diff options
Diffstat (limited to 'arch/s390')
78 files changed, 3156 insertions, 878 deletions
diff --git a/arch/s390/Kbuild b/arch/s390/Kbuild index 9858476fa0f..cc45d25487b 100644 --- a/arch/s390/Kbuild +++ b/arch/s390/Kbuild @@ -5,3 +5,4 @@ obj-$(CONFIG_CRYPTO_HW) += crypto/ obj-$(CONFIG_S390_HYPFS_FS) += hypfs/ obj-$(CONFIG_APPLDATA_BASE) += appldata/ obj-$(CONFIG_MATHEMU) += math-emu/ +obj-y += net/ diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index 107610e01a2..f9acddd9ace 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -49,10 +49,13 @@ config GENERIC_LOCKBREAK config PGSTE def_bool y if KVM -config VIRT_CPU_ACCOUNTING +config ARCH_SUPPORTS_DEBUG_PAGEALLOC def_bool y -config ARCH_SUPPORTS_DEBUG_PAGEALLOC +config KEXEC + def_bool y + +config AUDIT_ARCH def_bool y config S390 @@ -84,11 +87,15 @@ config S390 select HAVE_KERNEL_XZ select HAVE_ARCH_MUTEX_CPU_RELAX select HAVE_ARCH_JUMP_LABEL if !MARCH_G5 + select HAVE_BPF_JIT if 64BIT && PACK_STACK select ARCH_SAVE_PAGE_KEYS if HIBERNATION select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE select HAVE_MEMBLOCK select HAVE_MEMBLOCK_NODE_MAP select HAVE_CMPXCHG_LOCAL + select HAVE_CMPXCHG_DOUBLE + select HAVE_VIRT_CPU_ACCOUNTING + select VIRT_CPU_ACCOUNTING select ARCH_DISCARD_MEMBLOCK select BUILDTIME_EXTABLE_SORT select ARCH_INLINE_SPIN_TRYLOCK @@ -133,9 +140,79 @@ source "init/Kconfig" source "kernel/Kconfig.freezer" -menu "Base setup" +menu "Processor type and features" + +config HAVE_MARCH_Z900_FEATURES + def_bool n + +config HAVE_MARCH_Z990_FEATURES + def_bool n + select HAVE_MARCH_Z900_FEATURES + +config HAVE_MARCH_Z9_109_FEATURES + def_bool n + select HAVE_MARCH_Z990_FEATURES + +config HAVE_MARCH_Z10_FEATURES + def_bool n + select HAVE_MARCH_Z9_109_FEATURES + +config HAVE_MARCH_Z196_FEATURES + def_bool n + select HAVE_MARCH_Z10_FEATURES + +choice + prompt "Processor type" + default MARCH_G5 + +config MARCH_G5 + bool "System/390 model G5 and G6" + depends on !64BIT + help + Select this to build a 31 bit kernel that works + on all ESA/390 and z/Architecture machines. -comment "Processor type and features" +config MARCH_Z900 + bool "IBM zSeries model z800 and z900" + select HAVE_MARCH_Z900_FEATURES if 64BIT + help + Select this to enable optimizations for model z800/z900 (2064 and + 2066 series). This will enable some optimizations that are not + available on older ESA/390 (31 Bit) only CPUs. + +config MARCH_Z990 + bool "IBM zSeries model z890 and z990" + select HAVE_MARCH_Z990_FEATURES if 64BIT + help + Select this to enable optimizations for model z890/z990 (2084 and + 2086 series). The kernel will be slightly faster but will not work + on older machines. + +config MARCH_Z9_109 + bool "IBM System z9" + select HAVE_MARCH_Z9_109_FEATURES if 64BIT + help + Select this to enable optimizations for IBM System z9 (2094 and + 2096 series). The kernel will be slightly faster but will not work + on older machines. + +config MARCH_Z10 + bool "IBM System z10" + select HAVE_MARCH_Z10_FEATURES if 64BIT + help + Select this to enable optimizations for IBM System z10 (2097 and + 2098 series). The kernel will be slightly faster but will not work + on older machines. + +config MARCH_Z196 + bool "IBM zEnterprise 114 and 196" + select HAVE_MARCH_Z196_FEATURES if 64BIT + help + Select this to enable optimizations for IBM zEnterprise 114 and 196 + (2818 and 2817 series). The kernel will be slightly faster but will + not work on older machines. + +endchoice config 64BIT def_bool y @@ -147,6 +224,24 @@ config 64BIT config 32BIT def_bool y if !64BIT +config COMPAT + def_bool y + prompt "Kernel support for 31 bit emulation" + depends on 64BIT + select COMPAT_BINFMT_ELF if BINFMT_ELF + select ARCH_WANT_OLD_COMPAT_IPC + help + Select this option if you want to enable your system kernel to + handle system-calls from ELF binaries for 31 bit ESA. This option + (and some other stuff like libraries and such) is needed for + executing 31 bit applications. It is safe to say "Y". + +config SYSVIPC_COMPAT + def_bool y if COMPAT && SYSVIPC + +config KEYS_COMPAT + def_bool y if COMPAT && KEYS + config SMP def_bool y prompt "Symmetric multi-processing support" @@ -202,6 +297,8 @@ config SCHED_BOOK Book scheduler support improves the CPU scheduler's decision making when dealing with machines that have several books. +source kernel/Kconfig.preempt + config MATHEMU def_bool y prompt "IEEE FPU emulation" @@ -211,100 +308,35 @@ config MATHEMU on older ESA/390 machines. Say Y unless you know your machine doesn't need this. -config COMPAT - def_bool y - prompt "Kernel support for 31 bit emulation" - depends on 64BIT - select COMPAT_BINFMT_ELF if BINFMT_ELF - select ARCH_WANT_OLD_COMPAT_IPC - help - Select this option if you want to enable your system kernel to - handle system-calls from ELF binaries for 31 bit ESA. This option - (and some other stuff like libraries and such) is needed for - executing 31 bit applications. It is safe to say "Y". +source kernel/Kconfig.hz -config SYSVIPC_COMPAT - def_bool y if COMPAT && SYSVIPC +endmenu -config KEYS_COMPAT - def_bool y if COMPAT && KEYS +menu "Memory setup" -config AUDIT_ARCH +config ARCH_SPARSEMEM_ENABLE def_bool y + select SPARSEMEM_VMEMMAP_ENABLE + select SPARSEMEM_VMEMMAP + select SPARSEMEM_STATIC if !64BIT -config HAVE_MARCH_Z900_FEATURES - def_bool n - -config HAVE_MARCH_Z990_FEATURES - def_bool n - select HAVE_MARCH_Z900_FEATURES - -config HAVE_MARCH_Z9_109_FEATURES - def_bool n - select HAVE_MARCH_Z990_FEATURES - -config HAVE_MARCH_Z10_FEATURES - def_bool n - select HAVE_MARCH_Z9_109_FEATURES - -config HAVE_MARCH_Z196_FEATURES - def_bool n - select HAVE_MARCH_Z10_FEATURES - -comment "Code generation options" - -choice - prompt "Processor type" - default MARCH_G5 - -config MARCH_G5 - bool "System/390 model G5 and G6" - depends on !64BIT - help - Select this to build a 31 bit kernel that works - on all ESA/390 and z/Architecture machines. - -config MARCH_Z900 - bool "IBM zSeries model z800 and z900" - select HAVE_MARCH_Z900_FEATURES if 64BIT - help - Select this to enable optimizations for model z800/z900 (2064 and - 2066 series). This will enable some optimizations that are not - available on older ESA/390 (31 Bit) only CPUs. +config ARCH_SPARSEMEM_DEFAULT + def_bool y -config MARCH_Z990 - bool "IBM zSeries model z890 and z990" - select HAVE_MARCH_Z990_FEATURES if 64BIT - help - Select this to enable optimizations for model z890/z990 (2084 and - 2086 series). The kernel will be slightly faster but will not work - on older machines. +config ARCH_SELECT_MEMORY_MODEL + def_bool y -config MARCH_Z9_109 - bool "IBM System z9" - select HAVE_MARCH_Z9_109_FEATURES if 64BIT - help - Select this to enable optimizations for IBM System z9 (2094 and - 2096 series). The kernel will be slightly faster but will not work - on older machines. +config ARCH_ENABLE_MEMORY_HOTPLUG + def_bool y if SPARSEMEM -config MARCH_Z10 - bool "IBM System z10" - select HAVE_MARCH_Z10_FEATURES if 64BIT - help - Select this to enable optimizations for IBM System z10 (2097 and - 2098 series). The kernel will be slightly faster but will not work - on older machines. +config ARCH_ENABLE_MEMORY_HOTREMOVE + def_bool y -config MARCH_Z196 - bool "IBM zEnterprise 114 and 196" - select HAVE_MARCH_Z196_FEATURES if 64BIT - help - Select this to enable optimizations for IBM zEnterprise 114 and 196 - (2818 and 2817 series). The kernel will be slightly faster but will - not work on older machines. +config FORCE_MAX_ZONEORDER + int + default "9" -endchoice +source "mm/Kconfig" config PACK_STACK def_bool y @@ -368,34 +400,9 @@ config WARN_DYNAMIC_STACK Say N if you are unsure. -comment "Kernel preemption" - -source "kernel/Kconfig.preempt" - -config ARCH_SPARSEMEM_ENABLE - def_bool y - select SPARSEMEM_VMEMMAP_ENABLE - select SPARSEMEM_VMEMMAP - select SPARSEMEM_STATIC if !64BIT - -config ARCH_SPARSEMEM_DEFAULT - def_bool y - -config ARCH_SELECT_MEMORY_MODEL - def_bool y - -config ARCH_ENABLE_MEMORY_HOTPLUG - def_bool y if SPARSEMEM - -config ARCH_ENABLE_MEMORY_HOTREMOVE - def_bool y - -config ARCH_HIBERNATION_POSSIBLE - def_bool y if 64BIT - -source "mm/Kconfig" +endmenu -comment "I/O subsystem configuration" +menu "I/O subsystem" config QDIO def_tristate y @@ -426,13 +433,102 @@ config CHSC_SCH If unsure, say N. -comment "Misc" +config SCM_BUS + def_bool y + depends on 64BIT + prompt "SCM bus driver" + help + Bus driver for Storage Class Memory. + +config EADM_SCH + def_tristate m + prompt "Support for EADM subchannels" + depends on SCM_BUS + help + This driver allows usage of EADM subchannels. EADM subchannels act + as a communication vehicle for SCM increments. + + To compile this driver as a module, choose M here: the + module will be called eadm_sch. + +endmenu + +menu "Dump support" + +config CRASH_DUMP + bool "kernel crash dumps" + depends on 64BIT && SMP + select KEXEC + help + Generate crash dump after being started by kexec. + Crash dump kernels are loaded in the main kernel with kexec-tools + into a specially reserved region and then later executed after + a crash by kdump/kexec. + For more details see Documentation/kdump/kdump.txt + +config ZFCPDUMP + def_bool n + prompt "zfcpdump support" + select SMP + help + Select this option if you want to build an zfcpdump enabled kernel. + Refer to <file:Documentation/s390/zfcpdump.txt> for more details on this. + +endmenu + +menu "Executable file formats / Emulations" source "fs/Kconfig.binfmt" -config FORCE_MAX_ZONEORDER - int - default "9" +config SECCOMP + def_bool y + prompt "Enable seccomp to safely compute untrusted bytecode" + depends on PROC_FS + help + This kernel feature is useful for number crunching applications + that may need to compute untrusted bytecode during their + execution. By using pipes or other transports made available to + the process as file descriptors supporting the read/write + syscalls, it's possible to isolate those applications in + their own address space using seccomp. Once seccomp is + enabled via /proc/<pid>/seccomp, it cannot be disabled + and the task is only allowed to execute a few safe syscalls + defined by each seccomp mode. + + If unsure, say Y. + +endmenu + +menu "Power Management" + +config ARCH_HIBERNATION_POSSIBLE + def_bool y if 64BIT + +source "kernel/power/Kconfig" + +endmenu + +source "net/Kconfig" + +config PCMCIA + def_bool n + +config CCW + def_bool y + +source "drivers/Kconfig" + +source "fs/Kconfig" + +source "arch/s390/Kconfig.debug" + +source "security/Kconfig" + +source "crypto/Kconfig" + +source "lib/Kconfig" + +menu "Virtualization" config PFAULT def_bool y @@ -448,8 +544,8 @@ config PFAULT this option. config SHARED_KERNEL - def_bool y - prompt "VM shared kernel support" + bool "VM shared kernel support" + depends on !JUMP_LABEL help Select this option, if you want to share the text segment of the Linux kernel between different VM guests. This reduces memory @@ -544,8 +640,6 @@ config APPLDATA_NET_SUM This can also be compiled as a module, which will be called appldata_net_sum.o. -source kernel/Kconfig.hz - config S390_HYPFS_FS def_bool y prompt "s390 hypervisor file system support" @@ -554,90 +648,21 @@ config S390_HYPFS_FS This is a virtual file system intended to provide accounting information in an s390 hypervisor environment. -config KEXEC - def_bool n - prompt "kexec system call" - help - kexec is a system call that implements the ability to shutdown your - current kernel, and to start another kernel. It is like a reboot - but is independent of hardware/microcode support. - -config CRASH_DUMP - bool "kernel crash dumps" - depends on 64BIT && SMP - select KEXEC - help - Generate crash dump after being started by kexec. - Crash dump kernels are loaded in the main kernel with kexec-tools - into a specially reserved region and then later executed after - a crash by kdump/kexec. - For more details see Documentation/kdump/kdump.txt - -config ZFCPDUMP - def_bool n - prompt "zfcpdump support" - select SMP - help - Select this option if you want to build an zfcpdump enabled kernel. - Refer to <file:Documentation/s390/zfcpdump.txt> for more details on this. +source "arch/s390/kvm/Kconfig" config S390_GUEST def_bool y - prompt "s390 guest support for KVM (EXPERIMENTAL)" + prompt "s390 support for virtio devices (EXPERIMENTAL)" depends on 64BIT && EXPERIMENTAL select VIRTUALIZATION select VIRTIO select VIRTIO_RING select VIRTIO_CONSOLE help - Select this option if you want to run the kernel as a guest under - the KVM hypervisor. This will add detection for KVM as well as a - virtio transport. If KVM is detected, the virtio console will be - the default console. - -config SECCOMP - def_bool y - prompt "Enable seccomp to safely compute untrusted bytecode" - depends on PROC_FS - help - This kernel feature is useful for number crunching applications - that may need to compute untrusted bytecode during their - execution. By using pipes or other transports made available to - the process as file descriptors supporting the read/write - syscalls, it's possible to isolate those applications in - their own address space using seccomp. Once seccomp is - enabled via /proc/<pid>/seccomp, it cannot be disabled - and the task is only allowed to execute a few safe syscalls - defined by each seccomp mode. - - If unsure, say Y. - -endmenu + Enabling this option adds support for virtio based paravirtual device + drivers on s390. -menu "Power Management" - -source "kernel/power/Kconfig" + Select this option if you want to run the kernel as a guest under + the KVM hypervisor. endmenu - -source "net/Kconfig" - -config PCMCIA - def_bool n - -config CCW - def_bool y - -source "drivers/Kconfig" - -source "fs/Kconfig" - -source "arch/s390/Kconfig.debug" - -source "security/Kconfig" - -source "crypto/Kconfig" - -source "lib/Kconfig" - -source "arch/s390/kvm/Kconfig" diff --git a/arch/s390/boot/compressed/Makefile b/arch/s390/boot/compressed/Makefile index 10e22c4ec4a..3ad8f61c998 100644 --- a/arch/s390/boot/compressed/Makefile +++ b/arch/s390/boot/compressed/Makefile @@ -11,6 +11,7 @@ targets := vmlinux.lds vmlinux vmlinux.bin vmlinux.bin.gz vmlinux.bin.bz2 \ sizes.h head$(BITS).o KBUILD_CFLAGS := -m$(BITS) -D__KERNEL__ $(LINUX_INCLUDE) -O2 +KBUILD_CFLAGS += -DDISABLE_BRANCH_PROFILING KBUILD_CFLAGS += $(cflags-y) KBUILD_CFLAGS += $(call cc-option,-mpacked-stack) KBUILD_CFLAGS += $(call cc-option,-ffreestanding) diff --git a/arch/s390/boot/compressed/misc.c b/arch/s390/boot/compressed/misc.c index 465eca756fe..c4c6a1cf221 100644 --- a/arch/s390/boot/compressed/misc.c +++ b/arch/s390/boot/compressed/misc.c @@ -71,34 +71,37 @@ void *memset(void *s, int c, size_t n) { char *xs; - if (c == 0) - return __builtin_memset(s, 0, n); - - xs = (char *) s; - if (n > 0) - do { - *xs++ = c; - } while (--n > 0); + xs = s; + while (n--) + *xs++ = c; return s; } -void *memcpy(void *__dest, __const void *__src, size_t __n) +void *memcpy(void *dest, const void *src, size_t n) { - return __builtin_memcpy(__dest, __src, __n); + const char *s = src; + char *d = dest; + + while (n--) + *d++ = *s++; + return dest; } -void *memmove(void *__dest, __const void *__src, size_t __n) +void *memmove(void *dest, const void *src, size_t n) { - char *d; - const char *s; - - if (__dest <= __src) - return __builtin_memcpy(__dest, __src, __n); - d = __dest + __n; - s = __src + __n; - while (__n--) - *--d = *--s; - return __dest; + const char *s = src; + char *d = dest; + + if (d <= s) { + while (n--) + *d++ = *s++; + } else { + d += n; + s += n; + while (n--) + *--d = *--s; + } + return dest; } static void error(char *x) diff --git a/arch/s390/defconfig b/arch/s390/defconfig index f39cd710980..b74400e3e03 100644 --- a/arch/s390/defconfig +++ b/arch/s390/defconfig @@ -16,8 +16,8 @@ CONFIG_CGROUPS=y CONFIG_CPUSETS=y CONFIG_CGROUP_CPUACCT=y CONFIG_RESOURCE_COUNTERS=y -CONFIG_CGROUP_MEMCG=y -CONFIG_CGROUP_MEM_RES_CTLR_SWAP=y +CONFIG_MEMCG=y +CONFIG_MEMCG_SWAP=y CONFIG_CGROUP_SCHED=y CONFIG_RT_GROUP_SCHED=y CONFIG_BLK_CGROUP=y @@ -32,20 +32,19 @@ CONFIG_EXPERT=y CONFIG_PROFILING=y CONFIG_OPROFILE=y CONFIG_KPROBES=y +CONFIG_JUMP_LABEL=y CONFIG_MODULES=y CONFIG_MODULE_UNLOAD=y CONFIG_MODVERSIONS=y CONFIG_PARTITION_ADVANCED=y CONFIG_IBM_PARTITION=y CONFIG_DEFAULT_DEADLINE=y -CONFIG_PREEMPT=y +CONFIG_HZ_100=y CONFIG_MEMORY_HOTPLUG=y CONFIG_MEMORY_HOTREMOVE=y CONFIG_KSM=y -CONFIG_BINFMT_MISC=m -CONFIG_CMM=m -CONFIG_HZ_100=y CONFIG_CRASH_DUMP=y +CONFIG_BINFMT_MISC=m CONFIG_HIBERNATION=y CONFIG_PACKET=y CONFIG_UNIX=y @@ -75,6 +74,7 @@ CONFIG_NET_CLS_RSVP=m CONFIG_NET_CLS_RSVP6=m CONFIG_NET_CLS_ACT=y CONFIG_NET_ACT_POLICE=y +CONFIG_BPF_JIT=y CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug" CONFIG_DEVTMPFS=y CONFIG_BLK_DEV_LOOP=m @@ -121,7 +121,6 @@ CONFIG_DEBUG_NOTIFIERS=y CONFIG_RCU_TRACE=y CONFIG_KPROBES_SANITY_TEST=y CONFIG_DEBUG_FORCE_WEAK_PER_CPU=y -CONFIG_CPU_NOTIFIER_ERROR_INJECT=m CONFIG_LATENCYTOP=y CONFIG_DEBUG_PAGEALLOC=y CONFIG_BLK_DEV_IO_TRACE=y @@ -173,3 +172,4 @@ CONFIG_CRYPTO_SHA512_S390=m CONFIG_CRYPTO_DES_S390=m CONFIG_CRYPTO_AES_S390=m CONFIG_CRC7=m +CONFIG_CMM=m diff --git a/arch/s390/include/asm/appldata.h b/arch/s390/include/asm/appldata.h index f328294faea..32a70598715 100644 --- a/arch/s390/include/asm/appldata.h +++ b/arch/s390/include/asm/appldata.h @@ -70,7 +70,7 @@ static inline int appldata_asm(struct appldata_product_id *id, int ry; if (!MACHINE_IS_VM) - return -ENOSYS; + return -EOPNOTSUPP; parm_list.diag = 0xdc; parm_list.function = fn; parm_list.parlist_length = sizeof(parm_list); diff --git a/arch/s390/include/asm/chsc.h b/arch/s390/include/asm/chsc.h index bf115b49f44..aea451fd182 100644 --- a/arch/s390/include/asm/chsc.h +++ b/arch/s390/include/asm/chsc.h @@ -125,32 +125,4 @@ struct chsc_cpd_info { #define CHSC_INFO_CPD _IOWR(CHSC_IOCTL_MAGIC, 0x87, struct chsc_cpd_info) #define CHSC_INFO_DCAL _IOWR(CHSC_IOCTL_MAGIC, 0x88, struct chsc_dcal) -#ifdef __KERNEL__ - -struct css_general_char { - u64 : 12; - u32 dynio : 1; /* bit 12 */ - u32 : 28; - u32 aif : 1; /* bit 41 */ - u32 : 3; - u32 mcss : 1; /* bit 45 */ - u32 fcs : 1; /* bit 46 */ - u32 : 1; - u32 ext_mb : 1; /* bit 48 */ - u32 : 7; - u32 aif_tdd : 1; /* bit 56 */ - u32 : 1; - u32 qebsm : 1; /* bit 58 */ - u32 : 8; - u32 aif_osa : 1; /* bit 67 */ - u32 : 14; - u32 cib : 1; /* bit 82 */ - u32 : 5; - u32 fcx : 1; /* bit 88 */ - u32 : 7; -}__attribute__((packed)); - -extern struct css_general_char css_general_characteristics; - -#endif /* __KERNEL__ */ #endif diff --git a/arch/s390/include/asm/cio.h b/arch/s390/include/asm/cio.h index 77043aa44d6..55bde603521 100644 --- a/arch/s390/include/asm/cio.h +++ b/arch/s390/include/asm/cio.h @@ -80,6 +80,18 @@ struct erw { } __attribute__ ((packed)); /** + * struct erw_eadm - EADM Subchannel extended report word + * @b: aob error + * @r: arsb error + */ +struct erw_eadm { + __u32 : 16; + __u32 b : 1; + __u32 r : 1; + __u32 : 14; +} __packed; + +/** * struct sublog - subchannel logout area * @res0: reserved * @esf: extended status flags @@ -170,9 +182,22 @@ struct esw3 { } __attribute__ ((packed)); /** + * struct esw_eadm - EADM Subchannel Extended Status Word (ESW) + * @sublog: subchannel logout + * @erw: extended report word + */ +struct esw_eadm { + __u32 sublog; + struct erw_eadm erw; + __u32 : 32; + __u32 : 32; + __u32 : 32; +} __packed; + +/** * struct irb - interruption response block * @scsw: subchannel status word - * @esw: extened status word, 4 formats + * @esw: extened status word * @ecw: extended control word * * The irb that is handed to the device driver when an interrupt occurs. For @@ -191,6 +216,7 @@ struct irb { struct esw1 esw1; struct esw2 esw2; struct esw3 esw3; + struct esw_eadm eadm; } esw; __u8 ecw[32]; } __attribute__ ((packed,aligned(4))); diff --git a/arch/s390/include/asm/cmpxchg.h b/arch/s390/include/asm/cmpxchg.h index 8d798e962b6..0f636cbdf34 100644 --- a/arch/s390/include/asm/cmpxchg.h +++ b/arch/s390/include/asm/cmpxchg.h @@ -7,7 +7,9 @@ #ifndef __ASM_CMPXCHG_H #define __ASM_CMPXCHG_H +#include <linux/mmdebug.h> #include <linux/types.h> +#include <linux/bug.h> extern void __xchg_called_with_bad_pointer(void); @@ -203,6 +205,65 @@ static inline unsigned long long __cmpxchg64(void *ptr, }) #endif /* CONFIG_64BIT */ +#define __cmpxchg_double_op(p1, p2, o1, o2, n1, n2, insn) \ +({ \ + register __typeof__(*(p1)) __old1 asm("2") = (o1); \ + register __typeof__(*(p2)) __old2 asm("3") = (o2); \ + register __typeof__(*(p1)) __new1 asm("4") = (n1); \ + register __typeof__(*(p2)) __new2 asm("5") = (n2); \ + int cc; \ + asm volatile( \ + insn " %[old],%[new],%[ptr]\n" \ + " ipm %[cc]\n" \ + " srl %[cc],28" \ + : [cc] "=d" (cc), [old] "+d" (__old1), "+d" (__old2) \ + : [new] "d" (__new1), "d" (__new2), \ + [ptr] "Q" (*(p1)), "Q" (*(p2)) \ + : "memory", "cc"); \ + !cc; \ +}) + +#define __cmpxchg_double_4(p1, p2, o1, o2, n1, n2) \ + __cmpxchg_double_op(p1, p2, o1, o2, n1, n2, "cds") + +#define __cmpxchg_double_8(p1, p2, o1, o2, n1, n2) \ + __cmpxchg_double_op(p1, p2, o1, o2, n1, n2, "cdsg") + +extern void __cmpxchg_double_called_with_bad_pointer(void); + +#define __cmpxchg_double(p1, p2, o1, o2, n1, n2) \ +({ \ + int __ret; \ + switch (sizeof(*(p1))) { \ + case 4: \ + __ret = __cmpxchg_double_4(p1, p2, o1, o2, n1, n2); \ + break; \ + case 8: \ + __ret = __cmpxchg_double_8(p1, p2, o1, o2, n1, n2); \ + break; \ + default: \ + __cmpxchg_double_called_with_bad_pointer(); \ + } \ + __ret; \ +}) + +#define cmpxchg_double(p1, p2, o1, o2, n1, n2) \ +({ \ + __typeof__(p1) __p1 = (p1); \ + __typeof__(p2) __p2 = (p2); \ + int __ret; \ + BUILD_BUG_ON(sizeof(*(p1)) != sizeof(long)); \ + BUILD_BUG_ON(sizeof(*(p2)) != sizeof(long)); \ + VM_BUG_ON((unsigned long)((__p1) + 1) != (unsigned long)(__p2));\ + if (sizeof(long) == 4) \ + __ret = __cmpxchg_double_4(__p1, __p2, o1, o2, n1, n2); \ + else \ + __ret = __cmpxchg_double_8(__p1, __p2, o1, o2, n1, n2); \ + __ret; \ +}) + +#define system_has_cmpxchg_double() 1 + #include <asm-generic/cmpxchg-local.h> static inline unsigned long __cmpxchg_local(void *ptr, diff --git a/arch/s390/include/asm/cpu_mf.h b/arch/s390/include/asm/cpu_mf.h index a3afecdae14..35f0020b7ba 100644 --- a/arch/s390/include/asm/cpu_mf.h +++ b/arch/s390/include/asm/cpu_mf.h @@ -21,11 +21,15 @@ #define CPU_MF_INT_SF_LSDA (1 << 22) /* loss of sample data alert */ #define CPU_MF_INT_CF_CACA (1 << 7) /* counter auth. change alert */ #define CPU_MF_INT_CF_LCDA (1 << 6) /* loss of counter data alert */ +#define CPU_MF_INT_RI_HALTED (1 << 5) /* run-time instr. halted */ +#define CPU_MF_INT_RI_BUF_FULL (1 << 4) /* run-time instr. program + buffer full */ #define CPU_MF_INT_CF_MASK (CPU_MF_INT_CF_CACA|CPU_MF_INT_CF_LCDA) #define CPU_MF_INT_SF_MASK (CPU_MF_INT_SF_IAE|CPU_MF_INT_SF_ISE| \ CPU_MF_INT_SF_PRA|CPU_MF_INT_SF_SACA| \ CPU_MF_INT_SF_LSDA) +#define CPU_MF_INT_RI_MASK (CPU_MF_INT_RI_HALTED|CPU_MF_INT_RI_BUF_FULL) /* CPU measurement facility support */ static inline int cpum_cf_avail(void) diff --git a/arch/s390/include/asm/cputime.h b/arch/s390/include/asm/cputime.h index 8709bdef233..023d5ae2448 100644 --- a/arch/s390/include/asm/cputime.h +++ b/arch/s390/include/asm/cputime.h @@ -12,6 +12,9 @@ #include <linux/spinlock.h> #include <asm/div64.h> + +#define __ARCH_HAS_VTIME_ACCOUNT + /* We want to use full resolution of the CPU timer: 2**-12 micro-seconds. */ typedef unsigned long long __nocast cputime_t; diff --git a/arch/s390/include/asm/css_chars.h b/arch/s390/include/asm/css_chars.h new file mode 100644 index 00000000000..a06ebc2623f --- /dev/null +++ b/arch/s390/include/asm/css_chars.h @@ -0,0 +1,39 @@ +#ifndef _ASM_CSS_CHARS_H +#define _ASM_CSS_CHARS_H + +#include <linux/types.h> + +#ifdef __KERNEL__ + +struct css_general_char { + u64 : 12; + u32 dynio : 1; /* bit 12 */ + u32 : 4; + u32 eadm : 1; /* bit 17 */ + u32 : 23; + u32 aif : 1; /* bit 41 */ + u32 : 3; + u32 mcss : 1; /* bit 45 */ + u32 fcs : 1; /* bit 46 */ + u32 : 1; + u32 ext_mb : 1; /* bit 48 */ + u32 : 7; + u32 aif_tdd : 1; /* bit 56 */ + u32 : 1; + u32 qebsm : 1; /* bit 58 */ + u32 : 8; + u32 aif_osa : 1; /* bit 67 */ + u32 : 12; + u32 eadm_rf : 1; /* bit 80 */ + u32 : 1; + u32 cib : 1; /* bit 82 */ + u32 : 5; + u32 fcx : 1; /* bit 88 */ + u32 : 19; + u32 alt_ssi : 1; /* bit 108 */ +} __packed; + +extern struct css_general_char css_general_characteristics; + +#endif /* __KERNEL__ */ +#endif diff --git a/arch/s390/include/asm/eadm.h b/arch/s390/include/asm/eadm.h new file mode 100644 index 00000000000..8d4847191ec --- /dev/null +++ b/arch/s390/include/asm/eadm.h @@ -0,0 +1,124 @@ +#ifndef _ASM_S390_EADM_H +#define _ASM_S390_EADM_H + +#include <linux/types.h> +#include <linux/device.h> + +struct arqb { + u64 data; + u16 fmt:4; + u16:12; + u16 cmd_code; + u16:16; + u16 msb_count; + u32 reserved[12]; +} __packed; + +#define ARQB_CMD_MOVE 1 + +struct arsb { + u16 fmt:4; + u32:28; + u8 ef; + u8:8; + u8 ecbi; + u8:8; + u8 fvf; + u16:16; + u8 eqc; + u32:32; + u64 fail_msb; + u64 fail_aidaw; + u64 fail_ms; + u64 fail_scm; + u32 reserved[4]; +} __packed; + +struct msb { + u8 fmt:4; + u8 oc:4; + u8 flags; + u16:12; + u16 bs:4; + u32 blk_count; + u64 data_addr; + u64 scm_addr; + u64:64; +} __packed; + +struct aidaw { + u8 flags; + u32 :24; + u32 :32; + u64 data_addr; +} __packed; + +#define MSB_OC_CLEAR 0 +#define MSB_OC_READ 1 +#define MSB_OC_WRITE 2 +#define MSB_OC_RELEASE 3 + +#define MSB_FLAG_BNM 0x80 +#define MSB_FLAG_IDA 0x40 + +#define MSB_BS_4K 0 +#define MSB_BS_1M 1 + +#define AOB_NR_MSB 124 + +struct aob { + struct arqb request; + struct arsb response; + struct msb msb[AOB_NR_MSB]; +} __packed __aligned(PAGE_SIZE); + +struct aob_rq_header { + struct scm_device *scmdev; + char data[0]; +}; + +struct scm_device { + u64 address; + u64 size; + unsigned int nr_max_block; + struct device dev; + struct { + unsigned int persistence:4; + unsigned int oper_state:4; + unsigned int data_state:4; + unsigned int rank:4; + unsigned int release:1; + unsigned int res_id:8; + } __packed attrs; +}; + +#define OP_STATE_GOOD 1 +#define OP_STATE_TEMP_ERR 2 +#define OP_STATE_PERM_ERR 3 + +struct scm_driver { + struct device_driver drv; + int (*probe) (struct scm_device *scmdev); + int (*remove) (struct scm_device *scmdev); + void (*notify) (struct scm_device *scmdev); + void (*handler) (struct scm_device *scmdev, void *data, int error); +}; + +int scm_driver_register(struct scm_driver *scmdrv); +void scm_driver_unregister(struct scm_driver *scmdrv); + +int scm_start_aob(struct aob *aob); +void scm_irq_handler(struct aob *aob, int error); + +struct eadm_ops { + int (*eadm_start) (struct aob *aob); + struct module *owner; +}; + +int scm_get_ref(void); +void scm_put_ref(void); + +void register_eadm_ops(struct eadm_ops *ops); +void unregister_eadm_ops(struct eadm_ops *ops); + +#endif /* _ASM_S390_EADM_H */ diff --git a/arch/s390/include/asm/elf.h b/arch/s390/include/asm/elf.h index 9b94a160fe7..178ff966a8b 100644 --- a/arch/s390/include/asm/elf.h +++ b/arch/s390/include/asm/elf.h @@ -101,6 +101,7 @@ #define HWCAP_S390_HPAGE 128 #define HWCAP_S390_ETF3EH 256 #define HWCAP_S390_HIGH_GPRS 512 +#define HWCAP_S390_TE 1024 /* * These are used to set parameters in the core dumps. @@ -212,4 +213,6 @@ int arch_setup_additional_pages(struct linux_binprm *, int); extern unsigned long arch_randomize_brk(struct mm_struct *mm); #define arch_randomize_brk arch_randomize_brk +void *fill_cpu_elf_notes(void *ptr, struct save_area *sa); + #endif diff --git a/arch/s390/include/asm/etr.h b/arch/s390/include/asm/etr.h index a24b03b9fb6..629b79a9316 100644 --- a/arch/s390/include/asm/etr.h +++ b/arch/s390/include/asm/etr.h @@ -140,7 +140,7 @@ struct etr_ptff_qto { /* Inline assembly helper functions */ static inline int etr_setr(struct etr_eacr *ctrl) { - int rc = -ENOSYS; + int rc = -EOPNOTSUPP; asm volatile( " .insn s,0xb2160000,%1\n" @@ -154,7 +154,7 @@ static inline int etr_setr(struct etr_eacr *ctrl) /* Stores a format 1 aib with 64 bytes */ static inline int etr_stetr(struct etr_aib *aib) { - int rc = -ENOSYS; + int rc = -EOPNOTSUPP; asm volatile( " .insn s,0xb2170000,%1\n" @@ -169,7 +169,7 @@ static inline int etr_stetr(struct etr_aib *aib) static inline int etr_steai(struct etr_aib *aib, unsigned int func) { register unsigned int reg0 asm("0") = func; - int rc = -ENOSYS; + int rc = -EOPNOTSUPP; asm volatile( " .insn s,0xb2b30000,%1\n" @@ -190,7 +190,7 @@ static inline int etr_ptff(void *ptff_block, unsigned int func) { register unsigned int reg0 asm("0") = func; register unsigned long reg1 asm("1") = (unsigned long) ptff_block; - int rc = -ENOSYS; + int rc = -EOPNOTSUPP; asm volatile( " .word 0x0104\n" diff --git a/arch/s390/include/asm/hugetlb.h b/arch/s390/include/asm/hugetlb.h index 799ed0f1643..2d6e6e38056 100644 --- a/arch/s390/include/asm/hugetlb.h +++ b/arch/s390/include/asm/hugetlb.h @@ -66,16 +66,6 @@ static inline pte_t huge_ptep_get(pte_t *ptep) return pte; } -static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) -{ - pte_t pte = huge_ptep_get(ptep); - - mm->context.flush_mm = 1; - pmd_clear((pmd_t *) ptep); - return pte; -} - static inline void __pmd_csp(pmd_t *pmdp) { register unsigned long reg2 asm("2") = pmd_val(*pmdp); @@ -117,6 +107,15 @@ static inline void huge_ptep_invalidate(struct mm_struct *mm, __pmd_csp(pmdp); } +static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm, + unsigned long addr, pte_t *ptep) +{ + pte_t pte = huge_ptep_get(ptep); + + huge_ptep_invalidate(mm, addr, ptep); + return pte; +} + #define huge_ptep_set_access_flags(__vma, __addr, __ptep, __entry, __dirty) \ ({ \ int __changed = !pte_same(huge_ptep_get(__ptep), __entry); \ @@ -131,10 +130,7 @@ static inline void huge_ptep_invalidate(struct mm_struct *mm, ({ \ pte_t __pte = huge_ptep_get(__ptep); \ if (pte_write(__pte)) { \ - (__mm)->context.flush_mm = 1; \ - if (atomic_read(&(__mm)->context.attach_count) > 1 || \ - (__mm) != current->active_mm) \ - huge_ptep_invalidate(__mm, __addr, __ptep); \ + huge_ptep_invalidate(__mm, __addr, __ptep); \ set_huge_pte_at(__mm, __addr, __ptep, \ huge_pte_wrprotect(__pte)); \ } \ diff --git a/arch/s390/include/asm/irq.h b/arch/s390/include/asm/irq.h index 2b9d41899d2..6703dd986fd 100644 --- a/arch/s390/include/asm/irq.h +++ b/arch/s390/include/asm/irq.h @@ -19,6 +19,7 @@ enum interruption_class { EXTINT_IUC, EXTINT_CMS, EXTINT_CMC, + EXTINT_CMR, IOINT_CIO, IOINT_QAI, IOINT_DAS, @@ -30,6 +31,7 @@ enum interruption_class { IOINT_CLW, IOINT_CTC, IOINT_APB, + IOINT_ADM, IOINT_CSC, NMI_NMI, NR_IRQS, diff --git a/arch/s390/include/asm/isc.h b/arch/s390/include/asm/isc.h index 1420a111594..5ae606456b0 100644 --- a/arch/s390/include/asm/isc.h +++ b/arch/s390/include/asm/isc.h @@ -14,6 +14,7 @@ /* Regular I/O interrupts. */ #define IO_SCH_ISC 3 /* regular I/O subchannels */ #define CONSOLE_ISC 1 /* console I/O subchannel */ +#define EADM_SCH_ISC 4 /* EADM subchannels */ #define CHSC_SCH_ISC 7 /* CHSC subchannels */ /* Adapter interrupts. */ #define QDIO_AIRQ_ISC IO_SCH_ISC /* I/O subchannel in qdio mode */ diff --git a/arch/s390/include/asm/lowcore.h b/arch/s390/include/asm/lowcore.h index aab5555bbbd..bbf8141408c 100644 --- a/arch/s390/include/asm/lowcore.h +++ b/arch/s390/include/asm/lowcore.h @@ -329,9 +329,13 @@ struct _lowcore { __u8 pad_0x1338[0x1340-0x1338]; /* 0x1338 */ __u32 access_regs_save_area[16]; /* 0x1340 */ __u64 cregs_save_area[16]; /* 0x1380 */ + __u8 pad_0x1400[0x1800-0x1400]; /* 0x1400 */ + + /* Transaction abort diagnostic block */ + __u8 pgm_tdb[256]; /* 0x1800 */ /* align to the top of the prefix area */ - __u8 pad_0x1400[0x2000-0x1400]; /* 0x1400 */ + __u8 pad_0x1900[0x2000-0x1900]; /* 0x1900 */ } __packed; #endif /* CONFIG_32BIT */ diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h index b749c573365..084e7755ed9 100644 --- a/arch/s390/include/asm/mmu_context.h +++ b/arch/s390/include/asm/mmu_context.h @@ -57,7 +57,7 @@ static inline void update_mm(struct mm_struct *mm, struct task_struct *tsk) pgd_t *pgd = mm->pgd; S390_lowcore.user_asce = mm->context.asce_bits | __pa(pgd); - if (addressing_mode != HOME_SPACE_MODE) { + if (s390_user_mode != HOME_SPACE_MODE) { /* Load primary space page table origin. */ asm volatile(LCTL_OPCODE" 1,1,%0\n" : : "m" (S390_lowcore.user_asce) ); diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h index 6537e72e085..86fe0ee2cee 100644 --- a/arch/s390/include/asm/percpu.h +++ b/arch/s390/include/asm/percpu.h @@ -20,7 +20,7 @@ #endif #define arch_this_cpu_to_op(pcp, val, op) \ -do { \ +({ \ typedef typeof(pcp) pcp_op_T__; \ pcp_op_T__ old__, new__, prev__; \ pcp_op_T__ *ptr__; \ @@ -39,13 +39,19 @@ do { \ } \ } while (prev__ != old__); \ preempt_enable(); \ -} while (0) + new__; \ +}) #define this_cpu_add_1(pcp, val) arch_this_cpu_to_op(pcp, val, +) #define this_cpu_add_2(pcp, val) arch_this_cpu_to_op(pcp, val, +) #define this_cpu_add_4(pcp, val) arch_this_cpu_to_op(pcp, val, +) #define this_cpu_add_8(pcp, val) arch_this_cpu_to_op(pcp, val, +) +#define this_cpu_add_return_1(pcp, val) arch_this_cpu_to_op(pcp, val, +) +#define this_cpu_add_return_2(pcp, val) arch_this_cpu_to_op(pcp, val, +) +#define this_cpu_add_return_4(pcp, val) arch_this_cpu_to_op(pcp, val, +) +#define this_cpu_add_return_8(pcp, val) arch_this_cpu_to_op(pcp, val, +) + #define this_cpu_and_1(pcp, val) arch_this_cpu_to_op(pcp, val, &) #define this_cpu_and_2(pcp, val) arch_this_cpu_to_op(pcp, val, &) #define this_cpu_and_4(pcp, val) arch_this_cpu_to_op(pcp, val, &) @@ -61,7 +67,7 @@ do { \ #define this_cpu_xor_4(pcp, val) arch_this_cpu_to_op(pcp, val, ^) #define this_cpu_xor_8(pcp, val) arch_this_cpu_to_op(pcp, val, ^) -#define arch_this_cpu_cmpxchg(pcp, oval, nval) \ +#define arch_this_cpu_cmpxchg(pcp, oval, nval) \ ({ \ typedef typeof(pcp) pcp_op_T__; \ pcp_op_T__ ret__; \ @@ -84,6 +90,44 @@ do { \ #define this_cpu_cmpxchg_4(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval) #define this_cpu_cmpxchg_8(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval) +#define arch_this_cpu_xchg(pcp, nval) \ +({ \ + typeof(pcp) *ptr__; \ + typeof(pcp) ret__; \ + preempt_disable(); \ + ptr__ = __this_cpu_ptr(&(pcp)); \ + ret__ = xchg(ptr__, nval); \ + preempt_enable(); \ + ret__; \ +}) + +#define this_cpu_xchg_1(pcp, nval) arch_this_cpu_xchg(pcp, nval) +#define this_cpu_xchg_2(pcp, nval) arch_this_cpu_xchg(pcp, nval) +#define this_cpu_xchg_4(pcp, nval) arch_this_cpu_xchg(pcp, nval) +#ifdef CONFIG_64BIT +#define this_cpu_xchg_8(pcp, nval) arch_this_cpu_xchg(pcp, nval) +#endif + +#define arch_this_cpu_cmpxchg_double(pcp1, pcp2, o1, o2, n1, n2) \ +({ \ + typeof(pcp1) o1__ = (o1), n1__ = (n1); \ + typeof(pcp2) o2__ = (o2), n2__ = (n2); \ + typeof(pcp1) *p1__; \ + typeof(pcp2) *p2__; \ + int ret__; \ + preempt_disable(); \ + p1__ = __this_cpu_ptr(&(pcp1)); \ + p2__ = __this_cpu_ptr(&(pcp2)); \ + ret__ = __cmpxchg_double(p1__, p2__, o1__, o2__, n1__, n2__); \ + preempt_enable(); \ + ret__; \ +}) + +#define this_cpu_cmpxchg_double_4 arch_this_cpu_cmpxchg_double +#ifdef CONFIG_64BIT +#define this_cpu_cmpxchg_double_8 arch_this_cpu_cmpxchg_double +#endif + #include <asm-generic/percpu.h> #endif /* __ARCH_S390_PERCPU__ */ diff --git a/arch/s390/include/asm/processor.h b/arch/s390/include/asm/processor.h index 11e4e323693..f3e0aabfc6b 100644 --- a/arch/s390/include/asm/processor.h +++ b/arch/s390/include/asm/processor.h @@ -11,12 +11,15 @@ #ifndef __ASM_S390_PROCESSOR_H #define __ASM_S390_PROCESSOR_H +#ifndef __ASSEMBLY__ + #include <linux/linkage.h> #include <linux/irqflags.h> #include <asm/cpu.h> #include <asm/page.h> #include <asm/ptrace.h> #include <asm/setup.h> +#include <asm/runtime_instr.h> /* * Default implementation of macro that returns current @@ -75,11 +78,20 @@ struct thread_struct { unsigned long gmap_addr; /* address of last gmap fault. */ struct per_regs per_user; /* User specified PER registers */ struct per_event per_event; /* Cause of the last PER trap */ + unsigned long per_flags; /* Flags to control debug behavior */ /* pfault_wait is used to block the process on a pfault event */ unsigned long pfault_wait; struct list_head list; + /* cpu runtime instrumentation */ + struct runtime_instr_cb *ri_cb; + int ri_signum; +#ifdef CONFIG_64BIT + unsigned char trap_tdb[256]; /* Transaction abort diagnose block */ +#endif }; +#define PER_FLAG_NO_TE 1UL /* Flag to disable transactions. */ + typedef struct thread_struct thread_struct; /* @@ -130,6 +142,12 @@ struct task_struct; struct mm_struct; struct seq_file; +#ifdef CONFIG_64BIT +extern void show_cacheinfo(struct seq_file *m); +#else +static inline void show_cacheinfo(struct seq_file *m) { } +#endif + /* Free all resources held by a thread. */ extern void release_thread(struct task_struct *); extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags); @@ -140,6 +158,7 @@ extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags); extern unsigned long thread_saved_pc(struct task_struct *t); extern void show_code(struct pt_regs *regs); +extern void print_fn_code(unsigned char *code, unsigned long len); unsigned long get_wchan(struct task_struct *p); #define task_pt_regs(tsk) ((struct pt_regs *) \ @@ -331,23 +350,6 @@ extern void (*s390_base_ext_handler_fn)(void); #define ARCH_LOW_ADDRESS_LIMIT 0x7fffffffUL -/* - * Helper macro for exception table entries - */ -#ifndef CONFIG_64BIT -#define EX_TABLE(_fault,_target) \ - ".section __ex_table,\"a\"\n" \ - " .align 4\n" \ - " .long " #_fault "," #_target "\n" \ - ".previous\n" -#else -#define EX_TABLE(_fault,_target) \ - ".section __ex_table,\"a\"\n" \ - " .align 8\n" \ - " .quad " #_fault "," #_target "\n" \ - ".previous\n" -#endif - extern int memcpy_real(void *, void *, size_t); extern void memcpy_absolute(void *, void *, size_t); @@ -358,4 +360,25 @@ extern void memcpy_absolute(void *, void *, size_t); memcpy_absolute(&(dest), &__tmp, sizeof(__tmp)); \ } -#endif /* __ASM_S390_PROCESSOR_H */ +/* + * Helper macro for exception table entries + */ +#define EX_TABLE(_fault, _target) \ + ".section __ex_table,\"a\"\n" \ + ".align 4\n" \ + ".long (" #_fault ") - .\n" \ + ".long (" #_target ") - .\n" \ + ".previous\n" + +#else /* __ASSEMBLY__ */ + +#define EX_TABLE(_fault, _target) \ + .section __ex_table,"a" ; \ + .align 4 ; \ + .long (_fault) - . ; \ + .long (_target) - . ; \ + .previous + +#endif /* __ASSEMBLY__ */ + +#endif /* __ASM_S390_PROCESSOR_H */ diff --git a/arch/s390/include/asm/ptrace.h b/arch/s390/include/asm/ptrace.h index d5f08ea566e..ce20a53afe9 100644 --- a/arch/s390/include/asm/ptrace.h +++ b/arch/s390/include/asm/ptrace.h @@ -235,6 +235,7 @@ typedef struct #define PSW_MASK_ASC 0x0000C000UL #define PSW_MASK_CC 0x00003000UL #define PSW_MASK_PM 0x00000F00UL +#define PSW_MASK_RI 0x00000000UL #define PSW_MASK_EA 0x00000000UL #define PSW_MASK_BA 0x00000000UL @@ -264,10 +265,11 @@ typedef struct #define PSW_MASK_ASC 0x0000C00000000000UL #define PSW_MASK_CC 0x0000300000000000UL #define PSW_MASK_PM 0x00000F0000000000UL +#define PSW_MASK_RI 0x0000008000000000UL #define PSW_MASK_EA 0x0000000100000000UL #define PSW_MASK_BA 0x0000000080000000UL -#define PSW_MASK_USER 0x00003F0180000000UL +#define PSW_MASK_USER 0x00003F8180000000UL #define PSW_ADDR_AMODE 0x0000000000000000UL #define PSW_ADDR_INSN 0xFFFFFFFFFFFFFFFFUL @@ -359,17 +361,19 @@ struct per_struct_kernel { unsigned char access_id; /* PER trap access identification */ }; -#define PER_EVENT_MASK 0xE9000000UL +#define PER_EVENT_MASK 0xEB000000UL #define PER_EVENT_BRANCH 0x80000000UL #define PER_EVENT_IFETCH 0x40000000UL #define PER_EVENT_STORE 0x20000000UL #define PER_EVENT_STORE_REAL 0x08000000UL +#define PER_EVENT_TRANSACTION_END 0x02000000UL #define PER_EVENT_NULLIFICATION 0x01000000UL -#define PER_CONTROL_MASK 0x00a00000UL +#define PER_CONTROL_MASK 0x00e00000UL #define PER_CONTROL_BRANCH_ADDRESS 0x00800000UL +#define PER_CONTROL_SUSPENSION 0x00400000UL #define PER_CONTROL_ALTERATION 0x00200000UL #endif @@ -483,6 +487,8 @@ typedef struct #define PTRACE_GET_LAST_BREAK 0x5006 #define PTRACE_PEEK_SYSTEM_CALL 0x5007 #define PTRACE_POKE_SYSTEM_CALL 0x5008 +#define PTRACE_ENABLE_TE 0x5009 +#define PTRACE_DISABLE_TE 0x5010 /* * PT_PROT definition is loosely based on hppa bsd definition in diff --git a/arch/s390/include/asm/runtime_instr.h b/arch/s390/include/asm/runtime_instr.h new file mode 100644 index 00000000000..830da737ff8 --- /dev/null +++ b/arch/s390/include/asm/runtime_instr.h @@ -0,0 +1,98 @@ +#ifndef _RUNTIME_INSTR_H +#define _RUNTIME_INSTR_H + +#define S390_RUNTIME_INSTR_START 0x1 +#define S390_RUNTIME_INSTR_STOP 0x2 + +struct runtime_instr_cb { + __u64 buf_current; + __u64 buf_origin; + __u64 buf_limit; + + __u32 valid : 1; + __u32 pstate : 1; + __u32 pstate_set_buf : 1; + __u32 home_space : 1; + __u32 altered : 1; + __u32 : 3; + __u32 pstate_sample : 1; + __u32 sstate_sample : 1; + __u32 pstate_collect : 1; + __u32 sstate_collect : 1; + __u32 : 1; + __u32 halted_int : 1; + __u32 int_requested : 1; + __u32 buffer_full_int : 1; + __u32 key : 4; + __u32 : 9; + __u32 rgs : 3; + + __u32 mode : 4; + __u32 next : 1; + __u32 mae : 1; + __u32 : 2; + __u32 call_type_br : 1; + __u32 return_type_br : 1; + __u32 other_type_br : 1; + __u32 bc_other_type : 1; + __u32 emit : 1; + __u32 tx_abort : 1; + __u32 : 2; + __u32 bp_xn : 1; + __u32 bp_xt : 1; + __u32 bp_ti : 1; + __u32 bp_ni : 1; + __u32 suppr_y : 1; + __u32 suppr_z : 1; + + __u32 dc_miss_extra : 1; + __u32 lat_lev_ignore : 1; + __u32 ic_lat_lev : 4; + __u32 dc_lat_lev : 4; + + __u64 reserved1; + __u64 scaling_factor; + __u64 rsic; + __u64 reserved2; +} __packed __aligned(8); + +extern struct runtime_instr_cb runtime_instr_empty_cb; + +static inline void load_runtime_instr_cb(struct runtime_instr_cb *cb) +{ + asm volatile(".insn rsy,0xeb0000000060,0,0,%0" /* LRIC */ + : : "Q" (*cb)); +} + +static inline void store_runtime_instr_cb(struct runtime_instr_cb *cb) +{ + asm volatile(".insn rsy,0xeb0000000061,0,0,%0" /* STRIC */ + : "=Q" (*cb) : : "cc"); +} + +static inline void save_ri_cb(struct runtime_instr_cb *cb_prev) +{ +#ifdef CONFIG_64BIT + if (cb_prev) + store_runtime_instr_cb(cb_prev); +#endif +} + +static inline void restore_ri_cb(struct runtime_instr_cb *cb_next, + struct runtime_instr_cb *cb_prev) +{ +#ifdef CONFIG_64BIT + if (cb_next) + load_runtime_instr_cb(cb_next); + else if (cb_prev) + load_runtime_instr_cb(&runtime_instr_empty_cb); +#endif +} + +#ifdef CONFIG_64BIT +extern void exit_thread_runtime_instr(void); +#else +static inline void exit_thread_runtime_instr(void) { } +#endif + +#endif /* _RUNTIME_INSTR_H */ diff --git a/arch/s390/include/asm/scsw.h b/arch/s390/include/asm/scsw.h index 4071d00978c..4af99cdaddf 100644 --- a/arch/s390/include/asm/scsw.h +++ b/arch/s390/include/asm/scsw.h @@ -1,7 +1,7 @@ /* * Helper functions for scsw access. * - * Copyright IBM Corp. 2008, 2009 + * Copyright IBM Corp. 2008, 2012 * Author(s): Peter Oberparleiter <peter.oberparleiter@de.ibm.com> */ @@ -9,7 +9,7 @@ #define _ASM_S390_SCSW_H_ #include <linux/types.h> -#include <asm/chsc.h> +#include <asm/css_chars.h> #include <asm/cio.h> /** @@ -100,14 +100,46 @@ struct tm_scsw { } __attribute__ ((packed)); /** + * struct eadm_scsw - subchannel status word for eadm subchannels + * @key: subchannel key + * @eswf: esw format + * @cc: deferred condition code + * @ectl: extended control + * @fctl: function control + * @actl: activity control + * @stctl: status control + * @aob: AOB address + * @dstat: device status + * @cstat: subchannel status + */ +struct eadm_scsw { + u32 key:4; + u32:1; + u32 eswf:1; + u32 cc:2; + u32:6; + u32 ectl:1; + u32:2; + u32 fctl:3; + u32 actl:7; + u32 stctl:5; + u32 aob; + u32 dstat:8; + u32 cstat:8; + u32:16; +} __packed; + +/** * union scsw - subchannel status word * @cmd: command-mode SCSW * @tm: transport-mode SCSW + * @eadm: eadm SCSW */ union scsw { struct cmd_scsw cmd; struct tm_scsw tm; -} __attribute__ ((packed)); + struct eadm_scsw eadm; +} __packed; #define SCSW_FCTL_CLEAR_FUNC 0x1 #define SCSW_FCTL_HALT_FUNC 0x2 diff --git a/arch/s390/include/asm/setup.h b/arch/s390/include/asm/setup.h index e6859d16ee2..87b47ca954f 100644 --- a/arch/s390/include/asm/setup.h +++ b/arch/s390/include/asm/setup.h @@ -60,7 +60,7 @@ void create_mem_hole(struct mem_chunk memory_chunk[], unsigned long addr, #define SECONDARY_SPACE_MODE 2 #define HOME_SPACE_MODE 3 -extern unsigned int addressing_mode; +extern unsigned int s390_user_mode; /* * Machine features detected in head.S @@ -80,6 +80,7 @@ extern unsigned int addressing_mode; #define MACHINE_FLAG_LPAR (1UL << 12) #define MACHINE_FLAG_SPP (1UL << 13) #define MACHINE_FLAG_TOPOLOGY (1UL << 14) +#define MACHINE_FLAG_TE (1UL << 15) #define MACHINE_IS_VM (S390_lowcore.machine_flags & MACHINE_FLAG_VM) #define MACHINE_IS_KVM (S390_lowcore.machine_flags & MACHINE_FLAG_KVM) @@ -98,6 +99,7 @@ extern unsigned int addressing_mode; #define MACHINE_HAS_PFMF (0) #define MACHINE_HAS_SPP (0) #define MACHINE_HAS_TOPOLOGY (0) +#define MACHINE_HAS_TE (0) #else /* CONFIG_64BIT */ #define MACHINE_HAS_IEEE (1) #define MACHINE_HAS_CSP (1) @@ -109,6 +111,7 @@ extern unsigned int addressing_mode; #define MACHINE_HAS_PFMF (S390_lowcore.machine_flags & MACHINE_FLAG_PFMF) #define MACHINE_HAS_SPP (S390_lowcore.machine_flags & MACHINE_FLAG_SPP) #define MACHINE_HAS_TOPOLOGY (S390_lowcore.machine_flags & MACHINE_FLAG_TOPOLOGY) +#define MACHINE_HAS_TE (S390_lowcore.machine_flags & MACHINE_FLAG_TE) #endif /* CONFIG_64BIT */ #define ZFCPDUMP_HSA_SIZE (32UL<<20) diff --git a/arch/s390/include/asm/smp.h b/arch/s390/include/asm/smp.h index ce26ac3cb16..b64f15c3b4c 100644 --- a/arch/s390/include/asm/smp.h +++ b/arch/s390/include/asm/smp.h @@ -30,6 +30,8 @@ extern int smp_vcpu_scheduled(int cpu); extern void smp_yield_cpu(int cpu); extern void smp_yield(void); extern void smp_stop_cpu(void); +extern void smp_cpu_set_polarization(int cpu, int val); +extern int smp_cpu_get_polarization(int cpu); #else /* CONFIG_SMP */ @@ -43,7 +45,7 @@ static inline void smp_call_online_cpu(void (*func)(void *), void *data) func(data); } -static inline int smp_find_processor_id(int address) { return 0; } +static inline int smp_find_processor_id(u16 address) { return 0; } static inline int smp_store_status(int cpu) { return 0; } static inline int smp_vcpu_scheduled(int cpu) { return 1; } static inline void smp_yield_cpu(int cpu) { } diff --git a/arch/s390/include/asm/string.h b/arch/s390/include/asm/string.h index 1bd1352fa3b..7e2dcd7c57e 100644 --- a/arch/s390/include/asm/string.h +++ b/arch/s390/include/asm/string.h @@ -96,7 +96,6 @@ static inline char *strcat(char *dst, const char *src) static inline char *strcpy(char *dst, const char *src) { -#if __GNUC__ < 4 register int r0 asm("0") = 0; char *ret = dst; @@ -106,14 +105,10 @@ static inline char *strcpy(char *dst, const char *src) : "+&a" (dst), "+&a" (src) : "d" (r0) : "cc", "memory"); return ret; -#else - return __builtin_strcpy(dst, src); -#endif } static inline size_t strlen(const char *s) { -#if __GNUC__ < 4 register unsigned long r0 asm("0") = 0; const char *tmp = s; @@ -122,9 +117,6 @@ static inline size_t strlen(const char *s) " jo 0b" : "+d" (r0), "+a" (tmp) : : "cc"); return r0 - (unsigned long) s; -#else - return __builtin_strlen(s); -#endif } static inline size_t strnlen(const char * s, size_t n) diff --git a/arch/s390/include/asm/switch_to.h b/arch/s390/include/asm/switch_to.h index f223068b782..f3a9e0f9270 100644 --- a/arch/s390/include/asm/switch_to.h +++ b/arch/s390/include/asm/switch_to.h @@ -80,21 +80,19 @@ static inline void restore_access_regs(unsigned int *acrs) if (prev->mm) { \ save_fp_regs(&prev->thread.fp_regs); \ save_access_regs(&prev->thread.acrs[0]); \ + save_ri_cb(prev->thread.ri_cb); \ } \ if (next->mm) { \ restore_fp_regs(&next->thread.fp_regs); \ restore_access_regs(&next->thread.acrs[0]); \ + restore_ri_cb(next->thread.ri_cb, prev->thread.ri_cb); \ update_per_regs(next); \ } \ prev = __switch_to(prev,next); \ } while (0) -extern void account_vtime(struct task_struct *, struct task_struct *); -extern void account_tick_vtime(struct task_struct *); - #define finish_arch_switch(prev) do { \ set_fs(current->thread.mm_segment); \ - account_vtime(prev, current); \ } while (0) #endif /* __ASM_SWITCH_TO_H */ diff --git a/arch/s390/include/asm/sysinfo.h b/arch/s390/include/asm/sysinfo.h index 282ee36f616..f92428e459f 100644 --- a/arch/s390/include/asm/sysinfo.h +++ b/arch/s390/include/asm/sysinfo.h @@ -17,7 +17,10 @@ #include <asm/bitsperlong.h> struct sysinfo_1_1_1 { - unsigned short :16; + unsigned char p:1; + unsigned char :6; + unsigned char t:1; + unsigned char :8; unsigned char ccr; unsigned char cai; char reserved_0[28]; @@ -30,9 +33,14 @@ struct sysinfo_1_1_1 { char model[16]; char model_perm_cap[16]; char model_temp_cap[16]; - char model_cap_rating[4]; - char model_perm_cap_rating[4]; - char model_temp_cap_rating[4]; + unsigned int model_cap_rating; + unsigned int model_perm_cap_rating; + unsigned int model_temp_cap_rating; + unsigned char typepct[5]; + unsigned char reserved_2[3]; + unsigned int ncr; + unsigned int npr; + unsigned int ntr; }; struct sysinfo_1_2_1 { @@ -47,8 +55,9 @@ struct sysinfo_1_2_2 { char format; char reserved_0[1]; unsigned short acc_offset; - char reserved_1[24]; - unsigned int secondary_capability; + char reserved_1[20]; + unsigned int nominal_cap; + unsigned int secondary_cap; unsigned int capability; unsigned short cpus_total; unsigned short cpus_configured; @@ -109,6 +118,8 @@ struct sysinfo_3_2_2 { char reserved_544[3552]; }; +extern int topology_max_mnest; + #define TOPOLOGY_CPU_BITS 64 #define TOPOLOGY_NR_MAG 6 @@ -142,21 +153,7 @@ struct sysinfo_15_1_x { union topology_entry tle[0]; }; -static inline int stsi(void *sysinfo, int fc, int sel1, int sel2) -{ - register int r0 asm("0") = (fc << 28) | sel1; - register int r1 asm("1") = sel2; - - asm volatile( - " stsi 0(%2)\n" - "0: jz 2f\n" - "1: lhi %0,%3\n" - "2:\n" - EX_TABLE(0b, 1b) - : "+d" (r0) : "d" (r1), "a" (sysinfo), "K" (-ENOSYS) - : "cc", "memory"); - return r0; -} +int stsi(void *sysinfo, int fc, int sel1, int sel2); /* * Service level reporting interface. diff --git a/arch/s390/include/asm/tlbflush.h b/arch/s390/include/asm/tlbflush.h index 9fde315f3a7..1d8fe2b17ef 100644 --- a/arch/s390/include/asm/tlbflush.h +++ b/arch/s390/include/asm/tlbflush.h @@ -90,12 +90,10 @@ static inline void __tlb_flush_mm(struct mm_struct * mm) static inline void __tlb_flush_mm_cond(struct mm_struct * mm) { - spin_lock(&mm->page_table_lock); if (mm->context.flush_mm) { __tlb_flush_mm(mm); mm->context.flush_mm = 0; } - spin_unlock(&mm->page_table_lock); } /* diff --git a/arch/s390/include/asm/topology.h b/arch/s390/include/asm/topology.h index 0837de80c35..9ca30538376 100644 --- a/arch/s390/include/asm/topology.h +++ b/arch/s390/include/asm/topology.h @@ -2,8 +2,8 @@ #define _ASM_S390_TOPOLOGY_H #include <linux/cpumask.h> -#include <asm/sysinfo.h> +struct sysinfo_15_1_x; struct cpu; #ifdef CONFIG_SCHED_BOOK @@ -51,24 +51,6 @@ static inline void topology_expect_change(void) { } #define POLARIZATION_VM (2) #define POLARIZATION_VH (3) -extern int cpu_polarization[]; - -static inline void cpu_set_polarization(int cpu, int val) -{ -#ifdef CONFIG_SCHED_BOOK - cpu_polarization[cpu] = val; -#endif -} - -static inline int cpu_read_polarization(int cpu) -{ -#ifdef CONFIG_SCHED_BOOK - return cpu_polarization[cpu]; -#else - return POLARIZATION_HRZ; -#endif -} - #ifdef CONFIG_SCHED_BOOK void s390_init_cpu_topology(void); #else diff --git a/arch/s390/include/asm/uaccess.h b/arch/s390/include/asm/uaccess.h index a8ab18b18b5..34268df959a 100644 --- a/arch/s390/include/asm/uaccess.h +++ b/arch/s390/include/asm/uaccess.h @@ -76,9 +76,22 @@ static inline int __range_ok(unsigned long addr, unsigned long size) struct exception_table_entry { - unsigned long insn, fixup; + int insn, fixup; }; +static inline unsigned long extable_insn(const struct exception_table_entry *x) +{ + return (unsigned long)&x->insn + x->insn; +} + +static inline unsigned long extable_fixup(const struct exception_table_entry *x) +{ + return (unsigned long)&x->fixup + x->fixup; +} + +#define ARCH_HAS_SORT_EXTABLE +#define ARCH_HAS_SEARCH_EXTABLE + struct uaccess_ops { size_t (*copy_from_user)(size_t, const void __user *, void *); size_t (*copy_from_user_small)(size_t, const void __user *, void *); diff --git a/arch/s390/include/asm/unistd.h b/arch/s390/include/asm/unistd.h index 6756e78f480..4e64b5cd155 100644 --- a/arch/s390/include/asm/unistd.h +++ b/arch/s390/include/asm/unistd.h @@ -277,7 +277,9 @@ #define __NR_setns 339 #define __NR_process_vm_readv 340 #define __NR_process_vm_writev 341 -#define NR_syscalls 342 +#define __NR_s390_runtime_instr 342 +#define __NR_kcmp 343 +#define NR_syscalls 344 /* * There are some system calls that are not present on 64 bit, some diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile index 9733b3f0eb6..4da52fe3174 100644 --- a/arch/s390/kernel/Makefile +++ b/arch/s390/kernel/Makefile @@ -23,10 +23,11 @@ CFLAGS_sysinfo.o += -Iinclude/math-emu -Iarch/s390/math-emu -w obj-y := bitmap.o traps.o time.o process.o base.o early.o setup.o vtime.o \ processor.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o nmi.o \ debug.o irq.o ipl.o dis.o diag.o mem_detect.o sclp.o vdso.o \ - sysinfo.o jump_label.o lgr.o os_info.o + sysinfo.o jump_label.o lgr.o os_info.o machine_kexec.o obj-y += $(if $(CONFIG_64BIT),entry64.o,entry.o) obj-y += $(if $(CONFIG_64BIT),reipl64.o,reipl.o) +obj-y += $(if $(CONFIG_64BIT),relocate_kernel64.o,relocate_kernel.o) extra-y += head.o vmlinux.lds extra-y += $(if $(CONFIG_64BIT),head64.o,head31.o) @@ -48,12 +49,11 @@ obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o obj-$(CONFIG_FTRACE_SYSCALLS) += ftrace.o obj-$(CONFIG_CRASH_DUMP) += crash_dump.o -obj-$(CONFIG_PERF_EVENTS) += perf_event.o perf_cpum_cf.o -# Kexec part -S390_KEXEC_OBJS := machine_kexec.o crash.o -S390_KEXEC_OBJS += $(if $(CONFIG_64BIT),relocate_kernel64.o,relocate_kernel.o) -obj-$(CONFIG_KEXEC) += $(S390_KEXEC_OBJS) +ifdef CONFIG_64BIT +obj-$(CONFIG_PERF_EVENTS) += perf_event.o perf_cpum_cf.o +obj-y += runtime_instr.o cache.o +endif # vdso obj-$(CONFIG_64BIT) += vdso64/ diff --git a/arch/s390/kernel/asm-offsets.c b/arch/s390/kernel/asm-offsets.c index 45ef1a7b08f..fface87056e 100644 --- a/arch/s390/kernel/asm-offsets.c +++ b/arch/s390/kernel/asm-offsets.c @@ -157,6 +157,8 @@ int main(void) DEFINE(__LC_LAST_BREAK, offsetof(struct _lowcore, breaking_event_addr)); DEFINE(__LC_VDSO_PER_CPU, offsetof(struct _lowcore, vdso_per_cpu_data)); DEFINE(__LC_GMAP, offsetof(struct _lowcore, gmap)); + DEFINE(__LC_PGM_TDB, offsetof(struct _lowcore, pgm_tdb)); + DEFINE(__THREAD_trap_tdb, offsetof(struct task_struct, thread.trap_tdb)); DEFINE(__GMAP_ASCE, offsetof(struct gmap, asce)); #endif /* CONFIG_32BIT */ return 0; diff --git a/arch/s390/kernel/cache.c b/arch/s390/kernel/cache.c new file mode 100644 index 00000000000..8df8d8a19c9 --- /dev/null +++ b/arch/s390/kernel/cache.c @@ -0,0 +1,385 @@ +/* + * Extract CPU cache information and expose them via sysfs. + * + * Copyright IBM Corp. 2012 + * Author(s): Heiko Carstens <heiko.carstens@de.ibm.com> + */ + +#include <linux/notifier.h> +#include <linux/seq_file.h> +#include <linux/init.h> +#include <linux/list.h> +#include <linux/slab.h> +#include <linux/cpu.h> +#include <asm/facility.h> + +struct cache { + unsigned long size; + unsigned int line_size; + unsigned int associativity; + unsigned int nr_sets; + unsigned int level : 3; + unsigned int type : 2; + unsigned int private : 1; + struct list_head list; +}; + +struct cache_dir { + struct kobject *kobj; + struct cache_index_dir *index; +}; + +struct cache_index_dir { + struct kobject kobj; + int cpu; + struct cache *cache; + struct cache_index_dir *next; +}; + +enum { + CACHE_SCOPE_NOTEXISTS, + CACHE_SCOPE_PRIVATE, + CACHE_SCOPE_SHARED, + CACHE_SCOPE_RESERVED, +}; + +enum { + CACHE_TYPE_SEPARATE, + CACHE_TYPE_DATA, + CACHE_TYPE_INSTRUCTION, + CACHE_TYPE_UNIFIED, +}; + +enum { + EXTRACT_TOPOLOGY, + EXTRACT_LINE_SIZE, + EXTRACT_SIZE, + EXTRACT_ASSOCIATIVITY, +}; + +enum { + CACHE_TI_UNIFIED = 0, + CACHE_TI_INSTRUCTION = 0, + CACHE_TI_DATA, +}; + +struct cache_info { + unsigned char : 4; + unsigned char scope : 2; + unsigned char type : 2; +}; + +#define CACHE_MAX_LEVEL 8 + +union cache_topology { + struct cache_info ci[CACHE_MAX_LEVEL]; + unsigned long long raw; +}; + +static const char * const cache_type_string[] = { + "Data", + "Instruction", + "Unified", +}; + +static struct cache_dir *cache_dir_cpu[NR_CPUS]; +static LIST_HEAD(cache_list); + +void show_cacheinfo(struct seq_file *m) +{ + struct cache *cache; + int index = 0; + + list_for_each_entry(cache, &cache_list, list) { + seq_printf(m, "cache%-11d: ", index); + seq_printf(m, "level=%d ", cache->level); + seq_printf(m, "type=%s ", cache_type_string[cache->type]); + seq_printf(m, "scope=%s ", cache->private ? "Private" : "Shared"); + seq_printf(m, "size=%luK ", cache->size >> 10); + seq_printf(m, "line_size=%u ", cache->line_size); + seq_printf(m, "associativity=%d", cache->associativity); + seq_puts(m, "\n"); + index++; + } +} + +static inline unsigned long ecag(int ai, int li, int ti) +{ + unsigned long cmd, val; + + cmd = ai << 4 | li << 1 | ti; + asm volatile(".insn rsy,0xeb000000004c,%0,0,0(%1)" /* ecag */ + : "=d" (val) : "a" (cmd)); + return val; +} + +static int __init cache_add(int level, int private, int type) +{ + struct cache *cache; + int ti; + + cache = kzalloc(sizeof(*cache), GFP_KERNEL); + if (!cache) + return -ENOMEM; + ti = type == CACHE_TYPE_DATA ? CACHE_TI_DATA : CACHE_TI_UNIFIED; + cache->size = ecag(EXTRACT_SIZE, level, ti); + cache->line_size = ecag(EXTRACT_LINE_SIZE, level, ti); + cache->associativity = ecag(EXTRACT_ASSOCIATIVITY, level, ti); + cache->nr_sets = cache->size / cache->associativity; + cache->nr_sets /= cache->line_size; + cache->private = private; + cache->level = level + 1; + cache->type = type - 1; + list_add_tail(&cache->list, &cache_list); + return 0; +} + +static void __init cache_build_info(void) +{ + struct cache *cache, *next; + union cache_topology ct; + int level, private, rc; + + ct.raw = ecag(EXTRACT_TOPOLOGY, 0, 0); + for (level = 0; level < CACHE_MAX_LEVEL; level++) { + switch (ct.ci[level].scope) { + case CACHE_SCOPE_NOTEXISTS: + case CACHE_SCOPE_RESERVED: + return; + case CACHE_SCOPE_SHARED: + private = 0; + break; + case CACHE_SCOPE_PRIVATE: + private = 1; + break; + } + if (ct.ci[level].type == CACHE_TYPE_SEPARATE) { + rc = cache_add(level, private, CACHE_TYPE_DATA); + rc |= cache_add(level, private, CACHE_TYPE_INSTRUCTION); + } else { + rc = cache_add(level, private, ct.ci[level].type); + } + if (rc) + goto error; + } + return; +error: + list_for_each_entry_safe(cache, next, &cache_list, list) { + list_del(&cache->list); + kfree(cache); + } +} + +static struct cache_dir *__cpuinit cache_create_cache_dir(int cpu) +{ + struct cache_dir *cache_dir; + struct kobject *kobj = NULL; + struct device *dev; + + dev = get_cpu_device(cpu); + if (!dev) + goto out; + kobj = kobject_create_and_add("cache", &dev->kobj); + if (!kobj) + goto out; + cache_dir = kzalloc(sizeof(*cache_dir), GFP_KERNEL); + if (!cache_dir) + goto out; + cache_dir->kobj = kobj; + cache_dir_cpu[cpu] = cache_dir; + return cache_dir; +out: + kobject_put(kobj); + return NULL; +} + +static struct cache_index_dir *kobj_to_cache_index_dir(struct kobject *kobj) +{ + return container_of(kobj, struct cache_index_dir, kobj); +} + +static void cache_index_release(struct kobject *kobj) +{ + struct cache_index_dir *index; + + index = kobj_to_cache_index_dir(kobj); + kfree(index); +} + +static ssize_t cache_index_show(struct kobject *kobj, + struct attribute *attr, char *buf) +{ + struct kobj_attribute *kobj_attr; + + kobj_attr = container_of(attr, struct kobj_attribute, attr); + return kobj_attr->show(kobj, kobj_attr, buf); +} + +#define DEFINE_CACHE_ATTR(_name, _format, _value) \ +static ssize_t cache_##_name##_show(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + char *buf) \ +{ \ + struct cache_index_dir *index; \ + \ + index = kobj_to_cache_index_dir(kobj); \ + return sprintf(buf, _format, _value); \ +} \ +static struct kobj_attribute cache_##_name##_attr = \ + __ATTR(_name, 0444, cache_##_name##_show, NULL); + +DEFINE_CACHE_ATTR(size, "%luK\n", index->cache->size >> 10); +DEFINE_CACHE_ATTR(coherency_line_size, "%u\n", index->cache->line_size); +DEFINE_CACHE_ATTR(number_of_sets, "%u\n", index->cache->nr_sets); +DEFINE_CACHE_ATTR(ways_of_associativity, "%u\n", index->cache->associativity); +DEFINE_CACHE_ATTR(type, "%s\n", cache_type_string[index->cache->type]); +DEFINE_CACHE_ATTR(level, "%d\n", index->cache->level); + +static ssize_t shared_cpu_map_func(struct kobject *kobj, int type, char *buf) +{ + struct cache_index_dir *index; + int len; + + index = kobj_to_cache_index_dir(kobj); + len = type ? + cpulist_scnprintf(buf, PAGE_SIZE - 2, cpumask_of(index->cpu)) : + cpumask_scnprintf(buf, PAGE_SIZE - 2, cpumask_of(index->cpu)); + len += sprintf(&buf[len], "\n"); + return len; +} + +static ssize_t shared_cpu_map_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return shared_cpu_map_func(kobj, 0, buf); +} +static struct kobj_attribute cache_shared_cpu_map_attr = + __ATTR(shared_cpu_map, 0444, shared_cpu_map_show, NULL); + +static ssize_t shared_cpu_list_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return shared_cpu_map_func(kobj, 1, buf); +} +static struct kobj_attribute cache_shared_cpu_list_attr = + __ATTR(shared_cpu_list, 0444, shared_cpu_list_show, NULL); + +static struct attribute *cache_index_default_attrs[] = { + &cache_type_attr.attr, + &cache_size_attr.attr, + &cache_number_of_sets_attr.attr, + &cache_ways_of_associativity_attr.attr, + &cache_level_attr.attr, + &cache_coherency_line_size_attr.attr, + &cache_shared_cpu_map_attr.attr, + &cache_shared_cpu_list_attr.attr, + NULL, +}; + +static const struct sysfs_ops cache_index_ops = { + .show = cache_index_show, +}; + +static struct kobj_type cache_index_type = { + .sysfs_ops = &cache_index_ops, + .release = cache_index_release, + .default_attrs = cache_index_default_attrs, +}; + +static int __cpuinit cache_create_index_dir(struct cache_dir *cache_dir, + struct cache *cache, int index, + int cpu) +{ + struct cache_index_dir *index_dir; + int rc; + + index_dir = kzalloc(sizeof(*index_dir), GFP_KERNEL); + if (!index_dir) + return -ENOMEM; + index_dir->cache = cache; + index_dir->cpu = cpu; + rc = kobject_init_and_add(&index_dir->kobj, &cache_index_type, + cache_dir->kobj, "index%d", index); + if (rc) + goto out; + index_dir->next = cache_dir->index; + cache_dir->index = index_dir; + return 0; +out: + kfree(index_dir); + return rc; +} + +static int __cpuinit cache_add_cpu(int cpu) +{ + struct cache_dir *cache_dir; + struct cache *cache; + int rc, index = 0; + + if (list_empty(&cache_list)) + return 0; + cache_dir = cache_create_cache_dir(cpu); + if (!cache_dir) + return -ENOMEM; + list_for_each_entry(cache, &cache_list, list) { + if (!cache->private) + break; + rc = cache_create_index_dir(cache_dir, cache, index, cpu); + if (rc) + return rc; + index++; + } + return 0; +} + +static void __cpuinit cache_remove_cpu(int cpu) +{ + struct cache_index_dir *index, *next; + struct cache_dir *cache_dir; + + cache_dir = cache_dir_cpu[cpu]; + if (!cache_dir) + return; + index = cache_dir->index; + while (index) { + next = index->next; + kobject_put(&index->kobj); + index = next; + } + kobject_put(cache_dir->kobj); + kfree(cache_dir); + cache_dir_cpu[cpu] = NULL; +} + +static int __cpuinit cache_hotplug(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + int cpu = (long)hcpu; + int rc = 0; + + switch (action & ~CPU_TASKS_FROZEN) { + case CPU_ONLINE: + rc = cache_add_cpu(cpu); + if (rc) + cache_remove_cpu(cpu); + break; + case CPU_DEAD: + cache_remove_cpu(cpu); + break; + } + return rc ? NOTIFY_BAD : NOTIFY_OK; +} + +static int __init cache_init(void) +{ + int cpu; + + if (!test_facility(34)) + return 0; + cache_build_info(); + for_each_online_cpu(cpu) + cache_add_cpu(cpu); + hotcpu_notifier(cache_hotplug, 0); + return 0; +} +device_initcall(cache_init); diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S index 2d82cfcbce5..3afba804fe9 100644 --- a/arch/s390/kernel/compat_wrapper.S +++ b/arch/s390/kernel/compat_wrapper.S @@ -1646,3 +1646,16 @@ ENTRY(compat_sys_process_vm_writev_wrapper) llgf %r0,164(%r15) # unsigned long stg %r0,160(%r15) jg compat_sys_process_vm_writev + +ENTRY(sys_s390_runtime_instr_wrapper) + lgfr %r2,%r2 # int + lgfr %r3,%r3 # int + jg sys_s390_runtime_instr + +ENTRY(sys_kcmp_wrapper) + lgfr %r2,%r2 # pid_t + lgfr %r3,%r3 # pid_t + lgfr %r4,%r4 # int + llgfr %r5,%r5 # unsigned long + llgfr %r6,%r6 # unsigned long + jg sys_kcmp diff --git a/arch/s390/kernel/crash.c b/arch/s390/kernel/crash.c deleted file mode 100644 index 3819153de8b..00000000000 --- a/arch/s390/kernel/crash.c +++ /dev/null @@ -1,14 +0,0 @@ -/* - * Copyright IBM Corp. 2005 - * - * Author(s): Heiko Carstens <heiko.carstens@de.ibm.com> - * - */ - -#include <linux/threads.h> -#include <linux/kexec.h> -#include <linux/reboot.h> - -void machine_crash_shutdown(struct pt_regs *regs) -{ -} diff --git a/arch/s390/kernel/crash_dump.c b/arch/s390/kernel/crash_dump.c index cc1172b2687..fb8d8781a01 100644 --- a/arch/s390/kernel/crash_dump.c +++ b/arch/s390/kernel/crash_dump.c @@ -13,8 +13,9 @@ #include <linux/slab.h> #include <linux/bootmem.h> #include <linux/elf.h> -#include <asm/ipl.h> #include <asm/os_info.h> +#include <asm/elf.h> +#include <asm/ipl.h> #define PTR_ADD(x, y) (((char *) (x)) + ((unsigned long) (y))) #define PTR_SUB(x, y) (((char *) (x)) - ((unsigned long) (y))) diff --git a/arch/s390/kernel/dis.c b/arch/s390/kernel/dis.c index 619c5d35072..cc84a24c023 100644 --- a/arch/s390/kernel/dis.c +++ b/arch/s390/kernel/dis.c @@ -315,6 +315,11 @@ enum { LONG_INSN_POPCNT, LONG_INSN_RISBHG, LONG_INSN_RISBLG, + LONG_INSN_RINEXT, + LONG_INSN_RIEMIT, + LONG_INSN_TABORT, + LONG_INSN_TBEGIN, + LONG_INSN_TBEGINC, }; static char *long_insn_name[] = { @@ -329,7 +334,12 @@ static char *long_insn_name[] = { [LONG_INSN_LLGHRL] = "llghrl", [LONG_INSN_POPCNT] = "popcnt", [LONG_INSN_RISBHG] = "risbhg", - [LONG_INSN_RISBLG] = "risblk", + [LONG_INSN_RISBLG] = "risblg", + [LONG_INSN_RINEXT] = "rinext", + [LONG_INSN_RIEMIT] = "riemit", + [LONG_INSN_TABORT] = "tabort", + [LONG_INSN_TBEGIN] = "tbegin", + [LONG_INSN_TBEGINC] = "tbeginc", }; static struct insn opcode[] = { @@ -582,6 +592,17 @@ static struct insn opcode_a7[] = { { "", 0, INSTR_INVALID } }; +static struct insn opcode_aa[] = { +#ifdef CONFIG_64BIT + { { 0, LONG_INSN_RINEXT }, 0x00, INSTR_RI_RI }, + { "rion", 0x01, INSTR_RI_RI }, + { "tric", 0x02, INSTR_RI_RI }, + { "rioff", 0x03, INSTR_RI_RI }, + { { 0, LONG_INSN_RIEMIT }, 0x04, INSTR_RI_RI }, +#endif + { "", 0, INSTR_INVALID } +}; + static struct insn opcode_b2[] = { #ifdef CONFIG_64BIT { "sske", 0x2b, INSTR_RRF_M0RR }, @@ -594,6 +615,9 @@ static struct insn opcode_b2[] = { { "lpswe", 0xb2, INSTR_S_RD }, { "srnmt", 0xb9, INSTR_S_RD }, { "lfas", 0xbd, INSTR_S_RD }, + { "etndg", 0xec, INSTR_RRE_R0 }, + { { 0, LONG_INSN_TABORT }, 0xfc, INSTR_S_RD }, + { "tend", 0xf8, INSTR_S_RD }, #endif { "stidp", 0x02, INSTR_S_RD }, { "sck", 0x04, INSTR_S_RD }, @@ -1150,6 +1174,7 @@ static struct insn opcode_e3[] = { { "stfh", 0xcb, INSTR_RXY_RRRD }, { "chf", 0xcd, INSTR_RXY_RRRD }, { "clhf", 0xcf, INSTR_RXY_RRRD }, + { "ntstg", 0x25, INSTR_RXY_RRRD }, #endif { "lrv", 0x1e, INSTR_RXY_RRRD }, { "lrvh", 0x1f, INSTR_RXY_RRRD }, @@ -1173,6 +1198,8 @@ static struct insn opcode_e5[] = { { "mvhhi", 0x44, INSTR_SIL_RDI }, { "mvhi", 0x4c, INSTR_SIL_RDI }, { "mvghi", 0x48, INSTR_SIL_RDI }, + { { 0, LONG_INSN_TBEGIN }, 0x60, INSTR_SIL_RDU }, + { { 0, LONG_INSN_TBEGINC }, 0x61, INSTR_SIL_RDU }, #endif { "lasp", 0x00, INSTR_SSE_RDRD }, { "tprot", 0x01, INSTR_SSE_RDRD }, @@ -1210,6 +1237,9 @@ static struct insn opcode_eb[] = { { "cliy", 0x55, INSTR_SIY_URD }, { "oiy", 0x56, INSTR_SIY_URD }, { "xiy", 0x57, INSTR_SIY_URD }, + { "lric", 0x60, INSTR_RSY_RDRM }, + { "stric", 0x61, INSTR_RSY_RDRM }, + { "mric", 0x62, INSTR_RSY_RDRM }, { "icmh", 0x80, INSTR_RSE_RURD }, { "icmh", 0x80, INSTR_RSY_RURD }, { "icmy", 0x81, INSTR_RSY_RURD }, @@ -1408,6 +1438,9 @@ static struct insn *find_insn(unsigned char *code) case 0xa7: table = opcode_a7; break; + case 0xaa: + table = opcode_aa; + break; case 0xb2: table = opcode_b2; break; @@ -1601,3 +1634,26 @@ void show_code(struct pt_regs *regs) } printk("\n"); } + +void print_fn_code(unsigned char *code, unsigned long len) +{ + char buffer[64], *ptr; + int opsize, i; + + while (len) { + ptr = buffer; + opsize = insn_length(*code); + ptr += sprintf(ptr, "%p: ", code); + for (i = 0; i < opsize; i++) + ptr += sprintf(ptr, "%02x", code[i]); + *ptr++ = '\t'; + if (i < 4) + *ptr++ = '\t'; + ptr += print_insn(ptr, code, (unsigned long) code); + *ptr++ = '\n'; + *ptr++ = 0; + printk(buffer); + code += opsize; + len -= opsize; + } +} diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c index 83c3271c442..7f4717675c1 100644 --- a/arch/s390/kernel/early.c +++ b/arch/s390/kernel/early.c @@ -215,36 +215,54 @@ static noinline __init void init_kernel_storage_key(void) PAGE_DEFAULT_KEY, 0); } -static __initdata struct sysinfo_3_2_2 vmms __aligned(PAGE_SIZE); +static __initdata char sysinfo_page[PAGE_SIZE] __aligned(PAGE_SIZE); static noinline __init void detect_machine_type(void) { + struct sysinfo_3_2_2 *vmms = (struct sysinfo_3_2_2 *)&sysinfo_page; + /* Check current-configuration-level */ - if ((stsi(NULL, 0, 0, 0) >> 28) <= 2) { + if (stsi(NULL, 0, 0, 0) <= 2) { S390_lowcore.machine_flags |= MACHINE_FLAG_LPAR; return; } /* Get virtual-machine cpu information. */ - if (stsi(&vmms, 3, 2, 2) == -ENOSYS || !vmms.count) + if (stsi(vmms, 3, 2, 2) || !vmms->count) return; /* Running under KVM? If not we assume z/VM */ - if (!memcmp(vmms.vm[0].cpi, "\xd2\xe5\xd4", 3)) + if (!memcmp(vmms->vm[0].cpi, "\xd2\xe5\xd4", 3)) S390_lowcore.machine_flags |= MACHINE_FLAG_KVM; else S390_lowcore.machine_flags |= MACHINE_FLAG_VM; } +static __init void setup_topology(void) +{ +#ifdef CONFIG_64BIT + int max_mnest; + + if (!test_facility(11)) + return; + S390_lowcore.machine_flags |= MACHINE_FLAG_TOPOLOGY; + for (max_mnest = 6; max_mnest > 1; max_mnest--) { + if (stsi(&sysinfo_page, 15, 1, max_mnest) == 0) + break; + } + topology_max_mnest = max_mnest; +#endif +} + static void early_pgm_check_handler(void) { - unsigned long addr; const struct exception_table_entry *fixup; + unsigned long addr; addr = S390_lowcore.program_old_psw.addr; fixup = search_exception_tables(addr & PSW_ADDR_INSN); if (!fixup) disabled_wait(0); - S390_lowcore.program_old_psw.addr = fixup->fixup | PSW_ADDR_AMODE; + S390_lowcore.program_old_psw.addr = extable_fixup(fixup)|PSW_ADDR_AMODE; } static noinline __init void setup_lowcore_early(void) @@ -267,12 +285,10 @@ static noinline __init void setup_facility_list(void) static noinline __init void setup_hpage(void) { -#ifndef CONFIG_DEBUG_PAGEALLOC if (!test_facility(2) || !test_facility(8)) return; S390_lowcore.machine_flags |= MACHINE_FLAG_HPAGE; __ctl_set_bit(0, 23); -#endif } static __init void detect_mvpg(void) @@ -366,12 +382,12 @@ static __init void detect_machine_facilities(void) S390_lowcore.machine_flags |= MACHINE_FLAG_IDTE; if (test_facility(8)) S390_lowcore.machine_flags |= MACHINE_FLAG_PFMF; - if (test_facility(11)) - S390_lowcore.machine_flags |= MACHINE_FLAG_TOPOLOGY; if (test_facility(27)) S390_lowcore.machine_flags |= MACHINE_FLAG_MVCOS; if (test_facility(40)) S390_lowcore.machine_flags |= MACHINE_FLAG_SPP; + if (test_facility(50) && test_facility(73)) + S390_lowcore.machine_flags |= MACHINE_FLAG_TE; #endif } @@ -441,7 +457,6 @@ static void __init setup_boot_command_line(void) append_to_cmdline(append_ipl_scpdata); } - /* * Save ipl parameters, clear bss memory, initialize storage keys * and create a kernel NSS at startup if the SAVESYS= parm is defined @@ -468,6 +483,7 @@ void __init startup_init(void) detect_diag44(); detect_machine_facilities(); setup_hpage(); + setup_topology(); sclp_facilities_detect(); detect_memory_layout(memory_chunk); #ifdef CONFIG_DYNAMIC_FTRACE diff --git a/arch/s390/kernel/entry64.S b/arch/s390/kernel/entry64.S index 349b7eeb348..7549985402f 100644 --- a/arch/s390/kernel/entry64.S +++ b/arch/s390/kernel/entry64.S @@ -10,6 +10,7 @@ #include <linux/init.h> #include <linux/linkage.h> +#include <asm/processor.h> #include <asm/cache.h> #include <asm/errno.h> #include <asm/ptrace.h> @@ -412,6 +413,11 @@ ENTRY(pgm_check_handler) 1: UPDATE_VTIME %r14,__LC_SYNC_ENTER_TIMER LAST_BREAK %r14 lg %r15,__LC_KERNEL_STACK + lg %r14,__TI_task(%r12) + lghi %r13,__LC_PGM_TDB + tm __LC_PGM_ILC+2,0x02 # check for transaction abort + jz 2f + mvc __THREAD_trap_tdb(256,%r14),0(%r13) 2: aghi %r15,-(STACK_FRAME_OVERHEAD + __PT_SIZE) la %r11,STACK_FRAME_OVERHEAD(%r15) stmg %r0,%r7,__PT_R0(%r11) @@ -422,13 +428,12 @@ ENTRY(pgm_check_handler) stg %r10,__PT_ARGS(%r11) tm __LC_PGM_ILC+3,0x80 # check for per exception jz 0f - lg %r1,__TI_task(%r12) tmhh %r8,0x0001 # kernel per event ? jz pgm_kprobe oi __TI_flags+7(%r12),_TIF_PER_TRAP - mvc __THREAD_per_address(8,%r1),__LC_PER_ADDRESS - mvc __THREAD_per_cause(2,%r1),__LC_PER_CAUSE - mvc __THREAD_per_paid(1,%r1),__LC_PER_PAID + mvc __THREAD_per_address(8,%r14),__LC_PER_ADDRESS + mvc __THREAD_per_cause(2,%r14),__LC_PER_CAUSE + mvc __THREAD_per_paid(1,%r14),__LC_PER_PAID 0: REENABLE_IRQS xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) larl %r1,pgm_check_table @@ -1004,9 +1009,7 @@ sie_fault: .Lhost_id: .quad 0 - .section __ex_table,"a" - .quad sie_loop,sie_fault - .previous + EX_TABLE(sie_loop,sie_fault) #endif .section .rodata, "a" diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c index dd7630d8aab..6cdc55b26d6 100644 --- a/arch/s390/kernel/irq.c +++ b/arch/s390/kernel/irq.c @@ -30,33 +30,35 @@ struct irq_class { }; static const struct irq_class intrclass_names[] = { - {.name = "EXT" }, - {.name = "I/O" }, - {.name = "CLK", .desc = "[EXT] Clock Comparator" }, - {.name = "EXC", .desc = "[EXT] External Call" }, - {.name = "EMS", .desc = "[EXT] Emergency Signal" }, - {.name = "TMR", .desc = "[EXT] CPU Timer" }, - {.name = "TAL", .desc = "[EXT] Timing Alert" }, - {.name = "PFL", .desc = "[EXT] Pseudo Page Fault" }, - {.name = "DSD", .desc = "[EXT] DASD Diag" }, - {.name = "VRT", .desc = "[EXT] Virtio" }, - {.name = "SCP", .desc = "[EXT] Service Call" }, - {.name = "IUC", .desc = "[EXT] IUCV" }, - {.name = "CMS", .desc = "[EXT] CPU-Measurement: Sampling" }, - {.name = "CMC", .desc = "[EXT] CPU-Measurement: Counter" }, - {.name = "CIO", .desc = "[I/O] Common I/O Layer Interrupt" }, - {.name = "QAI", .desc = "[I/O] QDIO Adapter Interrupt" }, - {.name = "DAS", .desc = "[I/O] DASD" }, - {.name = "C15", .desc = "[I/O] 3215" }, - {.name = "C70", .desc = "[I/O] 3270" }, - {.name = "TAP", .desc = "[I/O] Tape" }, - {.name = "VMR", .desc = "[I/O] Unit Record Devices" }, - {.name = "LCS", .desc = "[I/O] LCS" }, - {.name = "CLW", .desc = "[I/O] CLAW" }, - {.name = "CTC", .desc = "[I/O] CTC" }, - {.name = "APB", .desc = "[I/O] AP Bus" }, - {.name = "CSC", .desc = "[I/O] CHSC Subchannel" }, - {.name = "NMI", .desc = "[NMI] Machine Check" }, + [EXTERNAL_INTERRUPT] = {.name = "EXT"}, + [IO_INTERRUPT] = {.name = "I/O"}, + [EXTINT_CLK] = {.name = "CLK", .desc = "[EXT] Clock Comparator"}, + [EXTINT_EXC] = {.name = "EXC", .desc = "[EXT] External Call"}, + [EXTINT_EMS] = {.name = "EMS", .desc = "[EXT] Emergency Signal"}, + [EXTINT_TMR] = {.name = "TMR", .desc = "[EXT] CPU Timer"}, + [EXTINT_TLA] = {.name = "TAL", .desc = "[EXT] Timing Alert"}, + [EXTINT_PFL] = {.name = "PFL", .desc = "[EXT] Pseudo Page Fault"}, + [EXTINT_DSD] = {.name = "DSD", .desc = "[EXT] DASD Diag"}, + [EXTINT_VRT] = {.name = "VRT", .desc = "[EXT] Virtio"}, + [EXTINT_SCP] = {.name = "SCP", .desc = "[EXT] Service Call"}, + [EXTINT_IUC] = {.name = "IUC", .desc = "[EXT] IUCV"}, + [EXTINT_CMS] = {.name = "CMS", .desc = "[EXT] CPU-Measurement: Sampling"}, + [EXTINT_CMC] = {.name = "CMC", .desc = "[EXT] CPU-Measurement: Counter"}, + [EXTINT_CMR] = {.name = "CMR", .desc = "[EXT] CPU-Measurement: RI"}, + [IOINT_CIO] = {.name = "CIO", .desc = "[I/O] Common I/O Layer Interrupt"}, + [IOINT_QAI] = {.name = "QAI", .desc = "[I/O] QDIO Adapter Interrupt"}, + [IOINT_DAS] = {.name = "DAS", .desc = "[I/O] DASD"}, + [IOINT_C15] = {.name = "C15", .desc = "[I/O] 3215"}, + [IOINT_C70] = {.name = "C70", .desc = "[I/O] 3270"}, + [IOINT_TAP] = {.name = "TAP", .desc = "[I/O] Tape"}, + [IOINT_VMR] = {.name = "VMR", .desc = "[I/O] Unit Record Devices"}, + [IOINT_LCS] = {.name = "LCS", .desc = "[I/O] LCS"}, + [IOINT_CLW] = {.name = "CLW", .desc = "[I/O] CLAW"}, + [IOINT_CTC] = {.name = "CTC", .desc = "[I/O] CTC"}, + [IOINT_APB] = {.name = "APB", .desc = "[I/O] AP Bus"}, + [IOINT_ADM] = {.name = "ADM", .desc = "[I/O] EADM Subchannel"}, + [IOINT_CSC] = {.name = "CSC", .desc = "[I/O] CHSC Subchannel"}, + [NMI_NMI] = {.name = "NMI", .desc = "[NMI] Machine Check"}, }; /* diff --git a/arch/s390/kernel/kprobes.c b/arch/s390/kernel/kprobes.c index 8aa634f5944..d1c7214e157 100644 --- a/arch/s390/kernel/kprobes.c +++ b/arch/s390/kernel/kprobes.c @@ -547,7 +547,7 @@ static int __kprobes kprobe_trap_handler(struct pt_regs *regs, int trapnr) */ entry = search_exception_tables(regs->psw.addr & PSW_ADDR_INSN); if (entry) { - regs->psw.addr = entry->fixup | PSW_ADDR_AMODE; + regs->psw.addr = extable_fixup(entry) | PSW_ADDR_AMODE; return 1; } diff --git a/arch/s390/kernel/lgr.c b/arch/s390/kernel/lgr.c index eca94e74d19..6ea6d69339b 100644 --- a/arch/s390/kernel/lgr.c +++ b/arch/s390/kernel/lgr.c @@ -51,16 +51,6 @@ static struct lgr_info lgr_info_cur; static struct debug_info *lgr_dbf; /* - * Return number of valid stsi levels - */ -static inline int stsi_0(void) -{ - int rc = stsi(NULL, 0, 0, 0); - - return rc == -ENOSYS ? rc : (((unsigned int) rc) >> 28); -} - -/* * Copy buffer and then convert it to ASCII */ static void cpascii(char *dst, char *src, int size) @@ -76,7 +66,7 @@ static void lgr_stsi_1_1_1(struct lgr_info *lgr_info) { struct sysinfo_1_1_1 *si = (void *) lgr_page; - if (stsi(si, 1, 1, 1) == -ENOSYS) + if (stsi(si, 1, 1, 1)) return; cpascii(lgr_info->manufacturer, si->manufacturer, sizeof(si->manufacturer)); @@ -93,7 +83,7 @@ static void lgr_stsi_2_2_2(struct lgr_info *lgr_info) { struct sysinfo_2_2_2 *si = (void *) lgr_page; - if (stsi(si, 2, 2, 2) == -ENOSYS) + if (stsi(si, 2, 2, 2)) return; cpascii(lgr_info->name, si->name, sizeof(si->name)); memcpy(&lgr_info->lpar_number, &si->lpar_number, @@ -108,7 +98,7 @@ static void lgr_stsi_3_2_2(struct lgr_info *lgr_info) struct sysinfo_3_2_2 *si = (void *) lgr_page; int i; - if (stsi(si, 3, 2, 2) == -ENOSYS) + if (stsi(si, 3, 2, 2)) return; for (i = 0; i < min_t(u8, si->count, VM_LEVEL_MAX); i++) { cpascii(lgr_info->vm[i].name, si->vm[i].name, @@ -124,16 +114,17 @@ static void lgr_stsi_3_2_2(struct lgr_info *lgr_info) */ static void lgr_info_get(struct lgr_info *lgr_info) { + int level; + memset(lgr_info, 0, sizeof(*lgr_info)); stfle(lgr_info->stfle_fac_list, ARRAY_SIZE(lgr_info->stfle_fac_list)); - lgr_info->level = stsi_0(); - if (lgr_info->level == -ENOSYS) - return; - if (lgr_info->level >= 1) + level = stsi(NULL, 0, 0, 0); + lgr_info->level = level; + if (level >= 1) lgr_stsi_1_1_1(lgr_info); - if (lgr_info->level >= 2) + if (level >= 2) lgr_stsi_2_2_2(lgr_info); - if (lgr_info->level >= 3) + if (level >= 3) lgr_stsi_3_2_2(lgr_info); } diff --git a/arch/s390/kernel/machine_kexec.c b/arch/s390/kernel/machine_kexec.c index 493304bdf1c..b3de2770001 100644 --- a/arch/s390/kernel/machine_kexec.c +++ b/arch/s390/kernel/machine_kexec.c @@ -21,6 +21,7 @@ #include <asm/reset.h> #include <asm/ipl.h> #include <asm/diag.h> +#include <asm/elf.h> #include <asm/asm-offsets.h> #include <asm/os_info.h> @@ -31,8 +32,6 @@ extern const unsigned long long relocate_kernel_len; #ifdef CONFIG_CRASH_DUMP -void *fill_cpu_elf_notes(void *ptr, struct save_area *sa); - /* * Create ELF notes for one CPU */ @@ -159,7 +158,7 @@ int machine_kexec_prepare(struct kimage *image) /* Can't replace kernel image since it is read-only. */ if (ipl_flags & IPL_NSS_VALID) - return -ENOSYS; + return -EOPNOTSUPP; if (image->type == KEXEC_TYPE_CRASH) return machine_kexec_prepare_kdump(); @@ -191,6 +190,10 @@ void machine_shutdown(void) { } +void machine_crash_shutdown(struct pt_regs *regs) +{ +} + /* * Do normal kexec */ diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c index 733175373a4..5024be27df4 100644 --- a/arch/s390/kernel/process.c +++ b/arch/s390/kernel/process.c @@ -26,10 +26,12 @@ #include <asm/io.h> #include <asm/processor.h> #include <asm/vtimer.h> +#include <asm/exec.h> #include <asm/irq.h> #include <asm/nmi.h> #include <asm/smp.h> #include <asm/switch_to.h> +#include <asm/runtime_instr.h> #include "entry.h" asmlinkage void ret_from_fork(void) asm ("ret_from_fork"); @@ -132,6 +134,7 @@ EXPORT_SYMBOL(kernel_thread); */ void exit_thread(void) { + exit_thread_runtime_instr(); } void flush_thread(void) @@ -170,6 +173,11 @@ int copy_thread(unsigned long clone_flags, unsigned long new_stackp, /* Save access registers to new thread structure. */ save_access_regs(&p->thread.acrs[0]); + /* Don't copy runtime instrumentation info */ + p->thread.ri_cb = NULL; + p->thread.ri_signum = 0; + frame->childregs.psw.mask &= ~PSW_MASK_RI; + #ifndef CONFIG_64BIT /* * save fprs to current->thread.fp_regs to merge them with diff --git a/arch/s390/kernel/processor.c b/arch/s390/kernel/processor.c index 572d4c9cb33..753c41d0ffd 100644 --- a/arch/s390/kernel/processor.c +++ b/arch/s390/kernel/processor.c @@ -39,9 +39,9 @@ void __cpuinit cpu_init(void) */ static int show_cpuinfo(struct seq_file *m, void *v) { - static const char *hwcap_str[10] = { + static const char *hwcap_str[] = { "esan3", "zarch", "stfle", "msa", "ldisp", "eimm", "dfp", - "edat", "etf3eh", "highgprs" + "edat", "etf3eh", "highgprs", "te" }; unsigned long n = (unsigned long) v - 1; int i; @@ -54,10 +54,11 @@ static int show_cpuinfo(struct seq_file *m, void *v) num_online_cpus(), loops_per_jiffy/(500000/HZ), (loops_per_jiffy/(5000/HZ))%100); seq_puts(m, "features\t: "); - for (i = 0; i < 10; i++) + for (i = 0; i < ARRAY_SIZE(hwcap_str); i++) if (hwcap_str[i] && (elf_hwcap & (1UL << i))) seq_printf(m, "%s ", hwcap_str[i]); seq_puts(m, "\n"); + show_cacheinfo(m); } get_online_cpus(); if (cpu_online(n)) { diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c index e4be113fbac..a314c57f4e9 100644 --- a/arch/s390/kernel/ptrace.c +++ b/arch/s390/kernel/ptrace.c @@ -42,6 +42,7 @@ enum s390_regset { REGSET_GENERAL, REGSET_FP, REGSET_LAST_BREAK, + REGSET_TDB, REGSET_SYSTEM_CALL, REGSET_GENERAL_EXTENDED, }; @@ -52,6 +53,22 @@ void update_per_regs(struct task_struct *task) struct thread_struct *thread = &task->thread; struct per_regs old, new; +#ifdef CONFIG_64BIT + /* Take care of the enable/disable of transactional execution. */ + if (MACHINE_HAS_TE) { + unsigned long cr0, cr0_new; + + __ctl_store(cr0, 0, 0); + /* set or clear transaction execution bits 8 and 9. */ + if (task->thread.per_flags & PER_FLAG_NO_TE) + cr0_new = cr0 & ~(3UL << 54); + else + cr0_new = cr0 | (3UL << 54); + /* Only load control register 0 if necessary. */ + if (cr0 != cr0_new) + __ctl_load(cr0_new, 0, 0); + } +#endif /* Copy user specified PER registers */ new.control = thread->per_user.control; new.start = thread->per_user.start; @@ -60,6 +77,10 @@ void update_per_regs(struct task_struct *task) /* merge TIF_SINGLE_STEP into user specified PER registers. */ if (test_tsk_thread_flag(task, TIF_SINGLE_STEP)) { new.control |= PER_EVENT_IFETCH; +#ifdef CONFIG_64BIT + new.control |= PER_CONTROL_SUSPENSION; + new.control |= PER_EVENT_TRANSACTION_END; +#endif new.start = 0; new.end = PSW_ADDR_INSN; } @@ -100,6 +121,7 @@ void ptrace_disable(struct task_struct *task) memset(&task->thread.per_event, 0, sizeof(task->thread.per_event)); clear_tsk_thread_flag(task, TIF_SINGLE_STEP); clear_tsk_thread_flag(task, TIF_PER_TRAP); + task->thread.per_flags = 0; } #ifndef CONFIG_64BIT @@ -416,6 +438,16 @@ long arch_ptrace(struct task_struct *child, long request, put_user(task_thread_info(child)->last_break, (unsigned long __user *) data); return 0; + case PTRACE_ENABLE_TE: + if (!MACHINE_HAS_TE) + return -EIO; + child->thread.per_flags &= ~PER_FLAG_NO_TE; + return 0; + case PTRACE_DISABLE_TE: + if (!MACHINE_HAS_TE) + return -EIO; + child->thread.per_flags |= PER_FLAG_NO_TE; + return 0; default: /* Removing high order bit from addr (only for 31 bit). */ addr &= PSW_ADDR_INSN; @@ -903,6 +935,28 @@ static int s390_last_break_set(struct task_struct *target, return 0; } +static int s390_tdb_get(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + void *kbuf, void __user *ubuf) +{ + struct pt_regs *regs = task_pt_regs(target); + unsigned char *data; + + if (!(regs->int_code & 0x200)) + return -ENODATA; + data = target->thread.trap_tdb; + return user_regset_copyout(&pos, &count, &kbuf, &ubuf, data, 0, 256); +} + +static int s390_tdb_set(struct task_struct *target, + const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + return 0; +} + #endif static int s390_system_call_get(struct task_struct *target, @@ -951,6 +1005,14 @@ static const struct user_regset s390_regsets[] = { .get = s390_last_break_get, .set = s390_last_break_set, }, + [REGSET_TDB] = { + .core_note_type = NT_S390_TDB, + .n = 1, + .size = 256, + .align = 1, + .get = s390_tdb_get, + .set = s390_tdb_set, + }, #endif [REGSET_SYSTEM_CALL] = { .core_note_type = NT_S390_SYSTEM_CALL, @@ -1148,6 +1210,14 @@ static const struct user_regset s390_compat_regsets[] = { .get = s390_compat_last_break_get, .set = s390_compat_last_break_set, }, + [REGSET_TDB] = { + .core_note_type = NT_S390_TDB, + .n = 1, + .size = 256, + .align = 1, + .get = s390_tdb_get, + .set = s390_tdb_set, + }, [REGSET_SYSTEM_CALL] = { .core_note_type = NT_S390_SYSTEM_CALL, .n = 1, diff --git a/arch/s390/kernel/runtime_instr.c b/arch/s390/kernel/runtime_instr.c new file mode 100644 index 00000000000..61066f6f71a --- /dev/null +++ b/arch/s390/kernel/runtime_instr.c @@ -0,0 +1,150 @@ +/* + * Copyright IBM Corp. 2012 + * Author(s): Jan Glauber <jang@linux.vnet.ibm.com> + */ + +#include <linux/kernel.h> +#include <linux/syscalls.h> +#include <linux/signal.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/init.h> +#include <linux/errno.h> +#include <linux/kernel_stat.h> +#include <asm/runtime_instr.h> +#include <asm/cpu_mf.h> +#include <asm/irq.h> + +/* empty control block to disable RI by loading it */ +struct runtime_instr_cb runtime_instr_empty_cb; + +static int runtime_instr_avail(void) +{ + return test_facility(64); +} + +static void disable_runtime_instr(void) +{ + struct pt_regs *regs = task_pt_regs(current); + + load_runtime_instr_cb(&runtime_instr_empty_cb); + + /* + * Make sure the RI bit is deleted from the PSW. If the user did not + * switch off RI before the system call the process will get a + * specification exception otherwise. + */ + regs->psw.mask &= ~PSW_MASK_RI; +} + +static void init_runtime_instr_cb(struct runtime_instr_cb *cb) +{ + cb->buf_limit = 0xfff; + if (s390_user_mode == HOME_SPACE_MODE) + cb->home_space = 1; + cb->int_requested = 1; + cb->pstate = 1; + cb->pstate_set_buf = 1; + cb->pstate_sample = 1; + cb->pstate_collect = 1; + cb->key = PAGE_DEFAULT_KEY; + cb->valid = 1; +} + +void exit_thread_runtime_instr(void) +{ + struct task_struct *task = current; + + if (!task->thread.ri_cb) + return; + disable_runtime_instr(); + kfree(task->thread.ri_cb); + task->thread.ri_signum = 0; + task->thread.ri_cb = NULL; +} + +static void runtime_instr_int_handler(struct ext_code ext_code, + unsigned int param32, unsigned long param64) +{ + struct siginfo info; + + if (!(param32 & CPU_MF_INT_RI_MASK)) + return; + + kstat_cpu(smp_processor_id()).irqs[EXTINT_CMR]++; + + if (!current->thread.ri_cb) + return; + if (current->thread.ri_signum < SIGRTMIN || + current->thread.ri_signum > SIGRTMAX) { + WARN_ON_ONCE(1); + return; + } + + memset(&info, 0, sizeof(info)); + info.si_signo = current->thread.ri_signum; + info.si_code = SI_QUEUE; + if (param32 & CPU_MF_INT_RI_BUF_FULL) + info.si_int = ENOBUFS; + else if (param32 & CPU_MF_INT_RI_HALTED) + info.si_int = ECANCELED; + else + return; /* unknown reason */ + + send_sig_info(current->thread.ri_signum, &info, current); +} + +SYSCALL_DEFINE2(s390_runtime_instr, int, command, int, signum) +{ + struct runtime_instr_cb *cb; + + if (!runtime_instr_avail()) + return -EOPNOTSUPP; + + if (command == S390_RUNTIME_INSTR_STOP) { + preempt_disable(); + exit_thread_runtime_instr(); + preempt_enable(); + return 0; + } + + if (command != S390_RUNTIME_INSTR_START || + (signum < SIGRTMIN || signum > SIGRTMAX)) + return -EINVAL; + + if (!current->thread.ri_cb) { + cb = kzalloc(sizeof(*cb), GFP_KERNEL); + if (!cb) + return -ENOMEM; + } else { + cb = current->thread.ri_cb; + memset(cb, 0, sizeof(*cb)); + } + + init_runtime_instr_cb(cb); + current->thread.ri_signum = signum; + + /* now load the control block to make it available */ + preempt_disable(); + current->thread.ri_cb = cb; + load_runtime_instr_cb(cb); + preempt_enable(); + return 0; +} + +static int __init runtime_instr_init(void) +{ + int rc; + + if (!runtime_instr_avail()) + return 0; + + measurement_alert_subclass_register(); + rc = register_external_interrupt(0x1407, runtime_instr_int_handler); + if (rc) + measurement_alert_subclass_unregister(); + else + pr_info("Runtime instrumentation facility initialized\n"); + return rc; +} +device_initcall(runtime_instr_init); diff --git a/arch/s390/kernel/s390_ksyms.c b/arch/s390/kernel/s390_ksyms.c index 57b536649b0..9bdbcef1da9 100644 --- a/arch/s390/kernel/s390_ksyms.c +++ b/arch/s390/kernel/s390_ksyms.c @@ -8,3 +8,5 @@ EXPORT_SYMBOL(_mcount); #if defined(CONFIG_KVM) || defined(CONFIG_KVM_MODULE) EXPORT_SYMBOL(sie64a); #endif +EXPORT_SYMBOL(memcpy); +EXPORT_SYMBOL(memset); diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c index f86c81e13c3..afa9fdba200 100644 --- a/arch/s390/kernel/setup.c +++ b/arch/s390/kernel/setup.c @@ -302,10 +302,10 @@ static int __init parse_vmalloc(char *arg) } early_param("vmalloc", parse_vmalloc); -unsigned int addressing_mode = HOME_SPACE_MODE; -EXPORT_SYMBOL_GPL(addressing_mode); +unsigned int s390_user_mode = PRIMARY_SPACE_MODE; +EXPORT_SYMBOL_GPL(s390_user_mode); -static int set_amode_primary(void) +static void __init set_user_mode_primary(void) { psw_kernel_bits = (psw_kernel_bits & ~PSW_MASK_ASC) | PSW_ASC_HOME; psw_user_bits = (psw_user_bits & ~PSW_MASK_ASC) | PSW_ASC_PRIMARY; @@ -313,48 +313,30 @@ static int set_amode_primary(void) psw32_user_bits = (psw32_user_bits & ~PSW32_MASK_ASC) | PSW32_ASC_PRIMARY; #endif - - if (MACHINE_HAS_MVCOS) { - memcpy(&uaccess, &uaccess_mvcos_switch, sizeof(uaccess)); - return 1; - } else { - memcpy(&uaccess, &uaccess_pt, sizeof(uaccess)); - return 0; - } -} - -/* - * Switch kernel/user addressing modes? - */ -static int __init early_parse_switch_amode(char *p) -{ - addressing_mode = PRIMARY_SPACE_MODE; - return 0; + uaccess = MACHINE_HAS_MVCOS ? uaccess_mvcos_switch : uaccess_pt; } -early_param("switch_amode", early_parse_switch_amode); static int __init early_parse_user_mode(char *p) { if (p && strcmp(p, "primary") == 0) - addressing_mode = PRIMARY_SPACE_MODE; + s390_user_mode = PRIMARY_SPACE_MODE; else if (!p || strcmp(p, "home") == 0) - addressing_mode = HOME_SPACE_MODE; + s390_user_mode = HOME_SPACE_MODE; else return 1; return 0; } early_param("user_mode", early_parse_user_mode); -static void setup_addressing_mode(void) +static void __init setup_addressing_mode(void) { - if (addressing_mode == PRIMARY_SPACE_MODE) { - if (set_amode_primary()) - pr_info("Address spaces switched, " - "mvcos available\n"); - else - pr_info("Address spaces switched, " - "mvcos not available\n"); - } + if (s390_user_mode != PRIMARY_SPACE_MODE) + return; + set_user_mode_primary(); + if (MACHINE_HAS_MVCOS) + pr_info("Address spaces switched, mvcos available\n"); + else + pr_info("Address spaces switched, mvcos not available\n"); } void *restart_stack __attribute__((__section__(".data"))); @@ -602,9 +584,7 @@ static void __init setup_memory_end(void) static void __init setup_vmcoreinfo(void) { -#ifdef CONFIG_KEXEC mem_assign_absolute(S390_lowcore.vmcore_info, paddr_vmcoreinfo_note()); -#endif } #ifdef CONFIG_CRASH_DUMP @@ -974,12 +954,20 @@ static void __init setup_hwcaps(void) if (MACHINE_HAS_HPAGE) elf_hwcap |= HWCAP_S390_HPAGE; +#if defined(CONFIG_64BIT) /* * 64-bit register support for 31-bit processes * HWCAP_S390_HIGH_GPRS is bit 9. */ elf_hwcap |= HWCAP_S390_HIGH_GPRS; + /* + * Transactional execution support HWCAP_S390_TE is bit 10. + */ + if (test_facility(50) && test_facility(73)) + elf_hwcap |= HWCAP_S390_TE; +#endif + get_cpu_id(&cpu_id); switch (cpu_id.machine) { case 0x9672: diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c index 720fda1620f..ea431e551c6 100644 --- a/arch/s390/kernel/smp.c +++ b/arch/s390/kernel/smp.c @@ -66,7 +66,7 @@ struct pcpu { unsigned long panic_stack; /* panic stack for the cpu */ unsigned long ec_mask; /* bit mask for ec_xxx functions */ int state; /* physical cpu state */ - u32 status; /* last status received via sigp */ + int polarization; /* physical polarization */ u16 address; /* physical cpu address */ }; @@ -74,6 +74,10 @@ static u8 boot_cpu_type; static u16 boot_cpu_address; static struct pcpu pcpu_devices[NR_CPUS]; +/* + * The smp_cpu_state_mutex must be held when changing the state or polarization + * member of a pcpu data structure within the pcpu_devices arreay. + */ DEFINE_MUTEX(smp_cpu_state_mutex); /* @@ -99,7 +103,7 @@ static inline int __pcpu_sigp_relax(u16 addr, u8 order, u32 parm, u32 *status) int cc; while (1) { - cc = __pcpu_sigp(addr, order, parm, status); + cc = __pcpu_sigp(addr, order, parm, NULL); if (cc != SIGP_CC_BUSY) return cc; cpu_relax(); @@ -111,7 +115,7 @@ static int pcpu_sigp_retry(struct pcpu *pcpu, u8 order, u32 parm) int cc, retry; for (retry = 0; ; retry++) { - cc = __pcpu_sigp(pcpu->address, order, parm, &pcpu->status); + cc = __pcpu_sigp(pcpu->address, order, parm, NULL); if (cc != SIGP_CC_BUSY) break; if (retry >= 3) @@ -122,16 +126,18 @@ static int pcpu_sigp_retry(struct pcpu *pcpu, u8 order, u32 parm) static inline int pcpu_stopped(struct pcpu *pcpu) { + u32 uninitialized_var(status); + if (__pcpu_sigp(pcpu->address, SIGP_SENSE, - 0, &pcpu->status) != SIGP_CC_STATUS_STORED) + 0, &status) != SIGP_CC_STATUS_STORED) return 0; - return !!(pcpu->status & (SIGP_STATUS_CHECK_STOP|SIGP_STATUS_STOPPED)); + return !!(status & (SIGP_STATUS_CHECK_STOP|SIGP_STATUS_STOPPED)); } static inline int pcpu_running(struct pcpu *pcpu) { if (__pcpu_sigp(pcpu->address, SIGP_SENSE_RUNNING, - 0, &pcpu->status) != SIGP_CC_STATUS_STORED) + 0, NULL) != SIGP_CC_STATUS_STORED) return 1; /* Status stored condition code is equivalent to cpu not running. */ return 0; @@ -586,6 +592,16 @@ static inline void smp_get_save_area(int cpu, u16 address) { } #endif /* CONFIG_ZFCPDUMP || CONFIG_CRASH_DUMP */ +void smp_cpu_set_polarization(int cpu, int val) +{ + pcpu_devices[cpu].polarization = val; +} + +int smp_cpu_get_polarization(int cpu) +{ + return pcpu_devices[cpu].polarization; +} + static struct sclp_cpu_info *smp_get_cpu_info(void) { static int use_sigp_detection; @@ -628,7 +644,7 @@ static int __devinit __smp_rescan_cpus(struct sclp_cpu_info *info, pcpu->address = info->cpu[i].address; pcpu->state = (cpu >= info->configured) ? CPU_STATE_STANDBY : CPU_STATE_CONFIGURED; - cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); + smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); set_cpu_present(cpu, true); if (sysfs_add && smp_add_present_cpu(cpu) != 0) set_cpu_present(cpu, false); @@ -796,7 +812,7 @@ void __init smp_prepare_boot_cpu(void) pcpu->async_stack = S390_lowcore.async_stack - ASYNC_SIZE; pcpu->panic_stack = S390_lowcore.panic_stack - PAGE_SIZE; S390_lowcore.percpu_offset = __per_cpu_offset[0]; - cpu_set_polarization(0, POLARIZATION_UNKNOWN); + smp_cpu_set_polarization(0, POLARIZATION_UNKNOWN); set_cpu_present(0, true); set_cpu_online(0, true); } @@ -862,7 +878,7 @@ static ssize_t cpu_configure_store(struct device *dev, if (rc) break; pcpu->state = CPU_STATE_STANDBY; - cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); + smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); topology_expect_change(); break; case 1: @@ -872,7 +888,7 @@ static ssize_t cpu_configure_store(struct device *dev, if (rc) break; pcpu->state = CPU_STATE_CONFIGURED; - cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); + smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); topology_expect_change(); break; default: @@ -959,23 +975,17 @@ static int __cpuinit smp_cpu_notify(struct notifier_block *self, struct device *s = &c->dev; int err = 0; - switch (action) { + switch (action & ~CPU_TASKS_FROZEN) { case CPU_ONLINE: - case CPU_ONLINE_FROZEN: err = sysfs_create_group(&s->kobj, &cpu_online_attr_group); break; case CPU_DEAD: - case CPU_DEAD_FROZEN: sysfs_remove_group(&s->kobj, &cpu_online_attr_group); break; } return notifier_from_errno(err); } -static struct notifier_block __cpuinitdata smp_cpu_nb = { - .notifier_call = smp_cpu_notify, -}; - static int __devinit smp_add_present_cpu(int cpu) { struct cpu *c = &pcpu_devices[cpu].cpu; @@ -1050,7 +1060,7 @@ static int __init s390_smp_init(void) { int cpu, rc; - register_cpu_notifier(&smp_cpu_nb); + hotcpu_notifier(smp_cpu_notify, 0); #ifdef CONFIG_HOTPLUG_CPU rc = device_create_file(cpu_subsys.dev_root, &dev_attr_rescan); if (rc) diff --git a/arch/s390/kernel/syscalls.S b/arch/s390/kernel/syscalls.S index bcab2f04ba5..48174850f3b 100644 --- a/arch/s390/kernel/syscalls.S +++ b/arch/s390/kernel/syscalls.S @@ -350,3 +350,5 @@ SYSCALL(sys_syncfs,sys_syncfs,sys_syncfs_wrapper) SYSCALL(sys_setns,sys_setns,sys_setns_wrapper) SYSCALL(sys_process_vm_readv,sys_process_vm_readv,compat_sys_process_vm_readv_wrapper) /* 340 */ SYSCALL(sys_process_vm_writev,sys_process_vm_writev,compat_sys_process_vm_writev_wrapper) +SYSCALL(sys_ni_syscall,sys_s390_runtime_instr,sys_s390_runtime_instr_wrapper) +SYSCALL(sys_kcmp,sys_kcmp,sys_kcmp_wrapper) diff --git a/arch/s390/kernel/sysinfo.c b/arch/s390/kernel/sysinfo.c index fa0eb238dac..62f89d98e88 100644 --- a/arch/s390/kernel/sysinfo.c +++ b/arch/s390/kernel/sysinfo.c @@ -22,17 +22,41 @@ #include <math-emu/soft-fp.h> #include <math-emu/single.h> -static inline int stsi_0(void) +int topology_max_mnest; + +/* + * stsi - store system information + * + * Returns the current configuration level if function code 0 was specified. + * Otherwise returns 0 on success or a negative value on error. + */ +int stsi(void *sysinfo, int fc, int sel1, int sel2) { - int rc = stsi(NULL, 0, 0, 0); - return rc == -ENOSYS ? rc : (((unsigned int) rc) >> 28); + register int r0 asm("0") = (fc << 28) | sel1; + register int r1 asm("1") = sel2; + int rc = 0; + + asm volatile( + " stsi 0(%3)\n" + "0: jz 2f\n" + "1: lhi %1,%4\n" + "2:\n" + EX_TABLE(0b, 1b) + : "+d" (r0), "+d" (rc) + : "d" (r1), "a" (sysinfo), "K" (-EOPNOTSUPP) + : "cc", "memory"); + if (rc) + return rc; + return fc ? 0 : ((unsigned int) r0) >> 28; } +EXPORT_SYMBOL(stsi); -static int stsi_1_1_1(struct sysinfo_1_1_1 *info, char *page, int len) +static void stsi_1_1_1(struct seq_file *m, struct sysinfo_1_1_1 *info) { - if (stsi(info, 1, 1, 1) == -ENOSYS) - return len; + int i; + if (stsi(info, 1, 1, 1)) + return; EBCASC(info->manufacturer, sizeof(info->manufacturer)); EBCASC(info->type, sizeof(info->type)); EBCASC(info->model, sizeof(info->model)); @@ -41,242 +65,197 @@ static int stsi_1_1_1(struct sysinfo_1_1_1 *info, char *page, int len) EBCASC(info->model_capacity, sizeof(info->model_capacity)); EBCASC(info->model_perm_cap, sizeof(info->model_perm_cap)); EBCASC(info->model_temp_cap, sizeof(info->model_temp_cap)); - len += sprintf(page + len, "Manufacturer: %-16.16s\n", - info->manufacturer); - len += sprintf(page + len, "Type: %-4.4s\n", - info->type); + seq_printf(m, "Manufacturer: %-16.16s\n", info->manufacturer); + seq_printf(m, "Type: %-4.4s\n", info->type); + /* + * Sigh: the model field has been renamed with System z9 + * to model_capacity and a new model field has been added + * after the plant field. To avoid confusing older programs + * the "Model:" prints "model_capacity model" or just + * "model_capacity" if the model string is empty . + */ + seq_printf(m, "Model: %-16.16s", info->model_capacity); if (info->model[0] != '\0') - /* - * Sigh: the model field has been renamed with System z9 - * to model_capacity and a new model field has been added - * after the plant field. To avoid confusing older programs - * the "Model:" prints "model_capacity model" or just - * "model_capacity" if the model string is empty . - */ - len += sprintf(page + len, - "Model: %-16.16s %-16.16s\n", - info->model_capacity, info->model); - else - len += sprintf(page + len, "Model: %-16.16s\n", - info->model_capacity); - len += sprintf(page + len, "Sequence Code: %-16.16s\n", - info->sequence); - len += sprintf(page + len, "Plant: %-4.4s\n", - info->plant); - len += sprintf(page + len, "Model Capacity: %-16.16s %08u\n", - info->model_capacity, *(u32 *) info->model_cap_rating); - if (info->model_perm_cap[0] != '\0') - len += sprintf(page + len, - "Model Perm. Capacity: %-16.16s %08u\n", - info->model_perm_cap, - *(u32 *) info->model_perm_cap_rating); - if (info->model_temp_cap[0] != '\0') - len += sprintf(page + len, - "Model Temp. Capacity: %-16.16s %08u\n", - info->model_temp_cap, - *(u32 *) info->model_temp_cap_rating); + seq_printf(m, " %-16.16s", info->model); + seq_putc(m, '\n'); + seq_printf(m, "Sequence Code: %-16.16s\n", info->sequence); + seq_printf(m, "Plant: %-4.4s\n", info->plant); + seq_printf(m, "Model Capacity: %-16.16s %08u\n", + info->model_capacity, info->model_cap_rating); + if (info->model_perm_cap_rating) + seq_printf(m, "Model Perm. Capacity: %-16.16s %08u\n", + info->model_perm_cap, + info->model_perm_cap_rating); + if (info->model_temp_cap_rating) + seq_printf(m, "Model Temp. Capacity: %-16.16s %08u\n", + info->model_temp_cap, + info->model_temp_cap_rating); + if (info->ncr) + seq_printf(m, "Nominal Cap. Rating: %08u\n", info->ncr); + if (info->npr) + seq_printf(m, "Nominal Perm. Rating: %08u\n", info->npr); + if (info->ntr) + seq_printf(m, "Nominal Temp. Rating: %08u\n", info->ntr); if (info->cai) { - len += sprintf(page + len, - "Capacity Adj. Ind.: %d\n", - info->cai); - len += sprintf(page + len, "Capacity Ch. Reason: %d\n", - info->ccr); + seq_printf(m, "Capacity Adj. Ind.: %d\n", info->cai); + seq_printf(m, "Capacity Ch. Reason: %d\n", info->ccr); + seq_printf(m, "Capacity Transient: %d\n", info->t); + } + if (info->p) { + for (i = 1; i <= ARRAY_SIZE(info->typepct); i++) { + seq_printf(m, "Type %d Percentage: %d\n", + i, info->typepct[i - 1]); + } } - return len; } -static int stsi_15_1_x(struct sysinfo_15_1_x *info, char *page, int len) +static void stsi_15_1_x(struct seq_file *m, struct sysinfo_15_1_x *info) { static int max_mnest; int i, rc; - len += sprintf(page + len, "\n"); + seq_putc(m, '\n'); if (!MACHINE_HAS_TOPOLOGY) - return len; - if (max_mnest) { - stsi(info, 15, 1, max_mnest); - } else { - for (max_mnest = 6; max_mnest > 1; max_mnest--) { - rc = stsi(info, 15, 1, max_mnest); - if (rc != -ENOSYS) - break; - } - } - len += sprintf(page + len, "CPU Topology HW: "); + return; + if (stsi(info, 15, 1, topology_max_mnest)) + return; + seq_printf(m, "CPU Topology HW: "); for (i = 0; i < TOPOLOGY_NR_MAG; i++) - len += sprintf(page + len, " %d", info->mag[i]); - len += sprintf(page + len, "\n"); + seq_printf(m, " %d", info->mag[i]); + seq_putc(m, '\n'); #ifdef CONFIG_SCHED_MC store_topology(info); - len += sprintf(page + len, "CPU Topology SW: "); + seq_printf(m, "CPU Topology SW: "); for (i = 0; i < TOPOLOGY_NR_MAG; i++) - len += sprintf(page + len, " %d", info->mag[i]); - len += sprintf(page + len, "\n"); + seq_printf(m, " %d", info->mag[i]); + seq_putc(m, '\n'); #endif - return len; } -static int stsi_1_2_2(struct sysinfo_1_2_2 *info, char *page, int len) +static void stsi_1_2_2(struct seq_file *m, struct sysinfo_1_2_2 *info) { struct sysinfo_1_2_2_extension *ext; int i; - if (stsi(info, 1, 2, 2) == -ENOSYS) - return len; + if (stsi(info, 1, 2, 2)) + return; ext = (struct sysinfo_1_2_2_extension *) ((unsigned long) info + info->acc_offset); - - len += sprintf(page + len, "CPUs Total: %d\n", - info->cpus_total); - len += sprintf(page + len, "CPUs Configured: %d\n", - info->cpus_configured); - len += sprintf(page + len, "CPUs Standby: %d\n", - info->cpus_standby); - len += sprintf(page + len, "CPUs Reserved: %d\n", - info->cpus_reserved); - - if (info->format == 1) { - /* - * Sigh 2. According to the specification the alternate - * capability field is a 32 bit floating point number - * if the higher order 8 bits are not zero. Printing - * a floating point number in the kernel is a no-no, - * always print the number as 32 bit unsigned integer. - * The user-space needs to know about the strange - * encoding of the alternate cpu capability. - */ - len += sprintf(page + len, "Capability: %u %u\n", - info->capability, ext->alt_capability); - for (i = 2; i <= info->cpus_total; i++) - len += sprintf(page + len, - "Adjustment %02d-way: %u %u\n", - i, info->adjustment[i-2], - ext->alt_adjustment[i-2]); - - } else { - len += sprintf(page + len, "Capability: %u\n", - info->capability); - for (i = 2; i <= info->cpus_total; i++) - len += sprintf(page + len, - "Adjustment %02d-way: %u\n", - i, info->adjustment[i-2]); + seq_printf(m, "CPUs Total: %d\n", info->cpus_total); + seq_printf(m, "CPUs Configured: %d\n", info->cpus_configured); + seq_printf(m, "CPUs Standby: %d\n", info->cpus_standby); + seq_printf(m, "CPUs Reserved: %d\n", info->cpus_reserved); + /* + * Sigh 2. According to the specification the alternate + * capability field is a 32 bit floating point number + * if the higher order 8 bits are not zero. Printing + * a floating point number in the kernel is a no-no, + * always print the number as 32 bit unsigned integer. + * The user-space needs to know about the strange + * encoding of the alternate cpu capability. + */ + seq_printf(m, "Capability: %u", info->capability); + if (info->format == 1) + seq_printf(m, " %u", ext->alt_capability); + seq_putc(m, '\n'); + if (info->nominal_cap) + seq_printf(m, "Nominal Capability: %d\n", info->nominal_cap); + if (info->secondary_cap) + seq_printf(m, "Secondary Capability: %d\n", info->secondary_cap); + for (i = 2; i <= info->cpus_total; i++) { + seq_printf(m, "Adjustment %02d-way: %u", + i, info->adjustment[i-2]); + if (info->format == 1) + seq_printf(m, " %u", ext->alt_adjustment[i-2]); + seq_putc(m, '\n'); } - - if (info->secondary_capability != 0) - len += sprintf(page + len, "Secondary Capability: %d\n", - info->secondary_capability); - return len; } -static int stsi_2_2_2(struct sysinfo_2_2_2 *info, char *page, int len) +static void stsi_2_2_2(struct seq_file *m, struct sysinfo_2_2_2 *info) { - if (stsi(info, 2, 2, 2) == -ENOSYS) - return len; - + if (stsi(info, 2, 2, 2)) + return; EBCASC(info->name, sizeof(info->name)); - - len += sprintf(page + len, "\n"); - len += sprintf(page + len, "LPAR Number: %d\n", - info->lpar_number); - - len += sprintf(page + len, "LPAR Characteristics: "); + seq_putc(m, '\n'); + seq_printf(m, "LPAR Number: %d\n", info->lpar_number); + seq_printf(m, "LPAR Characteristics: "); if (info->characteristics & LPAR_CHAR_DEDICATED) - len += sprintf(page + len, "Dedicated "); + seq_printf(m, "Dedicated "); if (info->characteristics & LPAR_CHAR_SHARED) - len += sprintf(page + len, "Shared "); + seq_printf(m, "Shared "); if (info->characteristics & LPAR_CHAR_LIMITED) - len += sprintf(page + len, "Limited "); - len += sprintf(page + len, "\n"); - - len += sprintf(page + len, "LPAR Name: %-8.8s\n", - info->name); - - len += sprintf(page + len, "LPAR Adjustment: %d\n", - info->caf); - - len += sprintf(page + len, "LPAR CPUs Total: %d\n", - info->cpus_total); - len += sprintf(page + len, "LPAR CPUs Configured: %d\n", - info->cpus_configured); - len += sprintf(page + len, "LPAR CPUs Standby: %d\n", - info->cpus_standby); - len += sprintf(page + len, "LPAR CPUs Reserved: %d\n", - info->cpus_reserved); - len += sprintf(page + len, "LPAR CPUs Dedicated: %d\n", - info->cpus_dedicated); - len += sprintf(page + len, "LPAR CPUs Shared: %d\n", - info->cpus_shared); - return len; + seq_printf(m, "Limited "); + seq_putc(m, '\n'); + seq_printf(m, "LPAR Name: %-8.8s\n", info->name); + seq_printf(m, "LPAR Adjustment: %d\n", info->caf); + seq_printf(m, "LPAR CPUs Total: %d\n", info->cpus_total); + seq_printf(m, "LPAR CPUs Configured: %d\n", info->cpus_configured); + seq_printf(m, "LPAR CPUs Standby: %d\n", info->cpus_standby); + seq_printf(m, "LPAR CPUs Reserved: %d\n", info->cpus_reserved); + seq_printf(m, "LPAR CPUs Dedicated: %d\n", info->cpus_dedicated); + seq_printf(m, "LPAR CPUs Shared: %d\n", info->cpus_shared); } -static int stsi_3_2_2(struct sysinfo_3_2_2 *info, char *page, int len) +static void stsi_3_2_2(struct seq_file *m, struct sysinfo_3_2_2 *info) { int i; - if (stsi(info, 3, 2, 2) == -ENOSYS) - return len; + if (stsi(info, 3, 2, 2)) + return; for (i = 0; i < info->count; i++) { EBCASC(info->vm[i].name, sizeof(info->vm[i].name)); EBCASC(info->vm[i].cpi, sizeof(info->vm[i].cpi)); - len += sprintf(page + len, "\n"); - len += sprintf(page + len, "VM%02d Name: %-8.8s\n", - i, info->vm[i].name); - len += sprintf(page + len, "VM%02d Control Program: %-16.16s\n", - i, info->vm[i].cpi); - - len += sprintf(page + len, "VM%02d Adjustment: %d\n", - i, info->vm[i].caf); - - len += sprintf(page + len, "VM%02d CPUs Total: %d\n", - i, info->vm[i].cpus_total); - len += sprintf(page + len, "VM%02d CPUs Configured: %d\n", - i, info->vm[i].cpus_configured); - len += sprintf(page + len, "VM%02d CPUs Standby: %d\n", - i, info->vm[i].cpus_standby); - len += sprintf(page + len, "VM%02d CPUs Reserved: %d\n", - i, info->vm[i].cpus_reserved); + seq_putc(m, '\n'); + seq_printf(m, "VM%02d Name: %-8.8s\n", i, info->vm[i].name); + seq_printf(m, "VM%02d Control Program: %-16.16s\n", i, info->vm[i].cpi); + seq_printf(m, "VM%02d Adjustment: %d\n", i, info->vm[i].caf); + seq_printf(m, "VM%02d CPUs Total: %d\n", i, info->vm[i].cpus_total); + seq_printf(m, "VM%02d CPUs Configured: %d\n", i, info->vm[i].cpus_configured); + seq_printf(m, "VM%02d CPUs Standby: %d\n", i, info->vm[i].cpus_standby); + seq_printf(m, "VM%02d CPUs Reserved: %d\n", i, info->vm[i].cpus_reserved); } - return len; } -static int proc_read_sysinfo(char *page, char **start, - off_t off, int count, - int *eof, void *data) +static int sysinfo_show(struct seq_file *m, void *v) { - unsigned long info = get_zeroed_page(GFP_KERNEL); - int level, len; + void *info = (void *)get_zeroed_page(GFP_KERNEL); + int level; if (!info) return 0; - - len = 0; - level = stsi_0(); + level = stsi(NULL, 0, 0, 0); if (level >= 1) - len = stsi_1_1_1((struct sysinfo_1_1_1 *) info, page, len); - + stsi_1_1_1(m, info); if (level >= 1) - len = stsi_15_1_x((struct sysinfo_15_1_x *) info, page, len); - + stsi_15_1_x(m, info); if (level >= 1) - len = stsi_1_2_2((struct sysinfo_1_2_2 *) info, page, len); - + stsi_1_2_2(m, info); if (level >= 2) - len = stsi_2_2_2((struct sysinfo_2_2_2 *) info, page, len); - + stsi_2_2_2(m, info); if (level >= 3) - len = stsi_3_2_2((struct sysinfo_3_2_2 *) info, page, len); + stsi_3_2_2(m, info); + free_page((unsigned long)info); + return 0; +} - free_page(info); - return len; +static int sysinfo_open(struct inode *inode, struct file *file) +{ + return single_open(file, sysinfo_show, NULL); } -static __init int create_proc_sysinfo(void) +static const struct file_operations sysinfo_fops = { + .open = sysinfo_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static int __init sysinfo_create_proc(void) { - create_proc_read_entry("sysinfo", 0444, NULL, - proc_read_sysinfo, NULL); + proc_create("sysinfo", 0444, NULL, &sysinfo_fops); return 0; } -device_initcall(create_proc_sysinfo); +device_initcall(sysinfo_create_proc); /* * Service levels interface. @@ -407,7 +386,7 @@ void s390_adjust_jiffies(void) if (!info) return; - if (stsi(info, 1, 2, 2) != -ENOSYS) { + if (stsi(info, 1, 2, 2) == 0) { /* * Major sigh. The cpu capability encoding is "special". * If the first 9 bits of info->capability are 0 then it diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c index dcec960fc72..2db1011b8b1 100644 --- a/arch/s390/kernel/time.c +++ b/arch/s390/kernel/time.c @@ -329,7 +329,7 @@ static unsigned long clock_sync_flags; * The synchronous get_clock function. It will write the current clock * value to the clock pointer and return 0 if the clock is in sync with * the external time source. If the clock mode is local it will return - * -ENOSYS and -EAGAIN if the clock is not in sync with the external + * -EOPNOTSUPP and -EAGAIN if the clock is not in sync with the external * reference. */ int get_sync_clock(unsigned long long *clock) @@ -347,7 +347,7 @@ int get_sync_clock(unsigned long long *clock) return 0; if (!test_bit(CLOCK_SYNC_HAS_ETR, &clock_sync_flags) && !test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags)) - return -ENOSYS; + return -EOPNOTSUPP; if (!test_bit(CLOCK_SYNC_ETR, &clock_sync_flags) && !test_bit(CLOCK_SYNC_STP, &clock_sync_flags)) return -EACCES; diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c index 05151e06c38..54d93f4b681 100644 --- a/arch/s390/kernel/topology.c +++ b/arch/s390/kernel/topology.c @@ -17,6 +17,7 @@ #include <linux/cpu.h> #include <linux/smp.h> #include <linux/mm.h> +#include <asm/sysinfo.h> #define PTF_HORIZONTAL (0UL) #define PTF_VERTICAL (1UL) @@ -44,9 +45,6 @@ static struct mask_info book_info; cpumask_t cpu_book_map[NR_CPUS]; unsigned char cpu_book_id[NR_CPUS]; -/* smp_cpu_state_mutex must be held when accessing this array */ -int cpu_polarization[NR_CPUS]; - static cpumask_t cpu_group_map(struct mask_info *info, unsigned int cpu) { cpumask_t mask; @@ -75,10 +73,7 @@ static struct mask_info *add_cpus_to_mask(struct topology_cpu *tl_cpu, { unsigned int cpu; - for (cpu = find_first_bit(&tl_cpu->mask[0], TOPOLOGY_CPU_BITS); - cpu < TOPOLOGY_CPU_BITS; - cpu = find_next_bit(&tl_cpu->mask[0], TOPOLOGY_CPU_BITS, cpu + 1)) - { + for_each_set_bit(cpu, &tl_cpu->mask[0], TOPOLOGY_CPU_BITS) { unsigned int rcpu; int lcpu; @@ -94,7 +89,7 @@ static struct mask_info *add_cpus_to_mask(struct topology_cpu *tl_cpu, } else { cpu_core_id[lcpu] = core->id; } - cpu_set_polarization(lcpu, tl_cpu->pp); + smp_cpu_set_polarization(lcpu, tl_cpu->pp); } } return core; @@ -201,7 +196,7 @@ static void topology_update_polarization_simple(void) mutex_lock(&smp_cpu_state_mutex); for_each_possible_cpu(cpu) - cpu_set_polarization(cpu, POLARIZATION_HRZ); + smp_cpu_set_polarization(cpu, POLARIZATION_HRZ); mutex_unlock(&smp_cpu_state_mutex); } @@ -231,7 +226,7 @@ int topology_set_cpu_management(int fc) if (rc) return -EBUSY; for_each_possible_cpu(cpu) - cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); + smp_cpu_set_polarization(cpu, POLARIZATION_UNKNOWN); return rc; } @@ -250,12 +245,10 @@ static void update_cpu_core_map(void) void store_topology(struct sysinfo_15_1_x *info) { - int rc; - - rc = stsi(info, 15, 1, 3); - if (rc != -ENOSYS) - return; - stsi(info, 15, 1, 2); + if (topology_max_mnest >= 3) + stsi(info, 15, 1, 3); + else + stsi(info, 15, 1, 2); } int arch_update_cpu_topology(void) @@ -415,7 +408,7 @@ static ssize_t cpu_polarization_show(struct device *dev, ssize_t count; mutex_lock(&smp_cpu_state_mutex); - switch (cpu_read_polarization(cpu)) { + switch (smp_cpu_get_polarization(cpu)) { case POLARIZATION_HRZ: count = sprintf(buf, "horizontal\n"); break; diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c index 01775c04a90..3d2b0fa37db 100644 --- a/arch/s390/kernel/traps.c +++ b/arch/s390/kernel/traps.c @@ -57,6 +57,23 @@ static int kstack_depth_to_print = 12; static int kstack_depth_to_print = 20; #endif /* CONFIG_64BIT */ +static inline void __user *get_trap_ip(struct pt_regs *regs) +{ +#ifdef CONFIG_64BIT + unsigned long address; + + if (regs->int_code & 0x200) + address = *(unsigned long *)(current->thread.trap_tdb + 24); + else + address = regs->psw.addr; + return (void __user *) + ((address - (regs->int_code >> 16)) & PSW_ADDR_INSN); +#else + return (void __user *) + ((regs->psw.addr - (regs->int_code >> 16)) & PSW_ADDR_INSN); +#endif +} + /* * For show_trace we have tree different stack to consider: * - the panic stack which is used if the kernel stack has overflown @@ -214,7 +231,6 @@ void show_registers(struct pt_regs *regs) void show_regs(struct pt_regs *regs) { - print_modules(); printk("CPU: %d %s %s %.*s\n", task_thread_info(current)->cpu, print_tainted(), init_utsname()->release, @@ -254,6 +270,7 @@ void die(struct pt_regs *regs, const char *str) #endif printk("\n"); notify_die(DIE_OOPS, str, regs, 0, regs->int_code & 0xffff, SIGSEGV); + print_modules(); show_regs(regs); bust_spinlocks(0); add_taint(TAINT_DIE); @@ -285,12 +302,6 @@ int is_valid_bugaddr(unsigned long addr) return 1; } -static inline void __user *get_psw_address(struct pt_regs *regs) -{ - return (void __user *) - ((regs->psw.addr - (regs->int_code >> 16)) & PSW_ADDR_INSN); -} - static void __kprobes do_trap(struct pt_regs *regs, int si_signo, int si_code, char *str) { @@ -304,14 +315,14 @@ static void __kprobes do_trap(struct pt_regs *regs, info.si_signo = si_signo; info.si_errno = 0; info.si_code = si_code; - info.si_addr = get_psw_address(regs); + info.si_addr = get_trap_ip(regs); force_sig_info(si_signo, &info, current); report_user_fault(regs, si_signo); } else { const struct exception_table_entry *fixup; fixup = search_exception_tables(regs->psw.addr & PSW_ADDR_INSN); if (fixup) - regs->psw.addr = fixup->fixup | PSW_ADDR_AMODE; + regs->psw.addr = extable_fixup(fixup) | PSW_ADDR_AMODE; else { enum bug_trap_type btt; @@ -381,6 +392,11 @@ DO_ERROR_INFO(special_op_exception, SIGILL, ILL_ILLOPN, DO_ERROR_INFO(translation_exception, SIGILL, ILL_ILLOPN, "translation exception") +#ifdef CONFIG_64BIT +DO_ERROR_INFO(transaction_exception, SIGILL, ILL_ILLOPN, + "transaction constraint exception") +#endif + static inline void do_fp_trap(struct pt_regs *regs, int fpc) { int si_code = 0; @@ -408,7 +424,7 @@ static void __kprobes illegal_op(struct pt_regs *regs) __u16 __user *location; int signal = 0; - location = get_psw_address(regs); + location = get_trap_ip(regs); if (user_mode(regs)) { if (get_user(*((__u16 *) opcode), (__u16 __user *) location)) @@ -476,7 +492,7 @@ void specification_exception(struct pt_regs *regs) __u16 __user *location = NULL; int signal = 0; - location = (__u16 __user *) get_psw_address(regs); + location = (__u16 __user *) get_trap_ip(regs); if (user_mode(regs)) { get_user(*((__u16 *) opcode), location); @@ -525,7 +541,7 @@ static void data_exception(struct pt_regs *regs) __u16 __user *location; int signal = 0; - location = get_psw_address(regs); + location = get_trap_ip(regs); if (MACHINE_HAS_IEEE) asm volatile("stfpc %0" : "=m" (current->thread.fp_regs.fpc)); @@ -641,6 +657,7 @@ void __init trap_init(void) pgm_check_table[0x12] = &translation_exception; pgm_check_table[0x13] = &special_op_exception; #ifdef CONFIG_64BIT + pgm_check_table[0x18] = &transaction_exception; pgm_check_table[0x38] = &do_asce_exception; pgm_check_table[0x39] = &do_dat_exception; pgm_check_table[0x3A] = &do_dat_exception; diff --git a/arch/s390/kernel/vdso.c b/arch/s390/kernel/vdso.c index 9a19ca367c1..d7776281cb6 100644 --- a/arch/s390/kernel/vdso.c +++ b/arch/s390/kernel/vdso.c @@ -85,7 +85,7 @@ struct vdso_data *vdso_data = &vdso_data_store.data; static void vdso_init_data(struct vdso_data *vd) { vd->ectg_available = - addressing_mode != HOME_SPACE_MODE && test_facility(31); + s390_user_mode != HOME_SPACE_MODE && test_facility(31); } #ifdef CONFIG_64BIT @@ -102,7 +102,7 @@ int vdso_alloc_per_cpu(struct _lowcore *lowcore) lowcore->vdso_per_cpu_data = __LC_PASTE; - if (addressing_mode == HOME_SPACE_MODE || !vdso_enabled) + if (s390_user_mode == HOME_SPACE_MODE || !vdso_enabled) return 0; segment_table = __get_free_pages(GFP_KERNEL, SEGMENT_ORDER); @@ -147,7 +147,7 @@ void vdso_free_per_cpu(struct _lowcore *lowcore) unsigned long segment_table, page_table, page_frame; u32 *psal, *aste; - if (addressing_mode == HOME_SPACE_MODE || !vdso_enabled) + if (s390_user_mode == HOME_SPACE_MODE || !vdso_enabled) return; psal = (u32 *)(addr_t) lowcore->paste[4]; @@ -165,7 +165,7 @@ static void vdso_init_cr5(void) { unsigned long cr5; - if (addressing_mode == HOME_SPACE_MODE || !vdso_enabled) + if (s390_user_mode == HOME_SPACE_MODE || !vdso_enabled) return; cr5 = offsetof(struct _lowcore, paste); __ctl_load(cr5, 5, 5); diff --git a/arch/s390/kernel/vtime.c b/arch/s390/kernel/vtime.c index 4fc97b40a6e..79033442789 100644 --- a/arch/s390/kernel/vtime.c +++ b/arch/s390/kernel/vtime.c @@ -99,7 +99,7 @@ static int do_account_vtime(struct task_struct *tsk, int hardirq_offset) return virt_timer_forward(user + system); } -void account_vtime(struct task_struct *prev, struct task_struct *next) +void vtime_task_switch(struct task_struct *prev) { struct thread_info *ti; @@ -107,7 +107,7 @@ void account_vtime(struct task_struct *prev, struct task_struct *next) ti = task_thread_info(prev); ti->user_timer = S390_lowcore.user_timer; ti->system_timer = S390_lowcore.system_timer; - ti = task_thread_info(next); + ti = task_thread_info(current); S390_lowcore.user_timer = ti->user_timer; S390_lowcore.system_timer = ti->system_timer; } @@ -122,7 +122,7 @@ void account_process_tick(struct task_struct *tsk, int user_tick) * Update process times based on virtual cpu times stored by entry.S * to the lowcore fields user_timer, system_timer & steal_clock. */ -void account_system_vtime(struct task_struct *tsk) +void vtime_account(struct task_struct *tsk) { struct thread_info *ti = task_thread_info(tsk); u64 timer, system; @@ -138,7 +138,7 @@ void account_system_vtime(struct task_struct *tsk) virt_timer_forward(system); } -EXPORT_SYMBOL_GPL(account_system_vtime); +EXPORT_SYMBOL_GPL(vtime_account); void __kprobes vtime_stop_cpu(void) { @@ -378,9 +378,8 @@ static int __cpuinit s390_nohz_notify(struct notifier_block *self, long cpu = (long) hcpu; idle = &per_cpu(s390_idle, cpu); - switch (action) { + switch (action & ~CPU_TASKS_FROZEN) { case CPU_DYING: - case CPU_DYING_FROZEN: idle->nohz_delay = 0; default: break; diff --git a/arch/s390/kvm/Kconfig b/arch/s390/kvm/Kconfig index 78eb9847008..9b04a32e569 100644 --- a/arch/s390/kvm/Kconfig +++ b/arch/s390/kvm/Kconfig @@ -5,7 +5,7 @@ source "virt/kvm/Kconfig" menuconfig VIRTUALIZATION def_bool y - prompt "Virtualization" + prompt "KVM" ---help--- Say Y here to get to see options for using your Linux host to run other operating systems inside virtual machines (guests). diff --git a/arch/s390/kvm/priv.c b/arch/s390/kvm/priv.c index 60da903d6f3..310be61bead 100644 --- a/arch/s390/kvm/priv.c +++ b/arch/s390/kvm/priv.c @@ -211,7 +211,7 @@ static void handle_stsi_3_2_2(struct kvm_vcpu *vcpu, struct sysinfo_3_2_2 *mem) spin_unlock(&fi->lock); /* deal with other level 3 hypervisors */ - if (stsi(mem, 3, 2, 2) == -ENOSYS) + if (stsi(mem, 3, 2, 2)) mem->count = 0; if (mem->count < 8) mem->count++; @@ -259,7 +259,7 @@ static int handle_stsi(struct kvm_vcpu *vcpu) mem = get_zeroed_page(GFP_KERNEL); if (!mem) goto out_fail; - if (stsi((void *) mem, fc, sel1, sel2) == -ENOSYS) + if (stsi((void *) mem, fc, sel1, sel2)) goto out_mem; break; case 3: diff --git a/arch/s390/lib/Makefile b/arch/s390/lib/Makefile index 761ab8b56af..6ab0d0b5cec 100644 --- a/arch/s390/lib/Makefile +++ b/arch/s390/lib/Makefile @@ -4,6 +4,7 @@ lib-y += delay.o string.o uaccess_std.o uaccess_pt.o obj-y += usercopy.o -obj-$(CONFIG_32BIT) += div64.o qrnnd.o ucmpdi2.o +obj-$(CONFIG_32BIT) += div64.o qrnnd.o ucmpdi2.o mem32.o +obj-$(CONFIG_64BIT) += mem64.o lib-$(CONFIG_64BIT) += uaccess_mvcos.o lib-$(CONFIG_SMP) += spinlock.o diff --git a/arch/s390/lib/mem32.S b/arch/s390/lib/mem32.S new file mode 100644 index 00000000000..14ca9244b61 --- /dev/null +++ b/arch/s390/lib/mem32.S @@ -0,0 +1,92 @@ +/* + * String handling functions. + * + * Copyright IBM Corp. 2012 + */ + +#include <linux/linkage.h> + +/* + * memset implementation + * + * This code corresponds to the C construct below. We do distinguish + * between clearing (c == 0) and setting a memory array (c != 0) simply + * because nearly all memset invocations in the kernel clear memory and + * the xc instruction is preferred in such cases. + * + * void *memset(void *s, int c, size_t n) + * { + * if (likely(c == 0)) + * return __builtin_memset(s, 0, n); + * return __builtin_memset(s, c, n); + * } + */ +ENTRY(memset) + basr %r5,%r0 +.Lmemset_base: + ltr %r4,%r4 + bzr %r14 + ltr %r3,%r3 + jnz .Lmemset_fill + ahi %r4,-1 + lr %r3,%r4 + srl %r3,8 + ltr %r3,%r3 + lr %r1,%r2 + je .Lmemset_clear_rest +.Lmemset_clear_loop: + xc 0(256,%r1),0(%r1) + la %r1,256(%r1) + brct %r3,.Lmemset_clear_loop +.Lmemset_clear_rest: + ex %r4,.Lmemset_xc-.Lmemset_base(%r5) + br %r14 +.Lmemset_fill: + stc %r3,0(%r2) + chi %r4,1 + lr %r1,%r2 + ber %r14 + ahi %r4,-2 + lr %r3,%r4 + srl %r3,8 + ltr %r3,%r3 + je .Lmemset_fill_rest +.Lmemset_fill_loop: + mvc 1(256,%r1),0(%r1) + la %r1,256(%r1) + brct %r3,.Lmemset_fill_loop +.Lmemset_fill_rest: + ex %r4,.Lmemset_mvc-.Lmemset_base(%r5) + br %r14 +.Lmemset_xc: + xc 0(1,%r1),0(%r1) +.Lmemset_mvc: + mvc 1(1,%r1),0(%r1) + +/* + * memcpy implementation + * + * void *memcpy(void *dest, const void *src, size_t n) + */ +ENTRY(memcpy) + basr %r5,%r0 +.Lmemcpy_base: + ltr %r4,%r4 + bzr %r14 + ahi %r4,-1 + lr %r0,%r4 + srl %r0,8 + ltr %r0,%r0 + lr %r1,%r2 + jnz .Lmemcpy_loop +.Lmemcpy_rest: + ex %r4,.Lmemcpy_mvc-.Lmemcpy_base(%r5) + br %r14 +.Lmemcpy_loop: + mvc 0(256,%r1),0(%r3) + la %r1,256(%r1) + la %r3,256(%r3) + brct %r0,.Lmemcpy_loop + j .Lmemcpy_rest +.Lmemcpy_mvc: + mvc 0(1,%r1),0(%r3) diff --git a/arch/s390/lib/mem64.S b/arch/s390/lib/mem64.S new file mode 100644 index 00000000000..c6d553e85ab --- /dev/null +++ b/arch/s390/lib/mem64.S @@ -0,0 +1,88 @@ +/* + * String handling functions. + * + * Copyright IBM Corp. 2012 + */ + +#include <linux/linkage.h> + +/* + * memset implementation + * + * This code corresponds to the C construct below. We do distinguish + * between clearing (c == 0) and setting a memory array (c != 0) simply + * because nearly all memset invocations in the kernel clear memory and + * the xc instruction is preferred in such cases. + * + * void *memset(void *s, int c, size_t n) + * { + * if (likely(c == 0)) + * return __builtin_memset(s, 0, n); + * return __builtin_memset(s, c, n); + * } + */ +ENTRY(memset) + ltgr %r4,%r4 + bzr %r14 + ltgr %r3,%r3 + jnz .Lmemset_fill + aghi %r4,-1 + srlg %r3,%r4,8 + ltgr %r3,%r3 + lgr %r1,%r2 + jz .Lmemset_clear_rest +.Lmemset_clear_loop: + xc 0(256,%r1),0(%r1) + la %r1,256(%r1) + brctg %r3,.Lmemset_clear_loop +.Lmemset_clear_rest: + larl %r3,.Lmemset_xc + ex %r4,0(%r3) + br %r14 +.Lmemset_fill: + stc %r3,0(%r2) + cghi %r4,1 + lgr %r1,%r2 + ber %r14 + aghi %r4,-2 + srlg %r3,%r4,8 + ltgr %r3,%r3 + jz .Lmemset_fill_rest +.Lmemset_fill_loop: + mvc 1(256,%r1),0(%r1) + la %r1,256(%r1) + brctg %r3,.Lmemset_fill_loop +.Lmemset_fill_rest: + larl %r3,.Lmemset_mvc + ex %r4,0(%r3) + br %r14 +.Lmemset_xc: + xc 0(1,%r1),0(%r1) +.Lmemset_mvc: + mvc 1(1,%r1),0(%r1) + +/* + * memcpy implementation + * + * void *memcpy(void *dest, const void *src, size_t n) + */ +ENTRY(memcpy) + ltgr %r4,%r4 + bzr %r14 + aghi %r4,-1 + srlg %r5,%r4,8 + ltgr %r5,%r5 + lgr %r1,%r2 + jnz .Lmemcpy_loop +.Lmemcpy_rest: + larl %r5,.Lmemcpy_mvc + ex %r4,0(%r5) + br %r14 +.Lmemcpy_loop: + mvc 0(256,%r1),0(%r3) + la %r1,256(%r1) + la %r3,256(%r3) + brctg %r5,.Lmemcpy_loop + j .Lmemcpy_rest +.Lmemcpy_mvc: + mvc 0(1,%r1),0(%r3) diff --git a/arch/s390/lib/string.c b/arch/s390/lib/string.c index 846ec64ab2c..b647d5ff0ad 100644 --- a/arch/s390/lib/string.c +++ b/arch/s390/lib/string.c @@ -43,11 +43,7 @@ static inline char *__strnend(const char *s, size_t n) */ size_t strlen(const char *s) { -#if __GNUC__ < 4 return __strend(s) - s; -#else - return __builtin_strlen(s); -#endif } EXPORT_SYMBOL(strlen); @@ -73,7 +69,6 @@ EXPORT_SYMBOL(strnlen); */ char *strcpy(char *dest, const char *src) { -#if __GNUC__ < 4 register int r0 asm("0") = 0; char *ret = dest; @@ -82,9 +77,6 @@ char *strcpy(char *dest, const char *src) : "+&a" (dest), "+&a" (src) : "d" (r0) : "cc", "memory" ); return ret; -#else - return __builtin_strcpy(dest, src); -#endif } EXPORT_SYMBOL(strcpy); @@ -106,7 +98,7 @@ size_t strlcpy(char *dest, const char *src, size_t size) if (size) { size_t len = (ret >= size) ? size-1 : ret; dest[len] = '\0'; - __builtin_memcpy(dest, src, len); + memcpy(dest, src, len); } return ret; } @@ -124,8 +116,8 @@ EXPORT_SYMBOL(strlcpy); char *strncpy(char *dest, const char *src, size_t n) { size_t len = __strnend(src, n) - src; - __builtin_memset(dest + len, 0, n - len); - __builtin_memcpy(dest, src, len); + memset(dest + len, 0, n - len); + memcpy(dest, src, len); return dest; } EXPORT_SYMBOL(strncpy); @@ -171,7 +163,7 @@ size_t strlcat(char *dest, const char *src, size_t n) if (len >= n) len = n - 1; dest[len] = '\0'; - __builtin_memcpy(dest, src, len); + memcpy(dest, src, len); } return res; } @@ -194,7 +186,7 @@ char *strncat(char *dest, const char *src, size_t n) char *p = __strend(dest); p[len] = '\0'; - __builtin_memcpy(p, src, len); + memcpy(p, src, len); return dest; } EXPORT_SYMBOL(strncat); @@ -348,41 +340,3 @@ void *memscan(void *s, int c, size_t n) return (void *) ret; } EXPORT_SYMBOL(memscan); - -/** - * memcpy - Copy one area of memory to another - * @dest: Where to copy to - * @src: Where to copy from - * @n: The size of the area. - * - * returns a pointer to @dest - */ -void *memcpy(void *dest, const void *src, size_t n) -{ - return __builtin_memcpy(dest, src, n); -} -EXPORT_SYMBOL(memcpy); - -/** - * memset - Fill a region of memory with the given value - * @s: Pointer to the start of the area. - * @c: The byte to fill the area with - * @n: The size of the area. - * - * returns a pointer to @s - */ -void *memset(void *s, int c, size_t n) -{ - char *xs; - - if (c == 0) - return __builtin_memset(s, 0, n); - - xs = (char *) s; - if (n > 0) - do { - *xs++ = c; - } while (--n > 0); - return s; -} -EXPORT_SYMBOL(memset); diff --git a/arch/s390/lib/uaccess_pt.c b/arch/s390/lib/uaccess_pt.c index 60ee2b88379..2d37bb861fa 100644 --- a/arch/s390/lib/uaccess_pt.c +++ b/arch/s390/lib/uaccess_pt.c @@ -2,69 +2,82 @@ * User access functions based on page table walks for enhanced * system layout without hardware support. * - * Copyright IBM Corp. 2006 + * Copyright IBM Corp. 2006, 2012 * Author(s): Gerald Schaefer (gerald.schaefer@de.ibm.com) */ #include <linux/errno.h> #include <linux/hardirq.h> #include <linux/mm.h> +#include <linux/hugetlb.h> #include <asm/uaccess.h> #include <asm/futex.h> #include "uaccess.h" -static inline pte_t *follow_table(struct mm_struct *mm, unsigned long addr) + +/* + * Returns kernel address for user virtual address. If the returned address is + * >= -4095 (IS_ERR_VALUE(x) returns true), a fault has occured and the address + * contains the (negative) exception code. + */ +static __always_inline unsigned long follow_table(struct mm_struct *mm, + unsigned long addr, int write) { pgd_t *pgd; pud_t *pud; pmd_t *pmd; + pte_t *ptep; pgd = pgd_offset(mm, addr); if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd))) - return (pte_t *) 0x3a; + return -0x3aUL; pud = pud_offset(pgd, addr); if (pud_none(*pud) || unlikely(pud_bad(*pud))) - return (pte_t *) 0x3b; + return -0x3bUL; pmd = pmd_offset(pud, addr); - if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd))) - return (pte_t *) 0x10; + if (pmd_none(*pmd)) + return -0x10UL; + if (pmd_huge(*pmd)) { + if (write && (pmd_val(*pmd) & _SEGMENT_ENTRY_RO)) + return -0x04UL; + return (pmd_val(*pmd) & HPAGE_MASK) + (addr & ~HPAGE_MASK); + } + if (unlikely(pmd_bad(*pmd))) + return -0x10UL; + + ptep = pte_offset_map(pmd, addr); + if (!pte_present(*ptep)) + return -0x11UL; + if (write && !pte_write(*ptep)) + return -0x04UL; - return pte_offset_map(pmd, addr); + return (pte_val(*ptep) & PAGE_MASK) + (addr & ~PAGE_MASK); } static __always_inline size_t __user_copy_pt(unsigned long uaddr, void *kptr, size_t n, int write_user) { struct mm_struct *mm = current->mm; - unsigned long offset, pfn, done, size; - pte_t *pte; + unsigned long offset, done, size, kaddr; void *from, *to; done = 0; retry: spin_lock(&mm->page_table_lock); do { - pte = follow_table(mm, uaddr); - if ((unsigned long) pte < 0x1000) + kaddr = follow_table(mm, uaddr, write_user); + if (IS_ERR_VALUE(kaddr)) goto fault; - if (!pte_present(*pte)) { - pte = (pte_t *) 0x11; - goto fault; - } else if (write_user && !pte_write(*pte)) { - pte = (pte_t *) 0x04; - goto fault; - } - pfn = pte_pfn(*pte); - offset = uaddr & (PAGE_SIZE - 1); + offset = uaddr & ~PAGE_MASK; size = min(n - done, PAGE_SIZE - offset); if (write_user) { - to = (void *)((pfn << PAGE_SHIFT) + offset); + to = (void *) kaddr; from = kptr + done; } else { - from = (void *)((pfn << PAGE_SHIFT) + offset); + from = (void *) kaddr; to = kptr + done; } memcpy(to, from, size); @@ -75,7 +88,7 @@ retry: return n - done; fault: spin_unlock(&mm->page_table_lock); - if (__handle_fault(uaddr, (unsigned long) pte, write_user)) + if (__handle_fault(uaddr, -kaddr, write_user)) return n - done; goto retry; } @@ -84,27 +97,22 @@ fault: * Do DAT for user address by page table walk, return kernel address. * This function needs to be called with current->mm->page_table_lock held. */ -static __always_inline unsigned long __dat_user_addr(unsigned long uaddr) +static __always_inline unsigned long __dat_user_addr(unsigned long uaddr, + int write) { struct mm_struct *mm = current->mm; - unsigned long pfn; - pte_t *pte; + unsigned long kaddr; int rc; retry: - pte = follow_table(mm, uaddr); - if ((unsigned long) pte < 0x1000) - goto fault; - if (!pte_present(*pte)) { - pte = (pte_t *) 0x11; + kaddr = follow_table(mm, uaddr, write); + if (IS_ERR_VALUE(kaddr)) goto fault; - } - pfn = pte_pfn(*pte); - return (pfn << PAGE_SHIFT) + (uaddr & (PAGE_SIZE - 1)); + return kaddr; fault: spin_unlock(&mm->page_table_lock); - rc = __handle_fault(uaddr, (unsigned long) pte, 0); + rc = __handle_fault(uaddr, -kaddr, write); spin_lock(&mm->page_table_lock); if (!rc) goto retry; @@ -159,11 +167,9 @@ static size_t clear_user_pt(size_t n, void __user *to) static size_t strnlen_user_pt(size_t count, const char __user *src) { - char *addr; unsigned long uaddr = (unsigned long) src; struct mm_struct *mm = current->mm; - unsigned long offset, pfn, done, len; - pte_t *pte; + unsigned long offset, done, len, kaddr; size_t len_str; if (segment_eq(get_fs(), KERNEL_DS)) @@ -172,19 +178,13 @@ static size_t strnlen_user_pt(size_t count, const char __user *src) retry: spin_lock(&mm->page_table_lock); do { - pte = follow_table(mm, uaddr); - if ((unsigned long) pte < 0x1000) - goto fault; - if (!pte_present(*pte)) { - pte = (pte_t *) 0x11; + kaddr = follow_table(mm, uaddr, 0); + if (IS_ERR_VALUE(kaddr)) goto fault; - } - pfn = pte_pfn(*pte); - offset = uaddr & (PAGE_SIZE-1); - addr = (char *)(pfn << PAGE_SHIFT) + offset; + offset = uaddr & ~PAGE_MASK; len = min(count - done, PAGE_SIZE - offset); - len_str = strnlen(addr, len); + len_str = strnlen((char *) kaddr, len); done += len_str; uaddr += len_str; } while ((len_str == len) && (done < count)); @@ -192,7 +192,7 @@ retry: return done + 1; fault: spin_unlock(&mm->page_table_lock); - if (__handle_fault(uaddr, (unsigned long) pte, 0)) + if (__handle_fault(uaddr, -kaddr, 0)) return 0; goto retry; } @@ -225,11 +225,10 @@ static size_t copy_in_user_pt(size_t n, void __user *to, const void __user *from) { struct mm_struct *mm = current->mm; - unsigned long offset_from, offset_to, offset_max, pfn_from, pfn_to, - uaddr, done, size, error_code; + unsigned long offset_max, uaddr, done, size, error_code; unsigned long uaddr_from = (unsigned long) from; unsigned long uaddr_to = (unsigned long) to; - pte_t *pte_from, *pte_to; + unsigned long kaddr_to, kaddr_from; int write_user; if (segment_eq(get_fs(), KERNEL_DS)) { @@ -242,38 +241,23 @@ retry: do { write_user = 0; uaddr = uaddr_from; - pte_from = follow_table(mm, uaddr_from); - error_code = (unsigned long) pte_from; - if (error_code < 0x1000) - goto fault; - if (!pte_present(*pte_from)) { - error_code = 0x11; + kaddr_from = follow_table(mm, uaddr_from, 0); + error_code = kaddr_from; + if (IS_ERR_VALUE(error_code)) goto fault; - } write_user = 1; uaddr = uaddr_to; - pte_to = follow_table(mm, uaddr_to); - error_code = (unsigned long) pte_to; - if (error_code < 0x1000) - goto fault; - if (!pte_present(*pte_to)) { - error_code = 0x11; + kaddr_to = follow_table(mm, uaddr_to, 1); + error_code = (unsigned long) kaddr_to; + if (IS_ERR_VALUE(error_code)) goto fault; - } else if (!pte_write(*pte_to)) { - error_code = 0x04; - goto fault; - } - pfn_from = pte_pfn(*pte_from); - pfn_to = pte_pfn(*pte_to); - offset_from = uaddr_from & (PAGE_SIZE-1); - offset_to = uaddr_from & (PAGE_SIZE-1); - offset_max = max(offset_from, offset_to); + offset_max = max(uaddr_from & ~PAGE_MASK, + uaddr_to & ~PAGE_MASK); size = min(n - done, PAGE_SIZE - offset_max); - memcpy((void *)(pfn_to << PAGE_SHIFT) + offset_to, - (void *)(pfn_from << PAGE_SHIFT) + offset_from, size); + memcpy((void *) kaddr_to, (void *) kaddr_from, size); done += size; uaddr_from += size; uaddr_to += size; @@ -282,7 +266,7 @@ retry: return n - done; fault: spin_unlock(&mm->page_table_lock); - if (__handle_fault(uaddr, error_code, write_user)) + if (__handle_fault(uaddr, -error_code, write_user)) return n - done; goto retry; } @@ -341,7 +325,7 @@ int futex_atomic_op_pt(int op, u32 __user *uaddr, int oparg, int *old) return __futex_atomic_op_pt(op, uaddr, oparg, old); spin_lock(¤t->mm->page_table_lock); uaddr = (u32 __force __user *) - __dat_user_addr((__force unsigned long) uaddr); + __dat_user_addr((__force unsigned long) uaddr, 1); if (!uaddr) { spin_unlock(¤t->mm->page_table_lock); return -EFAULT; @@ -378,7 +362,7 @@ int futex_atomic_cmpxchg_pt(u32 *uval, u32 __user *uaddr, return __futex_atomic_cmpxchg_pt(uval, uaddr, oldval, newval); spin_lock(¤t->mm->page_table_lock); uaddr = (u32 __force __user *) - __dat_user_addr((__force unsigned long) uaddr); + __dat_user_addr((__force unsigned long) uaddr, 1); if (!uaddr) { spin_unlock(¤t->mm->page_table_lock); return -EFAULT; diff --git a/arch/s390/mm/Makefile b/arch/s390/mm/Makefile index d98fe9004a5..0f5536b0c1a 100644 --- a/arch/s390/mm/Makefile +++ b/arch/s390/mm/Makefile @@ -3,7 +3,7 @@ # obj-y := init.o fault.o extmem.o mmap.o vmem.o pgtable.o maccess.o \ - page-states.o gup.o + page-states.o gup.o extable.o obj-$(CONFIG_CMM) += cmm.o obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o obj-$(CONFIG_DEBUG_SET_MODULE_RONX) += pageattr.o diff --git a/arch/s390/mm/extable.c b/arch/s390/mm/extable.c new file mode 100644 index 00000000000..4d1ee88864e --- /dev/null +++ b/arch/s390/mm/extable.c @@ -0,0 +1,81 @@ +#include <linux/module.h> +#include <linux/sort.h> +#include <asm/uaccess.h> + +/* + * Search one exception table for an entry corresponding to the + * given instruction address, and return the address of the entry, + * or NULL if none is found. + * We use a binary search, and thus we assume that the table is + * already sorted. + */ +const struct exception_table_entry * +search_extable(const struct exception_table_entry *first, + const struct exception_table_entry *last, + unsigned long value) +{ + const struct exception_table_entry *mid; + unsigned long addr; + + while (first <= last) { + mid = ((last - first) >> 1) + first; + addr = extable_insn(mid); + if (addr < value) + first = mid + 1; + else if (addr > value) + last = mid - 1; + else + return mid; + } + return NULL; +} + +/* + * The exception table needs to be sorted so that the binary + * search that we use to find entries in it works properly. + * This is used both for the kernel exception table and for + * the exception tables of modules that get loaded. + * + */ +static int cmp_ex(const void *a, const void *b) +{ + const struct exception_table_entry *x = a, *y = b; + + /* This compare is only valid after normalization. */ + return x->insn - y->insn; +} + +void sort_extable(struct exception_table_entry *start, + struct exception_table_entry *finish) +{ + struct exception_table_entry *p; + int i; + + /* Normalize entries to being relative to the start of the section */ + for (p = start, i = 0; p < finish; p++, i += 8) + p->insn += i; + sort(start, finish - start, sizeof(*start), cmp_ex, NULL); + /* Denormalize all entries */ + for (p = start, i = 0; p < finish; p++, i += 8) + p->insn -= i; +} + +#ifdef CONFIG_MODULES +/* + * If the exception table is sorted, any referring to the module init + * will be at the beginning or the end. + */ +void trim_init_extable(struct module *m) +{ + /* Trim the beginning */ + while (m->num_exentries && + within_module_init(extable_insn(&m->extable[0]), m)) { + m->extable++; + m->num_exentries--; + } + /* Trim the end */ + while (m->num_exentries && + within_module_init(extable_insn(&m->extable[m->num_exentries-1]), m)) + m->num_exentries--; +} +#endif /* CONFIG_MODULES */ diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c index 6c013f54414..ac9122ca115 100644 --- a/arch/s390/mm/fault.c +++ b/arch/s390/mm/fault.c @@ -111,7 +111,7 @@ static inline int user_space_fault(unsigned long trans_exc_code) if (trans_exc_code == 2) /* Access via secondary space, set_fs setting decides */ return current->thread.mm_segment.ar4; - if (addressing_mode == HOME_SPACE_MODE) + if (s390_user_mode == HOME_SPACE_MODE) /* User space if the access has been done via home space. */ return trans_exc_code == 3; /* @@ -163,7 +163,7 @@ static noinline void do_no_context(struct pt_regs *regs) /* Are we prepared to handle this kernel fault? */ fixup = search_exception_tables(regs->psw.addr & PSW_ADDR_INSN); if (fixup) { - regs->psw.addr = fixup->fixup | PSW_ADDR_AMODE; + regs->psw.addr = extable_fixup(fixup) | PSW_ADDR_AMODE; return; } @@ -628,9 +628,8 @@ static int __cpuinit pfault_cpu_notify(struct notifier_block *self, struct thread_struct *thread, *next; struct task_struct *tsk; - switch (action) { + switch (action & ~CPU_TASKS_FROZEN) { case CPU_DEAD: - case CPU_DEAD_FROZEN: spin_lock_irq(&pfault_lock); list_for_each_entry_safe(thread, next, &pfault_list, list) { thread->pfault_wait = 0; diff --git a/arch/s390/mm/gup.c b/arch/s390/mm/gup.c index 65cb06e2af4..eeaf8023851 100644 --- a/arch/s390/mm/gup.c +++ b/arch/s390/mm/gup.c @@ -154,6 +154,43 @@ static inline int gup_pud_range(pgd_t *pgdp, pgd_t pgd, unsigned long addr, return 1; } +/* + * Like get_user_pages_fast() except its IRQ-safe in that it won't fall + * back to the regular GUP. + */ +int __get_user_pages_fast(unsigned long start, int nr_pages, int write, + struct page **pages) +{ + struct mm_struct *mm = current->mm; + unsigned long addr, len, end; + unsigned long next, flags; + pgd_t *pgdp, pgd; + int nr = 0; + + start &= PAGE_MASK; + addr = start; + len = (unsigned long) nr_pages << PAGE_SHIFT; + end = start + len; + if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ, + (void __user *)start, len))) + return 0; + + local_irq_save(flags); + pgdp = pgd_offset(mm, addr); + do { + pgd = *pgdp; + barrier(); + next = pgd_addr_end(addr, end); + if (pgd_none(pgd)) + break; + if (!gup_pud_range(pgdp, pgd, addr, next, write, pages, &nr)) + break; + } while (pgdp++, addr = next, addr != end); + local_irq_restore(flags); + + return nr; +} + /** * get_user_pages_fast() - pin user pages in memory * @start: starting user address diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c index 6adbc082618..81e596c65de 100644 --- a/arch/s390/mm/init.c +++ b/arch/s390/mm/init.c @@ -42,7 +42,7 @@ pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__((__aligned__(PAGE_SIZE))); unsigned long empty_zero_page, zero_page_mask; EXPORT_SYMBOL(empty_zero_page); -static unsigned long setup_zero_pages(void) +static unsigned long __init setup_zero_pages(void) { struct cpuid cpu_id; unsigned int order; @@ -212,7 +212,7 @@ void free_initmem(void) } #ifdef CONFIG_BLK_DEV_INITRD -void free_initrd_mem(unsigned long start, unsigned long end) +void __init free_initrd_mem(unsigned long start, unsigned long end) { free_init_pages("initrd memory", start, end); } diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c index 18df31d1f2c..b402991e43d 100644 --- a/arch/s390/mm/pgtable.c +++ b/arch/s390/mm/pgtable.c @@ -609,8 +609,8 @@ static inline unsigned int atomic_xor_bits(atomic_t *v, unsigned int bits) */ unsigned long *page_table_alloc(struct mm_struct *mm, unsigned long vmaddr) { - struct page *page; - unsigned long *table; + unsigned long *uninitialized_var(table); + struct page *uninitialized_var(page); unsigned int mask, bit; if (mm_has_pgste(mm)) @@ -796,7 +796,7 @@ int s390_enable_sie(void) struct mm_struct *mm, *old_mm; /* Do we have switched amode? If no, we cannot do sie */ - if (addressing_mode == HOME_SPACE_MODE) + if (s390_user_mode == HOME_SPACE_MODE) return -EINVAL; /* Do we have pgstes? if yes, we are done */ diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c index 6f896e75ab4..c22abf900c9 100644 --- a/arch/s390/mm/vmem.c +++ b/arch/s390/mm/vmem.c @@ -107,7 +107,7 @@ static int vmem_add_mem(unsigned long start, unsigned long size, int ro) pte = mk_pte_phys(address, __pgprot(ro ? _PAGE_RO : 0)); pm_dir = pmd_offset(pu_dir, address); -#ifdef CONFIG_64BIT +#if defined(CONFIG_64BIT) && !defined(CONFIG_DEBUG_PAGEALLOC) if (MACHINE_HAS_HPAGE && !(address & ~HPAGE_MASK) && (address + HPAGE_SIZE <= start + size) && (address >= HPAGE_SIZE)) { diff --git a/arch/s390/net/Makefile b/arch/s390/net/Makefile new file mode 100644 index 00000000000..90568c33ddb --- /dev/null +++ b/arch/s390/net/Makefile @@ -0,0 +1,4 @@ +# +# Arch-specific network modules +# +obj-$(CONFIG_BPF_JIT) += bpf_jit.o bpf_jit_comp.o diff --git a/arch/s390/net/bpf_jit.S b/arch/s390/net/bpf_jit.S new file mode 100644 index 00000000000..7e45d13816c --- /dev/null +++ b/arch/s390/net/bpf_jit.S @@ -0,0 +1,130 @@ +/* + * BPF Jit compiler for s390, help functions. + * + * Copyright IBM Corp. 2012 + * + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + */ +#include <linux/linkage.h> + +/* + * Calling convention: + * registers %r2, %r6-%r8, %r10-%r11, %r13, %r15 are call saved + * %r2: skb pointer + * %r3: offset parameter + * %r5: BPF A accumulator + * %r8: return address + * %r9: save register for skb pointer + * %r10: skb->data + * %r11: skb->len - skb->data_len (headlen) + * %r12: BPF X accumulator + * + * skb_copy_bits takes 4 parameters: + * %r2 = skb pointer + * %r3 = offset into skb data + * %r4 = length to copy + * %r5 = pointer to temp buffer + */ +#define SKBDATA %r8 + + /* A = *(u32 *) (skb->data+K+X) */ +ENTRY(sk_load_word_ind) + ar %r3,%r12 # offset += X + bmr %r8 # < 0 -> return with cc + + /* A = *(u32 *) (skb->data+K) */ +ENTRY(sk_load_word) + llgfr %r1,%r3 # extend offset + ahi %r3,4 # offset + 4 + clr %r11,%r3 # hlen <= offset + 4 ? + jl sk_load_word_slow + l %r5,0(%r1,%r10) # get word from skb + xr %r1,%r1 # set cc to zero + br %r8 + +sk_load_word_slow: + lgr %r9,%r2 # save %r2 + lhi %r4,4 # 4 bytes + la %r5,160(%r15) # pointer to temp buffer + brasl %r14,skb_copy_bits # get data from skb + l %r5,160(%r15) # load result from temp buffer + ltgr %r2,%r2 # set cc to (%r2 != 0) + lgr %r2,%r9 # restore %r2 + br %r8 + + /* A = *(u16 *) (skb->data+K+X) */ +ENTRY(sk_load_half_ind) + ar %r3,%r12 # offset += X + bmr %r8 # < 0 -> return with cc + + /* A = *(u16 *) (skb->data+K) */ +ENTRY(sk_load_half) + llgfr %r1,%r3 # extend offset + ahi %r3,2 # offset + 2 + clr %r11,%r3 # hlen <= offset + 2 ? + jl sk_load_half_slow + llgh %r5,0(%r1,%r10) # get half from skb + xr %r1,%r1 # set cc to zero + br %r8 + +sk_load_half_slow: + lgr %r9,%r2 # save %r2 + lhi %r4,2 # 2 bytes + la %r5,162(%r15) # pointer to temp buffer + brasl %r14,skb_copy_bits # get data from skb + xc 160(2,%r15),160(%r15) + l %r5,160(%r15) # load result from temp buffer + ltgr %r2,%r2 # set cc to (%r2 != 0) + lgr %r2,%r9 # restore %r2 + br %r8 + + /* A = *(u8 *) (skb->data+K+X) */ +ENTRY(sk_load_byte_ind) + ar %r3,%r12 # offset += X + bmr %r8 # < 0 -> return with cc + + /* A = *(u8 *) (skb->data+K) */ +ENTRY(sk_load_byte) + llgfr %r1,%r3 # extend offset + clr %r11,%r3 # hlen < offset ? + jle sk_load_byte_slow + lhi %r5,0 + ic %r5,0(%r1,%r10) # get byte from skb + xr %r1,%r1 # set cc to zero + br %r8 + +sk_load_byte_slow: + lgr %r9,%r2 # save %r2 + lhi %r4,1 # 1 bytes + la %r5,163(%r15) # pointer to temp buffer + brasl %r14,skb_copy_bits # get data from skb + xc 160(3,%r15),160(%r15) + l %r5,160(%r15) # load result from temp buffer + ltgr %r2,%r2 # set cc to (%r2 != 0) + lgr %r2,%r9 # restore %r2 + br %r8 + + /* A = (*(u8 *)(skb->data+K) & 0xf) << 2 */ +ENTRY(sk_load_byte_msh) + llgfr %r1,%r3 # extend offset + clr %r11,%r3 # hlen < offset ? + jle sk_load_byte_slow + lhi %r12,0 + ic %r12,0(%r1,%r10) # get byte from skb + nill %r12,0x0f + sll %r12,2 + xr %r1,%r1 # set cc to zero + br %r8 + +sk_load_byte_msh_slow: + lgr %r9,%r2 # save %r2 + lhi %r4,2 # 2 bytes + la %r5,162(%r15) # pointer to temp buffer + brasl %r14,skb_copy_bits # get data from skb + xc 160(3,%r15),160(%r15) + l %r12,160(%r15) # load result from temp buffer + nill %r12,0x0f + sll %r12,2 + ltgr %r2,%r2 # set cc to (%r2 != 0) + lgr %r2,%r9 # restore %r2 + br %r8 diff --git a/arch/s390/net/bpf_jit_comp.c b/arch/s390/net/bpf_jit_comp.c new file mode 100644 index 00000000000..9b355b406af --- /dev/null +++ b/arch/s390/net/bpf_jit_comp.c @@ -0,0 +1,776 @@ +/* + * BPF Jit compiler for s390. + * + * Copyright IBM Corp. 2012 + * + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + */ +#include <linux/moduleloader.h> +#include <linux/netdevice.h> +#include <linux/filter.h> +#include <asm/cacheflush.h> +#include <asm/processor.h> +#include <asm/facility.h> + +/* + * Conventions: + * %r2 = skb pointer + * %r3 = offset parameter + * %r4 = scratch register / length parameter + * %r5 = BPF A accumulator + * %r8 = return address + * %r9 = save register for skb pointer + * %r10 = skb->data + * %r11 = skb->len - skb->data_len (headlen) + * %r12 = BPF X accumulator + * %r13 = literal pool pointer + * 0(%r15) - 63(%r15) scratch memory array with BPF_MEMWORDS + */ +int bpf_jit_enable __read_mostly; + +/* + * assembly code in arch/x86/net/bpf_jit.S + */ +extern u8 sk_load_word[], sk_load_half[], sk_load_byte[], sk_load_byte_msh[]; +extern u8 sk_load_word_ind[], sk_load_half_ind[], sk_load_byte_ind[]; + +struct bpf_jit { + unsigned int seen; + u8 *start; + u8 *prg; + u8 *mid; + u8 *lit; + u8 *end; + u8 *base_ip; + u8 *ret0_ip; + u8 *exit_ip; + unsigned int off_load_word; + unsigned int off_load_half; + unsigned int off_load_byte; + unsigned int off_load_bmsh; + unsigned int off_load_iword; + unsigned int off_load_ihalf; + unsigned int off_load_ibyte; +}; + +#define BPF_SIZE_MAX 4096 /* Max size for program */ + +#define SEEN_DATAREF 1 /* might call external helpers */ +#define SEEN_XREG 2 /* ebx is used */ +#define SEEN_MEM 4 /* use mem[] for temporary storage */ +#define SEEN_RET0 8 /* pc_ret0 points to a valid return 0 */ +#define SEEN_LITERAL 16 /* code uses literals */ +#define SEEN_LOAD_WORD 32 /* code uses sk_load_word */ +#define SEEN_LOAD_HALF 64 /* code uses sk_load_half */ +#define SEEN_LOAD_BYTE 128 /* code uses sk_load_byte */ +#define SEEN_LOAD_BMSH 256 /* code uses sk_load_byte_msh */ +#define SEEN_LOAD_IWORD 512 /* code uses sk_load_word_ind */ +#define SEEN_LOAD_IHALF 1024 /* code uses sk_load_half_ind */ +#define SEEN_LOAD_IBYTE 2048 /* code uses sk_load_byte_ind */ + +#define EMIT2(op) \ +({ \ + if (jit->prg + 2 <= jit->mid) \ + *(u16 *) jit->prg = op; \ + jit->prg += 2; \ +}) + +#define EMIT4(op) \ +({ \ + if (jit->prg + 4 <= jit->mid) \ + *(u32 *) jit->prg = op; \ + jit->prg += 4; \ +}) + +#define EMIT4_DISP(op, disp) \ +({ \ + unsigned int __disp = (disp) & 0xfff; \ + EMIT4(op | __disp); \ +}) + +#define EMIT4_IMM(op, imm) \ +({ \ + unsigned int __imm = (imm) & 0xffff; \ + EMIT4(op | __imm); \ +}) + +#define EMIT4_PCREL(op, pcrel) \ +({ \ + long __pcrel = ((pcrel) >> 1) & 0xffff; \ + EMIT4(op | __pcrel); \ +}) + +#define EMIT6(op1, op2) \ +({ \ + if (jit->prg + 6 <= jit->mid) { \ + *(u32 *) jit->prg = op1; \ + *(u16 *) (jit->prg + 4) = op2; \ + } \ + jit->prg += 6; \ +}) + +#define EMIT6_DISP(op1, op2, disp) \ +({ \ + unsigned int __disp = (disp) & 0xfff; \ + EMIT6(op1 | __disp, op2); \ +}) + +#define EMIT6_IMM(op, imm) \ +({ \ + unsigned int __imm = (imm); \ + EMIT6(op | (__imm >> 16), __imm & 0xffff); \ +}) + +#define EMIT_CONST(val) \ +({ \ + unsigned int ret; \ + ret = (unsigned int) (jit->lit - jit->base_ip); \ + jit->seen |= SEEN_LITERAL; \ + if (jit->lit + 4 <= jit->end) \ + *(u32 *) jit->lit = val; \ + jit->lit += 4; \ + ret; \ +}) + +#define EMIT_FN_CONST(bit, fn) \ +({ \ + unsigned int ret; \ + ret = (unsigned int) (jit->lit - jit->base_ip); \ + if (jit->seen & bit) { \ + jit->seen |= SEEN_LITERAL; \ + if (jit->lit + 8 <= jit->end) \ + *(void **) jit->lit = fn; \ + jit->lit += 8; \ + } \ + ret; \ +}) + +static void bpf_jit_prologue(struct bpf_jit *jit) +{ + /* Save registers and create stack frame if necessary */ + if (jit->seen & SEEN_DATAREF) { + /* stmg %r8,%r15,88(%r15) */ + EMIT6(0xeb8ff058, 0x0024); + /* lgr %r14,%r15 */ + EMIT4(0xb90400ef); + /* ahi %r15,<offset> */ + EMIT4_IMM(0xa7fa0000, (jit->seen & SEEN_MEM) ? -112 : -80); + /* stg %r14,152(%r15) */ + EMIT6(0xe3e0f098, 0x0024); + } else if ((jit->seen & SEEN_XREG) && (jit->seen & SEEN_LITERAL)) + /* stmg %r12,%r13,120(%r15) */ + EMIT6(0xebcdf078, 0x0024); + else if (jit->seen & SEEN_XREG) + /* stg %r12,120(%r15) */ + EMIT6(0xe3c0f078, 0x0024); + else if (jit->seen & SEEN_LITERAL) + /* stg %r13,128(%r15) */ + EMIT6(0xe3d0f080, 0x0024); + + /* Setup literal pool */ + if (jit->seen & SEEN_LITERAL) { + /* basr %r13,0 */ + EMIT2(0x0dd0); + jit->base_ip = jit->prg; + } + jit->off_load_word = EMIT_FN_CONST(SEEN_LOAD_WORD, sk_load_word); + jit->off_load_half = EMIT_FN_CONST(SEEN_LOAD_HALF, sk_load_half); + jit->off_load_byte = EMIT_FN_CONST(SEEN_LOAD_BYTE, sk_load_byte); + jit->off_load_bmsh = EMIT_FN_CONST(SEEN_LOAD_BMSH, sk_load_byte_msh); + jit->off_load_iword = EMIT_FN_CONST(SEEN_LOAD_IWORD, sk_load_word_ind); + jit->off_load_ihalf = EMIT_FN_CONST(SEEN_LOAD_IHALF, sk_load_half_ind); + jit->off_load_ibyte = EMIT_FN_CONST(SEEN_LOAD_IBYTE, sk_load_byte_ind); + + /* Filter needs to access skb data */ + if (jit->seen & SEEN_DATAREF) { + /* l %r11,<len>(%r2) */ + EMIT4_DISP(0x58b02000, offsetof(struct sk_buff, len)); + /* s %r11,<data_len>(%r2) */ + EMIT4_DISP(0x5bb02000, offsetof(struct sk_buff, data_len)); + /* lg %r10,<data>(%r2) */ + EMIT6_DISP(0xe3a02000, 0x0004, + offsetof(struct sk_buff, data)); + } +} + +static void bpf_jit_epilogue(struct bpf_jit *jit) +{ + /* Return 0 */ + if (jit->seen & SEEN_RET0) { + jit->ret0_ip = jit->prg; + /* lghi %r2,0 */ + EMIT4(0xa7290000); + } + jit->exit_ip = jit->prg; + /* Restore registers */ + if (jit->seen & SEEN_DATAREF) + /* lmg %r8,%r15,<offset>(%r15) */ + EMIT6_DISP(0xeb8ff000, 0x0004, + (jit->seen & SEEN_MEM) ? 200 : 168); + else if ((jit->seen & SEEN_XREG) && (jit->seen & SEEN_LITERAL)) + /* lmg %r12,%r13,120(%r15) */ + EMIT6(0xebcdf078, 0x0004); + else if (jit->seen & SEEN_XREG) + /* lg %r12,120(%r15) */ + EMIT6(0xe3c0f078, 0x0004); + else if (jit->seen & SEEN_LITERAL) + /* lg %r13,128(%r15) */ + EMIT6(0xe3d0f080, 0x0004); + /* br %r14 */ + EMIT2(0x07fe); +} + +/* + * make sure we dont leak kernel information to user + */ +static void bpf_jit_noleaks(struct bpf_jit *jit, struct sock_filter *filter) +{ + /* Clear temporary memory if (seen & SEEN_MEM) */ + if (jit->seen & SEEN_MEM) + /* xc 0(64,%r15),0(%r15) */ + EMIT6(0xd73ff000, 0xf000); + /* Clear X if (seen & SEEN_XREG) */ + if (jit->seen & SEEN_XREG) + /* lhi %r12,0 */ + EMIT4(0xa7c80000); + /* Clear A if the first register does not set it. */ + switch (filter[0].code) { + case BPF_S_LD_W_ABS: + case BPF_S_LD_H_ABS: + case BPF_S_LD_B_ABS: + case BPF_S_LD_W_LEN: + case BPF_S_LD_W_IND: + case BPF_S_LD_H_IND: + case BPF_S_LD_B_IND: + case BPF_S_LDX_B_MSH: + case BPF_S_LD_IMM: + case BPF_S_LD_MEM: + case BPF_S_MISC_TXA: + case BPF_S_ANC_PROTOCOL: + case BPF_S_ANC_PKTTYPE: + case BPF_S_ANC_IFINDEX: + case BPF_S_ANC_MARK: + case BPF_S_ANC_QUEUE: + case BPF_S_ANC_HATYPE: + case BPF_S_ANC_RXHASH: + case BPF_S_ANC_CPU: + case BPF_S_RET_K: + /* first instruction sets A register */ + break; + default: /* A = 0 */ + /* lhi %r5,0 */ + EMIT4(0xa7580000); + } +} + +static int bpf_jit_insn(struct bpf_jit *jit, struct sock_filter *filter, + unsigned int *addrs, int i, int last) +{ + unsigned int K; + int offset; + unsigned int mask; + + K = filter->k; + switch (filter->code) { + case BPF_S_ALU_ADD_X: /* A += X */ + jit->seen |= SEEN_XREG; + /* ar %r5,%r12 */ + EMIT2(0x1a5c); + break; + case BPF_S_ALU_ADD_K: /* A += K */ + if (!K) + break; + if (K <= 16383) + /* ahi %r5,<K> */ + EMIT4_IMM(0xa75a0000, K); + else if (test_facility(21)) + /* alfi %r5,<K> */ + EMIT6_IMM(0xc25b0000, K); + else + /* a %r5,<d(K)>(%r13) */ + EMIT4_DISP(0x5a50d000, EMIT_CONST(K)); + break; + case BPF_S_ALU_SUB_X: /* A -= X */ + jit->seen |= SEEN_XREG; + /* sr %r5,%r12 */ + EMIT2(0x1b5c); + break; + case BPF_S_ALU_SUB_K: /* A -= K */ + if (!K) + break; + if (K <= 16384) + /* ahi %r5,-K */ + EMIT4_IMM(0xa75a0000, -K); + else if (test_facility(21)) + /* alfi %r5,-K */ + EMIT6_IMM(0xc25b0000, -K); + else + /* s %r5,<d(K)>(%r13) */ + EMIT4_DISP(0x5b50d000, EMIT_CONST(K)); + break; + case BPF_S_ALU_MUL_X: /* A *= X */ + jit->seen |= SEEN_XREG; + /* msr %r5,%r12 */ + EMIT4(0xb252005c); + break; + case BPF_S_ALU_MUL_K: /* A *= K */ + if (K <= 16383) + /* mhi %r5,K */ + EMIT4_IMM(0xa75c0000, K); + else if (test_facility(34)) + /* msfi %r5,<K> */ + EMIT6_IMM(0xc2510000, K); + else + /* ms %r5,<d(K)>(%r13) */ + EMIT4_DISP(0x7150d000, EMIT_CONST(K)); + break; + case BPF_S_ALU_DIV_X: /* A /= X */ + jit->seen |= SEEN_XREG | SEEN_RET0; + /* ltr %r12,%r12 */ + EMIT2(0x12cc); + /* jz <ret0> */ + EMIT4_PCREL(0xa7840000, (jit->ret0_ip - jit->prg)); + /* lhi %r4,0 */ + EMIT4(0xa7480000); + /* dr %r4,%r12 */ + EMIT2(0x1d4c); + break; + case BPF_S_ALU_DIV_K: /* A = reciprocal_divide(A, K) */ + /* m %r4,<d(K)>(%r13) */ + EMIT4_DISP(0x5c40d000, EMIT_CONST(K)); + /* lr %r5,%r4 */ + EMIT2(0x1854); + break; + case BPF_S_ALU_AND_X: /* A &= X */ + jit->seen |= SEEN_XREG; + /* nr %r5,%r12 */ + EMIT2(0x145c); + break; + case BPF_S_ALU_AND_K: /* A &= K */ + if (test_facility(21)) + /* nilf %r5,<K> */ + EMIT6_IMM(0xc05b0000, K); + else + /* n %r5,<d(K)>(%r13) */ + EMIT4_DISP(0x5450d000, EMIT_CONST(K)); + break; + case BPF_S_ALU_OR_X: /* A |= X */ + jit->seen |= SEEN_XREG; + /* or %r5,%r12 */ + EMIT2(0x165c); + break; + case BPF_S_ALU_OR_K: /* A |= K */ + if (test_facility(21)) + /* oilf %r5,<K> */ + EMIT6_IMM(0xc05d0000, K); + else + /* o %r5,<d(K)>(%r13) */ + EMIT4_DISP(0x5650d000, EMIT_CONST(K)); + break; + case BPF_S_ANC_ALU_XOR_X: /* A ^= X; */ + jit->seen |= SEEN_XREG; + /* xr %r5,%r12 */ + EMIT2(0x175c); + break; + case BPF_S_ALU_LSH_X: /* A <<= X; */ + jit->seen |= SEEN_XREG; + /* sll %r5,0(%r12) */ + EMIT4(0x8950c000); + break; + case BPF_S_ALU_LSH_K: /* A <<= K */ + if (K == 0) + break; + /* sll %r5,K */ + EMIT4_DISP(0x89500000, K); + break; + case BPF_S_ALU_RSH_X: /* A >>= X; */ + jit->seen |= SEEN_XREG; + /* srl %r5,0(%r12) */ + EMIT4(0x8850c000); + break; + case BPF_S_ALU_RSH_K: /* A >>= K; */ + if (K == 0) + break; + /* srl %r5,K */ + EMIT4_DISP(0x88500000, K); + break; + case BPF_S_ALU_NEG: /* A = -A */ + /* lnr %r5,%r5 */ + EMIT2(0x1155); + break; + case BPF_S_JMP_JA: /* ip += K */ + offset = addrs[i + K] + jit->start - jit->prg; + EMIT4_PCREL(0xa7f40000, offset); + break; + case BPF_S_JMP_JGT_K: /* ip += (A > K) ? jt : jf */ + mask = 0x200000; /* jh */ + goto kbranch; + case BPF_S_JMP_JGE_K: /* ip += (A >= K) ? jt : jf */ + mask = 0xa00000; /* jhe */ + goto kbranch; + case BPF_S_JMP_JEQ_K: /* ip += (A == K) ? jt : jf */ + mask = 0x800000; /* je */ +kbranch: /* Emit compare if the branch targets are different */ + if (filter->jt != filter->jf) { + if (K <= 16383) + /* chi %r5,<K> */ + EMIT4_IMM(0xa75e0000, K); + else if (test_facility(21)) + /* clfi %r5,<K> */ + EMIT6_IMM(0xc25f0000, K); + else + /* c %r5,<d(K)>(%r13) */ + EMIT4_DISP(0x5950d000, EMIT_CONST(K)); + } +branch: if (filter->jt == filter->jf) { + if (filter->jt == 0) + break; + /* j <jt> */ + offset = addrs[i + filter->jt] + jit->start - jit->prg; + EMIT4_PCREL(0xa7f40000, offset); + break; + } + if (filter->jt != 0) { + /* brc <mask>,<jt> */ + offset = addrs[i + filter->jt] + jit->start - jit->prg; + EMIT4_PCREL(0xa7040000 | mask, offset); + } + if (filter->jf != 0) { + /* brc <mask^15>,<jf> */ + offset = addrs[i + filter->jf] + jit->start - jit->prg; + EMIT4_PCREL(0xa7040000 | (mask ^ 0xf00000), offset); + } + break; + case BPF_S_JMP_JSET_K: /* ip += (A & K) ? jt : jf */ + mask = 0x700000; /* jnz */ + /* Emit test if the branch targets are different */ + if (filter->jt != filter->jf) { + if (K > 65535) { + /* lr %r4,%r5 */ + EMIT2(0x1845); + /* n %r4,<d(K)>(%r13) */ + EMIT4_DISP(0x5440d000, EMIT_CONST(K)); + } else + /* tmll %r5,K */ + EMIT4_IMM(0xa7510000, K); + } + goto branch; + case BPF_S_JMP_JGT_X: /* ip += (A > X) ? jt : jf */ + mask = 0x200000; /* jh */ + goto xbranch; + case BPF_S_JMP_JGE_X: /* ip += (A >= X) ? jt : jf */ + mask = 0xa00000; /* jhe */ + goto xbranch; + case BPF_S_JMP_JEQ_X: /* ip += (A == X) ? jt : jf */ + mask = 0x800000; /* je */ +xbranch: /* Emit compare if the branch targets are different */ + if (filter->jt != filter->jf) { + jit->seen |= SEEN_XREG; + /* cr %r5,%r12 */ + EMIT2(0x195c); + } + goto branch; + case BPF_S_JMP_JSET_X: /* ip += (A & X) ? jt : jf */ + mask = 0x700000; /* jnz */ + /* Emit test if the branch targets are different */ + if (filter->jt != filter->jf) { + jit->seen |= SEEN_XREG; + /* lr %r4,%r5 */ + EMIT2(0x1845); + /* nr %r4,%r12 */ + EMIT2(0x144c); + } + goto branch; + case BPF_S_LD_W_ABS: /* A = *(u32 *) (skb->data+K) */ + jit->seen |= SEEN_DATAREF | SEEN_RET0 | SEEN_LOAD_WORD; + offset = jit->off_load_word; + goto load_abs; + case BPF_S_LD_H_ABS: /* A = *(u16 *) (skb->data+K) */ + jit->seen |= SEEN_DATAREF | SEEN_RET0 | SEEN_LOAD_HALF; + offset = jit->off_load_half; + goto load_abs; + case BPF_S_LD_B_ABS: /* A = *(u8 *) (skb->data+K) */ + jit->seen |= SEEN_DATAREF | SEEN_RET0 | SEEN_LOAD_BYTE; + offset = jit->off_load_byte; +load_abs: if ((int) K < 0) + goto out; +call_fn: /* lg %r1,<d(function)>(%r13) */ + EMIT6_DISP(0xe310d000, 0x0004, offset); + /* l %r3,<d(K)>(%r13) */ + EMIT4_DISP(0x5830d000, EMIT_CONST(K)); + /* basr %r8,%r1 */ + EMIT2(0x0d81); + /* jnz <ret0> */ + EMIT4_PCREL(0xa7740000, (jit->ret0_ip - jit->prg)); + break; + case BPF_S_LD_W_IND: /* A = *(u32 *) (skb->data+K+X) */ + jit->seen |= SEEN_DATAREF | SEEN_RET0 | SEEN_LOAD_IWORD; + offset = jit->off_load_iword; + goto call_fn; + case BPF_S_LD_H_IND: /* A = *(u16 *) (skb->data+K+X) */ + jit->seen |= SEEN_DATAREF | SEEN_RET0 | SEEN_LOAD_IHALF; + offset = jit->off_load_ihalf; + goto call_fn; + case BPF_S_LD_B_IND: /* A = *(u8 *) (skb->data+K+X) */ + jit->seen |= SEEN_DATAREF | SEEN_RET0 | SEEN_LOAD_IBYTE; + offset = jit->off_load_ibyte; + goto call_fn; + case BPF_S_LDX_B_MSH: + /* X = (*(u8 *)(skb->data+K) & 0xf) << 2 */ + jit->seen |= SEEN_RET0; + if ((int) K < 0) { + /* j <ret0> */ + EMIT4_PCREL(0xa7f40000, (jit->ret0_ip - jit->prg)); + break; + } + jit->seen |= SEEN_DATAREF | SEEN_LOAD_BMSH; + offset = jit->off_load_bmsh; + goto call_fn; + case BPF_S_LD_W_LEN: /* A = skb->len; */ + BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, len) != 4); + /* l %r5,<d(len)>(%r2) */ + EMIT4_DISP(0x58502000, offsetof(struct sk_buff, len)); + break; + case BPF_S_LDX_W_LEN: /* X = skb->len; */ + jit->seen |= SEEN_XREG; + /* l %r12,<d(len)>(%r2) */ + EMIT4_DISP(0x58c02000, offsetof(struct sk_buff, len)); + break; + case BPF_S_LD_IMM: /* A = K */ + if (K <= 16383) + /* lhi %r5,K */ + EMIT4_IMM(0xa7580000, K); + else if (test_facility(21)) + /* llilf %r5,<K> */ + EMIT6_IMM(0xc05f0000, K); + else + /* l %r5,<d(K)>(%r13) */ + EMIT4_DISP(0x5850d000, EMIT_CONST(K)); + break; + case BPF_S_LDX_IMM: /* X = K */ + jit->seen |= SEEN_XREG; + if (K <= 16383) + /* lhi %r12,<K> */ + EMIT4_IMM(0xa7c80000, K); + else if (test_facility(21)) + /* llilf %r12,<K> */ + EMIT6_IMM(0xc0cf0000, K); + else + /* l %r12,<d(K)>(%r13) */ + EMIT4_DISP(0x58c0d000, EMIT_CONST(K)); + break; + case BPF_S_LD_MEM: /* A = mem[K] */ + jit->seen |= SEEN_MEM; + /* l %r5,<K>(%r15) */ + EMIT4_DISP(0x5850f000, + (jit->seen & SEEN_DATAREF) ? 160 + K*4 : K*4); + break; + case BPF_S_LDX_MEM: /* X = mem[K] */ + jit->seen |= SEEN_XREG | SEEN_MEM; + /* l %r12,<K>(%r15) */ + EMIT4_DISP(0x58c0f000, + (jit->seen & SEEN_DATAREF) ? 160 + K*4 : K*4); + break; + case BPF_S_MISC_TAX: /* X = A */ + jit->seen |= SEEN_XREG; + /* lr %r12,%r5 */ + EMIT2(0x18c5); + break; + case BPF_S_MISC_TXA: /* A = X */ + jit->seen |= SEEN_XREG; + /* lr %r5,%r12 */ + EMIT2(0x185c); + break; + case BPF_S_RET_K: + if (K == 0) { + jit->seen |= SEEN_RET0; + if (last) + break; + /* j <ret0> */ + EMIT4_PCREL(0xa7f40000, jit->ret0_ip - jit->prg); + } else { + if (K <= 16383) + /* lghi %r2,K */ + EMIT4_IMM(0xa7290000, K); + else + /* llgf %r2,<K>(%r13) */ + EMIT6_DISP(0xe320d000, 0x0016, EMIT_CONST(K)); + /* j <exit> */ + if (last && !(jit->seen & SEEN_RET0)) + break; + EMIT4_PCREL(0xa7f40000, jit->exit_ip - jit->prg); + } + break; + case BPF_S_RET_A: + /* llgfr %r2,%r5 */ + EMIT4(0xb9160025); + /* j <exit> */ + EMIT4_PCREL(0xa7f40000, jit->exit_ip - jit->prg); + break; + case BPF_S_ST: /* mem[K] = A */ + jit->seen |= SEEN_MEM; + /* st %r5,<K>(%r15) */ + EMIT4_DISP(0x5050f000, + (jit->seen & SEEN_DATAREF) ? 160 + K*4 : K*4); + break; + case BPF_S_STX: /* mem[K] = X : mov %ebx,off8(%rbp) */ + jit->seen |= SEEN_XREG | SEEN_MEM; + /* st %r12,<K>(%r15) */ + EMIT4_DISP(0x50c0f000, + (jit->seen & SEEN_DATAREF) ? 160 + K*4 : K*4); + break; + case BPF_S_ANC_PROTOCOL: /* A = ntohs(skb->protocol); */ + BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, protocol) != 2); + /* lhi %r5,0 */ + EMIT4(0xa7580000); + /* icm %r5,3,<d(protocol)>(%r2) */ + EMIT4_DISP(0xbf532000, offsetof(struct sk_buff, protocol)); + break; + case BPF_S_ANC_IFINDEX: /* if (!skb->dev) return 0; + * A = skb->dev->ifindex */ + BUILD_BUG_ON(FIELD_SIZEOF(struct net_device, ifindex) != 4); + jit->seen |= SEEN_RET0; + /* lg %r1,<d(dev)>(%r2) */ + EMIT6_DISP(0xe3102000, 0x0004, offsetof(struct sk_buff, dev)); + /* ltgr %r1,%r1 */ + EMIT4(0xb9020011); + /* jz <ret0> */ + EMIT4_PCREL(0xa7840000, jit->ret0_ip - jit->prg); + /* l %r5,<d(ifindex)>(%r1) */ + EMIT4_DISP(0x58501000, offsetof(struct net_device, ifindex)); + break; + case BPF_S_ANC_MARK: /* A = skb->mark */ + BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, mark) != 4); + /* l %r5,<d(mark)>(%r2) */ + EMIT4_DISP(0x58502000, offsetof(struct sk_buff, mark)); + break; + case BPF_S_ANC_QUEUE: /* A = skb->queue_mapping */ + BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, queue_mapping) != 2); + /* lhi %r5,0 */ + EMIT4(0xa7580000); + /* icm %r5,3,<d(queue_mapping)>(%r2) */ + EMIT4_DISP(0xbf532000, offsetof(struct sk_buff, queue_mapping)); + break; + case BPF_S_ANC_HATYPE: /* if (!skb->dev) return 0; + * A = skb->dev->type */ + BUILD_BUG_ON(FIELD_SIZEOF(struct net_device, type) != 2); + jit->seen |= SEEN_RET0; + /* lg %r1,<d(dev)>(%r2) */ + EMIT6_DISP(0xe3102000, 0x0004, offsetof(struct sk_buff, dev)); + /* ltgr %r1,%r1 */ + EMIT4(0xb9020011); + /* jz <ret0> */ + EMIT4_PCREL(0xa7840000, jit->ret0_ip - jit->prg); + /* lhi %r5,0 */ + EMIT4(0xa7580000); + /* icm %r5,3,<d(type)>(%r1) */ + EMIT4_DISP(0xbf531000, offsetof(struct net_device, type)); + break; + case BPF_S_ANC_RXHASH: /* A = skb->rxhash */ + BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, rxhash) != 4); + /* l %r5,<d(rxhash)>(%r2) */ + EMIT4_DISP(0x58502000, offsetof(struct sk_buff, rxhash)); + break; + case BPF_S_ANC_CPU: /* A = smp_processor_id() */ +#ifdef CONFIG_SMP + /* l %r5,<d(cpu_nr)> */ + EMIT4_DISP(0x58500000, offsetof(struct _lowcore, cpu_nr)); +#else + /* lhi %r5,0 */ + EMIT4(0xa7580000); +#endif + break; + default: /* too complex, give up */ + goto out; + } + addrs[i] = jit->prg - jit->start; + return 0; +out: + return -1; +} + +void bpf_jit_compile(struct sk_filter *fp) +{ + unsigned long size, prg_len, lit_len; + struct bpf_jit jit, cjit; + unsigned int *addrs; + int pass, i; + + if (!bpf_jit_enable) + return; + addrs = kmalloc(fp->len * sizeof(*addrs), GFP_KERNEL); + if (addrs == NULL) + return; + memset(addrs, 0, fp->len * sizeof(*addrs)); + memset(&jit, 0, sizeof(cjit)); + memset(&cjit, 0, sizeof(cjit)); + + for (pass = 0; pass < 10; pass++) { + jit.prg = jit.start; + jit.lit = jit.mid; + + bpf_jit_prologue(&jit); + bpf_jit_noleaks(&jit, fp->insns); + for (i = 0; i < fp->len; i++) { + if (bpf_jit_insn(&jit, fp->insns + i, addrs, i, + i == fp->len - 1)) + goto out; + } + bpf_jit_epilogue(&jit); + if (jit.start) { + WARN_ON(jit.prg > cjit.prg || jit.lit > cjit.lit); + if (memcmp(&jit, &cjit, sizeof(jit)) == 0) + break; + } else if (jit.prg == cjit.prg && jit.lit == cjit.lit) { + prg_len = jit.prg - jit.start; + lit_len = jit.lit - jit.mid; + size = max_t(unsigned long, prg_len + lit_len, + sizeof(struct work_struct)); + if (size >= BPF_SIZE_MAX) + goto out; + jit.start = module_alloc(size); + if (!jit.start) + goto out; + jit.prg = jit.mid = jit.start + prg_len; + jit.lit = jit.end = jit.start + prg_len + lit_len; + jit.base_ip += (unsigned long) jit.start; + jit.exit_ip += (unsigned long) jit.start; + jit.ret0_ip += (unsigned long) jit.start; + } + cjit = jit; + } + if (bpf_jit_enable > 1) { + pr_err("flen=%d proglen=%lu pass=%d image=%p\n", + fp->len, jit.end - jit.start, pass, jit.start); + if (jit.start) { + printk(KERN_ERR "JIT code:\n"); + print_fn_code(jit.start, jit.mid - jit.start); + print_hex_dump(KERN_ERR, "JIT literals:\n", + DUMP_PREFIX_ADDRESS, 16, 1, + jit.mid, jit.end - jit.mid, false); + } + } + if (jit.start) + fp->bpf_func = (void *) jit.start; +out: + kfree(addrs); +} + +static void jit_free_defer(struct work_struct *arg) +{ + module_free(NULL, arg); +} + +/* run from softirq, we must use a work_struct to call + * module_free() from process context + */ +void bpf_jit_free(struct sk_filter *fp) +{ + struct work_struct *work; + + if (fp->bpf_func == sk_run_filter) + return; + work = (struct work_struct *)fp->bpf_func; + INIT_WORK(work, jit_free_defer); + schedule_work(work); +} |