diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2017-11-17 20:21:44 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2017-11-17 20:21:44 -0800 |
commit | 1deab8ce2c91e3b16563b7a7ea150f82334262ec (patch) | |
tree | 954b2d9d27df9765b9579b10f87054e1effd6cb3 | |
parent | 817002475046737877915d87889a012d851650fa (diff) | |
parent | 70f3c8b7c2e7ebcdde8354da004872e7c9184e97 (diff) | |
download | linux-1deab8ce2c91e3b16563b7a7ea150f82334262ec.tar.gz linux-1deab8ce2c91e3b16563b7a7ea150f82334262ec.tar.bz2 linux-1deab8ce2c91e3b16563b7a7ea150f82334262ec.zip |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc
Pull sparc updates from David Miller:
1) Add missing cmpxchg64() for 32-bit sparc.
2) Timer conversions from Allen Pais and Kees Cook.
3) vDSO support, from Nagarathnam Muthusamy.
4) Fix sparc64 huge page table walks based upon bug report by Al Viro,
from Nitin Gupta.
5) Optimized fls() for T4 and above, from Vijay Kumar.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
sparc64: Fix page table walk for PUD hugepages
sparc64: Convert timers to user timer_setup()
sparc64: convert mdesc_handle.refcnt from atomic_t to refcount_t
sparc/led: Convert timers to use timer_setup()
sparc64: Use sparc optimized fls and __fls for T4 and above
sparc64: SPARC optimized __fls function
sparc64: SPARC optimized fls function
sparc64: Define SPARC default __fls function
sparc64: Define SPARC default fls function
vDSO for sparc
sparc32: Add cmpxchg64().
sbus: char: Move D7S_MINOR to include/linux/miscdevice.h
sparc: time: Remove unneeded linux/miscdevice.h include
sparc64: mmu_context: Add missing include files
41 files changed, 1702 insertions, 18 deletions
diff --git a/arch/sparc/Kbuild b/arch/sparc/Kbuild index 675afa285ddb..b4d0f570cc00 100644 --- a/arch/sparc/Kbuild +++ b/arch/sparc/Kbuild @@ -7,3 +7,4 @@ obj-y += mm/ obj-y += math-emu/ obj-y += net/ obj-y += crypto/ +obj-$(CONFIG_SPARC64) += vdso/ diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig index 987a57502909..6bf594ace663 100644 --- a/arch/sparc/Kconfig +++ b/arch/sparc/Kconfig @@ -84,6 +84,8 @@ config SPARC64 select HAVE_REGS_AND_STACK_ACCESS_API select ARCH_USE_QUEUED_RWLOCKS select ARCH_USE_QUEUED_SPINLOCKS + select GENERIC_TIME_VSYSCALL + select ARCH_CLOCKSOURCE_DATA config ARCH_DEFCONFIG string diff --git a/arch/sparc/Makefile b/arch/sparc/Makefile index dbc448923f48..edac927e4952 100644 --- a/arch/sparc/Makefile +++ b/arch/sparc/Makefile @@ -81,6 +81,10 @@ install: archclean: $(Q)$(MAKE) $(clean)=$(boot) +PHONY += vdso_install +vdso_install: + $(Q)$(MAKE) $(build)=arch/sparc/vdso $@ + # This is the image used for packaging KBUILD_IMAGE := $(boot)/zImage diff --git a/arch/sparc/include/asm/bitops_64.h b/arch/sparc/include/asm/bitops_64.h index a90eea24b286..ca7ea5913494 100644 --- a/arch/sparc/include/asm/bitops_64.h +++ b/arch/sparc/include/asm/bitops_64.h @@ -23,10 +23,11 @@ void set_bit(unsigned long nr, volatile unsigned long *addr); void clear_bit(unsigned long nr, volatile unsigned long *addr); void change_bit(unsigned long nr, volatile unsigned long *addr); +int fls(unsigned int word); +int __fls(unsigned long word); + #include <asm-generic/bitops/non-atomic.h> -#include <asm-generic/bitops/fls.h> -#include <asm-generic/bitops/__fls.h> #include <asm-generic/bitops/fls64.h> #ifdef __KERNEL__ diff --git a/arch/sparc/include/asm/clocksource.h b/arch/sparc/include/asm/clocksource.h new file mode 100644 index 000000000000..d63ef224befe --- /dev/null +++ b/arch/sparc/include/asm/clocksource.h @@ -0,0 +1,17 @@ +/* + * Copyright (c) 2017 Oracle and/or its affiliates. All rights reserved. + */ + +#ifndef _ASM_SPARC_CLOCKSOURCE_H +#define _ASM_SPARC_CLOCKSOURCE_H + +/* VDSO clocksources */ +#define VCLOCK_NONE 0 /* Nothing userspace can do. */ +#define VCLOCK_TICK 1 /* Use %tick. */ +#define VCLOCK_STICK 2 /* Use %stick. */ + +struct arch_clocksource_data { + int vclock_mode; +}; + +#endif /* _ASM_SPARC_CLOCKSOURCE_H */ diff --git a/arch/sparc/include/asm/cmpxchg_32.h b/arch/sparc/include/asm/cmpxchg_32.h index 3e3823db303e..c73b5a3ab7b9 100644 --- a/arch/sparc/include/asm/cmpxchg_32.h +++ b/arch/sparc/include/asm/cmpxchg_32.h @@ -63,6 +63,9 @@ __cmpxchg(volatile void *ptr, unsigned long old, unsigned long new_, int size) (unsigned long)_n_, sizeof(*(ptr))); \ }) +u64 __cmpxchg_u64(u64 *ptr, u64 old, u64 new); +#define cmpxchg64(ptr, old, new) __cmpxchg_u64(ptr, old, new) + #include <asm-generic/cmpxchg-local.h> /* diff --git a/arch/sparc/include/asm/elf_64.h b/arch/sparc/include/asm/elf_64.h index 5894389f5ed5..25340df3570c 100644 --- a/arch/sparc/include/asm/elf_64.h +++ b/arch/sparc/include/asm/elf_64.h @@ -211,4 +211,18 @@ do { if ((ex).e_ident[EI_CLASS] == ELFCLASS32) \ (current->personality & (~PER_MASK))); \ } while (0) +extern unsigned int vdso_enabled; + +#define ARCH_DLINFO \ +do { \ + if (vdso_enabled) \ + NEW_AUX_ENT(AT_SYSINFO_EHDR, \ + (unsigned long)current->mm->context.vdso); \ +} while (0) + +struct linux_binprm; + +#define ARCH_HAS_SETUP_ADDITIONAL_PAGES 1 +extern int arch_setup_additional_pages(struct linux_binprm *bprm, + int uses_interp); #endif /* !(__ASM_SPARC64_ELF_H) */ diff --git a/arch/sparc/include/asm/mmu_64.h b/arch/sparc/include/asm/mmu_64.h index 5fe64a57b4ba..ad4fb93508ba 100644 --- a/arch/sparc/include/asm/mmu_64.h +++ b/arch/sparc/include/asm/mmu_64.h @@ -97,6 +97,7 @@ typedef struct { unsigned long thp_pte_count; struct tsb_config tsb_block[MM_NUM_TSBS]; struct hv_tsb_descr tsb_descr[MM_NUM_TSBS]; + void *vdso; } mm_context_t; #endif /* !__ASSEMBLY__ */ diff --git a/arch/sparc/include/asm/mmu_context_64.h b/arch/sparc/include/asm/mmu_context_64.h index e25d25b0a34b..b361702ef52a 100644 --- a/arch/sparc/include/asm/mmu_context_64.h +++ b/arch/sparc/include/asm/mmu_context_64.h @@ -8,9 +8,11 @@ #include <linux/spinlock.h> #include <linux/mm_types.h> +#include <linux/smp.h> #include <asm/spitfire.h> #include <asm-generic/mm_hooks.h> +#include <asm/percpu.h> static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk) { diff --git a/arch/sparc/include/asm/processor_64.h b/arch/sparc/include/asm/processor_64.h index c7c79fe8d265..aac23d4a4ddd 100644 --- a/arch/sparc/include/asm/processor_64.h +++ b/arch/sparc/include/asm/processor_64.h @@ -200,6 +200,13 @@ unsigned long get_wchan(struct task_struct *task); * To make a long story short, we are trying to yield the current cpu * strand during busy loops. */ +#ifdef BUILD_VDSO +#define cpu_relax() asm volatile("\n99:\n\t" \ + "rd %%ccr, %%g0\n\t" \ + "rd %%ccr, %%g0\n\t" \ + "rd %%ccr, %%g0\n\t" \ + ::: "memory") +#else /* ! BUILD_VDSO */ #define cpu_relax() asm volatile("\n99:\n\t" \ "rd %%ccr, %%g0\n\t" \ "rd %%ccr, %%g0\n\t" \ @@ -211,6 +218,7 @@ unsigned long get_wchan(struct task_struct *task); "nop\n\t" \ ".previous" \ ::: "memory") +#endif /* Prefetch support. This is tuned for UltraSPARC-III and later. * UltraSPARC-I will treat these as nops, and UltraSPARC-II has diff --git a/arch/sparc/include/asm/tsb.h b/arch/sparc/include/asm/tsb.h index 25b6abdb3908..522a677e050d 100644 --- a/arch/sparc/include/asm/tsb.h +++ b/arch/sparc/include/asm/tsb.h @@ -217,7 +217,7 @@ extern struct tsb_phys_patch_entry __tsb_phys_patch, __tsb_phys_patch_end; sllx REG2, 32, REG2; \ andcc REG1, REG2, %g0; \ be,pt %xcc, 700f; \ - sethi %hi(0x1ffc0000), REG2; \ + sethi %hi(0xffe00000), REG2; \ sllx REG2, 1, REG2; \ brgez,pn REG1, FAIL_LABEL; \ andn REG1, REG2, REG1; \ diff --git a/arch/sparc/include/asm/vdso.h b/arch/sparc/include/asm/vdso.h new file mode 100644 index 000000000000..93b628731a5e --- /dev/null +++ b/arch/sparc/include/asm/vdso.h @@ -0,0 +1,24 @@ +/* + * Copyright (c) 2017 Oracle and/or its affiliates. All rights reserved. + */ + +#ifndef _ASM_SPARC_VDSO_H +#define _ASM_SPARC_VDSO_H + +struct vdso_image { + void *data; + unsigned long size; /* Always a multiple of PAGE_SIZE */ + long sym_vvar_start; /* Negative offset to the vvar area */ + long sym_vread_tick; /* Start of vread_tick section */ + long sym_vread_tick_patch_start; /* Start of tick read */ + long sym_vread_tick_patch_end; /* End of tick read */ +}; + +#ifdef CONFIG_SPARC64 +extern const struct vdso_image vdso_image_64_builtin; +#endif +#ifdef CONFIG_COMPAT +extern const struct vdso_image vdso_image_32_builtin; +#endif + +#endif /* _ASM_SPARC_VDSO_H */ diff --git a/arch/sparc/include/asm/vvar.h b/arch/sparc/include/asm/vvar.h new file mode 100644 index 000000000000..0289503d1cb0 --- /dev/null +++ b/arch/sparc/include/asm/vvar.h @@ -0,0 +1,74 @@ +/* + * Copyright (c) 2017 Oracle and/or its affiliates. All rights reserved. + */ + +#ifndef _ASM_SPARC_VVAR_DATA_H +#define _ASM_SPARC_VVAR_DATA_H + +#include <asm/clocksource.h> +#include <linux/seqlock.h> +#include <linux/time.h> +#include <linux/types.h> + +struct vvar_data { + unsigned int seq; + + int vclock_mode; + struct { /* extract of a clocksource struct */ + u64 cycle_last; + u64 mask; + int mult; + int shift; + } clock; + /* open coded 'struct timespec' */ + u64 wall_time_sec; + u64 wall_time_snsec; + u64 monotonic_time_snsec; + u64 monotonic_time_sec; + u64 monotonic_time_coarse_sec; + u64 monotonic_time_coarse_nsec; + u64 wall_time_coarse_sec; + u64 wall_time_coarse_nsec; + + int tz_minuteswest; + int tz_dsttime; +}; + +extern struct vvar_data *vvar_data; +extern int vdso_fix_stick; + +static inline unsigned int vvar_read_begin(const struct vvar_data *s) +{ + unsigned int ret; + +repeat: + ret = READ_ONCE(s->seq); + if (unlikely(ret & 1)) { + cpu_relax(); + goto repeat; + } + smp_rmb(); /* Finish all reads before we return seq */ + return ret; +} + +static inline int vvar_read_retry(const struct vvar_data *s, + unsigned int start) +{ + smp_rmb(); /* Finish all reads before checking the value of seq */ + return unlikely(s->seq != start); +} + +static inline void vvar_write_begin(struct vvar_data *s) +{ + ++s->seq; + smp_wmb(); /* Makes sure that increment of seq is reflected */ +} + +static inline void vvar_write_end(struct vvar_data *s) +{ + smp_wmb(); /* Makes the value of seq current before we increment */ + ++s->seq; +} + + +#endif /* _ASM_SPARC_VVAR_DATA_H */ diff --git a/arch/sparc/include/uapi/asm/auxvec.h b/arch/sparc/include/uapi/asm/auxvec.h index ad6f360261f6..5f80a70cc901 100644 --- a/arch/sparc/include/uapi/asm/auxvec.h +++ b/arch/sparc/include/uapi/asm/auxvec.h @@ -1,4 +1,8 @@ #ifndef __ASMSPARC_AUXVEC_H #define __ASMSPARC_AUXVEC_H +#define AT_SYSINFO_EHDR 33 + +#define AT_VECTOR_SIZE_ARCH 1 + #endif /* !(__ASMSPARC_AUXVEC_H) */ diff --git a/arch/sparc/kernel/Makefile b/arch/sparc/kernel/Makefile index 8de9617589a5..cc97545737f0 100644 --- a/arch/sparc/kernel/Makefile +++ b/arch/sparc/kernel/Makefile @@ -43,6 +43,7 @@ obj-$(CONFIG_SPARC32) += systbls_32.o obj-y += time_$(BITS).o obj-$(CONFIG_SPARC32) += windows.o obj-y += cpu.o +obj-$(CONFIG_SPARC64) += vdso.o obj-$(CONFIG_SPARC32) += devices.o obj-y += ptrace_$(BITS).o obj-y += unaligned_$(BITS).o diff --git a/arch/sparc/kernel/head_64.S b/arch/sparc/kernel/head_64.S index 9e293de12052..a41e6e16eb36 100644 --- a/arch/sparc/kernel/head_64.S +++ b/arch/sparc/kernel/head_64.S @@ -641,6 +641,8 @@ niagara4_patch: nop call niagara4_patch_pageops nop + call niagara4_patch_fls + nop ba,a,pt %xcc, 80f nop diff --git a/arch/sparc/kernel/mdesc.c b/arch/sparc/kernel/mdesc.c index 1ef6156b1530..418592a09b41 100644 --- a/arch/sparc/kernel/mdesc.c +++ b/arch/sparc/kernel/mdesc.c @@ -13,6 +13,7 @@ #include <linux/miscdevice.h> #include <linux/bootmem.h> #include <linux/export.h> +#include <linux/refcount.h> #include <asm/cpudata.h> #include <asm/hypervisor.h> @@ -71,7 +72,7 @@ struct mdesc_handle { struct list_head list; struct mdesc_mem_ops *mops; void *self_base; - atomic_t refcnt; + refcount_t refcnt; unsigned int handle_size; struct mdesc_hdr mdesc; }; @@ -153,7 +154,7 @@ static void mdesc_handle_init(struct mdesc_handle *hp, memset(hp, 0, handle_size); INIT_LIST_HEAD(&hp->list); hp->self_base = base; - atomic_set(&hp->refcnt, 1); + refcount_set(&hp->refcnt, 1); hp->handle_size = handle_size; } @@ -183,7 +184,7 @@ static void __init mdesc_memblock_free(struct mdesc_handle *hp) unsigned int alloc_size; unsigned long start; - BUG_ON(atomic_read(&hp->refcnt) != 0); + BUG_ON(refcount_read(&hp->refcnt) != 0); BUG_ON(!list_empty(&hp->list)); alloc_size = PAGE_ALIGN(hp->handle_size); @@ -221,7 +222,7 @@ static struct mdesc_handle *mdesc_kmalloc(unsigned int mdesc_size) static void mdesc_kfree(struct mdesc_handle *hp) { - BUG_ON(atomic_read(&hp->refcnt) != 0); + BUG_ON(refcount_read(&hp->refcnt) != 0); BUG_ON(!list_empty(&hp->list)); kfree(hp->self_base); @@ -260,7 +261,7 @@ struct mdesc_handle *mdesc_grab(void) spin_lock_irqsave(&mdesc_lock, flags); hp = cur_mdesc; if (hp) - atomic_inc(&hp->refcnt); + refcount_inc(&hp->refcnt); spin_unlock_irqrestore(&mdesc_lock, flags); return hp; @@ -272,7 +273,7 @@ void mdesc_release(struct mdesc_handle *hp) unsigned long flags; spin_lock_irqsave(&mdesc_lock, flags); - if (atomic_dec_and_test(&hp->refcnt)) { + if (refcount_dec_and_test(&hp->refcnt)) { list_del_init(&hp->list); hp->mops->free(hp); } @@ -514,7 +515,7 @@ void mdesc_update(void) if (status != HV_EOK || real_len > len) { printk(KERN_ERR "MD: mdesc reread fails with %lu\n", status); - atomic_dec(&hp->refcnt); + refcount_dec(&hp->refcnt); mdesc_free(hp); goto out; } @@ -527,7 +528,7 @@ void mdesc_update(void) mdesc_notify_clients(orig_hp, hp); spin_lock_irqsave(&mdesc_lock, flags); - if (atomic_dec_and_test(&orig_hp->refcnt)) + if (refcount_dec_and_test(&orig_hp->refcnt)) mdesc_free(orig_hp); else list_add(&orig_hp->list, &mdesc_zombie_list); diff --git a/arch/sparc/kernel/time_64.c b/arch/sparc/kernel/time_64.c index 3b397081047a..2ef8cfa9677e 100644 --- a/arch/sparc/kernel/time_64.c +++ b/arch/sparc/kernel/time_64.c @@ -28,7 +28,6 @@ #include <linux/jiffies.h> #include <linux/cpufreq.h> #include <linux/percpu.h> -#include <linux/miscdevice.h> #include <linux/rtc/m48t59.h> #include <linux/kernel_stat.h> #include <linux/clockchips.h> @@ -54,6 +53,8 @@ DEFINE_SPINLOCK(rtc_lock); +unsigned int __read_mostly vdso_fix_stick; + #ifdef CONFIG_SMP unsigned long profile_pc(struct pt_regs *regs) { @@ -831,12 +832,17 @@ static void init_tick_ops(struct sparc64_tick_ops *ops) void __init time_init_early(void) { if (tlb_type == spitfire) { - if (is_hummingbird()) + if (is_hummingbird()) { init_tick_ops(&hbtick_operations); - else + clocksource_tick.archdata.vclock_mode = VCLOCK_NONE; + } else { init_tick_ops(&tick_operations); + clocksource_tick.archdata.vclock_mode = VCLOCK_TICK; + vdso_fix_stick = 1; + } } else { init_tick_ops(&stick_operations); + clocksource_tick.archdata.vclock_mode = VCLOCK_STICK; } } diff --git a/arch/sparc/kernel/vdso.c b/arch/sparc/kernel/vdso.c new file mode 100644 index 000000000000..58880662b271 --- /dev/null +++ b/arch/sparc/kernel/vdso.c @@ -0,0 +1,70 @@ +/* + * Copyright (C) 2001 Andrea Arcangeli <andrea@suse.de> SuSE + * Copyright 2003 Andi Kleen, SuSE Labs. + * + * Thanks to hpa@transmeta.com for some useful hint. + * Special thanks to Ingo Molnar for his early experience with + * a different vsyscall implementation for Linux/IA32 and for the name. + */ + +#include <linux/seqlock.h> +#include <linux/time.h> +#include <linux/timekeeper_internal.h> + +#include <asm/vvar.h> + +void update_vsyscall_tz(void) +{ + if (unlikely(vvar_data == NULL)) + return; + + vvar_data->tz_minuteswest = sys_tz.tz_minuteswest; + vvar_data->tz_dsttime = sys_tz.tz_dsttime; +} + +void update_vsyscall(struct timekeeper *tk) +{ + struct vvar_data *vdata = vvar_data; + + if (unlikely(vdata == NULL)) + return; + + vvar_write_begin(vdata); + vdata->vclock_mode = tk->tkr_mono.clock->archdata.vclock_mode; + vdata->clock.cycle_last = tk->tkr_mono.cycle_last; + vdata->clock.mask = tk->tkr_mono.mask; + vdata->clock.mult = tk->tkr_mono.mult; + vdata->clock.shift = tk->tkr_mono.shift; + + vdata->wall_time_sec = tk->xtime_sec; + vdata->wall_time_snsec = tk->tkr_mono.xtime_nsec; + + vdata->monotonic_time_sec = tk->xtime_sec + + tk->wall_to_monotonic.tv_sec; + vdata->monotonic_time_snsec = tk->tkr_mono.xtime_nsec + + (tk->wall_to_monotonic.tv_nsec << + tk->tkr_mono.shift); + + while (vdata->monotonic_time_snsec >= + (((u64)NSEC_PER_SEC) << tk->tkr_mono.shift)) { + vdata->monotonic_time_snsec -= + ((u64)NSEC_PER_SEC) << tk->tkr_mono.shift; + vdata->monotonic_time_sec++; + } + + vdata->wall_time_coarse_sec = tk->xtime_sec; + vdata->wall_time_coarse_nsec = + (long)(tk->tkr_mono.xtime_nsec >> tk->tkr_mono.shift); + + vdata->monotonic_time_coarse_sec = + vdata->wall_time_coarse_sec + tk->wall_to_monotonic.tv_sec; + vdata->monotonic_time_coarse_nsec = + vdata->wall_time_coarse_nsec + tk->wall_to_monotonic.tv_nsec; + + while (vdata->monotonic_time_coarse_nsec >= NSEC_PER_SEC) { + vdata->monotonic_time_coarse_nsec -= NSEC_PER_SEC; + vdata->monotonic_time_coarse_sec++; + } + + vvar_write_end(vdata); +} diff --git a/arch/sparc/kernel/viohs.c b/arch/sparc/kernel/viohs.c index c858f5f3ce2c..635d67ffc9a3 100644 --- a/arch/sparc/kernel/viohs.c +++ b/arch/sparc/kernel/viohs.c @@ -798,9 +798,9 @@ void vio_port_up(struct vio_driver_state *vio) } EXPORT_SYMBOL(vio_port_up); -static void vio_port_timer(unsigned long _arg) +static void vio_port_timer(struct timer_list *t) { - struct vio_driver_state *vio = (struct vio_driver_state *) _arg; + struct vio_driver_state *vio = from_timer(vio, t, timer); vio_port_up(vio); } @@ -849,7 +849,7 @@ int vio_driver_init(struct vio_driver_state *vio, struct vio_dev *vdev, vio->ops = ops; - setup_timer(&vio->timer, vio_port_timer, (unsigned long) vio); + timer_setup(&vio->timer, vio_port_timer, 0); return 0; } diff --git a/arch/sparc/lib/Makefile b/arch/sparc/lib/Makefile index 44829a8dc458..0f0f76b4f6cd 100644 --- a/arch/sparc/lib/Makefile +++ b/arch/sparc/lib/Makefile @@ -17,6 +17,9 @@ lib-$(CONFIG_SPARC64) += atomic_64.o lib-$(CONFIG_SPARC32) += lshrdi3.o ashldi3.o lib-$(CONFIG_SPARC32) += muldi3.o bitext.o cmpdi2.o lib-$(CONFIG_SPARC64) += multi3.o +lib-$(CONFIG_SPARC64) += fls.o +lib-$(CONFIG_SPARC64) += fls64.o +obj-$(CONFIG_SPARC64) += NG4fls.o lib-$(CONFIG_SPARC64) += copy_page.o clear_page.o bzero.o lib-$(CONFIG_SPARC64) += csum_copy.o csum_copy_from_user.o csum_copy_to_user.o diff --git a/arch/sparc/lib/NG4fls.S b/arch/sparc/lib/NG4fls.S new file mode 100644 index 000000000000..2d0991e5b034 --- /dev/null +++ b/arch/sparc/lib/NG4fls.S @@ -0,0 +1,30 @@ +/* NG4fls.S: SPARC optimized fls and __fls for T4 and above. + * + * Copyright (c) 2017, Oracle and/or its affiliates. All rights reserved. + */ + +#include <linux/linkage.h> + +#define LZCNT_O0_G2 \ + .word 0x85b002e8 + + .text + .register %g2, #scratch + .register %g3, #scratch + +ENTRY(NG4fls) + LZCNT_O0_G2 !lzcnt %o0, %g2 + mov 64, %g3 + retl + sub %g3, %g2, %o0 +ENDPROC(NG4fls) + +ENTRY(__NG4fls) + brz,pn %o0, 1f + LZCNT_O0_G2 !lzcnt %o0, %g2 + mov 63, %g3 + sub %g3, %g2, %o0 +1: + retl + nop +ENDPROC(__NG4fls) diff --git a/arch/sparc/lib/NG4patch.S b/arch/sparc/lib/NG4patch.S index aa58ab39f9a6..37866175c921 100644 --- a/arch/sparc/lib/NG4patch.S +++ b/arch/sparc/lib/NG4patch.S @@ -4,6 +4,8 @@ * Copyright (C) 2012 David S. Miller <davem@davemloft.net> */ +#include <linux/linkage.h> + #define BRANCH_ALWAYS 0x10680000 #define NOP 0x01000000 #define NG_DO_PATCH(OLD, NEW) \ @@ -53,3 +55,10 @@ niagara4_patch_pageops: retl nop .size niagara4_patch_pageops,.-niagara4_patch_pageops + +ENTRY(niagara4_patch_fls) + NG_DO_PATCH(fls, NG4fls) + NG_DO_PATCH(__fls, __NG4fls) + retl + nop +ENDPROC(niagara4_patch_fls) diff --git a/arch/sparc/lib/atomic32.c b/arch/sparc/lib/atomic32.c index 5010df497387..465a901a0ada 100644 --- a/arch/sparc/lib/atomic32.c +++ b/arch/sparc/lib/atomic32.c @@ -173,6 +173,20 @@ unsigned long __cmpxchg_u32(volatile u32 *ptr, u32 old, u32 new) } EXPORT_SYMBOL(__cmpxchg_u32); +u64 __cmpxchg_u64(u64 *ptr, u64 old, u64 new) +{ + unsigned long flags; + u64 prev; + + spin_lock_irqsave(ATOMIC_HASH(ptr), flags); + if ((prev = *ptr) == old) + *ptr = new; + spin_unlock_irqrestore(ATOMIC_HASH(ptr), flags); + + return prev; +} +EXPORT_SYMBOL(__cmpxchg_u64); + unsigned long __xchg_u32(volatile u32 *ptr, u32 new) { unsigned long flags; diff --git a/arch/sparc/lib/fls.S b/arch/sparc/lib/fls.S new file mode 100644 index 000000000000..06b8d300bcae --- /dev/null +++ b/arch/sparc/lib/fls.S @@ -0,0 +1,67 @@ +/* fls.S: SPARC default fls definition. + * + * SPARC default fls definition, which follows the same algorithm as + * in generic fls(). This function will be boot time patched on T4 + * and onward. + */ + +#include <linux/linkage.h> +#include <asm/export.h> + + .text + .register %g2, #scratch + .register %g3, #scratch +ENTRY(fls) + brz,pn %o0, 6f + mov 0, %o1 + sethi %hi(0xffff0000), %g3 + mov %o0, %g2 + andcc %o0, %g3, %g0 + be,pt %icc, 8f + mov 32, %o1 + sethi %hi(0xff000000), %g3 + andcc %g2, %g3, %g0 + bne,pt %icc, 3f + sethi %hi(0xf0000000), %g3 + sll %o0, 8, %o0 +1: + add %o1, -8, %o1 + sra %o0, 0, %o0 + mov %o0, %g2 +2: + sethi %hi(0xf0000000), %g3 +3: + andcc %g2, %g3, %g0 + bne,pt %icc, 4f + sethi %hi(0xc0000000), %g3 + sll %o0, 4, %o0 + add %o1, -4, %o1 + sra %o0, 0, %o0 + mov %o0, %g2 +4: + andcc %g2, %g3, %g0 + be,a,pt %icc, 7f + sll %o0, 2, %o0 +5: + xnor %g0, %o0, %o0 + srl %o0, 31, %o0 + sub %o1, %o0, %o1 +6: + jmp %o7 + 8 + sra %o1, 0, %o0 +7: + add %o1, -2, %o1 + ba,pt %xcc, 5b + sra %o0, 0, %o0 +8: + sll %o0, 16, %o0 + sethi %hi(0xff000000), %g3 + sra %o0, 0, %o0 + mov %o0, %g2 + andcc |