#include <linux/mm.h>
#include <linux/page-flags.h>
#include <linux/highmem.h>
+#include <linux/console.h>
#include <xen/interface/xen.h>
#include <xen/interface/physdev.h>
#include <asm/pgtable.h>
#include <asm/tlbflush.h>
#include <asm/reboot.h>
+#include <asm/pgalloc.h>
#include "xen-ops.h"
#include "mmu.h"
struct start_info *xen_start_info;
EXPORT_SYMBOL_GPL(xen_start_info);
-static /* __initdata */ struct shared_info dummy_shared_info;
+struct shared_info xen_dummy_shared_info;
/*
* Point at some empty memory to start with. We map the real shared_info
* page as soon as fixmap is up and running.
*/
-struct shared_info *HYPERVISOR_shared_info = (void *)&dummy_shared_info;
+struct shared_info *HYPERVISOR_shared_info = (void *)&xen_dummy_shared_info;
/*
* Flag to determine whether vcpu info placement is available on all
*
* 0: not available, 1: available
*/
-static int have_vcpu_info_placement = 0;
+static int have_vcpu_info_placement = 1;
-static void __init xen_vcpu_setup(int cpu)
+static void xen_vcpu_setup(int cpu)
{
struct vcpu_register_vcpu_info info;
int err;
struct vcpu_info *vcpup;
+ BUG_ON(HYPERVISOR_shared_info == &xen_dummy_shared_info);
per_cpu(xen_vcpu, cpu) = &HYPERVISOR_shared_info->vcpu_info[cpu];
if (!have_vcpu_info_placement)
}
}
+/*
+ * On restore, set the vcpu placement up again.
+ * If it fails, then we're in a bad state, since
+ * we can't back out from using it...
+ */
+void xen_vcpu_restore(void)
+{
+ if (have_vcpu_info_placement) {
+ int cpu;
+
+ for_each_online_cpu(cpu) {
+ bool other_cpu = (cpu != smp_processor_id());
+
+ if (other_cpu &&
+ HYPERVISOR_vcpu_op(VCPUOP_down, cpu, NULL))
+ BUG();
+
+ xen_vcpu_setup(cpu);
+
+ if (other_cpu &&
+ HYPERVISOR_vcpu_op(VCPUOP_up, cpu, NULL))
+ BUG();
+ }
+
+ BUG_ON(!have_vcpu_info_placement);
+ }
+}
+
static void __init xen_banner(void)
{
printk(KERN_INFO "Booting paravirtualized kernel on %s\n",
pv_info.name);
- printk(KERN_INFO "Hypervisor signature: %s\n", xen_start_info->magic);
+ printk(KERN_INFO "Hypervisor signature: %s%s\n",
+ xen_start_info->magic,
+ xen_feature(XENFEAT_mmu_pt_update_preserve_ad) ? " (preserve-AD)" : "");
}
static void xen_cpuid(unsigned int *ax, unsigned int *bx,
if (*ax == 1)
maskedx = ~((1 << X86_FEATURE_APIC) | /* disable APIC */
(1 << X86_FEATURE_ACPI) | /* disable ACPI */
+ (1 << X86_FEATURE_MCE) | /* disable MCE */
+ (1 << X86_FEATURE_MCA) | /* disable MCA */
(1 << X86_FEATURE_ACC)); /* thermal monitoring */
asm(XEN_EMULATE_PREFIX "cpuid"
{
struct vcpu_info *vcpu;
- /* There's a one instruction preempt window here. We need to
- make sure we're don't switch CPUs between getting the vcpu
- pointer and updating the mask. */
- preempt_disable();
+ /* We don't need to worry about being preempted here, since
+ either a) interrupts are disabled, so no preemption, or b)
+ the caller is confused and is trying to re-enable interrupts
+ on an indeterminate processor. */
+
vcpu = x86_read_percpu(xen_vcpu);
vcpu->evtchn_upcall_mask = 0;
- preempt_enable_no_resched();
/* Doesn't matter if we get preempted here, because any
pending event will get dealt with anyway. */
static void xen_safe_halt(void)
{
/* Blocking includes an implicit local_irq_enable(). */
- if (HYPERVISOR_sched_op(SCHEDOP_block, 0) != 0)
+ if (HYPERVISOR_sched_op(SCHEDOP_block, NULL) != 0)
BUG();
}
static void xen_flush_tlb(void)
{
struct mmuext_op *op;
- struct multicall_space mcs = xen_mc_entry(sizeof(*op));
+ struct multicall_space mcs;
+
+ preempt_disable();
+
+ mcs = xen_mc_entry(sizeof(*op));
op = mcs.args;
op->cmd = MMUEXT_TLB_FLUSH_LOCAL;
MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
xen_mc_issue(PARAVIRT_LAZY_MMU);
+
+ preempt_enable();
}
static void xen_flush_tlb_single(unsigned long addr)
{
struct mmuext_op *op;
- struct multicall_space mcs = xen_mc_entry(sizeof(*op));
+ struct multicall_space mcs;
+
+ preempt_disable();
+ mcs = xen_mc_entry(sizeof(*op));
op = mcs.args;
op->cmd = MMUEXT_INVLPG_LOCAL;
op->arg1.linear_addr = addr & PAGE_MASK;
MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
xen_mc_issue(PARAVIRT_LAZY_MMU);
+
+ preempt_enable();
}
static void xen_flush_tlb_others(const cpumask_t *cpus, struct mm_struct *mm,
xen_mc_issue(PARAVIRT_LAZY_MMU);
}
+static void xen_clts(void)
+{
+ struct multicall_space mcs;
+
+ mcs = xen_mc_entry(0);
+
+ MULTI_fpu_taskswitch(mcs.mc, 0);
+
+ xen_mc_issue(PARAVIRT_LAZY_CPU);
+}
+
+static void xen_write_cr0(unsigned long cr0)
+{
+ struct multicall_space mcs;
+
+ /* Only pay attention to cr0.TS; everything else is
+ ignored. */
+ mcs = xen_mc_entry(0);
+
+ MULTI_fpu_taskswitch(mcs.mc, (cr0 & X86_CR0_TS) != 0);
+
+ xen_mc_issue(PARAVIRT_LAZY_CPU);
+}
+
static void xen_write_cr2(unsigned long cr2)
{
x86_read_percpu(xen_vcpu)->arch.cr2 = cr2;
static void xen_write_cr4(unsigned long cr4)
{
- /* Just ignore cr4 changes; Xen doesn't allow us to do
- anything anyway. */
+ cr4 &= ~X86_CR4_PGE;
+ cr4 &= ~X86_CR4_PSE;
+
+ native_write_cr4(cr4);
}
static unsigned long xen_read_cr3(void)
/* Early in boot, while setting up the initial pagetable, assume
everything is pinned. */
-static __init void xen_alloc_pt_init(struct mm_struct *mm, u32 pfn)
+static __init void xen_alloc_pte_init(struct mm_struct *mm, u32 pfn)
{
+#ifdef CONFIG_FLATMEM
BUG_ON(mem_map); /* should only be used early */
+#endif
make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
}
-static void pin_pagetable_pfn(unsigned level, unsigned long pfn)
+/* Early release_pte assumes that all pts are pinned, since there's
+ only init_mm and anything attached to that is pinned. */
+static void xen_release_pte_init(u32 pfn)
+{
+ make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
+}
+
+static void pin_pagetable_pfn(unsigned cmd, unsigned long pfn)
{
struct mmuext_op op;
- op.cmd = level;
+ op.cmd = cmd;
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)
+static void xen_alloc_ptpage(struct mm_struct *mm, u32 pfn, unsigned level)
{
struct page *page = pfn_to_page(pfn);
if (!PageHighMem(page)) {
make_lowmem_page_readonly(__va(PFN_PHYS(pfn)));
- pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn);
+ if (level == PT_PTE)
+ pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn);
} else
/* make sure there are no stray mappings of
this page */
}
}
+static void xen_alloc_pte(struct mm_struct *mm, u32 pfn)
+{
+ xen_alloc_ptpage(mm, pfn, PT_PTE);
+}
+
+static void xen_alloc_pmd(struct mm_struct *mm, u32 pfn)
+{
+ xen_alloc_ptpage(mm, pfn, PT_PMD);
+}
+
/* This should never happen until we're OK to use struct page */
-static void xen_release_pt(u32 pfn)
+static void xen_release_ptpage(u32 pfn, unsigned level)
{
struct page *page = pfn_to_page(pfn);
if (PagePinned(page)) {
if (!PageHighMem(page)) {
- pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn);
+ if (level == PT_PTE)
+ pin_pagetable_pfn(MMUEXT_UNPIN_TABLE, pfn);
make_lowmem_page_readwrite(__va(PFN_PHYS(pfn)));
}
+ ClearPagePinned(page);
}
}
+static void xen_release_pte(u32 pfn)
+{
+ xen_release_ptpage(pfn, PT_PTE);
+}
+
+static void xen_release_pmd(u32 pfn)
+{
+ xen_release_ptpage(pfn, PT_PMD);
+}
+
#ifdef CONFIG_HIGHPTE
static void *xen_kmap_atomic_pte(struct page *page, enum km_type type)
{
static __init void xen_pagetable_setup_start(pgd_t *base)
{
pgd_t *xen_pgd = (pgd_t *)xen_start_info->pt_base;
+ int i;
/* special set_pte for pagetable initialization */
pv_mmu_ops.set_pte = xen_set_pte_init;
init_mm.pgd = base;
/*
- * copy top-level of Xen-supplied pagetable into place. For
- * !PAE we can use this as-is, but for PAE it is a stand-in
- * while we copy the pmd pages.
+ * copy top-level of Xen-supplied pagetable into place. This
+ * is a stand-in while we copy the pmd pages.
*/
memcpy(base, xen_pgd, PTRS_PER_PGD * sizeof(pgd_t));
- if (PTRS_PER_PMD > 1) {
- int i;
- /*
- * For PAE, need to allocate new pmds, rather than
- * share Xen's, since Xen doesn't like pmd's being
- * shared between address spaces.
- */
- for (i = 0; i < PTRS_PER_PGD; i++) {
- if (pgd_val_ma(xen_pgd[i]) & _PAGE_PRESENT) {
- pmd_t *pmd = (pmd_t *)alloc_bootmem_low_pages(PAGE_SIZE);
+ /*
+ * For PAE, need to allocate new pmds, rather than
+ * share Xen's, since Xen doesn't like pmd's being
+ * shared between address spaces.
+ */
+ for (i = 0; i < PTRS_PER_PGD; i++) {
+ if (pgd_val_ma(xen_pgd[i]) & _PAGE_PRESENT) {
+ pmd_t *pmd = (pmd_t *)alloc_bootmem_low_pages(PAGE_SIZE);
- memcpy(pmd, (void *)pgd_page_vaddr(xen_pgd[i]),
- PAGE_SIZE);
+ memcpy(pmd, (void *)pgd_page_vaddr(xen_pgd[i]),
+ PAGE_SIZE);
- make_lowmem_page_readonly(pmd);
+ make_lowmem_page_readonly(pmd);
- set_pgd(&base[i], __pgd(1 + __pa(pmd)));
- } else
- pgd_clear(&base[i]);
- }
+ set_pgd(&base[i], __pgd(1 + __pa(pmd)));
+ } else
+ pgd_clear(&base[i]);
}
/* make sure zero_page is mapped RO so we can use it in pagetables */
* added to the table can be prepared properly for Xen.
*/
xen_write_cr3(__pa(base));
+
+ /* Unpin initial Xen pagetable */
+ pin_pagetable_pfn(MMUEXT_UNPIN_TABLE,
+ PFN_DOWN(__pa(xen_start_info->pt_base)));
}
-static __init void xen_pagetable_setup_done(pgd_t *base)
+void xen_setup_shared_info(void)
{
- /* This will work as long as patching hasn't happened yet
- (which it hasn't) */
- pv_mmu_ops.alloc_pt = xen_alloc_pt;
- pv_mmu_ops.set_pte = xen_set_pte;
-
if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+ unsigned long addr = fix_to_virt(FIX_PARAVIRT_BOOTMAP);
+
/*
* Create a mapping for the shared info page.
* Should be set_fixmap(), but shared_info is a machine
* address with no corresponding pseudo-phys address.
*/
- set_pte_mfn(fix_to_virt(FIX_PARAVIRT_BOOTMAP),
+ set_pte_mfn(addr,
PFN_DOWN(xen_start_info->shared_info),
PAGE_KERNEL);
- HYPERVISOR_shared_info =
- (struct shared_info *)fix_to_virt(FIX_PARAVIRT_BOOTMAP);
-
+ HYPERVISOR_shared_info = (struct shared_info *)addr;
} else
HYPERVISOR_shared_info =
(struct shared_info *)__va(xen_start_info->shared_info);
+#ifndef CONFIG_SMP
+ /* In UP this is as good a place as any to set up shared info */
+ xen_setup_vcpu_info_placement();
+#endif
+
+ xen_setup_mfn_list_list();
+}
+
+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) */
+ pv_mmu_ops.alloc_pte = xen_alloc_pte;
+ pv_mmu_ops.alloc_pmd = xen_alloc_pmd;
+ pv_mmu_ops.release_pte = xen_release_pte;
+ pv_mmu_ops.release_pmd = xen_release_pmd;
+ pv_mmu_ops.set_pte = xen_set_pte;
+
+ xen_setup_shared_info();
+
/* Actually pin the pagetable down, but we can't set PG_pinned
yet because the page structures don't exist yet. */
- {
- unsigned level;
+ pin_pagetable_pfn(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(base)));
+}
-#ifdef CONFIG_X86_PAE
- level = MMUEXT_PIN_L3_TABLE;
-#else
- level = MMUEXT_PIN_L2_TABLE;
-#endif
+static __init void xen_post_allocator_init(void)
+{
+ pv_mmu_ops.set_pmd = xen_set_pmd;
+ pv_mmu_ops.set_pud = xen_set_pud;
- pin_pagetable_pfn(level, PFN_DOWN(__pa(base)));
- }
+ xen_mark_init_mm_pinned();
}
/* This is called once we have the cpu_possible_map */
-void __init xen_setup_vcpu_info_placement(void)
+void xen_setup_vcpu_info_placement(void)
{
int cpu;
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;
}
}
return ret;
}
+static void xen_set_fixmap(unsigned idx, unsigned long phys, pgprot_t prot)
+{
+ pte_t pte;
+
+ phys >>= PAGE_SHIFT;
+
+ switch (idx) {
+ case FIX_BTMAP_END ... FIX_BTMAP_BEGIN:
+#ifdef CONFIG_X86_F00F_BUG
+ case FIX_F00F_IDT:
+#endif
+ case FIX_WP_TEST:
+ case FIX_VDSO:
+#ifdef CONFIG_X86_LOCAL_APIC
+ case FIX_APIC_BASE: /* maps dummy local APIC */
+#endif
+ pte = pfn_pte(phys, prot);
+ break;
+
+ default:
+ pte = mfn_pte(phys, prot);
+ break;
+ }
+
+ __native_set_fixmap(idx, pte);
+}
+
static const struct pv_info xen_info __initdata = {
.paravirt_enabled = 1,
.shared_kernel_pmd = 0,
.banner = xen_banner,
.memory_setup = xen_memory_setup,
.arch_setup = xen_arch_setup,
- .post_allocator_init = xen_mark_init_mm_pinned,
+ .post_allocator_init = xen_post_allocator_init,
};
static const struct pv_time_ops xen_time_ops __initdata = {
.set_wallclock = xen_set_wallclock,
.get_wallclock = xen_get_wallclock,
- .get_cpu_khz = xen_cpu_khz,
+ .get_tsc_khz = xen_tsc_khz,
.sched_clock = xen_sched_clock,
};
.set_debugreg = xen_set_debugreg,
.get_debugreg = xen_get_debugreg,
- .clts = native_clts,
+ .clts = xen_clts,
.read_cr0 = native_read_cr0,
- .write_cr0 = native_write_cr0,
+ .write_cr0 = xen_write_cr0,
.read_cr4 = native_read_cr4,
.read_cr4_safe = native_read_cr4_safe,
.read_tsc = native_read_tsc,
.read_pmc = native_read_pmc,
- .iret = (void *)&hypercall_page[__HYPERVISOR_iret],
- .irq_enable_syscall_ret = NULL, /* never called */
+ .iret = xen_iret,
+ .irq_enable_sysexit = xen_sysexit,
.load_tr_desc = paravirt_nop,
.set_ldt = xen_set_ldt,
.irq_enable = xen_irq_enable,
.safe_halt = xen_safe_halt,
.halt = xen_halt,
+#ifdef CONFIG_X86_64
+ .adjust_exception_frame = paravirt_nop,
+#endif
};
static const struct pv_apic_ops xen_apic_ops __initdata = {
.pte_update = paravirt_nop,
.pte_update_defer = paravirt_nop,
- .alloc_pt = xen_alloc_pt_init,
- .release_pt = xen_release_pt,
- .alloc_pd = paravirt_nop,
- .alloc_pd_clone = paravirt_nop,
- .release_pd = paravirt_nop,
+ .pgd_alloc = __paravirt_pgd_alloc,
+ .pgd_free = paravirt_nop,
+
+ .alloc_pte = xen_alloc_pte_init,
+ .release_pte = xen_release_pte_init,
+ .alloc_pmd = xen_alloc_pte_init,
+ .alloc_pmd_clone = paravirt_nop,
+ .release_pmd = xen_release_pte_init,
#ifdef CONFIG_HIGHPTE
.kmap_atomic_pte = xen_kmap_atomic_pte,
.set_pte = NULL, /* see xen_pagetable_setup_* */
.set_pte_at = xen_set_pte_at,
- .set_pmd = xen_set_pmd,
+ .set_pmd = xen_set_pmd_hyper,
+
+ .ptep_modify_prot_start = __ptep_modify_prot_start,
+ .ptep_modify_prot_commit = __ptep_modify_prot_commit,
.pte_val = xen_pte_val,
+ .pte_flags = native_pte_val,
.pgd_val = xen_pgd_val,
.make_pte = xen_make_pte,
.make_pgd = xen_make_pgd,
-#ifdef CONFIG_X86_PAE
.set_pte_atomic = xen_set_pte_atomic,
.set_pte_present = xen_set_pte_at,
- .set_pud = xen_set_pud,
+ .set_pud = xen_set_pud_hyper,
.pte_clear = xen_pte_clear,
.pmd_clear = xen_pmd_clear,
.make_pmd = xen_make_pmd,
.pmd_val = xen_pmd_val,
-#endif /* PAE */
.activate_mm = xen_activate_mm,
.dup_mmap = xen_dup_mmap,
.enter = paravirt_enter_lazy_mmu,
.leave = xen_leave_lazy,
},
+
+ .set_fixmap = xen_set_fixmap,
};
#ifdef CONFIG_SMP
.smp_send_stop = xen_smp_send_stop,
.smp_send_reschedule = xen_smp_send_reschedule,
- .smp_call_function_mask = xen_smp_call_function_mask,
+
+ .send_call_func_ipi = xen_smp_send_call_function_ipi,
+ .send_call_func_single_ipi = xen_smp_send_call_function_single_ipi,
};
#endif /* CONFIG_SMP */
static void xen_reboot(int reason)
{
+ struct sched_shutdown r = { .reason = reason };
+
#ifdef CONFIG_SMP
smp_send_stop();
#endif
- if (HYPERVISOR_sched_op(SCHEDOP_shutdown, reason))
+ if (HYPERVISOR_sched_op(SCHEDOP_shutdown, &r))
BUG();
}
BUG_ON(memcmp(xen_start_info->magic, "xen-3", 5) != 0);
+ xen_setup_features();
+
/* Install Xen paravirt ops */
pv_info = xen_info;
pv_init_ops = xen_init_ops;
pv_apic_ops = xen_apic_ops;
pv_mmu_ops = xen_mmu_ops;
+ if (xen_feature(XENFEAT_mmu_pt_update_preserve_ad)) {
+ pv_mmu_ops.ptep_modify_prot_start = xen_ptep_modify_prot_start;
+ pv_mmu_ops.ptep_modify_prot_commit = xen_ptep_modify_prot_commit;
+ }
+
machine_ops = xen_machine_ops;
#ifdef CONFIG_SMP
smp_ops = xen_smp_ops;
#endif
- xen_setup_features();
-
/* Get mfn list */
if (!xen_feature(XENFEAT_auto_translated_physmap))
- phys_to_machine_mapping = (unsigned long *)xen_start_info->mfn_list;
+ xen_build_dynamic_phys_to_machine();
pgd = (pgd_t *)xen_start_info->pt_base;
+ init_pg_tables_start = __pa(pgd);
init_pg_tables_end = __pa(pgd) + xen_start_info->nr_pt_frames*PAGE_SIZE;
+ max_pfn_mapped = (init_pg_tables_end + 512*1024) >> PAGE_SHIFT;
init_mm.pgd = pgd; /* use the Xen pagetables to start */
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
- possible map. */
+ possible map and a non-dummy shared_info. */
per_cpu(xen_vcpu, 0) = &HYPERVISOR_shared_info->vcpu_info[0];
-#else
- /* May as well do it now, since there's no good time to call
- it later on UP. */
- xen_setup_vcpu_info_placement();
-#endif
pv_info.kernel_rpl = 1;
if (xen_feature(XENFEAT_supervisor_mode_kernel))
pv_info.kernel_rpl = 0;
+ /* Prevent unwanted bits from being set in PTEs. */
+ __supported_pte_mask &= ~_PAGE_GLOBAL;
+ if (!is_initial_xendomain())
+ __supported_pte_mask &= ~(_PAGE_PWT | _PAGE_PCD);
+
/* set the limit of our address space */
xen_reserve_top();
? __pa(xen_start_info->mod_start) : 0;
boot_params.hdr.ramdisk_size = xen_start_info->mod_len;
+ if (!is_initial_xendomain()) {
+ add_preferred_console("xenboot", 0, NULL);
+ add_preferred_console("tty", 0, NULL);
+ add_preferred_console("hvc", 0, NULL);
+ }
+
/* Start the world */
- start_kernel();
+ i386_start_kernel();
}