diff options
Diffstat (limited to 'arch/ia64')
39 files changed, 299 insertions, 284 deletions
diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index e23af4b6ae8c..6e41471449c0 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -468,7 +468,7 @@ config KEXEC help kexec is a system call that implements the ability to shutdown your current kernel, and to start another kernel. It is like a reboot - but it is indepedent of the system firmware. And like a reboot + but it is independent of the system firmware. And like a reboot you can start any kernel with it, not just Linux. The name comes from the similiarity to the exec system call. diff --git a/arch/ia64/hp/sim/boot/fw-emu.c b/arch/ia64/hp/sim/boot/fw-emu.c index 5a0a7afcfc3a..300acd913d9c 100644 --- a/arch/ia64/hp/sim/boot/fw-emu.c +++ b/arch/ia64/hp/sim/boot/fw-emu.c @@ -287,7 +287,7 @@ sys_fw_init (const char *args, int arglen) memset(efi_systab, 0, sizeof(efi_systab)); efi_systab->hdr.signature = EFI_SYSTEM_TABLE_SIGNATURE; - efi_systab->hdr.revision = EFI_SYSTEM_TABLE_REVISION; + efi_systab->hdr.revision = ((1 << 16) | 00); efi_systab->hdr.headersize = sizeof(efi_systab->hdr); efi_systab->fw_vendor = __pa("H\0e\0w\0l\0e\0t\0t\0-\0P\0a\0c\0k\0a\0r\0d\0\0"); efi_systab->fw_revision = 1; diff --git a/arch/ia64/ia32/ia32_entry.S b/arch/ia64/ia32/ia32_entry.S index 687e5fdc9683..99b665e2b1d5 100644 --- a/arch/ia64/ia32/ia32_entry.S +++ b/arch/ia64/ia32/ia32_entry.S @@ -52,43 +52,6 @@ ENTRY(ia32_clone) br.ret.sptk.many rp END(ia32_clone) -ENTRY(sys32_rt_sigsuspend) - .prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8) - alloc loc1=ar.pfs,8,2,3,0 // preserve all eight input regs - mov loc0=rp - mov out0=in0 // mask - mov out1=in1 // sigsetsize - mov out2=sp // out2 = &sigscratch - .fframe 16 - adds sp=-16,sp // allocate dummy "sigscratch" - ;; - .body - br.call.sptk.many rp=ia32_rt_sigsuspend -1: .restore sp - adds sp=16,sp - mov rp=loc0 - mov ar.pfs=loc1 - br.ret.sptk.many rp -END(sys32_rt_sigsuspend) - -ENTRY(sys32_sigsuspend) - .prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8) - alloc loc1=ar.pfs,8,2,3,0 // preserve all eight input regs - mov loc0=rp - mov out0=in2 // mask (first two args are ignored) - ;; - mov out1=sp // out1 = &sigscratch - .fframe 16 - adds sp=-16,sp // allocate dummy "sigscratch" - .body - br.call.sptk.many rp=ia32_sigsuspend -1: .restore sp - adds sp=16,sp - mov rp=loc0 - mov ar.pfs=loc1 - br.ret.sptk.many rp -END(sys32_sigsuspend) - GLOBAL_ENTRY(ia32_ret_from_clone) PT_REGS_UNWIND_INFO(0) { /* @@ -389,7 +352,7 @@ ia32_syscall_table: data8 sys_rt_sigpending data8 compat_sys_rt_sigtimedwait data8 sys32_rt_sigqueueinfo - data8 sys32_rt_sigsuspend + data8 compat_sys_rt_sigsuspend data8 sys32_pread /* 180 */ data8 sys32_pwrite data8 sys_chown /* 16-bit version */ diff --git a/arch/ia64/ia32/ia32_ldt.c b/arch/ia64/ia32/ia32_ldt.c index a152738c7d0d..16d51c146849 100644 --- a/arch/ia64/ia32/ia32_ldt.c +++ b/arch/ia64/ia32/ia32_ldt.c @@ -10,7 +10,6 @@ #include <linux/string.h> #include <linux/mm.h> #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/vmalloc.h> #include <asm/uaccess.h> diff --git a/arch/ia64/ia32/ia32_signal.c b/arch/ia64/ia32/ia32_signal.c index b3355a9ca2c3..85e82f32e480 100644 --- a/arch/ia64/ia32/ia32_signal.c +++ b/arch/ia64/ia32/ia32_signal.c @@ -18,7 +18,6 @@ #include <linux/sched.h> #include <linux/signal.h> #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/stddef.h> #include <linux/syscalls.h> #include <linux/unistd.h> @@ -452,59 +451,20 @@ sigact_set_handler (struct k_sigaction *sa, unsigned int handler, unsigned int r sa->sa.sa_handler = (__sighandler_t) (((unsigned long) restorer << 32) | handler); } -long -__ia32_rt_sigsuspend (compat_sigset_t *sset, unsigned int sigsetsize, struct sigscratch *scr) +asmlinkage long +sys32_sigsuspend (int history0, int history1, old_sigset_t mask) { - extern long ia64_do_signal (sigset_t *oldset, struct sigscratch *scr, long in_syscall); - sigset_t oldset, set; - - scr->scratch_unat = 0; /* avoid leaking kernel bits to user level */ - memset(&set, 0, sizeof(set)); - - memcpy(&set.sig, &sset->sig, sigsetsize); - - sigdelsetmask(&set, ~_BLOCKABLE); - + mask &= _BLOCKABLE; spin_lock_irq(¤t->sighand->siglock); - { - oldset = current->blocked; - current->blocked = set; - recalc_sigpending(); - } + current->saved_sigmask = current->blocked; + siginitset(¤t->blocked, mask); + recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); - /* - * The return below usually returns to the signal handler. We need to pre-set the - * correct error code here to ensure that the right values get saved in sigcontext - * by ia64_do_signal. - */ - scr->pt.r8 = -EINTR; - while (1) { - current->state = TASK_INTERRUPTIBLE; - schedule(); - if (ia64_do_signal(&oldset, scr, 1)) - return -EINTR; - } -} - -asmlinkage long -ia32_rt_sigsuspend (compat_sigset_t __user *uset, unsigned int sigsetsize, struct sigscratch *scr) -{ - compat_sigset_t set; - - if (sigsetsize > sizeof(compat_sigset_t)) - return -EINVAL; - - if (copy_from_user(&set.sig, &uset->sig, sigsetsize)) - return -EFAULT; - - return __ia32_rt_sigsuspend(&set, sigsetsize, scr); -} - -asmlinkage long -ia32_sigsuspend (unsigned int mask, struct sigscratch *scr) -{ - return __ia32_rt_sigsuspend((compat_sigset_t *) &mask, sizeof(mask), scr); + current->state = TASK_INTERRUPTIBLE; + schedule(); + set_thread_flag(TIF_RESTORE_SIGMASK); + return -ERESTARTNOHAND; } asmlinkage long @@ -811,7 +771,11 @@ get_sigframe (struct k_sigaction *ka, struct pt_regs * regs, size_t frame_size) } /* Legacy stack switching not supported */ - return (void __user *)((esp - frame_size) & -8ul); + esp -= frame_size; + /* Align the stack pointer according to the i386 ABI, + * i.e. so that on function entry ((sp + 4) & 15) == 0. */ + esp = ((esp + 4) & -16ul) - 4; + return (void __user *) esp; } static int diff --git a/arch/ia64/ia32/ia32_support.c b/arch/ia64/ia32/ia32_support.c index 6af400a12ca1..beea7a0b9dc6 100644 --- a/arch/ia64/ia32/ia32_support.c +++ b/arch/ia64/ia32/ia32_support.c @@ -252,10 +252,8 @@ ia32_init (void) extern struct kmem_cache *partial_page_cachep; partial_page_cachep = kmem_cache_create("partial_page_cache", - sizeof(struct partial_page), 0, 0, - NULL, NULL); - if (!partial_page_cachep) - panic("Cannot create partial page SLAB cache"); + sizeof(struct partial_page), + 0, SLAB_PANIC, NULL, NULL); } #endif return 0; diff --git a/arch/ia64/kernel/crash.c b/arch/ia64/kernel/crash.c index 80a94e707827..aeb79fb28f0b 100644 --- a/arch/ia64/kernel/crash.c +++ b/arch/ia64/kernel/crash.c @@ -16,8 +16,8 @@ #include <linux/elfcore.h> #include <linux/sysctl.h> #include <linux/init.h> +#include <linux/kdebug.h> -#include <asm/kdebug.h> #include <asm/mca.h> int kdump_status[NR_CPUS]; @@ -74,7 +74,7 @@ crash_save_this_cpu(void) buf = (u64 *) per_cpu_ptr(crash_notes, cpu); if (!buf) return; - buf = append_elf_note(buf, "CORE", NT_PRSTATUS, prstatus, + buf = append_elf_note(buf, KEXEC_CORE_NOTE_NAME, NT_PRSTATUS, prstatus, sizeof(*prstatus)); final_note(buf); } diff --git a/arch/ia64/kernel/efi.c b/arch/ia64/kernel/efi.c index 78d29b79947d..75ec3478d8a2 100644 --- a/arch/ia64/kernel/efi.c +++ b/arch/ia64/kernel/efi.c @@ -445,11 +445,11 @@ efi_init (void) panic("Woah! Can't find EFI system table.\n"); if (efi.systab->hdr.signature != EFI_SYSTEM_TABLE_SIGNATURE) panic("Woah! EFI system table signature incorrect\n"); - if ((efi.systab->hdr.revision ^ EFI_SYSTEM_TABLE_REVISION) >> 16 != 0) - printk(KERN_WARNING "Warning: EFI system table major version mismatch: " - "got %d.%02d, expected %d.%02d\n", - efi.systab->hdr.revision >> 16, efi.systab->hdr.revision & 0xffff, - EFI_SYSTEM_TABLE_REVISION >> 16, EFI_SYSTEM_TABLE_REVISION & 0xffff); + if ((efi.systab->hdr.revision >> 16) == 0) + printk(KERN_WARNING "Warning: EFI system table version " + "%d.%02d, expected 1.00 or greater\n", + efi.systab->hdr.revision >> 16, + efi.systab->hdr.revision & 0xffff); config_tables = __va(efi.systab->tables); diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S index 55fd2d5471e1..b50bf208678e 100644 --- a/arch/ia64/kernel/entry.S +++ b/arch/ia64/kernel/entry.S @@ -1199,32 +1199,6 @@ ENTRY(notify_resume_user) br.ret.sptk.many rp END(notify_resume_user) -GLOBAL_ENTRY(sys_rt_sigsuspend) - .prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8) - alloc loc1=ar.pfs,8,2,3,0 // preserve all eight input regs in case of syscall restart! - mov r9=ar.unat - mov loc0=rp // save return address - mov out0=in0 // mask - mov out1=in1 // sigsetsize - adds out2=8,sp // out2=&sigscratch->ar_pfs - ;; - .fframe 16 - .spillsp ar.unat, 16 - st8 [sp]=r9,-16 // allocate space for ar.unat and save it - st8 [out2]=loc1,-8 // save ar.pfs, out2=&sigscratch - .body - br.call.sptk.many rp=ia64_rt_sigsuspend -.ret17: .restore sp - adds sp=16,sp // pop scratch stack space - ;; - ld8 r9=[sp] // load new unat from sw->caller_unat - mov rp=loc0 - ;; - mov ar.unat=r9 - mov ar.pfs=loc1 - br.ret.sptk.many rp -END(sys_rt_sigsuspend) - ENTRY(sys_rt_sigreturn) PT_REGS_UNWIND_INFO(0) /* @@ -1598,8 +1572,8 @@ sys_call_table: data8 sys_readlinkat data8 sys_fchmodat data8 sys_faccessat - data8 sys_ni_syscall // reserved for pselect - data8 sys_ni_syscall // 1295 reserved for ppoll + data8 sys_pselect6 + data8 sys_ppoll data8 sys_unshare data8 sys_splice data8 sys_set_robust_list diff --git a/arch/ia64/kernel/err_inject.c b/arch/ia64/kernel/err_inject.c index d3e9f33e8bdd..6a49600cf337 100644 --- a/arch/ia64/kernel/err_inject.c +++ b/arch/ia64/kernel/err_inject.c @@ -236,9 +236,11 @@ static int __cpuinit err_inject_cpu_callback(struct notifier_block *nfb, sys_dev = get_cpu_sysdev(cpu); switch (action) { case CPU_ONLINE: + case CPU_ONLINE_FROZEN: err_inject_add_dev(sys_dev); break; case CPU_DEAD: + case CPU_DEAD_FROZEN: err_inject_remove_dev(sys_dev); break; } diff --git a/arch/ia64/kernel/iosapic.c b/arch/ia64/kernel/iosapic.c index dcfbf3e7a9ef..37f46527d233 100644 --- a/arch/ia64/kernel/iosapic.c +++ b/arch/ia64/kernel/iosapic.c @@ -87,7 +87,6 @@ #include <linux/list.h> #include <linux/pci.h> #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/string.h> #include <linux/bootmem.h> @@ -1013,7 +1012,7 @@ iosapic_register_platform_intr (u32 int_type, unsigned int gsi, /* * ACPI calls this when it finds an entry for a legacy ISA IRQ override. */ -void __init +void __devinit iosapic_override_isa_irq (unsigned int isa_irq, unsigned int gsi, unsigned long polarity, unsigned long trigger) diff --git a/arch/ia64/kernel/irq_ia64.c b/arch/ia64/kernel/irq_ia64.c index 456f57b087ca..bc47049f060f 100644 --- a/arch/ia64/kernel/irq_ia64.c +++ b/arch/ia64/kernel/irq_ia64.c @@ -27,7 +27,6 @@ #include <linux/random.h> /* for rand_initialize_irq() */ #include <linux/signal.h> #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/threads.h> #include <linux/bitops.h> #include <linux/irq.h> @@ -39,6 +38,7 @@ #include <asm/machvec.h> #include <asm/pgtable.h> #include <asm/system.h> +#include <asm/tlbflush.h> #ifdef CONFIG_PERFMON # include <asm/perfmon.h> @@ -127,8 +127,10 @@ void destroy_irq(unsigned int irq) #ifdef CONFIG_SMP # define IS_RESCHEDULE(vec) (vec == IA64_IPI_RESCHEDULE) +# define IS_LOCAL_TLB_FLUSH(vec) (vec == IA64_IPI_LOCAL_TLB_FLUSH) #else # define IS_RESCHEDULE(vec) (0) +# define IS_LOCAL_TLB_FLUSH(vec) (0) #endif /* * That's where the IVT branches when we get an external @@ -180,8 +182,11 @@ ia64_handle_irq (ia64_vector vector, struct pt_regs *regs) saved_tpr = ia64_getreg(_IA64_REG_CR_TPR); ia64_srlz_d(); while (vector != IA64_SPURIOUS_INT_VECTOR) { - if (unlikely(IS_RESCHEDULE(vector))) - kstat_this_cpu.irqs[vector]++; + if (unlikely(IS_LOCAL_TLB_FLUSH(vector))) { + smp_local_flush_tlb(); + kstat_this_cpu.irqs[vector]++; + } else if (unlikely(IS_RESCHEDULE(vector))) + kstat_this_cpu.irqs[vector]++; else { ia64_setreg(_IA64_REG_CR_TPR, vector); ia64_srlz_d(); @@ -227,8 +232,11 @@ void ia64_process_pending_intr(void) * Perform normal interrupt style processing */ while (vector != IA64_SPURIOUS_INT_VECTOR) { - if (unlikely(IS_RESCHEDULE(vector))) - kstat_this_cpu.irqs[vector]++; + if (unlikely(IS_LOCAL_TLB_FLUSH(vector))) { + smp_local_flush_tlb(); + kstat_this_cpu.irqs[vector]++; + } else if (unlikely(IS_RESCHEDULE(vector))) + kstat_this_cpu.irqs[vector]++; else { struct pt_regs *old_regs = set_irq_regs(NULL); @@ -260,12 +268,12 @@ void ia64_process_pending_intr(void) #ifdef CONFIG_SMP -extern irqreturn_t handle_IPI (int irq, void *dev_id); static irqreturn_t dummy_handler (int irq, void *dev_id) { BUG(); } +extern irqreturn_t handle_IPI (int irq, void *dev_id); static struct irqaction ipi_irqaction = { .handler = handle_IPI, @@ -278,6 +286,13 @@ static struct irqaction resched_irqaction = { .flags = IRQF_DISABLED, .name = "resched" }; + +static struct irqaction tlb_irqaction = { + .handler = dummy_handler, + .flags = IRQF_DISABLED, + .name = "tlb_flush" +}; + #endif void @@ -303,6 +318,7 @@ init_IRQ (void) #ifdef CONFIG_SMP register_percpu_irq(IA64_IPI_VECTOR, &ipi_irqaction); register_percpu_irq(IA64_IPI_RESCHEDULE, &resched_irqaction); + register_percpu_irq(IA64_IPI_LOCAL_TLB_FLUSH, &tlb_irqaction); #endif #ifdef CONFIG_PERFMON pfm_init_percpu(); diff --git a/arch/ia64/kernel/kprobes.c b/arch/ia64/kernel/kprobes.c index 6cb56dd4056d..4f5fd0960ba7 100644 --- a/arch/ia64/kernel/kprobes.c +++ b/arch/ia64/kernel/kprobes.c @@ -29,9 +29,9 @@ #include <linux/slab.h> #include <linux/preempt.h> #include <linux/moduleloader.h> +#include <linux/kdebug.h> #include <asm/pgtable.h> -#include <asm/kdebug.h> #include <asm/sections.h> #include <asm/uaccess.h> @@ -444,7 +444,8 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) break; } - BUG_ON(!orig_ret_address || (orig_ret_address == trampoline_address)); + kretprobe_assert(ri, orig_ret_address, trampoline_address); + regs->cr_iip = orig_ret_address; reset_current_kprobe(); @@ -464,23 +465,13 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) } /* Called with kretprobe_lock held */ -void __kprobes arch_prepare_kretprobe(struct kretprobe *rp, +void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri, struct pt_regs *regs) { - struct kretprobe_instance *ri; - - if ((ri = get_free_rp_inst(rp)) != NULL) { - ri->rp = rp; - ri->task = current; - ri->ret_addr = (kprobe_opcode_t *)regs->b0; - - /* Replace the return addr with trampoline addr */ - regs->b0 = ((struct fnptr *)kretprobe_trampoline)->ip; + ri->ret_addr = (kprobe_opcode_t *)regs->b0; - add_rp_inst(ri); - } else { - rp->nmissed++; - } + /* Replace the return addr with trampoline addr */ + regs->b0 = ((struct fnptr *)kretprobe_trampoline)->ip; } int __kprobes arch_prepare_kprobe(struct kprobe *p) @@ -1021,3 +1012,12 @@ int __init arch_init_kprobes(void) (kprobe_opcode_t *)((struct fnptr *)kretprobe_trampoline)->ip; return register_kprobe(&trampoline_p); } + +int __kprobes arch_trampoline_kprobe(struct kprobe *p) +{ + if (p->addr == + (kprobe_opcode_t *)((struct fnptr *)kretprobe_trampoline)->ip) + return 1; + + return 0; +} diff --git a/arch/ia64/kernel/mca.c b/arch/ia64/kernel/mca.c index 491687f84fb5..f8ae709de0b5 100644 --- a/arch/ia64/kernel/mca.c +++ b/arch/ia64/kernel/mca.c @@ -63,7 +63,6 @@ #include <linux/sched.h> #include <linux/interrupt.h> #include <linux/irq.h> -#include <linux/smp_lock.h> #include <linux/bootmem.h> #include <linux/acpi.h> #include <linux/timer.h> @@ -72,9 +71,9 @@ #include <linux/smp.h> #include <linux/workqueue.h> #include <linux/cpumask.h> +#include <linux/kdebug.h> #include <asm/delay.h> -#include <asm/kdebug.h> #include <asm/machvec.h> #include <asm/meminit.h> #include <asm/page.h> @@ -1690,7 +1689,7 @@ format_mca_init_stack(void *mca_data, unsigned long offset, ti->preempt_count = 1; ti->task = p; ti->cpu = cpu; - p->thread_info = ti; + p->stack = ti; p->state = TASK_UNINTERRUPTIBLE; cpu_set(cpu, p->cpus_allowed); INIT_LIST_HEAD(&p->tasks); diff --git a/arch/ia64/kernel/mca_drv.c b/arch/ia64/kernel/mca_drv.c index 832cf1e647e8..70b8bdbb7e6f 100644 --- a/arch/ia64/kernel/mca_drv.c +++ b/arch/ia64/kernel/mca_drv.c @@ -14,7 +14,6 @@ #include <linux/interrupt.h> #include <linux/irq.h> #include <linux/kallsyms.h> -#include <linux/smp_lock.h> #include <linux/bootmem.h> #include <linux/acpi.h> #include <linux/timer.h> diff --git a/arch/ia64/kernel/palinfo.c b/arch/ia64/kernel/palinfo.c index a71df9ae0397..85829e27785c 100644 --- a/arch/ia64/kernel/palinfo.c +++ b/arch/ia64/kernel/palinfo.c @@ -975,9 +975,11 @@ static int palinfo_cpu_callback(struct notifier_block *nfb, switch (action) { case CPU_ONLINE: + case CPU_ONLINE_FROZEN: create_palinfo_proc_entries(hotcpu); break; case CPU_DEAD: + case CPU_DEAD_FROZEN: remove_palinfo_proc_entries(hotcpu); break; } diff --git a/arch/ia64/kernel/perfmon.c b/arch/ia64/kernel/perfmon.c index abc7ad035886..e7191ca30b16 100644 --- a/arch/ia64/kernel/perfmon.c +++ b/arch/ia64/kernel/perfmon.c @@ -23,7 +23,6 @@ #include <linux/kernel.h> #include <linux/sched.h> #include <linux/interrupt.h> -#include <linux/smp_lock.h> #include <linux/proc_fs.h> #include <linux/seq_file.h> #include <linux/init.h> diff --git a/arch/ia64/kernel/process.c b/arch/ia64/kernel/process.c index ae96d4176995..d1c3ed9943e5 100644 --- a/arch/ia64/kernel/process.c +++ b/arch/ia64/kernel/process.c @@ -20,20 +20,19 @@ #include <linux/personality.h> #include <linux/sched.h> #include <linux/slab.h> -#include <linux/smp_lock.h> #include <linux/stddef.h> #include <linux/thread_info.h> #include <linux/unistd.h> #include <linux/efi.h> #include <linux/interrupt.h> #include <linux/delay.h> +#include <linux/kdebug.h> #include <asm/cpu.h> #include <asm/delay.h> #include <asm/elf.h> #include <asm/ia32.h> #include <asm/irq.h> -#include <asm/kdebug.h> #include <asm/kexec.h> #include <asm/pgalloc.h> #include <asm/processor.h> @@ -156,7 +155,7 @@ show_regs (struct pt_regs *regs) } void -do_notify_resume_user (sigset_t *oldset, struct sigscratch *scr, long in_syscall) +do_notify_resume_user (sigset_t *unused, struct sigscratch *scr, long in_syscall) { if (fsys_mode(current, &scr->pt)) { /* defer signal-handling etc. until we return to privilege-level 0. */ @@ -171,8 +170,8 @@ do_notify_resume_user (sigset_t *oldset, struct sigscratch *scr, long in_syscall #endif /* deal with pending signal delivery */ - if (test_thread_flag(TIF_SIGPENDING)) - ia64_do_signal(oldset, scr, in_syscall); + if (test_thread_flag(TIF_SIGPENDING)||test_thread_flag(TIF_RESTORE_SIGMASK)) + ia64_do_signal(scr, in_syscall); } static int pal_halt = 1; @@ -237,6 +236,7 @@ void cpu_idle_wait(void) { unsigned int cpu, this_cpu = get_cpu(); cpumask_t map; + cpumask_t tmp = current->cpus_allowed; set_cpus_allowed(current, cpumask_of_cpu(this_cpu)); put_cpu(); @@ -258,6 +258,7 @@ void cpu_idle_wait(void) } cpus_and(map, map, cpu_online_map); } while (!cpus_empty(map)); + set_cpus_allowed(current, tmp); } EXPORT_SYMBOL_GPL(cpu_idle_wait); diff --git a/arch/ia64/kernel/relocate_kernel.S b/arch/ia64/kernel/relocate_kernel.S index ae473e3f2a0d..903babd22d62 100644 --- a/arch/ia64/kernel/relocate_kernel.S +++ b/arch/ia64/kernel/relocate_kernel.S @@ -94,7 +94,7 @@ GLOBAL_ENTRY(relocate_new_kernel) 4: srlz.i ;; - //purge TR entry for kernel text and data + // purge TR entry for kernel text and data movl r16=KERNEL_START mov r18=KERNEL_TR_PAGE_SHIFT<<2 ;; @@ -104,15 +104,6 @@ GLOBAL_ENTRY(relocate_new_kernel) srlz.i ;; - // purge TR entry for percpu data - movl r16=PERCPU_ADDR - mov r18=PERCPU_PAGE_SHIFT<<2 - ;; - ptr.d r16,r18 - ;; - srlz.d - ;; - // purge TR entry for pal code mov r16=in3 mov r18=IA64_GRANULE_SHIFT<<2 diff --git a/arch/ia64/kernel/salinfo.c b/arch/ia64/kernel/salinfo.c index af9f8754d847..89f6b138a62c 100644 --- a/arch/ia64/kernel/salinfo.c +++ b/arch/ia64/kernel/salinfo.c @@ -42,7 +42,6 @@ #include <linux/proc_fs.h> #include <linux/module.h> #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/timer.h> #include <linux/vmalloc.h> @@ -583,6 +582,7 @@ salinfo_cpu_callback(struct notifier_block *nb, unsigned long action, void *hcpu struct salinfo_data *data; switch (action) { case CPU_ONLINE: + case CPU_ONLINE_FROZEN: spin_lock_irqsave(&data_saved_lock, flags); for (i = 0, data = salinfo_data; i < ARRAY_SIZE(salinfo_data); @@ -593,6 +593,7 @@ salinfo_cpu_callback(struct notifier_block *nb, unsigned long action, void *hcpu spin_unlock_irqrestore(&data_saved_lock, flags); break; case CPU_DEAD: + case CPU_DEAD_FROZEN: spin_lock_irqsave(&data_saved_lock, flags); for (i = 0, data = salinfo_data; i < ARRAY_SIZE(salinfo_data); diff --git a/arch/ia64/kernel/setup.c b/arch/ia64/kernel/setup.c index 6e19da122ae3..9df1efe7487d 100644 --- a/arch/ia64/kernel/setup.c +++ b/arch/ia64/kernel/setup.c @@ -786,7 +786,7 @@ identify_cpu (struct cpuinfo_ia64 *c) c->unimpl_pa_mask = ~((1L<<63) | ((1L << phys_addr_size) - 1)); } -void +void __init setup_per_cpu_areas (void) { /* start_kernel() requires this... */ diff --git a/arch/ia64/kernel/sigframe.h b/arch/ia64/kernel/sigframe.h index 37b986cb86e0..9fd9a1933b3d 100644 --- a/arch/ia64/kernel/sigframe.h +++ b/arch/ia64/kernel/sigframe.h @@ -22,4 +22,4 @@ struct sigframe { struct sigcontext sc; }; -extern long ia64_do_signal (sigset_t *, struct sigscratch *, long); +extern void ia64_do_signal (struct sigscratch *, long); diff --git a/arch/ia64/kernel/signal.c b/arch/ia64/kernel/signal.c index 77f8b49c7882..aeec8184e862 100644 --- a/arch/ia64/kernel/signal.c +++ b/arch/ia64/kernel/signal.c @@ -14,7 +14,6 @@ #include <linux/sched.h> #include <linux/signal.h> #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/stddef.h> #include <linux/tty.h> #include <linux/binfmts.h> @@ -41,47 +40,6 @@ # define GET_SIGSET(k,u) __get_user((k)->sig[0], &(u)->sig[0]) #endif -long -ia64_rt_sigsuspend (sigset_t __user *uset, size_t sigsetsize, struct sigscratch *scr) -{ - sigset_t oldset, set; - - /* XXX: Don't preclude handling different sized sigset_t's. */ - if (sigsetsize != sizeof(sigset_t)) - return -EINVAL; - - if (!access_ok(VERIFY_READ, uset, sigsetsize)) - return -EFAULT; - - if (GET_SIGSET(&set, uset)) - return -EFAULT; - - sigdelsetmask(&set, ~_BLOCKABLE); - - spin_lock_irq(¤t->sighand->siglock); - { - oldset = current->blocked; - current->blocked = set; - recalc_sigpending(); - } - spin_unlock_irq(¤t->sighand->siglock); - - /* - * The return below usually returns to the signal handler. We need to - * pre-set the correct error code here to ensure that the right values - * get saved in sigcontext by ia64_do_signal. - */ - scr->pt.r8 = EINTR; - scr->pt.r10 = -1; - - while (1) { - current->state = TASK_INTERRUPTIBLE; - schedule(); - if (ia64_do_signal(&oldset, scr, 1)) - return -EINTR; - } -} - asmlinkage long sys_sigaltstack (const stack_t __user *uss, stack_t __user *uoss, long arg2, long arg3, long arg4, long arg5, long arg6, long arg7, @@ -478,10 +436,11 @@ handle_signal (unsigned long sig, struct k_sigaction *ka, siginfo_t *info, sigse * Note that `init' is a special process: it doesn't get signals it doesn't want to * handle. Thus you cannot kill init even with a SIGKILL even by mistake. */ -long -ia64_do_signal (sigset_t *oldset, struct sigscratch *scr, long in_syscall) +void +ia64_do_signal (struct sigscratch *scr, long in_syscall) { struct k_sigaction ka; + sigset_t *oldset; siginfo_t info; long restart = in_syscall; long errno = scr->pt.r8; @@ -493,9 +452,11 @@ ia64_do_signal (sigset_t *oldset, struct sigscratch *scr, long in_syscall) * doing anything if so. */ if (!user_mode(&scr->pt)) - return 0; + return; - if (!oldset) + if (test_thread_flag(TIF_RESTORE_SIGMASK)) + oldset = ¤t->saved_sigmask; + else oldset = ¤t->blocked; /* @@ -558,8 +519,15 @@ ia64_do_signal (sigset_t *oldset, struct sigscratch *scr, long in_syscall) * Whee! Actually deliver the signal. If the delivery failed, we need to * continue to iterate in this loop so we can deliver the SIGSEGV... */ - if (handle_signal(signr, &ka, &info, oldset, scr)) - return 1; + if (handle_signal(signr, &ka, &info, oldset, scr)) { + /* a signal was successfully delivered; the saved + * sigmask will have been stored in the signal frame, + * and will be restored by sigreturn, so we can simply + * clear the TIF_RESTORE_SIGMASK flag */ + if (test_thread_flag(TIF_RESTORE_SIGMASK)) + clear_thread_flag(TIF_RESTORE_SIGMASK); + return; + } } /* Did we come from a system call? */ @@ -585,5 +553,11 @@ ia64_do_signal (sigset_t *oldset, struct sigscratch *scr, long in_syscall) } } } - return 0; + + /* if there's no signal to deliver, we just put the saved sigmask + * back */ + if (test_thread_flag(TIF_RESTORE_SIGMASK)) { + clear_thread_flag(TIF_RESTORE_SIGMASK); + sigprocmask(SIG_SETMASK, ¤t->saved_sigmask, NULL); + } } diff --git a/arch/ia64/kernel/smp.c b/arch/ia64/kernel/smp.c index 55ddd809b02d..221de3804560 100644 --- a/arch/ia64/kernel/smp.c +++ b/arch/ia64/kernel/smp.c @@ -50,6 +50,18 @@ #include <asm/mca.h> /* + * Note: alignment of 4 entries/cacheline was empirically determined + * to be a good tradeoff between hot cachelines & spreading the array + * across too many cacheline. + */ +static struct local_tlb_flush_counts { + unsigned int count; +} __attribute__((__aligned__(32))) local_tlb_flush_counts[NR_CPUS]; + +static DEFINE_PER_CPU(unsigned int, shadow_flush_counts[NR_CPUS]) ____cacheline_aligned; + + +/* * Structure and data for smp_call_function(). This is designed to minimise static memory * requirements. It also looks cleaner. */ @@ -248,6 +260,62 @@ smp_send_reschedule (int cpu) platform_send_ipi(cpu, IA64_IPI_RESCHEDULE, IA64_IPI_DM_INT, 0); } +/* + * Called with preeemption disabled. + */ +static void +smp_send_local_flush_tlb (int cpu) +{ + platform_send_ipi(cpu, IA64_IPI_LOCAL_TLB_FLUSH, IA64_IPI_DM_INT, 0); +} + +void +smp_local_flush_tlb(void) +{ + /* + * Use atomic ops. Otherwise, the load/increment/store sequence from + * a "++" operation can have the line stolen between the load & store. + * The overhead of the atomic op in negligible in this case & offers + * significant benefit for the brief periods where lots of cpus + * are simultaneously flushing TLBs. + */ + ia64_fetchadd(1, &local_tlb_flush_counts[smp_processor_id()].count, acq); + local_flush_tlb_all(); +} + +#define FLUSH_DELAY 5 /* Usec backoff to eliminate excessive cacheline bouncing */ + +void +smp_flush_tlb_cpumask(cpumask_t xcpumask) +{ + unsigned int *counts = __ia64_per_cpu_var(shadow_flush_counts); + cpumask_t cpumask = xcpumask; + int mycpu, cpu, flush_mycpu = 0; + + preempt_disable(); + mycpu = smp_processor_id(); + + for_each_cpu_mask(cpu, cpumask) + counts[cpu] = local_tlb_flush_counts[cpu].count; + + mb(); + for_each_cpu_mask(cpu, cpumask) { + if (cpu == mycpu) + flush_mycpu = 1; + else + smp_send_local_flush_tlb(cpu); + } + + if (flush_mycpu) + smp_local_flush_tlb(); + + for_each_cpu_mask(cpu, cpumask) + while(counts[cpu] == local_tlb_flush_counts[cpu].count) + udelay(FLUSH_DELAY); + + preempt_enable(); +} + void smp_flush_tlb_all (void) { diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c index ff7df439da6d..a44792d0f3a9 100644 --- a/arch/ia64/kernel/smpboot.c +++ b/arch/ia64/kernel/smpboot.c @@ -35,7 +35,6 @@ #include <linux/mm.h> #include <linux/notifier.h> #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/spinlock.h> #include <linux/efi.h> #include <linux/percpu.h> diff --git a/arch/ia64/kernel/sys_ia64.c b/arch/ia64/kernel/sys_ia64.c index 2fcaa2051aa3..1eda194b9559 100644 --- a/arch/ia64/kernel/sys_ia64.c +++ b/arch/ia64/kernel/sys_ia64.c @@ -13,7 +13,6 @@ #include <linux/shm.h> #include <linux/file.h> /* doh, must come after sched.h... */ #include <linux/smp.h> -#include <linux/smp_lock.h> #include <linux/syscalls.h> #include <linux/highuid.h> #include <linux/hugetlb.h> diff --git a/arch/ia64/kernel/time.c b/arch/ia64/kernel/time.c index 39e0cd3a0884..a06667c7acc0 100644 --- a/arch/ia64/kernel/time.c +++ b/arch/ia64/kernel/time.c @@ -235,7 +235,7 @@ ia64_init_itm (void) static struct irqaction timer_irqaction = { .handler = timer_interrupt, - .flags = IRQF_DISABLED, + .flags = IRQF_DISABLED | IRQF_IRQPOLL, .name = "timer" }; diff --git a/arch/ia64/kernel/topology.c b/arch/ia64/kernel/topology.c index 687500ddb4b8..94ae3c87d828 100644 --- a/arch/ia64/kernel/topology.c +++ b/arch/ia64/kernel/topology.c @@ -412,9 +412,11 @@ static int __cpuinit cache_cpu_callback(struct notifier_block *nfb, sys_dev = get_cpu_sysdev(cpu); switch (action) { case CPU_ONLINE: + case CPU_ONLINE_FROZEN: cache_add_dev(sys_dev); break; case CPU_DEAD: + case CPU_DEAD_FROZEN: cache_remove_dev(sys_dev); break; } diff --git a/arch/ia64/kernel/traps.c b/arch/ia64/kernel/traps.c index 765cbe5ba6ae..b8e0d70bf989 100644 --- a/arch/ia64/kernel/traps.c +++ b/arch/ia64/kernel/traps.c @@ -16,33 +16,17 @@ #include <linux/hardirq.h> #include <linux/kprobes.h> #include <linux/delay.h> /* for ssleep() */ +#include <linux/kdebug.h> #include <asm/fpswa.h> #include <asm/ia32.h> #include <asm/intrinsics.h> #include <asm/processor.h> #include <asm/uaccess.h> -#include <asm/kdebug.h> fpswa_interface_t *fpswa_interface; EXPORT_SYMBOL(fpswa_interface); -ATOMIC_NOTIFIER_HEAD(ia64die_chain); - -int -register_die_notifier(struct notifier_block *nb) -{ - return atomic_notifier_chain_register(&ia64die_chain, nb); -} -EXPORT_SYMBOL_GPL(register_die_notifier); - -int -unregister_die_notifier(struct notifier_block *nb) -{ - return atomic_notifier_chain_unregister(&ia64die_chain, nb); -} -EXPORT_SYMBOL_GPL(unregister_die_notifier); - void __init trap_init (void) { @@ -59,9 +43,9 @@ die (const char *str, struct pt_regs *regs, long err) u32 lock_owner; int lock_owner_depth; } die = { - .lock = SPIN_LOCK_UNLOCKED, - .lock_owner = -1, - .lock_owner_depth = 0 + .lock = __SPIN_LOCK_UNLOCKED(die.lock), + .lock_owner = -1, + .lock_owner_depth = 0 }; static int die_counter; int cpu = get_cpu(); diff --git a/arch/ia64/kernel/unaligned.c b/arch/ia64/kernel/unaligned.c index 1e357550c776..fe6aa5a9f8fa 100644 --- a/arch/ia64/kernel/unaligned.c +++ b/arch/ia64/kernel/unaligned.c @@ -15,7 +15,6 @@ */ #include <linux/kernel.h> #include <linux/sched.h> -#include <linux/smp_lock.h> #include <linux/tty.h> #include <asm/intrinsics.h> diff --git a/arch/ia64/kernel/unwind.c b/arch/ia64/kernel/unwind.c index 93d5a3b41f69..fe1426266b9b 100644 --- a/arch/ia64/kernel/unwind.c +++ b/arch/ia64/kernel/unwind.c @@ -60,6 +60,7 @@ # define UNW_DEBUG_ON(n) unw_debug_level >= n /* Do not code a printk level, not all debug lines end in newline */ # define UNW_DPRINT(n, ...) if (UNW_DEBUG_ON(n)) printk(__VA_ARGS__) +# undef inline # define inline #else /* !UNW_DEBUG */ # define UNW_DEBUG_ON(n) 0 @@ -145,7 +146,7 @@ static struct { # endif } unw = { .tables = &unw.kernel_table, - .lock = SPIN_LOCK_UNLOCKED, + .lock = __SPIN_LOCK_UNLOCKED(unw.lock), .save_order = { UNW_REG_RP, UNW_REG_PFS, UNW_REG_PSP, UNW_REG_PR, UNW_REG_UNAT, UNW_REG_LC, UNW_REG_FPSR, UNW_REG_PRI_UNAT_GR @@ -1943,9 +1944,9 @@ EXPORT_SYMBOL(unw_unwind); int unw_unwind_to_user (struct unw_frame_info *info) { - unsigned long ip, sp, pr = 0; + unsigned long ip, sp, pr = info->pr; - while (unw_unwind(info) >= 0) { + do { unw_get_sp(info, &sp); if ((long)((unsigned long)info->task + IA64_STK_OFFSET - sp) < IA64_PT_REGS_SIZE) { @@ -1963,7 +1964,7 @@ unw_unwind_to_user (struct unw_frame_info *info) __FUNCTION__, ip); return -1; } - } + } while (unw_unwind(info) >= 0); unw_get_ip(info, &ip); UNW_DPRINT(0, "unwind.%s: failed to unwind to user-level (ip=0x%lx)\n", __FUNCTION__, ip); diff --git a/arch/ia64/mm/discontig.c b/arch/ia64/mm/discontig.c index 872da7a2accd..94844442812a 100644 --- a/arch/ia64/mm/discontig.c +++ b/arch/ia64/mm/discontig.c @@ -693,6 +693,7 @@ void __init paging_init(void) zero_page_memmap_ptr = virt_to_page(ia64_imva(empty_zero_page)); } +#ifdef CONFIG_MEMORY_HOTPLUG pg_data_t *arch_alloc_nodedata(int nid) { unsigned long size = compute_pernodesize(nid); @@ -710,3 +711,4 @@ void arch_refresh_nodedata(int update_node, pg_data_t *update_pgdat) pgdat_list[update_node] = update_pgdat; scatter_node_data(); } +#endif diff --git a/arch/ia64/mm/fault.c b/arch/ia64/mm/fault.c index 59f3ab937615..21658e02116c 100644 --- a/arch/ia64/mm/fault.c +++ b/arch/ia64/mm/fault.c @@ -7,15 +7,14 @@ #include <linux/sched.h> #include <linux/kernel.h> #include <linux/mm.h> -#include <linux/smp_lock.h> #include <linux/interrupt.h> #include <linux/kprobes.h> +#include <linux/kdebug.h> #include <asm/pgtable.h> #include <asm/processor.h> #include <asm/system.h> #include <asm/uaccess.h> -#include <asm/kdebug.h> extern void die (char *, struct pt_regs *, long); diff --git a/arch/ia64/mm/hugetlbpage.c b/arch/ia64/mm/hugetlbpage.c index c7c90f4f4298..1346b7f05397 100644 --- a/arch/ia64/mm/hugetlbpage.c +++ b/arch/ia64/mm/hugetlbpage.c @@ -13,7 +13,6 @@ #include <linux/mm.h> #include <linux/hugetlb.h> #include <linux/pagemap.h> -#include <linux/smp_lock.h> #include <linux/slab.h> #include <linux/sysctl.h> #include <asm/mman.h> diff --git a/arch/ia64/mm/tlb.c b/arch/ia64/mm/tlb.c index ffad7624436c..fa4e6d4810f3 100644 --- a/arch/ia64/mm/tlb.c +++ b/arch/ia64/mm/tlb.c @@ -32,9 +32,9 @@ static struct { } purge; struct ia64_ctx ia64_ctx = { - .lock = SPIN_LOCK_UNLOCKED, - .next = 1, - .max_ctx = ~0U + .lock = __SPIN_LOCK_UNLOCKED(ia64_ctx.lock), + .next = 1, + .max_ctx = ~0U }; DEFINE_PER_CPU(u8, ia64_need_tlb_flush); diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index 9f635896d252..3549f3b42592 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -18,7 +18,6 @@ #include <linux/init.h> #include <linux/ioport.h> #include <linux/slab.h> -#include <linux/smp_lock.h> #include <linux/spinlock.h> #include <asm/machvec.h> diff --git a/arch/ia64/sn/kernel/irq.c b/arch/ia64/sn/kernel/irq.c index 8d2a1bfbfe7c..7f6d2360a262 100644 --- a/arch/ia64/sn/kernel/irq.c +++ b/arch/ia64/sn/kernel/irq.c @@ -59,6 +59,22 @@ void sn_intr_free(nasid_t local_nasid, int local_widget, (u64) sn_irq_info->irq_cookie, 0, 0); } +u64 sn_intr_redirect(nasid_t local_nasid, int local_widget, + struct sn_irq_info *sn_irq_info, + nasid_t req_nasid, int req_slice) +{ + struct ia64_sal_retval ret_stuff; + ret_stuff.status = 0; + ret_stuff.v0 = 0; + + SAL_CALL_NOLOCK(ret_stuff, (u64) SN_SAL_IOIF_INTERRUPT, + (u64) SAL_INTR_REDIRECT, (u64) local_nasid, + (u64) local_widget, __pa(sn_irq_info), + (u64) req_nasid, (u64) req_slice, 0); + + return ret_stuff.status; +} + static unsigned int sn_startup_irq(unsigned int irq) { return 0; @@ -127,15 +143,8 @@ struct sn_irq_info *sn_retarget_vector(struct sn_irq_info *sn_irq_info, struct sn_irq_info *new_irq_info; struct sn_pcibus_provider *pci_provider; - new_irq_info = kmalloc(sizeof(struct sn_irq_info), GFP_ATOMIC); - if (new_irq_info == NULL) - return NULL; - - memcpy(new_irq_info, sn_irq_info, sizeof(struct sn_irq_info)); - - bridge = (u64) new_irq_info->irq_bridge; + bridge = (u64) sn_irq_info->irq_bridge; if (!bridge) { - kfree(new_irq_info); return NULL; /* irq is not a device interrupt */ } @@ -145,8 +154,25 @@ struct sn_irq_info *sn_retarget_vector(struct sn_irq_info *sn_irq_info, local_widget = TIO_SWIN_WIDGETNUM(bridge); else local_widget = SWIN_WIDGETNUM(bridge); - vector = sn_irq_info->irq_irq; + + /* Make use of SAL_INTR_REDIRECT if PROM supports it */ + status = sn_intr_redirect(local_nasid, local_widget, sn_irq_info, nasid, slice); + if (!status) { + new_irq_info = sn_irq_info; + goto finish_up; + } + + /* + * PROM does not support SAL_INTR_REDIRECT, or it failed. + * Revert to old method. + */ + new_irq_info = kmalloc(sizeof(struct sn_irq_info), GFP_ATOMIC); + if (new_irq_info == NULL) + return NULL; + + memcpy(new_irq_info, sn_irq_info, sizeof(struct sn_irq_info)); + /* Free the old PROM new_irq_info structure */ sn_intr_free(local_nasid, local_widget, new_irq_info); unregister_intr_pda(new_irq_info); @@ -162,11 +188,18 @@ struct sn_irq_info *sn_retarget_vector(struct sn_irq_info *sn_irq_info, return NULL; } + register_intr_pda(new_irq_info); + spin_lock(&sn_irq_info_lock); + list_replace_rcu(&sn_irq_info->list, &new_irq_info->list); + spin_unlock(&sn_irq_info_lock); + call_rcu(&sn_irq_info->rcu, sn_irq_info_free); + + +finish_up: /* Update kernels new_irq_info with new target info */ cpuid = nasid_slice_to_cpuid(new_irq_info->irq_nasid, new_irq_info->irq_slice); new_irq_info->irq_cpuid = cpuid; - register_intr_pda(new_irq_info); pci_provider = sn_pci_provider[new_irq_info->irq_bridge_type]; @@ -178,11 +211,6 @@ struct sn_irq_info *sn_retarget_vector(struct sn_irq_info *sn_irq_info, pci_provider && pci_provider->target_interrupt) (pci_provider->target_interrupt)(new_irq_info); - spin_lock(&sn_irq_info_lock); - list_replace_rcu(&sn_irq_info->list, &new_irq_info->list); - spin_unlock(&sn_irq_info_lock); - call_rcu(&sn_irq_info->rcu, sn_irq_info_free); - #ifdef CONFIG_SMP cpuphys = cpu_physical_id(cpuid); set_irq_affinity_info((vector & 0xff), cpuphys, 0); diff --git a/arch/ia64/sn/kernel/sn2/sn2_smp.c b/arch/ia64/sn/kernel/sn2/sn2_smp.c index 601747b1e22a..5d318b579fb1 100644 --- a/arch/ia64/sn/kernel/sn2/sn2_smp.c +++ b/arch/ia64/sn/kernel/sn2/sn2_smp.c @@ -46,6 +46,9 @@ DECLARE_PER_CPU(struct ptc_stats, ptcstats); static __cacheline_aligned DEFINE_SPINLOCK(sn2_global_ptc_lock); +/* 0 = old algorithm (no IPI flushes), 1 = ipi deadlock flush, 2 = ipi instead of SHUB ptc, >2 = always ipi */ +static int sn2_flush_opt = 0; + extern unsigned long sn2_ptc_deadlock_recovery_core(volatile unsigned long *, unsigned long, volatile unsigned long *, unsigned long, @@ -76,6 +79,8 @@ struct ptc_stats { unsigned long shub_itc_clocks; unsigned long shub_itc_clocks_max; unsigned long shub_ptc_flushes_not_my_mm; + unsigned long shub_ipi_flushes; + unsigned long shub_ipi_flushes_itc_clocks; }; #define sn2_ptctest 0 @@ -121,6 +126,18 @@ void sn_tlb_migrate_finish(struct mm_struct *mm) flush_tlb_mm(mm); } +static void +sn2_ipi_flush_all_tlb(struct mm_struct *mm) +{ + unsigned long itc; + + itc = ia64_get_itc(); + smp_flush_tlb_cpumask(mm->cpu_vm_mask); + itc = ia64_get_itc() - itc; + __get_cpu_var(ptcstats).shub_ipi_flushes_itc_clocks += itc; + __get_cpu_var(ptcstats).shub_ipi_flushes++; +} + /** * sn2_global_tlb_purge - globally purge translation cache of virtual address range * @mm: mm_struct containing virtual address range @@ -154,7 +171,12 @@ sn2_global_tlb_purge(struct mm_struct *mm, unsigned long start, unsigned long itc, itc2, flags, data0 = 0, data1 = 0, rr_value, old_rr = 0; short nasids[MAX_NUMNODES], nix; nodemask_t nodes_flushed; - int active, max_active, deadlock; + int active, max_active, deadlock, flush_opt = sn2_flush_opt; + + if (flush_opt > 2) { + sn2_ipi_flush_all_tlb(mm); + return; + } nodes_clear(nodes_flushed); i = 0; @@ -189,6 +211,12 @@ sn2_global_tlb_purge(struct mm_struct *mm, unsigned long start, return; } + if (flush_opt == 2) { + sn2_ipi_flush_all_tlb(mm); + preempt_enable(); + return; + } + itc = ia64_get_itc(); nix = 0; for_each_node_mask(cnode, nodes_flushed) @@ -256,6 +284,8 @@ sn2_global_tlb_purge(struct mm_struct *mm, unsigned long start, } if (active >= max_active || i == (nix - 1)) { if ((deadlock = wait_piowc())) { + if (flush_opt == 1) + goto done; sn2_ptc_deadlock_recovery(nasids, ibegin, i, mynasid, ptc0, data0, ptc1, data1); if (reset_max_active_on_deadlock()) max_active = 1; @@ -267,6 +297,7 @@ sn2_global_tlb_purge(struct mm_struct *mm, unsigned long start, start += (1UL << nbits); } while (start < end); +done: itc2 = ia64_get_itc() - itc2; __get_cpu_var(ptcstats).shub_itc_clocks += itc2; if (itc2 > __get_cpu_var(ptcstats).shub_itc_clocks_max) @@ -279,6 +310,11 @@ sn2_global_tlb_purge(struct mm_struct *mm, unsigned long start, spin_unlock_irqrestore(PTC_LOCK(shub1), flags); + if (flush_opt == 1 && deadlock) { + __get_cpu_var(ptcstats).deadlocks++; + sn2_ipi_flush_all_tlb(mm); + } + preempt_enable(); } @@ -425,24 +461,42 @@ static int sn2_ptc_seq_show(struct seq_file *file, void *data) if (!cpu) { seq_printf(file, - "# cpu ptc_l newrid ptc_flushes nodes_flushed deadlocks lock_nsec shub_nsec shub_nsec_max not_my_mm deadlock2\n"); - seq_printf(file, "# ptctest %d\n", sn2_ptctest); + "# cpu ptc_l newrid ptc_flushes nodes_flushed deadlocks lock_nsec shub_nsec shub_nsec_max not_my_mm deadlock2 ipi_fluches ipi_nsec\n"); + seq_printf(file, "# ptctest %d, flushopt %d\n", sn2_ptctest, sn2_flush_opt); } if (cpu < NR_CPUS && cpu_online(cpu)) { stat = &per_cpu(ptcstats, cpu); - seq_printf(file, "cpu %d %ld %ld %ld %ld %ld %ld %ld %ld %ld %ld\n", cpu, stat->ptc_l, + seq_printf(file, "cpu %d %ld %ld %ld %ld %ld %ld %ld %ld %ld %ld %ld %ld\n", cpu, stat->ptc_l, stat->change_rid, stat->shub_ptc_flushes, stat->nodes_flushed, stat->deadlocks, 1000 * stat->lock_itc_clocks / per_cpu(cpu_info, cpu).cyc_per_usec, 1000 * stat->shub_itc_clocks / per_cpu(cpu_info, cpu).cyc_per_usec, 1000 * stat->shub_itc_clocks_max / per_cpu(cpu_info, cpu).cyc_per_usec, stat->shub_ptc_flushes_not_my_mm, - stat->deadlocks2); + stat->deadlocks2, + stat->shub_ipi_flushes, + 1000 * stat->shub_ipi_flushes_itc_clocks / per_cpu(cpu_info, cpu).cyc_per_usec); } return 0; } +static ssize_t sn2_ptc_proc_write(struct file *file, const char __user *user, size_t count, loff_t *data) +{ + int cpu; + char optstr[64]; + + if (copy_from_user(optstr, user, count)) + return -EFAULT; + optstr[count - 1] = '\0'; + sn2_flush_opt = simple_strtoul(optstr, NULL, 0); + + for_each_online_cpu(cpu) + memset(&per_cpu(ptcstats, cpu), 0, sizeof(struct ptc_stats)); + + return count; +} + static struct seq_operations sn2_ptc_seq_ops = { .start = sn2_ptc_seq_start, .next = sn2_ptc_seq_next, @@ -458,6 +512,7 @@ static int sn2_ptc_proc_open(struct inode *inode, struct file *file) static const struct file_operations proc_sn2_ptc_operations = { .open = sn2_ptc_proc_open, .read = seq_read, + .write = sn2_ptc_proc_write, .llseek = seq_lseek, .release = seq_release, }; diff --git a/arch/ia64/sn/kernel/xpc_main.c b/arch/ia64/sn/kernel/xpc_main.c index 68355ef6f841..e336e1692a73 100644 --- a/arch/ia64/sn/kernel/xpc_main.c +++ b/arch/ia64/sn/kernel/xpc_main.c @@ -55,9 +55,9 @@ #include <linux/delay.h> #include <linux/reboot.h> #include <linux/completion.h> +#include <linux/kdebug.h> #include <asm/sn/intr.h> #include <asm/sn/sn_sal.h> -#include <asm/kdebug.h> #include <asm/uaccess.h> #include <asm/sn/xpc.h> @@ -1332,7 +1332,7 @@ xpc_init(void) dev_warn(xpc_part, "can't register reboot notifier\n"); } - /* add ourselves to the die_notifier list (i.e., ia64die_chain) */ + /* add ourselves to the die_notifier list */ ret = register_die_notifier(&xpc_die_notifier); if (ret != 0) { dev_warn(xpc_part, "can't register die notifier\n"); |