Jeremy Fitzhardinge
2009-May-07 20:32 UTC
[Xen-devel] [GIT PULL] x86: paravirt_ops updates
Hi Ingo,
This series has a some changes to x86 paravirt_ops:
x86/paravirt: split paravirt definitions into paravirt_types.h
Move all the type-like definitions into asm/paravirt_types.h
x86: split out core __math_state_restore
Split out the core of math_state_restore() into __math_state_restore(),
which just does what''s needed for a context switch (ie, assumes
TS is already clear, and that the fpu context exists).
x86-32: make sure clts is batched during context switch
x86-64: move unlazy_fpu() into lazy cpu state part of context switch
x86-64: move clts into batch cpu state updates when preloading fpu
Make sure the TS state updates during context switch are performed in
the lazy CPU update region so that they can be batched when running
paravirtualized.
The following changes since commit 9ce5424d75e56891905b77d1589924765e62059a:
Ingo Molnar (1):
Merge branch ''x86/urgent''
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen.git x86/paravirt
Jeremy Fitzhardinge (5):
x86/paravirt: split paravirt definitions into paravirt_types.h
x86: split out core __math_state_restore
x86-32: make sure clts is batched during context switch
x86-64: move unlazy_fpu() into lazy cpu state part of context switch
x86-64: move clts into batch cpu state updates when preloading fpu
arch/x86/include/asm/i387.h | 1 +
arch/x86/include/asm/paravirt.h | 711 +--------------------------------
arch/x86/include/asm/paravirt_types.h | 720 +++++++++++++++++++++++++++++++++
arch/x86/kernel/process_32.c | 27 +-
arch/x86/kernel/process_64.c | 34 +-
arch/x86/kernel/traps.c | 33 +-
6 files changed, 783 insertions(+), 743 deletions(-)
create mode 100644 arch/x86/include/asm/paravirt_types.h
Thanks,
J
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2009-May-07 20:32 UTC
[Xen-devel] [PATCH 1/5] x86/paravirt: split paravirt definitions into paravirt_types.h
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Create a definitions-only asm/paravirt_types.h header.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/include/asm/paravirt.h | 711 +--------------------------------
arch/x86/include/asm/paravirt_types.h | 720 +++++++++++++++++++++++++++++++++
2 files changed, 721 insertions(+), 710 deletions(-)
create mode 100644 arch/x86/include/asm/paravirt_types.h
diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index 1fe5837..dad43ff 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -7,689 +7,11 @@
#include <asm/pgtable_types.h>
#include <asm/asm.h>
-/* Bitmask of what can be clobbered: usually at least eax. */
-#define CLBR_NONE 0
-#define CLBR_EAX (1 << 0)
-#define CLBR_ECX (1 << 1)
-#define CLBR_EDX (1 << 2)
-#define CLBR_EDI (1 << 3)
-
-#ifdef CONFIG_X86_32
-/* CLBR_ANY should match all regs platform has. For i386, that''s just
it */
-#define CLBR_ANY ((1 << 4) - 1)
-
-#define CLBR_ARG_REGS (CLBR_EAX | CLBR_EDX | CLBR_ECX)
-#define CLBR_RET_REG (CLBR_EAX | CLBR_EDX)
-#define CLBR_SCRATCH (0)
-#else
-#define CLBR_RAX CLBR_EAX
-#define CLBR_RCX CLBR_ECX
-#define CLBR_RDX CLBR_EDX
-#define CLBR_RDI CLBR_EDI
-#define CLBR_RSI (1 << 4)
-#define CLBR_R8 (1 << 5)
-#define CLBR_R9 (1 << 6)
-#define CLBR_R10 (1 << 7)
-#define CLBR_R11 (1 << 8)
-
-#define CLBR_ANY ((1 << 9) - 1)
-
-#define CLBR_ARG_REGS (CLBR_RDI | CLBR_RSI | CLBR_RDX | \
- CLBR_RCX | CLBR_R8 | CLBR_R9)
-#define CLBR_RET_REG (CLBR_RAX)
-#define CLBR_SCRATCH (CLBR_R10 | CLBR_R11)
-
-#include <asm/desc_defs.h>
-#endif /* X86_64 */
-
-#define CLBR_CALLEE_SAVE ((CLBR_ARG_REGS | CLBR_SCRATCH) & ~CLBR_RET_REG)
+#include <asm/paravirt_types.h>
#ifndef __ASSEMBLY__
#include <linux/types.h>
#include <linux/cpumask.h>
-#include <asm/kmap_types.h>
-#include <asm/desc_defs.h>
-
-struct page;
-struct thread_struct;
-struct desc_ptr;
-struct tss_struct;
-struct mm_struct;
-struct desc_struct;
-struct task_struct;
-
-/*
- * Wrapper type for pointers to code which uses the non-standard
- * calling convention. See PV_CALL_SAVE_REGS_THUNK below.
- */
-struct paravirt_callee_save {
- void *func;
-};
-
-/* general info */
-struct pv_info {
- unsigned int kernel_rpl;
- int shared_kernel_pmd;
- 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.
- */
- unsigned (*patch)(u8 type, u16 clobber, void *insnbuf,
- unsigned long addr, unsigned len);
-
- /* Basic arch-specific setup */
- void (*arch_setup)(void);
- char *(*memory_setup)(void);
- void (*post_allocator_init)(void);
-
- /* Print a banner to identify the environment */
- void (*banner)(void);
-};
-
-
-struct pv_lazy_ops {
- /* Set deferred update mode, used for batching operations. */
- void (*enter)(void);
- void (*leave)(void);
-};
-
-struct pv_time_ops {
- void (*time_init)(void);
-
- /* Set and set time of day */
- unsigned long (*get_wallclock)(void);
- int (*set_wallclock)(unsigned long);
-
- unsigned long long (*sched_clock)(void);
- unsigned long (*get_tsc_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);
-
- void (*clts)(void);
-
- unsigned long (*read_cr0)(void);
- void (*write_cr0)(unsigned long);
-
- unsigned long (*read_cr4_safe)(void);
- unsigned long (*read_cr4)(void);
- void (*write_cr4)(unsigned long);
-
-#ifdef CONFIG_X86_64
- unsigned long (*read_cr8)(void);
- void (*write_cr8)(unsigned long);
-#endif
-
- /* Segment descriptor handling */
- void (*load_tr_desc)(void);
- void (*load_gdt)(const struct desc_ptr *);
- void (*load_idt)(const struct desc_ptr *);
- void (*store_gdt)(struct desc_ptr *);
- void (*store_idt)(struct desc_ptr *);
- void (*set_ldt)(const void *desc, unsigned entries);
- unsigned long (*store_tr)(void);
- void (*load_tls)(struct thread_struct *t, unsigned int cpu);
-#ifdef CONFIG_X86_64
- void (*load_gs_index)(unsigned int idx);
-#endif
- void (*write_ldt_entry)(struct desc_struct *ldt, int entrynum,
- const void *desc);
- void (*write_gdt_entry)(struct desc_struct *,
- int entrynum, const void *desc, int size);
- void (*write_idt_entry)(gate_desc *,
- int entrynum, const gate_desc *gate);
- void (*alloc_ldt)(struct desc_struct *ldt, unsigned entries);
- void (*free_ldt)(struct desc_struct *ldt, unsigned entries);
-
- void (*load_sp0)(struct tss_struct *tss, struct thread_struct *t);
-
- void (*set_iopl_mask)(unsigned mask);
-
- void (*wbinvd)(void);
- void (*io_delay)(void);
-
- /* 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_amd)(unsigned int msr, int *err);
- u64 (*read_msr)(unsigned int msr, int *err);
- int (*write_msr)(unsigned int msr, unsigned low, unsigned high);
-
- u64 (*read_tsc)(void);
- u64 (*read_pmc)(int counter);
- unsigned long long (*read_tscp)(unsigned int *aux);
-
- /*
- * Atomically enable interrupts and return to userspace. This
- * is only ever used to return to 32-bit processes; in a
- * 64-bit kernel, it''s used for 32-on-64 compat processes, but
- * never native 64-bit processes. (Jump, not call.)
- */
- void (*irq_enable_sysexit)(void);
-
- /*
- * Switch to usermode gs and return to 64-bit usermode using
- * sysret. Only used in 64-bit kernels to return to 64-bit
- * processes. Usermode register state, including %rsp, must
- * already be restored.
- */
- void (*usergs_sysret64)(void);
-
- /*
- * Switch to usermode gs and return to 32-bit usermode using
- * sysret. Used to return to 32-on-64 compat processes.
- * Other usermode register state, including %esp, must already
- * be restored.
- */
- void (*usergs_sysret32)(void);
-
- /* Normal iret. Jump to this with the standard iret stack
- frame set up. */
- void (*iret)(void);
-
- void (*swapgs)(void);
-
- void (*start_context_switch)(struct task_struct *prev);
- void (*end_context_switch)(struct task_struct *next);
-};
-
-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.
- *
- * NOTE: These functions callers expect the callee to preserve
- * more registers than the standard C calling convention.
- */
- struct paravirt_callee_save save_fl;
- struct paravirt_callee_save restore_fl;
- struct paravirt_callee_save irq_disable;
- struct paravirt_callee_save irq_enable;
-
- void (*safe_halt)(void);
- void (*halt)(void);
-
-#ifdef CONFIG_X86_64
- void (*adjust_exception_frame)(void);
-#endif
-};
-
-struct pv_apic_ops {
-#ifdef CONFIG_X86_LOCAL_APIC
- void (*setup_boot_clock)(void);
- void (*setup_secondary_clock)(void);
-
- void (*startup_ipi_hook)(int phys_apicid,
- 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);
- void (*flush_tlb_kernel)(void);
- void (*flush_tlb_single)(unsigned long addr);
- void (*flush_tlb_others)(const struct cpumask *cpus,
- struct mm_struct *mm,
- unsigned long va);
-
- /* Hooks for allocating and freeing a pagetable top-level */
- int (*pgd_alloc)(struct mm_struct *mm);
- void (*pgd_free)(struct mm_struct *mm, pgd_t *pgd);
-
- /*
- * Hooks for allocating/releasing pagetable pages when they''re
- * attached to a pagetable
- */
- void (*alloc_pte)(struct mm_struct *mm, unsigned long pfn);
- void (*alloc_pmd)(struct mm_struct *mm, unsigned long pfn);
- void (*alloc_pmd_clone)(unsigned long pfn, unsigned long clonepfn, unsigned
long start, unsigned long count);
- void (*alloc_pud)(struct mm_struct *mm, unsigned long pfn);
- void (*release_pte)(unsigned long pfn);
- void (*release_pmd)(unsigned long pfn);
- void (*release_pud)(unsigned long pfn);
-
- /* Pagetable manipulation functions */
- void (*set_pte)(pte_t *ptep, pte_t pteval);
- void (*set_pte_at)(struct mm_struct *mm, unsigned long addr,
- pte_t *ptep, pte_t pteval);
- void (*set_pmd)(pmd_t *pmdp, pmd_t pmdval);
- void (*pte_update)(struct mm_struct *mm, unsigned long addr,
- pte_t *ptep);
- void (*pte_update_defer)(struct mm_struct *mm,
- unsigned long addr, pte_t *ptep);
-
- pte_t (*ptep_modify_prot_start)(struct mm_struct *mm, unsigned long addr,
- pte_t *ptep);
- void (*ptep_modify_prot_commit)(struct mm_struct *mm, unsigned long addr,
- pte_t *ptep, pte_t pte);
-
- struct paravirt_callee_save pte_val;
- struct paravirt_callee_save make_pte;
-
- struct paravirt_callee_save pgd_val;
- struct paravirt_callee_save make_pgd;
-
-#if PAGETABLE_LEVELS >= 3
-#ifdef CONFIG_X86_PAE
- void (*set_pte_atomic)(pte_t *ptep, pte_t pteval);
- void (*pte_clear)(struct mm_struct *mm, unsigned long addr,
- pte_t *ptep);
- void (*pmd_clear)(pmd_t *pmdp);
-
-#endif /* CONFIG_X86_PAE */
-
- void (*set_pud)(pud_t *pudp, pud_t pudval);
-
- struct paravirt_callee_save pmd_val;
- struct paravirt_callee_save make_pmd;
-
-#if PAGETABLE_LEVELS == 4
- struct paravirt_callee_save pud_val;
- struct paravirt_callee_save make_pud;
-
- void (*set_pgd)(pgd_t *pudp, pgd_t pgdval);
-#endif /* PAGETABLE_LEVELS == 4 */
-#endif /* PAGETABLE_LEVELS >= 3 */
-
-#ifdef CONFIG_HIGHPTE
- void *(*kmap_atomic_pte)(struct page *page, enum km_type type);
-#endif
-
- struct pv_lazy_ops lazy_mode;
-
- /* dom0 ops */
-
- /* Sometimes the physical address is a pfn, and sometimes its
- an mfn. We can tell which is which from the index. */
- void (*set_fixmap)(unsigned /* enum fixed_addresses */ idx,
- phys_addr_t phys, pgprot_t flags);
-};
-
-struct raw_spinlock;
-struct pv_lock_ops {
- int (*spin_is_locked)(struct raw_spinlock *lock);
- int (*spin_is_contended)(struct raw_spinlock *lock);
- void (*spin_lock)(struct raw_spinlock *lock);
- void (*spin_lock_flags)(struct raw_spinlock *lock, unsigned long flags);
- int (*spin_trylock)(struct raw_spinlock *lock);
- void (*spin_unlock)(struct raw_spinlock *lock);
-};
-
-/* 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_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;
- struct pv_lock_ops pv_lock_ops;
-};
-
-extern struct pv_info pv_info;
-extern struct pv_init_ops pv_init_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;
-extern struct pv_lock_ops pv_lock_ops;
-
-#define PARAVIRT_PATCH(x) \
- (offsetof(struct paravirt_patch_template, x) / sizeof(void *))
-
-#define paravirt_type(op) \
- [paravirt_typenum] "i" (PARAVIRT_PATCH(op)), \
- [paravirt_opptr] "i" (&(op))
-#define paravirt_clobber(clobber) \
- [paravirt_clobber] "i" (clobber)
-
-/*
- * Generate some code, and mark it as patchable by the
- * apply_paravirt() alternate instruction patcher.
- */
-#define _paravirt_alt(insn_string, type, clobber) \
- "771:\n\t" insn_string "\n" "772:\n" \
- ".pushsection .parainstructions,\"a\"\n" \
- _ASM_ALIGN "\n" \
- _ASM_PTR " 771b\n" \
- " .byte " type "\n" \
- " .byte 772b-771b\n" \
- " .short " clobber "\n" \
- ".popsection\n"
-
-/* Generate patchable code, with the default asm parameters. */
-#define paravirt_alt(insn_string) \
- _paravirt_alt(insn_string, "%c[paravirt_typenum]",
"%c[paravirt_clobber]")
-
-/* Simple instruction patching code. */
-#define DEF_NATIVE(ops, name, code) \
- extern const char start_##ops##_##name[], end_##ops##_##name[]; \
- asm("start_" #ops "_" #name ": " code ";
end_" #ops "_" #name ":")
-
-unsigned paravirt_patch_nop(void);
-unsigned paravirt_patch_ident_32(void *insnbuf, unsigned len);
-unsigned paravirt_patch_ident_64(void *insnbuf, unsigned len);
-unsigned paravirt_patch_ignore(unsigned len);
-unsigned paravirt_patch_call(void *insnbuf,
- const void *target, u16 tgt_clobbers,
- unsigned long addr, u16 site_clobbers,
- unsigned len);
-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);
-
-unsigned paravirt_patch_insns(void *insnbuf, unsigned len,
- const char *start, const char *end);
-
-unsigned native_patch(u8 type, u16 clobbers, void *ibuf,
- unsigned long addr, unsigned len);
-
-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_patch_template structure, and can therefore be
- * freely converted back into a structure offset.
- */
-#define PARAVIRT_CALL "call *%c[paravirt_opptr];"
-
-/*
- * 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:
- * (pv_op_struct.operations)(args...).
- *
- * Unfortunately, this is a relatively slow operation for modern CPUs,
- * because it cannot necessarily determine what the destination
- * address is. In this case, the address is a runtime constant, so at
- * the very least we can patch the call to e a simple direct call, or
- * ideally, patch an inline implementation into the callsite. (Direct
- * calls are essentially free, because the call and return addresses
- * are completely predictable.)
- *
- * For i386, these macros rely on the standard gcc "regparm(3)"
calling
- * convention, in which the first three arguments are placed in %eax,
- * %edx, %ecx (in that order), and the remaining arguments are placed
- * on the stack. All caller-save registers (eax,edx,ecx) are expected
- * to be modified (either clobbered or used for return values).
- * X86_64, on the other hand, already specifies a register-based calling
- * conventions, returning at %rax, with parameteres going on %rdi, %rsi,
- * %rdx, and %rcx. Note that for this reason, x86_64 does not need any
- * special handling for dealing with 4 arguments, unlike i386.
- * However, x86_64 also have to clobber all caller saved registers, which
- * unfortunately, are quite a bit (r8 - r11)
- *
- * 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 pv_init_ops
- * implementation.
- *
- * Unfortunately there''s no way to get gcc to generate the args setup
- * for the call, and then allow the call itself to be generated by an
- * inline asm. Because of this, we must do the complete arg setup and
- * return value handling from within these macros. This is fairly
- * cumbersome.
- *
- * There are 5 sets of PVOP_* macros for dealing with 0-4 arguments.
- * It could be extended to more arguments, but there would be little
- * to be gained from that. For each number of arguments, there are
- * the two VCALL and CALL variants for void and non-void functions.
- *
- * When there is a return value, the invoker of the macro must specify
- * the return type. The macro then uses sizeof() on that type to
- * determine whether its a 32 or 64 bit value, and places the return
- * in the right register(s) (just %eax for 32-bit, and %edx:%eax for
- * 64-bit). For x86_64 machines, it just returns at %rax regardless of
- * the return value size.
- *
- * 64-bit arguments are passed as a pair of adjacent 32-bit arguments
- * i386 also passes 64-bit arguments as a pair of adjacent 32-bit arguments
- * in low,high order
- *
- * Small structures are passed and returned in registers. The macro
- * calling convention can''t directly deal with this, so the wrapper
- * functions must do this.
- *
- * These PVOP_* macros are only defined within this header. This
- * means that all uses must be wrapped in inline functions. This also
- * makes sure the incoming and outgoing types are always correct.
- */
-#ifdef CONFIG_X86_32
-#define PVOP_VCALL_ARGS \
- unsigned long __eax = __eax, __edx = __edx, __ecx = __ecx
-#define PVOP_CALL_ARGS PVOP_VCALL_ARGS
-
-#define PVOP_CALL_ARG1(x) "a" ((unsigned long)(x))
-#define PVOP_CALL_ARG2(x) "d" ((unsigned long)(x))
-#define PVOP_CALL_ARG3(x) "c" ((unsigned long)(x))
-
-#define PVOP_VCALL_CLOBBERS "=a" (__eax), "=d" (__edx), \
- "=c" (__ecx)
-#define PVOP_CALL_CLOBBERS PVOP_VCALL_CLOBBERS
-
-#define PVOP_VCALLEE_CLOBBERS "=a" (__eax), "=d" (__edx)
-#define PVOP_CALLEE_CLOBBERS PVOP_VCALLEE_CLOBBERS
-
-#define EXTRA_CLOBBERS
-#define VEXTRA_CLOBBERS
-#else /* CONFIG_X86_64 */
-#define PVOP_VCALL_ARGS \
- unsigned long __edi = __edi, __esi = __esi, \
- __edx = __edx, __ecx = __ecx
-#define PVOP_CALL_ARGS PVOP_VCALL_ARGS, __eax
-
-#define PVOP_CALL_ARG1(x) "D" ((unsigned long)(x))
-#define PVOP_CALL_ARG2(x) "S" ((unsigned long)(x))
-#define PVOP_CALL_ARG3(x) "d" ((unsigned long)(x))
-#define PVOP_CALL_ARG4(x) "c" ((unsigned long)(x))
-
-#define PVOP_VCALL_CLOBBERS "=D" (__edi), \
- "=S" (__esi), "=d" (__edx), \
- "=c" (__ecx)
-#define PVOP_CALL_CLOBBERS PVOP_VCALL_CLOBBERS, "=a" (__eax)
-
-#define PVOP_VCALLEE_CLOBBERS "=a" (__eax)
-#define PVOP_CALLEE_CLOBBERS PVOP_VCALLEE_CLOBBERS
-
-#define EXTRA_CLOBBERS , "r8", "r9", "r10",
"r11"
-#define VEXTRA_CLOBBERS , "rax", "r8", "r9",
"r10", "r11"
-#endif /* CONFIG_X86_32 */
-
-#ifdef CONFIG_PARAVIRT_DEBUG
-#define PVOP_TEST_NULL(op) BUG_ON(op == NULL)
-#else
-#define PVOP_TEST_NULL(op) ((void)op)
-#endif
-
-#define ____PVOP_CALL(rettype, op, clbr, call_clbr, extra_clbr, \
- pre, post, ...) \
- ({ \
- rettype __ret; \
- PVOP_CALL_ARGS; \
- PVOP_TEST_NULL(op); \
- /* This is 32-bit specific, but is okay in 64-bit */ \
- /* since this condition will never hold */ \
- if (sizeof(rettype) > sizeof(unsigned long)) { \
- asm volatile(pre \
- paravirt_alt(PARAVIRT_CALL) \
- post \
- : call_clbr \
- : paravirt_type(op), \
- paravirt_clobber(clbr), \
- ##__VA_ARGS__ \
- : "memory", "cc" extra_clbr); \
- __ret = (rettype)((((u64)__edx) << 32) | __eax); \
- } else { \
- asm volatile(pre \
- paravirt_alt(PARAVIRT_CALL) \
- post \
- : call_clbr \
- : paravirt_type(op), \
- paravirt_clobber(clbr), \
- ##__VA_ARGS__ \
- : "memory", "cc" extra_clbr); \
- __ret = (rettype)__eax; \
- } \
- __ret; \
- })
-
-#define __PVOP_CALL(rettype, op, pre, post, ...) \
- ____PVOP_CALL(rettype, op, CLBR_ANY, PVOP_CALL_CLOBBERS, \
- EXTRA_CLOBBERS, pre, post, ##__VA_ARGS__)
-
-#define __PVOP_CALLEESAVE(rettype, op, pre, post, ...) \
- ____PVOP_CALL(rettype, op.func, CLBR_RET_REG, \
- PVOP_CALLEE_CLOBBERS, , \
- pre, post, ##__VA_ARGS__)
-
-
-#define ____PVOP_VCALL(op, clbr, call_clbr, extra_clbr, pre, post, ...) \
- ({ \
- PVOP_VCALL_ARGS; \
- PVOP_TEST_NULL(op); \
- asm volatile(pre \
- paravirt_alt(PARAVIRT_CALL) \
- post \
- : call_clbr \
- : paravirt_type(op), \
- paravirt_clobber(clbr), \
- ##__VA_ARGS__ \
- : "memory", "cc" extra_clbr); \
- })
-
-#define __PVOP_VCALL(op, pre, post, ...) \
- ____PVOP_VCALL(op, CLBR_ANY, PVOP_VCALL_CLOBBERS, \
- VEXTRA_CLOBBERS, \
- pre, post, ##__VA_ARGS__)
-
-#define __PVOP_VCALLEESAVE(rettype, op, pre, post, ...) \
- ____PVOP_CALL(rettype, op.func, CLBR_RET_REG, \
- PVOP_VCALLEE_CLOBBERS, , \
- pre, post, ##__VA_ARGS__)
-
-
-
-#define PVOP_CALL0(rettype, op) \
- __PVOP_CALL(rettype, op, "", "")
-#define PVOP_VCALL0(op) \
- __PVOP_VCALL(op, "", "")
-
-#define PVOP_CALLEE0(rettype, op) \
- __PVOP_CALLEESAVE(rettype, op, "", "")
-#define PVOP_VCALLEE0(op) \
- __PVOP_VCALLEESAVE(op, "", "")
-
-
-#define PVOP_CALL1(rettype, op, arg1) \
- __PVOP_CALL(rettype, op, "", "", PVOP_CALL_ARG1(arg1))
-#define PVOP_VCALL1(op, arg1) \
- __PVOP_VCALL(op, "", "", PVOP_CALL_ARG1(arg1))
-
-#define PVOP_CALLEE1(rettype, op, arg1) \
- __PVOP_CALLEESAVE(rettype, op, "", "",
PVOP_CALL_ARG1(arg1))
-#define PVOP_VCALLEE1(op, arg1) \
- __PVOP_VCALLEESAVE(op, "", "", PVOP_CALL_ARG1(arg1))
-
-
-#define PVOP_CALL2(rettype, op, arg1, arg2) \
- __PVOP_CALL(rettype, op, "", "", PVOP_CALL_ARG1(arg1), \
- PVOP_CALL_ARG2(arg2))
-#define PVOP_VCALL2(op, arg1, arg2) \
- __PVOP_VCALL(op, "", "", PVOP_CALL_ARG1(arg1), \
- PVOP_CALL_ARG2(arg2))
-
-#define PVOP_CALLEE2(rettype, op, arg1, arg2) \
- __PVOP_CALLEESAVE(rettype, op, "", "",
PVOP_CALL_ARG1(arg1), \
- PVOP_CALL_ARG2(arg2))
-#define PVOP_VCALLEE2(op, arg1, arg2) \
- __PVOP_VCALLEESAVE(op, "", "", PVOP_CALL_ARG1(arg1), \
- PVOP_CALL_ARG2(arg2))
-
-
-#define PVOP_CALL3(rettype, op, arg1, arg2, arg3) \
- __PVOP_CALL(rettype, op, "", "", PVOP_CALL_ARG1(arg1), \
- PVOP_CALL_ARG2(arg2), PVOP_CALL_ARG3(arg3))
-#define PVOP_VCALL3(op, arg1, arg2, arg3) \
- __PVOP_VCALL(op, "", "", PVOP_CALL_ARG1(arg1), \
- PVOP_CALL_ARG2(arg2), PVOP_CALL_ARG3(arg3))
-
-/* This is the only difference in x86_64. We can make it much simpler */
-#ifdef CONFIG_X86_32
-#define PVOP_CALL4(rettype, op, arg1, arg2, arg3, arg4) \
- __PVOP_CALL(rettype, op, \
- "push %[_arg4];", "lea 4(%%esp),%%esp;", \
- PVOP_CALL_ARG1(arg1), PVOP_CALL_ARG2(arg2), \
- PVOP_CALL_ARG3(arg3), [_arg4] "mr" ((u32)(arg4)))
-#define PVOP_VCALL4(op, arg1, arg2, arg3, arg4) \
- __PVOP_VCALL(op, \
- "push %[_arg4];", "lea 4(%%esp),%%esp;", \
- "0" ((u32)(arg1)), "1" ((u32)(arg2)), \
- "2" ((u32)(arg3)), [_arg4] "mr" ((u32)(arg4)))
-#else
-#define PVOP_CALL4(rettype, op, arg1, arg2, arg3, arg4) \
- __PVOP_CALL(rettype, op, "", "", \
- PVOP_CALL_ARG1(arg1), PVOP_CALL_ARG2(arg2), \
- PVOP_CALL_ARG3(arg3), PVOP_CALL_ARG4(arg4))
-#define PVOP_VCALL4(op, arg1, arg2, arg3, arg4) \
- __PVOP_VCALL(op, "", "", \
- PVOP_CALL_ARG1(arg1), PVOP_CALL_ARG2(arg2), \
- PVOP_CALL_ARG3(arg3), PVOP_CALL_ARG4(arg4))
-#endif
static inline int paravirt_enabled(void)
{
@@ -1393,20 +715,6 @@ static inline void pmd_clear(pmd_t *pmdp)
}
#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_start_context_switch(struct task_struct *prev);
-void paravirt_end_context_switch(struct task_struct *next);
-
-void paravirt_enter_lazy_mmu(void);
-void paravirt_leave_lazy_mmu(void);
-
#define __HAVE_ARCH_START_CONTEXT_SWITCH
static inline void arch_start_context_switch(struct task_struct *prev)
{
@@ -1437,12 +745,6 @@ static inline void __set_fixmap(unsigned /* enum
fixed_addresses */ idx,
pv_mmu_ops.set_fixmap(idx, phys, flags);
}
-void _paravirt_nop(void);
-u32 _paravirt_ident_32(u32);
-u64 _paravirt_ident_64(u64);
-
-#define paravirt_nop ((void *)_paravirt_nop)
-
#ifdef CONFIG_SMP
static inline int __raw_spin_is_locked(struct raw_spinlock *lock)
@@ -1479,17 +781,6 @@ static __always_inline void __raw_spin_unlock(struct
raw_spinlock *lock)
#endif
-/* These all sit in the .parainstructions section to tell us what to patch. */
-struct paravirt_patch_site {
- u8 *instr; /* original instructions */
- u8 instrtype; /* type of this instruction */
- u8 len; /* length of original instruction */
- u16 clobbers; /* what registers you may clobber */
-};
-
-extern struct paravirt_patch_site __parainstructions[],
- __parainstructions_end[];
-
#ifdef CONFIG_X86_32
#define PV_SAVE_REGS "pushl %ecx; pushl %edx;"
#define PV_RESTORE_REGS "popl %edx; popl %ecx;"
diff --git a/arch/x86/include/asm/paravirt_types.h
b/arch/x86/include/asm/paravirt_types.h
new file mode 100644
index 0000000..2b3371b
--- /dev/null
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -0,0 +1,720 @@
+#ifndef _ASM_X86_PARAVIRT_TYPES_H
+#define _ASM_X86_PARAVIRT_TYPES_H
+
+/* Bitmask of what can be clobbered: usually at least eax. */
+#define CLBR_NONE 0
+#define CLBR_EAX (1 << 0)
+#define CLBR_ECX (1 << 1)
+#define CLBR_EDX (1 << 2)
+#define CLBR_EDI (1 << 3)
+
+#ifdef CONFIG_X86_32
+/* CLBR_ANY should match all regs platform has. For i386, that''s just
it */
+#define CLBR_ANY ((1 << 4) - 1)
+
+#define CLBR_ARG_REGS (CLBR_EAX | CLBR_EDX | CLBR_ECX)
+#define CLBR_RET_REG (CLBR_EAX | CLBR_EDX)
+#define CLBR_SCRATCH (0)
+#else
+#define CLBR_RAX CLBR_EAX
+#define CLBR_RCX CLBR_ECX
+#define CLBR_RDX CLBR_EDX
+#define CLBR_RDI CLBR_EDI
+#define CLBR_RSI (1 << 4)
+#define CLBR_R8 (1 << 5)
+#define CLBR_R9 (1 << 6)
+#define CLBR_R10 (1 << 7)
+#define CLBR_R11 (1 << 8)
+
+#define CLBR_ANY ((1 << 9) - 1)
+
+#define CLBR_ARG_REGS (CLBR_RDI | CLBR_RSI | CLBR_RDX | \
+ CLBR_RCX | CLBR_R8 | CLBR_R9)
+#define CLBR_RET_REG (CLBR_RAX)
+#define CLBR_SCRATCH (CLBR_R10 | CLBR_R11)
+
+#endif /* X86_64 */
+
+#define CLBR_CALLEE_SAVE ((CLBR_ARG_REGS | CLBR_SCRATCH) & ~CLBR_RET_REG)
+
+#ifndef __ASSEMBLY__
+
+#include <asm/desc_defs.h>
+#include <asm/kmap_types.h>
+
+struct page;
+struct thread_struct;
+struct desc_ptr;
+struct tss_struct;
+struct mm_struct;
+struct desc_struct;
+struct task_struct;
+struct cpumask;
+
+/*
+ * Wrapper type for pointers to code which uses the non-standard
+ * calling convention. See PV_CALL_SAVE_REGS_THUNK below.
+ */
+struct paravirt_callee_save {
+ void *func;
+};
+
+/* general info */
+struct pv_info {
+ unsigned int kernel_rpl;
+ int shared_kernel_pmd;
+ 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.
+ */
+ unsigned (*patch)(u8 type, u16 clobber, void *insnbuf,
+ unsigned long addr, unsigned len);
+
+ /* Basic arch-specific setup */
+ void (*arch_setup)(void);
+ char *(*memory_setup)(void);
+ void (*post_allocator_init)(void);
+
+ /* Print a banner to identify the environment */
+ void (*banner)(void);
+};
+
+
+struct pv_lazy_ops {
+ /* Set deferred update mode, used for batching operations. */
+ void (*enter)(void);
+ void (*leave)(void);
+};
+
+struct pv_time_ops {
+ void (*time_init)(void);
+
+ /* Set and set time of day */
+ unsigned long (*get_wallclock)(void);
+ int (*set_wallclock)(unsigned long);
+
+ unsigned long long (*sched_clock)(void);
+ unsigned long (*get_tsc_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);
+
+ void (*clts)(void);
+
+ unsigned long (*read_cr0)(void);
+ void (*write_cr0)(unsigned long);
+
+ unsigned long (*read_cr4_safe)(void);
+ unsigned long (*read_cr4)(void);
+ void (*write_cr4)(unsigned long);
+
+#ifdef CONFIG_X86_64
+ unsigned long (*read_cr8)(void);
+ void (*write_cr8)(unsigned long);
+#endif
+
+ /* Segment descriptor handling */
+ void (*load_tr_desc)(void);
+ void (*load_gdt)(const struct desc_ptr *);
+ void (*load_idt)(const struct desc_ptr *);
+ void (*store_gdt)(struct desc_ptr *);
+ void (*store_idt)(struct desc_ptr *);
+ void (*set_ldt)(const void *desc, unsigned entries);
+ unsigned long (*store_tr)(void);
+ void (*load_tls)(struct thread_struct *t, unsigned int cpu);
+#ifdef CONFIG_X86_64
+ void (*load_gs_index)(unsigned int idx);
+#endif
+ void (*write_ldt_entry)(struct desc_struct *ldt, int entrynum,
+ const void *desc);
+ void (*write_gdt_entry)(struct desc_struct *,
+ int entrynum, const void *desc, int size);
+ void (*write_idt_entry)(gate_desc *,
+ int entrynum, const gate_desc *gate);
+ void (*alloc_ldt)(struct desc_struct *ldt, unsigned entries);
+ void (*free_ldt)(struct desc_struct *ldt, unsigned entries);
+
+ void (*load_sp0)(struct tss_struct *tss, struct thread_struct *t);
+
+ void (*set_iopl_mask)(unsigned mask);
+
+ void (*wbinvd)(void);
+ void (*io_delay)(void);
+
+ /* 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_amd)(unsigned int msr, int *err);
+ u64 (*read_msr)(unsigned int msr, int *err);
+ int (*write_msr)(unsigned int msr, unsigned low, unsigned high);
+
+ u64 (*read_tsc)(void);
+ u64 (*read_pmc)(int counter);
+ unsigned long long (*read_tscp)(unsigned int *aux);
+
+ /*
+ * Atomically enable interrupts and return to userspace. This
+ * is only ever used to return to 32-bit processes; in a
+ * 64-bit kernel, it''s used for 32-on-64 compat processes, but
+ * never native 64-bit processes. (Jump, not call.)
+ */
+ void (*irq_enable_sysexit)(void);
+
+ /*
+ * Switch to usermode gs and return to 64-bit usermode using
+ * sysret. Only used in 64-bit kernels to return to 64-bit
+ * processes. Usermode register state, including %rsp, must
+ * already be restored.
+ */
+ void (*usergs_sysret64)(void);
+
+ /*
+ * Switch to usermode gs and return to 32-bit usermode using
+ * sysret. Used to return to 32-on-64 compat processes.
+ * Other usermode register state, including %esp, must already
+ * be restored.
+ */
+ void (*usergs_sysret32)(void);
+
+ /* Normal iret. Jump to this with the standard iret stack
+ frame set up. */
+ void (*iret)(void);
+
+ void (*swapgs)(void);
+
+ void (*start_context_switch)(struct task_struct *prev);
+ void (*end_context_switch)(struct task_struct *next);
+};
+
+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.
+ *
+ * NOTE: These functions callers expect the callee to preserve
+ * more registers than the standard C calling convention.
+ */
+ struct paravirt_callee_save save_fl;
+ struct paravirt_callee_save restore_fl;
+ struct paravirt_callee_save irq_disable;
+ struct paravirt_callee_save irq_enable;
+
+ void (*safe_halt)(void);
+ void (*halt)(void);
+
+#ifdef CONFIG_X86_64
+ void (*adjust_exception_frame)(void);
+#endif
+};
+
+struct pv_apic_ops {
+#ifdef CONFIG_X86_LOCAL_APIC
+ void (*setup_boot_clock)(void);
+ void (*setup_secondary_clock)(void);
+
+ void (*startup_ipi_hook)(int phys_apicid,
+ 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);
+ void (*flush_tlb_kernel)(void);
+ void (*flush_tlb_single)(unsigned long addr);
+ void (*flush_tlb_others)(const struct cpumask *cpus,
+ struct mm_struct *mm,
+ unsigned long va);
+
+ /* Hooks for allocating and freeing a pagetable top-level */
+ int (*pgd_alloc)(struct mm_struct *mm);
+ void (*pgd_free)(struct mm_struct *mm, pgd_t *pgd);
+
+ /*
+ * Hooks for allocating/releasing pagetable pages when they''re
+ * attached to a pagetable
+ */
+ void (*alloc_pte)(struct mm_struct *mm, unsigned long pfn);
+ void (*alloc_pmd)(struct mm_struct *mm, unsigned long pfn);
+ void (*alloc_pmd_clone)(unsigned long pfn, unsigned long clonepfn, unsigned
long start, unsigned long count);
+ void (*alloc_pud)(struct mm_struct *mm, unsigned long pfn);
+ void (*release_pte)(unsigned long pfn);
+ void (*release_pmd)(unsigned long pfn);
+ void (*release_pud)(unsigned long pfn);
+
+ /* Pagetable manipulation functions */
+ void (*set_pte)(pte_t *ptep, pte_t pteval);
+ void (*set_pte_at)(struct mm_struct *mm, unsigned long addr,
+ pte_t *ptep, pte_t pteval);
+ void (*set_pmd)(pmd_t *pmdp, pmd_t pmdval);
+ void (*pte_update)(struct mm_struct *mm, unsigned long addr,
+ pte_t *ptep);
+ void (*pte_update_defer)(struct mm_struct *mm,
+ unsigned long addr, pte_t *ptep);
+
+ pte_t (*ptep_modify_prot_start)(struct mm_struct *mm, unsigned long addr,
+ pte_t *ptep);
+ void (*ptep_modify_prot_commit)(struct mm_struct *mm, unsigned long addr,
+ pte_t *ptep, pte_t pte);
+
+ struct paravirt_callee_save pte_val;
+ struct paravirt_callee_save make_pte;
+
+ struct paravirt_callee_save pgd_val;
+ struct paravirt_callee_save make_pgd;
+
+#if PAGETABLE_LEVELS >= 3
+#ifdef CONFIG_X86_PAE
+ void (*set_pte_atomic)(pte_t *ptep, pte_t pteval);
+ void (*pte_clear)(struct mm_struct *mm, unsigned long addr,
+ pte_t *ptep);
+ void (*pmd_clear)(pmd_t *pmdp);
+
+#endif /* CONFIG_X86_PAE */
+
+ void (*set_pud)(pud_t *pudp, pud_t pudval);
+
+ struct paravirt_callee_save pmd_val;
+ struct paravirt_callee_save make_pmd;
+
+#if PAGETABLE_LEVELS == 4
+ struct paravirt_callee_save pud_val;
+ struct paravirt_callee_save make_pud;
+
+ void (*set_pgd)(pgd_t *pudp, pgd_t pgdval);
+#endif /* PAGETABLE_LEVELS == 4 */
+#endif /* PAGETABLE_LEVELS >= 3 */
+
+#ifdef CONFIG_HIGHPTE
+ void *(*kmap_atomic_pte)(struct page *page, enum km_type type);
+#endif
+
+ struct pv_lazy_ops lazy_mode;
+
+ /* dom0 ops */
+
+ /* Sometimes the physical address is a pfn, and sometimes its
+ an mfn. We can tell which is which from the index. */
+ void (*set_fixmap)(unsigned /* enum fixed_addresses */ idx,
+ phys_addr_t phys, pgprot_t flags);
+};
+
+struct raw_spinlock;
+struct pv_lock_ops {
+ int (*spin_is_locked)(struct raw_spinlock *lock);
+ int (*spin_is_contended)(struct raw_spinlock *lock);
+ void (*spin_lock)(struct raw_spinlock *lock);
+ void (*spin_lock_flags)(struct raw_spinlock *lock, unsigned long flags);
+ int (*spin_trylock)(struct raw_spinlock *lock);
+ void (*spin_unlock)(struct raw_spinlock *lock);
+};
+
+/* 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_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;
+ struct pv_lock_ops pv_lock_ops;
+};
+
+extern struct pv_info pv_info;
+extern struct pv_init_ops pv_init_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;
+extern struct pv_lock_ops pv_lock_ops;
+
+#define PARAVIRT_PATCH(x) \
+ (offsetof(struct paravirt_patch_template, x) / sizeof(void *))
+
+#define paravirt_type(op) \
+ [paravirt_typenum] "i" (PARAVIRT_PATCH(op)), \
+ [paravirt_opptr] "i" (&(op))
+#define paravirt_clobber(clobber) \
+ [paravirt_clobber] "i" (clobber)
+
+/*
+ * Generate some code, and mark it as patchable by the
+ * apply_paravirt() alternate instruction patcher.
+ */
+#define _paravirt_alt(insn_string, type, clobber) \
+ "771:\n\t" insn_string "\n" "772:\n" \
+ ".pushsection .parainstructions,\"a\"\n" \
+ _ASM_ALIGN "\n" \
+ _ASM_PTR " 771b\n" \
+ " .byte " type "\n" \
+ " .byte 772b-771b\n" \
+ " .short " clobber "\n" \
+ ".popsection\n"
+
+/* Generate patchable code, with the default asm parameters. */
+#define paravirt_alt(insn_string) \
+ _paravirt_alt(insn_string, "%c[paravirt_typenum]",
"%c[paravirt_clobber]")
+
+/* Simple instruction patching code. */
+#define DEF_NATIVE(ops, name, code) \
+ extern const char start_##ops##_##name[], end_##ops##_##name[]; \
+ asm("start_" #ops "_" #name ": " code ";
end_" #ops "_" #name ":")
+
+unsigned paravirt_patch_nop(void);
+unsigned paravirt_patch_ident_32(void *insnbuf, unsigned len);
+unsigned paravirt_patch_ident_64(void *insnbuf, unsigned len);
+unsigned paravirt_patch_ignore(unsigned len);
+unsigned paravirt_patch_call(void *insnbuf,
+ const void *target, u16 tgt_clobbers,
+ unsigned long addr, u16 site_clobbers,
+ unsigned len);
+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);
+
+unsigned paravirt_patch_insns(void *insnbuf, unsigned len,
+ const char *start, const char *end);
+
+unsigned native_patch(u8 type, u16 clobbers, void *ibuf,
+ unsigned long addr, unsigned len);
+
+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_patch_template structure, and can therefore be
+ * freely converted back into a structure offset.
+ */
+#define PARAVIRT_CALL "call *%c[paravirt_opptr];"
+
+/*
+ * 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:
+ * (pv_op_struct.operations)(args...).
+ *
+ * Unfortunately, this is a relatively slow operation for modern CPUs,
+ * because it cannot necessarily determine what the destination
+ * address is. In this case, the address is a runtime constant, so at
+ * the very least we can patch the call to e a simple direct call, or
+ * ideally, patch an inline implementation into the callsite. (Direct
+ * calls are essentially free, because the call and return addresses
+ * are completely predictable.)
+ *
+ * For i386, these macros rely on the standard gcc "regparm(3)"
calling
+ * convention, in which the first three arguments are placed in %eax,
+ * %edx, %ecx (in that order), and the remaining arguments are placed
+ * on the stack. All caller-save registers (eax,edx,ecx) are expected
+ * to be modified (either clobbered or used for return values).
+ * X86_64, on the other hand, already specifies a register-based calling
+ * conventions, returning at %rax, with parameteres going on %rdi, %rsi,
+ * %rdx, and %rcx. Note that for this reason, x86_64 does not need any
+ * special handling for dealing with 4 arguments, unlike i386.
+ * However, x86_64 also have to clobber all caller saved registers, which
+ * unfortunately, are quite a bit (r8 - r11)
+ *
+ * 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 pv_init_ops
+ * implementation.
+ *
+ * Unfortunately there''s no way to get gcc to generate the args setup
+ * for the call, and then allow the call itself to be generated by an
+ * inline asm. Because of this, we must do the complete arg setup and
+ * return value handling from within these macros. This is fairly
+ * cumbersome.
+ *
+ * There are 5 sets of PVOP_* macros for dealing with 0-4 arguments.
+ * It could be extended to more arguments, but there would be little
+ * to be gained from that. For each number of arguments, there are
+ * the two VCALL and CALL variants for void and non-void functions.
+ *
+ * When there is a return value, the invoker of the macro must specify
+ * the return type. The macro then uses sizeof() on that type to
+ * determine whether its a 32 or 64 bit value, and places the return
+ * in the right register(s) (just %eax for 32-bit, and %edx:%eax for
+ * 64-bit). For x86_64 machines, it just returns at %rax regardless of
+ * the return value size.
+ *
+ * 64-bit arguments are passed as a pair of adjacent 32-bit arguments
+ * i386 also passes 64-bit arguments as a pair of adjacent 32-bit arguments
+ * in low,high order
+ *
+ * Small structures are passed and returned in registers. The macro
+ * calling convention can''t directly deal with this, so the wrapper
+ * functions must do this.
+ *
+ * These PVOP_* macros are only defined within this header. This
+ * means that all uses must be wrapped in inline functions. This also
+ * makes sure the incoming and outgoing types are always correct.
+ */
+#ifdef CONFIG_X86_32
+#define PVOP_VCALL_ARGS \
+ unsigned long __eax = __eax, __edx = __edx, __ecx = __ecx
+#define PVOP_CALL_ARGS PVOP_VCALL_ARGS
+
+#define PVOP_CALL_ARG1(x) "a" ((unsigned long)(x))
+#define PVOP_CALL_ARG2(x) "d" ((unsigned long)(x))
+#define PVOP_CALL_ARG3(x) "c" ((unsigned long)(x))
+
+#define PVOP_VCALL_CLOBBERS "=a" (__eax), "=d" (__edx), \
+ "=c" (__ecx)
+#define PVOP_CALL_CLOBBERS PVOP_VCALL_CLOBBERS
+
+#define PVOP_VCALLEE_CLOBBERS "=a" (__eax), "=d" (__edx)
+#define PVOP_CALLEE_CLOBBERS PVOP_VCALLEE_CLOBBERS
+
+#define EXTRA_CLOBBERS
+#define VEXTRA_CLOBBERS
+#else /* CONFIG_X86_64 */
+#define PVOP_VCALL_ARGS \
+ unsigned long __edi = __edi, __esi = __esi, \
+ __edx = __edx, __ecx = __ecx
+#define PVOP_CALL_ARGS PVOP_VCALL_ARGS, __eax
+
+#define PVOP_CALL_ARG1(x) "D" ((unsigned long)(x))
+#define PVOP_CALL_ARG2(x) "S" ((unsigned long)(x))
+#define PVOP_CALL_ARG3(x) "d" ((unsigned long)(x))
+#define PVOP_CALL_ARG4(x) "c" ((unsigned long)(x))
+
+#define PVOP_VCALL_CLOBBERS "=D" (__edi), \
+ "=S" (__esi), "=d" (__edx), \
+ "=c" (__ecx)
+#define PVOP_CALL_CLOBBERS PVOP_VCALL_CLOBBERS, "=a" (__eax)
+
+#define PVOP_VCALLEE_CLOBBERS "=a" (__eax)
+#define PVOP_CALLEE_CLOBBERS PVOP_VCALLEE_CLOBBERS
+
+#define EXTRA_CLOBBERS , "r8", "r9", "r10",
"r11"
+#define VEXTRA_CLOBBERS , "rax", "r8", "r9",
"r10", "r11"
+#endif /* CONFIG_X86_32 */
+
+#ifdef CONFIG_PARAVIRT_DEBUG
+#define PVOP_TEST_NULL(op) BUG_ON(op == NULL)
+#else
+#define PVOP_TEST_NULL(op) ((void)op)
+#endif
+
+#define ____PVOP_CALL(rettype, op, clbr, call_clbr, extra_clbr, \
+ pre, post, ...) \
+ ({ \
+ rettype __ret; \
+ PVOP_CALL_ARGS; \
+ PVOP_TEST_NULL(op); \
+ /* This is 32-bit specific, but is okay in 64-bit */ \
+ /* since this condition will never hold */ \
+ if (sizeof(rettype) > sizeof(unsigned long)) { \
+ asm volatile(pre \
+ paravirt_alt(PARAVIRT_CALL) \
+ post \
+ : call_clbr \
+ : paravirt_type(op), \
+ paravirt_clobber(clbr), \
+ ##__VA_ARGS__ \
+ : "memory", "cc" extra_clbr); \
+ __ret = (rettype)((((u64)__edx) << 32) | __eax); \
+ } else { \
+ asm volatile(pre \
+ paravirt_alt(PARAVIRT_CALL) \
+ post \
+ : call_clbr \
+ : paravirt_type(op), \
+ paravirt_clobber(clbr), \
+ ##__VA_ARGS__ \
+ : "memory", "cc" extra_clbr); \
+ __ret = (rettype)__eax; \
+ } \
+ __ret; \
+ })
+
+#define __PVOP_CALL(rettype, op, pre, post, ...) \
+ ____PVOP_CALL(rettype, op, CLBR_ANY, PVOP_CALL_CLOBBERS, \
+ EXTRA_CLOBBERS, pre, post, ##__VA_ARGS__)
+
+#define __PVOP_CALLEESAVE(rettype, op, pre, post, ...) \
+ ____PVOP_CALL(rettype, op.func, CLBR_RET_REG, \
+ PVOP_CALLEE_CLOBBERS, , \
+ pre, post, ##__VA_ARGS__)
+
+
+#define ____PVOP_VCALL(op, clbr, call_clbr, extra_clbr, pre, post, ...) \
+ ({ \
+ PVOP_VCALL_ARGS; \
+ PVOP_TEST_NULL(op); \
+ asm volatile(pre \
+ paravirt_alt(PARAVIRT_CALL) \
+ post \
+ : call_clbr \
+ : paravirt_type(op), \
+ paravirt_clobber(clbr), \
+ ##__VA_ARGS__ \
+ : "memory", "cc" extra_clbr); \
+ })
+
+#define __PVOP_VCALL(op, pre, post, ...) \
+ ____PVOP_VCALL(op, CLBR_ANY, PVOP_VCALL_CLOBBERS, \
+ VEXTRA_CLOBBERS, \
+ pre, post, ##__VA_ARGS__)
+
+#define __PVOP_VCALLEESAVE(rettype, op, pre, post, ...) \
+ ____PVOP_CALL(rettype, op.func, CLBR_RET_REG, \
+ PVOP_VCALLEE_CLOBBERS, , \
+ pre, post, ##__VA_ARGS__)
+
+
+
+#define PVOP_CALL0(rettype, op) \
+ __PVOP_CALL(rettype, op, "", "")
+#define PVOP_VCALL0(op) \
+ __PVOP_VCALL(op, "", "")
+
+#define PVOP_CALLEE0(rettype, op) \
+ __PVOP_CALLEESAVE(rettype, op, "", "")
+#define PVOP_VCALLEE0(op) \
+ __PVOP_VCALLEESAVE(op, "", "")
+
+
+#define PVOP_CALL1(rettype, op, arg1) \
+ __PVOP_CALL(rettype, op, "", "", PVOP_CALL_ARG1(arg1))
+#define PVOP_VCALL1(op, arg1) \
+ __PVOP_VCALL(op, "", "", PVOP_CALL_ARG1(arg1))
+
+#define PVOP_CALLEE1(rettype, op, arg1) \
+ __PVOP_CALLEESAVE(rettype, op, "", "",
PVOP_CALL_ARG1(arg1))
+#define PVOP_VCALLEE1(op, arg1) \
+ __PVOP_VCALLEESAVE(op, "", "", PVOP_CALL_ARG1(arg1))
+
+
+#define PVOP_CALL2(rettype, op, arg1, arg2) \
+ __PVOP_CALL(rettype, op, "", "", PVOP_CALL_ARG1(arg1), \
+ PVOP_CALL_ARG2(arg2))
+#define PVOP_VCALL2(op, arg1, arg2) \
+ __PVOP_VCALL(op, "", "", PVOP_CALL_ARG1(arg1), \
+ PVOP_CALL_ARG2(arg2))
+
+#define PVOP_CALLEE2(rettype, op, arg1, arg2) \
+ __PVOP_CALLEESAVE(rettype, op, "", "",
PVOP_CALL_ARG1(arg1), \
+ PVOP_CALL_ARG2(arg2))
+#define PVOP_VCALLEE2(op, arg1, arg2) \
+ __PVOP_VCALLEESAVE(op, "", "", PVOP_CALL_ARG1(arg1), \
+ PVOP_CALL_ARG2(arg2))
+
+
+#define PVOP_CALL3(rettype, op, arg1, arg2, arg3) \
+ __PVOP_CALL(rettype, op, "", "", PVOP_CALL_ARG1(arg1), \
+ PVOP_CALL_ARG2(arg2), PVOP_CALL_ARG3(arg3))
+#define PVOP_VCALL3(op, arg1, arg2, arg3) \
+ __PVOP_VCALL(op, "", "", PVOP_CALL_ARG1(arg1), \
+ PVOP_CALL_ARG2(arg2), PVOP_CALL_ARG3(arg3))
+
+/* This is the only difference in x86_64. We can make it much simpler */
+#ifdef CONFIG_X86_32
+#define PVOP_CALL4(rettype, op, arg1, arg2, arg3, arg4) \
+ __PVOP_CALL(rettype, op, \
+ "push %[_arg4];", "lea 4(%%esp),%%esp;", \
+ PVOP_CALL_ARG1(arg1), PVOP_CALL_ARG2(arg2), \
+ PVOP_CALL_ARG3(arg3), [_arg4] "mr" ((u32)(arg4)))
+#define PVOP_VCALL4(op, arg1, arg2, arg3, arg4) \
+ __PVOP_VCALL(op, \
+ "push %[_arg4];", "lea 4(%%esp),%%esp;", \
+ "0" ((u32)(arg1)), "1" ((u32)(arg2)), \
+ "2" ((u32)(arg3)), [_arg4] "mr" ((u32)(arg4)))
+#else
+#define PVOP_CALL4(rettype, op, arg1, arg2, arg3, arg4) \
+ __PVOP_CALL(rettype, op, "", "", \
+ PVOP_CALL_ARG1(arg1), PVOP_CALL_ARG2(arg2), \
+ PVOP_CALL_ARG3(arg3), PVOP_CALL_ARG4(arg4))
+#define PVOP_VCALL4(op, arg1, arg2, arg3, arg4) \
+ __PVOP_VCALL(op, "", "", \
+ PVOP_CALL_ARG1(arg1), PVOP_CALL_ARG2(arg2), \
+ PVOP_CALL_ARG3(arg3), PVOP_CALL_ARG4(arg4))
+#endif
+
+/* 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_start_context_switch(struct task_struct *prev);
+void paravirt_end_context_switch(struct task_struct *next);
+
+void paravirt_enter_lazy_mmu(void);
+void paravirt_leave_lazy_mmu(void);
+
+void _paravirt_nop(void);
+u32 _paravirt_ident_32(u32);
+u64 _paravirt_ident_64(u64);
+
+#define paravirt_nop ((void *)_paravirt_nop)
+
+/* These all sit in the .parainstructions section to tell us what to patch. */
+struct paravirt_patch_site {
+ u8 *instr; /* original instructions */
+ u8 instrtype; /* type of this instruction */
+ u8 len; /* length of original instruction */
+ u16 clobbers; /* what registers you may clobber */
+};
+
+extern struct paravirt_patch_site __parainstructions[],
+ __parainstructions_end[];
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* _ASM_X86_PARAVIRT_TYPES_H */
--
1.6.0.6
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2009-May-07 20:32 UTC
[Xen-devel] [PATCH 2/5] x86: split out core __math_state_restore
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Split the core fpu state restoration out into __math_state_restore, which
assumes that cr0.TS is clear and that the fpu context has been initialized.
This will be used during context switch. There are two reasons this is
desireable:
- There''s a small clarification. When __switch_to() calls
math_state_restore,
it relies on the fact that tsk_used_math() returns true, and so will
never do a blocking init_fpu(). __math_state_restore() does not have
(or need) that logic, so the question never arises.
- It allows the clts() to be moved earler in __switch_to() so it can be
performed
while cpu context updates are batched (will be done in a later patch).
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/include/asm/i387.h | 1 +
arch/x86/kernel/traps.c | 33 +++++++++++++++++++++++----------
2 files changed, 24 insertions(+), 10 deletions(-)
diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h
index 63d1850..6eda5d2 100644
--- a/arch/x86/include/asm/i387.h
+++ b/arch/x86/include/asm/i387.h
@@ -26,6 +26,7 @@ extern void fpu_init(void);
extern void mxcsr_feature_mask_init(void);
extern int init_fpu(struct task_struct *child);
extern asmlinkage void math_state_restore(void);
+extern void __math_state_restore(void);
extern void init_thread_xstate(void);
extern int dump_fpu(struct pt_regs *, struct user_i387_struct *);
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index f07ada4..4ea50f7 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -814,6 +814,28 @@ asmlinkage void __attribute__((weak))
mce_threshold_interrupt(void)
#endif
/*
+ * __math_state_restore assumes that cr0.TS is already clear and the
+ * fpu state is all ready for use. Used during context switch.
+ */
+inline void __math_state_restore(void)
+{
+ struct thread_info *thread = current_thread_info();
+ struct task_struct *tsk = thread->task;
+
+ /*
+ * Paranoid restore. send a SIGSEGV if we fail to restore the state.
+ */
+ if (unlikely(restore_fpu_checking(tsk))) {
+ stts();
+ force_sig(SIGSEGV, tsk);
+ return;
+ }
+
+ thread->status |= TS_USEDFPU; /* So we fnsave on switch_to() */
+ tsk->fpu_counter++;
+}
+
+/*
* ''math_state_restore()'' saves the current math information
in the
* old math state array, and gets the new ones from the current task
*
@@ -844,17 +866,8 @@ asmlinkage void math_state_restore(void)
}
clts(); /* Allow maths ops (or we recurse) */
- /*
- * Paranoid restore. send a SIGSEGV if we fail to restore the state.
- */
- if (unlikely(restore_fpu_checking(tsk))) {
- stts();
- force_sig(SIGSEGV, tsk);
- return;
- }
- thread->status |= TS_USEDFPU; /* So we fnsave on switch_to() */
- tsk->fpu_counter++;
+ __math_state_restore();
}
EXPORT_SYMBOL_GPL(math_state_restore);
--
1.6.0.6
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2009-May-07 20:32 UTC
[Xen-devel] [PATCH 3/5] x86-32: make sure clts is batched during context switch
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> If we''re preloading the fpu state during context switch, make sure the clts happens while we''re batching the cpu context update, then do the actual __math_state_restore once the updates are flushed. This allows more efficient context switches when running paravirtualized, as all the hypercalls can be folded together into one. Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> --- arch/x86/kernel/process_32.c | 27 ++++++++++++++++----------- 1 files changed, 16 insertions(+), 11 deletions(-) diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 5463858..022671c 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -353,14 +353,21 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) *next = &next_p->thread; int cpu = smp_processor_id(); struct tss_struct *tss = &per_cpu(init_tss, cpu); + bool preload_fpu; /* never put a printk in __switch_to... printk() calls wake_up*() indirectly */ - __unlazy_fpu(prev_p); + /* + * If the task has used fpu the last 5 timeslices, just do a full + * restore of the math state immediately to avoid the trap; the + * chances of needing FPU soon are obviously high now + */ + preload_fpu = tsk_used_math(next_p) && next_p->fpu_counter > 5; + __unlazy_fpu(prev_p); /* we''re going to use this soon, after a few expensive things */ - if (next_p->fpu_counter > 5) + if (preload_fpu) prefetch(next->xstate); /* @@ -401,6 +408,11 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) task_thread_info(next_p)->flags & _TIF_WORK_CTXSW_NEXT)) __switch_to_xtra(prev_p, next_p, tss); + /* If we''re going to preload the fpu context, make sure clts + is run while we''re batching the cpu state updates. */ + if (preload_fpu) + clts(); + /* * Leave lazy mode, flushing any hypercalls made here. * This must be done before restoring TLS segments so @@ -410,15 +422,8 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) */ arch_end_context_switch(next_p); - /* If the task has used fpu the last 5 timeslices, just do a full - * restore of the math state immediately to avoid the trap; the - * chances of needing FPU soon are obviously high now - * - * tsk_used_math() checks prevent calling math_state_restore(), - * which can sleep in the case of !tsk_used_math() - */ - if (tsk_used_math(next_p) && next_p->fpu_counter > 5) - math_state_restore(); + if (preload_fpu) + __math_state_restore(); /* * Restore %gs if needed (which is common) -- 1.6.0.6 _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2009-May-07 20:32 UTC
[Xen-devel] [PATCH 4/5] x86-64: move unlazy_fpu() into lazy cpu state part of context switch
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Make sure that unlazy_fpu()''s stts gets batched along with the other cpu state changes during context switch. (32-bit already does this.) This makes sure it gets batched when running paravirtualized. Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> --- arch/x86/kernel/process_64.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 2b659f2..534859b 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -422,6 +422,9 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) load_TLS(next, cpu); + /* Must be after DS reload */ + unlazy_fpu(prev_p); + /* * Leave lazy mode, flushing any hypercalls made here. * This must be done before restoring TLS segments so @@ -462,9 +465,6 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) wrmsrl(MSR_KERNEL_GS_BASE, next->gs); prev->gsindex = gsindex; - /* Must be after DS reload */ - unlazy_fpu(prev_p); - /* * Switch the PDA and FPU contexts. */ -- 1.6.0.6 _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2009-May-07 20:32 UTC
[Xen-devel] [PATCH 5/5] x86-64: move clts into batch cpu state updates when preloading fpu
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> When a task is likely to be using the fpu, we preload its state during the context switch, rather than waiting for it to run an fpu instruction. Make sure the clts() happens while we''re doing batched fpu state updates to optimise paravirtualized context switches. Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> --- arch/x86/kernel/process_64.c | 28 +++++++++++++++++++--------- 1 files changed, 19 insertions(+), 9 deletions(-) diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 534859b..45c5eb8 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -389,9 +389,17 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) int cpu = smp_processor_id(); struct tss_struct *tss = &per_cpu(init_tss, cpu); unsigned fsindex, gsindex; + bool preload_fpu; + + /* + * If the task has used fpu the last 5 timeslices, just do a full + * restore of the math state immediately to avoid the trap; the + * chances of needing FPU soon are obviously high now + */ + preload_fpu = tsk_used_math(next_p) && next_p->fpu_counter > 5; /* we''re going to use this soon, after a few expensive things */ - if (next_p->fpu_counter > 5) + if (preload_fpu) prefetch(next->xstate); /* @@ -425,6 +433,10 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) /* Must be after DS reload */ unlazy_fpu(prev_p); + /* Make sure cpu is ready for new context */ + if (preload_fpu) + clts(); + /* * Leave lazy mode, flushing any hypercalls made here. * This must be done before restoring TLS segments so @@ -483,15 +495,13 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) task_thread_info(prev_p)->flags & _TIF_WORK_CTXSW_PREV)) __switch_to_xtra(prev_p, next_p, tss); - /* If the task has used fpu the last 5 timeslices, just do a full - * restore of the math state immediately to avoid the trap; the - * chances of needing FPU soon are obviously high now - * - * tsk_used_math() checks prevent calling math_state_restore(), - * which can sleep in the case of !tsk_used_math() + /* + * Preload the FPU context, now that we''ve determined that the + * task is likely to be using it. */ - if (tsk_used_math(next_p) && next_p->fpu_counter > 5) - math_state_restore(); + if (preload_fpu) + __math_state_restore(); + return prev_p; } -- 1.6.0.6 _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2009-May-12 00:25 UTC
Re: [Xen-devel] [GIT PULL] x86: paravirt_ops updates
Jeremy Fitzhardinge wrote:> Hi Ingo, > > This series has a some changes to x86 paravirt_ops: >Ping? Did you overlook these changes, or have you just not had a chance to look at them? Thanks, J> x86/paravirt: split paravirt definitions into paravirt_types.h > > Move all the type-like definitions into asm/paravirt_types.h > > x86: split out core __math_state_restore > > Split out the core of math_state_restore() into __math_state_restore(), > which just does what''s needed for a context switch (ie, assumes > TS is already clear, and that the fpu context exists). > > x86-32: make sure clts is batched during context switch > x86-64: move unlazy_fpu() into lazy cpu state part of context switch > x86-64: move clts into batch cpu state updates when preloading fpu > > Make sure the TS state updates during context switch are performed in > the lazy CPU update region so that they can be batched when running > paravirtualized. > > The following changes since commit 9ce5424d75e56891905b77d1589924765e62059a: > Ingo Molnar (1): > Merge branch ''x86/urgent'' > > are available in the git repository at: > > git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen.git x86/paravirt > > Jeremy Fitzhardinge (5): > x86/paravirt: split paravirt definitions into paravirt_types.h > x86: split out core __math_state_restore > x86-32: make sure clts is batched during context switch > x86-64: move unlazy_fpu() into lazy cpu state part of context switch > x86-64: move clts into batch cpu state updates when preloading fpu > > arch/x86/include/asm/i387.h | 1 + > arch/x86/include/asm/paravirt.h | 711 +-------------------------------- > arch/x86/include/asm/paravirt_types.h | 720 +++++++++++++++++++++++++++++++++ > arch/x86/kernel/process_32.c | 27 +- > arch/x86/kernel/process_64.c | 34 +- > arch/x86/kernel/traps.c | 33 +- > 6 files changed, 783 insertions(+), 743 deletions(-) > create mode 100644 arch/x86/include/asm/paravirt_types.h > > Thanks, > J > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel