diff options
Diffstat (limited to 'arch/s390/lib')
29 files changed, 2277 insertions, 1864 deletions
diff --git a/arch/s390/lib/Makefile b/arch/s390/lib/Makefile index 20b0e97a7df2..f43f897d3fc0 100644 --- a/arch/s390/lib/Makefile +++ b/arch/s390/lib/Makefile @@ -1,9 +1,26 @@ +# SPDX-License-Identifier: GPL-2.0 # # Makefile for s390-specific library files.. # -lib-y += delay.o string.o uaccess_std.o uaccess_pt.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 +lib-y += delay.o string.o uaccess.o find.o spinlock.o tishift.o +lib-y += csum-partial.o +obj-y += mem.o xor.o +lib-$(CONFIG_KPROBES) += probes.o +lib-$(CONFIG_UPROBES) += probes.o +obj-$(CONFIG_S390_KPROBES_SANITY_TEST) += test_kprobes_s390.o +test_kprobes_s390-objs += test_kprobes_asm.o test_kprobes.o + +# Instrumenting memory accesses to __user data (in different address space) +# produce false positives +KASAN_SANITIZE_uaccess.o := n + +obj-$(CONFIG_S390_UNWIND_SELFTEST) += test_unwind.o +CFLAGS_test_unwind.o += -fno-optimize-sibling-calls + +obj-$(CONFIG_S390_MODULES_SANITY_TEST) += test_modules.o +obj-$(CONFIG_S390_MODULES_SANITY_TEST_HELPERS) += test_modules_helpers.o + +lib-$(CONFIG_FUNCTION_ERROR_INJECTION) += error-inject.o + +obj-$(CONFIG_EXPOLINE_EXTERN) += expoline.o diff --git a/arch/s390/lib/csum-partial.c b/arch/s390/lib/csum-partial.c new file mode 100644 index 000000000000..458abd9bac70 --- /dev/null +++ b/arch/s390/lib/csum-partial.c @@ -0,0 +1,91 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <linux/export.h> +#include <asm/checksum.h> +#include <asm/fpu.h> + +/* + * Computes the checksum of a memory block at src, length len, + * and adds in "sum" (32-bit). If copy is true copies to dst. + * + * Returns a 32-bit number suitable for feeding into itself + * or csum_tcpudp_magic. + * + * This function must be called with even lengths, except + * for the last fragment, which may be odd. + * + * It's best to have src and dst aligned on a 64-bit boundary. + */ +static __always_inline __wsum csum_copy(void *dst, const void *src, int len, __wsum sum, bool copy) +{ + DECLARE_KERNEL_FPU_ONSTACK8(vxstate); + + if (!cpu_has_vx()) { + if (copy) + memcpy(dst, src, len); + return cksm(dst, len, sum); + } + kernel_fpu_begin(&vxstate, KERNEL_VXR_V16V23); + fpu_vlvgf(16, (__force u32)sum, 1); + fpu_vzero(17); + fpu_vzero(18); + fpu_vzero(19); + while (len >= 64) { + fpu_vlm(20, 23, src); + if (copy) { + fpu_vstm(20, 23, dst); + dst += 64; + } + fpu_vcksm(16, 20, 16); + fpu_vcksm(17, 21, 17); + fpu_vcksm(18, 22, 18); + fpu_vcksm(19, 23, 19); + src += 64; + len -= 64; + } + while (len >= 32) { + fpu_vlm(20, 21, src); + if (copy) { + fpu_vstm(20, 21, dst); + dst += 32; + } + fpu_vcksm(16, 20, 16); + fpu_vcksm(17, 21, 17); + src += 32; + len -= 32; + } + while (len >= 16) { + fpu_vl(20, src); + if (copy) { + fpu_vst(20, dst); + dst += 16; + } + fpu_vcksm(16, 20, 16); + src += 16; + len -= 16; + } + if (len) { + fpu_vll(20, len - 1, src); + if (copy) + fpu_vstl(20, len - 1, dst); + fpu_vcksm(16, 20, 16); + } + fpu_vcksm(18, 19, 18); + fpu_vcksm(16, 17, 16); + fpu_vcksm(16, 18, 16); + sum = (__force __wsum)fpu_vlgvf(16, 1); + kernel_fpu_end(&vxstate, KERNEL_VXR_V16V23); + return sum; +} + +__wsum csum_partial(const void *buff, int len, __wsum sum) +{ + return csum_copy(NULL, buff, len, sum, false); +} +EXPORT_SYMBOL(csum_partial); + +__wsum csum_partial_copy_nocheck(const void *src, void *dst, int len) +{ + return csum_copy(dst, src, len, 0, true); +} +EXPORT_SYMBOL(csum_partial_copy_nocheck); diff --git a/arch/s390/lib/delay.c b/arch/s390/lib/delay.c index c61b9fad43cc..c1ea14e3c927 100644 --- a/arch/s390/lib/delay.c +++ b/arch/s390/lib/delay.c @@ -1,131 +1,46 @@ +// SPDX-License-Identifier: GPL-2.0 /* * Precise Delay Loops for S390 * * Copyright IBM Corp. 1999, 2008 * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>, - * Heiko Carstens <heiko.carstens@de.ibm.com>, */ -#include <linux/sched.h> +#include <linux/processor.h> +#include <linux/export.h> #include <linux/delay.h> -#include <linux/timex.h> -#include <linux/module.h> -#include <linux/irqflags.h> -#include <linux/interrupt.h> -#include <asm/vtimer.h> #include <asm/div64.h> +#include <asm/timex.h> void __delay(unsigned long loops) { - /* - * To end the bloody studid and useless discussion about the - * BogoMips number I took the liberty to define the __delay - * function in a way that that resulting BogoMips number will - * yield the megahertz number of the cpu. The important function - * is udelay and that is done using the tod clock. -- martin. - */ + /* + * Loop 'loops' times. Callers must not assume a specific + * amount of time passes before this function returns. + */ asm volatile("0: brct %0,0b" : : "d" ((loops/2) + 1)); } +EXPORT_SYMBOL(__delay); -static void __udelay_disabled(unsigned long long usecs) +static void delay_loop(unsigned long delta) { - unsigned long cr0, cr6, new; - u64 clock_saved, end; + unsigned long end; - end = get_tod_clock() + (usecs << 12); - clock_saved = local_tick_disable(); - __ctl_store(cr0, 0, 0); - __ctl_store(cr6, 6, 6); - new = (cr0 & 0xffff00e0) | 0x00000800; - __ctl_load(new , 0, 0); - new = 0; - __ctl_load(new, 6, 6); - lockdep_off(); - do { - set_clock_comparator(end); - vtime_stop_cpu(); - local_irq_disable(); - } while (get_tod_clock() < end); - lockdep_on(); - __ctl_load(cr0, 0, 0); - __ctl_load(cr6, 6, 6); - local_tick_enable(clock_saved); -} - -static void __udelay_enabled(unsigned long long usecs) -{ - u64 clock_saved, end; - - end = get_tod_clock() + (usecs << 12); - do { - clock_saved = 0; - if (end < S390_lowcore.clock_comparator) { - clock_saved = local_tick_disable(); - set_clock_comparator(end); - } - vtime_stop_cpu(); - local_irq_disable(); - if (clock_saved) - local_tick_enable(clock_saved); - } while (get_tod_clock() < end); + end = get_tod_clock_monotonic() + delta; + while (!tod_after(get_tod_clock_monotonic(), end)) + cpu_relax(); } -/* - * Waits for 'usecs' microseconds using the TOD clock comparator. - */ -void __udelay(unsigned long long usecs) +void __udelay(unsigned long usecs) { - unsigned long flags; - - preempt_disable(); - local_irq_save(flags); - if (in_irq()) { - __udelay_disabled(usecs); - goto out; - } - if (in_softirq()) { - if (raw_irqs_disabled_flags(flags)) - __udelay_disabled(usecs); - else - __udelay_enabled(usecs); - goto out; - } - if (raw_irqs_disabled_flags(flags)) { - local_bh_disable(); - __udelay_disabled(usecs); - _local_bh_enable(); - goto out; - } - __udelay_enabled(usecs); -out: - local_irq_restore(flags); - preempt_enable(); + delay_loop(usecs << 12); } EXPORT_SYMBOL(__udelay); -/* - * Simple udelay variant. To be used on startup and reboot - * when the interrupt handler isn't working. - */ -void udelay_simple(unsigned long long usecs) -{ - u64 end; - - end = get_tod_clock() + (usecs << 12); - while (get_tod_clock() < end) - cpu_relax(); -} - -void __ndelay(unsigned long long nsecs) +void __ndelay(unsigned long nsecs) { - u64 end; - nsecs <<= 9; do_div(nsecs, 125); - end = get_tod_clock() + nsecs; - if (nsecs & ~0xfffUL) - __udelay(nsecs >> 12); - while (get_tod_clock() < end) - barrier(); + delay_loop(nsecs); } EXPORT_SYMBOL(__ndelay); diff --git a/arch/s390/lib/div64.c b/arch/s390/lib/div64.c deleted file mode 100644 index 261152f83242..000000000000 --- a/arch/s390/lib/div64.c +++ /dev/null @@ -1,147 +0,0 @@ -/* - * __div64_32 implementation for 31 bit. - * - * Copyright IBM Corp. 2006 - * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), - */ - -#include <linux/types.h> -#include <linux/module.h> - -#ifdef CONFIG_MARCH_G5 - -/* - * Function to divide an unsigned 64 bit integer by an unsigned - * 31 bit integer using signed 64/32 bit division. - */ -static uint32_t __div64_31(uint64_t *n, uint32_t base) -{ - register uint32_t reg2 asm("2"); - register uint32_t reg3 asm("3"); - uint32_t *words = (uint32_t *) n; - uint32_t tmp; - - /* Special case base==1, remainder = 0, quotient = n */ - if (base == 1) - return 0; - /* - * Special case base==0 will cause a fixed point divide exception - * on the dr instruction and may not happen anyway. For the - * following calculation we can assume base > 1. The first - * signed 64 / 32 bit division with an upper half of 0 will - * give the correct upper half of the 64 bit quotient. - */ - reg2 = 0UL; - reg3 = words[0]; - asm volatile( - " dr %0,%2\n" - : "+d" (reg2), "+d" (reg3) : "d" (base) : "cc" ); - words[0] = reg3; - reg3 = words[1]; - /* - * To get the lower half of the 64 bit quotient and the 32 bit - * remainder we have to use a little trick. Since we only have - * a signed division the quotient can get too big. To avoid this - * the 64 bit dividend is halved, then the signed division will - * work. Afterwards the quotient and the remainder are doubled. - * If the last bit of the dividend has been one the remainder - * is increased by one then checked against the base. If the - * remainder has overflown subtract base and increase the - * quotient. Simple, no ? - */ - asm volatile( - " nr %2,%1\n" - " srdl %0,1\n" - " dr %0,%3\n" - " alr %0,%0\n" - " alr %1,%1\n" - " alr %0,%2\n" - " clr %0,%3\n" - " jl 0f\n" - " slr %0,%3\n" - " ahi %1,1\n" - "0:\n" - : "+d" (reg2), "+d" (reg3), "=d" (tmp) - : "d" (base), "2" (1UL) : "cc" ); - words[1] = reg3; - return reg2; -} - -/* - * Function to divide an unsigned 64 bit integer by an unsigned - * 32 bit integer using the unsigned 64/31 bit division. - */ -uint32_t __div64_32(uint64_t *n, uint32_t base) -{ - uint32_t r; - - /* - * If the most significant bit of base is set, divide n by - * (base/2). That allows to use 64/31 bit division and gives a - * good approximation of the result: n = (base/2)*q + r. The - * result needs to be corrected with two simple transformations. - * If base is already < 2^31-1 __div64_31 can be used directly. - */ - r = __div64_31(n, ((signed) base < 0) ? (base/2) : base); - if ((signed) base < 0) { - uint64_t q = *n; - /* - * First transformation: - * n = (base/2)*q + r - * = ((base/2)*2)*(q/2) + ((q&1) ? (base/2) : 0) + r - * Since r < (base/2), r + (base/2) < base. - * With q1 = (q/2) and r1 = r + ((q&1) ? (base/2) : 0) - * n = ((base/2)*2)*q1 + r1 with r1 < base. - */ - if (q & 1) - r += base/2; - q >>= 1; - /* - * Second transformation. ((base/2)*2) could have lost the - * last bit. - * n = ((base/2)*2)*q1 + r1 - * = base*q1 - ((base&1) ? q1 : 0) + r1 - */ - if (base & 1) { - int64_t rx = r - q; - /* - * base is >= 2^31. The worst case for the while - * loop is n=2^64-1 base=2^31+1. That gives a - * maximum for q=(2^64-1)/2^31 = 0x1ffffffff. Since - * base >= 2^31 the loop is finished after a maximum - * of three iterations. - */ - while (rx < 0) { - rx += base; - q--; - } - r = rx; - } - *n = q; - } - return r; -} - -#else /* MARCH_G5 */ - -uint32_t __div64_32(uint64_t *n, uint32_t base) -{ - register uint32_t reg2 asm("2"); - register uint32_t reg3 asm("3"); - uint32_t *words = (uint32_t *) n; - - reg2 = 0UL; - reg3 = words[0]; - asm volatile( - " dlr %0,%2\n" - : "+d" (reg2), "+d" (reg3) : "d" (base) : "cc" ); - words[0] = reg3; - reg3 = words[1]; - asm volatile( - " dlr %0,%2\n" - : "+d" (reg2), "+d" (reg3) : "d" (base) : "cc" ); - words[1] = reg3; - return reg2; -} - -#endif /* MARCH_G5 */ diff --git a/arch/s390/lib/error-inject.c b/arch/s390/lib/error-inject.c new file mode 100644 index 000000000000..8c9d4da87eef --- /dev/null +++ b/arch/s390/lib/error-inject.c @@ -0,0 +1,14 @@ +// SPDX-License-Identifier: GPL-2.0+ +#include <asm/ptrace.h> +#include <linux/error-injection.h> +#include <linux/kprobes.h> + +void override_function_with_return(struct pt_regs *regs) +{ + /* + * Emulate 'br 14'. 'regs' is captured by kprobes on entry to some + * kernel function. + */ + regs->psw.addr = regs->gprs[14]; +} +NOKPROBE_SYMBOL(override_function_with_return); diff --git a/arch/s390/lib/expoline.S b/arch/s390/lib/expoline.S new file mode 100644 index 000000000000..92ed8409a7a4 --- /dev/null +++ b/arch/s390/lib/expoline.S @@ -0,0 +1,12 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#include <asm/nospec-insn.h> +#include <linux/linkage.h> + +.macro GEN_ALL_BR_THUNK_EXTERN + .irp r1,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 + GEN_BR_THUNK_EXTERN %r\r1 + .endr +.endm + +GEN_ALL_BR_THUNK_EXTERN diff --git a/arch/s390/lib/find.c b/arch/s390/lib/find.c new file mode 100644 index 000000000000..96a8a2e2d067 --- /dev/null +++ b/arch/s390/lib/find.c @@ -0,0 +1,76 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * MSB0 numbered special bitops handling. + * + * The bits are numbered: + * |0..............63|64............127|128...........191|192...........255| + * + * The reason for this bit numbering is the fact that the hardware sets bits + * in a bitmap starting at bit 0 (MSB) and we don't want to scan the bitmap + * from the 'wrong end'. + */ + +#include <linux/compiler.h> +#include <linux/bitops.h> +#include <linux/export.h> + +unsigned long find_first_bit_inv(const unsigned long *addr, unsigned long size) +{ + const unsigned long *p = addr; + unsigned long result = 0; + unsigned long tmp; + + while (size & ~(BITS_PER_LONG - 1)) { + if ((tmp = *(p++))) + goto found; + result += BITS_PER_LONG; + size -= BITS_PER_LONG; + } + if (!size) + return result; + tmp = (*p) & (~0UL << (BITS_PER_LONG - size)); + if (!tmp) /* Are any bits set? */ + return result + size; /* Nope. */ +found: + return result + (__fls(tmp) ^ (BITS_PER_LONG - 1)); +} +EXPORT_SYMBOL(find_first_bit_inv); + +unsigned long find_next_bit_inv(const unsigned long *addr, unsigned long size, + unsigned long offset) +{ + const unsigned long *p = addr + (offset / BITS_PER_LONG); + unsigned long result = offset & ~(BITS_PER_LONG - 1); + unsigned long tmp; + + if (offset >= size) + return size; + size -= result; + offset %= BITS_PER_LONG; + if (offset) { + tmp = *(p++); + tmp &= (~0UL >> offset); + if (size < BITS_PER_LONG) + goto found_first; + if (tmp) + goto found_middle; + size -= BITS_PER_LONG; + result += BITS_PER_LONG; + } + while (size & ~(BITS_PER_LONG-1)) { + if ((tmp = *(p++))) + goto found_middle; + result += BITS_PER_LONG; + size -= BITS_PER_LONG; + } + if (!size) + return result; + tmp = *p; +found_first: + tmp &= (~0UL << (BITS_PER_LONG - size)); + if (!tmp) /* Are any bits set? */ + return result + size; /* Nope. */ +found_middle: + return result + (__fls(tmp) ^ (BITS_PER_LONG - 1)); +} +EXPORT_SYMBOL(find_next_bit_inv); diff --git a/arch/s390/lib/mem.S b/arch/s390/lib/mem.S new file mode 100644 index 000000000000..d026debf250c --- /dev/null +++ b/arch/s390/lib/mem.S @@ -0,0 +1,192 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * String handling functions. + * + * Copyright IBM Corp. 2012 + */ + +#include <linux/export.h> +#include <linux/linkage.h> +#include <asm/nospec-insn.h> + + GEN_BR_THUNK %r14 + +/* + * void *memmove(void *dest, const void *src, size_t n) + */ +SYM_FUNC_START(__memmove) + ltgr %r4,%r4 + lgr %r1,%r2 + jz .Lmemmove_exit + aghi %r4,-1 + clgr %r2,%r3 + jnh .Lmemmove_forward + la %r5,1(%r4,%r3) + clgr %r2,%r5 + jl .Lmemmove_reverse +.Lmemmove_forward: + srlg %r0,%r4,8 + ltgr %r0,%r0 + jz .Lmemmove_forward_remainder +.Lmemmove_forward_loop: + mvc 0(256,%r1),0(%r3) + la %r1,256(%r1) + la %r3,256(%r3) + brctg %r0,.Lmemmove_forward_loop +.Lmemmove_forward_remainder: + exrl %r4,.Lmemmove_mvc +.Lmemmove_exit: + BR_EX %r14 +.Lmemmove_reverse: + ic %r0,0(%r4,%r3) + stc %r0,0(%r4,%r1) + brctg %r4,.Lmemmove_reverse + ic %r0,0(%r4,%r3) + stc %r0,0(%r4,%r1) + BR_EX %r14 +.Lmemmove_mvc: + mvc 0(1,%r1),0(%r3) +SYM_FUNC_END(__memmove) +EXPORT_SYMBOL(__memmove) + +SYM_FUNC_ALIAS(memmove, __memmove) +EXPORT_SYMBOL(memmove) + +/* + * 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); + * } + */ +SYM_FUNC_START(__memset) + ltgr %r4,%r4 + jz .Lmemset_exit + ltgr %r3,%r3 + jnz .Lmemset_fill + aghi %r4,-1 + srlg %r3,%r4,8 + ltgr %r3,%r3 + lgr %r1,%r2 + jz .Lmemset_clear_remainder +.Lmemset_clear_loop: + xc 0(256,%r1),0(%r1) + la %r1,256(%r1) + brctg %r3,.Lmemset_clear_loop +.Lmemset_clear_remainder: + exrl %r4,.Lmemset_xc +.Lmemset_exit: + BR_EX %r14 +.Lmemset_fill: + cghi %r4,1 + lgr %r1,%r2 + je .Lmemset_fill_exit + aghi %r4,-2 + srlg %r5,%r4,8 + ltgr %r5,%r5 + jz .Lmemset_fill_remainder +.Lmemset_fill_loop: + stc %r3,0(%r1) + mvc 1(255,%r1),0(%r1) + la %r1,256(%r1) + brctg %r5,.Lmemset_fill_loop +.Lmemset_fill_remainder: + stc %r3,0(%r1) + exrl %r4,.Lmemset_mvc + BR_EX %r14 +.Lmemset_fill_exit: + stc %r3,0(%r1) + BR_EX %r14 +.Lmemset_xc: + xc 0(1,%r1),0(%r1) +.Lmemset_mvc: + mvc 1(1,%r1),0(%r1) +SYM_FUNC_END(__memset) +EXPORT_SYMBOL(__memset) + +SYM_FUNC_ALIAS(memset, __memset) +EXPORT_SYMBOL(memset) + +/* + * memcpy implementation + * + * void *memcpy(void *dest, const void *src, size_t n) + */ +SYM_FUNC_START(__memcpy) + ltgr %r4,%r4 + jz .Lmemcpy_exit + aghi %r4,-1 + srlg %r5,%r4,8 + ltgr %r5,%r5 + lgr %r1,%r2 + jnz .Lmemcpy_loop +.Lmemcpy_remainder: + exrl %r4,.Lmemcpy_mvc +.Lmemcpy_exit: + BR_EX %r14 +.Lmemcpy_loop: + mvc 0(256,%r1),0(%r3) + la %r1,256(%r1) + la %r3,256(%r3) + brctg %r5,.Lmemcpy_loop + j .Lmemcpy_remainder +.Lmemcpy_mvc: + mvc 0(1,%r1),0(%r3) +SYM_FUNC_END(__memcpy) +EXPORT_SYMBOL(__memcpy) + +SYM_FUNC_ALIAS(memcpy, __memcpy) +EXPORT_SYMBOL(memcpy) + +/* + * __memset16/32/64 + * + * void *__memset16(uint16_t *s, uint16_t v, size_t count) + * void *__memset32(uint32_t *s, uint32_t v, size_t count) + * void *__memset64(uint64_t *s, uint64_t v, size_t count) + */ +.macro __MEMSET bits,bytes,insn +SYM_FUNC_START(__memset\bits) + ltgr %r4,%r4 + jz .L__memset_exit\bits + cghi %r4,\bytes + je .L__memset_store\bits + aghi %r4,-(\bytes+1) + srlg %r5,%r4,8 + ltgr %r5,%r5 + lgr %r1,%r2 + jz .L__memset_remainder\bits +.L__memset_loop\bits: + \insn %r3,0(%r1) + mvc \bytes(256-\bytes,%r1),0(%r1) + la %r1,256(%r1) + brctg %r5,.L__memset_loop\bits +.L__memset_remainder\bits: + \insn %r3,0(%r1) + exrl %r4,.L__memset_mvc\bits + BR_EX %r14 +.L__memset_store\bits: + \insn %r3,0(%r2) +.L__memset_exit\bits: + BR_EX %r14 +.L__memset_mvc\bits: + mvc \bytes(1,%r1),0(%r1) +SYM_FUNC_END(__memset\bits) +.endm + +__MEMSET 16,2,sth +EXPORT_SYMBOL(__memset16) + +__MEMSET 32,4,st +EXPORT_SYMBOL(__memset32) + +__MEMSET 64,8,stg +EXPORT_SYMBOL(__memset64) diff --git a/arch/s390/lib/mem32.S b/arch/s390/lib/mem32.S deleted file mode 100644 index 14ca9244b615..000000000000 --- a/arch/s390/lib/mem32.S +++ /dev/null @@ -1,92 +0,0 @@ -/* - * 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 deleted file mode 100644 index c6d553e85ab1..000000000000 --- a/arch/s390/lib/mem64.S +++ /dev/null @@ -1,88 +0,0 @@ -/* - * 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/probes.c b/arch/s390/lib/probes.c new file mode 100644 index 000000000000..1e184a03442d --- /dev/null +++ b/arch/s390/lib/probes.c @@ -0,0 +1,161 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Common helper functions for kprobes and uprobes + * + * Copyright IBM Corp. 2014 + */ + +#include <linux/errno.h> +#include <asm/kprobes.h> +#include <asm/dis.h> + +int probe_is_prohibited_opcode(u16 *insn) +{ + if (!is_known_insn((unsigned char *)insn)) + return -EINVAL; + switch (insn[0] >> 8) { + case 0x0c: /* bassm */ + case 0x0b: /* bsm */ + case 0x83: /* diag */ + case 0x44: /* ex */ + case 0xac: /* stnsm */ + case 0xad: /* stosm */ + return -EINVAL; + case 0xc6: + switch (insn[0] & 0x0f) { + case 0x00: /* exrl */ + return -EINVAL; + } + } + switch (insn[0]) { + case 0x0101: /* pr */ + case 0xb25a: /* bsa */ + case 0xb240: /* bakr */ + case 0xb258: /* bsg */ + case 0xb218: /* pc */ + case 0xb228: /* pt */ + case 0xb98d: /* epsw */ + case 0xe560: /* tbegin */ + case 0xe561: /* tbeginc */ + case 0xb2f8: /* tend */ + return -EINVAL; + } + return 0; +} + +int probe_get_fixup_type(u16 *insn) +{ + /* default fixup method */ + int fixup = FIXUP_PSW_NORMAL; + + switch (insn[0] >> 8) { + case 0x05: /* balr */ + case 0x0d: /* basr */ + fixup = FIXUP_RETURN_REGISTER; + /* if r2 = 0, no branch will be taken */ + if ((insn[0] & 0x0f) == 0) + fixup |= FIXUP_BRANCH_NOT_TAKEN; + break; + case 0x06: /* bctr */ + case 0x07: /* bcr */ + fixup = FIXUP_BRANCH_NOT_TAKEN; + break; + case 0x45: /* bal */ + case 0x4d: /* bas */ + fixup = FIXUP_RETURN_REGISTER; + break; + case 0x47: /* bc */ + case 0x46: /* bct */ + case 0x86: /* bxh */ + case 0x87: /* bxle */ + fixup = FIXUP_BRANCH_NOT_TAKEN; + break; + case 0x82: /* lpsw */ + fixup = FIXUP_NOT_REQUIRED; + break; + case 0xb2: /* lpswe */ + if ((insn[0] & 0xff) == 0xb2) + fixup = FIXUP_NOT_REQUIRED; + break; + case 0xa7: /* bras */ + if ((insn[0] & 0x0f) == 0x05) + fixup |= FIXUP_RETURN_REGISTER; + break; + case 0xc0: + if ((insn[0] & 0x0f) == 0x05) /* brasl */ + fixup |= FIXUP_RETURN_REGISTER; + break; + case 0xeb: + switch (insn[2] & 0xff) { + case 0x44: /* bxhg */ + case 0x45: /* bxleg */ + fixup = FIXUP_BRANCH_NOT_TAKEN; + break; + } + break; + case 0xe3: /* bctg */ + if ((insn[2] & 0xff) == 0x46) + fixup = FIXUP_BRANCH_NOT_TAKEN; + break; + case 0xec: + switch (insn[2] & 0xff) { + case 0xe5: /* clgrb */ + case 0xe6: /* cgrb */ + case 0xf6: /* crb */ + case 0xf7: /* clrb */ + case 0xfc: /* cgib */ + case 0xfd: /* cglib */ + case 0xfe: /* cib */ + case 0xff: /* clib */ + fixup = FIXUP_BRANCH_NOT_TAKEN; + break; + } + break; + } + return fixup; +} + +int probe_is_insn_relative_long(u16 *insn) +{ + /* Check if we have a RIL-b or RIL-c format instruction which + * we need to modify in order to avoid instruction emulation. */ + switch (insn[0] >> 8) { + case 0xc0: + if ((insn[0] & 0x0f) == 0x00) /* larl */ + return true; + break; + case 0xc4: + switch (insn[0] & 0x0f) { + case 0x02: /* llhrl */ + case 0x04: /* lghrl */ + case 0x05: /* lhrl */ + case 0x06: /* llghrl */ + case 0x07: /* sthrl */ + case 0x08: /* lgrl */ + case 0x0b: /* stgrl */ + case 0x0c: /* lgfrl */ + case 0x0d: /* lrl */ + case 0x0e: /* llgfrl */ + case 0x0f: /* strl */ + return true; + } + break; + case 0xc6: + switch (insn[0] & 0x0f) { + case 0x02: /* pfdrl */ + case 0x04: /* cghrl */ + case 0x05: /* chrl */ + case 0x06: /* clghrl */ + case 0x07: /* clhrl */ + case 0x08: /* cgrl */ + case 0x0a: /* clgrl */ + case 0x0c: /* cgfrl */ + case 0x0d: /* crl */ + case 0x0e: /* clgfrl */ + case 0x0f: /* clrl */ + return true; + } + break; + } + return false; +} diff --git a/arch/s390/lib/qrnnd.S b/arch/s390/lib/qrnnd.S deleted file mode 100644 index d321329130ec..000000000000 --- a/arch/s390/lib/qrnnd.S +++ /dev/null @@ -1,78 +0,0 @@ -# S/390 __udiv_qrnnd - -#include <linux/linkage.h> - -# r2 : &__r -# r3 : upper half of 64 bit word n -# r4 : lower half of 64 bit word n -# r5 : divisor d -# the reminder r of the division is to be stored to &__r and -# the quotient q is to be returned - - .text -ENTRY(__udiv_qrnnd) - st %r2,24(%r15) # store pointer to reminder for later - lr %r0,%r3 # reload n - lr %r1,%r4 - ltr %r2,%r5 # reload and test divisor - jp 5f - # divisor >= 0x80000000 - srdl %r0,2 # n/4 - srl %r2,1 # d/2 - slr %r1,%r2 # special case if last bit of d is set - brc 3,0f # (n/4) div (n/2) can overflow by 1 - ahi %r0,-1 # trick: subtract n/2, then divide -0: dr %r0,%r2 # signed division - ahi %r1,1 # trick part 2: add 1 to the quotient - # now (n >> 2) = (d >> 1) * %r1 + %r0 - lhi %r3,1 - nr %r3,%r1 # test last bit of q - jz 1f - alr %r0,%r2 # add (d>>1) to r -1: srl %r1,1 # q >>= 1 - # now (n >> 2) = (d&-2) * %r1 + %r0 - lhi %r3,1 - nr %r3,%r5 # test last bit of d - jz 2f - slr %r0,%r1 # r -= q - brc 3,2f # borrow ? - alr %r0,%r5 # r += d - ahi %r1,-1 -2: # now (n >> 2) = d * %r1 + %r0 - alr %r1,%r1 # q <<= 1 - alr %r0,%r0 # r <<= 1 - brc 12,3f # overflow on r ? - slr %r0,%r5 # r -= d - ahi %r1,1 # q += 1 -3: lhi %r3,2 - nr %r3,%r4 # test next to last bit of n - jz 4f - ahi %r0,1 # r += 1 -4: clr %r0,%r5 # r >= d ? - jl 6f - slr %r0,%r5 # r -= d - ahi %r1,1 # q += 1 - # now (n >> 1) = d * %r1 + %r0 - j 6f -5: # divisor < 0x80000000 - srdl %r0,1 - dr %r0,%r2 # signed division - # now (n >> 1) = d * %r1 + %r0 -6: alr %r1,%r1 # q <<= 1 - alr %r0,%r0 # r <<= 1 - brc 12,7f # overflow on r ? - slr %r0,%r5 # r -= d - ahi %r1,1 # q += 1 -7: lhi %r3,1 - nr %r3,%r4 # isolate last bit of n - alr %r0,%r3 # r += (n & 1) - clr %r0,%r5 # r >= d ? - jl 8f - slr %r0,%r5 # r -= d - ahi %r1,1 # q += 1 -8: # now n = d * %r1 + %r0 - l %r2,24(%r15) - st %r0,0(%r2) - lr %r2,%r1 - br %r14 - .end __udiv_qrnnd diff --git a/arch/s390/lib/spinlock.c b/arch/s390/lib/spinlock.c index f709983f41f8..10db1e56a811 100644 --- a/arch/s390/lib/spinlock.c +++ b/arch/s390/lib/spinlock.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: GPL-2.0 /* * Out of line spinlock code. * @@ -6,15 +7,29 @@ */ #include <linux/types.h> -#include <linux/module.h> +#include <linux/export.h> #include <linux/spinlock.h> +#include <linux/jiffies.h> +#include <linux/sysctl.h> #include <linux/init.h> #include <linux/smp.h> -#include <asm/io.h> +#include <linux/percpu.h> +#include <linux/io.h> +#include <asm/alternative.h> +#include <asm/machine.h> +#include <asm/asm.h> -int spin_retry = 1000; +int spin_retry = -1; -/** +static int __init spin_retry_init(void) +{ + if (spin_retry < 0) + spin_retry = 1000; + return 0; +} +early_initcall(spin_retry_init); + +/* * spin_retry= parameter */ static int __init spin_retry_setup(char *str) @@ -24,195 +39,325 @@ static int __init spin_retry_setup(char *str) } __setup("spin_retry=", spin_retry_setup); -void arch_spin_lock_wait(arch_spinlock_t *lp) +static const struct ctl_table s390_spin_sysctl_table[] = { + { + .procname = "spin_retry", + .data = &spin_retry, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec, + }, +}; + +static int __init init_s390_spin_sysctls(void) { - int count = spin_retry; - unsigned int cpu = ~smp_processor_id(); - unsigned int owner; + register_sysctl_init("kernel", s390_spin_sysctl_table); + return 0; +} +arch_initcall(init_s390_spin_sysctls); - while (1) { - owner = lp->owner_cpu; - if (!owner || smp_vcpu_scheduled(~owner)) { - for (count = spin_retry; count > 0; count--) { - if (arch_spin_is_locked(lp)) - continue; - if (_raw_compare_and_swap(&lp->owner_cpu, 0, - cpu) == 0) - return; - } - if (MACHINE_IS_LPAR) - continue; - } - owner = lp->owner_cpu; - if (owner) - smp_yield_cpu(~owner); - if (_raw_compare_and_swap(&lp->owner_cpu, 0, cpu) == 0) - return; +struct spin_wait { + struct spin_wait *next, *prev; + int node_id; +} __aligned(32); + +static DEFINE_PER_CPU_ALIGNED(struct spin_wait, spin_wait[4]); + +#define _Q_LOCK_CPU_OFFSET 0 +#define _Q_LOCK_STEAL_OFFSET 16 +#define _Q_TAIL_IDX_OFFSET 18 +#define _Q_TAIL_CPU_OFFSET 20 + +#define _Q_LOCK_CPU_MASK 0x0000ffff +#define _Q_LOCK_STEAL_ADD 0x00010000 +#define _Q_LOCK_STEAL_MASK 0x00030000 +#define _Q_TAIL_IDX_MASK 0x000c0000 +#define _Q_TAIL_CPU_MASK 0xfff00000 + +#define _Q_LOCK_MASK (_Q_LOCK_CPU_MASK | _Q_LOCK_STEAL_MASK) +#define _Q_TAIL_MASK (_Q_TAIL_IDX_MASK | _Q_TAIL_CPU_MASK) + +void arch_spin_lock_setup(int cpu) +{ + struct spin_wait *node; + int ix; + + node = per_cpu_ptr(&spin_wait[0], cpu); + for (ix = 0; ix < 4; ix++, node++) { + memset(node, 0, sizeof(*node)); + node->node_id = ((cpu + 1) << _Q_TAIL_CPU_OFFSET) + + (ix << _Q_TAIL_IDX_OFFSET); } } -EXPORT_SYMBOL(arch_spin_lock_wait); -void arch_spin_lock_wait_flags(arch_spinlock_t *lp, unsigned long flags) +static inline int arch_load_niai4(int *lock) { - int count = spin_retry; - unsigned int cpu = ~smp_processor_id(); - unsigned int owner; + int owner; - local_irq_restore(flags); - while (1) { - owner = lp->owner_cpu; - if (!owner || smp_vcpu_scheduled(~owner)) { - for (count = spin_retry; count > 0; count--) { - if (arch_spin_is_locked(lp)) - continue; - local_irq_disable(); - if (_raw_compare_and_swap(&lp->owner_cpu, 0, - cpu) == 0) - return; - local_irq_restore(flags); - } - if (MACHINE_IS_LPAR) - continue; - } - owner = lp->owner_cpu; - if (owner) - smp_yield_cpu(~owner); - local_irq_disable(); - if (_raw_compare_and_swap(&lp->owner_cpu, 0, cpu) == 0) - return; - local_irq_restore(flags); - } + asm_inline volatile( + ALTERNATIVE("nop", ".insn rre,0xb2fa0000,4,0", ALT_FACILITY(49)) /* NIAI 4 */ + " l %[owner],%[lock]" + : [owner] "=d" (owner) : [lock] "R" (*lock) : "memory"); + return owner; } -EXPORT_SYMBOL(arch_spin_lock_wait_flags); -int arch_spin_trylock_retry(arch_spinlock_t *lp) +#ifdef __HAVE_ASM_FLAG_OUTPUTS__ + +static inline int arch_try_cmpxchg_niai8(int *lock, int old, int new) { - unsigned int cpu = ~smp_processor_id(); - int count; + int cc; - for (count = spin_retry; count > 0; count--) { - if (arch_spin_is_locked(lp)) - continue; - if (_raw_compare_and_swap(&lp->owner_cpu, 0, cpu) == 0) - return 1; - } - return 0; + asm_inline volatile( + ALTERNATIVE("nop", ".insn rre,0xb2fa0000,8,0", ALT_FACILITY(49)) /* NIAI 8 */ + " cs %[old],%[new],%[lock]" + : [old] "+d" (old), [lock] "+Q" (*lock), "=@cc" (cc) + : [new] "d" (new) + : "memory"); + return cc == 0; } -EXPORT_SYMBOL(arch_spin_trylock_retry); -void arch_spin_relax(arch_spinlock_t *lock) +#else /* __HAVE_ASM_FLAG_OUTPUTS__ */ + +static inline int arch_try_cmpxchg_niai8(int *lock, int old, int new) { - unsigned int cpu = lock->owner_cpu; - if (cpu != 0) { - if (MACHINE_IS_VM || MACHINE_IS_KVM || - !smp_vcpu_scheduled(~cpu)) - smp_yield_cpu(~cpu); - } + int expected = old; + + asm_inline volatile( + ALTERNATIVE("nop", ".insn rre,0xb2fa0000,8,0", ALT_FACILITY(49)) /* NIAI 8 */ + " cs %[old],%[new],%[lock]" + : [old] "+d" (old), [lock] "+Q" (*lock) + : [new] "d" (new) + : "cc", "memory"); + return expected == old; } -EXPORT_SYMBOL(arch_spin_relax); -void _raw_read_lock_wait(arch_rwlock_t *rw) +#endif /* __HAVE_ASM_FLAG_OUTPUTS__ */ + +static inline struct spin_wait *arch_spin_decode_tail(int lock) { - unsigned int old; - int count = spin_retry; + int ix, cpu; + + ix = (lock & _Q_TAIL_IDX_MASK) >> _Q_TAIL_IDX_OFFSET; + cpu = (lock & _Q_TAIL_CPU_MASK) >> _Q_TAIL_CPU_OFFSET; + return per_cpu_ptr(&spin_wait[ix], cpu - 1); +} + +static inline int arch_spin_yield_target(int lock, struct spin_wait *node) +{ + if (lock & _Q_LOCK_CPU_MASK) + return lock & _Q_LOCK_CPU_MASK; + if (node == NULL || node->prev == NULL) + return 0; /* 0 -> no target cpu */ + while (node->prev) + node = node->prev; + return node->node_id >> _Q_TAIL_CPU_OFFSET; +} + +static inline void arch_spin_lock_queued(arch_spinlock_t *lp) +{ + struct spin_wait *node, *next; + int lockval, ix, node_id, tail_id, old, new, owner, count; + + ix = get_lowcore()->spinlock_index++; + barrier(); + lockval = spinlock_lockval(); /* cpu + 1 */ + node = this_cpu_ptr(&spin_wait[ix]); + node->prev = node->next = NULL; + node_id = node->node_id; + /* Enqueue the node for this CPU in the spinlock wait queue */ + old = READ_ONCE(lp->lock); while (1) { - if (count-- <= 0) { - smp_yield(); + if ((old & _Q_LOCK_CPU_MASK) == 0 && + (old & _Q_LOCK_STEAL_MASK) != _Q_LOCK_STEAL_MASK) { + /* + * The lock is free but there may be waiters. + * With no waiters simply take the lock, if there + * are waiters try to steal the lock. The lock may + * be stolen three times before the next queued + * waiter will get the lock. + */ + new = (old ? (old + _Q_LOCK_STEAL_ADD) : 0) | lockval; + if (arch_try_cmpxchg(&lp->lock, &old, new)) + /* Got the lock */ + goto out; + /* lock passing in progress */ + continue; + } + /* Make the node of this CPU the new tail. */ + new = node_id | (old & _Q_LOCK_MASK); + if (arch_try_cmpxchg(&lp->lock, &old, new)) + break; + } + /* Set the 'next' pointer of the tail node in the queue */ + tail_id = old & _Q_TAIL_MASK; + if (tail_id != 0) { + node->prev = arch_spin_decode_tail(tail_id); + WRITE_ONCE(node->prev->next, node); + } + + /* Pass the virtual CPU to the lock holder if it is not running */ + owner = arch_spin_yield_target(old, node); + if (owner && arch_vcpu_is_preempted(owner - 1)) + smp_yield_cpu(owner - 1); + + /* Spin on the CPU local node->prev pointer */ + if (tail_id != 0) { + count = spin_retry; + while (READ_ONCE(node->prev) != NULL) { + if (count-- >= 0) + continue; count = spin_retry; + /* Query running state of lock holder again. */ + owner = arch_spin_yield_target(old, node); + if (owner && arch_vcpu_is_preempted(owner - 1)) + smp_yield_cpu(owner - 1); + } + } + + /* Spin on the lock value in the spinlock_t */ + count = spin_retry; + while (1) { + old = READ_ONCE(lp->lock); + owner = old & _Q_LOCK_CPU_MASK; + if (!owner) { + tail_id = old & _Q_TAIL_MASK; + new = ((tail_id != node_id) ? tail_id : 0) | lockval; + if (arch_try_cmpxchg(&lp->lock, &old, new)) + /* Got the lock */ + break; + continue; } - if (!arch_read_can_lock(rw)) + if (count-- >= 0) continue; - old = rw->lock & 0x7fffffffU; - if (_raw_compare_and_swap(&rw->lock, old, old + 1) == old) - return; + count = spin_retry; + if (!machine_is_lpar() || arch_vcpu_is_preempted(owner - 1)) + smp_yield_cpu(owner - 1); } + + /* Pass lock_spin job to next CPU in the queue */ + if (node_id && tail_id != node_id) { + /* Wait until the next CPU has set up the 'next' pointer */ + while ((next = READ_ONCE(node->next)) == NULL) + ; + next->prev = NULL; + } + + out: + get_lowcore()->spinlock_index--; } -EXPORT_SYMBOL(_raw_read_lock_wait); -void _raw_read_lock_wait_flags(arch_rwlock_t *rw, unsigned long flags) +static inline void arch_spin_lock_classic(arch_spinlock_t *lp) { - unsigned int old; - int count = spin_retry; + int lockval, old, new, owner, count; - local_irq_restore(flags); + lockval = spinlock_lockval(); /* cpu + 1 */ + + /* Pass the virtual CPU to the lock holder if it is not running */ + owner = arch_spin_yield_target(READ_ONCE(lp->lock), NULL); + if (owner && arch_vcpu_is_preempted(owner - 1)) + smp_yield_cpu(owner - 1); + + count = spin_retry; while (1) { - if (count-- <= 0) { - smp_yield(); - count = spin_retry; + old = arch_load_niai4(&lp->lock); + owner = old & _Q_LOCK_CPU_MASK; + /* Try to get the lock if it is free. */ + if (!owner) { + new = (old & _Q_TAIL_MASK) | lockval; + if (arch_try_cmpxchg_niai8(&lp->lock, old, new)) { + /* Got the lock */ + return; + } + continue; } - if (!arch_read_can_lock(rw)) + if (count-- >= 0) continue; - old = rw->lock & 0x7fffffffU; - local_irq_disable(); - if (_raw_compare_and_swap(&rw->lock, old, old + 1) == old) - return; + count = spin_retry; + if (!machine_is_lpar() || arch_vcpu_is_preempted(owner - 1)) + smp_yield_cpu(owner - 1); } } -EXPORT_SYMBOL(_raw_read_lock_wait_flags); -int _raw_read_trylock_retry(arch_rwlock_t *rw) +void arch_spin_lock_wait(arch_spinlock_t *lp) +{ + if (test_cpu_flag(CIF_DEDICATED_CPU)) + arch_spin_lock_queued(lp); + else + arch_spin_lock_classic(lp); +} +EXPORT_SYMBOL(arch_spin_lock_wait); + +int arch_spin_trylock_retry(arch_spinlock_t *lp) { - unsigned int old; - int count = spin_retry; + int cpu = spinlock_lockval(); + int owner, count; - while (count-- > 0) { - if (!arch_read_can_lock(rw)) - continue; - old = rw->lock & 0x7fffffffU; - if (_raw_compare_and_swap(&rw->lock, old, old + 1) == old) - return 1; + for (count = spin_retry; count > 0; count--) { + owner = READ_ONCE(lp->lock); + /* Try to get the lock if it is free. */ + if (!owner) { + if (arch_try_cmpxchg(&lp->lock, &owner, cpu)) + return 1; + } } return 0; } -EXPORT_SYMBOL(_raw_read_trylock_retry); +EXPORT_SYMBOL(arch_spin_trylock_retry); -void _raw_write_lock_wait(arch_rwlock_t *rw) +void arch_read_lock_wait(arch_rwlock_t *rw) { - int count = spin_retry; - - while (1) { - if (count-- <= 0) { - smp_yield(); - count = spin_retry; - } - if (!arch_write_can_lock(rw)) - continue; - if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000) == 0) - return; + if (unlikely(in_interrupt())) { + while (READ_ONCE(rw->cnts) & 0x10000) + barrier(); + return; } + + /* Remove this reader again to allow recursive read locking */ + __atomic_add_const(-1, &rw->cnts); + /* Put the reader into the wait queue */ + arch_spin_lock(&rw->wait); + /* Now add this reader to the count value again */ + __atomic_add_const(1, &rw->cnts); + /* Loop until the writer is done */ + while (READ_ONCE(rw->cnts) & 0x10000) + barrier(); + arch_spin_unlock(&rw->wait); } -EXPORT_SYMBOL(_raw_write_lock_wait); +EXPORT_SYMBOL(arch_read_lock_wait); -void _raw_write_lock_wait_flags(arch_rwlock_t *rw, unsigned long flags) +void arch_write_lock_wait(arch_rwlock_t *rw) { - int count = spin_retry; + int old; + + /* Add this CPU to the write waiters */ + __atomic_add(0x20000, &rw->cnts); + + /* Put the writer into the wait queue */ + arch_spin_lock(&rw->wait); - local_irq_restore(flags); while (1) { - if (count-- <= 0) { - smp_yield(); - count = spin_retry; - } - if (!arch_write_can_lock(rw)) - continue; - local_irq_disable(); - if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000) == 0) - return; + old = READ_ONCE(rw->cnts); + if ((old & 0x1ffff) == 0 && + arch_try_cmpxchg(&rw->cnts, &old, old | 0x10000)) + /* Got the lock */ + break; + barrier(); } + + arch_spin_unlock(&rw->wait); } -EXPORT_SYMBOL(_raw_write_lock_wait_flags); +EXPORT_SYMBOL(arch_write_lock_wait); -int _raw_write_trylock_retry(arch_rwlock_t *rw) +void arch_spin_relax(arch_spinlock_t *lp) { - int count = spin_retry; + int cpu; - while (count-- > 0) { - if (!arch_write_can_lock(rw)) - continue; - if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000) == 0) - return 1; - } - return 0; + cpu = READ_ONCE(lp->lock) & _Q_LOCK_CPU_MASK; + if (!cpu) + return; + if (machine_is_lpar() && !arch_vcpu_is_preempted(cpu - 1)) + return; + smp_yield_cpu(cpu - 1); } -EXPORT_SYMBOL(_raw_write_trylock_retry); +EXPORT_SYMBOL(arch_spin_relax); diff --git a/arch/s390/lib/string.c b/arch/s390/lib/string.c index b647d5ff0ad9..757f58960198 100644 --- a/arch/s390/lib/string.c +++ b/arch/s390/lib/string.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: GPL-2.0 /* * Optimized string functions * @@ -7,32 +8,44 @@ */ #define IN_ARCH_STRING_C 1 +#ifndef __NO_FORTIFY +# define __NO_FORTIFY +#endif #include <linux/types.h> -#include <linux/module.h> +#include <linux/string.h> +#include <linux/export.h> +#include <asm/asm.h> /* * Helper functions to find the end of a string */ static inline char *__strend(const char *s) { - register unsigned long r0 asm("0") = 0; - - asm volatile ("0: srst %0,%1\n" - " jo 0b" - : "+d" (r0), "+a" (s) : : "cc" ); - return (char *) r0; + unsigned long e = 0; + + asm volatile( + " lghi 0,0\n" + "0: srst %[e],%[s]\n" + " jo 0b" + : [e] "+&a" (e), [s] "+&a" (s) + : + : "cc", "memory", "0"); + return (char *)e; } static inline char *__strnend(const char *s, size_t n) { - register unsigned long r0 asm("0") = 0; const char *p = s + n; - asm volatile ("0: srst %0,%1\n" - " jo 0b" - : "+d" (p), "+a" (s) : "d" (r0) : "cc" ); - return (char *) p; + asm volatile( + " lghi 0,0\n" + "0: srst %[p],%[s]\n" + " jo 0b" + : [p] "+&d" (p), [s] "+&a" (s) + : + : "cc", "memory", "0"); + return (char *)p; } /** @@ -41,11 +54,13 @@ static inline char *__strnend(const char *s, size_t n) * * returns the length of @s */ +#ifdef __HAVE_ARCH_STRLEN size_t strlen(const char *s) { return __strend(s) - s; } EXPORT_SYMBOL(strlen); +#endif /** * strnlen - Find the length of a length-limited string @@ -54,73 +69,13 @@ EXPORT_SYMBOL(strlen); * * returns the minimum of the length of @s and @n */ -size_t strnlen(const char * s, size_t n) +#ifdef __HAVE_ARCH_STRNLEN +size_t strnlen(const char *s, size_t n) { return __strnend(s, n) - s; } EXPORT_SYMBOL(strnlen); - -/** - * strcpy - Copy a %NUL terminated string - * @dest: Where to copy the string to - * @src: Where to copy the string from - * - * returns a pointer to @dest - */ -char *strcpy(char *dest, const char *src) -{ - register int r0 asm("0") = 0; - char *ret = dest; - - asm volatile ("0: mvst %0,%1\n" - " jo 0b" - : "+&a" (dest), "+&a" (src) : "d" (r0) - : "cc", "memory" ); - return ret; -} -EXPORT_SYMBOL(strcpy); - -/** - * strlcpy - Copy a %NUL terminated string into a sized buffer - * @dest: Where to copy the string to - * @src: Where to copy the string from - * @size: size of destination buffer - * - * Compatible with *BSD: the result is always a valid - * NUL-terminated string that fits in the buffer (unless, - * of course, the buffer size is zero). It does not pad - * out the result like strncpy() does. - */ -size_t strlcpy(char *dest, const char *src, size_t size) -{ - size_t ret = __strend(src) - src; - - if (size) { - size_t len = (ret >= size) ? size-1 : ret; - dest[len] = '\0'; - memcpy(dest, src, len); - } - return ret; -} -EXPORT_SYMBOL(strlcpy); - -/** - * strncpy - Copy a length-limited, %NUL-terminated string - * @dest: Where to copy the string to - * @src: Where to copy the string from - * @n: The maximum number of bytes to copy - * - * The result is not %NUL-terminated if the source exceeds - * @n bytes. - */ -char *strncpy(char *dest, const char *src, size_t n) -{ - size_t len = __strnend(src, n) - src; - memset(dest + len, 0, n - len); - memcpy(dest, src, len); - return dest; -} -EXPORT_SYMBOL(strncpy); +#endif /** * strcat - Append one %NUL-terminated string to another @@ -129,21 +84,25 @@ EXPORT_SYMBOL(strncpy); * * returns a pointer to @dest */ +#ifdef __HAVE_ARCH_STRCAT char *strcat(char *dest, const char *src) { - register int r0 asm("0") = 0; - unsigned long dummy; + unsigned long dummy = 0; char *ret = dest; - asm volatile ("0: srst %0,%1\n" - " jo 0b\n" - "1: mvst %0,%2\n" - " jo 1b" - : "=&a" (dummy), "+a" (dest), "+a" (src) - : "d" (r0), "0" (0UL) : "cc", "memory" ); + asm volatile( + " lghi 0,0\n" + "0: srst %[dummy],%[dest]\n" + " jo 0b\n" + "1: mvst %[dummy],%[src]\n" + " jo 1b" + : [dummy] "+&a" (dummy), [dest] "+&a" (dest), [src] "+&a" (src) + : + : "cc", "memory", "0"); return ret; } EXPORT_SYMBOL(strcat); +#endif /** * strlcat - Append a length-limited, %NUL-terminated string to another @@ -151,6 +110,7 @@ EXPORT_SYMBOL(strcat); * @src: The string to append to it * @n: The size of the destination buffer. */ +#ifdef __HAVE_ARCH_STRLCAT size_t strlcat(char *dest, const char *src, size_t n) { size_t dsize = __strend(dest) - dest; @@ -168,6 +128,7 @@ size_t strlcat(char *dest, const char *src, size_t n) return res; } EXPORT_SYMBOL(strlcat); +#endif /** * strncat - Append a length-limited, %NUL-terminated string to another @@ -176,10 +137,8 @@ EXPORT_SYMBOL(strlcat); * @n: The maximum numbers of bytes to copy * * returns a pointer to @dest - * - * Note that in contrast to strncpy, strncat ensures the result is - * terminated. */ +#ifdef __HAVE_ARCH_STRNCAT char *strncat(char *dest, const char *src, size_t n) { size_t len = __strnend(src, n) - src; @@ -190,58 +149,63 @@ char *strncat(char *dest, const char *src, size_t n) return dest; } EXPORT_SYMBOL(strncat); +#endif /** * strcmp - Compare two strings - * @cs: One string - * @ct: Another string + * @s1: One string + * @s2: Another string * - * returns 0 if @cs and @ct are equal, - * < 0 if @cs is less than @ct - * > 0 if @cs is greater than @ct + * returns 0 if @s1 and @s2 are equal, + * < 0 if @s1 is less than @s2 + * > 0 if @s1 is greater than @s2 */ -int strcmp(const char *cs, const char *ct) +#ifdef __HAVE_ARCH_STRCMP +int strcmp(const char *s1, const char *s2) { - register int r0 asm("0") = 0; int ret = 0; - asm volatile ("0: clst %2,%3\n" - " jo 0b\n" - " je 1f\n" - " ic %0,0(%2)\n" - " ic %1,0(%3)\n" - " sr %0,%1\n" - "1:" - : "+d" (ret), "+d" (r0), "+a" (cs), "+a" (ct) - : : "cc" ); + asm volatile( + " lghi 0,0\n" + "0: clst %[s1],%[s2]\n" + " jo 0b\n" + " je 1f\n" + " ic %[ret],0(%[s1])\n" + " ic 0,0(%[s2])\n" + " sr %[ret],0\n" + "1:" + : [ret] "+&d" (ret), [s1] "+&a" (s1), [s2] "+&a" (s2) + : + : "cc", "memory", "0"); return ret; } EXPORT_SYMBOL(strcmp); +#endif -/** - * strrchr - Find the last occurrence of a character in a string - * @s: The string to be searched - * @c: The character to search for - */ -char * strrchr(const char * s, int c) +static inline int clcle(const char *s1, unsigned long l1, + const char *s2, unsigned long l2) { - size_t len = __strend(s) - s; - - if (len) - do { - if (s[len] == (char) c) - return (char *) s + len; - } while (--len > 0); - return NULL; + union register_pair r1 = { .even = (unsigned long)s1, .odd = l1, }; + union register_pair r3 = { .even = (unsigned long)s2, .odd = l2, }; + int cc; + + asm volatile( + "0: clcle %[r1],%[r3],0\n" + " jo 0b\n" + CC_IPM(cc) + : CC_OUT(cc, cc), [r1] "+d" (r1.pair), [r3] "+d" (r3.pair) + : + : CC_CLOBBER_LIST("memory")); + return CC_TRANSFORM(cc); } -EXPORT_SYMBOL(strrchr); /** * strstr - Find the first substring in a %NUL terminated string * @s1: The string to be searched * @s2: The string to search for */ -char * strstr(const char * s1,const char * s2) +#ifdef __HAVE_ARCH_STRSTR +char *strstr(const char *s1, const char *s2) { int l1, l2; @@ -250,18 +214,9 @@ char * strstr(const char * s1,const char * s2) return (char *) s1; l1 = __strend(s1) - s1; while (l1-- >= l2) { - register unsigned long r2 asm("2") = (unsigned long) s1; - register unsigned long r3 asm("3") = (unsigned long) l2; - register unsigned long r4 asm("4") = (unsigned long) s2; - register unsigned long r5 asm("5") = (unsigned long) l2; int cc; - asm volatile ("0: clcle %1,%3,0\n" - " jo 0b\n" - " ipm %0\n" - " srl %0,28" - : "=&d" (cc), "+a" (r2), "+a" (r3), - "+a" (r4), "+a" (r5) : : "cc" ); + cc = clcle(s1, l2, s2, l2); if (!cc) return (char *) s1; s1++; @@ -269,6 +224,7 @@ char * strstr(const char * s1,const char * s2) return NULL; } EXPORT_SYMBOL(strstr); +#endif /** * memchr - Find a character in an area of memory. @@ -279,46 +235,44 @@ EXPORT_SYMBOL(strstr); * returns the address of the first occurrence of @c, or %NULL * if @c is not found */ +#ifdef __HAVE_ARCH_MEMCHR void *memchr(const void *s, int c, size_t n) { - register int r0 asm("0") = (char) c; const void *ret = s + n; - asm volatile ("0: srst %0,%1\n" - " jo 0b\n" - " jl 1f\n" - " la %0,0\n" - "1:" - : "+a" (ret), "+&a" (s) : "d" (r0) : "cc" ); + asm volatile( + " lgr 0,%[c]\n" + "0: srst %[ret],%[s]\n" + " jo 0b\n" + " jl 1f\n" + " la %[ret],0\n" + "1:" + : [ret] "+&a" (ret), [s] "+&a" (s) + : [c] "d" (c) + : "cc", "memory", "0"); return (void *) ret; } EXPORT_SYMBOL(memchr); +#endif /** * memcmp - Compare two areas of memory - * @cs: One area of memory - * @ct: Another area of memory - * @count: The size of the area. + * @s1: One area of memory + * @s2: Another area of memory + * @n: The size of the area. */ -int memcmp(const void *cs, const void *ct, size_t n) +#ifdef __HAVE_ARCH_MEMCMP +int memcmp(const void *s1, const void *s2, size_t n) { - register unsigned long r2 asm("2") = (unsigned long) cs; - register unsigned long r3 asm("3") = (unsigned long) n; - register unsigned long r4 asm("4") = (unsigned long) ct; - register unsigned long r5 asm("5") = (unsigned long) n; int ret; - asm volatile ("0: clcle %1,%3,0\n" - " jo 0b\n" - " ipm %0\n" - " srl %0,28" - : "=&d" (ret), "+a" (r2), "+a" (r3), "+a" (r4), "+a" (r5) - : : "cc" ); + ret = clcle(s1, n, s2, n); if (ret) - ret = *(char *) r2 - *(char *) r4; + ret = ret == 1 ? -1 : 1; return ret; } EXPORT_SYMBOL(memcmp); +#endif /** * memscan - Find a character in an area of memory. @@ -329,14 +283,19 @@ EXPORT_SYMBOL(memcmp); * returns the address of the first occurrence of @c, or 1 byte past * the area if @c is not found */ +#ifdef __HAVE_ARCH_MEMSCAN void *memscan(void *s, int c, size_t n) { - register int r0 asm("0") = (char) c; const void *ret = s + n; - asm volatile ("0: srst %0,%1\n" - " jo 0b\n" - : "+a" (ret), "+&a" (s) : "d" (r0) : "cc" ); - return (void *) ret; + asm volatile( + " lgr 0,%[c]\n" + "0: srst %[ret],%[s]\n" + " jo 0b" + : [ret] "+&a" (ret), [s] "+&a" (s) + : [c] "d" (c) + : "cc", "memory", "0"); + return (void *)ret; } EXPORT_SYMBOL(memscan); +#endif diff --git a/arch/s390/lib/test_kprobes.c b/arch/s390/lib/test_kprobes.c new file mode 100644 index 000000000000..9021298c3e8a --- /dev/null +++ b/arch/s390/lib/test_kprobes.c @@ -0,0 +1,76 @@ +// SPDX-License-Identifier: GPL-2.0+ + +#include <linux/kernel.h> +#include <linux/kprobes.h> +#include <linux/random.h> +#include <kunit/test.h> +#include "test_kprobes.h" + +static struct kprobe kp; + +static void setup_kprobe(struct kunit *test, struct kprobe *kp, + const char *symbol, int offset) +{ + kp->offset = offset; + kp->addr = NULL; + kp->symbol_name = symbol; +} + +static void test_kprobe_offset(struct kunit *test, struct kprobe *kp, + const char *target, int offset) +{ + int ret; + + setup_kprobe(test, kp, target, 0); + ret = register_kprobe(kp); + if (!ret) + unregister_kprobe(kp); + KUNIT_EXPECT_EQ(test, 0, ret); + setup_kprobe(test, kp, target, offset); + ret = register_kprobe(kp); + KUNIT_EXPECT_EQ(test, -EINVAL, ret); + if (!ret) + unregister_kprobe(kp); +} + +static void test_kprobe_odd(struct kunit *test) +{ + test_kprobe_offset(test, &kp, "kprobes_target_odd", + kprobes_target_odd_offs); +} + +static void test_kprobe_in_insn4(struct kunit *test) +{ + test_kprobe_offset(test, &kp, "kprobes_target_in_insn4", + kprobes_target_in_insn4_offs); +} + +static void test_kprobe_in_insn6_lo(struct kunit *test) +{ + test_kprobe_offset(test, &kp, "kprobes_target_in_insn6_lo", + kprobes_target_in_insn6_lo_offs); +} + +static void test_kprobe_in_insn6_hi(struct kunit *test) +{ + test_kprobe_offset(test, &kp, "kprobes_target_in_insn6_hi", + kprobes_target_in_insn6_hi_offs); +} + +static struct kunit_case kprobes_testcases[] = { + KUNIT_CASE(test_kprobe_odd), + KUNIT_CASE(test_kprobe_in_insn4), + KUNIT_CASE(test_kprobe_in_insn6_lo), + KUNIT_CASE(test_kprobe_in_insn6_hi), + {} +}; + +static struct kunit_suite kprobes_test_suite = { + .name = "kprobes_test_s390", + .test_cases = kprobes_testcases, +}; + +kunit_test_suites(&kprobes_test_suite); + +MODULE_DESCRIPTION("KUnit tests for kprobes"); +MODULE_LICENSE("GPL"); diff --git a/arch/s390/lib/test_kprobes.h b/arch/s390/lib/test_kprobes.h new file mode 100644 index 000000000000..2b4c9bc337f1 --- /dev/null +++ b/arch/s390/lib/test_kprobes.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +#ifndef TEST_KPROBES_H +#define TEST_KPROBES_H + +extern unsigned long kprobes_target_odd_offs; +extern unsigned long kprobes_target_in_insn4_offs; +extern unsigned long kprobes_target_in_insn6_lo_offs; +extern unsigned long kprobes_target_in_insn6_hi_offs; + +#endif diff --git a/arch/s390/lib/test_kprobes_asm.S b/arch/s390/lib/test_kprobes_asm.S new file mode 100644 index 000000000000..ade7a3042334 --- /dev/null +++ b/arch/s390/lib/test_kprobes_asm.S @@ -0,0 +1,45 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ + +#include <linux/linkage.h> +#include <asm/ftrace.h> + +#define KPROBES_TARGET_START(name) \ + SYM_FUNC_START(name); \ + FTRACE_GEN_NOP_ASM(name) + +#define KPROBES_TARGET_END(name) \ + SYM_FUNC_END(name); \ + SYM_DATA(name##_offs, .quad 1b - name) + +KPROBES_TARGET_START(kprobes_target_in_insn4) + .word 0x4700 // bc 0,0 +1: .word 0x0000 + br %r14 +KPROBES_TARGET_END(kprobes_target_in_insn4) + +KPROBES_TARGET_START(kprobes_target_in_insn6_lo) + .word 0xe310 // ly 1,0 +1: .word 0x0000 + .word 0x0058 + br %r14 +KPROBES_TARGET_END(kprobes_target_in_insn6_lo) + +KPROBES_TARGET_START(kprobes_target_in_insn6_hi) + .word 0xe310 // ly 1,0 + .word 0x0000 +1: .word 0x0058 + br %r14 +KPROBES_TARGET_END(kprobes_target_in_insn6_hi) + +KPROBES_TARGET_START(kprobes_target_bp) + nop + .word 0x0000 + nop +1: br %r14 +KPROBES_TARGET_END(kprobes_target_bp) + +KPROBES_TARGET_START(kprobes_target_odd) + .byte 0x07 +1: .byte 0x07 + br %r14 +KPROBES_TARGET_END(kprobes_target_odd) diff --git a/arch/s390/lib/test_modules.c b/arch/s390/lib/test_modules.c new file mode 100644 index 000000000000..f96b6a3737e7 --- /dev/null +++ b/arch/s390/lib/test_modules.c @@ -0,0 +1,33 @@ +// SPDX-License-Identifier: GPL-2.0+ + +#include <kunit/test.h> +#include <linux/module.h> + +#include "test_modules.h" + +/* + * Test that modules with many relocations are loaded properly. + */ +static void test_modules_many_vmlinux_relocs(struct kunit *test) +{ + int result = 0; + +#define CALL_RETURN(i) result += test_modules_return_ ## i() + REPEAT_10000(CALL_RETURN); + KUNIT_ASSERT_EQ(test, result, 49995000); +} + +static struct kunit_case modules_testcases[] = { + KUNIT_CASE(test_modules_many_vmlinux_relocs), + {} +}; + +static struct kunit_suite modules_test_suite = { + .name = "modules_test_s390", + .test_cases = modules_testcases, +}; + +kunit_test_suites(&modules_test_suite); + +MODULE_DESCRIPTION("KUnit test that modules with many relocations are loaded properly"); +MODULE_LICENSE("GPL"); diff --git a/arch/s390/lib/test_modules.h b/arch/s390/lib/test_modules.h new file mode 100644 index 000000000000..6371fcf17684 --- /dev/null +++ b/arch/s390/lib/test_modules.h @@ -0,0 +1,53 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +#ifndef TEST_MODULES_H +#define TEST_MODULES_H + +#define __REPEAT_10000_3(f, x) \ + f(x ## 0); \ + f(x ## 1); \ + f(x ## 2); \ + f(x ## 3); \ + f(x ## 4); \ + f(x ## 5); \ + f(x ## 6); \ + f(x ## 7); \ + f(x ## 8); \ + f(x ## 9) +#define __REPEAT_10000_2(f, x) \ + __REPEAT_10000_3(f, x ## 0); \ + __REPEAT_10000_3(f, x ## 1); \ + __REPEAT_10000_3(f, x ## 2); \ + __REPEAT_10000_3(f, x ## 3); \ + __REPEAT_10000_3(f, x ## 4); \ + __REPEAT_10000_3(f, x ## 5); \ + __REPEAT_10000_3(f, x ## 6); \ + __REPEAT_10000_3(f, x ## 7); \ + __REPEAT_10000_3(f, x ## 8); \ + __REPEAT_10000_3(f, x ## 9) +#define __REPEAT_10000_1(f, x) \ + __REPEAT_10000_2(f, x ## 0); \ + __REPEAT_10000_2(f, x ## 1); \ + __REPEAT_10000_2(f, x ## 2); \ + __REPEAT_10000_2(f, x ## 3); \ + __REPEAT_10000_2(f, x ## 4); \ + __REPEAT_10000_2(f, x ## 5); \ + __REPEAT_10000_2(f, x ## 6); \ + __REPEAT_10000_2(f, x ## 7); \ + __REPEAT_10000_2(f, x ## 8); \ + __REPEAT_10000_2(f, x ## 9) +#define REPEAT_10000(f) \ + __REPEAT_10000_1(f, 0); \ + __REPEAT_10000_1(f, 1); \ + __REPEAT_10000_1(f, 2); \ + __REPEAT_10000_1(f, 3); \ + __REPEAT_10000_1(f, 4); \ + __REPEAT_10000_1(f, 5); \ + __REPEAT_10000_1(f, 6); \ + __REPEAT_10000_1(f, 7); \ + __REPEAT_10000_1(f, 8); \ + __REPEAT_10000_1(f, 9) + +#define DECLARE_RETURN(i) int test_modules_return_ ## i(void) +REPEAT_10000(DECLARE_RETURN); + +#endif diff --git a/arch/s390/lib/test_modules_helpers.c b/arch/s390/lib/test_modules_helpers.c new file mode 100644 index 000000000000..1670349a03eb --- /dev/null +++ b/arch/s390/lib/test_modules_helpers.c @@ -0,0 +1,13 @@ +// SPDX-License-Identifier: GPL-2.0+ + +#include <linux/export.h> + +#include "test_modules.h" + +#define DEFINE_RETURN(i) \ + int test_modules_return_ ## i(void) \ + { \ + return 1 ## i - 10000; \ + } \ + EXPORT_SYMBOL_GPL(test_modules_return_ ## i) +REPEAT_10000(DEFINE_RETURN); diff --git a/arch/s390/lib/test_unwind.c b/arch/s390/lib/test_unwind.c new file mode 100644 index 000000000000..6bb3fa5bf925 --- /dev/null +++ b/arch/s390/lib/test_unwind.c @@ -0,0 +1,523 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Test module for unwind_for_each_frame + */ + +#include <kunit/test.h> +#include <asm/unwind.h> +#include <linux/completion.h> +#include <linux/kallsyms.h> +#include <linux/kthread.h> +#include <linux/ftrace.h> +#include <linux/module.h> +#include <linux/timer.h> +#include <linux/slab.h> +#include <linux/string.h> +#include <linux/kprobes.h> +#include <linux/wait.h> +#include <asm/irq.h> + +static struct kunit *current_test; + +#define BT_BUF_SIZE (PAGE_SIZE * 4) + +static bool force_bt; +module_param_named(backtrace, force_bt, bool, 0444); +MODULE_PARM_DESC(backtrace, "print backtraces for all tests"); + +/* + * To avoid printk line limit split backtrace by lines + */ +static void print_backtrace(char *bt) +{ + char *p; + + while (true) { + p = strsep(&bt, "\n"); + if (!p) + break; + kunit_err(current_test, "%s\n", p); + } +} + +/* + * Calls unwind_for_each_frame(task, regs, sp) and verifies that the result + * contains unwindme_func2 followed by unwindme_func1. + */ +static noinline int test_unwind(struct task_struct *task, struct pt_regs *regs, + unsigned long sp) +{ + int frame_count, prev_is_func2, seen_func2_func1, seen_arch_rethook_trampoline; + const int max_frames = 128; + struct unwind_state state; + size_t bt_pos = 0; + int ret = 0; + char *bt; + + bt = kmalloc(BT_BUF_SIZE, GFP_ATOMIC); + if (!bt) { + kunit_err(current_test, "failed to allocate backtrace buffer\n"); + return -ENOMEM; + } + /* Unwind. */ + frame_count = 0; + prev_is_func2 = 0; + seen_func2_func1 = 0; + seen_arch_rethook_trampoline = 0; + unwind_for_each_frame(&state, task, regs, sp) { + unsigned long addr = unwind_get_return_address(&state); + char sym[KSYM_SYMBOL_LEN]; + + if (frame_count++ == max_frames) + break; + if (state.reliable && !addr) { + kunit_err(current_test, "unwind state reliable but addr is 0\n"); + ret = -EINVAL; + break; + } + sprint_symbol(sym, addr); + if (bt_pos < BT_BUF_SIZE) { + bt_pos += snprintf(bt + bt_pos, BT_BUF_SIZE - bt_pos, + state.reliable ? " [%-7s%px] %pSR\n" : + "([%-7s%px] %pSR)\n", + stack_type_name(state.stack_info.type), + (void *)state.sp, (void *)state.ip); + if (bt_pos >= BT_BUF_SIZE) + kunit_err(current_test, "backtrace buffer is too small\n"); + } + frame_count += 1; + if (prev_is_func2 && str_has_prefix(sym, "unwindme_func1")) + seen_func2_func1 = 1; + prev_is_func2 = str_has_prefix(sym, "unwindme_func2"); + if (str_has_prefix(sym, "arch_rethook_trampoline+0x0/")) + seen_arch_rethook_trampoline = 1; + } + + /* Check the results. */ + if (unwind_error(&state)) { + kunit_err(current_test, "unwind error\n"); + ret = -EINVAL; + } + if (!seen_func2_func1) { + kunit_err(current_test, "unwindme_func2 and unwindme_func1 not found\n"); + ret = -EINVAL; + } + if (frame_count == max_frames) { + kunit_err(current_test, "Maximum number of frames exceeded\n"); + ret = -EINVAL; + } + if (seen_arch_rethook_trampoline) { + kunit_err(current_test, "arch_rethook_trampoline+0x0 in unwinding results\n"); + ret = -EINVAL; + } + if (ret || force_bt) + print_backtrace(bt); + kfree(bt); + return ret; +} + +/* State of the task being unwound. */ +struct unwindme { + int flags; + int ret; + struct task_struct *task; + struct completion task_ready; + wait_queue_head_t task_wq; + unsigned long sp; +}; + +static struct unwindme *unwindme; + +/* Values of unwindme.flags. */ +#define UWM_DEFAULT 0x0 +#define UWM_THREAD 0x1 /* Unwind a separate task. */ +#define UWM_REGS 0x2 /* Pass regs to test_unwind(). */ +#define UWM_SP 0x4 /* Pass sp to test_unwind(). */ +#define UWM_CALLER 0x8 /* Unwind starting from caller. */ +#define UWM_SWITCH_STACK 0x10 /* Use call_on_stack. */ +#define UWM_IRQ 0x20 /* Unwind from irq context. */ +#define UWM_PGM 0x40 /* Unwind from program check handler */ +#define UWM_KPROBE_ON_FTRACE 0x80 /* Unwind from kprobe handler called via ftrace. */ +#define UWM_FTRACE 0x100 /* Unwind from ftrace handler. */ +#define UWM_KRETPROBE 0x200 /* Unwind through kretprobed function. */ +#define UWM_KRETPROBE_HANDLER 0x400 /* Unwind from kretprobe handler. */ + +static __always_inline struct pt_regs fake_pt_regs(void) +{ + struct pt_regs regs; + + memset(®s, 0, sizeof(regs)); + regs.gprs[15] = current_stack_pointer; + + asm volatile( + "basr %[psw_addr],0" + : [psw_addr] "=d" (regs.psw.addr)); + return regs; +} + +static int kretprobe_ret_handler(struct kretprobe_instance *ri, struct pt_regs *regs) +{ + struct unwindme *u = unwindme; + + if (!(u->flags & UWM_KRETPROBE_HANDLER)) + return 0; + + u->ret = test_unwind(NULL, (u->flags & UWM_REGS) ? regs : NULL, + (u->flags & UWM_SP) ? u->sp : 0); + + return 0; +} + +static noinline notrace int test_unwind_kretprobed_func(struct unwindme *u) +{ + struct pt_regs regs; + + if (!(u->flags & UWM_KRETPROBE)) + return 0; + + regs = fake_pt_regs(); + return test_unwind(NULL, (u->flags & UWM_REGS) ? ®s : NULL, + (u->flags & UWM_SP) ? u->sp : 0); +} + +static noinline int test_unwind_kretprobed_func_caller(struct unwindme *u) +{ + return test_unwind_kretprobed_func(u); +} + +static int test_unwind_kretprobe(struct unwindme *u) +{ + int ret; + struct kretprobe my_kretprobe; + + if (!IS_ENABLED(CONFIG_KPROBES)) + kunit_skip(current_test, "requires CONFIG_KPROBES"); + + u->ret = -1; /* make sure kprobe is called */ + unwindme = u; + + memset(&my_kretprobe, 0, sizeof(my_kretprobe)); + my_kretprobe.handler = kretprobe_ret_handler; + my_kretprobe.maxactive = 1; + my_kretprobe.kp.addr = (kprobe_opcode_t *)test_unwind_kretprobed_func; + + ret = register_kretprobe(&my_kretprobe); + + if (ret < 0) { + kunit_err(current_test, "register_kretprobe failed %d\n", ret); + return -EINVAL; + } + + ret = test_unwind_kretprobed_func_caller(u); + unregister_kretprobe(&my_kretprobe); + unwindme = NULL; + if (u->flags & UWM_KRETPROBE_HANDLER) + ret = u->ret; + return ret; +} + +static int kprobe_pre_handler(struct kprobe *p, struct pt_regs *regs) +{ + struct unwindme *u = unwindme; + + u->ret = test_unwind(NULL, (u->flags & UWM_REGS) ? regs : NULL, + (u->flags & UWM_SP) ? u->sp : 0); + return 0; +} + +extern const char test_unwind_kprobed_insn[]; + +static noinline void test_unwind_kprobed_func(void) +{ + asm volatile( + " nopr %%r7\n" + "test_unwind_kprobed_insn:\n" + " nopr %%r7" + :); +} + +static int test_unwind_kprobe(struct unwindme *u) +{ + struct kprobe kp; + int ret; + + if (!IS_ENABLED(CONFIG_KPROBES)) + kunit_skip(current_test, "requires CONFIG_KPROBES"); + if (!IS_ENABLED(CONFIG_KPROBES_ON_FTRACE) && u->flags & UWM_KPROBE_ON_FTRACE) + kunit_skip(current_test, "requires CONFIG_KPROBES_ON_FTRACE"); + + u->ret = -1; /* make sure kprobe is called */ + unwindme = u; + memset(&kp, 0, sizeof(kp)); + kp.pre_handler = kprobe_pre_handler; + kp.addr = u->flags & UWM_KPROBE_ON_FTRACE ? + (kprobe_opcode_t *)test_unwind_kprobed_func : + (kprobe_opcode_t *)test_unwind_kprobed_insn; + ret = register_kprobe(&kp); + if (ret < 0) { + kunit_err(current_test, "register_kprobe failed %d\n", ret); + return -EINVAL; + } + + test_unwind_kprobed_func(); + unregister_kprobe(&kp); + unwindme = NULL; + return u->ret; +} + +static void notrace __used test_unwind_ftrace_handler(unsigned long ip, + unsigned long parent_ip, + struct ftrace_ops *fops, + struct ftrace_regs *fregs) +{ + struct unwindme *u = (struct unwindme *)arch_ftrace_regs(fregs)->regs.gprs[2]; + + u->ret = test_unwind(NULL, (u->flags & UWM_REGS) ? &arch_ftrace_regs(fregs)->regs : NULL, + (u->flags & UWM_SP) ? u->sp : 0); +} + +static noinline int test_unwind_ftraced_func(struct unwindme *u) +{ + return READ_ONCE(u)->ret; +} + +static int test_unwind_ftrace(struct unwindme *u) +{ + int ret; +#ifdef CONFIG_DYNAMIC_FTRACE + struct ftrace_ops *fops; + + fops = kunit_kzalloc(current_test, sizeof(*fops), GFP_KERNEL); + fops->func = test_unwind_ftrace_handler; + fops->flags = FTRACE_OPS_FL_DYNAMIC | + FTRACE_OPS_FL_RECURSION | + FTRACE_OPS_FL_SAVE_REGS | + FTRACE_OPS_FL_PERMANENT; +#else + kunit_skip(current_test, "requires CONFIG_DYNAMIC_FTRACE"); +#endif + + ret = ftrace_set_filter_ip(fops, (unsigned long)test_unwind_ftraced_func, 0, 0); + if (ret) { + kunit_err(current_test, "failed to set ftrace filter (%d)\n", ret); + return -1; + } + + ret = register_ftrace_function(fops); + if (!ret) { + ret = test_unwind_ftraced_func(u); + unregister_ftrace_function(fops); + } else { + kunit_err(current_test, "failed to register ftrace handler (%d)\n", ret); + } + + ftrace_set_filter_ip(fops, (unsigned long)test_unwind_ftraced_func, 1, 0); + return ret; +} + +/* This function may or may not appear in the backtrace. */ +static noinline int unwindme_func4(struct unwindme *u) +{ + if (!(u->flags & UWM_CALLER)) + u->sp = current_frame_address(); + if (u->flags & UWM_THREAD) { + complete(&u->task_ready); + wait_event(u->task_wq, kthread_should_park()); + kthread_parkme(); + return 0; + } else if (u->flags & (UWM_PGM | UWM_KPROBE_ON_FTRACE)) { + return test_unwind_kprobe(u); + } else if (u->flags & (UWM_KRETPROBE | UWM_KRETPROBE_HANDLER)) { + return test_unwind_kretprobe(u); + } else if (u->flags & UWM_FTRACE) { + return test_unwind_ftrace(u); + } else { + struct pt_regs regs = fake_pt_regs(); + + return test_unwind(NULL, + (u->flags & UWM_REGS) ? ®s : NULL, + (u->flags & UWM_SP) ? u->sp : 0); + } +} + +/* This function may or may not appear in the backtrace. */ +static noinline int unwindme_func3(struct unwindme *u) +{ + u->sp = current_frame_address(); + return unwindme_func4(u); +} + +/* This function must appear in the backtrace. */ +static noinline int unwindme_func2(struct unwindme *u) +{ + unsigned long flags, mflags; + int rc; + + if (u->flags & UWM_SWITCH_STACK) { + local_irq_save(flags); + local_mcck_save(mflags); + rc = call_on_stack(1, get_lowcore()->nodat_stack, + int, unwindme_func3, struct unwindme *, u); + local_mcck_restore(mflags); + local_irq_restore(flags); + return rc; + } else { + return unwindme_func3(u); + } +} + +/* This function must follow unwindme_func2 in the backtrace. */ +static noinline int unwindme_func1(void *u) +{ + return unwindme_func2((struct unwindme *)u); +} + +static void unwindme_timer_fn(struct timer_list *unused) +{ + struct unwindme *u = READ_ONCE(unwindme); + + if (u) { + unwindme = NULL; + u->task = NULL; + u->ret = unwindme_func1(u); + complete(&u->task_ready); + } +} + +static struct timer_list unwind_timer; + +static int test_unwind_irq(struct unwindme *u) +{ + unwindme = u; + init_completion(&u->task_ready); + timer_setup(&unwind_timer, unwindme_timer_fn, 0); + mod_timer(&unwind_timer, jiffies + 1); + wait_for_completion(&u->task_ready); + return u->ret; +} + +/* Spawns a task and passes it to test_unwind(). */ +static int test_unwind_task(struct unwindme *u) +{ + struct task_struct *task; + int ret; + + /* Initialize thread-related fields. */ + init_completion(&u->task_ready); + init_waitqueue_head(&u->task_wq); + + /* + * Start the task and wait until it reaches unwindme_func4() and sleeps + * in (task_ready, unwind_done] range. + */ + task = kthread_run(unwindme_func1, u, "%s", __func__); + if (IS_ERR(task)) { + kunit_err(current_test, "kthread_run() failed\n"); + return PTR_ERR(task); + } + /* + * Make sure task reaches unwindme_func4 before parking it, + * we might park it before kthread function has been executed otherwise + */ + wait_for_completion(&u->task_ready); + kthread_park(task); + /* Unwind. */ + ret = test_unwind(task, NULL, (u->flags & UWM_SP) ? u->sp : 0); + kthread_stop(task); + return ret; +} + +struct test_params { + int flags; + char *name; +}; + +/* + * Create required parameter list for tests + */ +#define TEST_WITH_FLAGS(f) { .flags = f, .name = #f } +static const struct test_params param_list[] = { + TEST_WITH_FLAGS(UWM_DEFAULT), + TEST_WITH_FLAGS(UWM_SP), + TEST_WITH_FLAGS(UWM_REGS), + TEST_WITH_FLAGS(UWM_SWITCH_STACK), + TEST_WITH_FLAGS(UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_CALLER | UWM_SP), + TEST_WITH_FLAGS(UWM_CALLER | UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_CALLER | UWM_SP | UWM_REGS | UWM_SWITCH_STACK), + TEST_WITH_FLAGS(UWM_THREAD), + TEST_WITH_FLAGS(UWM_THREAD | UWM_SP), + TEST_WITH_FLAGS(UWM_THREAD | UWM_CALLER | UWM_SP), + TEST_WITH_FLAGS(UWM_IRQ), + TEST_WITH_FLAGS(UWM_IRQ | UWM_SWITCH_STACK), + TEST_WITH_FLAGS(UWM_IRQ | UWM_SP), + TEST_WITH_FLAGS(UWM_IRQ | UWM_REGS), + TEST_WITH_FLAGS(UWM_IRQ | UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_IRQ | UWM_CALLER | UWM_SP), + TEST_WITH_FLAGS(UWM_IRQ | UWM_CALLER | UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_IRQ | UWM_CALLER | UWM_SP | UWM_REGS | UWM_SWITCH_STACK), + TEST_WITH_FLAGS(UWM_PGM), + TEST_WITH_FLAGS(UWM_PGM | UWM_SP), + TEST_WITH_FLAGS(UWM_PGM | UWM_REGS), + TEST_WITH_FLAGS(UWM_PGM | UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_KPROBE_ON_FTRACE), + TEST_WITH_FLAGS(UWM_KPROBE_ON_FTRACE | UWM_SP), + TEST_WITH_FLAGS(UWM_KPROBE_ON_FTRACE | UWM_REGS), + TEST_WITH_FLAGS(UWM_KPROBE_ON_FTRACE | UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_FTRACE), + TEST_WITH_FLAGS(UWM_FTRACE | UWM_SP), + TEST_WITH_FLAGS(UWM_FTRACE | UWM_REGS), + TEST_WITH_FLAGS(UWM_FTRACE | UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_KRETPROBE), + TEST_WITH_FLAGS(UWM_KRETPROBE | UWM_SP), + TEST_WITH_FLAGS(UWM_KRETPROBE | UWM_REGS), + TEST_WITH_FLAGS(UWM_KRETPROBE | UWM_SP | UWM_REGS), + TEST_WITH_FLAGS(UWM_KRETPROBE_HANDLER), + TEST_WITH_FLAGS(UWM_KRETPROBE_HANDLER | UWM_SP), + TEST_WITH_FLAGS(UWM_KRETPROBE_HANDLER | UWM_REGS), + TEST_WITH_FLAGS(UWM_KRETPROBE_HANDLER | UWM_SP | UWM_REGS), +}; + +/* + * Parameter description generator: required for KUNIT_ARRAY_PARAM() + */ +static void get_desc(const struct test_params *params, char *desc) +{ + strscpy(desc, params->name, KUNIT_PARAM_DESC_SIZE); +} + +/* + * Create test_unwind_gen_params + */ +KUNIT_ARRAY_PARAM(test_unwind, param_list, get_desc); + +static void test_unwind_flags(struct kunit *test) +{ + struct unwindme u; + const struct test_params *params; + + current_test = test; + params = (const struct test_params *)test->param_value; + u.flags = params->flags; + if (u.flags & UWM_THREAD) + KUNIT_EXPECT_EQ(test, 0, test_unwind_task(&u)); + else if (u.flags & UWM_IRQ) + KUNIT_EXPECT_EQ(test, 0, test_unwind_irq(&u)); + else + KUNIT_EXPECT_EQ(test, 0, unwindme_func1(&u)); +} + +static struct kunit_case unwind_test_cases[] = { + KUNIT_CASE_PARAM(test_unwind_flags, test_unwind_gen_params), + {} +}; + +static struct kunit_suite test_unwind_suite = { + .name = "test_unwind", + .test_cases = unwind_test_cases, +}; + +kunit_test_suites(&test_unwind_suite); + +MODULE_DESCRIPTION("KUnit test for unwind_for_each_frame"); +MODULE_LICENSE("GPL"); diff --git a/arch/s390/lib/tishift.S b/arch/s390/lib/tishift.S new file mode 100644 index 000000000000..96214f51f49b --- /dev/null +++ b/arch/s390/lib/tishift.S @@ -0,0 +1,63 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#include <linux/export.h> +#include <linux/linkage.h> +#include <asm/nospec-insn.h> + + .section .noinstr.text, "ax" + + GEN_BR_THUNK %r14 + +SYM_FUNC_START(__ashlti3) + lmg %r0,%r1,0(%r3) + cije %r4,0,1f + lhi %r3,64 + sr %r3,%r4 + jnh 0f + srlg %r3,%r1,0(%r3) + sllg %r0,%r0,0(%r4) + sllg %r1,%r1,0(%r4) + ogr %r0,%r3 + j 1f +0: sllg %r0,%r1,-64(%r4) + lghi %r1,0 +1: stmg %r0,%r1,0(%r2) + BR_EX %r14 +SYM_FUNC_END(__ashlti3) +EXPORT_SYMBOL(__ashlti3) + +SYM_FUNC_START(__ashrti3) + lmg %r0,%r1,0(%r3) + cije %r4,0,1f + lhi %r3,64 + sr %r3,%r4 + jnh 0f + sllg %r3,%r0,0(%r3) + srlg %r1,%r1,0(%r4) + srag %r0,%r0,0(%r4) + ogr %r1,%r3 + j 1f +0: srag %r1,%r0,-64(%r4) + srag %r0,%r0,63 +1: stmg %r0,%r1,0(%r2) + BR_EX %r14 +SYM_FUNC_END(__ashrti3) +EXPORT_SYMBOL(__ashrti3) + +SYM_FUNC_START(__lshrti3) + lmg %r0,%r1,0(%r3) + cije %r4,0,1f + lhi %r3,64 + sr %r3,%r4 + jnh 0f + sllg %r3,%r0,0(%r3) + srlg %r1,%r1,0(%r4) + srlg %r0,%r0,0(%r4) + ogr %r1,%r3 + j 1f +0: srlg %r1,%r0,-64(%r4) + lghi %r0,0 +1: stmg %r0,%r1,0(%r2) + BR_EX %r14 +SYM_FUNC_END(__lshrti3) +EXPORT_SYMBOL(__lshrti3) diff --git a/arch/s390/lib/uaccess.c b/arch/s390/lib/uaccess.c new file mode 100644 index 000000000000..1a6ba105e071 --- /dev/null +++ b/arch/s390/lib/uaccess.c @@ -0,0 +1,335 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Standard user space access functions based on mvcp/mvcs and doing + * interesting things in the secondary space mode. + * + * Copyright IBM Corp. 2006,2014 + * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), + * Gerald Schaefer (gerald.schaefer@de.ibm.com) + */ + +#include <linux/kprobes.h> +#include <linux/uaccess.h> +#include <linux/export.h> +#include <linux/mm.h> +#include <asm/asm-extable.h> +#include <asm/ctlreg.h> +#include <asm/skey.h> + +#ifdef CONFIG_DEBUG_ENTRY +void debug_user_asce(int exit) +{ + struct lowcore *lc = get_lowcore(); + struct ctlreg cr1, cr7; + + local_ctl_store(1, &cr1); + local_ctl_store(7, &cr7); + if (cr1.val == lc->user_asce.val && cr7.val == lc->user_asce.val) + return; + panic("incorrect ASCE on kernel %s\n" + "cr1: %016lx cr7: %016lx\n" + "kernel: %016lx user: %016lx\n", + exit ? "exit" : "entry", cr1.val, cr7.val, + lc->kernel_asce.val, lc->user_asce.val); +} +#endif /*CONFIG_DEBUG_ENTRY */ + +union oac { + unsigned int val; + struct { + struct { + unsigned short key : 4; + unsigned short : 4; + unsigned short as : 2; + unsigned short : 4; + unsigned short k : 1; + unsigned short a : 1; + } oac1; + struct { + unsigned short key : 4; + unsigned short : 4; + unsigned short as : 2; + unsigned short : 4; + unsigned short k : 1; + unsigned short a : 1; + } oac2; + }; +}; + +static uaccess_kmsan_or_inline __must_check unsigned long +raw_copy_from_user_key(void *to, const void __user *from, unsigned long size, unsigned long key) +{ + unsigned long osize; + union oac spec = { + .oac2.key = key, + .oac2.as = PSW_BITS_AS_SECONDARY, + .oac2.k = 1, + .oac2.a = 1, + }; + int cc; + + while (1) { + osize = size; + asm_inline volatile( + " lr %%r0,%[spec]\n" + "0: mvcos %[to],%[from],%[size]\n" + "1: nopr %%r7\n" + CC_IPM(cc) + EX_TABLE_UA_MVCOS_FROM(0b, 0b) + EX_TABLE_UA_MVCOS_FROM(1b, 0b) + : CC_OUT(cc, cc), [size] "+d" (size), [to] "=Q" (*(char *)to) + : [spec] "d" (spec.val), [from] "Q" (*(const char __user *)from) + : CC_CLOBBER_LIST("memory", "0")); + if (CC_TRANSFORM(cc) == 0) + return osize - size; + size -= 4096; + to += 4096; + from += 4096; + } +} + +unsigned long _copy_from_user_key(void *to, const void __user *from, + unsigned long n, unsigned long key) +{ + unsigned long res = n; + + might_fault(); + if (!should_fail_usercopy()) { + instrument_copy_from_user_before(to, from, n); + res = raw_copy_from_user_key(to, from, n, key); + instrument_copy_from_user_after(to, from, n, res); + } + if (unlikely(res)) + memset(to + (n - res), 0, res); + return res; +} +EXPORT_SYMBOL(_copy_from_user_key); + +static uaccess_kmsan_or_inline __must_check unsigned long +raw_copy_to_user_key(void __user *to, const void *from, unsigned long size, unsigned long key) +{ + unsigned long osize; + union oac spec = { + .oac1.key = key, + .oac1.as = PSW_BITS_AS_SECONDARY, + .oac1.k = 1, + .oac1.a = 1, + }; + int cc; + + while (1) { + osize = size; + asm_inline volatile( + " lr %%r0,%[spec]\n" + "0: mvcos %[to],%[from],%[size]\n" + "1: nopr %%r7\n" + CC_IPM(cc) + EX_TABLE_UA_MVCOS_TO(0b, 0b) + EX_TABLE_UA_MVCOS_TO(1b, 0b) + : CC_OUT(cc, cc), [size] "+d" (size), [to] "=Q" (*(char __user *)to) + : [spec] "d" (spec.val), [from] "Q" (*(const char *)from) + : CC_CLOBBER_LIST("memory", "0")); + if (CC_TRANSFORM(cc) == 0) + return osize - size; + size -= 4096; + to += 4096; + from += 4096; + } +} + +unsigned long _copy_to_user_key(void __user *to, const void *from, + unsigned long n, unsigned long key) +{ + might_fault(); + if (should_fail_usercopy()) + return n; + instrument_copy_to_user(to, from, n); + return raw_copy_to_user_key(to, from, n, key); +} +EXPORT_SYMBOL(_copy_to_user_key); + +#define CMPXCHG_USER_KEY_MAX_LOOPS 128 + +static nokprobe_inline int __cmpxchg_user_key_small(unsigned long address, unsigned int *uval, + unsigned int old, unsigned int new, + unsigned int mask, unsigned long key) +{ + unsigned long count; + unsigned int prev; + bool sacf_flag; + int rc = 0; + + skey_regions_initialize(); + sacf_flag = enable_sacf_uaccess(); + asm_inline volatile( + "20: spka 0(%[key])\n" + " sacf 256\n" + " llill %[count],%[max_loops]\n" + "0: l %[prev],%[address]\n" + "1: nr %[prev],%[mask]\n" + " xilf %[mask],0xffffffff\n" + " or %[new],%[prev]\n" + " or %[prev],%[tmp]\n" + "2: lr %[tmp],%[prev]\n" + "3: cs %[prev],%[new],%[address]\n" + "4: jnl 5f\n" + " xr %[tmp],%[prev]\n" + " xr %[new],%[tmp]\n" + " nr %[tmp],%[mask]\n" + " jnz 5f\n" + " brct %[count],2b\n" + "5: sacf 768\n" + " spka %[default_key]\n" + "21:\n" + EX_TABLE_UA_LOAD_REG(0b, 5b, %[rc], %[prev]) + EX_TABLE_UA_LOAD_REG(1b, 5b, %[rc], %[prev]) + EX_TABLE_UA_LOAD_REG(3b, 5b, %[rc], %[prev]) + EX_TABLE_UA_LOAD_REG(4b, 5b, %[rc], %[prev]) + SKEY_REGION(20b, 21b) + : [rc] "+&d" (rc), + [prev] "=&d" (prev), + [address] "+Q" (*(int *)address), + [tmp] "+&d" (old), + [new] "+&d" (new), + [mask] "+&d" (mask), + [count] "=a" (count) + : [key] "%[count]" (key << 4), + [default_key] "J" (PAGE_DEFAULT_KEY), + [max_loops] "J" (CMPXCHG_USER_KEY_MAX_LOOPS) + : "memory", "cc"); + disable_sacf_uaccess(sacf_flag); + *uval = prev; + if (!count) + rc = -EAGAIN; + return rc; +} + +int __kprobes __cmpxchg_user_key1(unsigned long address, unsigned char *uval, + unsigned char old, unsigned char new, unsigned long key) +{ + unsigned int prev, shift, mask, _old, _new; + int rc; + + shift = (3 ^ (address & 3)) << 3; + address ^= address & 3; + _old = (unsigned int)old << shift; + _new = (unsigned int)new << shift; + mask = ~(0xff << shift); + rc = __cmpxchg_user_key_small(address, &prev, _old, _new, mask, key); + *uval = prev >> shift; + return rc; +} +EXPORT_SYMBOL(__cmpxchg_user_key1); + +int __kprobes __cmpxchg_user_key2(unsigned long address, unsigned short *uval, + unsigned short old, unsigned short new, unsigned long key) +{ + unsigned int prev, shift, mask, _old, _new; + int rc; + + shift = (2 ^ (address & 2)) << 3; + address ^= address & 2; + _old = (unsigned int)old << shift; + _new = (unsigned int)new << shift; + mask = ~(0xffff << shift); + rc = __cmpxchg_user_key_small(address, &prev, _old, _new, mask, key); + *uval = prev >> shift; + return rc; +} +EXPORT_SYMBOL(__cmpxchg_user_key2); + +int __kprobes __cmpxchg_user_key4(unsigned long address, unsigned int *uval, + unsigned int old, unsigned int new, unsigned long key) +{ + unsigned int prev = old; + bool sacf_flag; + int rc = 0; + + skey_regions_initialize(); + sacf_flag = enable_sacf_uaccess(); + asm_inline volatile( + "20: spka 0(%[key])\n" + " sacf 256\n" + "0: cs %[prev],%[new],%[address]\n" + "1: sacf 768\n" + " spka %[default_key]\n" + "21:\n" + EX_TABLE_UA_LOAD_REG(0b, 1b, %[rc], %[prev]) + EX_TABLE_UA_LOAD_REG(1b, 1b, %[rc], %[prev]) + SKEY_REGION(20b, 21b) + : [rc] "+&d" (rc), + [prev] "+&d" (prev), + [address] "+Q" (*(int *)address) + : [new] "d" (new), + [key] "a" (key << 4), + [default_key] "J" (PAGE_DEFAULT_KEY) + : "memory", "cc"); + disable_sacf_uaccess(sacf_flag); + *uval = prev; + return rc; +} +EXPORT_SYMBOL(__cmpxchg_user_key4); + +int __kprobes __cmpxchg_user_key8(unsigned long address, unsigned long *uval, + unsigned long old, unsigned long new, unsigned long key) +{ + unsigned long prev = old; + bool sacf_flag; + int rc = 0; + + skey_regions_initialize(); + sacf_flag = enable_sacf_uaccess(); + asm_inline volatile( + "20: spka 0(%[key])\n" + " sacf 256\n" + "0: csg %[prev],%[new],%[address]\n" + "1: sacf 768\n" + " spka %[default_key]\n" + "21:\n" + EX_TABLE_UA_LOAD_REG(0b, 1b, %[rc], %[prev]) + EX_TABLE_UA_LOAD_REG(1b, 1b, %[rc], %[prev]) + SKEY_REGION(20b, 21b) + : [rc] "+&d" (rc), + [prev] "+&d" (prev), + [address] "+QS" (*(long *)address) + : [new] "d" (new), + [key] "a" (key << 4), + [default_key] "J" (PAGE_DEFAULT_KEY) + : "memory", "cc"); + disable_sacf_uaccess(sacf_flag); + *uval = prev; + return rc; +} +EXPORT_SYMBOL(__cmpxchg_user_key8); + +int __kprobes __cmpxchg_user_key16(unsigned long address, __uint128_t *uval, + __uint128_t old, __uint128_t new, unsigned long key) +{ + __uint128_t prev = old; + bool sacf_flag; + int rc = 0; + + skey_regions_initialize(); + sacf_flag = enable_sacf_uaccess(); + asm_inline volatile( + "20: spka 0(%[key])\n" + " sacf 256\n" + "0: cdsg %[prev],%[new],%[address]\n" + "1: sacf 768\n" + " spka %[default_key]\n" + "21:\n" + EX_TABLE_UA_LOAD_REGPAIR(0b, 1b, %[rc], %[prev]) + EX_TABLE_UA_LOAD_REGPAIR(1b, 1b, %[rc], %[prev]) + SKEY_REGION(20b, 21b) + : [rc] "+&d" (rc), + [prev] "+&d" (prev), + [address] "+QS" (*(__int128_t *)address) + : [new] "d" (new), + [key] "a" (key << 4), + [default_key] "J" (PAGE_DEFAULT_KEY) + : "memory", "cc"); + disable_sacf_uaccess(sacf_flag); + *uval = prev; + return rc; +} +EXPORT_SYMBOL(__cmpxchg_user_key16); diff --git a/arch/s390/lib/uaccess.h b/arch/s390/lib/uaccess.h deleted file mode 100644 index 315dbe09983e..000000000000 --- a/arch/s390/lib/uaccess.h +++ /dev/null @@ -1,21 +0,0 @@ -/* - * Copyright IBM Corp. 2007 - * - */ - -#ifndef __ARCH_S390_LIB_UACCESS_H -#define __ARCH_S390_LIB_UACCESS_H - -extern size_t copy_from_user_std(size_t, const void __user *, void *); -extern size_t copy_to_user_std(size_t, void __user *, const void *); -extern size_t strnlen_user_std(size_t, const char __user *); -extern size_t strncpy_from_user_std(size_t, const char __user *, char *); -extern int futex_atomic_cmpxchg_std(u32 *, u32 __user *, u32, u32); -extern int futex_atomic_op_std(int, u32 __user *, int, int *); - -extern size_t copy_from_user_pt(size_t, const void __user *, void *); -extern size_t copy_to_user_pt(size_t, void __user *, const void *); -extern int futex_atomic_op_pt(int, u32 __user *, int, int *); -extern int futex_atomic_cmpxchg_pt(u32 *, u32 __user *, u32, u32); - -#endif /* __ARCH_S390_LIB_UACCESS_H */ diff --git a/arch/s390/lib/uaccess_mvcos.c b/arch/s390/lib/uaccess_mvcos.c deleted file mode 100644 index 1829742bf479..000000000000 --- a/arch/s390/lib/uaccess_mvcos.c +++ /dev/null @@ -1,227 +0,0 @@ -/* - * Optimized user space space access functions based on mvcos. - * - * Copyright IBM Corp. 2006 - * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), - * Gerald Schaefer (gerald.schaefer@de.ibm.com) - */ - -#include <linux/errno.h> -#include <linux/mm.h> -#include <asm/uaccess.h> -#include <asm/futex.h> -#include "uaccess.h" - -#ifndef CONFIG_64BIT -#define AHI "ahi" -#define ALR "alr" -#define CLR "clr" -#define LHI "lhi" -#define SLR "slr" -#else -#define AHI "aghi" -#define ALR "algr" -#define CLR "clgr" -#define LHI "lghi" -#define SLR "slgr" -#endif - -static size_t copy_from_user_mvcos(size_t size, const void __user *ptr, void *x) -{ - register unsigned long reg0 asm("0") = 0x81UL; - unsigned long tmp1, tmp2; - - tmp1 = -4096UL; - asm volatile( - "0: .insn ss,0xc80000000000,0(%0,%2),0(%1),0\n" - "9: jz 7f\n" - "1:"ALR" %0,%3\n" - " "SLR" %1,%3\n" - " "SLR" %2,%3\n" - " j 0b\n" - "2: la %4,4095(%1)\n"/* %4 = ptr + 4095 */ - " nr %4,%3\n" /* %4 = (ptr + 4095) & -4096 */ - " "SLR" %4,%1\n" - " "CLR" %0,%4\n" /* copy crosses next page boundary? */ - " jnh 4f\n" - "3: .insn ss,0xc80000000000,0(%4,%2),0(%1),0\n" - "10:"SLR" %0,%4\n" - " "ALR" %2,%4\n" - "4:"LHI" %4,-1\n" - " "ALR" %4,%0\n" /* copy remaining size, subtract 1 */ - " bras %3,6f\n" /* memset loop */ - " xc 0(1,%2),0(%2)\n" - "5: xc 0(256,%2),0(%2)\n" - " la %2,256(%2)\n" - "6:"AHI" %4,-256\n" - " jnm 5b\n" - " ex %4,0(%3)\n" - " j 8f\n" - "7:"SLR" %0,%0\n" - "8: \n" - EX_TABLE(0b,2b) EX_TABLE(3b,4b) EX_TABLE(9b,2b) EX_TABLE(10b,4b) - : "+a" (size), "+a" (ptr), "+a" (x), "+a" (tmp1), "=a" (tmp2) - : "d" (reg0) : "cc", "memory"); - return size; -} - -static size_t copy_from_user_mvcos_check(size_t size, const void __user *ptr, void *x) -{ - if (size <= 256) - return copy_from_user_std(size, ptr, x); - return copy_from_user_mvcos(size, ptr, x); -} - -static size_t copy_to_user_mvcos(size_t size, void __user *ptr, const void *x) -{ - register unsigned long reg0 asm("0") = 0x810000UL; - unsigned long tmp1, tmp2; - - tmp1 = -4096UL; - asm volatile( - "0: .insn ss,0xc80000000000,0(%0,%1),0(%2),0\n" - "6: jz 4f\n" - "1:"ALR" %0,%3\n" - " "SLR" %1,%3\n" - " "SLR" %2,%3\n" - " j 0b\n" - "2: la %4,4095(%1)\n"/* %4 = ptr + 4095 */ - " nr %4,%3\n" /* %4 = (ptr + 4095) & -4096 */ - " "SLR" %4,%1\n" - " "CLR" %0,%4\n" /* copy crosses next page boundary? */ - " jnh 5f\n" - "3: .insn ss,0xc80000000000,0(%4,%1),0(%2),0\n" - "7:"SLR" %0,%4\n" - " j 5f\n" - "4:"SLR" %0,%0\n" - "5: \n" - EX_TABLE(0b,2b) EX_TABLE(3b,5b) EX_TABLE(6b,2b) EX_TABLE(7b,5b) - : "+a" (size), "+a" (ptr), "+a" (x), "+a" (tmp1), "=a" (tmp2) - : "d" (reg0) : "cc", "memory"); - return size; -} - -static size_t copy_to_user_mvcos_check(size_t size, void __user *ptr, - const void *x) -{ - if (size <= 256) - return copy_to_user_std(size, ptr, x); - return copy_to_user_mvcos(size, ptr, x); -} - -static size_t copy_in_user_mvcos(size_t size, void __user *to, - const void __user *from) -{ - register unsigned long reg0 asm("0") = 0x810081UL; - unsigned long tmp1, tmp2; - - tmp1 = -4096UL; - /* FIXME: copy with reduced length. */ - asm volatile( - "0: .insn ss,0xc80000000000,0(%0,%1),0(%2),0\n" - " jz 2f\n" - "1:"ALR" %0,%3\n" - " "SLR" %1,%3\n" - " "SLR" %2,%3\n" - " j 0b\n" - "2:"SLR" %0,%0\n" - "3: \n" - EX_TABLE(0b,3b) - : "+a" (size), "+a" (to), "+a" (from), "+a" (tmp1), "=a" (tmp2) - : "d" (reg0) : "cc", "memory"); - return size; -} - -static size_t clear_user_mvcos(size_t size, void __user *to) -{ - register unsigned long reg0 asm("0") = 0x810000UL; - unsigned long tmp1, tmp2; - - tmp1 = -4096UL; - asm volatile( - "0: .insn ss,0xc80000000000,0(%0,%1),0(%4),0\n" - " jz 4f\n" - "1:"ALR" %0,%2\n" - " "SLR" %1,%2\n" - " j 0b\n" - "2: la %3,4095(%1)\n"/* %4 = to + 4095 */ - " nr %3,%2\n" /* %4 = (to + 4095) & -4096 */ - " "SLR" %3,%1\n" - " "CLR" %0,%3\n" /* copy crosses next page boundary? */ - " jnh 5f\n" - "3: .insn ss,0xc80000000000,0(%3,%1),0(%4),0\n" - " "SLR" %0,%3\n" - " j 5f\n" - "4:"SLR" %0,%0\n" - "5: \n" - EX_TABLE(0b,2b) EX_TABLE(3b,5b) - : "+a" (size), "+a" (to), "+a" (tmp1), "=a" (tmp2) - : "a" (empty_zero_page), "d" (reg0) : "cc", "memory"); - return size; -} - -static size_t strnlen_user_mvcos(size_t count, const char __user *src) -{ - size_t done, len, offset, len_str; - char buf[256]; - - done = 0; - do { - offset = (size_t)src & ~PAGE_MASK; - len = min(256UL, PAGE_SIZE - offset); - len = min(count - done, len); - if (copy_from_user_mvcos(len, src, buf)) - return 0; - len_str = strnlen(buf, len); - done += len_str; - src += len_str; - } while ((len_str == len) && (done < count)); - return done + 1; -} - -static size_t strncpy_from_user_mvcos(size_t count, const char __user *src, - char *dst) -{ - size_t done, len, offset, len_str; - - if (unlikely(!count)) - return 0; - done = 0; - do { - offset = (size_t)src & ~PAGE_MASK; - len = min(count - done, PAGE_SIZE - offset); - if (copy_from_user_mvcos(len, src, dst)) - return -EFAULT; - len_str = strnlen(dst, len); - done += len_str; - src += len_str; - dst += len_str; - } while ((len_str == len) && (done < count)); - return done; -} - -struct uaccess_ops uaccess_mvcos = { - .copy_from_user = copy_from_user_mvcos_check, - .copy_from_user_small = copy_from_user_std, - .copy_to_user = copy_to_user_mvcos_check, - .copy_to_user_small = copy_to_user_std, - .copy_in_user = copy_in_user_mvcos, - .clear_user = clear_user_mvcos, - .strnlen_user = strnlen_user_std, - .strncpy_from_user = strncpy_from_user_std, - .futex_atomic_op = futex_atomic_op_std, - .futex_atomic_cmpxchg = futex_atomic_cmpxchg_std, -}; - -struct uaccess_ops uaccess_mvcos_switch = { - .copy_from_user = copy_from_user_mvcos, - .copy_from_user_small = copy_from_user_mvcos, - .copy_to_user = copy_to_user_mvcos, - .copy_to_user_small = copy_to_user_mvcos, - .copy_in_user = copy_in_user_mvcos, - .clear_user = clear_user_mvcos, - .strnlen_user = strnlen_user_mvcos, - .strncpy_from_user = strncpy_from_user_mvcos, - .futex_atomic_op = futex_atomic_op_pt, - .futex_atomic_cmpxchg = futex_atomic_cmpxchg_pt, -}; diff --git a/arch/s390/lib/uaccess_pt.c b/arch/s390/lib/uaccess_pt.c deleted file mode 100644 index 50ea137a2d3c..000000000000 --- a/arch/s390/lib/uaccess_pt.c +++ /dev/null @@ -1,473 +0,0 @@ -/* - * User access functions based on page table walks for enhanced - * system layout without hardware support. - * - * 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" - -#ifndef CONFIG_64BIT -#define AHI "ahi" -#define SLR "slr" -#else -#define AHI "aghi" -#define SLR "slgr" -#endif - -static size_t strnlen_kernel(size_t count, const char __user *src) -{ - register unsigned long reg0 asm("0") = 0UL; - unsigned long tmp1, tmp2; - - asm volatile( - " la %2,0(%1)\n" - " la %3,0(%0,%1)\n" - " "SLR" %0,%0\n" - "0: srst %3,%2\n" - " jo 0b\n" - " la %0,1(%3)\n" /* strnlen_kernel results includes \0 */ - " "SLR" %0,%1\n" - "1:\n" - EX_TABLE(0b,1b) - : "+a" (count), "+a" (src), "=a" (tmp1), "=a" (tmp2) - : "d" (reg0) : "cc", "memory"); - return count; -} - -static size_t copy_in_kernel(size_t count, void __user *to, - const void __user *from) -{ - unsigned long tmp1; - - asm volatile( - " "AHI" %0,-1\n" - " jo 5f\n" - " bras %3,3f\n" - "0:"AHI" %0,257\n" - "1: mvc 0(1,%1),0(%2)\n" - " la %1,1(%1)\n" - " la %2,1(%2)\n" - " "AHI" %0,-1\n" - " jnz 1b\n" - " j 5f\n" - "2: mvc 0(256,%1),0(%2)\n" - " la %1,256(%1)\n" - " la %2,256(%2)\n" - "3:"AHI" %0,-256\n" - " jnm 2b\n" - "4: ex %0,1b-0b(%3)\n" - "5:"SLR" %0,%0\n" - "6:\n" - EX_TABLE(1b,6b) EX_TABLE(2b,0b) EX_TABLE(4b,0b) - : "+a" (count), "+a" (to), "+a" (from), "=a" (tmp1) - : : "cc", "memory"); - return count; -} - -/* - * 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. - */ -#ifdef CONFIG_64BIT -static unsigned long follow_table(struct mm_struct *mm, - unsigned long address, int write) -{ - unsigned long *table = (unsigned long *)__pa(mm->pgd); - - switch (mm->context.asce_bits & _ASCE_TYPE_MASK) { - case _ASCE_TYPE_REGION1: - table = table + ((address >> 53) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INV)) - return -0x39UL; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - /* fallthrough */ - case _ASCE_TYPE_REGION2: - table = table + ((address >> 42) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INV)) - return -0x3aUL; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - /* fallthrough */ - case _ASCE_TYPE_REGION3: - table = table + ((address >> 31) & 0x7ff); - if (unlikely(*table & _REGION_ENTRY_INV)) - return -0x3bUL; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); - /* fallthrough */ - case _ASCE_TYPE_SEGMENT: - table = table + ((address >> 20) & 0x7ff); - if (unlikely(*table & _SEGMENT_ENTRY_INV)) - return -0x10UL; - if (unlikely(*table & _SEGMENT_ENTRY_LARGE)) { - if (write && (*table & _SEGMENT_ENTRY_RO)) - return -0x04UL; - return (*table & _SEGMENT_ENTRY_ORIGIN_LARGE) + - (address & ~_SEGMENT_ENTRY_ORIGIN_LARGE); - } - table = (unsigned long *)(*table & _SEGMENT_ENTRY_ORIGIN); - } - table = table + ((address >> 12) & 0xff); - if (unlikely(*table & _PAGE_INVALID)) - return -0x11UL; - if (write && (*table & _PAGE_RO)) - return -0x04UL; - return (*table & PAGE_MASK) + (address & ~PAGE_MASK); -} - -#else /* CONFIG_64BIT */ - -static unsigned long follow_table(struct mm_struct *mm, - unsigned long address, int write) -{ - unsigned long *table = (unsigned long *)__pa(mm->pgd); - - table = table + ((address >> 20) & 0x7ff); - if (unlikely(*table & _SEGMENT_ENTRY_INV)) - return -0x10UL; - table = (unsigned long *)(*table & _SEGMENT_ENTRY_ORIGIN); - table = table + ((address >> 12) & 0xff); - if (unlikely(*table & _PAGE_INVALID)) - return -0x11UL; - if (write && (*table & _PAGE_RO)) - return -0x04UL; - return (*table & PAGE_MASK) + (address & ~PAGE_MASK); -} - -#endif /* CONFIG_64BIT */ - -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, done, size, kaddr; - void *from, *to; - - done = 0; -retry: - spin_lock(&mm->page_table_lock); - do { - kaddr = follow_table(mm, uaddr, write_user); - if (IS_ERR_VALUE(kaddr)) - goto fault; - - offset = uaddr & ~PAGE_MASK; - size = min(n - done, PAGE_SIZE - offset); - if (write_user) { - to = (void *) kaddr; - from = kptr + done; - } else { - from = (void *) kaddr; - to = kptr + done; - } - memcpy(to, from, size); - done += size; - uaddr += size; - } while (done < n); - spin_unlock(&mm->page_table_lock); - return n - done; -fault: - spin_unlock(&mm->page_table_lock); - if (__handle_fault(uaddr, -kaddr, write_user)) - return n - done; - goto retry; -} - -/* - * 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, - int write) -{ - struct mm_struct *mm = current->mm; - unsigned long kaddr; - int rc; - -retry: - kaddr = follow_table(mm, uaddr, write); - if (IS_ERR_VALUE(kaddr)) - goto fault; - - return kaddr; -fault: - spin_unlock(&mm->page_table_lock); - rc = __handle_fault(uaddr, -kaddr, write); - spin_lock(&mm->page_table_lock); - if (!rc) - goto retry; - return 0; -} - -size_t copy_from_user_pt(size_t n, const void __user *from, void *to) -{ - size_t rc; - - if (segment_eq(get_fs(), KERNEL_DS)) - return copy_in_kernel(n, (void __user *) to, from); - rc = __user_copy_pt((unsigned long) from, to, n, 0); - if (unlikely(rc)) - memset(to + n - rc, 0, rc); - return rc; -} - -size_t copy_to_user_pt(size_t n, void __user *to, const void *from) -{ - if (segment_eq(get_fs(), KERNEL_DS)) - return copy_in_kernel(n, to, (void __user *) from); - return __user_copy_pt((unsigned long) to, (void *) from, n, 1); -} - -static size_t clear_user_pt(size_t n, void __user *to) -{ - void *zpage = (void *) empty_zero_page; - long done, size, ret; - - done = 0; - do { - if (n - done > PAGE_SIZE) - size = PAGE_SIZE; - else - size = n - done; - if (segment_eq(get_fs(), KERNEL_DS)) - ret = copy_in_kernel(n, to, (void __user *) zpage); - else - ret = __user_copy_pt((unsigned long) to, zpage, size, 1); - done += size; - to += size; - if (ret) - return ret + n - done; - } while (done < n); - return 0; -} - -static size_t strnlen_user_pt(size_t count, const char __user *src) -{ - unsigned long uaddr = (unsigned long) src; - struct mm_struct *mm = current->mm; - unsigned long offset, done, len, kaddr; - size_t len_str; - - if (unlikely(!count)) - return 0; - if (segment_eq(get_fs(), KERNEL_DS)) - return strnlen_kernel(count, src); - done = 0; -retry: - spin_lock(&mm->page_table_lock); - do { - kaddr = follow_table(mm, uaddr, 0); - if (IS_ERR_VALUE(kaddr)) - goto fault; - - offset = uaddr & ~PAGE_MASK; - len = min(count - done, PAGE_SIZE - offset); - len_str = strnlen((char *) kaddr, len); - done += len_str; - uaddr += len_str; - } while ((len_str == len) && (done < count)); - spin_unlock(&mm->page_table_lock); - return done + 1; -fault: - spin_unlock(&mm->page_table_lock); - if (__handle_fault(uaddr, -kaddr, 0)) - return 0; - goto retry; -} - -static size_t strncpy_from_user_pt(size_t count, const char __user *src, - char *dst) -{ - size_t done, len, offset, len_str; - - if (unlikely(!count)) - return 0; - done = 0; - do { - offset = (size_t)src & ~PAGE_MASK; - len = min(count - done, PAGE_SIZE - offset); - if (segment_eq(get_fs(), KERNEL_DS)) { - if (copy_in_kernel(len, (void __user *) dst, src)) - return -EFAULT; - } else { - if (__user_copy_pt((unsigned long) src, dst, len, 0)) - return -EFAULT; - } - len_str = strnlen(dst, len); - done += len_str; - src += len_str; - dst += len_str; - } while ((len_str == len) && (done < count)); - return done; -} - -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_max, uaddr, done, size, error_code; - unsigned long uaddr_from = (unsigned long) from; - unsigned long uaddr_to = (unsigned long) to; - unsigned long kaddr_to, kaddr_from; - int write_user; - - if (segment_eq(get_fs(), KERNEL_DS)) - return copy_in_kernel(n, to, from); - done = 0; -retry: - spin_lock(&mm->page_table_lock); - do { - write_user = 0; - uaddr = uaddr_from; - 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; - kaddr_to = follow_table(mm, uaddr_to, 1); - error_code = (unsigned long) kaddr_to; - if (IS_ERR_VALUE(error_code)) - goto fault; - - offset_max = max(uaddr_from & ~PAGE_MASK, - uaddr_to & ~PAGE_MASK); - size = min(n - done, PAGE_SIZE - offset_max); - - memcpy((void *) kaddr_to, (void *) kaddr_from, size); - done += size; - uaddr_from += size; - uaddr_to += size; - } while (done < n); - spin_unlock(&mm->page_table_lock); - return n - done; -fault: - spin_unlock(&mm->page_table_lock); - if (__handle_fault(uaddr, -error_code, write_user)) - return n - done; - goto retry; -} - -#define __futex_atomic_op(insn, ret, oldval, newval, uaddr, oparg) \ - asm volatile("0: l %1,0(%6)\n" \ - "1: " insn \ - "2: cs %1,%2,0(%6)\n" \ - "3: jl 1b\n" \ - " lhi %0,0\n" \ - "4:\n" \ - EX_TABLE(0b,4b) EX_TABLE(2b,4b) EX_TABLE(3b,4b) \ - : "=d" (ret), "=&d" (oldval), "=&d" (newval), \ - "=m" (*uaddr) \ - : "0" (-EFAULT), "d" (oparg), "a" (uaddr), \ - "m" (*uaddr) : "cc" ); - -static int __futex_atomic_op_pt(int op, u32 __user *uaddr, int oparg, int *old) -{ - int oldval = 0, newval, ret; - - switch (op) { - case FUTEX_OP_SET: - __futex_atomic_op("lr %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_ADD: - __futex_atomic_op("lr %2,%1\nar %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_OR: - __futex_atomic_op("lr %2,%1\nor %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_ANDN: - __futex_atomic_op("lr %2,%1\nnr %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_XOR: - __futex_atomic_op("lr %2,%1\nxr %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - default: - ret = -ENOSYS; - } - if (ret == 0) - *old = oldval; - return ret; -} - -int futex_atomic_op_pt(int op, u32 __user *uaddr, int oparg, int *old) -{ - int ret; - - if (segment_eq(get_fs(), KERNEL_DS)) - 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, 1); - if (!uaddr) { - spin_unlock(¤t->mm->page_table_lock); - return -EFAULT; - } - get_page(virt_to_page(uaddr)); - spin_unlock(¤t->mm->page_table_lock); - ret = __futex_atomic_op_pt(op, uaddr, oparg, old); - put_page(virt_to_page(uaddr)); - return ret; -} - -static int __futex_atomic_cmpxchg_pt(u32 *uval, u32 __user *uaddr, - u32 oldval, u32 newval) -{ - int ret; - - asm volatile("0: cs %1,%4,0(%5)\n" - "1: la %0,0\n" - "2:\n" - EX_TABLE(0b,2b) EX_TABLE(1b,2b) - : "=d" (ret), "+d" (oldval), "=m" (*uaddr) - : "0" (-EFAULT), "d" (newval), "a" (uaddr), "m" (*uaddr) - : "cc", "memory" ); - *uval = oldval; - return ret; -} - -int futex_atomic_cmpxchg_pt(u32 *uval, u32 __user *uaddr, - u32 oldval, u32 newval) -{ - int ret; - - if (segment_eq(get_fs(), KERNEL_DS)) - 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, 1); - if (!uaddr) { - spin_unlock(¤t->mm->page_table_lock); - return -EFAULT; - } - get_page(virt_to_page(uaddr)); - spin_unlock(¤t->mm->page_table_lock); - ret = __futex_atomic_cmpxchg_pt(uval, uaddr, oldval, newval); - put_page(virt_to_page(uaddr)); - return ret; -} - -struct uaccess_ops uaccess_pt = { - .copy_from_user = copy_from_user_pt, - .copy_from_user_small = copy_from_user_pt, - .copy_to_user = copy_to_user_pt, - .copy_to_user_small = copy_to_user_pt, - .copy_in_user = copy_in_user_pt, - .clear_user = clear_user_pt, - .strnlen_user = strnlen_user_pt, - .strncpy_from_user = strncpy_from_user_pt, - .futex_atomic_op = futex_atomic_op_pt, - .futex_atomic_cmpxchg = futex_atomic_cmpxchg_pt, -}; diff --git a/arch/s390/lib/uaccess_std.c b/arch/s390/lib/uaccess_std.c deleted file mode 100644 index 4a75d475b06a..000000000000 --- a/arch/s390/lib/uaccess_std.c +++ /dev/null @@ -1,305 +0,0 @@ -/* - * Standard user space access functions based on mvcp/mvcs and doing - * interesting things in the secondary space mode. - * - * Copyright IBM Corp. 2006 - * Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com), - * Gerald Schaefer (gerald.schaefer@de.ibm.com) - */ - -#include <linux/errno.h> -#include <linux/mm.h> -#include <linux/uaccess.h> -#include <asm/futex.h> -#include "uaccess.h" - -#ifndef CONFIG_64BIT -#define AHI "ahi" -#define ALR "alr" -#define CLR "clr" -#define LHI "lhi" -#define SLR "slr" -#else -#define AHI "aghi" -#define ALR "algr" -#define CLR "clgr" -#define LHI "lghi" -#define SLR "slgr" -#endif - -size_t copy_from_user_std(size_t size, const void __user *ptr, void *x) -{ - unsigned long tmp1, tmp2; - - tmp1 = -256UL; - asm volatile( - "0: mvcp 0(%0,%2),0(%1),%3\n" - "10:jz 8f\n" - "1:"ALR" %0,%3\n" - " la %1,256(%1)\n" - " la %2,256(%2)\n" - "2: mvcp 0(%0,%2),0(%1),%3\n" - "11:jnz 1b\n" - " j 8f\n" - "3: la %4,255(%1)\n" /* %4 = ptr + 255 */ - " "LHI" %3,-4096\n" - " nr %4,%3\n" /* %4 = (ptr + 255) & -4096 */ - " "SLR" %4,%1\n" - " "CLR" %0,%4\n" /* copy crosses next page boundary? */ - " jnh 5f\n" - "4: mvcp 0(%4,%2),0(%1),%3\n" - "12:"SLR" %0,%4\n" - " "ALR" %2,%4\n" - "5:"LHI" %4,-1\n" - " "ALR" %4,%0\n" /* copy remaining size, subtract 1 */ - " bras %3,7f\n" /* memset loop */ - " xc 0(1,%2),0(%2)\n" - "6: xc 0(256,%2),0(%2)\n" - " la %2,256(%2)\n" - "7:"AHI" %4,-256\n" - " jnm 6b\n" - " ex %4,0(%3)\n" - " j 9f\n" - "8:"SLR" %0,%0\n" - "9: \n" - EX_TABLE(0b,3b) EX_TABLE(2b,3b) EX_TABLE(4b,5b) - EX_TABLE(10b,3b) EX_TABLE(11b,3b) EX_TABLE(12b,5b) - : "+a" (size), "+a" (ptr), "+a" (x), "+a" (tmp1), "=a" (tmp2) - : : "cc", "memory"); - return size; -} - -static size_t copy_from_user_std_check(size_t size, const void __user *ptr, - void *x) -{ - if (size <= 1024) - return copy_from_user_std(size, ptr, x); - return copy_from_user_pt(size, ptr, x); -} - -size_t copy_to_user_std(size_t size, void __user *ptr, const void *x) -{ - unsigned long tmp1, tmp2; - - tmp1 = -256UL; - asm volatile( - "0: mvcs 0(%0,%1),0(%2),%3\n" - "7: jz 5f\n" - "1:"ALR" %0,%3\n" - " la %1,256(%1)\n" - " la %2,256(%2)\n" - "2: mvcs 0(%0,%1),0(%2),%3\n" - "8: jnz 1b\n" - " j 5f\n" - "3: la %4,255(%1)\n" /* %4 = ptr + 255 */ - " "LHI" %3,-4096\n" - " nr %4,%3\n" /* %4 = (ptr + 255) & -4096 */ - " "SLR" %4,%1\n" - " "CLR" %0,%4\n" /* copy crosses next page boundary? */ - " jnh 6f\n" - "4: mvcs 0(%4,%1),0(%2),%3\n" - "9:"SLR" %0,%4\n" - " j 6f\n" - "5:"SLR" %0,%0\n" - "6: \n" - EX_TABLE(0b,3b) EX_TABLE(2b,3b) EX_TABLE(4b,6b) - EX_TABLE(7b,3b) EX_TABLE(8b,3b) EX_TABLE(9b,6b) - : "+a" (size), "+a" (ptr), "+a" (x), "+a" (tmp1), "=a" (tmp2) - : : "cc", "memory"); - return size; -} - -static size_t copy_to_user_std_check(size_t size, void __user *ptr, - const void *x) -{ - if (size <= 1024) - return copy_to_user_std(size, ptr, x); - return copy_to_user_pt(size, ptr, x); -} - -static size_t copy_in_user_std(size_t size, void __user *to, - const void __user *from) -{ - unsigned long tmp1; - - asm volatile( - " sacf 256\n" - " "AHI" %0,-1\n" - " jo 5f\n" - " bras %3,3f\n" - "0:"AHI" %0,257\n" - "1: mvc 0(1,%1),0(%2)\n" - " la %1,1(%1)\n" - " la %2,1(%2)\n" - " "AHI" %0,-1\n" - " jnz 1b\n" - " j 5f\n" - "2: mvc 0(256,%1),0(%2)\n" - " la %1,256(%1)\n" - " la %2,256(%2)\n" - "3:"AHI" %0,-256\n" - " jnm 2b\n" - "4: ex %0,1b-0b(%3)\n" - "5: "SLR" %0,%0\n" - "6: sacf 0\n" - EX_TABLE(1b,6b) EX_TABLE(2b,0b) EX_TABLE(4b,0b) - : "+a" (size), "+a" (to), "+a" (from), "=a" (tmp1) - : : "cc", "memory"); - return size; -} - -static size_t clear_user_std(size_t size, void __user *to) -{ - unsigned long tmp1, tmp2; - - asm volatile( - " sacf 256\n" - " "AHI" %0,-1\n" - " jo 5f\n" - " bras %3,3f\n" - " xc 0(1,%1),0(%1)\n" - "0:"AHI" %0,257\n" - " la %2,255(%1)\n" /* %2 = ptr + 255 */ - " srl %2,12\n" - " sll %2,12\n" /* %2 = (ptr + 255) & -4096 */ - " "SLR" %2,%1\n" - " "CLR" %0,%2\n" /* clear crosses next page boundary? */ - " jnh 5f\n" - " "AHI" %2,-1\n" - "1: ex %2,0(%3)\n" - " "AHI" %2,1\n" - " "SLR" %0,%2\n" - " j 5f\n" - "2: xc 0(256,%1),0(%1)\n" - " la %1,256(%1)\n" - "3:"AHI" %0,-256\n" - " jnm 2b\n" - "4: ex %0,0(%3)\n" - "5: "SLR" %0,%0\n" - "6: sacf 0\n" - EX_TABLE(1b,6b) EX_TABLE(2b,0b) EX_TABLE(4b,0b) - : "+a" (size), "+a" (to), "=a" (tmp1), "=a" (tmp2) - : : "cc", "memory"); - return size; -} - -size_t strnlen_user_std(size_t size, const char __user *src) -{ - register unsigned long reg0 asm("0") = 0UL; - unsigned long tmp1, tmp2; - - if (unlikely(!size)) - return 0; - asm volatile( - " la %2,0(%1)\n" - " la %3,0(%0,%1)\n" - " "SLR" %0,%0\n" - " sacf 256\n" - "0: srst %3,%2\n" - " jo 0b\n" - " la %0,1(%3)\n" /* strnlen_user results includes \0 */ - " "SLR" %0,%1\n" - "1: sacf 0\n" - EX_TABLE(0b,1b) - : "+a" (size), "+a" (src), "=a" (tmp1), "=a" (tmp2) - : "d" (reg0) : "cc", "memory"); - return size; -} - -size_t strncpy_from_user_std(size_t count, const char __user *src, char *dst) -{ - size_t done, len, offset, len_str; - - if (unlikely(!count)) - return 0; - done = 0; - do { - offset = (size_t)src & ~PAGE_MASK; - len = min(count - done, PAGE_SIZE - offset); - if (copy_from_user_std(len, src, dst)) - return -EFAULT; - len_str = strnlen(dst, len); - done += len_str; - src += len_str; - dst += len_str; - } while ((len_str == len) && (done < count)); - return done; -} - -#define __futex_atomic_op(insn, ret, oldval, newval, uaddr, oparg) \ - asm volatile( \ - " sacf 256\n" \ - "0: l %1,0(%6)\n" \ - "1:"insn \ - "2: cs %1,%2,0(%6)\n" \ - "3: jl 1b\n" \ - " lhi %0,0\n" \ - "4: sacf 0\n" \ - EX_TABLE(0b,4b) EX_TABLE(2b,4b) EX_TABLE(3b,4b) \ - : "=d" (ret), "=&d" (oldval), "=&d" (newval), \ - "=m" (*uaddr) \ - : "0" (-EFAULT), "d" (oparg), "a" (uaddr), \ - "m" (*uaddr) : "cc"); - -int futex_atomic_op_std(int op, u32 __user *uaddr, int oparg, int *old) -{ - int oldval = 0, newval, ret; - - switch (op) { - case FUTEX_OP_SET: - __futex_atomic_op("lr %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_ADD: - __futex_atomic_op("lr %2,%1\nar %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_OR: - __futex_atomic_op("lr %2,%1\nor %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_ANDN: - __futex_atomic_op("lr %2,%1\nnr %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - case FUTEX_OP_XOR: - __futex_atomic_op("lr %2,%1\nxr %2,%5\n", - ret, oldval, newval, uaddr, oparg); - break; - default: - ret = -ENOSYS; - } - *old = oldval; - return ret; -} - -int futex_atomic_cmpxchg_std(u32 *uval, u32 __user *uaddr, - u32 oldval, u32 newval) -{ - int ret; - - asm volatile( - " sacf 256\n" - "0: cs %1,%4,0(%5)\n" - "1: la %0,0\n" - "2: sacf 0\n" - EX_TABLE(0b,2b) EX_TABLE(1b,2b) - : "=d" (ret), "+d" (oldval), "=m" (*uaddr) - : "0" (-EFAULT), "d" (newval), "a" (uaddr), "m" (*uaddr) - : "cc", "memory" ); - *uval = oldval; - return ret; -} - -struct uaccess_ops uaccess_std = { - .copy_from_user = copy_from_user_std_check, - .copy_from_user_small = copy_from_user_std, - .copy_to_user = copy_to_user_std_check, - .copy_to_user_small = copy_to_user_std, - .copy_in_user = copy_in_user_std, - .clear_user = clear_user_std, - .strnlen_user = strnlen_user_std, - .strncpy_from_user = strncpy_from_user_std, - .futex_atomic_op = futex_atomic_op_std, - .futex_atomic_cmpxchg = futex_atomic_cmpxchg_std, -}; diff --git a/arch/s390/lib/ucmpdi2.c b/arch/s390/lib/ucmpdi2.c deleted file mode 100644 index 3e05ff532582..000000000000 --- a/arch/s390/lib/ucmpdi2.c +++ /dev/null @@ -1,26 +0,0 @@ -#include <linux/module.h> - -union ull_union { - unsigned long long ull; - struct { - unsigned int high; - unsigned int low; - } ui; -}; - -int __ucmpdi2(unsigned long long a, unsigned long long b) -{ - union ull_union au = {.ull = a}; - union ull_union bu = {.ull = b}; - - if (au.ui.high < bu.ui.high) - return 0; - else if (au.ui.high > bu.ui.high) - return 2; - if (au.ui.low < bu.ui.low) - return 0; - else if (au.ui.low > bu.ui.low) - return 2; - return 1; -} -EXPORT_SYMBOL(__ucmpdi2); diff --git a/arch/s390/lib/xor.c b/arch/s390/lib/xor.c new file mode 100644 index 000000000000..1721b73b7803 --- /dev/null +++ b/arch/s390/lib/xor.c @@ -0,0 +1,137 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Optimized xor_block operation for RAID4/5 + * + * Copyright IBM Corp. 2016 + * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com> + */ + +#include <linux/types.h> +#include <linux/export.h> +#include <linux/raid/xor.h> +#include <asm/xor.h> + +static void xor_xc_2(unsigned long bytes, unsigned long * __restrict p1, + const unsigned long * __restrict p2) +{ + asm volatile( + " aghi %0,-1\n" + " jm 3f\n" + " srlg 0,%0,8\n" + " ltgr 0,0\n" + " jz 1f\n" + "0: xc 0(256,%1),0(%2)\n" + " la %1,256(%1)\n" + " la %2,256(%2)\n" + " brctg 0,0b\n" + "1: exrl %0,2f\n" + " j 3f\n" + "2: xc 0(1,%1),0(%2)\n" + "3:" + : : "d" (bytes), "a" (p1), "a" (p2) + : "0", "cc", "memory"); +} + +static void xor_xc_3(unsigned long bytes, unsigned long * __restrict p1, + const unsigned long * __restrict p2, + const unsigned long * __restrict p3) +{ + asm volatile( + " aghi %0,-1\n" + " jm 4f\n" + " srlg 0,%0,8\n" + " ltgr 0,0\n" + " jz 1f\n" + "0: xc 0(256,%1),0(%2)\n" + " xc 0(256,%1),0(%3)\n" + " la %1,256(%1)\n" + " la %2,256(%2)\n" + " la %3,256(%3)\n" + " brctg 0,0b\n" + "1: exrl %0,2f\n" + " exrl %0,3f\n" + " j 4f\n" + "2: xc 0(1,%1),0(%2)\n" + "3: xc 0(1,%1),0(%3)\n" + "4:" + : "+d" (bytes), "+a" (p1), "+a" (p2), "+a" (p3) + : : "0", "cc", "memory"); +} + +static void xor_xc_4(unsigned long bytes, unsigned long * __restrict p1, + const unsigned long * __restrict p2, + const unsigned long * __restrict p3, + const unsigned long * __restrict p4) +{ + asm volatile( + " aghi %0,-1\n" + " jm 5f\n" + " srlg 0,%0,8\n" + " ltgr 0,0\n" + " jz 1f\n" + "0: xc 0(256,%1),0(%2)\n" + " xc 0(256,%1),0(%3)\n" + " xc 0(256,%1),0(%4)\n" + " la %1,256(%1)\n" + " la %2,256(%2)\n" + " la %3,256(%3)\n" + " la %4,256(%4)\n" + " brctg 0,0b\n" + "1: exrl %0,2f\n" + " exrl %0,3f\n" + " exrl %0,4f\n" + " j 5f\n" + "2: xc 0(1,%1),0(%2)\n" + "3: xc 0(1,%1),0(%3)\n" + "4: xc 0(1,%1),0(%4)\n" + "5:" + : "+d" (bytes), "+a" (p1), "+a" (p2), "+a" (p3), "+a" (p4) + : : "0", "cc", "memory"); +} + +static void xor_xc_5(unsigned long bytes, unsigned long * __restrict p1, + const unsigned long * __restrict p2, + const unsigned long * __restrict p3, + const unsigned long * __restrict p4, + const unsigned long * __restrict p5) +{ + asm volatile( + " larl 1,2f\n" + " aghi %0,-1\n" + " jm 6f\n" + " srlg 0,%0,8\n" + " ltgr 0,0\n" + " jz 1f\n" + "0: xc 0(256,%1),0(%2)\n" + " xc 0(256,%1),0(%3)\n" + " xc 0(256,%1),0(%4)\n" + " xc 0(256,%1),0(%5)\n" + " la %1,256(%1)\n" + " la %2,256(%2)\n" + " la %3,256(%3)\n" + " la %4,256(%4)\n" + " la %5,256(%5)\n" + " brctg 0,0b\n" + "1: exrl %0,2f\n" + " exrl %0,3f\n" + " exrl %0,4f\n" + " exrl %0,5f\n" + " j 6f\n" + "2: xc 0(1,%1),0(%2)\n" + "3: xc 0(1,%1),0(%3)\n" + "4: xc 0(1,%1),0(%4)\n" + "5: xc 0(1,%1),0(%5)\n" + "6:" + : "+d" (bytes), "+a" (p1), "+a" (p2), "+a" (p3), "+a" (p4), + "+a" (p5) + : : "0", "cc", "memory"); +} + +struct xor_block_template xor_block_xc = { + .name = "xc", + .do_2 = xor_xc_2, + .do_3 = xor_xc_3, + .do_4 = xor_xc_4, + .do_5 = xor_xc_5, +}; +EXPORT_SYMBOL(xor_block_xc); |