -\name: - .global \name - .size \name,\size - .type \name,object - .skip \size,0 -.endm - -GLOBAL_VARIABLE purgatory_sha256_digest,32 -GLOBAL_VARIABLE purgatory_sha_regions,16*__KEXEC_SHA_REGION_SIZE -GLOBAL_VARIABLE kernel_entry,8 -GLOBAL_VARIABLE kernel_type,8 -GLOBAL_VARIABLE crash_start,8 -GLOBAL_VARIABLE crash_size,8 - - .align PAGE_SIZE -stack: +SYM_CODE_END(purgatory_start) + +SYM_DATA_LOCAL(load_psw_mask, .long 0x00080000,0x80000000) + .balign 8 +SYM_DATA_LOCAL(disabled_wait_psw, .quad 0x0002000180000000,.do_checksum_verification) +SYM_DATA_LOCAL(gprregs, .fill 10,8,0) +SYM_DATA(purgatory_sha256_digest, .skip 32) +SYM_DATA(purgatory_sha_regions, .skip 16*__KEXEC_SHA_REGION_SIZE) +SYM_DATA(kernel_entry, .skip 8) +SYM_DATA(kernel_type, .skip 8) +SYM_DATA(crash_start, .skip 8) +SYM_DATA(crash_size, .skip 8) + .balign PAGE_SIZE +SYM_DATA_START_LOCAL(stack) /* The buffer to move this code must be as big as the code. */ .skip stack-purgatory_start - .align PAGE_SIZE -purgatory_end: + .balign PAGE_SIZE +SYM_DATA_END_LABEL(stack, SYM_L_LOCAL, purgatory_end) diff --git a/arch/s390/purgatory/kexec-purgatory.S b/arch/s390/purgatory/kexec-purgatory.S index 8293753100ae..25f512b1de12 100644 --- a/arch/s390/purgatory/kexec-purgatory.S +++ b/arch/s390/purgatory/kexec-purgatory.S @@ -1,14 +1,12 @@ /* SPDX-License-Identifier: GPL-2.0 */ +#include <linux/linkage.h> .section .rodata, "a" - .align 8 -kexec_purgatory: - .globl kexec_purgatory + .balign 8 +SYM_DATA_START(kexec_purgatory) .incbin "arch/s390/purgatory/purgatory.ro" -.Lkexec_purgatroy_end: +SYM_DATA_END_LABEL(kexec_purgatory, SYM_L_LOCAL, kexec_purgatory_end) - .align 8 -kexec_purgatory_size: - .globl kexec_purgatory_size - .quad .Lkexec_purgatroy_end - kexec_purgatory + .balign 8 +SYM_DATA(kexec_purgatory_size, .quad kexec_purgatory_end-kexec_purgatory) |