From 93b1eab3d29e7ea32ee583de3362da84db06ded8 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:29 -0700 Subject: [PATCH 01/12] paravirt: refactor struct paravirt_ops into smaller pv_*_ops This patch refactors the paravirt_ops structure into groups of functionally related ops: pv_info - random info, rather than function entrypoints pv_init_ops - functions used at boot time (some for module_init too) pv_misc_ops - lazy mode, which didn't fit well anywhere else pv_time_ops - time-related functions pv_cpu_ops - various privileged instruction ops pv_irq_ops - operations for managing interrupt state pv_apic_ops - APIC operations pv_mmu_ops - operations for managing pagetables There are several motivations for this: 1. Some of these ops will be general to all x86, and some will be i386/x86-64 specific. This makes it easier to share common stuff while allowing separate implementations where needed. 2. At the moment we must export all of paravirt_ops, but modules only need selected parts of it. This allows us to export on a case by case basis (and also choose which export license we want to apply). 3. Functional groupings make things a bit more readable. Struct paravirt_ops is now only used as a template to generate patch-site identifiers, and to extract function pointers for inserting into jmp/calls when patching. It is only instantiated when needed. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Rusty Russell Cc: Andi Kleen Cc: Zach Amsden Cc: Avi Kivity Cc: Anthony Liguory Cc: "Glauber de Oliveira Costa" Cc: Jun Nakajima --- arch/x86/kernel/alternative.c | 4 +- arch/x86/kernel/asm-offsets_32.c | 14 +- arch/x86/kernel/entry_32.S | 2 +- arch/x86/kernel/paravirt_32.c | 180 ++++++---- arch/x86/kernel/vmi_32.c | 158 ++++----- arch/x86/xen/enlighten.c | 98 ++++-- drivers/char/hvc_lguest.c | 2 +- drivers/lguest/core.c | 6 +- drivers/lguest/lguest.c | 124 ++++--- drivers/lguest/lguest_bus.c | 2 +- include/asm-x86/paravirt.h | 463 ++++++++++++++------------ include/asm-x86/pgtable-3level-defs.h | 2 +- 12 files changed, 590 insertions(+), 465 deletions(-) diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index bd72d94e713..63c55148dd0 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -368,8 +368,8 @@ void apply_paravirt(struct paravirt_patch_site *start, BUG_ON(p->len > MAX_PATCH_LEN); /* prep the buffer with the original instructions */ memcpy(insnbuf, p->instr, p->len); - used = paravirt_ops.patch(p->instrtype, p->clobbers, insnbuf, - (unsigned long)p->instr, p->len); + used = pv_init_ops.patch(p->instrtype, p->clobbers, insnbuf, + (unsigned long)p->instr, p->len); BUG_ON(used > p->len); diff --git a/arch/x86/kernel/asm-offsets_32.c b/arch/x86/kernel/asm-offsets_32.c index 8029742c0fc..f1b7cdda82b 100644 --- a/arch/x86/kernel/asm-offsets_32.c +++ b/arch/x86/kernel/asm-offsets_32.c @@ -116,12 +116,14 @@ void foo(void) #ifdef CONFIG_PARAVIRT BLANK(); - OFFSET(PARAVIRT_enabled, paravirt_ops, paravirt_enabled); - OFFSET(PARAVIRT_irq_disable, paravirt_ops, irq_disable); - OFFSET(PARAVIRT_irq_enable, paravirt_ops, irq_enable); - OFFSET(PARAVIRT_irq_enable_sysexit, paravirt_ops, irq_enable_sysexit); - OFFSET(PARAVIRT_iret, paravirt_ops, iret); - OFFSET(PARAVIRT_read_cr0, paravirt_ops, read_cr0); + OFFSET(PARAVIRT_enabled, pv_info, paravirt_enabled); + OFFSET(PARAVIRT_PATCH_pv_cpu_ops, paravirt_patch_template, pv_cpu_ops); + OFFSET(PARAVIRT_PATCH_pv_irq_ops, paravirt_patch_template, pv_irq_ops); + OFFSET(PV_IRQ_irq_disable, pv_irq_ops, irq_disable); + OFFSET(PV_IRQ_irq_enable, pv_irq_ops, irq_enable); + OFFSET(PV_CPU_iret, pv_cpu_ops, iret); + OFFSET(PV_CPU_irq_enable_sysexit, pv_cpu_ops, irq_enable_sysexit); + OFFSET(PV_CPU_read_cr0, pv_cpu_ops, read_cr0); #endif #ifdef CONFIG_XEN diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S index 290b7bc82da..1f2062e94d8 100644 --- a/arch/x86/kernel/entry_32.S +++ b/arch/x86/kernel/entry_32.S @@ -434,7 +434,7 @@ ldt_ss: * is still available to implement the setting of the high * 16-bits in the INTERRUPT_RETURN paravirt-op. */ - cmpl $0, paravirt_ops+PARAVIRT_enabled + cmpl $0, pv_info+PARAVIRT_enabled jne restore_nocheck #endif diff --git a/arch/x86/kernel/paravirt_32.c b/arch/x86/kernel/paravirt_32.c index 739cfb207dd..fa412515af7 100644 --- a/arch/x86/kernel/paravirt_32.c +++ b/arch/x86/kernel/paravirt_32.c @@ -42,32 +42,33 @@ void _paravirt_nop(void) static void __init default_banner(void) { printk(KERN_INFO "Booting paravirtualized kernel on %s\n", - paravirt_ops.name); + pv_info.name); } char *memory_setup(void) { - return paravirt_ops.memory_setup(); + return pv_init_ops.memory_setup(); } /* Simple instruction patching code. */ -#define DEF_NATIVE(name, code) \ - extern const char start_##name[], end_##name[]; \ - asm("start_" #name ": " code "; end_" #name ":") +#define DEF_NATIVE(ops, name, code) \ + extern const char start_##ops##_##name[], end_##ops##_##name[]; \ + asm("start_" #ops "_" #name ": " code "; end_" #ops "_" #name ":") -DEF_NATIVE(irq_disable, "cli"); -DEF_NATIVE(irq_enable, "sti"); -DEF_NATIVE(restore_fl, "push %eax; popf"); -DEF_NATIVE(save_fl, "pushf; pop %eax"); -DEF_NATIVE(iret, "iret"); -DEF_NATIVE(irq_enable_sysexit, "sti; sysexit"); -DEF_NATIVE(read_cr2, "mov %cr2, %eax"); -DEF_NATIVE(write_cr3, "mov %eax, %cr3"); -DEF_NATIVE(read_cr3, "mov %cr3, %eax"); -DEF_NATIVE(clts, "clts"); -DEF_NATIVE(read_tsc, "rdtsc"); +DEF_NATIVE(pv_irq_ops, irq_disable, "cli"); +DEF_NATIVE(pv_irq_ops, irq_enable, "sti"); +DEF_NATIVE(pv_irq_ops, restore_fl, "push %eax; popf"); +DEF_NATIVE(pv_irq_ops, save_fl, "pushf; pop %eax"); +DEF_NATIVE(pv_cpu_ops, iret, "iret"); +DEF_NATIVE(pv_cpu_ops, irq_enable_sysexit, "sti; sysexit"); +DEF_NATIVE(pv_mmu_ops, read_cr2, "mov %cr2, %eax"); +DEF_NATIVE(pv_mmu_ops, write_cr3, "mov %eax, %cr3"); +DEF_NATIVE(pv_mmu_ops, read_cr3, "mov %cr3, %eax"); +DEF_NATIVE(pv_cpu_ops, clts, "clts"); +DEF_NATIVE(pv_cpu_ops, read_tsc, "rdtsc"); -DEF_NATIVE(ud2a, "ud2a"); +/* Undefined instruction for dealing with missing ops pointers. */ +static const unsigned char ud2a[] = { 0x0f, 0x0b }; static unsigned native_patch(u8 type, u16 clobbers, void *ibuf, unsigned long addr, unsigned len) @@ -76,37 +77,29 @@ static unsigned native_patch(u8 type, u16 clobbers, void *ibuf, unsigned ret; switch(type) { -#define SITE(x) case PARAVIRT_PATCH(x): start = start_##x; end = end_##x; goto patch_site - SITE(irq_disable); - SITE(irq_enable); - SITE(restore_fl); - SITE(save_fl); - SITE(iret); - SITE(irq_enable_sysexit); - SITE(read_cr2); - SITE(read_cr3); - SITE(write_cr3); - SITE(clts); - SITE(read_tsc); +#define SITE(ops, x) \ + case PARAVIRT_PATCH(ops.x): \ + start = start_##ops##_##x; \ + end = end_##ops##_##x; \ + goto patch_site + + SITE(pv_irq_ops, irq_disable); + SITE(pv_irq_ops, irq_enable); + SITE(pv_irq_ops, restore_fl); + SITE(pv_irq_ops, save_fl); + SITE(pv_cpu_ops, iret); + SITE(pv_cpu_ops, irq_enable_sysexit); + SITE(pv_mmu_ops, read_cr2); + SITE(pv_mmu_ops, read_cr3); + SITE(pv_mmu_ops, write_cr3); + SITE(pv_cpu_ops, clts); + SITE(pv_cpu_ops, read_tsc); #undef SITE patch_site: ret = paravirt_patch_insns(ibuf, len, start, end); break; - case PARAVIRT_PATCH(make_pgd): - case PARAVIRT_PATCH(make_pte): - case PARAVIRT_PATCH(pgd_val): - case PARAVIRT_PATCH(pte_val): -#ifdef CONFIG_X86_PAE - case PARAVIRT_PATCH(make_pmd): - case PARAVIRT_PATCH(pmd_val): -#endif - /* These functions end up returning exactly what - they're passed, in the same registers. */ - ret = paravirt_patch_nop(); - break; - default: ret = paravirt_patch_default(type, clobbers, ibuf, addr, len); break; @@ -150,7 +143,7 @@ unsigned paravirt_patch_call(void *insnbuf, return 5; } -unsigned paravirt_patch_jmp(const void *target, void *insnbuf, +unsigned paravirt_patch_jmp(void *insnbuf, const void *target, unsigned long addr, unsigned len) { struct branch *b = insnbuf; @@ -165,22 +158,38 @@ unsigned paravirt_patch_jmp(const void *target, void *insnbuf, return 5; } +/* Neat trick to map patch type back to the call within the + * corresponding structure. */ +static void *get_call_destination(u8 type) +{ + struct paravirt_patch_template tmpl = { + .pv_init_ops = pv_init_ops, + .pv_misc_ops = pv_misc_ops, + .pv_time_ops = pv_time_ops, + .pv_cpu_ops = pv_cpu_ops, + .pv_irq_ops = pv_irq_ops, + .pv_apic_ops = pv_apic_ops, + .pv_mmu_ops = pv_mmu_ops, + }; + return *((void **)&tmpl + type); +} + unsigned paravirt_patch_default(u8 type, u16 clobbers, void *insnbuf, unsigned long addr, unsigned len) { - void *opfunc = *((void **)¶virt_ops + type); + void *opfunc = get_call_destination(type); unsigned ret; if (opfunc == NULL) /* If there's no function, patch it with a ud2a (BUG) */ - ret = paravirt_patch_insns(insnbuf, len, start_ud2a, end_ud2a); + ret = paravirt_patch_insns(insnbuf, len, ud2a, ud2a+sizeof(ud2a)); else if (opfunc == paravirt_nop) /* If the operation is a nop, then nop the callsite */ ret = paravirt_patch_nop(); - else if (type == PARAVIRT_PATCH(iret) || - type == PARAVIRT_PATCH(irq_enable_sysexit)) + else if (type == PARAVIRT_PATCH(pv_cpu_ops.iret) || + type == PARAVIRT_PATCH(pv_cpu_ops.irq_enable_sysexit)) /* If operation requires a jmp, then jmp */ - ret = paravirt_patch_jmp(opfunc, insnbuf, addr, len); + ret = paravirt_patch_jmp(insnbuf, opfunc, addr, len); else /* Otherwise call the function; assume target could clobber any caller-save reg */ @@ -205,7 +214,7 @@ unsigned paravirt_patch_insns(void *insnbuf, unsigned len, void init_IRQ(void) { - paravirt_ops.init_IRQ(); + pv_irq_ops.init_IRQ(); } static void native_flush_tlb(void) @@ -233,7 +242,7 @@ extern void native_irq_enable_sysexit(void); static int __init print_banner(void) { - paravirt_ops.banner(); + pv_init_ops.banner(); return 0; } core_initcall(print_banner); @@ -273,47 +282,53 @@ int paravirt_disable_iospace(void) return ret; } -struct paravirt_ops paravirt_ops = { +struct pv_info pv_info = { .name = "bare hardware", .paravirt_enabled = 0, .kernel_rpl = 0, .shared_kernel_pmd = 1, /* Only used when CONFIG_X86_PAE is set */ +}; - .patch = native_patch, +struct pv_init_ops pv_init_ops = { + .patch = native_patch, .banner = default_banner, .arch_setup = paravirt_nop, .memory_setup = machine_specific_memory_setup, +}; + +struct pv_time_ops pv_time_ops = { + .time_init = hpet_time_init, .get_wallclock = native_get_wallclock, .set_wallclock = native_set_wallclock, - .time_init = hpet_time_init, - .init_IRQ = native_init_IRQ, + .sched_clock = native_sched_clock, + .get_cpu_khz = native_calculate_cpu_khz, +}; - .cpuid = native_cpuid, - .get_debugreg = native_get_debugreg, - .set_debugreg = native_set_debugreg, - .clts = native_clts, - .read_cr0 = native_read_cr0, - .write_cr0 = native_write_cr0, - .read_cr2 = native_read_cr2, - .write_cr2 = native_write_cr2, - .read_cr3 = native_read_cr3, - .write_cr3 = native_write_cr3, - .read_cr4 = native_read_cr4, - .read_cr4_safe = native_read_cr4_safe, - .write_cr4 = native_write_cr4, +struct pv_irq_ops pv_irq_ops = { + .init_IRQ = native_init_IRQ, .save_fl = native_save_fl, .restore_fl = native_restore_fl, .irq_disable = native_irq_disable, .irq_enable = native_irq_enable, .safe_halt = native_safe_halt, .halt = native_halt, +}; + +struct pv_cpu_ops pv_cpu_ops = { + .cpuid = native_cpuid, + .get_debugreg = native_get_debugreg, + .set_debugreg = native_set_debugreg, + .clts = native_clts, + .read_cr0 = native_read_cr0, + .write_cr0 = native_write_cr0, + .read_cr4 = native_read_cr4, + .read_cr4_safe = native_read_cr4_safe, + .write_cr4 = native_write_cr4, .wbinvd = native_wbinvd, .read_msr = native_read_msr_safe, .write_msr = native_write_msr_safe, .read_tsc = native_read_tsc, .read_pmc = native_read_pmc, - .sched_clock = native_sched_clock, - .get_cpu_khz = native_calculate_cpu_khz, .load_tr_desc = native_load_tr_desc, .set_ldt = native_set_ldt, .load_gdt = native_load_gdt, @@ -327,9 +342,14 @@ struct paravirt_ops paravirt_ops = { .write_idt_entry = write_dt_entry, .load_esp0 = native_load_esp0, + .irq_enable_sysexit = native_irq_enable_sysexit, + .iret = native_iret, + .set_iopl_mask = native_set_iopl_mask, .io_delay = native_io_delay, +}; +struct pv_apic_ops pv_apic_ops = { #ifdef CONFIG_X86_LOCAL_APIC .apic_write = native_apic_write, .apic_write_atomic = native_apic_write_atomic, @@ -338,11 +358,21 @@ struct paravirt_ops paravirt_ops = { .setup_secondary_clock = setup_secondary_APIC_clock, .startup_ipi_hook = paravirt_nop, #endif - .set_lazy_mode = paravirt_nop, +}; +struct pv_misc_ops pv_misc_ops = { + .set_lazy_mode = paravirt_nop, +}; + +struct pv_mmu_ops pv_mmu_ops = { .pagetable_setup_start = native_pagetable_setup_start, .pagetable_setup_done = native_pagetable_setup_done, + .read_cr2 = native_read_cr2, + .write_cr2 = native_write_cr2, + .read_cr3 = native_read_cr3, + .write_cr3 = native_write_cr3, + .flush_tlb_user = native_flush_tlb, .flush_tlb_kernel = native_flush_tlb_global, .flush_tlb_single = native_flush_tlb_single, @@ -381,12 +411,14 @@ struct paravirt_ops paravirt_ops = { .make_pte = native_make_pte, .make_pgd = native_make_pgd, - .irq_enable_sysexit = native_irq_enable_sysexit, - .iret = native_iret, - .dup_mmap = paravirt_nop, .exit_mmap = paravirt_nop, .activate_mm = paravirt_nop, }; -EXPORT_SYMBOL(paravirt_ops); +EXPORT_SYMBOL_GPL(pv_time_ops); +EXPORT_SYMBOL_GPL(pv_cpu_ops); +EXPORT_SYMBOL_GPL(pv_mmu_ops); +EXPORT_SYMBOL_GPL(pv_apic_ops); +EXPORT_SYMBOL_GPL(pv_info); +EXPORT_SYMBOL (pv_irq_ops); diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c index 18673e0f193..67cea5c2e3e 100644 --- a/arch/x86/kernel/vmi_32.c +++ b/arch/x86/kernel/vmi_32.c @@ -134,21 +134,21 @@ static unsigned vmi_patch(u8 type, u16 clobbers, void *insns, unsigned long eip, unsigned len) { switch (type) { - case PARAVIRT_PATCH(irq_disable): + case PARAVIRT_PATCH(pv_irq_ops.irq_disable): return patch_internal(VMI_CALL_DisableInterrupts, len, insns, eip); - case PARAVIRT_PATCH(irq_enable): + case PARAVIRT_PATCH(pv_irq_ops.irq_enable): return patch_internal(VMI_CALL_EnableInterrupts, len, insns, eip); - case PARAVIRT_PATCH(restore_fl): + case PARAVIRT_PATCH(pv_irq_ops.restore_fl): return patch_internal(VMI_CALL_SetInterruptMask, len, insns, eip); - case PARAVIRT_PATCH(save_fl): + case PARAVIRT_PATCH(pv_irq_ops.save_fl): return patch_internal(VMI_CALL_GetInterruptMask, len, insns, eip); - case PARAVIRT_PATCH(iret): + case PARAVIRT_PATCH(pv_cpu_ops.iret): return patch_internal(VMI_CALL_IRET, len, insns, eip); - case PARAVIRT_PATCH(irq_enable_sysexit): + case PARAVIRT_PATCH(pv_cpu_ops.irq_enable_sysexit): return patch_internal(VMI_CALL_SYSEXIT, len, insns, eip); default: break; @@ -690,9 +690,9 @@ do { \ reloc = call_vrom_long_func(vmi_rom, get_reloc, \ VMI_CALL_##vmicall); \ if (rel->type == VMI_RELOCATION_CALL_REL) \ - paravirt_ops.opname = (void *)rel->eip; \ + opname = (void *)rel->eip; \ else if (rel->type == VMI_RELOCATION_NOP) \ - paravirt_ops.opname = (void *)vmi_nop; \ + opname = (void *)vmi_nop; \ else if (rel->type != VMI_RELOCATION_NONE) \ printk(KERN_WARNING "VMI: Unknown relocation " \ "type %d for " #vmicall"\n",\ @@ -712,7 +712,7 @@ do { \ VMI_CALL_##vmicall); \ BUG_ON(rel->type == VMI_RELOCATION_JUMP_REL); \ if (rel->type == VMI_RELOCATION_CALL_REL) { \ - paravirt_ops.opname = wrapper; \ + opname = wrapper; \ vmi_ops.cache = (void *)rel->eip; \ } \ } while (0) @@ -732,11 +732,11 @@ static inline int __init activate_vmi(void) } savesegment(cs, kernel_cs); - paravirt_ops.paravirt_enabled = 1; - paravirt_ops.kernel_rpl = kernel_cs & SEGMENT_RPL_MASK; + pv_info.paravirt_enabled = 1; + pv_info.kernel_rpl = kernel_cs & SEGMENT_RPL_MASK; + pv_info.name = "vmi"; - paravirt_ops.patch = vmi_patch; - paravirt_ops.name = "vmi"; + pv_init_ops.patch = vmi_patch; /* * Many of these operations are ABI compatible with VMI. @@ -754,26 +754,26 @@ static inline int __init activate_vmi(void) */ /* CPUID is special, so very special it gets wrapped like a present */ - para_wrap(cpuid, vmi_cpuid, cpuid, CPUID); + para_wrap(pv_cpu_ops.cpuid, vmi_cpuid, cpuid, CPUID); - para_fill(clts, CLTS); - para_fill(get_debugreg, GetDR); - para_fill(set_debugreg, SetDR); - para_fill(read_cr0, GetCR0); - para_fill(read_cr2, GetCR2); - para_fill(read_cr3, GetCR3); - para_fill(read_cr4, GetCR4); - para_fill(write_cr0, SetCR0); - para_fill(write_cr2, SetCR2); - para_fill(write_cr3, SetCR3); - para_fill(write_cr4, SetCR4); - para_fill(save_fl, GetInterruptMask); - para_fill(restore_fl, SetInterruptMask); - para_fill(irq_disable, DisableInterrupts); - para_fill(irq_enable, EnableInterrupts); + para_fill(pv_cpu_ops.clts, CLTS); + para_fill(pv_cpu_ops.get_debugreg, GetDR); + para_fill(pv_cpu_ops.set_debugreg, SetDR); + para_fill(pv_cpu_ops.read_cr0, GetCR0); + para_fill(pv_mmu_ops.read_cr2, GetCR2); + para_fill(pv_mmu_ops.read_cr3, GetCR3); + para_fill(pv_cpu_ops.read_cr4, GetCR4); + para_fill(pv_cpu_ops.write_cr0, SetCR0); + para_fill(pv_mmu_ops.write_cr2, SetCR2); + para_fill(pv_mmu_ops.write_cr3, SetCR3); + para_fill(pv_cpu_ops.write_cr4, SetCR4); + para_fill(pv_irq_ops.save_fl, GetInterruptMask); + para_fill(pv_irq_ops.restore_fl, SetInterruptMask); + para_fill(pv_irq_ops.irq_disable, DisableInterrupts); + para_fill(pv_irq_ops.irq_enable, EnableInterrupts); - para_fill(wbinvd, WBINVD); - para_fill(read_tsc, RDTSC); + para_fill(pv_cpu_ops.wbinvd, WBINVD); + para_fill(pv_cpu_ops.read_tsc, RDTSC); /* The following we emulate with trap and emulate for now */ /* paravirt_ops.read_msr = vmi_rdmsr */ @@ -781,29 +781,29 @@ static inline int __init activate_vmi(void) /* paravirt_ops.rdpmc = vmi_rdpmc */ /* TR interface doesn't pass TR value, wrap */ - para_wrap(load_tr_desc, vmi_set_tr, set_tr, SetTR); + para_wrap(pv_cpu_ops.load_tr_desc, vmi_set_tr, set_tr, SetTR); /* LDT is special, too */ - para_wrap(set_ldt, vmi_set_ldt, _set_ldt, SetLDT); + para_wrap(pv_cpu_ops.set_ldt, vmi_set_ldt, _set_ldt, SetLDT); - para_fill(load_gdt, SetGDT); - para_fill(load_idt, SetIDT); - para_fill(store_gdt, GetGDT); - para_fill(store_idt, GetIDT); - para_fill(store_tr, GetTR); - paravirt_ops.load_tls = vmi_load_tls; - para_fill(write_ldt_entry, WriteLDTEntry); - para_fill(write_gdt_entry, WriteGDTEntry); - para_fill(write_idt_entry, WriteIDTEntry); - para_wrap(load_esp0, vmi_load_esp0, set_kernel_stack, UpdateKernelStack); - para_fill(set_iopl_mask, SetIOPLMask); - para_fill(io_delay, IODelay); - para_wrap(set_lazy_mode, vmi_set_lazy_mode, set_lazy_mode, SetLazyMode); + para_fill(pv_cpu_ops.load_gdt, SetGDT); + para_fill(pv_cpu_ops.load_idt, SetIDT); + para_fill(pv_cpu_ops.store_gdt, GetGDT); + para_fill(pv_cpu_ops.store_idt, GetIDT); + para_fill(pv_cpu_ops.store_tr, GetTR); + pv_cpu_ops.load_tls = vmi_load_tls; + para_fill(pv_cpu_ops.write_ldt_entry, WriteLDTEntry); + para_fill(pv_cpu_ops.write_gdt_entry, WriteGDTEntry); + para_fill(pv_cpu_ops.write_idt_entry, WriteIDTEntry); + para_wrap(pv_cpu_ops.load_esp0, vmi_load_esp0, set_kernel_stack, UpdateKernelStack); + para_fill(pv_cpu_ops.set_iopl_mask, SetIOPLMask); + para_fill(pv_cpu_ops.io_delay, IODelay); + para_wrap(pv_misc_ops.set_lazy_mode, vmi_set_lazy_mode, set_lazy_mode, SetLazyMode); /* user and kernel flush are just handled with different flags to FlushTLB */ - para_wrap(flush_tlb_user, vmi_flush_tlb_user, _flush_tlb, FlushTLB); - para_wrap(flush_tlb_kernel, vmi_flush_tlb_kernel, _flush_tlb, FlushTLB); - para_fill(flush_tlb_single, InvalPage); + para_wrap(pv_mmu_ops.flush_tlb_user, vmi_flush_tlb_user, _flush_tlb, FlushTLB); + para_wrap(pv_mmu_ops.flush_tlb_kernel, vmi_flush_tlb_kernel, _flush_tlb, FlushTLB); + para_fill(pv_mmu_ops.flush_tlb_single, InvalPage); /* * Until a standard flag format can be agreed on, we need to @@ -819,41 +819,41 @@ static inline int __init activate_vmi(void) #endif if (vmi_ops.set_pte) { - paravirt_ops.set_pte = vmi_set_pte; - paravirt_ops.set_pte_at = vmi_set_pte_at; - paravirt_ops.set_pmd = vmi_set_pmd; + pv_mmu_ops.set_pte = vmi_set_pte; + pv_mmu_ops.set_pte_at = vmi_set_pte_at; + pv_mmu_ops.set_pmd = vmi_set_pmd; #ifdef CONFIG_X86_PAE - paravirt_ops.set_pte_atomic = vmi_set_pte_atomic; - paravirt_ops.set_pte_present = vmi_set_pte_present; - paravirt_ops.set_pud = vmi_set_pud; - paravirt_ops.pte_clear = vmi_pte_clear; - paravirt_ops.pmd_clear = vmi_pmd_clear; + pv_mmu_ops.set_pte_atomic = vmi_set_pte_atomic; + pv_mmu_ops.set_pte_present = vmi_set_pte_present; + pv_mmu_ops.set_pud = vmi_set_pud; + pv_mmu_ops.pte_clear = vmi_pte_clear; + pv_mmu_ops.pmd_clear = vmi_pmd_clear; #endif } if (vmi_ops.update_pte) { - paravirt_ops.pte_update = vmi_update_pte; - paravirt_ops.pte_update_defer = vmi_update_pte_defer; + pv_mmu_ops.pte_update = vmi_update_pte; + pv_mmu_ops.pte_update_defer = vmi_update_pte_defer; } vmi_ops.allocate_page = vmi_get_function(VMI_CALL_AllocatePage); if (vmi_ops.allocate_page) { - paravirt_ops.alloc_pt = vmi_allocate_pt; - paravirt_ops.alloc_pd = vmi_allocate_pd; - paravirt_ops.alloc_pd_clone = vmi_allocate_pd_clone; + pv_mmu_ops.alloc_pt = vmi_allocate_pt; + pv_mmu_ops.alloc_pd = vmi_allocate_pd; + pv_mmu_ops.alloc_pd_clone = vmi_allocate_pd_clone; } vmi_ops.release_page = vmi_get_function(VMI_CALL_ReleasePage); if (vmi_ops.release_page) { - paravirt_ops.release_pt = vmi_release_pt; - paravirt_ops.release_pd = vmi_release_pd; + pv_mmu_ops.release_pt = vmi_release_pt; + pv_mmu_ops.release_pd = vmi_release_pd; } /* Set linear is needed in all cases */ vmi_ops.set_linear_mapping = vmi_get_function(VMI_CALL_SetLinearMapping); #ifdef CONFIG_HIGHPTE if (vmi_ops.set_linear_mapping) - paravirt_ops.kmap_atomic_pte = vmi_kmap_atomic_pte; + pv_mmu_ops.kmap_atomic_pte = vmi_kmap_atomic_pte; #endif /* @@ -863,17 +863,17 @@ static inline int __init activate_vmi(void) * the backend. They are performance critical anyway, so requiring * a patch is not a big problem. */ - paravirt_ops.irq_enable_sysexit = (void *)0xfeedbab0; - paravirt_ops.iret = (void *)0xbadbab0; + pv_cpu_ops.irq_enable_sysexit = (void *)0xfeedbab0; + pv_cpu_ops.iret = (void *)0xbadbab0; #ifdef CONFIG_SMP - para_wrap(startup_ipi_hook, vmi_startup_ipi_hook, set_initial_ap_state, SetInitialAPState); + para_wrap(pv_apic_ops.startup_ipi_hook, vmi_startup_ipi_hook, set_initial_ap_state, SetInitialAPState); #endif #ifdef CONFIG_X86_LOCAL_APIC - para_fill(apic_read, APICRead); - para_fill(apic_write, APICWrite); - para_fill(apic_write_atomic, APICWrite); + para_fill(pv_apic_ops.apic_read, APICRead); + para_fill(pv_apic_ops.apic_write, APICWrite); + para_fill(pv_apic_ops.apic_write_atomic, APICWrite); #endif /* @@ -891,15 +891,15 @@ static inline int __init activate_vmi(void) vmi_timer_ops.set_alarm = vmi_get_function(VMI_CALL_SetAlarm); vmi_timer_ops.cancel_alarm = vmi_get_function(VMI_CALL_CancelAlarm); - paravirt_ops.time_init = vmi_time_init; - paravirt_ops.get_wallclock = vmi_get_wallclock; - paravirt_ops.set_wallclock = vmi_set_wallclock; + pv_time_ops.time_init = vmi_time_init; + pv_time_ops.get_wallclock = vmi_get_wallclock; + pv_time_ops.set_wallclock = vmi_set_wallclock; #ifdef CONFIG_X86_LOCAL_APIC - paravirt_ops.setup_boot_clock = vmi_time_bsp_init; - paravirt_ops.setup_secondary_clock = vmi_time_ap_init; + pv_apic_ops.setup_boot_clock = vmi_time_bsp_init; + pv_apic_ops.setup_secondary_clock = vmi_time_ap_init; #endif - paravirt_ops.sched_clock = vmi_sched_clock; - paravirt_ops.get_cpu_khz = vmi_cpu_khz; + pv_time_ops.sched_clock = vmi_sched_clock; + pv_time_ops.get_cpu_khz = vmi_cpu_khz; /* We have true wallclock functions; disable CMOS clock sync */ no_sync_cmos_clock = 1; @@ -908,7 +908,7 @@ static inline int __init activate_vmi(void) disable_vmi_timer = 1; } - para_fill(safe_halt, Halt); + para_fill(pv_irq_ops.safe_halt, Halt); /* * Alternative instruction rewriting doesn't happen soon enough diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index f01bfcd4bde..3d3bf05dec7 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -124,7 +124,7 @@ static void __init xen_vcpu_setup(int cpu) static void __init xen_banner(void) { printk(KERN_INFO "Booting paravirtualized kernel on %s\n", - paravirt_ops.name); + pv_info.name); printk(KERN_INFO "Hypervisor signature: %s\n", xen_start_info->magic); } @@ -738,7 +738,7 @@ static __init void xen_pagetable_setup_start(pgd_t *base) pgd_t *xen_pgd = (pgd_t *)xen_start_info->pt_base; /* special set_pte for pagetable initialization */ - paravirt_ops.set_pte = xen_set_pte_init; + pv_mmu_ops.set_pte = xen_set_pte_init; init_mm.pgd = base; /* @@ -785,8 +785,8 @@ static __init void xen_pagetable_setup_done(pgd_t *base) { /* This will work as long as patching hasn't happened yet (which it hasn't) */ - paravirt_ops.alloc_pt = xen_alloc_pt; - paravirt_ops.set_pte = xen_set_pte; + pv_mmu_ops.alloc_pt = xen_alloc_pt; + pv_mmu_ops.set_pte = xen_set_pte; if (!xen_feature(XENFEAT_auto_translated_physmap)) { /* @@ -833,12 +833,12 @@ void __init xen_setup_vcpu_info_placement(void) if (have_vcpu_info_placement) { printk(KERN_INFO "Xen: using vcpu_info placement\n"); - paravirt_ops.save_fl = xen_save_fl_direct; - paravirt_ops.restore_fl = xen_restore_fl_direct; - paravirt_ops.irq_disable = xen_irq_disable_direct; - paravirt_ops.irq_enable = xen_irq_enable_direct; - paravirt_ops.read_cr2 = xen_read_cr2_direct; - paravirt_ops.iret = xen_iret_direct; + pv_irq_ops.save_fl = xen_save_fl_direct; + pv_irq_ops.restore_fl = xen_restore_fl_direct; + pv_irq_ops.irq_disable = xen_irq_disable_direct; + pv_irq_ops.irq_enable = xen_irq_enable_direct; + pv_mmu_ops.read_cr2 = xen_read_cr2_direct; + pv_cpu_ops.iret = xen_iret_direct; } } @@ -850,8 +850,8 @@ static unsigned xen_patch(u8 type, u16 clobbers, void *insnbuf, start = end = reloc = NULL; -#define SITE(x) \ - case PARAVIRT_PATCH(x): \ +#define SITE(op, x) \ + case PARAVIRT_PATCH(op.x): \ if (have_vcpu_info_placement) { \ start = (char *)xen_##x##_direct; \ end = xen_##x##_direct_end; \ @@ -860,10 +860,10 @@ static unsigned xen_patch(u8 type, u16 clobbers, void *insnbuf, goto patch_site switch (type) { - SITE(irq_enable); - SITE(irq_disable); - SITE(save_fl); - SITE(restore_fl); + SITE(pv_irq_ops, irq_enable); + SITE(pv_irq_ops, irq_disable); + SITE(pv_irq_ops, save_fl); + SITE(pv_irq_ops, restore_fl); #undef SITE patch_site: @@ -895,26 +895,32 @@ static unsigned xen_patch(u8 type, u16 clobbers, void *insnbuf, return ret; } -static const struct paravirt_ops xen_paravirt_ops __initdata = { +static const struct pv_info xen_info __initdata = { .paravirt_enabled = 1, .shared_kernel_pmd = 0, .name = "Xen", - .banner = xen_banner, +}; +static const struct pv_init_ops xen_init_ops __initdata = { .patch = xen_patch, + .banner = xen_banner, .memory_setup = xen_memory_setup, .arch_setup = xen_arch_setup, - .init_IRQ = xen_init_IRQ, .post_allocator_init = xen_mark_init_mm_pinned, +}; +static const struct pv_time_ops xen_time_ops __initdata = { .time_init = xen_time_init, + .set_wallclock = xen_set_wallclock, .get_wallclock = xen_get_wallclock, .get_cpu_khz = xen_cpu_khz, .sched_clock = xen_sched_clock, +}; +static const struct pv_cpu_ops xen_cpu_ops __initdata = { .cpuid = xen_cpuid, .set_debugreg = xen_set_debugreg, @@ -925,22 +931,10 @@ static const struct paravirt_ops xen_paravirt_ops __initdata = { .read_cr0 = native_read_cr0, .write_cr0 = native_write_cr0, - .read_cr2 = xen_read_cr2, - .write_cr2 = xen_write_cr2, - - .read_cr3 = xen_read_cr3, - .write_cr3 = xen_write_cr3, - .read_cr4 = native_read_cr4, .read_cr4_safe = native_read_cr4_safe, .write_cr4 = xen_write_cr4, - .save_fl = xen_save_fl, - .restore_fl = xen_restore_fl, - .irq_disable = xen_irq_disable, - .irq_enable = xen_irq_enable, - .safe_halt = xen_safe_halt, - .halt = xen_halt, .wbinvd = native_wbinvd, .read_msr = native_read_msr_safe, @@ -968,7 +962,19 @@ static const struct paravirt_ops xen_paravirt_ops __initdata = { .set_iopl_mask = xen_set_iopl_mask, .io_delay = xen_io_delay, +}; +static const struct pv_irq_ops xen_irq_ops __initdata = { + .init_IRQ = xen_init_IRQ, + .save_fl = xen_save_fl, + .restore_fl = xen_restore_fl, + .irq_disable = xen_irq_disable, + .irq_enable = xen_irq_enable, + .safe_halt = xen_safe_halt, + .halt = xen_halt, +}; + +static const struct pv_apic_ops xen_apic_ops __initdata = { #ifdef CONFIG_X86_LOCAL_APIC .apic_write = xen_apic_write, .apic_write_atomic = xen_apic_write, @@ -977,6 +983,17 @@ static const struct paravirt_ops xen_paravirt_ops __initdata = { .setup_secondary_clock = paravirt_nop, .startup_ipi_hook = paravirt_nop, #endif +}; + +static const struct pv_mmu_ops xen_mmu_ops __initdata = { + .pagetable_setup_start = xen_pagetable_setup_start, + .pagetable_setup_done = xen_pagetable_setup_done, + + .read_cr2 = xen_read_cr2, + .write_cr2 = xen_write_cr2, + + .read_cr3 = xen_read_cr3, + .write_cr3 = xen_write_cr3, .flush_tlb_user = xen_flush_tlb, .flush_tlb_kernel = xen_flush_tlb, @@ -986,9 +1003,6 @@ static const struct paravirt_ops xen_paravirt_ops __initdata = { .pte_update = paravirt_nop, .pte_update_defer = paravirt_nop, - .pagetable_setup_start = xen_pagetable_setup_start, - .pagetable_setup_done = xen_pagetable_setup_done, - .alloc_pt = xen_alloc_pt_init, .release_pt = xen_release_pt, .alloc_pd = paravirt_nop, @@ -1023,7 +1037,9 @@ static const struct paravirt_ops xen_paravirt_ops __initdata = { .activate_mm = xen_activate_mm, .dup_mmap = xen_dup_mmap, .exit_mmap = xen_exit_mmap, +}; +static const struct pv_misc_ops xen_misc_ops __initdata = { .set_lazy_mode = xen_set_lazy_mode, }; @@ -1091,7 +1107,15 @@ asmlinkage void __init xen_start_kernel(void) BUG_ON(memcmp(xen_start_info->magic, "xen-3.0", 7) != 0); /* Install Xen paravirt ops */ - paravirt_ops = xen_paravirt_ops; + pv_info = xen_info; + pv_init_ops = xen_init_ops; + pv_time_ops = xen_time_ops; + pv_cpu_ops = xen_cpu_ops; + pv_irq_ops = xen_irq_ops; + pv_apic_ops = xen_apic_ops; + pv_mmu_ops = xen_mmu_ops; + pv_misc_ops = xen_misc_ops; + machine_ops = xen_machine_ops; #ifdef CONFIG_SMP @@ -1124,9 +1148,9 @@ asmlinkage void __init xen_start_kernel(void) xen_setup_vcpu_info_placement(); #endif - paravirt_ops.kernel_rpl = 1; + pv_info.kernel_rpl = 1; if (xen_feature(XENFEAT_supervisor_mode_kernel)) - paravirt_ops.kernel_rpl = 0; + pv_info.kernel_rpl = 0; /* set the limit of our address space */ reserve_top_address(-HYPERVISOR_VIRT_START + 2 * PAGE_SIZE); diff --git a/drivers/char/hvc_lguest.c b/drivers/char/hvc_lguest.c index 3d6bd0baa56..efccb215583 100644 --- a/drivers/char/hvc_lguest.c +++ b/drivers/char/hvc_lguest.c @@ -115,7 +115,7 @@ static struct hv_ops lguest_cons = { * (0), and the struct hv_ops containing the put_chars() function. */ static int __init cons_init(void) { - if (strcmp(paravirt_ops.name, "lguest") != 0) + if (strcmp(pv_info.name, "lguest") != 0) return 0; return hvc_instantiate(0, 0, &lguest_cons); diff --git a/drivers/lguest/core.c b/drivers/lguest/core.c index 4a315f08a56..a0788c12b39 100644 --- a/drivers/lguest/core.c +++ b/drivers/lguest/core.c @@ -248,8 +248,8 @@ static void unmap_switcher(void) } /*H:130 Our Guest is usually so well behaved; it never tries to do things it - * isn't allowed to. Unfortunately, "struct paravirt_ops" isn't quite - * complete, because it doesn't contain replacements for the Intel I/O + * isn't allowed to. Unfortunately, Linux's paravirtual infrastructure isn't + * quite complete, because it doesn't contain replacements for the Intel I/O * instructions. As a result, the Guest sometimes fumbles across one during * the boot process as it probes for various things which are usually attached * to a PC. @@ -694,7 +694,7 @@ static int __init init(void) /* Lguest can't run under Xen, VMI or itself. It does Tricky Stuff. */ if (paravirt_enabled()) { - printk("lguest is afraid of %s\n", paravirt_ops.name); + printk("lguest is afraid of %s\n", pv_info.name); return -EPERM; } diff --git a/drivers/lguest/lguest.c b/drivers/lguest/lguest.c index ee1c6d05c3d..ca9b844f37c 100644 --- a/drivers/lguest/lguest.c +++ b/drivers/lguest/lguest.c @@ -23,7 +23,7 @@ * * So how does the kernel know it's a Guest? The Guest starts at a special * entry point marked with a magic string, which sets up a few things then - * calls here. We replace the native functions in "struct paravirt_ops" + * calls here. We replace the native functions various "paravirt" structures * with our Guest versions, then boot like normal. :*/ /* @@ -331,7 +331,7 @@ static void lguest_load_tls(struct thread_struct *t, unsigned int cpu) } /*G:038 That's enough excitement for now, back to ploughing through each of - * the paravirt_ops (we're about 1/3 of the way through). + * the different pv_ops structures (we're about 1/3 of the way through). * * This is the Local Descriptor Table, another weird Intel thingy. Linux only * uses this for some strange applications like Wine. We don't do anything @@ -558,7 +558,7 @@ static void lguest_set_pte(pte_t *ptep, pte_t pteval) lazy_hcall(LHCALL_FLUSH_TLB, 1, 0, 0); } -/* Unfortunately for Lguest, the paravirt_ops for page tables were based on +/* Unfortunately for Lguest, the pv_mmu_ops for page tables were based on * native page table operations. On native hardware you can set a new page * table entry whenever you want, but if you want to remove one you have to do * a TLB flush (a TLB is a little cache of page table entries kept by the CPU). @@ -782,7 +782,7 @@ static void lguest_time_init(void) clocksource_register(&lguest_clock); /* Now we've set up our clock, we can use it as the scheduler clock */ - paravirt_ops.sched_clock = lguest_sched_clock; + pv_time_ops.sched_clock = lguest_sched_clock; /* We can't set cpumask in the initializer: damn C limitations! Set it * here and register our timer device. */ @@ -902,7 +902,7 @@ static __init char *lguest_memory_setup(void) /*G:050 * Patching (Powerfully Placating Performance Pedants) * - * We have already seen that "struct paravirt_ops" lets us replace simple + * We have already seen that pv_ops structures let us replace simple * native instructions with calls to the appropriate back end all throughout * the kernel. This allows the same kernel to run as a Guest and as a native * kernel, but it's slow because of all the indirect branches. @@ -927,10 +927,10 @@ static const struct lguest_insns { const char *start, *end; } lguest_insns[] = { - [PARAVIRT_PATCH(irq_disable)] = { lgstart_cli, lgend_cli }, - [PARAVIRT_PATCH(irq_enable)] = { lgstart_sti, lgend_sti }, - [PARAVIRT_PATCH(restore_fl)] = { lgstart_popf, lgend_popf }, - [PARAVIRT_PATCH(save_fl)] = { lgstart_pushf, lgend_pushf }, + [PARAVIRT_PATCH(pv_irq_ops.irq_disable)] = { lgstart_cli, lgend_cli }, + [PARAVIRT_PATCH(pv_irq_ops.irq_enable)] = { lgstart_sti, lgend_sti }, + [PARAVIRT_PATCH(pv_irq_ops.restore_fl)] = { lgstart_popf, lgend_popf }, + [PARAVIRT_PATCH(pv_irq_ops.save_fl)] = { lgstart_pushf, lgend_pushf }, }; /* Now our patch routine is fairly simple (based on the native one in @@ -957,9 +957,9 @@ static unsigned lguest_patch(u8 type, u16 clobber, void *ibuf, return insn_len; } -/*G:030 Once we get to lguest_init(), we know we're a Guest. The paravirt_ops - * structure in the kernel provides a single point for (almost) every routine - * we have to override to avoid privileged instructions. */ +/*G:030 Once we get to lguest_init(), we know we're a Guest. The pv_ops + * structures in the kernel provide points for (almost) every routine we have + * to override to avoid privileged instructions. */ __init void lguest_init(void *boot) { /* Copy boot parameters first: the Launcher put the physical location @@ -974,54 +974,68 @@ __init void lguest_init(void *boot) /* We're under lguest, paravirt is enabled, and we're running at * privilege level 1, not 0 as normal. */ - paravirt_ops.name = "lguest"; - paravirt_ops.paravirt_enabled = 1; - paravirt_ops.kernel_rpl = 1; + pv_info.name = "lguest"; + pv_info.paravirt_enabled = 1; + pv_info.kernel_rpl = 1; /* We set up all the lguest overrides for sensitive operations. These * are detailed with the operations themselves. */ - paravirt_ops.save_fl = save_fl; - paravirt_ops.restore_fl = restore_fl; - paravirt_ops.irq_disable = irq_disable; - paravirt_ops.irq_enable = irq_enable; - paravirt_ops.load_gdt = lguest_load_gdt; - paravirt_ops.memory_setup = lguest_memory_setup; - paravirt_ops.cpuid = lguest_cpuid; - paravirt_ops.write_cr3 = lguest_write_cr3; - paravirt_ops.flush_tlb_user = lguest_flush_tlb_user; - paravirt_ops.flush_tlb_single = lguest_flush_tlb_single; - paravirt_ops.flush_tlb_kernel = lguest_flush_tlb_kernel; - paravirt_ops.set_pte = lguest_set_pte; - paravirt_ops.set_pte_at = lguest_set_pte_at; - paravirt_ops.set_pmd = lguest_set_pmd; + + /* interrupt-related operations */ + pv_irq_ops.init_IRQ = lguest_init_IRQ; + pv_irq_ops.save_fl = save_fl; + pv_irq_ops.restore_fl = restore_fl; + pv_irq_ops.irq_disable = irq_disable; + pv_irq_ops.irq_enable = irq_enable; + pv_irq_ops.safe_halt = lguest_safe_halt; + + /* init-time operations */ + pv_init_ops.memory_setup = lguest_memory_setup; + pv_init_ops.patch = lguest_patch; + + /* Intercepts of various cpu instructions */ + pv_cpu_ops.load_gdt = lguest_load_gdt; + pv_cpu_ops.cpuid = lguest_cpuid; + pv_cpu_ops.load_idt = lguest_load_idt; + pv_cpu_ops.iret = lguest_iret; + pv_cpu_ops.load_esp0 = lguest_load_esp0; + pv_cpu_ops.load_tr_desc = lguest_load_tr_desc; + pv_cpu_ops.set_ldt = lguest_set_ldt; + pv_cpu_ops.load_tls = lguest_load_tls; + pv_cpu_ops.set_debugreg = lguest_set_debugreg; + pv_cpu_ops.clts = lguest_clts; + pv_cpu_ops.read_cr0 = lguest_read_cr0; + pv_cpu_ops.write_cr0 = lguest_write_cr0; + pv_cpu_ops.read_cr4 = lguest_read_cr4; + pv_cpu_ops.write_cr4 = lguest_write_cr4; + pv_cpu_ops.write_gdt_entry = lguest_write_gdt_entry; + pv_cpu_ops.write_idt_entry = lguest_write_idt_entry; + pv_cpu_ops.wbinvd = lguest_wbinvd; + + /* pagetable management */ + pv_mmu_ops.write_cr3 = lguest_write_cr3; + pv_mmu_ops.flush_tlb_user = lguest_flush_tlb_user; + pv_mmu_ops.flush_tlb_single = lguest_flush_tlb_single; + pv_mmu_ops.flush_tlb_kernel = lguest_flush_tlb_kernel; + pv_mmu_ops.set_pte = lguest_set_pte; + pv_mmu_ops.set_pte_at = lguest_set_pte_at; + pv_mmu_ops.set_pmd = lguest_set_pmd; + pv_mmu_ops.read_cr2 = lguest_read_cr2; + pv_mmu_ops.read_cr3 = lguest_read_cr3; + #ifdef CONFIG_X86_LOCAL_APIC - paravirt_ops.apic_write = lguest_apic_write; - paravirt_ops.apic_write_atomic = lguest_apic_write; - paravirt_ops.apic_read = lguest_apic_read; + /* apic read/write intercepts */ + pv_apic_ops.apic_write = lguest_apic_write; + pv_apic_ops.apic_write_atomic = lguest_apic_write; + pv_apic_ops.apic_read = lguest_apic_read; #endif - paravirt_ops.load_idt = lguest_load_idt; - paravirt_ops.iret = lguest_iret; - paravirt_ops.load_esp0 = lguest_load_esp0; - paravirt_ops.load_tr_desc = lguest_load_tr_desc; - paravirt_ops.set_ldt = lguest_set_ldt; - paravirt_ops.load_tls = lguest_load_tls; - paravirt_ops.set_debugreg = lguest_set_debugreg; - paravirt_ops.clts = lguest_clts; - paravirt_ops.read_cr0 = lguest_read_cr0; - paravirt_ops.write_cr0 = lguest_write_cr0; - paravirt_ops.init_IRQ = lguest_init_IRQ; - paravirt_ops.read_cr2 = lguest_read_cr2; - paravirt_ops.read_cr3 = lguest_read_cr3; - paravirt_ops.read_cr4 = lguest_read_cr4; - paravirt_ops.write_cr4 = lguest_write_cr4; - paravirt_ops.write_gdt_entry = lguest_write_gdt_entry; - paravirt_ops.write_idt_entry = lguest_write_idt_entry; - paravirt_ops.patch = lguest_patch; - paravirt_ops.safe_halt = lguest_safe_halt; - paravirt_ops.get_wallclock = lguest_get_wallclock; - paravirt_ops.time_init = lguest_time_init; - paravirt_ops.set_lazy_mode = lguest_lazy_mode; - paravirt_ops.wbinvd = lguest_wbinvd; + + /* time operations */ + pv_time_ops.get_wallclock = lguest_get_wallclock; + pv_time_ops.time_init = lguest_time_init; + + pv_misc_ops.set_lazy_mode = lguest_lazy_mode; + /* Now is a good time to look at the implementations of these functions * before returning to the rest of lguest_init(). */ diff --git a/drivers/lguest/lguest_bus.c b/drivers/lguest/lguest_bus.c index 9e7752cc800..57329788f8a 100644 --- a/drivers/lguest/lguest_bus.c +++ b/drivers/lguest/lguest_bus.c @@ -201,7 +201,7 @@ static void scan_devices(void) * "struct lguest_device_desc" array. */ static int __init lguest_bus_init(void) { - if (strcmp(paravirt_ops.name, "lguest") != 0) + if (strcmp(pv_info.name, "lguest") != 0) return 0; /* Devices are in a single page above top of "normal" mem */ diff --git a/include/asm-x86/paravirt.h b/include/asm-x86/paravirt.h index 9fa3fa9e62d..19726e12051 100644 --- a/include/asm-x86/paravirt.h +++ b/include/asm-x86/paravirt.h @@ -33,19 +33,23 @@ enum paravirt_lazy_mode { PARAVIRT_LAZY_FLUSH = 3, }; -struct paravirt_ops -{ + +/* general info */ +struct pv_info { unsigned int kernel_rpl; int shared_kernel_pmd; - int paravirt_enabled; + int paravirt_enabled; const char *name; +}; +struct pv_init_ops { /* - * Patch may replace one of the defined code sequences with arbitrary - * code, subject to the same register constraints. This generally - * means the code is not free to clobber any registers other than EAX. - * The patch function should return the number of bytes of code - * generated, as we nop pad the rest in generic code. + * Patch may replace one of the defined code sequences with + * arbitrary code, subject to the same register constraints. + * This generally means the code is not free to clobber any + * registers other than EAX. The patch function should return + * the number of bytes of code generated, as we nop pad the + * rest in generic code. */ unsigned (*patch)(u8 type, u16 clobber, void *insnbuf, unsigned long addr, unsigned len); @@ -55,29 +59,28 @@ struct paravirt_ops char *(*memory_setup)(void); void (*post_allocator_init)(void); - void (*init_IRQ)(void); - void (*time_init)(void); - - /* - * Called before/after init_mm pagetable setup. setup_start - * may reset %cr3, and may pre-install parts of the pagetable; - * pagetable setup is expected to preserve any existing - * mapping. - */ - void (*pagetable_setup_start)(pgd_t *pgd_base); - void (*pagetable_setup_done)(pgd_t *pgd_base); - /* Print a banner to identify the environment */ void (*banner)(void); +}; + + +struct pv_misc_ops { + /* Set deferred update mode, used for batching operations. */ + void (*set_lazy_mode)(enum paravirt_lazy_mode mode); +}; + +struct pv_time_ops { + void (*time_init)(void); /* Set and set time of day */ unsigned long (*get_wallclock)(void); int (*set_wallclock)(unsigned long); - /* cpuid emulation, mostly so that caps bits can be disabled */ - void (*cpuid)(unsigned int *eax, unsigned int *ebx, - unsigned int *ecx, unsigned int *edx); + unsigned long long (*sched_clock)(void); + unsigned long (*get_cpu_khz)(void); +}; +struct pv_cpu_ops { /* hooks for various privileged instructions */ unsigned long (*get_debugreg)(int regno); void (*set_debugreg)(int regno, unsigned long value); @@ -87,41 +90,10 @@ struct paravirt_ops unsigned long (*read_cr0)(void); void (*write_cr0)(unsigned long); - unsigned long (*read_cr2)(void); - void (*write_cr2)(unsigned long); - - unsigned long (*read_cr3)(void); - void (*write_cr3)(unsigned long); - unsigned long (*read_cr4_safe)(void); unsigned long (*read_cr4)(void); void (*write_cr4)(unsigned long); - /* - * Get/set interrupt state. save_fl and restore_fl are only - * expected to use X86_EFLAGS_IF; all other bits - * returned from save_fl are undefined, and may be ignored by - * restore_fl. - */ - unsigned long (*save_fl)(void); - void (*restore_fl)(unsigned long); - void (*irq_disable)(void); - void (*irq_enable)(void); - void (*safe_halt)(void); - void (*halt)(void); - - void (*wbinvd)(void); - - /* MSR, PMC and TSR operations. - err = 0/-EFAULT. wrmsr returns 0/-EFAULT. */ - u64 (*read_msr)(unsigned int msr, int *err); - int (*write_msr)(unsigned int msr, u64 val); - - u64 (*read_tsc)(void); - u64 (*read_pmc)(void); - unsigned long long (*sched_clock)(void); - unsigned long (*get_cpu_khz)(void); - /* Segment descriptor handling */ void (*load_tr_desc)(void); void (*load_gdt)(const struct Xgt_desc_struct *); @@ -140,18 +112,45 @@ struct paravirt_ops void (*load_esp0)(struct tss_struct *tss, struct thread_struct *t); void (*set_iopl_mask)(unsigned mask); + + void (*wbinvd)(void); void (*io_delay)(void); - /* - * Hooks for intercepting the creation/use/destruction of an - * mm_struct. - */ - void (*activate_mm)(struct mm_struct *prev, - struct mm_struct *next); - void (*dup_mmap)(struct mm_struct *oldmm, - struct mm_struct *mm); - void (*exit_mmap)(struct mm_struct *mm); + /* cpuid emulation, mostly so that caps bits can be disabled */ + void (*cpuid)(unsigned int *eax, unsigned int *ebx, + unsigned int *ecx, unsigned int *edx); + /* MSR, PMC and TSR operations. + err = 0/-EFAULT. wrmsr returns 0/-EFAULT. */ + u64 (*read_msr)(unsigned int msr, int *err); + int (*write_msr)(unsigned int msr, u64 val); + + u64 (*read_tsc)(void); + u64 (*read_pmc)(void); + + /* These two are jmp to, not actually called. */ + void (*irq_enable_sysexit)(void); + void (*iret)(void); +}; + +struct pv_irq_ops { + void (*init_IRQ)(void); + + /* + * Get/set interrupt state. save_fl and restore_fl are only + * expected to use X86_EFLAGS_IF; all other bits + * returned from save_fl are undefined, and may be ignored by + * restore_fl. + */ + unsigned long (*save_fl)(void); + void (*restore_fl)(unsigned long); + void (*irq_disable)(void); + void (*irq_enable)(void); + void (*safe_halt)(void); + void (*halt)(void); +}; + +struct pv_apic_ops { #ifdef CONFIG_X86_LOCAL_APIC /* * Direct APIC operations, principally for VMI. Ideally @@ -167,6 +166,34 @@ struct paravirt_ops unsigned long start_eip, unsigned long start_esp); #endif +}; + +struct pv_mmu_ops { + /* + * Called before/after init_mm pagetable setup. setup_start + * may reset %cr3, and may pre-install parts of the pagetable; + * pagetable setup is expected to preserve any existing + * mapping. + */ + void (*pagetable_setup_start)(pgd_t *pgd_base); + void (*pagetable_setup_done)(pgd_t *pgd_base); + + unsigned long (*read_cr2)(void); + void (*write_cr2)(unsigned long); + + unsigned long (*read_cr3)(void); + void (*write_cr3)(unsigned long); + + /* + * Hooks for intercepting the creation/use/destruction of an + * mm_struct. + */ + void (*activate_mm)(struct mm_struct *prev, + struct mm_struct *next); + void (*dup_mmap)(struct mm_struct *oldmm, + struct mm_struct *mm); + void (*exit_mmap)(struct mm_struct *mm); + /* TLB operations */ void (*flush_tlb_user)(void); @@ -191,15 +218,12 @@ struct paravirt_ops void (*pte_update_defer)(struct mm_struct *mm, unsigned long addr, pte_t *ptep); -#ifdef CONFIG_HIGHPTE - void *(*kmap_atomic_pte)(struct page *page, enum km_type type); -#endif - #ifdef CONFIG_X86_PAE void (*set_pte_atomic)(pte_t *ptep, pte_t pteval); - void (*set_pte_present)(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte); + void (*set_pte_present)(struct mm_struct *mm, unsigned long addr, + pte_t *ptep, pte_t pte); void (*set_pud)(pud_t *pudp, pud_t pudval); - void (*pte_clear)(struct mm_struct *mm, unsigned long addr, pte_t *ptep); + void (*pte_clear)(struct mm_struct *mm, unsigned long addr, pte_t *ptep); void (*pmd_clear)(pmd_t *pmdp); unsigned long long (*pte_val)(pte_t); @@ -217,21 +241,40 @@ struct paravirt_ops pgd_t (*make_pgd)(unsigned long pgd); #endif - /* Set deferred update mode, used for batching operations. */ - void (*set_lazy_mode)(enum paravirt_lazy_mode mode); - - /* These two are jmp to, not actually called. */ - void (*irq_enable_sysexit)(void); - void (*iret)(void); +#ifdef CONFIG_HIGHPTE + void *(*kmap_atomic_pte)(struct page *page, enum km_type type); +#endif }; -extern struct paravirt_ops paravirt_ops; +/* This contains all the paravirt structures: we get a convenient + * number for each function using the offset which we use to indicate + * what to patch. */ +struct paravirt_patch_template +{ + struct pv_init_ops pv_init_ops; + struct pv_misc_ops pv_misc_ops; + struct pv_time_ops pv_time_ops; + struct pv_cpu_ops pv_cpu_ops; + struct pv_irq_ops pv_irq_ops; + struct pv_apic_ops pv_apic_ops; + struct pv_mmu_ops pv_mmu_ops; +}; + +extern struct pv_info pv_info; +extern struct pv_init_ops pv_init_ops; +extern struct pv_misc_ops pv_misc_ops; +extern struct pv_time_ops pv_time_ops; +extern struct pv_cpu_ops pv_cpu_ops; +extern struct pv_irq_ops pv_irq_ops; +extern struct pv_apic_ops pv_apic_ops; +extern struct pv_mmu_ops pv_mmu_ops; #define PARAVIRT_PATCH(x) \ - (offsetof(struct paravirt_ops, x) / sizeof(void *)) + (offsetof(struct paravirt_patch_template, x) / sizeof(void *)) -#define paravirt_type(type) \ - [paravirt_typenum] "i" (PARAVIRT_PATCH(type)) +#define paravirt_type(op) \ + [paravirt_typenum] "i" (PARAVIRT_PATCH(op)), \ + [paravirt_opptr] "m" (op) #define paravirt_clobber(clobber) \ [paravirt_clobber] "i" (clobber) @@ -258,7 +301,7 @@ unsigned paravirt_patch_call(void *insnbuf, const void *target, u16 tgt_clobbers, unsigned long addr, u16 site_clobbers, unsigned len); -unsigned paravirt_patch_jmp(const void *target, void *insnbuf, +unsigned paravirt_patch_jmp(void *insnbuf, const void *target, unsigned long addr, unsigned len); unsigned paravirt_patch_default(u8 type, u16 clobbers, void *insnbuf, unsigned long addr, unsigned len); @@ -271,14 +314,14 @@ int paravirt_disable_iospace(void); /* * This generates an indirect call based on the operation type number. * The type number, computed in PARAVIRT_PATCH, is derived from the - * offset into the paravirt_ops structure, and can therefore be freely - * converted back into a structure offset. + * offset into the paravirt_patch_template structure, and can therefore be + * freely converted back into a structure offset. */ -#define PARAVIRT_CALL "call *(paravirt_ops+%c[paravirt_typenum]*4);" +#define PARAVIRT_CALL "call *%[paravirt_opptr];" /* - * These macros are intended to wrap calls into a paravirt_ops - * operation, so that they can be later identified and patched at + * These macros are intended to wrap calls through one of the paravirt + * ops structs, so that they can be later identified and patched at * runtime. * * Normally, a call to a pv_op function is a simple indirect call: @@ -301,7 +344,7 @@ int paravirt_disable_iospace(void); * The call instruction itself is marked by placing its start address * and size into the .parainstructions section, so that * apply_paravirt() in arch/i386/kernel/alternative.c can do the - * appropriate patching under the control of the backend paravirt_ops + * appropriate patching under the control of the backend pv_init_ops * implementation. * * Unfortunately there's no way to get gcc to generate the args setup @@ -409,36 +452,36 @@ int paravirt_disable_iospace(void); static inline int paravirt_enabled(void) { - return paravirt_ops.paravirt_enabled; + return pv_info.paravirt_enabled; } static inline void load_esp0(struct tss_struct *tss, struct thread_struct *thread) { - PVOP_VCALL2(load_esp0, tss, thread); + PVOP_VCALL2(pv_cpu_ops.load_esp0, tss, thread); } -#define ARCH_SETUP paravirt_ops.arch_setup(); +#define ARCH_SETUP pv_init_ops.arch_setup(); static inline unsigned long get_wallclock(void) { - return PVOP_CALL0(unsigned long, get_wallclock); + return PVOP_CALL0(unsigned long, pv_time_ops.get_wallclock); } static inline int set_wallclock(unsigned long nowtime) { - return PVOP_CALL1(int, set_wallclock, nowtime); + return PVOP_CALL1(int, pv_time_ops.set_wallclock, nowtime); } static inline void (*choose_time_init(void))(void) { - return paravirt_ops.time_init; + return pv_time_ops.time_init; } /* The paravirtualized CPUID instruction. */ static inline void __cpuid(unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx) { - PVOP_VCALL4(cpuid, eax, ebx, ecx, edx); + PVOP_VCALL4(pv_cpu_ops.cpuid, eax, ebx, ecx, edx); } /* @@ -446,87 +489,87 @@ static inline void __cpuid(unsigned int *eax, unsigned int *ebx, */ static inline unsigned long paravirt_get_debugreg(int reg) { - return PVOP_CALL1(unsigned long, get_debugreg, reg); + return PVOP_CALL1(unsigned long, pv_cpu_ops.get_debugreg, reg); } #define get_debugreg(var, reg) var = paravirt_get_debugreg(reg) static inline void set_debugreg(unsigned long val, int reg) { - PVOP_VCALL2(set_debugreg, reg, val); + PVOP_VCALL2(pv_cpu_ops.set_debugreg, reg, val); } static inline void clts(void) { - PVOP_VCALL0(clts); + PVOP_VCALL0(pv_cpu_ops.clts); } static inline unsigned long read_cr0(void) { - return PVOP_CALL0(unsigned long, read_cr0); + return PVOP_CALL0(unsigned long, pv_cpu_ops.read_cr0); } static inline void write_cr0(unsigned long x) { - PVOP_VCALL1(write_cr0, x); + PVOP_VCALL1(pv_cpu_ops.write_cr0, x); } static inline unsigned long read_cr2(void) { - return PVOP_CALL0(unsigned long, read_cr2); + return PVOP_CALL0(unsigned long, pv_mmu_ops.read_cr2); } static inline void write_cr2(unsigned long x) { - PVOP_VCALL1(write_cr2, x); + PVOP_VCALL1(pv_mmu_ops.write_cr2, x); } static inline unsigned long read_cr3(void) { - return PVOP_CALL0(unsigned long, read_cr3); + return PVOP_CALL0(unsigned long, pv_mmu_ops.read_cr3); } static inline void write_cr3(unsigned long x) { - PVOP_VCALL1(write_cr3, x); + PVOP_VCALL1(pv_mmu_ops.write_cr3, x); } static inline unsigned long read_cr4(void) { - return PVOP_CALL0(unsigned long, read_cr4); + return PVOP_CALL0(unsigned long, pv_cpu_ops.read_cr4); } static inline unsigned long read_cr4_safe(void) { - return PVOP_CALL0(unsigned long, read_cr4_safe); + return PVOP_CALL0(unsigned long, pv_cpu_ops.read_cr4_safe); } static inline void write_cr4(unsigned long x) { - PVOP_VCALL1(write_cr4, x); + PVOP_VCALL1(pv_cpu_ops.write_cr4, x); } static inline void raw_safe_halt(void) { - PVOP_VCALL0(safe_halt); + PVOP_VCALL0(pv_irq_ops.safe_halt); } static inline void halt(void) { - PVOP_VCALL0(safe_halt); + PVOP_VCALL0(pv_irq_ops.safe_halt); } static inline void wbinvd(void) { - PVOP_VCALL0(wbinvd); + PVOP_VCALL0(pv_cpu_ops.wbinvd); } -#define get_kernel_rpl() (paravirt_ops.kernel_rpl) +#define get_kernel_rpl() (pv_info.kernel_rpl) static inline u64 paravirt_read_msr(unsigned msr, int *err) { - return PVOP_CALL2(u64, read_msr, msr, err); + return PVOP_CALL2(u64, pv_cpu_ops.read_msr, msr, err); } static inline int paravirt_write_msr(unsigned msr, unsigned low, unsigned high) { - return PVOP_CALL3(int, write_msr, msr, low, high); + return PVOP_CALL3(int, pv_cpu_ops.write_msr, msr, low, high); } /* These should all do BUG_ON(_err), but our headers are too tangled. */ @@ -560,7 +603,7 @@ static inline int paravirt_write_msr(unsigned msr, unsigned low, unsigned high) static inline u64 paravirt_read_tsc(void) { - return PVOP_CALL0(u64, read_tsc); + return PVOP_CALL0(u64, pv_cpu_ops.read_tsc); } #define rdtscl(low) do { \ @@ -572,15 +615,15 @@ static inline u64 paravirt_read_tsc(void) static inline unsigned long long paravirt_sched_clock(void) { - return PVOP_CALL0(unsigned long long, sched_clock); + return PVOP_CALL0(unsigned long long, pv_time_ops.sched_clock); } -#define calculate_cpu_khz() (paravirt_ops.get_cpu_khz()) +#define calculate_cpu_khz() (pv_time_ops.get_cpu_khz()) #define write_tsc(val1,val2) wrmsr(0x10, val1, val2) static inline unsigned long long paravirt_read_pmc(int counter) { - return PVOP_CALL1(u64, read_pmc, counter); + return PVOP_CALL1(u64, pv_cpu_ops.read_pmc, counter); } #define rdpmc(counter,low,high) do { \ @@ -591,61 +634,61 @@ static inline unsigned long long paravirt_read_pmc(int counter) static inline void load_TR_desc(void) { - PVOP_VCALL0(load_tr_desc); + PVOP_VCALL0(pv_cpu_ops.load_tr_desc); } static inline void load_gdt(const struct Xgt_desc_struct *dtr) { - PVOP_VCALL1(load_gdt, dtr); + PVOP_VCALL1(pv_cpu_ops.load_gdt, dtr); } static inline void load_idt(const struct Xgt_desc_struct *dtr) { - PVOP_VCALL1(load_idt, dtr); + PVOP_VCALL1(pv_cpu_ops.load_idt, dtr); } static inline void set_ldt(const void *addr, unsigned entries) { - PVOP_VCALL2(set_ldt, addr, entries); + PVOP_VCALL2(pv_cpu_ops.set_ldt, addr, entries); } static inline void store_gdt(struct Xgt_desc_struct *dtr) { - PVOP_VCALL1(store_gdt, dtr); + PVOP_VCALL1(pv_cpu_ops.store_gdt, dtr); } static inline void store_idt(struct Xgt_desc_struct *dtr) { - PVOP_VCALL1(store_idt, dtr); + PVOP_VCALL1(pv_cpu_ops.store_idt, dtr); } static inline unsigned long paravirt_store_tr(void) { - return PVOP_CALL0(unsigned long, store_tr); + return PVOP_CALL0(unsigned long, pv_cpu_ops.store_tr); } #define store_tr(tr) ((tr) = paravirt_store_tr()) static inline void load_TLS(struct thread_struct *t, unsigned cpu) { - PVOP_VCALL2(load_tls, t, cpu); + PVOP_VCALL2(pv_cpu_ops.load_tls, t, cpu); } static inline void write_ldt_entry(void *dt, int entry, u32 low, u32 high) { - PVOP_VCALL4(write_ldt_entry, dt, entry, low, high); + PVOP_VCALL4(pv_cpu_ops.write_ldt_entry, dt, entry, low, high); } static inline void write_gdt_entry(void *dt, int entry, u32 low, u32 high) { - PVOP_VCALL4(write_gdt_entry, dt, entry, low, high); + PVOP_VCALL4(pv_cpu_ops.write_gdt_entry, dt, entry, low, high); } static inline void write_idt_entry(void *dt, int entry, u32 low, u32 high) { - PVOP_VCALL4(write_idt_entry, dt, entry, low, high); + PVOP_VCALL4(pv_cpu_ops.write_idt_entry, dt, entry, low, high); } static inline void set_iopl_mask(unsigned mask) { - PVOP_VCALL1(set_iopl_mask, mask); + PVOP_VCALL1(pv_cpu_ops.set_iopl_mask, mask); } /* The paravirtualized I/O functions */ static inline void slow_down_io(void) { - paravirt_ops.io_delay(); + pv_cpu_ops.io_delay(); #ifdef REALLY_SLOW_IO - paravirt_ops.io_delay(); - paravirt_ops.io_delay(); - paravirt_ops.io_delay(); + pv_cpu_ops.io_delay(); + pv_cpu_ops.io_delay(); + pv_cpu_ops.io_delay(); #endif } @@ -655,121 +698,120 @@ static inline void slow_down_io(void) { */ static inline void apic_write(unsigned long reg, unsigned long v) { - PVOP_VCALL2(apic_write, reg, v); + PVOP_VCALL2(pv_apic_ops.apic_write, reg, v); } static inline void apic_write_atomic(unsigned long reg, unsigned long v) { - PVOP_VCALL2(apic_write_atomic, reg, v); + PVOP_VCALL2(pv_apic_ops.apic_write_atomic, reg, v); } static inline unsigned long apic_read(unsigned long reg) { - return PVOP_CALL1(unsigned long, apic_read, reg); + return PVOP_CALL1(unsigned long, pv_apic_ops.apic_read, reg); } static inline void setup_boot_clock(void) { - PVOP_VCALL0(setup_boot_clock); + PVOP_VCALL0(pv_apic_ops.setup_boot_clock); } static inline void setup_secondary_clock(void) { - PVOP_VCALL0(setup_secondary_clock); + PVOP_VCALL0(pv_apic_ops.setup_secondary_clock); } #endif static inline void paravirt_post_allocator_init(void) { - if (paravirt_ops.post_allocator_init) - (*paravirt_ops.post_allocator_init)(); + if (pv_init_ops.post_allocator_init) + (*pv_init_ops.post_allocator_init)(); } static inline void paravirt_pagetable_setup_start(pgd_t *base) { - if (paravirt_ops.pagetable_setup_start) - (*paravirt_ops.pagetable_setup_start)(base); + (*pv_mmu_ops.pagetable_setup_start)(base); } static inline void paravirt_pagetable_setup_done(pgd_t *base) { - if (paravirt_ops.pagetable_setup_done) - (*paravirt_ops.pagetable_setup_done)(base); + (*pv_mmu_ops.pagetable_setup_done)(base); } #ifdef CONFIG_SMP static inline void startup_ipi_hook(int phys_apicid, unsigned long start_eip, unsigned long start_esp) { - PVOP_VCALL3(startup_ipi_hook, phys_apicid, start_eip, start_esp); + PVOP_VCALL3(pv_apic_ops.startup_ipi_hook, + phys_apicid, start_eip, start_esp); } #endif static inline void paravirt_activate_mm(struct mm_struct *prev, struct mm_struct *next) { - PVOP_VCALL2(activate_mm, prev, next); + PVOP_VCALL2(pv_mmu_ops.activate_mm, prev, next); } static inline void arch_dup_mmap(struct mm_struct *oldmm, struct mm_struct *mm) { - PVOP_VCALL2(dup_mmap, oldmm, mm); + PVOP_VCALL2(pv_mmu_ops.dup_mmap, oldmm, mm); } static inline void arch_exit_mmap(struct mm_struct *mm) { - PVOP_VCALL1(exit_mmap, mm); + PVOP_VCALL1(pv_mmu_ops.exit_mmap, mm); } static inline void __flush_tlb(void) { - PVOP_VCALL0(flush_tlb_user); + PVOP_VCALL0(pv_mmu_ops.flush_tlb_user); } static inline void __flush_tlb_global(void) { - PVOP_VCALL0(flush_tlb_kernel); + PVOP_VCALL0(pv_mmu_ops.flush_tlb_kernel); } static inline void __flush_tlb_single(unsigned long addr) { - PVOP_VCALL1(flush_tlb_single, addr); + PVOP_VCALL1(pv_mmu_ops.flush_tlb_single, addr); } static inline void flush_tlb_others(cpumask_t cpumask, struct mm_struct *mm, unsigned long va) { - PVOP_VCALL3(flush_tlb_others, &cpumask, mm, va); + PVOP_VCALL3(pv_mmu_ops.flush_tlb_others, &cpumask, mm, va); } static inline void paravirt_alloc_pt(struct mm_struct *mm, unsigned pfn) { - PVOP_VCALL2(alloc_pt, mm, pfn); + PVOP_VCALL2(pv_mmu_ops.alloc_pt, mm, pfn); } static inline void paravirt_release_pt(unsigned pfn) { - PVOP_VCALL1(release_pt, pfn); + PVOP_VCALL1(pv_mmu_ops.release_pt, pfn); } static inline void paravirt_alloc_pd(unsigned pfn) { - PVOP_VCALL1(alloc_pd, pfn); + PVOP_VCALL1(pv_mmu_ops.alloc_pd, pfn); } static inline void paravirt_alloc_pd_clone(unsigned pfn, unsigned clonepfn, unsigned start, unsigned count) { - PVOP_VCALL4(alloc_pd_clone, pfn, clonepfn, start, count); + PVOP_VCALL4(pv_mmu_ops.alloc_pd_clone, pfn, clonepfn, start, count); } static inline void paravirt_release_pd(unsigned pfn) { - PVOP_VCALL1(release_pd, pfn); + PVOP_VCALL1(pv_mmu_ops.release_pd, pfn); } #ifdef CONFIG_HIGHPTE static inline void *kmap_atomic_pte(struct page *page, enum km_type type) { unsigned long ret; - ret = PVOP_CALL2(unsigned long, kmap_atomic_pte, page, type); + ret = PVOP_CALL2(unsigned long, pv_mmu_ops.kmap_atomic_pte, page, type); return (void *)ret; } #endif @@ -777,162 +819,171 @@ static inline void *kmap_atomic_pte(struct page *page, enum km_type type) static inline void pte_update(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - PVOP_VCALL3(pte_update, mm, addr, ptep); + PVOP_VCALL3(pv_mmu_ops.pte_update, mm, addr, ptep); } static inline void pte_update_defer(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - PVOP_VCALL3(pte_update_defer, mm, addr, ptep); + PVOP_VCALL3(pv_mmu_ops.pte_update_defer, mm, addr, ptep); } #ifdef CONFIG_X86_PAE static inline pte_t __pte(unsigned long long val) { - unsigned long long ret = PVOP_CALL2(unsigned long long, make_pte, + unsigned long long ret = PVOP_CALL2(unsigned long long, + pv_mmu_ops.make_pte, val, val >> 32); return (pte_t) { ret, ret >> 32 }; } static inline pmd_t __pmd(unsigned long long val) { - return (pmd_t) { PVOP_CALL2(unsigned long long, make_pmd, val, val >> 32) }; + return (pmd_t) { PVOP_CALL2(unsigned long long, pv_mmu_ops.make_pmd, + val, val >> 32) }; } static inline pgd_t __pgd(unsigned long long val) { - return (pgd_t) { PVOP_CALL2(unsigned long long, make_pgd, val, val >> 32) }; + return (pgd_t) { PVOP_CALL2(unsigned long long, pv_mmu_ops.make_pgd, + val, val >> 32) }; } static inline unsigned long long pte_val(pte_t x) { - return PVOP_CALL2(unsigned long long, pte_val, x.pte_low, x.pte_high); + return PVOP_CALL2(unsigned long long, pv_mmu_ops.pte_val, + x.pte_low, x.pte_high); } static inline unsigned long long pmd_val(pmd_t x) { - return PVOP_CALL2(unsigned long long, pmd_val, x.pmd, x.pmd >> 32); + return PVOP_CALL2(unsigned long long, pv_mmu_ops.pmd_val, + x.pmd, x.pmd >> 32); } static inline unsigned long long pgd_val(pgd_t x) { - return PVOP_CALL2(unsigned long long, pgd_val, x.pgd, x.pgd >> 32); + return PVOP_CALL2(unsigned long long, pv_mmu_ops.pgd_val, + x.pgd, x.pgd >> 32); } static inline void set_pte(pte_t *ptep, pte_t pteval) { - PVOP_VCALL3(set_pte, ptep, pteval.pte_low, pteval.pte_high); + PVOP_VCALL3(pv_mmu_ops.set_pte, ptep, pteval.pte_low, pteval.pte_high); } static inline void set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pteval) { /* 5 arg words */ - paravirt_ops.set_pte_at(mm, addr, ptep, pteval); + pv_mmu_ops.set_pte_at(mm, addr, ptep, pteval); } static inline void set_pte_atomic(pte_t *ptep, pte_t pteval) { - PVOP_VCALL3(set_pte_atomic, ptep, pteval.pte_low, pteval.pte_high); + PVOP_VCALL3(pv_mmu_ops.set_pte_atomic, ptep, + pteval.pte_low, pteval.pte_high); } static inline void set_pte_present(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) { /* 5 arg words */ - paravirt_ops.set_pte_present(mm, addr, ptep, pte); + pv_mmu_ops.set_pte_present(mm, addr, ptep, pte); } static inline void set_pmd(pmd_t *pmdp, pmd_t pmdval) { - PVOP_VCALL3(set_pmd, pmdp, pmdval.pmd, pmdval.pmd >> 32); + PVOP_VCALL3(pv_mmu_ops.set_pmd, pmdp, + pmdval.pmd, pmdval.pmd >> 32); } static inline void set_pud(pud_t *pudp, pud_t pudval) { - PVOP_VCALL3(set_pud, pudp, pudval.pgd.pgd, pudval.pgd.pgd >> 32); + PVOP_VCALL3(pv_mmu_ops.set_pud, pudp, + pudval.pgd.pgd, pudval.pgd.pgd >> 32); } static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - PVOP_VCALL3(pte_clear, mm, addr, ptep); + PVOP_VCALL3(pv_mmu_ops.pte_clear, mm, addr, ptep); } static inline void pmd_clear(pmd_t *pmdp) { - PVOP_VCALL1(pmd_clear, pmdp); + PVOP_VCALL1(pv_mmu_ops.pmd_clear, pmdp); } #else /* !CONFIG_X86_PAE */ static inline pte_t __pte(unsigned long val) { - return (pte_t) { PVOP_CALL1(unsigned long, make_pte, val) }; + return (pte_t) { PVOP_CALL1(unsigned long, pv_mmu_ops.make_pte, val) }; } static inline pgd_t __pgd(unsigned long val) { - return (pgd_t) { PVOP_CALL1(unsigned long, make_pgd, val) }; + return (pgd_t) { PVOP_CALL1(unsigned long, pv_mmu_ops.make_pgd, val) }; } static inline unsigned long pte_val(pte_t x) { - return PVOP_CALL1(unsigned long, pte_val, x.pte_low); + return PVOP_CALL1(unsigned long, pv_mmu_ops.pte_val, x.pte_low); } static inline unsigned long pgd_val(pgd_t x) { - return PVOP_CALL1(unsigned long, pgd_val, x.pgd); + return PVOP_CALL1(unsigned long, pv_mmu_ops.pgd_val, x.pgd); } static inline void set_pte(pte_t *ptep, pte_t pteval) { - PVOP_VCALL2(set_pte, ptep, pteval.pte_low); + PVOP_VCALL2(pv_mmu_ops.set_pte, ptep, pteval.pte_low); } static inline void set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pteval) { - PVOP_VCALL4(set_pte_at, mm, addr, ptep, pteval.pte_low); + PVOP_VCALL4(pv_mmu_ops.set_pte_at, mm, addr, ptep, pteval.pte_low); } static inline void set_pmd(pmd_t *pmdp, pmd_t pmdval) { - PVOP_VCALL2(set_pmd, pmdp, pmdval.pud.pgd.pgd); + PVOP_VCALL2(pv_mmu_ops.set_pmd, pmdp, pmdval.pud.pgd.pgd); } #endif /* CONFIG_X86_PAE */ #define __HAVE_ARCH_ENTER_LAZY_CPU_MODE static inline void arch_enter_lazy_cpu_mode(void) { - PVOP_VCALL1(set_lazy_mode, PARAVIRT_LAZY_CPU); + PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_CPU); } static inline void arch_leave_lazy_cpu_mode(void) { - PVOP_VCALL1(set_lazy_mode, PARAVIRT_LAZY_NONE); + PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_NONE); } static inline void arch_flush_lazy_cpu_mode(void) { - PVOP_VCALL1(set_lazy_mode, PARAVIRT_LAZY_FLUSH); + PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_FLUSH); } #define __HAVE_ARCH_ENTER_LAZY_MMU_MODE static inline void arch_enter_lazy_mmu_mode(void) { - PVOP_VCALL1(set_lazy_mode, PARAVIRT_LAZY_MMU); + PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_MMU); } static inline void arch_leave_lazy_mmu_mode(void) { - PVOP_VCALL1(set_lazy_mode, PARAVIRT_LAZY_NONE); + PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_NONE); } static inline void arch_flush_lazy_mmu_mode(void) { - PVOP_VCALL1(set_lazy_mode, PARAVIRT_LAZY_FLUSH); + PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_FLUSH); } void _paravirt_nop(void); @@ -957,7 +1008,7 @@ static inline unsigned long __raw_local_save_flags(void) PARAVIRT_CALL "popl %%edx; popl %%ecx") : "=a"(f) - : paravirt_type(save_fl), + : paravirt_type(pv_irq_ops.save_fl), paravirt_clobber(CLBR_EAX) : "memory", "cc"); return f; @@ -970,7 +1021,7 @@ static inline void raw_local_irq_restore(unsigned long f) "popl %%edx; popl %%ecx") : "=a"(f) : "0"(f), - paravirt_type(restore_fl), + paravirt_type(pv_irq_ops.restore_fl), paravirt_clobber(CLBR_EAX) : "memory", "cc"); } @@ -981,7 +1032,7 @@ static inline void raw_local_irq_disable(void) PARAVIRT_CALL "popl %%edx; popl %%ecx") : - : paravirt_type(irq_disable), + : paravirt_type(pv_irq_ops.irq_disable), paravirt_clobber(CLBR_EAX) : "memory", "eax", "cc"); } @@ -992,7 +1043,7 @@ static inline void raw_local_irq_enable(void) PARAVIRT_CALL "popl %%edx; popl %%ecx") : - : paravirt_type(irq_enable), + : paravirt_type(pv_irq_ops.irq_enable), paravirt_clobber(CLBR_EAX) : "memory", "eax", "cc"); } @@ -1008,21 +1059,23 @@ static inline unsigned long __raw_local_irq_save(void) #define CLI_STRING \ _paravirt_alt("pushl %%ecx; pushl %%edx;" \ - "call *paravirt_ops+%c[paravirt_cli_type]*4;" \ + "call *%[paravirt_cli_opptr];" \ "popl %%edx; popl %%ecx", \ "%c[paravirt_cli_type]", "%c[paravirt_clobber]") #define STI_STRING \ _paravirt_alt("pushl %%ecx; pushl %%edx;" \ - "call *paravirt_ops+%c[paravirt_sti_type]*4;" \ + "call *%[paravirt_sti_opptr];" \ "popl %%edx; popl %%ecx", \ "%c[paravirt_sti_type]", "%c[paravirt_clobber]") #define CLI_STI_CLOBBERS , "%eax" #define CLI_STI_INPUT_ARGS \ , \ - [paravirt_cli_type] "i" (PARAVIRT_PATCH(irq_disable)), \ - [paravirt_sti_type] "i" (PARAVIRT_PATCH(irq_enable)), \ + [paravirt_cli_type] "i" (PARAVIRT_PATCH(pv_irq_ops.irq_disable)), \ + [paravirt_cli_opptr] "m" (pv_irq_ops.irq_disable), \ + [paravirt_sti_type] "i" (PARAVIRT_PATCH(pv_irq_ops.irq_enable)), \ + [paravirt_sti_opptr] "m" (pv_irq_ops.irq_enable), \ paravirt_clobber(CLBR_EAX) /* Make sure as little as possible of this mess escapes. */ @@ -1042,7 +1095,7 @@ static inline unsigned long __raw_local_irq_save(void) #else /* __ASSEMBLY__ */ -#define PARA_PATCH(off) ((off) / 4) +#define PARA_PATCH(struct, off) ((PARAVIRT_PATCH_##struct + (off)) / 4) #define PARA_SITE(ptype, clobbers, ops) \ 771:; \ @@ -1055,29 +1108,29 @@ static inline unsigned long __raw_local_irq_save(void) .short clobbers; \ .popsection -#define INTERRUPT_RETURN \ - PARA_SITE(PARA_PATCH(PARAVIRT_iret), CLBR_NONE, \ - jmp *%cs:paravirt_ops+PARAVIRT_iret) +#define INTERRUPT_RETURN \ + PARA_SITE(PARA_PATCH(pv_cpu_ops, PV_CPU_iret), CLBR_NONE, \ + jmp *%cs:pv_cpu_ops+PV_CPU_iret) #define DISABLE_INTERRUPTS(clobbers) \ - PARA_SITE(PARA_PATCH(PARAVIRT_irq_disable), clobbers, \ + PARA_SITE(PARA_PATCH(pv_irq_ops, PV_IRQ_irq_disable), clobbers, \ pushl %eax; pushl %ecx; pushl %edx; \ - call *%cs:paravirt_ops+PARAVIRT_irq_disable; \ + call *%cs:pv_irq_ops+PV_IRQ_irq_disable; \ popl %edx; popl %ecx; popl %eax) \ #define ENABLE_INTERRUPTS(clobbers) \ - PARA_SITE(PARA_PATCH(PARAVIRT_irq_enable), clobbers, \ + PARA_SITE(PARA_PATCH(pv_irq_ops, PV_IRQ_irq_enable), clobbers, \ pushl %eax; pushl %ecx; pushl %edx; \ - call *%cs:paravirt_ops+PARAVIRT_irq_enable; \ + call *%cs:pv_irq_ops+PV_IRQ_irq_enable; \ popl %edx; popl %ecx; popl %eax) -#define ENABLE_INTERRUPTS_SYSEXIT \ - PARA_SITE(PARA_PATCH(PARAVIRT_irq_enable_sysexit), CLBR_NONE, \ - jmp *%cs:paravirt_ops+PARAVIRT_irq_enable_sysexit) +#define ENABLE_INTERRUPTS_SYSEXIT \ + PARA_SITE(PARA_PATCH(pv_cpu_ops, PV_CPU_irq_enable_sysexit), CLBR_NONE,\ + jmp *%cs:pv_cpu_ops+PV_CPU_irq_enable_sysexit) #define GET_CR0_INTO_EAX \ push %ecx; push %edx; \ - call *paravirt_ops+PARAVIRT_read_cr0; \ + call *pv_cpu_ops+PV_CPU_read_cr0; \ pop %edx; pop %ecx #endif /* __ASSEMBLY__ */ diff --git a/include/asm-x86/pgtable-3level-defs.h b/include/asm-x86/pgtable-3level-defs.h index c0df89f66e8..448ac951631 100644 --- a/include/asm-x86/pgtable-3level-defs.h +++ b/include/asm-x86/pgtable-3level-defs.h @@ -2,7 +2,7 @@ #define _I386_PGTABLE_3LEVEL_DEFS_H #ifdef CONFIG_PARAVIRT -#define SHARED_KERNEL_PMD (paravirt_ops.shared_kernel_pmd) +#define SHARED_KERNEL_PMD (pv_info.shared_kernel_pmd) #else #define SHARED_KERNEL_PMD 1 #endif From 8965c1c0950d459d99b8b81dfc1ab02e3d2cfb08 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:29 -0700 Subject: [PATCH 02/12] paravirt: clean up lazy mode handling Currently, the set_lazy_mode pv_op is overloaded with 5 functions: 1. enter lazy cpu mode 2. leave lazy cpu mode 3. enter lazy mmu mode 4. leave lazy mmu mode 5. flush pending batched operations This complicates each paravirt backend, since it needs to deal with all the possible state transitions, handling flushing, etc. In particular, flushing is quite distinct from the other 4 functions, and seems to just cause complication. This patch removes the set_lazy_mode operation, and adds "enter" and "leave" lazy mode operations on mmu_ops and cpu_ops. All the logic associated with enter and leaving lazy states is now in common code (basically BUG_ONs to make sure that no mode is current when entering a lazy mode, and make sure that the mode is current when leaving). Also, flush is handled in a common way, by simply leaving and re-entering the lazy mode. The result is that the Xen, lguest and VMI lazy mode implementations are much simpler. Signed-off-by: Jeremy Fitzhardinge Cc: Andi Kleen Cc: Zach Amsden Cc: Rusty Russell Cc: Avi Kivity Cc: Anthony Liguory Cc: "Glauber de Oliveira Costa" Cc: Jun Nakajima --- arch/x86/kernel/paravirt_32.c | 58 ++++++++++++++++++++++++++++++++--- arch/x86/kernel/vmi_32.c | 39 +++++++++++++---------- arch/x86/xen/enlighten.c | 40 ++++++++---------------- arch/x86/xen/mmu.c | 2 +- arch/x86/xen/multicalls.h | 2 +- arch/x86/xen/xen-ops.h | 7 ----- drivers/lguest/lguest.c | 32 +++++++------------ include/asm-x86/paravirt.h | 52 +++++++++++++++++++------------ 8 files changed, 134 insertions(+), 98 deletions(-) diff --git a/arch/x86/kernel/paravirt_32.c b/arch/x86/kernel/paravirt_32.c index fa412515af7..6a80d67c212 100644 --- a/arch/x86/kernel/paravirt_32.c +++ b/arch/x86/kernel/paravirt_32.c @@ -164,7 +164,6 @@ static void *get_call_destination(u8 type) { struct paravirt_patch_template tmpl = { .pv_init_ops = pv_init_ops, - .pv_misc_ops = pv_misc_ops, .pv_time_ops = pv_time_ops, .pv_cpu_ops = pv_cpu_ops, .pv_irq_ops = pv_irq_ops, @@ -282,6 +281,49 @@ int paravirt_disable_iospace(void) return ret; } +static DEFINE_PER_CPU(enum paravirt_lazy_mode, paravirt_lazy_mode) = PARAVIRT_LAZY_NONE; + +static inline void enter_lazy(enum paravirt_lazy_mode mode) +{ + BUG_ON(x86_read_percpu(paravirt_lazy_mode) != PARAVIRT_LAZY_NONE); + BUG_ON(preemptible()); + + x86_write_percpu(paravirt_lazy_mode, mode); +} + +void paravirt_leave_lazy(enum paravirt_lazy_mode mode) +{ + BUG_ON(x86_read_percpu(paravirt_lazy_mode) != mode); + BUG_ON(preemptible()); + + x86_write_percpu(paravirt_lazy_mode, PARAVIRT_LAZY_NONE); +} + +void paravirt_enter_lazy_mmu(void) +{ + enter_lazy(PARAVIRT_LAZY_MMU); +} + +void paravirt_leave_lazy_mmu(void) +{ + paravirt_leave_lazy(PARAVIRT_LAZY_MMU); +} + +void paravirt_enter_lazy_cpu(void) +{ + enter_lazy(PARAVIRT_LAZY_CPU); +} + +void paravirt_leave_lazy_cpu(void) +{ + paravirt_leave_lazy(PARAVIRT_LAZY_CPU); +} + +enum paravirt_lazy_mode paravirt_get_lazy_mode(void) +{ + return x86_read_percpu(paravirt_lazy_mode); +} + struct pv_info pv_info = { .name = "bare hardware", .paravirt_enabled = 0, @@ -347,6 +389,11 @@ struct pv_cpu_ops pv_cpu_ops = { .set_iopl_mask = native_set_iopl_mask, .io_delay = native_io_delay, + + .lazy_mode = { + .enter = paravirt_nop, + .leave = paravirt_nop, + }, }; struct pv_apic_ops pv_apic_ops = { @@ -360,10 +407,6 @@ struct pv_apic_ops pv_apic_ops = { #endif }; -struct pv_misc_ops pv_misc_ops = { - .set_lazy_mode = paravirt_nop, -}; - struct pv_mmu_ops pv_mmu_ops = { .pagetable_setup_start = native_pagetable_setup_start, .pagetable_setup_done = native_pagetable_setup_done, @@ -414,6 +457,11 @@ struct pv_mmu_ops pv_mmu_ops = { .dup_mmap = paravirt_nop, .exit_mmap = paravirt_nop, .activate_mm = paravirt_nop, + + .lazy_mode = { + .enter = paravirt_nop, + .leave = paravirt_nop, + }, }; EXPORT_SYMBOL_GPL(pv_time_ops); diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c index 67cea5c2e3e..f02bad68aba 100644 --- a/arch/x86/kernel/vmi_32.c +++ b/arch/x86/kernel/vmi_32.c @@ -552,24 +552,22 @@ vmi_startup_ipi_hook(int phys_apicid, unsigned long start_eip, } #endif -static void vmi_set_lazy_mode(enum paravirt_lazy_mode mode) +static void vmi_enter_lazy_cpu(void) { - static DEFINE_PER_CPU(enum paravirt_lazy_mode, lazy_mode); + paravirt_enter_lazy_cpu(); + vmi_ops.set_lazy_mode(2); +} - if (!vmi_ops.set_lazy_mode) - return; +static void vmi_enter_lazy_mmu(void) +{ + paravirt_enter_lazy_mmu(); + vmi_ops.set_lazy_mode(1); +} - /* Modes should never nest or overlap */ - BUG_ON(__get_cpu_var(lazy_mode) && !(mode == PARAVIRT_LAZY_NONE || - mode == PARAVIRT_LAZY_FLUSH)); - - if (mode == PARAVIRT_LAZY_FLUSH) { - vmi_ops.set_lazy_mode(0); - vmi_ops.set_lazy_mode(__get_cpu_var(lazy_mode)); - } else { - vmi_ops.set_lazy_mode(mode); - __get_cpu_var(lazy_mode) = mode; - } +static void vmi_leave_lazy(void) +{ + paravirt_leave_lazy(paravirt_get_lazy_mode()); + vmi_ops.set_lazy_mode(0); } static inline int __init check_vmi_rom(struct vrom_header *rom) @@ -798,7 +796,16 @@ static inline int __init activate_vmi(void) para_wrap(pv_cpu_ops.load_esp0, vmi_load_esp0, set_kernel_stack, UpdateKernelStack); para_fill(pv_cpu_ops.set_iopl_mask, SetIOPLMask); para_fill(pv_cpu_ops.io_delay, IODelay); - para_wrap(pv_misc_ops.set_lazy_mode, vmi_set_lazy_mode, set_lazy_mode, SetLazyMode); + + para_wrap(pv_cpu_ops.lazy_mode.enter, vmi_enter_lazy_cpu, + set_lazy_mode, SetLazyMode); + para_wrap(pv_cpu_ops.lazy_mode.leave, vmi_leave_lazy, + set_lazy_mode, SetLazyMode); + + para_wrap(pv_mmu_ops.lazy_mode.enter, vmi_enter_lazy_mmu, + set_lazy_mode, SetLazyMode); + para_wrap(pv_mmu_ops.lazy_mode.leave, vmi_leave_lazy, + set_lazy_mode, SetLazyMode); /* user and kernel flush are just handled with different flags to FlushTLB */ para_wrap(pv_mmu_ops.flush_tlb_user, vmi_flush_tlb_user, _flush_tlb, FlushTLB); diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 3d3bf05dec7..7171a073607 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -52,8 +52,6 @@ EXPORT_SYMBOL_GPL(hypercall_page); -DEFINE_PER_CPU(enum paravirt_lazy_mode, xen_lazy_mode); - DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu); DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info); DEFINE_PER_CPU(unsigned long, xen_cr3); @@ -249,29 +247,10 @@ static void xen_halt(void) xen_safe_halt(); } -static void xen_set_lazy_mode(enum paravirt_lazy_mode mode) +static void xen_leave_lazy(void) { - BUG_ON(preemptible()); - - switch (mode) { - case PARAVIRT_LAZY_NONE: - BUG_ON(x86_read_percpu(xen_lazy_mode) == PARAVIRT_LAZY_NONE); - break; - - case PARAVIRT_LAZY_MMU: - case PARAVIRT_LAZY_CPU: - BUG_ON(x86_read_percpu(xen_lazy_mode) != PARAVIRT_LAZY_NONE); - break; - - case PARAVIRT_LAZY_FLUSH: - /* flush if necessary, but don't change state */ - if (x86_read_percpu(xen_lazy_mode) != PARAVIRT_LAZY_NONE) - xen_mc_flush(); - return; - } - + paravirt_leave_lazy(paravirt_get_lazy_mode()); xen_mc_flush(); - x86_write_percpu(xen_lazy_mode, mode); } static unsigned long xen_store_tr(void) @@ -358,7 +337,7 @@ static void xen_load_tls(struct thread_struct *t, unsigned int cpu) * loaded properly. This will go away as soon as Xen has been * modified to not save/restore %gs for normal hypercalls. */ - if (xen_get_lazy_mode() == PARAVIRT_LAZY_CPU) + if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_CPU) loadsegment(gs, 0); } @@ -962,6 +941,11 @@ static const struct pv_cpu_ops xen_cpu_ops __initdata = { .set_iopl_mask = xen_set_iopl_mask, .io_delay = xen_io_delay, + + .lazy_mode = { + .enter = paravirt_enter_lazy_cpu, + .leave = xen_leave_lazy, + }, }; static const struct pv_irq_ops xen_irq_ops __initdata = { @@ -1037,10 +1021,11 @@ static const struct pv_mmu_ops xen_mmu_ops __initdata = { .activate_mm = xen_activate_mm, .dup_mmap = xen_dup_mmap, .exit_mmap = xen_exit_mmap, -}; -static const struct pv_misc_ops xen_misc_ops __initdata = { - .set_lazy_mode = xen_set_lazy_mode, + .lazy_mode = { + .enter = paravirt_enter_lazy_mmu, + .leave = xen_leave_lazy, + }, }; #ifdef CONFIG_SMP @@ -1114,7 +1099,6 @@ asmlinkage void __init xen_start_kernel(void) pv_irq_ops = xen_irq_ops; pv_apic_ops = xen_apic_ops; pv_mmu_ops = xen_mmu_ops; - pv_misc_ops = xen_misc_ops; machine_ops = xen_machine_ops; diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 874db0cd1d2..2061bdd3e7d 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -155,7 +155,7 @@ void xen_set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pteval) { if (mm == current->mm || mm == &init_mm) { - if (xen_get_lazy_mode() == PARAVIRT_LAZY_MMU) { + if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU) { struct multicall_space mcs; mcs = xen_mc_entry(0); diff --git a/arch/x86/xen/multicalls.h b/arch/x86/xen/multicalls.h index e6f7530b156..5d96a5fa210 100644 --- a/arch/x86/xen/multicalls.h +++ b/arch/x86/xen/multicalls.h @@ -35,7 +35,7 @@ void xen_mc_flush(void); /* Issue a multicall if we're not in a lazy mode */ static inline void xen_mc_issue(unsigned mode) { - if ((xen_get_lazy_mode() & mode) == 0) + if ((paravirt_get_lazy_mode() & mode) == 0) xen_mc_flush(); /* restore flags saved in xen_mc_batch */ diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h index b9aaea45f07..b5697bae52d 100644 --- a/arch/x86/xen/xen-ops.h +++ b/arch/x86/xen/xen-ops.h @@ -29,13 +29,6 @@ unsigned long long xen_sched_clock(void); void xen_mark_init_mm_pinned(void); -DECLARE_PER_CPU(enum paravirt_lazy_mode, xen_lazy_mode); - -static inline unsigned xen_get_lazy_mode(void) -{ - return x86_read_percpu(xen_lazy_mode); -} - void __init xen_fill_possible_map(void); void __init xen_setup_vcpu_info_placement(void); diff --git a/drivers/lguest/lguest.c b/drivers/lguest/lguest.c index ca9b844f37c..c302629e089 100644 --- a/drivers/lguest/lguest.c +++ b/drivers/lguest/lguest.c @@ -97,29 +97,17 @@ static cycle_t clock_base; * them as a batch when lazy_mode is eventually turned off. Because hypercalls * are reasonably expensive, batching them up makes sense. For example, a * large mmap might update dozens of page table entries: that code calls - * lguest_lazy_mode(PARAVIRT_LAZY_MMU), does the dozen updates, then calls - * lguest_lazy_mode(PARAVIRT_LAZY_NONE). + * paravirt_enter_lazy_mmu(), does the dozen updates, then calls + * lguest_leave_lazy_mode(). * * So, when we're in lazy mode, we call async_hypercall() to store the call for * future processing. When lazy mode is turned off we issue a hypercall to * flush the stored calls. - * - * There's also a hack where "mode" is set to "PARAVIRT_LAZY_FLUSH" which - * indicates we're to flush any outstanding calls immediately. This is used - * when an interrupt handler does a kmap_atomic(): the page table changes must - * happen immediately even if we're in the middle of a batch. Usually we're - * not, though, so there's nothing to do. */ -static enum paravirt_lazy_mode lazy_mode; /* Note: not SMP-safe! */ -static void lguest_lazy_mode(enum paravirt_lazy_mode mode) + */ +static void lguest_leave_lazy_mode(void) { - if (mode == PARAVIRT_LAZY_FLUSH) { - if (unlikely(lazy_mode != PARAVIRT_LAZY_NONE)) - hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0); - } else { - lazy_mode = mode; - if (mode == PARAVIRT_LAZY_NONE) - hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0); - } + paravirt_leave_lazy(paravirt_get_lazy_mode()); + hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0); } static void lazy_hcall(unsigned long call, @@ -127,7 +115,7 @@ static void lazy_hcall(unsigned long call, unsigned long arg2, unsigned long arg3) { - if (lazy_mode == PARAVIRT_LAZY_NONE) + if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) hcall(call, arg1, arg2, arg3); else async_hcall(call, arg1, arg2, arg3); @@ -1011,6 +999,8 @@ __init void lguest_init(void *boot) pv_cpu_ops.write_gdt_entry = lguest_write_gdt_entry; pv_cpu_ops.write_idt_entry = lguest_write_idt_entry; pv_cpu_ops.wbinvd = lguest_wbinvd; + pv_cpu_ops.lazy_mode.enter = paravirt_enter_lazy_cpu; + pv_cpu_ops.lazy_mode.leave = lguest_leave_lazy_mode; /* pagetable management */ pv_mmu_ops.write_cr3 = lguest_write_cr3; @@ -1022,6 +1012,8 @@ __init void lguest_init(void *boot) pv_mmu_ops.set_pmd = lguest_set_pmd; pv_mmu_ops.read_cr2 = lguest_read_cr2; pv_mmu_ops.read_cr3 = lguest_read_cr3; + pv_mmu_ops.lazy_mode.enter = paravirt_enter_lazy_mmu; + pv_mmu_ops.lazy_mode.leave = lguest_leave_lazy_mode; #ifdef CONFIG_X86_LOCAL_APIC /* apic read/write intercepts */ @@ -1034,8 +1026,6 @@ __init void lguest_init(void *boot) pv_time_ops.get_wallclock = lguest_get_wallclock; pv_time_ops.time_init = lguest_time_init; - pv_misc_ops.set_lazy_mode = lguest_lazy_mode; - /* Now is a good time to look at the implementations of these functions * before returning to the rest of lguest_init(). */ diff --git a/include/asm-x86/paravirt.h b/include/asm-x86/paravirt.h index 19726e12051..f59d370c5df 100644 --- a/include/asm-x86/paravirt.h +++ b/include/asm-x86/paravirt.h @@ -25,15 +25,6 @@ struct tss_struct; struct mm_struct; struct desc_struct; -/* Lazy mode for batching updates / context switch */ -enum paravirt_lazy_mode { - PARAVIRT_LAZY_NONE = 0, - PARAVIRT_LAZY_MMU = 1, - PARAVIRT_LAZY_CPU = 2, - PARAVIRT_LAZY_FLUSH = 3, -}; - - /* general info */ struct pv_info { unsigned int kernel_rpl; @@ -64,9 +55,10 @@ struct pv_init_ops { }; -struct pv_misc_ops { +struct pv_lazy_ops { /* Set deferred update mode, used for batching operations. */ - void (*set_lazy_mode)(enum paravirt_lazy_mode mode); + void (*enter)(void); + void (*leave)(void); }; struct pv_time_ops { @@ -131,6 +123,8 @@ struct pv_cpu_ops { /* These two are jmp to, not actually called. */ void (*irq_enable_sysexit)(void); void (*iret)(void); + + struct pv_lazy_ops lazy_mode; }; struct pv_irq_ops { @@ -244,6 +238,8 @@ struct pv_mmu_ops { #ifdef CONFIG_HIGHPTE void *(*kmap_atomic_pte)(struct page *page, enum km_type type); #endif + + struct pv_lazy_ops lazy_mode; }; /* This contains all the paravirt structures: we get a convenient @@ -252,7 +248,6 @@ struct pv_mmu_ops { struct paravirt_patch_template { struct pv_init_ops pv_init_ops; - struct pv_misc_ops pv_misc_ops; struct pv_time_ops pv_time_ops; struct pv_cpu_ops pv_cpu_ops; struct pv_irq_ops pv_irq_ops; @@ -262,7 +257,6 @@ struct paravirt_patch_template extern struct pv_info pv_info; extern struct pv_init_ops pv_init_ops; -extern struct pv_misc_ops pv_misc_ops; extern struct pv_time_ops pv_time_ops; extern struct pv_cpu_ops pv_cpu_ops; extern struct pv_irq_ops pv_irq_ops; @@ -953,37 +947,57 @@ static inline void set_pmd(pmd_t *pmdp, pmd_t pmdval) } #endif /* CONFIG_X86_PAE */ +/* Lazy mode for batching updates / context switch */ +enum paravirt_lazy_mode { + PARAVIRT_LAZY_NONE, + PARAVIRT_LAZY_MMU, + PARAVIRT_LAZY_CPU, +}; + +enum paravirt_lazy_mode paravirt_get_lazy_mode(void); +void paravirt_enter_lazy_cpu(void); +void paravirt_leave_lazy_cpu(void); +void paravirt_enter_lazy_mmu(void); +void paravirt_leave_lazy_mmu(void); +void paravirt_leave_lazy(enum paravirt_lazy_mode mode); + #define __HAVE_ARCH_ENTER_LAZY_CPU_MODE static inline void arch_enter_lazy_cpu_mode(void) { - PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_CPU); + PVOP_VCALL0(pv_cpu_ops.lazy_mode.enter); } static inline void arch_leave_lazy_cpu_mode(void) { - PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_NONE); + PVOP_VCALL0(pv_cpu_ops.lazy_mode.leave); } static inline void arch_flush_lazy_cpu_mode(void) { - PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_FLUSH); + if (unlikely(paravirt_get_lazy_mode() == PARAVIRT_LAZY_CPU)) { + arch_leave_lazy_cpu_mode(); + arch_enter_lazy_cpu_mode(); + } } #define __HAVE_ARCH_ENTER_LAZY_MMU_MODE static inline void arch_enter_lazy_mmu_mode(void) { - PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_MMU); + PVOP_VCALL0(pv_mmu_ops.lazy_mode.enter); } static inline void arch_leave_lazy_mmu_mode(void) { - PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_NONE); + PVOP_VCALL0(pv_mmu_ops.lazy_mode.leave); } static inline void arch_flush_lazy_mmu_mode(void) { - PVOP_VCALL1(pv_misc_ops.set_lazy_mode, PARAVIRT_LAZY_FLUSH); + if (unlikely(paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU)) { + arch_leave_lazy_mmu_mode(); + arch_enter_lazy_mmu_mode(); + } } void _paravirt_nop(void); From 4f8178477400cfc49a0657b96e873914d95081f8 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:29 -0700 Subject: [PATCH 03/12] remove dead code in pgtable_cache_init The conversion from using a slab cache to quicklist left some residual dead code. I note that in the conversion it now always allocates a whole page for the pgd, rather than the 32 bytes needed for a PAE pgd. Was this intended? Signed-off-by: Jeremy Fitzhardinge Cc: Christoph Lameter Cc: Andi Kleen Cc: William Lee Irwin III Cc: Ingo Molnar Cc: Thomas Gleixner --- arch/x86/mm/init_32.c | 22 +++++----------------- 1 file changed, 5 insertions(+), 17 deletions(-) diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 730a5b177b1..aa4708fda3e 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -746,24 +746,12 @@ struct kmem_cache *pmd_cache; void __init pgtable_cache_init(void) { - size_t pgd_size = PTRS_PER_PGD*sizeof(pgd_t); - - if (PTRS_PER_PMD > 1) { + if (PTRS_PER_PMD > 1) pmd_cache = kmem_cache_create("pmd", - PTRS_PER_PMD*sizeof(pmd_t), - PTRS_PER_PMD*sizeof(pmd_t), - SLAB_PANIC, - pmd_ctor); - if (!SHARED_KERNEL_PMD) { - /* If we're in PAE mode and have a non-shared - kernel pmd, then the pgd size must be a - page size. This is because the pgd_list - links through the page structure, so there - can only be one pgd per page for this to - work. */ - pgd_size = PAGE_SIZE; - } - } + PTRS_PER_PMD*sizeof(pmd_t), + PTRS_PER_PMD*sizeof(pmd_t), + SLAB_PANIC, + pmd_ctor); } /* From d626a1f1cbbdfac90c529216e40a4fa2a22ecc3d Mon Sep 17 00:00:00 2001 From: Jesper Juhl Date: Tue, 16 Oct 2007 11:51:29 -0700 Subject: [PATCH 04/12] Clean up duplicate includes in arch/i386/xen/ This patch cleans up duplicate includes in arch/i386/xen/ Signed-off-by: Jesper Juhl Signed-off-by: Jeremy Fitzhardinge --- arch/x86/xen/enlighten.c | 1 - arch/x86/xen/mmu.c | 1 - 2 files changed, 2 deletions(-) diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 7171a073607..75fd36971d0 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -25,7 +25,6 @@ #include #include #include -#include #include #include diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 2061bdd3e7d..c4a391f8898 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -41,7 +41,6 @@ #include #include #include -#include #include #include From f0d733942750c1ee6358c3a4a1a5d7ba73b7122f Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:30 -0700 Subject: [PATCH 05/12] xen: yield to IPI target if necessary When sending a call-function IPI to a vcpu, yield if the vcpu isn't running. Signed-off-by: Jeremy Fitzhardinge --- arch/x86/xen/smp.c | 14 ++++++++++---- arch/x86/xen/time.c | 6 ++++++ arch/x86/xen/xen-ops.h | 2 ++ 3 files changed, 18 insertions(+), 4 deletions(-) diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c index 557b8e24706..865953e6f34 100644 --- a/arch/x86/xen/smp.c +++ b/arch/x86/xen/smp.c @@ -360,7 +360,8 @@ int xen_smp_call_function_mask(cpumask_t mask, void (*func)(void *), void *info, int wait) { struct call_data_struct data; - int cpus; + int cpus, cpu; + bool yield; /* Holding any lock stops cpus from going down. */ spin_lock(&call_lock); @@ -389,9 +390,14 @@ int xen_smp_call_function_mask(cpumask_t mask, void (*func)(void *), /* Send a message to other CPUs and wait for them to respond */ xen_send_IPI_mask(mask, XEN_CALL_FUNCTION_VECTOR); - /* Make sure other vcpus get a chance to run. - XXX too severe? Maybe we should check the other CPU's states? */ - HYPERVISOR_sched_op(SCHEDOP_yield, 0); + /* Make sure other vcpus get a chance to run if they need to. */ + yield = false; + for_each_cpu_mask(cpu, mask) + if (xen_vcpu_stolen(cpu)) + yield = true; + + if (yield) + HYPERVISOR_sched_op(SCHEDOP_yield, 0); /* Wait for response */ while (atomic_read(&data.started) != cpus || diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index dfd6db69ead..d083ff5ef08 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -105,6 +105,12 @@ static void get_runstate_snapshot(struct vcpu_runstate_info *res) } while (get64(&state->state_entry_time) != state_time); } +/* return true when a vcpu could run but has no real cpu to run on */ +bool xen_vcpu_stolen(int vcpu) +{ + return per_cpu(runstate, vcpu).state == RUNSTATE_runnable; +} + static void setup_runstate_info(int cpu) { struct vcpu_register_runstate_memory_area area; diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h index b5697bae52d..3847eed0bb0 100644 --- a/arch/x86/xen/xen-ops.h +++ b/arch/x86/xen/xen-ops.h @@ -27,6 +27,8 @@ unsigned long xen_get_wallclock(void); int xen_set_wallclock(unsigned long time); unsigned long long xen_sched_clock(void); +bool xen_vcpu_stolen(int vcpu); + void xen_mark_init_mm_pinned(void); void __init xen_fill_possible_map(void); From 91e0c5f3dad47838cb2ecc1865ce789a0b7182b1 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:30 -0700 Subject: [PATCH 06/12] xen: add batch completion callbacks This adds a mechanism to register a callback function to be called once a batch of hypercalls has been issued. This is typically used to unlock things which must remain locked until the hypercall has taken place. [ Stable folks: pre-req for 2.6.23 bugfix "xen: deal with stale cr3 values when unpinning pagetables" ] Signed-off-by: Jeremy Fitzhardinge Cc: Stable Kernel --- arch/x86/xen/multicalls.c | 29 ++++++++++++++++++++++++++--- arch/x86/xen/multicalls.h | 3 +++ 2 files changed, 29 insertions(+), 3 deletions(-) diff --git a/arch/x86/xen/multicalls.c b/arch/x86/xen/multicalls.c index c837e8e463d..ce9c4b41f02 100644 --- a/arch/x86/xen/multicalls.c +++ b/arch/x86/xen/multicalls.c @@ -32,7 +32,11 @@ struct mc_buffer { struct multicall_entry entries[MC_BATCH]; u64 args[MC_ARGS]; - unsigned mcidx, argidx; + struct callback { + void (*fn)(void *); + void *data; + } callbacks[MC_BATCH]; + unsigned mcidx, argidx, cbidx; }; static DEFINE_PER_CPU(struct mc_buffer, mc_buffer); @@ -43,6 +47,7 @@ void xen_mc_flush(void) struct mc_buffer *b = &__get_cpu_var(mc_buffer); int ret = 0; unsigned long flags; + int i; BUG_ON(preemptible()); @@ -51,8 +56,6 @@ void xen_mc_flush(void) local_irq_save(flags); if (b->mcidx) { - int i; - if (HYPERVISOR_multicall(b->entries, b->mcidx) != 0) BUG(); for (i = 0; i < b->mcidx; i++) @@ -65,6 +68,13 @@ void xen_mc_flush(void) local_irq_restore(flags); + for(i = 0; i < b->cbidx; i++) { + struct callback *cb = &b->callbacks[i]; + + (*cb->fn)(cb->data); + } + b->cbidx = 0; + BUG_ON(ret); } @@ -88,3 +98,16 @@ struct multicall_space __xen_mc_entry(size_t args) return ret; } + +void xen_mc_callback(void (*fn)(void *), void *data) +{ + struct mc_buffer *b = &__get_cpu_var(mc_buffer); + struct callback *cb; + + if (b->cbidx == MC_BATCH) + xen_mc_flush(); + + cb = &b->callbacks[b->cbidx++]; + cb->fn = fn; + cb->data = data; +} diff --git a/arch/x86/xen/multicalls.h b/arch/x86/xen/multicalls.h index 5d96a5fa210..8bae996d99a 100644 --- a/arch/x86/xen/multicalls.h +++ b/arch/x86/xen/multicalls.h @@ -42,4 +42,7 @@ static inline void xen_mc_issue(unsigned mode) local_irq_restore(x86_read_percpu(xen_mc_irq_flags)); } +/* Set up a callback to be called when the current batch is flushed */ +void xen_mc_callback(void (*fn)(void *), void *data); + #endif /* _XEN_MULTICALLS_H */ From 9f79991d4186089e228274196413572cc000143b Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:30 -0700 Subject: [PATCH 07/12] xen: deal with stale cr3 values when unpinning pagetables When a pagetable is no longer in use, it must be unpinned so that its pages can be freed. However, this is only possible if there are no stray uses of the pagetable. The code currently deals with all the usual cases, but there's a rare case where a vcpu is changing cr3, but is doing so lazily, and the change hasn't actually happened by the time the pagetable is unpinned, even though it appears to have been completed. This change adds a second per-cpu cr3 variable - xen_current_cr3 - which tracks the actual state of the vcpu cr3. It is only updated once the actual hypercall to set cr3 has been completed. Other processors wishing to unpin a pagetable can check other vcpu's xen_current_cr3 values to see if any cross-cpu IPIs are needed to clean things up. [ Stable folks: 2.6.23 bugfix ] Signed-off-by: Jeremy Fitzhardinge Cc: Stable Kernel --- arch/x86/xen/enlighten.c | 55 +++++++++++++++++++++++++++------------- arch/x86/xen/mmu.c | 29 ++++++++++++++++++--- arch/x86/xen/xen-ops.h | 1 + 3 files changed, 65 insertions(+), 20 deletions(-) diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 75fd36971d0..e334bf7cb32 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -53,7 +53,23 @@ EXPORT_SYMBOL_GPL(hypercall_page); DEFINE_PER_CPU(struct vcpu_info *, xen_vcpu); DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info); -DEFINE_PER_CPU(unsigned long, xen_cr3); + +/* + * Note about cr3 (pagetable base) values: + * + * xen_cr3 contains the current logical cr3 value; it contains the + * last set cr3. This may not be the current effective cr3, because + * its update may be being lazily deferred. However, a vcpu looking + * at its own cr3 can use this value knowing that it everything will + * be self-consistent. + * + * xen_current_cr3 contains the actual vcpu cr3; it is set once the + * hypercall to set the vcpu cr3 is complete (so it may be a little + * out of date, but it will never be set early). If one vcpu is + * looking at another vcpu's cr3 value, it should use this variable. + */ +DEFINE_PER_CPU(unsigned long, xen_cr3); /* cr3 stored as physaddr */ +DEFINE_PER_CPU(unsigned long, xen_current_cr3); /* actual vcpu cr3 */ struct start_info *xen_start_info; EXPORT_SYMBOL_GPL(xen_start_info); @@ -610,32 +626,36 @@ static unsigned long xen_read_cr3(void) return x86_read_percpu(xen_cr3); } +static void set_current_cr3(void *v) +{ + x86_write_percpu(xen_current_cr3, (unsigned long)v); +} + static void xen_write_cr3(unsigned long cr3) { + struct mmuext_op *op; + struct multicall_space mcs; + unsigned long mfn = pfn_to_mfn(PFN_DOWN(cr3)); + BUG_ON(preemptible()); - if (cr3 == x86_read_percpu(xen_cr3)) { - /* just a simple tlb flush */ - xen_flush_tlb(); - return; - } + mcs = xen_mc_entry(sizeof(*op)); /* disables interrupts */ + /* Update while interrupts are disabled, so its atomic with + respect to ipis */ x86_write_percpu(xen_cr3, cr3); + op = mcs.args; + op->cmd = MMUEXT_NEW_BASEPTR; + op->arg1.mfn = mfn; - { - struct mmuext_op *op; - struct multicall_space mcs = xen_mc_entry(sizeof(*op)); - unsigned long mfn = pfn_to_mfn(PFN_DOWN(cr3)); + MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF); - op = mcs.args; - op->cmd = MMUEXT_NEW_BASEPTR; - op->arg1.mfn = mfn; + /* Update xen_update_cr3 once the batch has actually + been submitted. */ + xen_mc_callback(set_current_cr3, (void *)cr3); - MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF); - - xen_mc_issue(PARAVIRT_LAZY_CPU); - } + xen_mc_issue(PARAVIRT_LAZY_CPU); /* interrupts restored */ } /* Early in boot, while setting up the initial pagetable, assume @@ -1120,6 +1140,7 @@ asmlinkage void __init xen_start_kernel(void) /* keep using Xen gdt for now; no urgent need to change it */ x86_write_percpu(xen_cr3, __pa(pgd)); + x86_write_percpu(xen_current_cr3, __pa(pgd)); #ifdef CONFIG_SMP /* Don't do the full vcpu_info placement stuff until we have a diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index c4a391f8898..72f08ab43a4 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -514,20 +514,43 @@ static void drop_other_mm_ref(void *info) if (__get_cpu_var(cpu_tlbstate).active_mm == mm) leave_mm(smp_processor_id()); + + /* If this cpu still has a stale cr3 reference, then make sure + it has been flushed. */ + if (x86_read_percpu(xen_current_cr3) == __pa(mm->pgd)) { + load_cr3(swapper_pg_dir); + arch_flush_lazy_cpu_mode(); + } } static void drop_mm_ref(struct mm_struct *mm) { + cpumask_t mask; + unsigned cpu; + if (current->active_mm == mm) { if (current->mm == mm) load_cr3(swapper_pg_dir); else leave_mm(smp_processor_id()); + arch_flush_lazy_cpu_mode(); } - if (!cpus_empty(mm->cpu_vm_mask)) - xen_smp_call_function_mask(mm->cpu_vm_mask, drop_other_mm_ref, - mm, 1); + /* Get the "official" set of cpus referring to our pagetable. */ + mask = mm->cpu_vm_mask; + + /* It's possible that a vcpu may have a stale reference to our + cr3, because its in lazy mode, and it hasn't yet flushed + its set of pending hypercalls yet. In this case, we can + look at its actual current cr3 value, and force it to flush + if needed. */ + for_each_online_cpu(cpu) { + if (per_cpu(xen_current_cr3, cpu) == __pa(mm->pgd)) + cpu_set(cpu, mask); + } + + if (!cpus_empty(mask)) + xen_smp_call_function_mask(mask, drop_other_mm_ref, mm, 1); } #else static void drop_mm_ref(struct mm_struct *mm) diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h index 3847eed0bb0..b02a909bfd4 100644 --- a/arch/x86/xen/xen-ops.h +++ b/arch/x86/xen/xen-ops.h @@ -11,6 +11,7 @@ void xen_copy_trap_info(struct trap_info *traps); DECLARE_PER_CPU(struct vcpu_info *, xen_vcpu); DECLARE_PER_CPU(unsigned long, xen_cr3); +DECLARE_PER_CPU(unsigned long, xen_current_cr3); extern struct start_info *xen_start_info; extern struct shared_info *HYPERVISOR_shared_info; From 74260714c56de4f967fcb2f17a8656bc574b75be Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:30 -0700 Subject: [PATCH 08/12] xen: lock pte pages while pinning/unpinning When a pagetable is created, it is made globally visible in the rmap prio tree before it is pinned via arch_dup_mmap(), and remains in the rmap tree while it is unpinned with arch_exit_mmap(). This means that other CPUs may race with the pinning/unpinning process, and see a pte between when it gets marked RO and actually pinned, causing any pte updates to fail with write-protect faults. As a result, all pte pages must be properly locked, and only unlocked once the pinning/unpinning process has finished. In order to avoid taking spinlocks for the whole pagetable - which may overflow the PREEMPT_BITS portion of preempt counter - it locks and pins each pte page individually, and then finally pins the whole pagetable. Signed-off-by: Jeremy Fitzhardinge Cc: Rik van Riel Cc: Hugh Dickens Cc: David Rientjes Cc: Andrew Morton Cc: Andi Kleen Cc: Keir Fraser Cc: Jan Beulich --- arch/x86/xen/enlighten.c | 30 +++++++---- arch/x86/xen/mmu.c | 113 ++++++++++++++++++++++++++++----------- mm/Kconfig | 1 - 3 files changed, 103 insertions(+), 41 deletions(-) diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index e334bf7cb32..4186cb6a7f5 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -666,6 +666,15 @@ static __init void xen_alloc_pt_init(struct mm_struct *mm, u32 pfn) make_lowmem_page_readonly(__va(PFN_PHYS(pfn))); } +static void pin_pagetable_pfn(unsigned level, unsigned long pfn) +{ + struct mmuext_op op; + op.cmd = level; + op.arg1.mfn = pfn_to_mfn(pfn); + if (HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF)) + BUG(); +} + /* This needs to make sure the new pte page is pinned iff its being attached to a pinned pagetable. */ static void xen_alloc_pt(struct mm_struct *mm, u32 pfn) @@ -675,9 +684,10 @@ static void xen_alloc_pt(struct mm_struct *mm, u32 pfn) if (PagePinned(virt_to_page(mm->pgd))) { SetPagePinned(page); - if (!PageHighMem(page)) + if (!PageHighMem(page)) { make_lowmem_page_readonly(__va(PFN_PHYS(pfn))); - else + pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn); + } else /* make sure there are no stray mappings of this page */ kmap_flush_unused(); @@ -690,8 +700,10 @@ static void xen_release_pt(u32 pfn) struct page *page = pfn_to_page(pfn); if (PagePinned(page)) { - if (!PageHighMem(page)) + if (!PageHighMem(page)) { + pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn); make_lowmem_page_readwrite(__va(PFN_PHYS(pfn))); + } } } @@ -806,15 +818,15 @@ static __init void xen_pagetable_setup_done(pgd_t *base) /* Actually pin the pagetable down, but we can't set PG_pinned yet because the page structures don't exist yet. */ { - struct mmuext_op op; + unsigned level; + #ifdef CONFIG_X86_PAE - op.cmd = MMUEXT_PIN_L3_TABLE; + level = MMUEXT_PIN_L3_TABLE; #else - op.cmd = MMUEXT_PIN_L3_TABLE; + level = MMUEXT_PIN_L2_TABLE; #endif - op.arg1.mfn = pfn_to_mfn(PFN_DOWN(__pa(base))); - if (HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF)) - BUG(); + + pin_pagetable_pfn(level, PFN_DOWN(__pa(base))); } } diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 72f08ab43a4..b2e32f9d007 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -303,7 +303,12 @@ pgd_t xen_make_pgd(unsigned long pgd) } #endif /* CONFIG_X86_PAE */ - +enum pt_level { + PT_PGD, + PT_PUD, + PT_PMD, + PT_PTE +}; /* (Yet another) pagetable walker. This one is intended for pinning a @@ -315,7 +320,7 @@ pgd_t xen_make_pgd(unsigned long pgd) FIXADDR_TOP. But the important bit is that we don't pin beyond there, because then we start getting into Xen's ptes. */ -static int pgd_walk(pgd_t *pgd_base, int (*func)(struct page *, unsigned), +static int pgd_walk(pgd_t *pgd_base, int (*func)(struct page *, enum pt_level), unsigned long limit) { pgd_t *pgd = pgd_base; @@ -340,7 +345,7 @@ static int pgd_walk(pgd_t *pgd_base, int (*func)(struct page *, unsigned), pud = pud_offset(pgd, 0); if (PTRS_PER_PUD > 1) /* not folded */ - flush |= (*func)(virt_to_page(pud), 0); + flush |= (*func)(virt_to_page(pud), PT_PUD); for (; addr != pud_limit; pud++, addr = pud_next) { pmd_t *pmd; @@ -359,7 +364,7 @@ static int pgd_walk(pgd_t *pgd_base, int (*func)(struct page *, unsigned), pmd = pmd_offset(pud, 0); if (PTRS_PER_PMD > 1) /* not folded */ - flush |= (*func)(virt_to_page(pmd), 0); + flush |= (*func)(virt_to_page(pmd), PT_PMD); for (; addr != pmd_limit; pmd++) { addr += (PAGE_SIZE * PTRS_PER_PTE); @@ -371,17 +376,47 @@ static int pgd_walk(pgd_t *pgd_base, int (*func)(struct page *, unsigned), if (pmd_none(*pmd)) continue; - flush |= (*func)(pmd_page(*pmd), 0); + flush |= (*func)(pmd_page(*pmd), PT_PTE); } } } - flush |= (*func)(virt_to_page(pgd_base), UVMF_TLB_FLUSH); + flush |= (*func)(virt_to_page(pgd_base), PT_PGD); return flush; } -static int pin_page(struct page *page, unsigned flags) +static spinlock_t *lock_pte(struct page *page) +{ + spinlock_t *ptl = NULL; + +#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS + ptl = __pte_lockptr(page); + spin_lock(ptl); +#endif + + return ptl; +} + +static void do_unlock(void *v) +{ + spinlock_t *ptl = v; + spin_unlock(ptl); +} + +static void xen_do_pin(unsigned level, unsigned long pfn) +{ + struct mmuext_op *op; + struct multicall_space mcs; + + mcs = __xen_mc_entry(sizeof(*op)); + op = mcs.args; + op->cmd = level; + op->arg1.mfn = pfn_to_mfn(pfn); + MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF); +} + +static int pin_page(struct page *page, enum pt_level level) { unsigned pgfl = test_and_set_bit(PG_pinned, &page->flags); int flush; @@ -396,12 +431,26 @@ static int pin_page(struct page *page, unsigned flags) void *pt = lowmem_page_address(page); unsigned long pfn = page_to_pfn(page); struct multicall_space mcs = __xen_mc_entry(0); + spinlock_t *ptl; flush = 0; + ptl = NULL; + if (level == PT_PTE) + ptl = lock_pte(page); + MULTI_update_va_mapping(mcs.mc, (unsigned long)pt, pfn_pte(pfn, PAGE_KERNEL_RO), - flags); + level == PT_PGD ? UVMF_TLB_FLUSH : 0); + + if (level == PT_PTE) + xen_do_pin(MMUEXT_PIN_L1_TABLE, pfn); + + if (ptl) { + /* Queue a deferred unlock for when this batch + is completed. */ + xen_mc_callback(do_unlock, ptl); + } } return flush; @@ -412,8 +461,7 @@ static int pin_page(struct page *page, unsigned flags) read-only, and can be pinned. */ void xen_pgd_pin(pgd_t *pgd) { - struct multicall_space mcs; - struct mmuext_op *op; + unsigned level; xen_mc_batch(); @@ -424,16 +472,13 @@ void xen_pgd_pin(pgd_t *pgd) xen_mc_batch(); } - mcs = __xen_mc_entry(sizeof(*op)); - op = mcs.args; - #ifdef CONFIG_X86_PAE - op->cmd = MMUEXT_PIN_L3_TABLE; + level = MMUEXT_PIN_L3_TABLE; #else - op->cmd = MMUEXT_PIN_L2_TABLE; + level = MMUEXT_PIN_L2_TABLE; #endif - op->arg1.mfn = pfn_to_mfn(PFN_DOWN(__pa(pgd))); - MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF); + + xen_do_pin(level, PFN_DOWN(__pa(pgd))); xen_mc_issue(0); } @@ -441,7 +486,7 @@ void xen_pgd_pin(pgd_t *pgd) /* The init_mm pagetable is really pinned as soon as its created, but that's before we have page structures to store the bits. So do all the book-keeping now. */ -static __init int mark_pinned(struct page *page, unsigned flags) +static __init int mark_pinned(struct page *page, enum pt_level level) { SetPagePinned(page); return 0; @@ -452,18 +497,32 @@ void __init xen_mark_init_mm_pinned(void) pgd_walk(init_mm.pgd, mark_pinned, FIXADDR_TOP); } -static int unpin_page(struct page *page, unsigned flags) +static int unpin_page(struct page *page, enum pt_level level) { unsigned pgfl = test_and_clear_bit(PG_pinned, &page->flags); if (pgfl && !PageHighMem(page)) { void *pt = lowmem_page_address(page); unsigned long pfn = page_to_pfn(page); - struct multicall_space mcs = __xen_mc_entry(0); + spinlock_t *ptl = NULL; + struct multicall_space mcs; + + if (level == PT_PTE) { + ptl = lock_pte(page); + + xen_do_pin(MMUEXT_UNPIN_TABLE, pfn); + } + + mcs = __xen_mc_entry(0); MULTI_update_va_mapping(mcs.mc, (unsigned long)pt, pfn_pte(pfn, PAGE_KERNEL), - flags); + level == PT_PGD ? UVMF_TLB_FLUSH : 0); + + if (ptl) { + /* unlock when batch completed */ + xen_mc_callback(do_unlock, ptl); + } } return 0; /* never need to flush on unpin */ @@ -472,18 +531,9 @@ static int unpin_page(struct page *page, unsigned flags) /* Release a pagetables pages back as normal RW */ static void xen_pgd_unpin(pgd_t *pgd) { - struct mmuext_op *op; - struct multicall_space mcs; - xen_mc_batch(); - mcs = __xen_mc_entry(sizeof(*op)); - - op = mcs.args; - op->cmd = MMUEXT_UNPIN_TABLE; - op->arg1.mfn = pfn_to_mfn(PFN_DOWN(__pa(pgd))); - - MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF); + xen_do_pin(MMUEXT_UNPIN_TABLE, PFN_DOWN(__pa(pgd))); pgd_walk(pgd, unpin_page, TASK_SIZE); @@ -585,5 +635,6 @@ void xen_exit_mmap(struct mm_struct *mm) /* pgd may not be pinned in the error exit path of execve */ if (PagePinned(virt_to_page(mm->pgd))) xen_pgd_unpin(mm->pgd); + spin_unlock(&mm->page_table_lock); } diff --git a/mm/Kconfig b/mm/Kconfig index a7609cbcb00..e24d348083c 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -137,7 +137,6 @@ config SPLIT_PTLOCK_CPUS int default "4096" if ARM && !CPU_CACHE_VIPT default "4096" if PARISC && !PA20 - default "4096" if XEN default "4" # From fb1d84043ca73212b08ff57608f51b372529e6d6 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:31 -0700 Subject: [PATCH 09/12] xen: ask the hypervisor how much space it needs reserved Ask the hypervisor how much space it needs reserved, since 32-on-64 doesn't need any space, and it may change in future. Signed-off-by: Jeremy Fitzhardinge --- arch/x86/xen/enlighten.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 4186cb6a7f5..bc7bf5f8d07 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -1112,6 +1112,17 @@ static const struct machine_ops __initdata xen_machine_ops = { }; +static void __init xen_reserve_top(void) +{ + unsigned long top = HYPERVISOR_VIRT_START; + struct xen_platform_parameters pp; + + if (HYPERVISOR_xen_version(XENVER_platform_parameters, &pp) == 0) + top = pp.virt_start; + + reserve_top_address(-top + 2 * PAGE_SIZE); +} + /* First C function to be called on Xen boot */ asmlinkage void __init xen_start_kernel(void) { @@ -1169,7 +1180,7 @@ asmlinkage void __init xen_start_kernel(void) pv_info.kernel_rpl = 0; /* set the limit of our address space */ - reserve_top_address(-HYPERVISOR_VIRT_START + 2 * PAGE_SIZE); + xen_reserve_top(); /* set up basic CPUID stuff */ cpu_detect(&new_cpu_data); From e3d2697669abbe26c08dc9b95e2a71c634d096ed Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:31 -0700 Subject: [PATCH 10/12] xen: fix incorrect vcpu_register_vcpu_info hypercall argument The kernel's copy of struct vcpu_register_vcpu_info was out of date, at best causing the hypercall to fail and the guest kernel to fall back to the old mechanism, or worse, causing random memory corruption. [ Stable folks: applies to 2.6.23 ] Signed-off-by: Jeremy Fitzhardinge Cc: Stable Kernel Cc: Morten =?utf-8?q?B=C3=B8geskov?= Cc: Mark Williamson --- arch/x86/xen/enlighten.c | 2 +- include/xen/interface/vcpu.h | 5 +++-- 2 files changed, 4 insertions(+), 3 deletions(-) diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index bc7bf5f8d07..c89e5b407f9 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -113,7 +113,7 @@ static void __init xen_vcpu_setup(int cpu) info.mfn = virt_to_mfn(vcpup); info.offset = offset_in_page(vcpup); - printk(KERN_DEBUG "trying to map vcpu_info %d at %p, mfn %x, offset %d\n", + printk(KERN_DEBUG "trying to map vcpu_info %d at %p, mfn %llx, offset %d\n", cpu, vcpup, info.mfn, info.offset); /* Check to see if the hypervisor will put the vcpu_info diff --git a/include/xen/interface/vcpu.h b/include/xen/interface/vcpu.h index ff61ea36599..b05d8a6d914 100644 --- a/include/xen/interface/vcpu.h +++ b/include/xen/interface/vcpu.h @@ -160,8 +160,9 @@ struct vcpu_set_singleshot_timer { */ #define VCPUOP_register_vcpu_info 10 /* arg == struct vcpu_info */ struct vcpu_register_vcpu_info { - uint32_t mfn; /* mfn of page to place vcpu_info */ - uint32_t offset; /* offset within page */ + uint64_t mfn; /* mfn of page to place vcpu_info */ + uint32_t offset; /* offset within page */ + uint32_t rsvd; /* unused */ }; #endif /* __XEN_PUBLIC_VCPU_H__ */ From a122d6230e8d8ac7cffdf0bc9cc4b256b928fe49 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:31 -0700 Subject: [PATCH 11/12] xen: add some debug output for failed multicalls Multicalls are expected to never fail, and the normal response to a failed multicall is very terse. In the interests of better debuggability, add some more verbose output. It may be worth turning this off once it all seems more tested. Signed-off-by: Jeremy Fitzhardinge --- arch/x86/xen/multicalls.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) diff --git a/arch/x86/xen/multicalls.c b/arch/x86/xen/multicalls.c index ce9c4b41f02..5e6f36f6d87 100644 --- a/arch/x86/xen/multicalls.c +++ b/arch/x86/xen/multicalls.c @@ -26,11 +26,16 @@ #include "multicalls.h" +#define MC_DEBUG 1 + #define MC_BATCH 32 #define MC_ARGS (MC_BATCH * 16 / sizeof(u64)) struct mc_buffer { struct multicall_entry entries[MC_BATCH]; +#if MC_DEBUG + struct multicall_entry debug[MC_BATCH]; +#endif u64 args[MC_ARGS]; struct callback { void (*fn)(void *); @@ -56,11 +61,31 @@ void xen_mc_flush(void) local_irq_save(flags); if (b->mcidx) { +#if MC_DEBUG + memcpy(b->debug, b->entries, + b->mcidx * sizeof(struct multicall_entry)); +#endif + if (HYPERVISOR_multicall(b->entries, b->mcidx) != 0) BUG(); for (i = 0; i < b->mcidx; i++) if (b->entries[i].result < 0) ret++; + +#if MC_DEBUG + if (ret) { + printk(KERN_ERR "%d multicall(s) failed: cpu %d\n", + ret, smp_processor_id()); + for(i = 0; i < b->mcidx; i++) { + printk(" call %2d/%d: op=%lu arg=[%lx] result=%ld\n", + i+1, b->mcidx, + b->debug[i].op, + b->debug[i].args[0], + b->entries[i].result); + } + } +#endif + b->mcidx = 0; b->argidx = 0; } else From ace2e92e193126711cb3a83a3752b2c5b8396950 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 16 Oct 2007 11:51:31 -0700 Subject: [PATCH 12/12] xfs: eagerly remove vmap mappings to avoid upsetting Xen XFS leaves stray mappings around when it vmaps memory to make it virtually contigious. This upsets Xen if one of those pages is being recycled into a pagetable, since it finds an extra writable mapping of the page. This patch solves the problem in a brute force way, by making XFS always eagerly unmap its mappings. David Chinner says this shouldn't have any performance impact on filesystems with default block sizes; it will only affect filesystems with large block sizes. Signed-off-by: Jeremy Fitzhardinge Acked-by: David Chinner Cc: Nick Piggin Cc: XFS masters Cc: Stable kernel Cc: Morten =?utf-8?q?B=C3=B8geskov?= Cc: Mark Williamson --- fs/xfs/linux-2.6/xfs_buf.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 39f44ee572e..455e042e0c1 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -187,6 +187,19 @@ free_address( { a_list_t *aentry; +#ifdef CONFIG_XEN + /* + * Xen needs to be able to make sure it can get an exclusive + * RO mapping of pages it wants to turn into a pagetable. If + * a newly allocated page is also still being vmap()ed by xfs, + * it will cause pagetable construction to fail. This is a + * quick workaround to always eagerly unmap pages so that Xen + * is happy. + */ + vunmap(addr); + return; +#endif + aentry = kmalloc(sizeof(a_list_t), GFP_NOWAIT); if (likely(aentry)) { spin_lock(&as_lock);