More x86-64 stuff.
3ddb79bcSC_LvnmFlX-T5iTgaR0SKg xen/arch/x86/boot/x86_32.S
40e42bdbNu4MjI750THP_8J1S-Sa0g xen/arch/x86/boot/x86_64.S
3ddb79bcUrk2EIaM5VsT6wUudH1kkg xen/arch/x86/delay.c
-3e32af9aRnYGl4GMOaDKp7JdfhOGhg xen/arch/x86/domain_page.c
-3ddb79bcecupHj56ZbTa3B0FxDowMg xen/arch/x86/entry.S
+3ddb79bc1_2bAt67x9MFCP4AZrQnvQ xen/arch/x86/domain.c
3ddb79bcY5zW7KhvI9gvfuPi3ZumEg xen/arch/x86/extable.c
3fe443fdDDb0Sw6NQBCk4GQapayfTA xen/arch/x86/flushtlb.c
3ddb79bcesE5E-lS4QhRhlqXxqj9cA xen/arch/x86/i387.c
3ddb79bcBit4xJXbwtX0kb1hh2uO1Q xen/arch/x86/idle0_task.c
3ddb79bcKIkRR0kqWaJhe5VUDkMdxg xen/arch/x86/io_apic.c
3ddb79bdqfIcjkz_h9Hvtp8Tk_19Zw xen/arch/x86/irq.c
-3ddb79bcHwuCQDjBICDTSis52hWguw xen/arch/x86/mm.c
3ddb79bdS4UeWWXDH-FaBKqcpMFcnw xen/arch/x86/mpparse.c
3f12cff65EV3qOG2j37Qm0ShgvXGRw xen/arch/x86/nmi.c
3ddb79bdHe6_Uij4-glW91vInNtBYQ xen/arch/x86/pci-irq.c
3ddb79bdIKgipvGoqExEQ7jawfVowA xen/arch/x86/pci-x86.h
40a4dfced2dnSzbKgJFlD3chKHexjQ xen/arch/x86/pdb-linux.c
4022a73czgX7d-2zfF_cb33oVemApQ xen/arch/x86/pdb-stub.c
-3ddb79bc1_2bAt67x9MFCP4AZrQnvQ xen/arch/x86/process.c
3ddb79bc7KxGCEJsgBnkDX7XjD_ZEQ xen/arch/x86/rwlock.c
3ddb79bcrD6Z_rUvSDgrvjyb4846Eg xen/arch/x86/setup.c
3ddb79bcSx2e8JSR3pdSGa8x1ScYzA xen/arch/x86/smp.c
3ddb79bc-Udq7ol-NX4q9XsYnN7A2Q xen/arch/x86/time.c
3ddb79bccYVzXZJyVaxuv5T42Z1Fsw xen/arch/x86/trampoline.S
3ddb79bcOftONV9h4QCxXOfiT0h91w xen/arch/x86/traps.c
-3ddb79bc4nTpGQOe6_-MbyZzkhlhFQ xen/arch/x86/usercopy.c
-3ddb79bcOMCu9-5mKpjIh5d0qqBDPg xen/arch/x86/xen.lds
+3e32af9aRnYGl4GMOaDKp7JdfhOGhg xen/arch/x86/x86_32/domain_page.c
+3ddb79bcecupHj56ZbTa3B0FxDowMg xen/arch/x86/x86_32/entry.S
+3ddb79bcHwuCQDjBICDTSis52hWguw xen/arch/x86/x86_32/mm.c
+3ddb79bc4nTpGQOe6_-MbyZzkhlhFQ xen/arch/x86/x86_32/usercopy.c
+3ddb79bcOMCu9-5mKpjIh5d0qqBDPg xen/arch/x86/x86_32/xen.lds
+40e96d3aLDI-nViMuYneD7VKYlZrVg xen/arch/x86/x86_64/entry.S
+40e96d3ahBTZqbTViInnq0lM03vs7A xen/arch/x86/x86_64/usercopy.c
+40e96d3akN3Hu_J5Bk-WXD8OGscrYQ xen/arch/x86/x86_64/xen.lds
3ddb79bdff-gj-jFGKjOejeHLqL8Lg xen/common/Makefile
3e397e66AyyD5fYraAySWuwi9uqSXg xen/common/ac_timer.c
4022a73c_BbDFd2YJ_NQYVvKX5Oz7w xen/common/debug-linux.c
OBJS := $(subst pdb-stub.o,,$(OBJS))
endif
+OBJS += $(patsubst %.S,%.o,$(wildcard $(TARGET_SUBARCH)/*.S))
+OBJS += $(patsubst %.c,%.o,$(wildcard $(TARGET_SUBARCH)/*.c))
+
LINK_BASE := 0xFC500000 # Xen is linked here
LOAD_BASE := 0x00100000 # Xen is loaded here
clean:
rm -f *.o *~ core boot/*.o boot/*~ boot/core
+ rm -f $(TARGET_SUBARCH)/*.o $(TARGET_SUBARCH)/*~
+ rm -f $(TARGETSUBARCH)/core
CFLAGS += -iwithprefix include -Wall -Werror -fomit-frame-pointer
CFLAGS += -I$(BASEDIR)/include -Wno-pointer-arith -Wredundant-decls
-LDFLAGS := -T xen.lds -N
+LDFLAGS := -T $(TARGET_SUBARCH)/xen.lds -N
ifeq ($(TARGET_SUBARCH),x86_32)
CFLAGS += -m32 -march=i686
*****************************************************************************/
/* used for system time scaling */
-static unsigned int bus_freq;
-static u32 bus_cycle; /* length of one bus cycle in pico-seconds */
-static u32 bus_scale; /* scaling factor convert ns to bus cycles */
+static unsigned long bus_freq; /* KAF: pointer-size avoids compile warns. */
+static u32 bus_cycle; /* length of one bus cycle in pico-seconds */
+static u32 bus_scale; /* scaling factor convert ns to bus cycles */
/*
* The timer chip is already set up at HZ interrupts per second here,
--- /dev/null
+/*
+ * Copyright (C) 1995 Linus Torvalds
+ *
+ * Pentium III FXSR, SSE support
+ * Gareth Hughes <gareth@valinux.com>, May 2000
+ */
+
+#include <xen/config.h>
+#include <xen/lib.h>
+#include <xen/errno.h>
+#include <xen/sched.h>
+#include <xen/smp.h>
+#include <xen/delay.h>
+#include <xen/softirq.h>
+#include <asm/ptrace.h>
+#include <asm/mc146818rtc.h>
+#include <asm/system.h>
+#include <asm/io.h>
+#include <asm/processor.h>
+#include <asm/desc.h>
+#include <asm/i387.h>
+#include <asm/mpspec.h>
+#include <asm/ldt.h>
+#include <xen/irq.h>
+#include <xen/event.h>
+#include <xen/shadow.h>
+
+int hlt_counter;
+
+void disable_hlt(void)
+{
+ hlt_counter++;
+}
+
+void enable_hlt(void)
+{
+ hlt_counter--;
+}
+
+/*
+ * We use this if we don't have any better
+ * idle routine..
+ */
+static void default_idle(void)
+{
+ if ( hlt_counter == 0 )
+ {
+ __cli();
+ if ( !softirq_pending(smp_processor_id()) )
+ safe_halt();
+ else
+ __sti();
+ }
+}
+
+void continue_cpu_idle_loop(void)
+{
+ int cpu = smp_processor_id();
+ for ( ; ; )
+ {
+ irq_stat[cpu].idle_timestamp = jiffies;
+ while ( !softirq_pending(cpu) )
+ default_idle();
+ do_softirq();
+ }
+}
+
+void startup_cpu_idle_loop(void)
+{
+ /* Just some sanity to ensure that the scheduler is set up okay. */
+ ASSERT(current->domain == IDLE_DOMAIN_ID);
+ domain_unpause_by_systemcontroller(current);
+ __enter_scheduler();
+
+ /*
+ * Declares CPU setup done to the boot processor.
+ * Therefore memory barrier to ensure state is visible.
+ */
+ smp_mb();
+ init_idle();
+
+ continue_cpu_idle_loop();
+}
+
+static long no_idt[2];
+static int reboot_mode;
+int reboot_thru_bios = 0;
+
+#ifdef CONFIG_SMP
+int reboot_smp = 0;
+static int reboot_cpu = -1;
+/* shamelessly grabbed from lib/vsprintf.c for readability */
+#define is_digit(c) ((c) >= '0' && (c) <= '9')
+#endif
+
+
+static inline void kb_wait(void)
+{
+ int i;
+
+ for (i=0; i<0x10000; i++)
+ if ((inb_p(0x64) & 0x02) == 0)
+ break;
+}
+
+
+void machine_restart(char * __unused)
+{
+ extern int opt_noreboot;
+#ifdef CONFIG_SMP
+ int cpuid;
+#endif
+
+ if ( opt_noreboot )
+ {
+ printk("Reboot disabled on cmdline: require manual reset\n");
+ for ( ; ; ) __asm__ __volatile__ ("hlt");
+ }
+
+#ifdef CONFIG_SMP
+ cpuid = GET_APIC_ID(apic_read(APIC_ID));
+
+ /* KAF: Need interrupts enabled for safe IPI. */
+ __sti();
+
+ if (reboot_smp) {
+
+ /* check to see if reboot_cpu is valid
+ if its not, default to the BSP */
+ if ((reboot_cpu == -1) ||
+ (reboot_cpu > (NR_CPUS -1)) ||
+ !(phys_cpu_present_map & (1<<cpuid)))
+ reboot_cpu = boot_cpu_physical_apicid;
+
+ reboot_smp = 0; /* use this as a flag to only go through this once*/
+ /* re-run this function on the other CPUs
+ it will fall though this section since we have
+ cleared reboot_smp, and do the reboot if it is the
+ correct CPU, otherwise it halts. */
+ if (reboot_cpu != cpuid)
+ smp_call_function((void *)machine_restart , NULL, 1, 0);
+ }
+
+ /* if reboot_cpu is still -1, then we want a tradional reboot,
+ and if we are not running on the reboot_cpu,, halt */
+ if ((reboot_cpu != -1) && (cpuid != reboot_cpu)) {
+ for (;;)
+ __asm__ __volatile__ ("hlt");
+ }
+ /*
+ * Stop all CPUs and turn off local APICs and the IO-APIC, so
+ * other OSs see a clean IRQ state.
+ */
+ smp_send_stop();
+ disable_IO_APIC();
+#endif
+
+ if(!reboot_thru_bios) {
+ /* rebooting needs to touch the page at absolute addr 0 */
+ *((unsigned short *)__va(0x472)) = reboot_mode;
+ for (;;) {
+ int i;
+ for (i=0; i<100; i++) {
+ kb_wait();
+ udelay(50);
+ outb(0xfe,0x64); /* pulse reset low */
+ udelay(50);
+ }
+ /* That didn't work - force a triple fault.. */
+ __asm__ __volatile__("lidt %0": "=m" (no_idt));
+ __asm__ __volatile__("int3");
+ }
+ }
+
+ panic("Need to reinclude BIOS reboot code\n");
+}
+
+void machine_halt(void)
+{
+ machine_restart(0);
+}
+
+void machine_power_off(void)
+{
+ machine_restart(0);
+}
+
+#if defined(__i386__)
+
+void new_thread(struct domain *p,
+ unsigned long start_pc,
+ unsigned long start_stack,
+ unsigned long start_info)
+{
+ execution_context_t *ec = &p->shared_info->execution_context;
+
+ /*
+ * Initial register values:
+ * DS,ES,FS,GS = FLAT_RING1_DS
+ * CS:EIP = FLAT_RING1_CS:start_pc
+ * SS:ESP = FLAT_RING1_DS:start_stack
+ * ESI = start_info
+ * [EAX,EBX,ECX,EDX,EDI,EBP are zero]
+ */
+ ec->ds = ec->es = ec->fs = ec->gs = ec->ss = FLAT_RING1_DS;
+ ec->cs = FLAT_RING1_CS;
+ ec->eip = start_pc;
+ ec->esp = start_stack;
+ ec->esi = start_info;
+
+ __save_flags(ec->eflags);
+ ec->eflags |= X86_EFLAGS_IF;
+
+ /* No fast trap at start of day. */
+ SET_DEFAULT_FAST_TRAP(&p->thread);
+}
+
+
+/*
+ * This special macro can be used to load a debugging register
+ */
+#define loaddebug(thread,register) \
+ __asm__("movl %0,%%db" #register \
+ : /* no output */ \
+ :"r" (thread->debugreg[register]))
+
+
+void switch_to(struct domain *prev_p, struct domain *next_p)
+{
+ struct thread_struct *next = &next_p->thread;
+ struct tss_struct *tss = init_tss + smp_processor_id();
+ execution_context_t *stack_ec = get_execution_context();
+ int i;
+
+ __cli();
+
+ /* Switch guest general-register state. */
+ if ( !is_idle_task(prev_p) )
+ {
+ memcpy(&prev_p->shared_info->execution_context,
+ stack_ec,
+ sizeof(*stack_ec));
+ unlazy_fpu(prev_p);
+ CLEAR_FAST_TRAP(&prev_p->thread);
+ }
+
+ if ( !is_idle_task(next_p) )
+ {
+ memcpy(stack_ec,
+ &next_p->shared_info->execution_context,
+ sizeof(*stack_ec));
+
+ /*
+ * This is sufficient! If the descriptor DPL differs from CS RPL then
+ * we'll #GP. If DS, ES, FS, GS are DPL 0 then they'll be cleared
+ * automatically. If SS RPL or DPL differs from CS RPL then we'll #GP.
+ */
+ if ( (stack_ec->cs & 3) == 0 )
+ stack_ec->cs = FLAT_RING1_CS;
+ if ( (stack_ec->ss & 3) == 0 )
+ stack_ec->ss = FLAT_RING1_DS;
+
+ SET_FAST_TRAP(&next_p->thread);
+
+ /* Switch the guest OS ring-1 stack. */
+ tss->esp1 = next->guestos_sp;
+ tss->ss1 = next->guestos_ss;
+
+ /* Maybe switch the debug registers. */
+ if ( unlikely(next->debugreg[7]) )
+ {
+ loaddebug(next, 0);
+ loaddebug(next, 1);
+ loaddebug(next, 2);
+ loaddebug(next, 3);
+ /* no 4 and 5 */
+ loaddebug(next, 6);
+ loaddebug(next, 7);
+ }
+
+ /* Switch page tables. */
+ write_ptbase(&next_p->mm);
+ tlb_clocktick();
+ }
+
+ if ( unlikely(prev_p->io_bitmap != NULL) ||
+ unlikely(next_p->io_bitmap != NULL) )
+ {
+ if ( next_p->io_bitmap != NULL )
+ {
+ /* Copy in the appropriate parts of the IO bitmap. We use the
+ * selector to copy only the interesting parts of the bitmap. */
+
+ u64 old_sel = ~0ULL; /* IO bitmap selector for previous task. */
+
+ if ( prev_p->io_bitmap != NULL)
+ {
+ old_sel = prev_p->io_bitmap_sel;
+
+ /* Replace any areas of the IO bitmap that had bits cleared. */
+ for ( i = 0; i < sizeof(prev_p->io_bitmap_sel) * 8; i++ )
+ if ( !test_bit(i, &prev_p->io_bitmap_sel) )
+ memcpy(&tss->io_bitmap[i * IOBMP_SELBIT_LWORDS],
+ &next_p->io_bitmap[i * IOBMP_SELBIT_LWORDS],
+ IOBMP_SELBIT_LWORDS * sizeof(unsigned long));
+ }
+
+ /* Copy in any regions of the new task's bitmap that have bits
+ * clear and we haven't already dealt with. */
+ for ( i = 0; i < sizeof(prev_p->io_bitmap_sel) * 8; i++ )
+ {
+ if ( test_bit(i, &old_sel)
+ && !test_bit(i, &next_p->io_bitmap_sel) )
+ memcpy(&tss->io_bitmap[i * IOBMP_SELBIT_LWORDS],
+ &next_p->io_bitmap[i * IOBMP_SELBIT_LWORDS],
+ IOBMP_SELBIT_LWORDS * sizeof(unsigned long));
+ }
+
+ tss->bitmap = IO_BITMAP_OFFSET;
+
+ }
+ else
+ {
+ /* In this case, we're switching FROM a task with IO port access,
+ * to a task that doesn't use the IO bitmap. We set any TSS bits
+ * that might have been cleared, ready for future use. */
+ for ( i = 0; i < sizeof(prev_p->io_bitmap_sel) * 8; i++ )
+ if ( !test_bit(i, &prev_p->io_bitmap_sel) )
+ memset(&tss->io_bitmap[i * IOBMP_SELBIT_LWORDS],
+ 0xFF, IOBMP_SELBIT_LWORDS * sizeof(unsigned long));
+
+ /*
+ * a bitmap offset pointing outside of the TSS limit
+ * causes a nicely controllable SIGSEGV if a process
+ * tries to use a port IO instruction. The first
+ * sys_ioperm() call sets up the bitmap properly.
+ */
+ tss->bitmap = INVALID_IO_BITMAP_OFFSET;
+ }
+ }
+
+ set_current(next_p);
+
+ /* Switch GDT and LDT. */
+ __asm__ __volatile__ ("lgdt %0" : "=m" (*next_p->mm.gdt));
+ load_LDT(next_p);
+
+ __sti();
+}
+
+
+/* XXX Currently the 'domain' field is ignored! XXX */
+long do_iopl(domid_t domain, unsigned int new_io_pl)
+{
+ execution_context_t *ec = get_execution_context();
+ ec->eflags = (ec->eflags & 0xffffcfff) | ((new_io_pl&3) << 12);
+ return 0;
+}
+
+#endif
+++ /dev/null
-/******************************************************************************
- * domain_page.h
- *
- * Allow temporary mapping of domain pages. Based on ideas from the
- * Linux PKMAP code -- the copyrights and credits are retained below.
- */
-
-/*
- * (C) 1999 Andrea Arcangeli, SuSE GmbH, andrea@suse.de
- * Gerhard Wichert, Siemens AG, Gerhard.Wichert@pdb.siemens.de *
- * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
- */
-
-#include <xen/config.h>
-#include <xen/sched.h>
-#include <xen/mm.h>
-#include <xen/perfc.h>
-#include <asm/domain_page.h>
-#include <asm/flushtlb.h>
-
-unsigned long *mapcache;
-static unsigned int map_idx, shadow_map_idx[NR_CPUS];
-static spinlock_t map_lock = SPIN_LOCK_UNLOCKED;
-
-/* Use a spare PTE bit to mark entries ready for recycling. */
-#define READY_FOR_TLB_FLUSH (1<<10)
-
-static void flush_all_ready_maps(void)
-{
- unsigned long *cache = mapcache;
-
- /* A bit skanky -- depends on having an aligned PAGE_SIZE set of PTEs. */
- do { if ( (*cache & READY_FOR_TLB_FLUSH) ) *cache = 0; }
- while ( ((unsigned long)(++cache) & ~PAGE_MASK) != 0 );
-
- perfc_incrc(domain_page_tlb_flush);
- local_flush_tlb();
-}
-
-
-void *map_domain_mem(unsigned long pa)
-{
- unsigned long va;
- unsigned int idx, cpu = smp_processor_id();
- unsigned long *cache = mapcache;
- unsigned long flags;
-
- perfc_incrc(map_domain_mem_count);
-
- spin_lock_irqsave(&map_lock, flags);
-
- /* Has some other CPU caused a wrap? We must flush if so. */
- if ( map_idx < shadow_map_idx[cpu] )
- {
- perfc_incrc(domain_page_tlb_flush);
- local_flush_tlb();
- }
-
- for ( ; ; )
- {
- idx = map_idx = (map_idx + 1) & (MAPCACHE_ENTRIES - 1);
- if ( idx == 0 ) flush_all_ready_maps();
- if ( cache[idx] == 0 ) break;
- }
-
- cache[idx] = (pa & PAGE_MASK) | __PAGE_HYPERVISOR;
-
- spin_unlock_irqrestore(&map_lock, flags);
-
- shadow_map_idx[cpu] = idx;
-
- va = MAPCACHE_VIRT_START + (idx << PAGE_SHIFT) + (pa & ~PAGE_MASK);
- return (void *)va;
-}
-
-void unmap_domain_mem(void *va)
-{
- unsigned int idx;
- idx = ((unsigned long)va - MAPCACHE_VIRT_START) >> PAGE_SHIFT;
- mapcache[idx] |= READY_FOR_TLB_FLUSH;
-}
+++ /dev/null
-/*
- * Hypercall and fault low-level handling routines.
- *
- * Copyright (c) 2002-2004, K A Fraser
- * Copyright (c) 1991, 1992 Linus Torvalds
- */
-
-/*
- * The idea for callbacks to guest OSes
- * ====================================
- *
- * First, we require that all callbacks (either via a supplied
- * interrupt-descriptor-table, or via the special event or failsafe callbacks
- * in the shared-info-structure) are to ring 1. This just makes life easier,
- * in that it means we don't have to do messy GDT/LDT lookups to find
- * out which the privilege-level of the return code-selector. That code
- * would just be a hassle to write, and would need to account for running
- * off the end of the GDT/LDT, for example. For all callbacks we check
- * that the provided
- * return CS is not == __HYPERVISOR_{CS,DS}. Apart from that we're safe as
- * don't allow a guest OS to install ring-0 privileges into the GDT/LDT.
- * It's up to the guest OS to ensure all returns via the IDT are to ring 1.
- * If not, we load incorrect SS/ESP values from the TSS (for ring 1 rather
- * than the correct ring) and bad things are bound to ensue -- IRET is
- * likely to fault, and we may end up killing the domain (no harm can
- * come to Xen, though).
- *
- * When doing a callback, we check if the return CS is in ring 0. If so,
- * callback is delayed until next return to ring != 0.
- * If return CS is in ring 1, then we create a callback frame
- * starting at return SS/ESP. The base of the frame does an intra-privilege
- * interrupt-return.
- * If return CS is in ring > 1, we create a callback frame starting
- * at SS/ESP taken from appropriate section of the current TSS. The base
- * of the frame does an inter-privilege interrupt-return.
- *
- * Note that the "failsafe callback" uses a special stackframe:
- * { return_DS, return_ES, return_FS, return_GS, return_EIP,
- * return_CS, return_EFLAGS[, return_ESP, return_SS] }
- * That is, original values for DS/ES/FS/GS are placed on stack rather than
- * in DS/ES/FS/GS themselves. Why? It saves us loading them, only to have them
- * saved/restored in guest OS. Furthermore, if we load them we may cause
- * a fault if they are invalid, which is a hassle to deal with. We avoid
- * that problem if we don't load them :-) This property allows us to use
- * the failsafe callback as a fallback: if we ever fault on loading DS/ES/FS/GS
- * on return to ring != 0, we can simply package it up as a return via
- * the failsafe callback, and let the guest OS sort it out (perhaps by
- * killing an application process). Note that we also do this for any
- * faulting IRET -- just let the guest OS handle it via the event
- * callback.
- *
- * We terminate a domain in the following cases:
- * - creating a callback stack frame (due to bad ring-1 stack).
- * - faulting IRET on entry to failsafe callback handler.
- * So, each domain must keep its ring-1 %ss/%esp and failsafe callback
- * handler in good order (absolutely no faults allowed!).
- */
-
-#include <xen/config.h>
-#include <xen/errno.h>
-#include <hypervisor-ifs/hypervisor-if.h>
-
-EBX = 0x00
-ECX = 0x04
-EDX = 0x08
-ESI = 0x0C
-EDI = 0x10
-EBP = 0x14
-EAX = 0x18
-DS = 0x1C
-ES = 0x20
-FS = 0x24
-GS = 0x28
-ORIG_EAX = 0x2C
-EIP = 0x30
-CS = 0x34
-EFLAGS = 0x38
-OLDESP = 0x3C
-OLDSS = 0x40
-
-/* Offsets in domain structure */
-PROCESSOR = 0
-SHARED_INFO = 4
-EVENT_SEL = 8
-EVENT_ADDR = 12
-FAILSAFE_BUFFER = 16
-FAILSAFE_SEL = 32
-FAILSAFE_ADDR = 36
-
-/* Offsets in shared_info_t */
-#define UPCALL_PENDING /* 0 */
-#define UPCALL_MASK 1
-
-/* Offsets in guest_trap_bounce */
-GTB_ERROR_CODE = 0
-GTB_CR2 = 4
-GTB_FLAGS = 8
-GTB_CS = 10
-GTB_EIP = 12
-GTBF_TRAP = 1
-GTBF_TRAP_NOCODE = 2
-GTBF_TRAP_CR2 = 4
-
-CF_MASK = 0x00000001
-IF_MASK = 0x00000200
-NT_MASK = 0x00004000
-
-#define SAVE_ALL_NOSEGREGS \
- cld; \
- pushl %gs; \
- pushl %fs; \
- pushl %es; \
- pushl %ds; \
- pushl %eax; \
- pushl %ebp; \
- pushl %edi; \
- pushl %esi; \
- pushl %edx; \
- pushl %ecx; \
- pushl %ebx; \
-
-#define SAVE_ALL \
- SAVE_ALL_NOSEGREGS \
- movl $(__HYPERVISOR_DS),%edx; \
- movl %edx,%ds; \
- movl %edx,%es; \
- movl %edx,%fs; \
- movl %edx,%gs; \
- sti;
-
-#define GET_CURRENT(reg) \
- movl $4096-4, reg; \
- orl %esp, reg; \
- andl $~3,reg; \
- movl (reg),reg;
-
-ENTRY(continue_nonidle_task)
- GET_CURRENT(%ebx)
- jmp test_all_events
-
- ALIGN
-/*
- * HYPERVISOR_multicall(call_list, nr_calls)
- * Execute a list of 'nr_calls' hypercalls, pointed at by 'call_list'.
- * This is fairly easy except that:
- * 1. We may fault reading the call list, and must patch that up; and
- * 2. We cannot recursively call HYPERVISOR_multicall, or a malicious
- * caller could cause our stack to blow up.
- */
-do_multicall:
- popl %eax
- cmpl $SYMBOL_NAME(multicall_return_from_call),%eax
- je multicall_return_from_call
- pushl %ebx
- movl 4(%esp),%ebx /* EBX == call_list */
- movl 8(%esp),%ecx /* ECX == nr_calls */
-multicall_loop:
- pushl %ecx
-multicall_fault1:
- pushl 20(%ebx) # args[4]
-multicall_fault2:
- pushl 16(%ebx) # args[3]
-multicall_fault3:
- pushl 12(%ebx) # args[2]
-multicall_fault4:
- pushl 8(%ebx) # args[1]
-multicall_fault5:
- pushl 4(%ebx) # args[0]
-multicall_fault6:
- movl (%ebx),%eax # op
- andl $(NR_hypercalls-1),%eax
- call *SYMBOL_NAME(hypercall_table)(,%eax,4)
-multicall_return_from_call:
-multicall_fault7:
- movl %eax,24(%ebx) # args[5] == result
- addl $20,%esp
- popl %ecx
- addl $(ARGS_PER_MULTICALL_ENTRY*4),%ebx
- loop multicall_loop
- popl %ebx
- xorl %eax,%eax
- jmp ret_from_hypercall
-
-.section __ex_table,"a"
- .align 4
- .long multicall_fault1, multicall_fixup1
- .long multicall_fault2, multicall_fixup2
- .long multicall_fault3, multicall_fixup3
- .long multicall_fault4, multicall_fixup4
- .long multicall_fault5, multicall_fixup5
- .long multicall_fault6, multicall_fixup6
-.previous
-
-.section .fixup,"ax"
-multicall_fixup6:
- addl $4,%esp
-multicall_fixup5:
- addl $4,%esp
-multicall_fixup4:
- addl $4,%esp
-multicall_fixup3:
- addl $4,%esp
-multicall_fixup2:
- addl $4,%esp
-multicall_fixup1:
- addl $4,%esp
- popl %ebx
- movl $-EFAULT,%eax
- jmp ret_from_hypercall
-.previous
-
- ALIGN
-restore_all_guest:
- # First, may need to restore %ds if clobbered by create_bounce_frame
- pushl %ss
- popl %ds
- # Second, create a failsafe copy of DS,ES,FS,GS in case any are bad
- leal DS(%esp),%esi
- leal FAILSAFE_BUFFER(%ebx),%edi
- movsl
- movsl
- movsl
- movsl
- # Finally, restore guest registers -- faults will cause failsafe
- popl %ebx
- popl %ecx
- popl %edx
- popl %esi
- popl %edi
- popl %ebp
- popl %eax
-1: popl %ds
-2: popl %es
-3: popl %fs
-4: popl %gs
- addl $4,%esp
-5: iret
-.section .fixup,"ax"
-10: subl $4,%esp
- pushl %gs
-9: pushl %fs
-8: pushl %es
-7: pushl %ds
-6: pushl %eax
- pushl %ebp
- pushl %edi
- pushl %esi
- pushl %edx
- pushl %ecx
- pushl %ebx
- pushl %ss
- popl %ds
- pushl %ss
- popl %es
- jmp failsafe_callback
-.previous
-.section __ex_table,"a"
- .align 4
- .long 1b,6b
- .long 2b,7b
- .long 3b,8b
- .long 4b,9b
- .long 5b,10b
-.previous
-
-/* No special register assumptions */
-failsafe_callback:
- GET_CURRENT(%ebx)
- movl PROCESSOR(%ebx),%eax
- shl $4,%eax
- lea guest_trap_bounce(%eax),%edx
- movl FAILSAFE_ADDR(%ebx),%eax
- movl %eax,GTB_EIP(%edx)
- movl FAILSAFE_SEL(%ebx),%eax
- movw %ax,GTB_CS(%edx)
- call create_bounce_frame
- subl $16,%esi # add DS/ES/FS/GS to failsafe stack frame
- leal FAILSAFE_BUFFER(%ebx),%ebp
- movl 0(%ebp),%eax # DS
-FAULT1: movl %eax,(%esi)
- movl 4(%ebp),%eax # ES
-FAULT2: movl %eax,4(%esi)
- movl 8(%ebp),%eax # FS
-FAULT3: movl %eax,8(%esi)
- movl 12(%ebp),%eax # GS
-FAULT4: movl %eax,12(%esi)
- movl %esi,OLDESP(%esp)
- popl %ebx
- popl %ecx
- popl %edx
- popl %esi
- popl %edi
- popl %ebp
- popl %eax
- addl $20,%esp # skip DS/ES/FS/GS/ORIG_EAX
-FAULT5: iret
-
-
- ALIGN
-# Simple restore -- we should never fault as we we will only interrupt ring 0
-# when sane values have been placed in all registers. The only exception is
-# NMI, which may interrupt before good values have been placed in DS-GS.
-# The NMI return code deals with this problem itself.
-restore_all_xen:
- popl %ebx
- popl %ecx
- popl %edx
- popl %esi
- popl %edi
- popl %ebp
- popl %eax
- popl %ds
- popl %es
- popl %fs
- popl %gs
- addl $4,%esp
- iret
-
- ALIGN
-ENTRY(hypercall)
- pushl %eax # save orig_eax
- SAVE_ALL
- GET_CURRENT(%ebx)
- andl $(NR_hypercalls-1),%eax
- call *SYMBOL_NAME(hypercall_table)(,%eax,4)
-
-ret_from_hypercall:
- movl %eax,EAX(%esp) # save the return value
-
-test_all_events:
- xorl %ecx,%ecx
- notl %ecx
- cli # tests must not race interrupts
-/*test_softirqs:*/
- movl PROCESSOR(%ebx),%eax
- shl $6,%eax # sizeof(irq_cpustat) == 64
- test %ecx,SYMBOL_NAME(irq_stat)(%eax,1)
- jnz process_softirqs
-/*test_guest_events:*/
- movl SHARED_INFO(%ebx),%eax
- testb $0xFF,UPCALL_MASK(%eax)
- jnz restore_all_guest
- testb $0xFF,UPCALL_PENDING(%eax)
- jz restore_all_guest
- movb $1,UPCALL_MASK(%eax) # Upcalls are masked during delivery
-/*process_guest_events:*/
- movl PROCESSOR(%ebx),%edx
- shl $4,%edx # sizeof(guest_trap_bounce) == 16
- lea guest_trap_bounce(%edx),%edx
- movl EVENT_ADDR(%ebx),%eax
- movl %eax,GTB_EIP(%edx)
- movl EVENT_SEL(%ebx),%eax
- movw %ax,GTB_CS(%edx)
- call create_bounce_frame
- jmp restore_all_guest
-
- ALIGN
-process_softirqs:
- sti
- call SYMBOL_NAME(do_softirq)
- jmp test_all_events
-
-/* CREATE A BASIC EXCEPTION FRAME ON GUEST OS (RING-1) STACK: */
-/* {EIP, CS, EFLAGS, [ESP, SS]} */
-/* %edx == guest_trap_bounce, %ebx == task_struct */
-/* %eax,%ecx are clobbered. %ds:%esi contain new OLDSS/OLDESP. */
-create_bounce_frame:
- mov CS+4(%esp),%cl
- test $2,%cl
- jz 1f /* jump if returning to an existing ring-1 activation */
- /* obtain ss/esp from TSS -- no current ring-1 activations */
- movl PROCESSOR(%ebx),%eax
- /* next 4 lines multiply %eax by 8320, which is sizeof(tss_struct) */
- movl %eax, %ecx
- shll $7, %ecx
- shll $13, %eax
- addl %ecx,%eax
- addl $init_tss + 12,%eax
- movl (%eax),%esi /* tss->esp1 */
-FAULT6: movl 4(%eax),%ds /* tss->ss1 */
- /* base of stack frame must contain ss/esp (inter-priv iret) */
- subl $8,%esi
- movl OLDESP+4(%esp),%eax
-FAULT7: movl %eax,(%esi)
- movl OLDSS+4(%esp),%eax
-FAULT8: movl %eax,4(%esi)
- jmp 2f
-1: /* obtain ss/esp from oldss/oldesp -- a ring-1 activation exists */
- movl OLDESP+4(%esp),%esi
-FAULT9: movl OLDSS+4(%esp),%ds
-2: /* Construct a stack frame: EFLAGS, CS/EIP */
- subl $12,%esi
- movl EIP+4(%esp),%eax
-FAULT10:movl %eax,(%esi)
- movl CS+4(%esp),%eax
-FAULT11:movl %eax,4(%esi)
- movl EFLAGS+4(%esp),%eax
-FAULT12:movl %eax,8(%esi)
- /* Rewrite our stack frame and return to ring 1. */
- /* IA32 Ref. Vol. 3: TF, VM, RF and NT flags are cleared on trap. */
- andl $0xfffcbeff,%eax
- movl %eax,EFLAGS+4(%esp)
- movl %ds,OLDSS+4(%esp)
- movl %esi,OLDESP+4(%esp)
- movzwl %es:GTB_CS(%edx),%eax
- movl %eax,CS+4(%esp)
- movl %es:GTB_EIP(%edx),%eax
- movl %eax,EIP+4(%esp)
- ret
-
-
-.section __ex_table,"a"
- .align 4
- .long FAULT1, crash_domain_fixup3 # Fault writing to ring-1 stack
- .long FAULT2, crash_domain_fixup3 # Fault writing to ring-1 stack
- .long FAULT3, crash_domain_fixup3 # Fault writing to ring-1 stack
- .long FAULT4, crash_domain_fixup3 # Fault writing to ring-1 stack
- .long FAULT5, crash_domain_fixup1 # Fault executing failsafe iret
- .long FAULT6, crash_domain_fixup2 # Fault loading ring-1 stack selector
- .long FAULT7, crash_domain_fixup2 # Fault writing to ring-1 stack
- .long FAULT8, crash_domain_fixup2 # Fault writing to ring-1 stack
- .long FAULT9, crash_domain_fixup2 # Fault loading ring-1 stack selector
- .long FAULT10,crash_domain_fixup2 # Fault writing to ring-1 stack
- .long FAULT11,crash_domain_fixup2 # Fault writing to ring-1 stack
- .long FAULT12,crash_domain_fixup2 # Fault writing to ring-1 stack
- .long FAULT13,crash_domain_fixup3 # Fault writing to ring-1 stack
- .long FAULT14,crash_domain_fixup3 # Fault writing to ring-1 stack
-.previous
-
-# This handler kills domains which experience unrecoverable faults.
-.section .fixup,"ax"
-crash_domain_fixup1:
- subl $4,%esp
- SAVE_ALL
- jmp domain_crash
-crash_domain_fixup2:
- addl $4,%esp
-crash_domain_fixup3:
- pushl %ss
- popl %ds
- jmp domain_crash
-.previous
-
- ALIGN
-process_guest_exception_and_events:
- movl PROCESSOR(%ebx),%eax
- shl $4,%eax
- lea guest_trap_bounce(%eax),%edx
- testb $~0,GTB_FLAGS(%edx)
- jz test_all_events
- call create_bounce_frame # just the basic frame
- mov %es:GTB_FLAGS(%edx),%cl
- test $GTBF_TRAP_NOCODE,%cl
- jnz 2f
- subl $4,%esi # push error_code onto guest frame
- movl %es:GTB_ERROR_CODE(%edx),%eax
-FAULT13:movl %eax,(%esi)
- test $GTBF_TRAP_CR2,%cl
- jz 1f
- subl $4,%esi # push %cr2 onto guest frame
- movl %es:GTB_CR2(%edx),%eax
-FAULT14:movl %eax,(%esi)
-1: movl %esi,OLDESP(%esp)
-2: push %es # unclobber %ds
- pop %ds
- movb $0,GTB_FLAGS(%edx)
- jmp test_all_events
-
- ALIGN
-ENTRY(ret_from_intr)
- GET_CURRENT(%ebx)
- movb CS(%esp),%al
- testb $3,%al # return to non-supervisor?
- jne test_all_events
- jmp restore_all_xen
-
-ENTRY(divide_error)
- pushl $0 # no error code
- pushl $ SYMBOL_NAME(do_divide_error)
- ALIGN
-error_code:
- pushl %fs
- pushl %es
- pushl %ds
- pushl %eax
- xorl %eax,%eax
- pushl %ebp
- pushl %edi
- pushl %esi
- pushl %edx
- decl %eax # eax = -1
- pushl %ecx
- pushl %ebx
- cld
- movl %gs,%ecx
- movl ORIG_EAX(%esp), %esi # get the error code
- movl GS(%esp), %edi # get the function address
- movl %eax, ORIG_EAX(%esp)
- movl %ecx, GS(%esp)
- movl $(__HYPERVISOR_DS),%edx
- movl %edx,%ds
- movl %edx,%es
- movl %edx,%fs
- movl %edx,%gs
- movl %esp,%edx
- pushl %esi # push the error code
- pushl %edx # push the pt_regs pointer
- GET_CURRENT(%ebx)
- call *%edi
- addl $8,%esp
- movb CS(%esp),%al
- testb $3,%al
- je restore_all_xen
- jmp process_guest_exception_and_events
-
-ENTRY(coprocessor_error)
- pushl $0
- pushl $ SYMBOL_NAME(do_coprocessor_error)
- jmp error_code
-
-ENTRY(simd_coprocessor_error)
- pushl $0
- pushl $ SYMBOL_NAME(do_simd_coprocessor_error)
- jmp error_code
-
-ENTRY(device_not_available)
- pushl $0
- pushl $SYMBOL_NAME(math_state_restore)
- jmp error_code
-
-ENTRY(debug)
- pushl $0
- pushl $ SYMBOL_NAME(do_debug)
- jmp error_code
-
-ENTRY(int3)
- pushl $0
- pushl $ SYMBOL_NAME(do_int3)
- jmp error_code
-
-ENTRY(overflow)
- pushl $0
- pushl $ SYMBOL_NAME(do_overflow)
- jmp error_code
-
-ENTRY(bounds)
- pushl $0
- pushl $ SYMBOL_NAME(do_bounds)
- jmp error_code
-
-ENTRY(invalid_op)
- pushl $0
- pushl $ SYMBOL_NAME(do_invalid_op)
- jmp error_code
-
-ENTRY(coprocessor_segment_overrun)
- pushl $0
- pushl $ SYMBOL_NAME(do_coprocessor_segment_overrun)
- jmp error_code
-
-ENTRY(invalid_TSS)
- pushl $ SYMBOL_NAME(do_invalid_TSS)
- jmp error_code
-
-ENTRY(segment_not_present)
- pushl $ SYMBOL_NAME(do_segment_not_present)
- jmp error_code
-
-ENTRY(stack_segment)
- pushl $ SYMBOL_NAME(do_stack_segment)
- jmp error_code
-
-ENTRY(general_protection)
- pushl $ SYMBOL_NAME(do_general_protection)
- jmp error_code
-
-ENTRY(alignment_check)
- pushl $ SYMBOL_NAME(do_alignment_check)
- jmp error_code
-
-ENTRY(page_fault)
- pushl $ SYMBOL_NAME(do_page_fault)
- jmp error_code
-
-ENTRY(machine_check)
- pushl $0
- pushl $ SYMBOL_NAME(do_machine_check)
- jmp error_code
-
-ENTRY(spurious_interrupt_bug)
- pushl $0
- pushl $ SYMBOL_NAME(do_spurious_interrupt_bug)
- jmp error_code
-
-ENTRY(nmi)
- # Save state but do not trash the segment registers!
- # We may otherwise be unable to reload them or copy them to ring 1.
- pushl %eax
- SAVE_ALL_NOSEGREGS
-
- # Check for hardware problems. These are always fatal so we can
- # reload DS and ES when handling them.
- inb $0x61,%al
- testb $0x80,%al
- jne nmi_parity_err
- testb $0x40,%al
- jne nmi_io_err
- movl %eax,%ebx
-
- # Okay, its almost a normal NMI tick. We can only process it if:
- # A. We are the outermost Xen activation (in which case we have
- # the selectors safely saved on our stack)
- # B. DS-GS all contain sane Xen values.
- # In all other cases we bail without touching DS-GS, as we have
- # interrupted an enclosing Xen activation in tricky prologue or
- # epilogue code.
- movb CS(%esp),%al
- testb $3,%al
- jne do_watchdog_tick
- movl DS(%esp),%eax
- cmpw $(__HYPERVISOR_DS),%ax
- jne nmi_badseg
- movl ES(%esp),%eax
- cmpw $(__HYPERVISOR_DS),%ax
- jne nmi_badseg
- movl FS(%esp),%eax
- cmpw $(__HYPERVISOR_DS),%ax
- jne nmi_badseg
- movl GS(%esp),%eax
- cmpw $(__HYPERVISOR_DS),%ax
- jne nmi_badseg
-
-do_watchdog_tick:
- movl $(__HYPERVISOR_DS),%edx
- movl %edx,%ds
- movl %edx,%es
- movl %esp,%edx
- pushl %ebx # reason
- pushl %edx # regs
- call SYMBOL_NAME(do_nmi)
- addl $8,%esp
- movb CS(%esp),%al
- testb $3,%al
- je restore_all_xen
- GET_CURRENT(%ebx)
- jmp restore_all_guest
-
-nmi_badseg:
- popl %ebx
- popl %ecx
- popl %edx
- popl %esi
- popl %edi
- popl %ebp
- popl %eax
- addl $20,%esp
- iret
-
-nmi_parity_err:
- movl $(__HYPERVISOR_DS),%edx
- movl %edx,%ds
- movl %edx,%es
- jmp SYMBOL_NAME(mem_parity_error)
-
-nmi_io_err:
- movl $(__HYPERVISOR_DS),%edx
- movl %edx,%ds
- movl %edx,%es
- jmp SYMBOL_NAME(io_check_error)
-
-.data
-ENTRY(hypercall_table)
- .long SYMBOL_NAME(do_set_trap_table) /* 0 */
- .long SYMBOL_NAME(do_mmu_update)
- .long SYMBOL_NAME(do_set_gdt)
- .long SYMBOL_NAME(do_stack_switch)
- .long SYMBOL_NAME(do_set_callbacks)
- .long SYMBOL_NAME(do_fpu_taskswitch) /* 5 */
- .long SYMBOL_NAME(do_sched_op)
- .long SYMBOL_NAME(do_dom0_op)
- .long SYMBOL_NAME(do_set_debugreg)
- .long SYMBOL_NAME(do_get_debugreg)
- .long SYMBOL_NAME(do_update_descriptor) /* 10 */
- .long SYMBOL_NAME(do_set_fast_trap)
- .long SYMBOL_NAME(do_dom_mem_op)
- .long SYMBOL_NAME(do_multicall)
- .long SYMBOL_NAME(do_update_va_mapping)
- .long SYMBOL_NAME(do_set_timer_op) /* 15 */
- .long SYMBOL_NAME(do_event_channel_op)
- .long SYMBOL_NAME(do_xen_version)
- .long SYMBOL_NAME(do_console_io)
- .long SYMBOL_NAME(do_physdev_op)
- .long SYMBOL_NAME(do_update_va_mapping_otherdomain) /* 20 */
- .rept NR_hypercalls-((.-hypercall_table)/4)
- .long SYMBOL_NAME(do_ni_hypercall)
- .endr
asmlinkage void do_IRQ(struct pt_regs regs)
{
+#if defined(__i386__)
unsigned int irq = regs.orig_eax;
+#else
+ unsigned int irq = 0; /* XXX */
+#endif
irq_desc_t *desc = &irq_desc[irq];
struct irqaction *action;
+++ /dev/null
-/******************************************************************************
- * arch/i386/mm.c
- *
- * Modifications to Linux original are copyright (c) 2002-2003, K A Fraser
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
- */
-
-#include <xen/config.h>
-#include <xen/lib.h>
-#include <xen/init.h>
-#include <xen/mm.h>
-#include <asm/page.h>
-#include <asm/flushtlb.h>
-#include <asm/fixmap.h>
-#include <asm/domain_page.h>
-
-static inline void set_pte_phys(unsigned long vaddr,
- l1_pgentry_t entry)
-{
- l2_pgentry_t *l2ent;
- l1_pgentry_t *l1ent;
-
- l2ent = &idle_pg_table[l2_table_offset(vaddr)];
- l1ent = l2_pgentry_to_l1(*l2ent) + l1_table_offset(vaddr);
- *l1ent = entry;
-
- /* It's enough to flush this one mapping. */
- __flush_tlb_one(vaddr);
-}
-
-
-void __set_fixmap(enum fixed_addresses idx,
- l1_pgentry_t entry)
-{
- unsigned long address = fix_to_virt(idx);
-
- if ( likely(idx < __end_of_fixed_addresses) )
- set_pte_phys(address, entry);
- else
- printk("Invalid __set_fixmap\n");
-}
-
-
-static void __init fixrange_init(unsigned long start,
- unsigned long end,
- l2_pgentry_t *pg_base)
-{
- l2_pgentry_t *l2e;
- int i;
- unsigned long vaddr, page;
-
- vaddr = start;
- i = l2_table_offset(vaddr);
- l2e = pg_base + i;
-
- for ( ; (i < ENTRIES_PER_L2_PAGETABLE) && (vaddr != end); l2e++, i++ )
- {
- if ( !l2_pgentry_empty(*l2e) )
- continue;
- page = (unsigned long)get_free_page();
- clear_page(page);
- *l2e = mk_l2_pgentry(__pa(page) | __PAGE_HYPERVISOR);
- vaddr += 1 << L2_PAGETABLE_SHIFT;
- }
-}
-
-void __init paging_init(void)
-{
- unsigned long addr;
- void *ioremap_pt;
- int i;
-
- /* Idle page table 1:1 maps the first part of physical memory. */
- for ( i = 0; i < DOMAIN_ENTRIES_PER_L2_PAGETABLE; i++ )
- idle_pg_table[i] =
- mk_l2_pgentry((i << L2_PAGETABLE_SHIFT) |
- __PAGE_HYPERVISOR | _PAGE_PSE);
-
- /*
- * Fixed mappings, only the page table structure has to be
- * created - mappings will be set by set_fixmap():
- */
- addr = FIXADDR_START & ~((1<<L2_PAGETABLE_SHIFT)-1);
- fixrange_init(addr, 0, idle_pg_table);
-
- /* Create page table for ioremap(). */
- ioremap_pt = (void *)get_free_page();
- clear_page(ioremap_pt);
- idle_pg_table[IOREMAP_VIRT_START >> L2_PAGETABLE_SHIFT] =
- mk_l2_pgentry(__pa(ioremap_pt) | __PAGE_HYPERVISOR);
-
- /* Create read-only mapping of MPT for guest-OS use. */
- idle_pg_table[RO_MPT_VIRT_START >> L2_PAGETABLE_SHIFT] =
- idle_pg_table[RDWR_MPT_VIRT_START >> L2_PAGETABLE_SHIFT];
- mk_l2_readonly(idle_pg_table +
- (RO_MPT_VIRT_START >> L2_PAGETABLE_SHIFT));
-
- /* Set up mapping cache for domain pages. */
- mapcache = (unsigned long *)get_free_page();
- clear_page(mapcache);
- idle_pg_table[MAPCACHE_VIRT_START >> L2_PAGETABLE_SHIFT] =
- mk_l2_pgentry(__pa(mapcache) | __PAGE_HYPERVISOR);
-
- /* Set up linear page table mapping. */
- idle_pg_table[LINEAR_PT_VIRT_START >> L2_PAGETABLE_SHIFT] =
- mk_l2_pgentry(__pa(idle_pg_table) | __PAGE_HYPERVISOR);
-
-}
-
-void __init zap_low_mappings(void)
-{
- int i;
- for ( i = 0; i < DOMAIN_ENTRIES_PER_L2_PAGETABLE; i++ )
- idle_pg_table[i] = mk_l2_pgentry(0);
- flush_tlb_all_pge();
-}
-
-
-long do_stack_switch(unsigned long ss, unsigned long esp)
-{
- int nr = smp_processor_id();
- struct tss_struct *t = &init_tss[nr];
-
- /* We need to do this check as we load and use SS on guest's behalf. */
- if ( (ss & 3) == 0 )
- return -EPERM;
-
- current->thread.guestos_ss = ss;
- current->thread.guestos_sp = esp;
- t->ss1 = ss;
- t->esp1 = esp;
-
- return 0;
-}
-
-
-/* Returns TRUE if given descriptor is valid for GDT or LDT. */
-int check_descriptor(unsigned long a, unsigned long b)
-{
- unsigned long base, limit;
-
- /* A not-present descriptor will always fault, so is safe. */
- if ( !(b & _SEGMENT_P) )
- goto good;
-
- /*
- * We don't allow a DPL of zero. There is no legitimate reason for
- * specifying DPL==0, and it gets rather dangerous if we also accept call
- * gates (consider a call gate pointing at another guestos descriptor with
- * DPL 0 -- this would get the OS ring-0 privileges).
- */
- if ( (b & _SEGMENT_DPL) == 0 )
- goto bad;
-
- if ( !(b & _SEGMENT_S) )
- {
- /*
- * System segment:
- * 1. Don't allow interrupt or trap gates as they belong in the IDT.
- * 2. Don't allow TSS descriptors or task gates as we don't
- * virtualise x86 tasks.
- * 3. Don't allow LDT descriptors because they're unnecessary and
- * I'm uneasy about allowing an LDT page to contain LDT
- * descriptors. In any case, Xen automatically creates the
- * required descriptor when reloading the LDT register.
- * 4. We allow call gates but they must not jump to a private segment.
- */
-
- /* Disallow everything but call gates. */
- if ( (b & _SEGMENT_TYPE) != 0xc00 )
- goto bad;
-
- /* Can't allow far jump to a Xen-private segment. */
- if ( !VALID_CODESEL(a>>16) )
- goto bad;
-
- /* Reserved bits must be zero. */
- if ( (b & 0xe0) != 0 )
- goto bad;
-
- /* No base/limit check is needed for a call gate. */
- goto good;
- }
-
- /* Check that base/limit do not overlap Xen-private space. */
- base = (b&(0xff<<24)) | ((b&0xff)<<16) | (a>>16);
- limit = (b&0xf0000) | (a&0xffff);
- limit++; /* We add one because limit is inclusive. */
- if ( (b & _SEGMENT_G) )
- limit <<= 12;
- if ( ((base + limit) <= base) ||
- ((base + limit) > PAGE_OFFSET) )
- goto bad;
-
- good:
- return 1;
- bad:
- return 0;
-}
-
-
-long set_gdt(struct domain *p,
- unsigned long *frames,
- unsigned int entries)
-{
- /* NB. There are 512 8-byte entries per GDT page. */
- int i, nr_pages = (entries + 511) / 512;
- unsigned long pfn;
- struct desc_struct *vgdt;
-
- /* Check the new GDT. */
- for ( i = 0; i < nr_pages; i++ )
- {
- if ( unlikely(frames[i] >= max_page) ||
- unlikely(!get_page_and_type(&frame_table[frames[i]],
- p, PGT_gdt_page)) )
- goto fail;
- }
-
- /* Copy reserved GDT entries to the new GDT. */
- vgdt = map_domain_mem(frames[0] << PAGE_SHIFT);
- memcpy(vgdt + FIRST_RESERVED_GDT_ENTRY,
- gdt_table + FIRST_RESERVED_GDT_ENTRY,
- NR_RESERVED_GDT_ENTRIES*8);
- unmap_domain_mem(vgdt);
-
- /* Tear down the old GDT. */
- for ( i = 0; i < 16; i++ )
- {
- if ( (pfn = l1_pgentry_to_pagenr(p->mm.perdomain_pt[i])) != 0 )
- put_page_and_type(&frame_table[pfn]);
- p->mm.perdomain_pt[i] = mk_l1_pgentry(0);
- }
-
- /* Install the new GDT. */
- for ( i = 0; i < nr_pages; i++ )
- p->mm.perdomain_pt[i] =
- mk_l1_pgentry((frames[i] << PAGE_SHIFT) | __PAGE_HYPERVISOR);
-
- SET_GDT_ADDRESS(p, GDT_VIRT_START);
- SET_GDT_ENTRIES(p, (entries*8)-1);
-
- return 0;
-
- fail:
- while ( i-- > 0 )
- put_page_and_type(&frame_table[frames[i]]);
- return -EINVAL;
-}
-
-
-long do_set_gdt(unsigned long *frame_list, unsigned int entries)
-{
- int nr_pages = (entries + 511) / 512;
- unsigned long frames[16];
- long ret;
-
- if ( (entries <= LAST_RESERVED_GDT_ENTRY) || (entries > 8192) )
- return -EINVAL;
-
- if ( copy_from_user(frames, frame_list, nr_pages * sizeof(unsigned long)) )
- return -EFAULT;
-
- if ( (ret = set_gdt(current, frames, entries)) == 0 )
- {
- local_flush_tlb();
- __asm__ __volatile__ ("lgdt %0" : "=m" (*current->mm.gdt));
- }
-
- return ret;
-}
-
-
-long do_update_descriptor(
- unsigned long pa, unsigned long word1, unsigned long word2)
-{
- unsigned long *gdt_pent, pfn = pa >> PAGE_SHIFT;
- struct pfn_info *page;
- long ret = -EINVAL;
-
- if ( (pa & 7) || (pfn >= max_page) || !check_descriptor(word1, word2) )
- return -EINVAL;
-
- page = &frame_table[pfn];
- if ( unlikely(!get_page(page, current)) )
- goto out;
-
- /* Check if the given frame is in use in an unsafe context. */
- switch ( page->type_and_flags & PGT_type_mask )
- {
- case PGT_gdt_page:
- /* Disallow updates of Xen-reserved descriptors in the current GDT. */
- if ( (l1_pgentry_to_pagenr(current->mm.perdomain_pt[0]) == pfn) &&
- (((pa&(PAGE_SIZE-1))>>3) >= FIRST_RESERVED_GDT_ENTRY) &&
- (((pa&(PAGE_SIZE-1))>>3) <= LAST_RESERVED_GDT_ENTRY) )
- goto out;
- if ( unlikely(!get_page_type(page, PGT_gdt_page)) )
- goto out;
- break;
- case PGT_ldt_page:
- if ( unlikely(!get_page_type(page, PGT_ldt_page)) )
- goto out;
- break;
- default:
- if ( unlikely(!get_page_type(page, PGT_writeable_page)) )
- goto out;
- break;
- }
-
- /* All is good so make the update. */
- gdt_pent = map_domain_mem(pa);
- gdt_pent[0] = word1;
- gdt_pent[1] = word2;
- unmap_domain_mem(gdt_pent);
-
- put_page_type(page);
-
- ret = 0; /* success */
-
- out:
- put_page(page);
- return ret;
-}
-
-#ifdef MEMORY_GUARD
-
-void *memguard_init(void *heap_start)
-{
- l1_pgentry_t *l1;
- int i, j;
-
- /* Round the allocation pointer up to a page boundary. */
- heap_start = (void *)(((unsigned long)heap_start + (PAGE_SIZE-1)) &
- PAGE_MASK);
-
- /* Memory guarding is incompatible with super pages. */
- for ( i = 0; i < (xenheap_phys_end >> L2_PAGETABLE_SHIFT); i++ )
- {
- l1 = (l1_pgentry_t *)heap_start;
- heap_start = (void *)((unsigned long)heap_start + PAGE_SIZE);
- for ( j = 0; j < ENTRIES_PER_L1_PAGETABLE; j++ )
- l1[j] = mk_l1_pgentry((i << L2_PAGETABLE_SHIFT) |
- (j << L1_PAGETABLE_SHIFT) |
- __PAGE_HYPERVISOR);
- idle_pg_table[i] = idle_pg_table[i + l2_table_offset(PAGE_OFFSET)] =
- mk_l2_pgentry(virt_to_phys(l1) | __PAGE_HYPERVISOR);
- }
-
- return heap_start;
-}
-
-static void __memguard_change_range(void *p, unsigned long l, int guard)
-{
- l1_pgentry_t *l1;
- l2_pgentry_t *l2;
- unsigned long _p = (unsigned long)p;
- unsigned long _l = (unsigned long)l;
-
- /* Ensure we are dealing with a page-aligned whole number of pages. */
- ASSERT((_p&PAGE_MASK) != 0);
- ASSERT((_l&PAGE_MASK) != 0);
- ASSERT((_p&~PAGE_MASK) == 0);
- ASSERT((_l&~PAGE_MASK) == 0);
-
- while ( _l != 0 )
- {
- l2 = &idle_pg_table[l2_table_offset(_p)];
- l1 = l2_pgentry_to_l1(*l2) + l1_table_offset(_p);
- if ( guard )
- *l1 = mk_l1_pgentry(l1_pgentry_val(*l1) & ~_PAGE_PRESENT);
- else
- *l1 = mk_l1_pgentry(l1_pgentry_val(*l1) | _PAGE_PRESENT);
- _p += PAGE_SIZE;
- _l -= PAGE_SIZE;
- }
-}
-
-void memguard_guard_range(void *p, unsigned long l)
-{
- __memguard_change_range(p, l, 1);
- local_flush_tlb();
-}
-
-void memguard_unguard_range(void *p, unsigned long l)
-{
- __memguard_change_range(p, l, 0);
-}
-
-int memguard_is_guarded(void *p)
-{
- l1_pgentry_t *l1;
- l2_pgentry_t *l2;
- unsigned long _p = (unsigned long)p;
- l2 = &idle_pg_table[l2_table_offset(_p)];
- l1 = l2_pgentry_to_l1(*l2) + l1_table_offset(_p);
- return !(l1_pgentry_val(*l1) & _PAGE_PRESENT);
-}
-
-#endif
if ((clustered_apic_mode == CLUSTERED_APIC_NUMAQ) && mpc->mpc_oemptr) {
/* We need to process the oem mpc tables to tell us which quad things are in ... */
mpc_record = 0;
- smp_read_mpc_oem((struct mp_config_oemtable *) mpc->mpc_oemptr, mpc->mpc_oemsize);
+ smp_read_mpc_oem((struct mp_config_oemtable *)(unsigned long)mpc->mpc_oemptr, mpc->mpc_oemsize);
mpc_record = 0;
}
* Read the physical hardware table. Anything here will
* override the defaults.
*/
- if (!smp_read_mpc((void *)mpf->mpf_physptr)) {
+ if (!smp_read_mpc((void *)(unsigned long)mpf->mpf_physptr)) {
smp_found_config = 0;
printk(KERN_ERR "BIOS bug, MP table errors detected!...\n");
printk(KERN_ERR "... disabling SMP support. (tell your hw vendor)\n");
+++ /dev/null
-/*
- * linux/arch/i386/kernel/process.c
- *
- * Copyright (C) 1995 Linus Torvalds
- *
- * Pentium III FXSR, SSE support
- * Gareth Hughes <gareth@valinux.com>, May 2000
- */
-
-/*
- * This file handles the architecture-dependent parts of process handling..
- */
-
-#define __KERNEL_SYSCALLS__
-#include <xen/config.h>
-#include <xen/lib.h>
-#include <xen/errno.h>
-#include <xen/sched.h>
-#include <xen/smp.h>
-#include <xen/delay.h>
-#include <xen/softirq.h>
-#include <asm/ptrace.h>
-#include <asm/mc146818rtc.h>
-#include <asm/system.h>
-#include <asm/io.h>
-#include <asm/processor.h>
-#include <asm/desc.h>
-#include <asm/i387.h>
-#include <asm/mpspec.h>
-#include <asm/ldt.h>
-#include <xen/irq.h>
-#include <xen/event.h>
-#include <xen/shadow.h>
-
-int hlt_counter;
-
-void disable_hlt(void)
-{
- hlt_counter++;
-}
-
-void enable_hlt(void)
-{
- hlt_counter--;
-}
-
-/*
- * We use this if we don't have any better
- * idle routine..
- */
-static void default_idle(void)
-{
- if ( hlt_counter == 0 )
- {
- __cli();
- if ( !softirq_pending(smp_processor_id()) )
- safe_halt();
- else
- __sti();
- }
-}
-
-void continue_cpu_idle_loop(void)
-{
- int cpu = smp_processor_id();
- for ( ; ; )
- {
- irq_stat[cpu].idle_timestamp = jiffies;
- while ( !softirq_pending(cpu) )
- default_idle();
- do_softirq();
- }
-}
-
-void startup_cpu_idle_loop(void)
-{
- /* Just some sanity to ensure that the scheduler is set up okay. */
- ASSERT(current->domain == IDLE_DOMAIN_ID);
- domain_unpause_by_systemcontroller(current);
- __enter_scheduler();
-
- /*
- * Declares CPU setup done to the boot processor.
- * Therefore memory barrier to ensure state is visible.
- */
- smp_mb();
- init_idle();
-
- continue_cpu_idle_loop();
-}
-
-static long no_idt[2];
-static int reboot_mode;
-int reboot_thru_bios = 0;
-
-#ifdef CONFIG_SMP
-int reboot_smp = 0;
-static int reboot_cpu = -1;
-/* shamelessly grabbed from lib/vsprintf.c for readability */
-#define is_digit(c) ((c) >= '0' && (c) <= '9')
-#endif
-
-
-static inline void kb_wait(void)
-{
- int i;
-
- for (i=0; i<0x10000; i++)
- if ((inb_p(0x64) & 0x02) == 0)
- break;
-}
-
-
-void machine_restart(char * __unused)
-{
- extern int opt_noreboot;
-#ifdef CONFIG_SMP
- int cpuid;
-#endif
-
- if ( opt_noreboot )
- {
- printk("Reboot disabled on cmdline: require manual reset\n");
- for ( ; ; ) __asm__ __volatile__ ("hlt");
- }
-
-#ifdef CONFIG_SMP
- cpuid = GET_APIC_ID(apic_read(APIC_ID));
-
- /* KAF: Need interrupts enabled for safe IPI. */
- __sti();
-
- if (reboot_smp) {
-
- /* check to see if reboot_cpu is valid
- if its not, default to the BSP */
- if ((reboot_cpu == -1) ||
- (reboot_cpu > (NR_CPUS -1)) ||
- !(phys_cpu_present_map & (1<<cpuid)))
- reboot_cpu = boot_cpu_physical_apicid;
-
- reboot_smp = 0; /* use this as a flag to only go through this once*/
- /* re-run this function on the other CPUs
- it will fall though this section since we have
- cleared reboot_smp, and do the reboot if it is the
- correct CPU, otherwise it halts. */
- if (reboot_cpu != cpuid)
- smp_call_function((void *)machine_restart , NULL, 1, 0);
- }
-
- /* if reboot_cpu is still -1, then we want a tradional reboot,
- and if we are not running on the reboot_cpu,, halt */
- if ((reboot_cpu != -1) && (cpuid != reboot_cpu)) {
- for (;;)
- __asm__ __volatile__ ("hlt");
- }
- /*
- * Stop all CPUs and turn off local APICs and the IO-APIC, so
- * other OSs see a clean IRQ state.
- */
- smp_send_stop();
- disable_IO_APIC();
-#endif
-
- if(!reboot_thru_bios) {
- /* rebooting needs to touch the page at absolute addr 0 */
- *((unsigned short *)__va(0x472)) = reboot_mode;
- for (;;) {
- int i;
- for (i=0; i<100; i++) {
- kb_wait();
- udelay(50);
- outb(0xfe,0x64); /* pulse reset low */
- udelay(50);
- }
- /* That didn't work - force a triple fault.. */
- __asm__ __volatile__("lidt %0": "=m" (no_idt));
- __asm__ __volatile__("int3");
- }
- }
-
- panic("Need to reinclude BIOS reboot code\n");
-}
-
-void machine_halt(void)
-{
- machine_restart(0);
-}
-
-void machine_power_off(void)
-{
- machine_restart(0);
-}
-
-void new_thread(struct domain *p,
- unsigned long start_pc,
- unsigned long start_stack,
- unsigned long start_info)
-{
- execution_context_t *ec = &p->shared_info->execution_context;
-
- /*
- * Initial register values:
- * DS,ES,FS,GS = FLAT_RING1_DS
- * CS:EIP = FLAT_RING1_CS:start_pc
- * SS:ESP = FLAT_RING1_DS:start_stack
- * ESI = start_info
- * [EAX,EBX,ECX,EDX,EDI,EBP are zero]
- */
- ec->ds = ec->es = ec->fs = ec->gs = ec->ss = FLAT_RING1_DS;
- ec->cs = FLAT_RING1_CS;
- ec->eip = start_pc;
- ec->esp = start_stack;
- ec->esi = start_info;
-
- __save_flags(ec->eflags);
- ec->eflags |= X86_EFLAGS_IF;
-
- /* No fast trap at start of day. */
- SET_DEFAULT_FAST_TRAP(&p->thread);
-}
-
-
-/*
- * This special macro can be used to load a debugging register
- */
-#define loaddebug(thread,register) \
- __asm__("movl %0,%%db" #register \
- : /* no output */ \
- :"r" (thread->debugreg[register]))
-
-
-void switch_to(struct domain *prev_p, struct domain *next_p)
-{
- struct thread_struct *next = &next_p->thread;
- struct tss_struct *tss = init_tss + smp_processor_id();
- execution_context_t *stack_ec = get_execution_context();
- int i;
-
- __cli();
-
- /* Switch guest general-register state. */
- if ( !is_idle_task(prev_p) )
- {
- memcpy(&prev_p->shared_info->execution_context,
- stack_ec,
- sizeof(*stack_ec));
- unlazy_fpu(prev_p);
- CLEAR_FAST_TRAP(&prev_p->thread);
- }
-
- if ( !is_idle_task(next_p) )
- {
- memcpy(stack_ec,
- &next_p->shared_info->execution_context,
- sizeof(*stack_ec));
-
- /*
- * This is sufficient! If the descriptor DPL differs from CS RPL then
- * we'll #GP. If DS, ES, FS, GS are DPL 0 then they'll be cleared
- * automatically. If SS RPL or DPL differs from CS RPL then we'll #GP.
- */
- if ( (stack_ec->cs & 3) == 0 )
- stack_ec->cs = FLAT_RING1_CS;
- if ( (stack_ec->ss & 3) == 0 )
- stack_ec->ss = FLAT_RING1_DS;
-
- SET_FAST_TRAP(&next_p->thread);
-
- /* Switch the guest OS ring-1 stack. */
- tss->esp1 = next->guestos_sp;
- tss->ss1 = next->guestos_ss;
-
- /* Maybe switch the debug registers. */
- if ( unlikely(next->debugreg[7]) )
- {
- loaddebug(next, 0);
- loaddebug(next, 1);
- loaddebug(next, 2);
- loaddebug(next, 3);
- /* no 4 and 5 */
- loaddebug(next, 6);
- loaddebug(next, 7);
- }
-
- /* Switch page tables. */
- write_ptbase(&next_p->mm);
- tlb_clocktick();
- }
-
- if ( unlikely(prev_p->io_bitmap != NULL) ||
- unlikely(next_p->io_bitmap != NULL) )
- {
- if ( next_p->io_bitmap != NULL )
- {
- /* Copy in the appropriate parts of the IO bitmap. We use the
- * selector to copy only the interesting parts of the bitmap. */
-
- u64 old_sel = ~0ULL; /* IO bitmap selector for previous task. */
-
- if ( prev_p->io_bitmap != NULL)
- {
- old_sel = prev_p->io_bitmap_sel;
-
- /* Replace any areas of the IO bitmap that had bits cleared. */
- for ( i = 0; i < sizeof(prev_p->io_bitmap_sel) * 8; i++ )
- if ( !test_bit(i, &prev_p->io_bitmap_sel) )
- memcpy(&tss->io_bitmap[i * IOBMP_SELBIT_LWORDS],
- &next_p->io_bitmap[i * IOBMP_SELBIT_LWORDS],
- IOBMP_SELBIT_LWORDS * sizeof(unsigned long));
- }
-
- /* Copy in any regions of the new task's bitmap that have bits
- * clear and we haven't already dealt with. */
- for ( i = 0; i < sizeof(prev_p->io_bitmap_sel) * 8; i++ )
- {
- if ( test_bit(i, &old_sel)
- && !test_bit(i, &next_p->io_bitmap_sel) )
- memcpy(&tss->io_bitmap[i * IOBMP_SELBIT_LWORDS],
- &next_p->io_bitmap[i * IOBMP_SELBIT_LWORDS],
- IOBMP_SELBIT_LWORDS * sizeof(unsigned long));
- }
-
- tss->bitmap = IO_BITMAP_OFFSET;
-
- }
- else
- {
- /* In this case, we're switching FROM a task with IO port access,
- * to a task that doesn't use the IO bitmap. We set any TSS bits
- * that might have been cleared, ready for future use. */
- for ( i = 0; i < sizeof(prev_p->io_bitmap_sel) * 8; i++ )
- if ( !test_bit(i, &prev_p->io_bitmap_sel) )
- memset(&tss->io_bitmap[i * IOBMP_SELBIT_LWORDS],
- 0xFF, IOBMP_SELBIT_LWORDS * sizeof(unsigned long));
-
- /*
- * a bitmap offset pointing outside of the TSS limit
- * causes a nicely controllable SIGSEGV if a process
- * tries to use a port IO instruction. The first
- * sys_ioperm() call sets up the bitmap properly.
- */
- tss->bitmap = INVALID_IO_BITMAP_OFFSET;
- }
- }
-
- set_current(next_p);
-
- /* Switch GDT and LDT. */
- __asm__ __volatile__ ("lgdt %0" : "=m" (*next_p->mm.gdt));
- load_LDT(next_p);
-
- __sti();
-}
-
-
-/* XXX Currently the 'domain' field is ignored! XXX */
-long do_iopl(domid_t domain, unsigned int new_io_pl)
-{
- execution_context_t *ec = get_execution_context();
- ec->eflags = (ec->eflags & 0xffffcfff) | ((new_io_pl&3) << 12);
- return 0;
-}
".align 4\n"
".globl __write_lock_failed\n"
"__write_lock_failed:\n"
-" " LOCK "addl $" RW_LOCK_BIAS_STR ",(%eax)\n"
+" " LOCK "addl $" RW_LOCK_BIAS_STR ",(%"__OP"ax)\n"
"1: rep; nop\n"
-" cmpl $" RW_LOCK_BIAS_STR ",(%eax)\n"
+" cmpl $" RW_LOCK_BIAS_STR ",(%"__OP"ax)\n"
" jne 1b\n"
-" " LOCK "subl $" RW_LOCK_BIAS_STR ",(%eax)\n"
+" " LOCK "subl $" RW_LOCK_BIAS_STR ",(%"__OP"ax)\n"
" jnz __write_lock_failed\n"
" ret\n"
".align 4\n"
".globl __read_lock_failed\n"
"__read_lock_failed:\n"
-" lock ; incl (%eax)\n"
+" lock ; incl (%"__OP"ax)\n"
"1: rep; nop\n"
-" cmpl $1,(%eax)\n"
+" cmpl $1,(%"__OP"ax)\n"
" js 1b\n"
-" lock ; decl (%eax)\n"
+" lock ; decl (%"__OP"ax)\n"
" js __read_lock_failed\n"
" ret\n"
);
int phys_proc_id[NR_CPUS];
int logical_proc_id[NR_CPUS];
+#if defined(__i386__)
+
/* Standard macro to see if a specific flag is changeable */
static inline int flag_is_changeable_p(u32 flag)
{
return flag_is_changeable_p(X86_EFLAGS_ID);
}
+#elif defined(__x86_64__)
+
+#define have_cpuid_p() (1)
+
+#endif
+
void __init get_cpu_vendor(struct cpuinfo_x86 *c)
{
char *v = c->x86_vendor_id;
unsigned long cpu_initialized;
void __init cpu_init(void)
{
+#if defined(__i386__) /* XXX */
int nr = smp_processor_id();
struct tss_struct * t = &init_tss[nr];
write_ptbase(¤t->mm);
init_idle_task();
+#endif
}
static void __init do_initcalls(void)
map_cpu_to_boot_apicid(cpu, apicid);
+#if defined(__i386__)
SET_DEFAULT_FAST_TRAP(&idle->thread);
+#endif
idle_task[cpu] = idle;
static unsigned long __init calibrate_tsc(void)
{
- unsigned long startlow, starthigh, endlow, endhigh, count;
+ u64 start, end, diff;
+ unsigned long count;
/* Set the Gate high, disable speaker */
outb((inb(0x61) & ~0x02) | 0x01, 0x61);
outb(CALIBRATE_LATCH & 0xff, 0x42); /* LSB of count */
outb(CALIBRATE_LATCH >> 8, 0x42); /* MSB of count */
- rdtsc(startlow, starthigh);
+ rdtscll(start);
for ( count = 0; (inb(0x61) & 0x20) == 0; count++ )
continue;
- rdtsc(endlow, endhigh);
+ rdtscll(end);
/* Error if the CTC doesn't behave itself. */
if ( count == 0 )
return 0;
- /* [endhigh:endlow] = [endhigh:endlow] - [starthigh:startlow] */
- __asm__( "subl %2,%0 ; sbbl %3,%1"
- : "=a" (endlow), "=d" (endhigh)
- : "g" (startlow), "g" (starthigh), "0" (endlow), "1" (endhigh) );
+ diff = end - start;
+#if defined(_i386__)
/* If quotient doesn't fit in 32 bits then we return error (zero). */
- return endhigh ? 0 : endlow;
+ if ( diff & ~0xffffffffULL )
+ return 0;
+#endif
+
+ return (unsigned long)diff;
}
lmsw %ax # into protected mode
jmp flush_instr
flush_instr:
- ljmpl $__HYPERVISOR_CS, $0x100000 # 1MB
+#if defined(__x86_64__)
+ ljmpl $__HYPERVISOR_CS32, $0x100000 # 1MB
+#else
+ ljmpl $__HYPERVISOR_CS, $0x100000 # 1MB
+#endif
idt_48:
.word 0 # idt limit = 0
unsigned long eip; /* 12 */
} guest_trap_bounce[NR_CPUS] = { { 0 } };
+#if defined(__i386__)
+
#define DOUBLEFAULT_STACK_SIZE 1024
static struct tss_struct doublefault_tss;
static unsigned char doublefault_stack[DOUBLEFAULT_STACK_SIZE];
if ( (reg < 0) || (reg > 7) ) return -EINVAL;
return current->thread.debugreg[reg];
}
+
+#endif /* __i386__ */
+++ /dev/null
-/*
- * User address space access functions.
- * The non inlined parts of asm-i386/uaccess.h are here.
- *
- * Copyright 1997 Andi Kleen <ak@muc.de>
- * Copyright 1997 Linus Torvalds
- */
-#include <xen/config.h>
-#include <asm/uaccess.h>
-//#include <asm/mmx.h>
-
-#ifdef CONFIG_X86_USE_3DNOW_AND_WORKS
-
-unsigned long
-__generic_copy_to_user(void *to, const void *from, unsigned long n)
-{
- if (access_ok(VERIFY_WRITE, to, n))
- {
- if(n<512)
- __copy_user(to,from,n);
- else
- mmx_copy_user(to,from,n);
- }
- return n;
-}
-
-unsigned long
-__generic_copy_from_user(void *to, const void *from, unsigned long n)
-{
- if (access_ok(VERIFY_READ, from, n))
- {
- if(n<512)
- __copy_user_zeroing(to,from,n);
- else
- mmx_copy_user_zeroing(to, from, n);
- }
- else
- memset(to, 0, n);
- return n;
-}
-
-#else
-
-unsigned long
-__generic_copy_to_user(void *to, const void *from, unsigned long n)
-{
- prefetch(from);
- if (access_ok(VERIFY_WRITE, to, n))
- __copy_user(to,from,n);
- return n;
-}
-
-unsigned long
-__generic_copy_from_user(void *to, const void *from, unsigned long n)
-{
- prefetchw(to);
- if (access_ok(VERIFY_READ, from, n))
- __copy_user_zeroing(to,from,n);
- else
- memset(to, 0, n);
- return n;
-}
-
-#endif
-
-/*
- * Copy a null terminated string from userspace.
- */
-
-#define __do_strncpy_from_user(dst,src,count,res) \
-do { \
- int __d0, __d1, __d2; \
- __asm__ __volatile__( \
- " testl %1,%1\n" \
- " jz 2f\n" \
- "0: lodsb\n" \
- " stosb\n" \
- " testb %%al,%%al\n" \
- " jz 1f\n" \
- " decl %1\n" \
- " jnz 0b\n" \
- "1: subl %1,%0\n" \
- "2:\n" \
- ".section .fixup,\"ax\"\n" \
- "3: movl %5,%0\n" \
- " jmp 2b\n" \
- ".previous\n" \
- ".section __ex_table,\"a\"\n" \
- " .align 4\n" \
- " .long 0b,3b\n" \
- ".previous" \
- : "=d"(res), "=c"(count), "=&a" (__d0), "=&S" (__d1), \
- "=&D" (__d2) \
- : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \
- : "memory"); \
-} while (0)
-
-long
-__strncpy_from_user(char *dst, const char *src, long count)
-{
- long res;
- __do_strncpy_from_user(dst, src, count, res);
- return res;
-}
-
-long
-strncpy_from_user(char *dst, const char *src, long count)
-{
- long res = -EFAULT;
- if (access_ok(VERIFY_READ, src, 1))
- __do_strncpy_from_user(dst, src, count, res);
- return res;
-}
-
-
-/*
- * Zero Userspace
- */
-
-#define __do_clear_user(addr,size) \
-do { \
- int __d0; \
- __asm__ __volatile__( \
- "0: rep; stosl\n" \
- " movl %2,%0\n" \
- "1: rep; stosb\n" \
- "2:\n" \
- ".section .fixup,\"ax\"\n" \
- "3: lea 0(%2,%0,4),%0\n" \
- " jmp 2b\n" \
- ".previous\n" \
- ".section __ex_table,\"a\"\n" \
- " .align 4\n" \
- " .long 0b,3b\n" \
- " .long 1b,2b\n" \
- ".previous" \
- : "=&c"(size), "=&D" (__d0) \
- : "r"(size & 3), "0"(size / 4), "1"(addr), "a"(0)); \
-} while (0)
-
-unsigned long
-clear_user(void *to, unsigned long n)
-{
- if (access_ok(VERIFY_WRITE, to, n))
- __do_clear_user(to, n);
- return n;
-}
-
-unsigned long
-__clear_user(void *to, unsigned long n)
-{
- __do_clear_user(to, n);
- return n;
-}
-
-/*
- * Return the size of a string (including the ending 0)
- *
- * Return 0 on exception, a value greater than N if too long
- */
-
-long strnlen_user(const char *s, long n)
-{
- unsigned long mask = -__addr_ok(s);
- unsigned long res, tmp;
-
- __asm__ __volatile__(
- " testl %0, %0\n"
- " jz 3f\n"
- " andl %0,%%ecx\n"
- "0: repne; scasb\n"
- " setne %%al\n"
- " subl %%ecx,%0\n"
- " addl %0,%%eax\n"
- "1:\n"
- ".section .fixup,\"ax\"\n"
- "2: xorl %%eax,%%eax\n"
- " jmp 1b\n"
- "3: movb $1,%%al\n"
- " jmp 1b\n"
- ".previous\n"
- ".section __ex_table,\"a\"\n"
- " .align 4\n"
- " .long 0b,2b\n"
- ".previous"
- :"=r" (n), "=D" (s), "=a" (res), "=c" (tmp)
- :"0" (n), "1" (s), "2" (0), "3" (mask)
- :"cc");
- return res & mask;
-}
--- /dev/null
+/******************************************************************************
+ * domain_page.h
+ *
+ * Allow temporary mapping of domain pages. Based on ideas from the
+ * Linux PKMAP code -- the copyrights and credits are retained below.
+ */
+
+/*
+ * (C) 1999 Andrea Arcangeli, SuSE GmbH, andrea@suse.de
+ * Gerhard Wichert, Siemens AG, Gerhard.Wichert@pdb.siemens.de *
+ * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
+ */
+
+#include <xen/config.h>
+#include <xen/sched.h>
+#include <xen/mm.h>
+#include <xen/perfc.h>
+#include <asm/domain_page.h>
+#include <asm/flushtlb.h>
+
+unsigned long *mapcache;
+static unsigned int map_idx, shadow_map_idx[NR_CPUS];
+static spinlock_t map_lock = SPIN_LOCK_UNLOCKED;
+
+/* Use a spare PTE bit to mark entries ready for recycling. */
+#define READY_FOR_TLB_FLUSH (1<<10)
+
+static void flush_all_ready_maps(void)
+{
+ unsigned long *cache = mapcache;
+
+ /* A bit skanky -- depends on having an aligned PAGE_SIZE set of PTEs. */
+ do { if ( (*cache & READY_FOR_TLB_FLUSH) ) *cache = 0; }
+ while ( ((unsigned long)(++cache) & ~PAGE_MASK) != 0 );
+
+ perfc_incrc(domain_page_tlb_flush);
+ local_flush_tlb();
+}
+
+
+void *map_domain_mem(unsigned long pa)
+{
+ unsigned long va;
+ unsigned int idx, cpu = smp_processor_id();
+ unsigned long *cache = mapcache;
+ unsigned long flags;
+
+ perfc_incrc(map_domain_mem_count);
+
+ spin_lock_irqsave(&map_lock, flags);
+
+ /* Has some other CPU caused a wrap? We must flush if so. */
+ if ( map_idx < shadow_map_idx[cpu] )
+ {
+ perfc_incrc(domain_page_tlb_flush);
+ local_flush_tlb();
+ }
+
+ for ( ; ; )
+ {
+ idx = map_idx = (map_idx + 1) & (MAPCACHE_ENTRIES - 1);
+ if ( idx == 0 ) flush_all_ready_maps();
+ if ( cache[idx] == 0 ) break;
+ }
+
+ cache[idx] = (pa & PAGE_MASK) | __PAGE_HYPERVISOR;
+
+ spin_unlock_irqrestore(&map_lock, flags);
+
+ shadow_map_idx[cpu] = idx;
+
+ va = MAPCACHE_VIRT_START + (idx << PAGE_SHIFT) + (pa & ~PAGE_MASK);
+ return (void *)va;
+}
+
+void unmap_domain_mem(void *va)
+{
+ unsigned int idx;
+ idx = ((unsigned long)va - MAPCACHE_VIRT_START) >> PAGE_SHIFT;
+ mapcache[idx] |= READY_FOR_TLB_FLUSH;
+}
--- /dev/null
+/*
+ * Hypercall and fault low-level handling routines.
+ *
+ * Copyright (c) 2002-2004, K A Fraser
+ * Copyright (c) 1991, 1992 Linus Torvalds
+ */
+
+/*
+ * The idea for callbacks to guest OSes
+ * ====================================
+ *
+ * First, we require that all callbacks (either via a supplied
+ * interrupt-descriptor-table, or via the special event or failsafe callbacks
+ * in the shared-info-structure) are to ring 1. This just makes life easier,
+ * in that it means we don't have to do messy GDT/LDT lookups to find
+ * out which the privilege-level of the return code-selector. That code
+ * would just be a hassle to write, and would need to account for running
+ * off the end of the GDT/LDT, for example. For all callbacks we check
+ * that the provided
+ * return CS is not == __HYPERVISOR_{CS,DS}. Apart from that we're safe as
+ * don't allow a guest OS to install ring-0 privileges into the GDT/LDT.
+ * It's up to the guest OS to ensure all returns via the IDT are to ring 1.
+ * If not, we load incorrect SS/ESP values from the TSS (for ring 1 rather
+ * than the correct ring) and bad things are bound to ensue -- IRET is
+ * likely to fault, and we may end up killing the domain (no harm can
+ * come to Xen, though).
+ *
+ * When doing a callback, we check if the return CS is in ring 0. If so,
+ * callback is delayed until next return to ring != 0.
+ * If return CS is in ring 1, then we create a callback frame
+ * starting at return SS/ESP. The base of the frame does an intra-privilege
+ * interrupt-return.
+ * If return CS is in ring > 1, we create a callback frame starting
+ * at SS/ESP taken from appropriate section of the current TSS. The base
+ * of the frame does an inter-privilege interrupt-return.
+ *
+ * Note that the "failsafe callback" uses a special stackframe:
+ * { return_DS, return_ES, return_FS, return_GS, return_EIP,
+ * return_CS, return_EFLAGS[, return_ESP, return_SS] }
+ * That is, original values for DS/ES/FS/GS are placed on stack rather than
+ * in DS/ES/FS/GS themselves. Why? It saves us loading them, only to have them
+ * saved/restored in guest OS. Furthermore, if we load them we may cause
+ * a fault if they are invalid, which is a hassle to deal with. We avoid
+ * that problem if we don't load them :-) This property allows us to use
+ * the failsafe callback as a fallback: if we ever fault on loading DS/ES/FS/GS
+ * on return to ring != 0, we can simply package it up as a return via
+ * the failsafe callback, and let the guest OS sort it out (perhaps by
+ * killing an application process). Note that we also do this for any
+ * faulting IRET -- just let the guest OS handle it via the event
+ * callback.
+ *
+ * We terminate a domain in the following cases:
+ * - creating a callback stack frame (due to bad ring-1 stack).
+ * - faulting IRET on entry to failsafe callback handler.
+ * So, each domain must keep its ring-1 %ss/%esp and failsafe callback
+ * handler in good order (absolutely no faults allowed!).
+ */
+
+#include <xen/config.h>
+#include <xen/errno.h>
+#include <hypervisor-ifs/hypervisor-if.h>
+
+EBX = 0x00
+ECX = 0x04
+EDX = 0x08
+ESI = 0x0C
+EDI = 0x10
+EBP = 0x14
+EAX = 0x18
+DS = 0x1C
+ES = 0x20
+FS = 0x24
+GS = 0x28
+ORIG_EAX = 0x2C
+EIP = 0x30
+CS = 0x34
+EFLAGS = 0x38
+OLDESP = 0x3C
+OLDSS = 0x40
+
+/* Offsets in domain structure */
+PROCESSOR = 0
+SHARED_INFO = 4
+EVENT_SEL = 8
+EVENT_ADDR = 12
+FAILSAFE_BUFFER = 16
+FAILSAFE_SEL = 32
+FAILSAFE_ADDR = 36
+
+/* Offsets in shared_info_t */
+#define UPCALL_PENDING /* 0 */
+#define UPCALL_MASK 1
+
+/* Offsets in guest_trap_bounce */
+GTB_ERROR_CODE = 0
+GTB_CR2 = 4
+GTB_FLAGS = 8
+GTB_CS = 10
+GTB_EIP = 12
+GTBF_TRAP = 1
+GTBF_TRAP_NOCODE = 2
+GTBF_TRAP_CR2 = 4
+
+CF_MASK = 0x00000001
+IF_MASK = 0x00000200
+NT_MASK = 0x00004000
+
+#define SAVE_ALL_NOSEGREGS \
+ cld; \
+ pushl %gs; \
+ pushl %fs; \
+ pushl %es; \
+ pushl %ds; \
+ pushl %eax; \
+ pushl %ebp; \
+ pushl %edi; \
+ pushl %esi; \
+ pushl %edx; \
+ pushl %ecx; \
+ pushl %ebx; \
+
+#define SAVE_ALL \
+ SAVE_ALL_NOSEGREGS \
+ movl $(__HYPERVISOR_DS),%edx; \
+ movl %edx,%ds; \
+ movl %edx,%es; \
+ movl %edx,%fs; \
+ movl %edx,%gs; \
+ sti;
+
+#define GET_CURRENT(reg) \
+ movl $4096-4, reg; \
+ orl %esp, reg; \
+ andl $~3,reg; \
+ movl (reg),reg;
+
+ENTRY(continue_nonidle_task)
+ GET_CURRENT(%ebx)
+ jmp test_all_events
+
+ ALIGN
+/*
+ * HYPERVISOR_multicall(call_list, nr_calls)
+ * Execute a list of 'nr_calls' hypercalls, pointed at by 'call_list'.
+ * This is fairly easy except that:
+ * 1. We may fault reading the call list, and must patch that up; and
+ * 2. We cannot recursively call HYPERVISOR_multicall, or a malicious
+ * caller could cause our stack to blow up.
+ */
+do_multicall:
+ popl %eax
+ cmpl $SYMBOL_NAME(multicall_return_from_call),%eax
+ je multicall_return_from_call
+ pushl %ebx
+ movl 4(%esp),%ebx /* EBX == call_list */
+ movl 8(%esp),%ecx /* ECX == nr_calls */
+multicall_loop:
+ pushl %ecx
+multicall_fault1:
+ pushl 20(%ebx) # args[4]
+multicall_fault2:
+ pushl 16(%ebx) # args[3]
+multicall_fault3:
+ pushl 12(%ebx) # args[2]
+multicall_fault4:
+ pushl 8(%ebx) # args[1]
+multicall_fault5:
+ pushl 4(%ebx) # args[0]
+multicall_fault6:
+ movl (%ebx),%eax # op
+ andl $(NR_hypercalls-1),%eax
+ call *SYMBOL_NAME(hypercall_table)(,%eax,4)
+multicall_return_from_call:
+multicall_fault7:
+ movl %eax,24(%ebx) # args[5] == result
+ addl $20,%esp
+ popl %ecx
+ addl $(ARGS_PER_MULTICALL_ENTRY*4),%ebx
+ loop multicall_loop
+ popl %ebx
+ xorl %eax,%eax
+ jmp ret_from_hypercall
+
+.section __ex_table,"a"
+ .align 4
+ .long multicall_fault1, multicall_fixup1
+ .long multicall_fault2, multicall_fixup2
+ .long multicall_fault3, multicall_fixup3
+ .long multicall_fault4, multicall_fixup4
+ .long multicall_fault5, multicall_fixup5
+ .long multicall_fault6, multicall_fixup6
+.previous
+
+.section .fixup,"ax"
+multicall_fixup6:
+ addl $4,%esp
+multicall_fixup5:
+ addl $4,%esp
+multicall_fixup4:
+ addl $4,%esp
+multicall_fixup3:
+ addl $4,%esp
+multicall_fixup2:
+ addl $4,%esp
+multicall_fixup1:
+ addl $4,%esp
+ popl %ebx
+ movl $-EFAULT,%eax
+ jmp ret_from_hypercall
+.previous
+
+ ALIGN
+restore_all_guest:
+ # First, may need to restore %ds if clobbered by create_bounce_frame
+ pushl %ss
+ popl %ds
+ # Second, create a failsafe copy of DS,ES,FS,GS in case any are bad
+ leal DS(%esp),%esi
+ leal FAILSAFE_BUFFER(%ebx),%edi
+ movsl
+ movsl
+ movsl
+ movsl
+ # Finally, restore guest registers -- faults will cause failsafe
+ popl %ebx
+ popl %ecx
+ popl %edx
+ popl %esi
+ popl %edi
+ popl %ebp
+ popl %eax
+1: popl %ds
+2: popl %es
+3: popl %fs
+4: popl %gs
+ addl $4,%esp
+5: iret
+.section .fixup,"ax"
+10: subl $4,%esp
+ pushl %gs
+9: pushl %fs
+8: pushl %es
+7: pushl %ds
+6: pushl %eax
+ pushl %ebp
+ pushl %edi
+ pushl %esi
+ pushl %edx
+ pushl %ecx
+ pushl %ebx
+ pushl %ss
+ popl %ds
+ pushl %ss
+ popl %es
+ jmp failsafe_callback
+.previous
+.section __ex_table,"a"
+ .align 4
+ .long 1b,6b
+ .long 2b,7b
+ .long 3b,8b
+ .long 4b,9b
+ .long 5b,10b
+.previous
+
+/* No special register assumptions */
+failsafe_callback:
+ GET_CURRENT(%ebx)
+ movl PROCESSOR(%ebx),%eax
+ shl $4,%eax
+ lea guest_trap_bounce(%eax),%edx
+ movl FAILSAFE_ADDR(%ebx),%eax
+ movl %eax,GTB_EIP(%edx)
+ movl FAILSAFE_SEL(%ebx),%eax
+ movw %ax,GTB_CS(%edx)
+ call create_bounce_frame
+ subl $16,%esi # add DS/ES/FS/GS to failsafe stack frame
+ leal FAILSAFE_BUFFER(%ebx),%ebp
+ movl 0(%ebp),%eax # DS
+FAULT1: movl %eax,(%esi)
+ movl 4(%ebp),%eax # ES
+FAULT2: movl %eax,4(%esi)
+ movl 8(%ebp),%eax # FS
+FAULT3: movl %eax,8(%esi)
+ movl 12(%ebp),%eax # GS
+FAULT4: movl %eax,12(%esi)
+ movl %esi,OLDESP(%esp)
+ popl %ebx
+ popl %ecx
+ popl %edx
+ popl %esi
+ popl %edi
+ popl %ebp
+ popl %eax
+ addl $20,%esp # skip DS/ES/FS/GS/ORIG_EAX
+FAULT5: iret
+
+
+ ALIGN
+# Simple restore -- we should never fault as we we will only interrupt ring 0
+# when sane values have been placed in all registers. The only exception is
+# NMI, which may interrupt before good values have been placed in DS-GS.
+# The NMI return code deals with this problem itself.
+restore_all_xen:
+ popl %ebx
+ popl %ecx
+ popl %edx
+ popl %esi
+ popl %edi
+ popl %ebp
+ popl %eax
+ popl %ds
+ popl %es
+ popl %fs
+ popl %gs
+ addl $4,%esp
+ iret
+
+ ALIGN
+ENTRY(hypercall)
+ pushl %eax # save orig_eax
+ SAVE_ALL
+ GET_CURRENT(%ebx)
+ andl $(NR_hypercalls-1),%eax
+ call *SYMBOL_NAME(hypercall_table)(,%eax,4)
+
+ret_from_hypercall:
+ movl %eax,EAX(%esp) # save the return value
+
+test_all_events:
+ xorl %ecx,%ecx
+ notl %ecx
+ cli # tests must not race interrupts
+/*test_softirqs:*/
+ movl PROCESSOR(%ebx),%eax
+ shl $6,%eax # sizeof(irq_cpustat) == 64
+ test %ecx,SYMBOL_NAME(irq_stat)(%eax,1)
+ jnz process_softirqs
+/*test_guest_events:*/
+ movl SHARED_INFO(%ebx),%eax
+ testb $0xFF,UPCALL_MASK(%eax)
+ jnz restore_all_guest
+ testb $0xFF,UPCALL_PENDING(%eax)
+ jz restore_all_guest
+ movb $1,UPCALL_MASK(%eax) # Upcalls are masked during delivery
+/*process_guest_events:*/
+ movl PROCESSOR(%ebx),%edx
+ shl $4,%edx # sizeof(guest_trap_bounce) == 16
+ lea guest_trap_bounce(%edx),%edx
+ movl EVENT_ADDR(%ebx),%eax
+ movl %eax,GTB_EIP(%edx)
+ movl EVENT_SEL(%ebx),%eax
+ movw %ax,GTB_CS(%edx)
+ call create_bounce_frame
+ jmp restore_all_guest
+
+ ALIGN
+process_softirqs:
+ sti
+ call SYMBOL_NAME(do_softirq)
+ jmp test_all_events
+
+/* CREATE A BASIC EXCEPTION FRAME ON GUEST OS (RING-1) STACK: */
+/* {EIP, CS, EFLAGS, [ESP, SS]} */
+/* %edx == guest_trap_bounce, %ebx == task_struct */
+/* %eax,%ecx are clobbered. %ds:%esi contain new OLDSS/OLDESP. */
+create_bounce_frame:
+ mov CS+4(%esp),%cl
+ test $2,%cl
+ jz 1f /* jump if returning to an existing ring-1 activation */
+ /* obtain ss/esp from TSS -- no current ring-1 activations */
+ movl PROCESSOR(%ebx),%eax
+ /* next 4 lines multiply %eax by 8320, which is sizeof(tss_struct) */
+ movl %eax, %ecx
+ shll $7, %ecx
+ shll $13, %eax
+ addl %ecx,%eax
+ addl $init_tss + 12,%eax
+ movl (%eax),%esi /* tss->esp1 */
+FAULT6: movl 4(%eax),%ds /* tss->ss1 */
+ /* base of stack frame must contain ss/esp (inter-priv iret) */
+ subl $8,%esi
+ movl OLDESP+4(%esp),%eax
+FAULT7: movl %eax,(%esi)
+ movl OLDSS+4(%esp),%eax
+FAULT8: movl %eax,4(%esi)
+ jmp 2f
+1: /* obtain ss/esp from oldss/oldesp -- a ring-1 activation exists */
+ movl OLDESP+4(%esp),%esi
+FAULT9: movl OLDSS+4(%esp),%ds
+2: /* Construct a stack frame: EFLAGS, CS/EIP */
+ subl $12,%esi
+ movl EIP+4(%esp),%eax
+FAULT10:movl %eax,(%esi)
+ movl CS+4(%esp),%eax
+FAULT11:movl %eax,4(%esi)
+ movl EFLAGS+4(%esp),%eax
+FAULT12:movl %eax,8(%esi)
+ /* Rewrite our stack frame and return to ring 1. */
+ /* IA32 Ref. Vol. 3: TF, VM, RF and NT flags are cleared on trap. */
+ andl $0xfffcbeff,%eax
+ movl %eax,EFLAGS+4(%esp)
+ movl %ds,OLDSS+4(%esp)
+ movl %esi,OLDESP+4(%esp)
+ movzwl %es:GTB_CS(%edx),%eax
+ movl %eax,CS+4(%esp)
+ movl %es:GTB_EIP(%edx),%eax
+ movl %eax,EIP+4(%esp)
+ ret
+
+
+.section __ex_table,"a"
+ .align 4
+ .long FAULT1, crash_domain_fixup3 # Fault writing to ring-1 stack
+ .long FAULT2, crash_domain_fixup3 # Fault writing to ring-1 stack
+ .long FAULT3, crash_domain_fixup3 # Fault writing to ring-1 stack
+ .long FAULT4, crash_domain_fixup3 # Fault writing to ring-1 stack
+ .long FAULT5, crash_domain_fixup1 # Fault executing failsafe iret
+ .long FAULT6, crash_domain_fixup2 # Fault loading ring-1 stack selector
+ .long FAULT7, crash_domain_fixup2 # Fault writing to ring-1 stack
+ .long FAULT8, crash_domain_fixup2 # Fault writing to ring-1 stack
+ .long FAULT9, crash_domain_fixup2 # Fault loading ring-1 stack selector
+ .long FAULT10,crash_domain_fixup2 # Fault writing to ring-1 stack
+ .long FAULT11,crash_domain_fixup2 # Fault writing to ring-1 stack
+ .long FAULT12,crash_domain_fixup2 # Fault writing to ring-1 stack
+ .long FAULT13,crash_domain_fixup3 # Fault writing to ring-1 stack
+ .long FAULT14,crash_domain_fixup3 # Fault writing to ring-1 stack
+.previous
+
+# This handler kills domains which experience unrecoverable faults.
+.section .fixup,"ax"
+crash_domain_fixup1:
+ subl $4,%esp
+ SAVE_ALL
+ jmp domain_crash
+crash_domain_fixup2:
+ addl $4,%esp
+crash_domain_fixup3:
+ pushl %ss
+ popl %ds
+ jmp domain_crash
+.previous
+
+ ALIGN
+process_guest_exception_and_events:
+ movl PROCESSOR(%ebx),%eax
+ shl $4,%eax
+ lea guest_trap_bounce(%eax),%edx
+ testb $~0,GTB_FLAGS(%edx)
+ jz test_all_events
+ call create_bounce_frame # just the basic frame
+ mov %es:GTB_FLAGS(%edx),%cl
+ test $GTBF_TRAP_NOCODE,%cl
+ jnz 2f
+ subl $4,%esi # push error_code onto guest frame
+ movl %es:GTB_ERROR_CODE(%edx),%eax
+FAULT13:movl %eax,(%esi)
+ test $GTBF_TRAP_CR2,%cl
+ jz 1f
+ subl $4,%esi # push %cr2 onto guest frame
+ movl %es:GTB_CR2(%edx),%eax
+FAULT14:movl %eax,(%esi)
+1: movl %esi,OLDESP(%esp)
+2: push %es # unclobber %ds
+ pop %ds
+ movb $0,GTB_FLAGS(%edx)
+ jmp test_all_events
+
+ ALIGN
+ENTRY(ret_from_intr)
+ GET_CURRENT(%ebx)
+ movb CS(%esp),%al
+ testb $3,%al # return to non-supervisor?
+ jne test_all_events
+ jmp restore_all_xen
+
+ENTRY(divide_error)
+ pushl $0 # no error code
+ pushl $ SYMBOL_NAME(do_divide_error)
+ ALIGN
+error_code:
+ pushl %fs
+ pushl %es
+ pushl %ds
+ pushl %eax
+ xorl %eax,%eax
+ pushl %ebp
+ pushl %edi
+ pushl %esi
+ pushl %edx
+ decl %eax # eax = -1
+ pushl %ecx
+ pushl %ebx
+ cld
+ movl %gs,%ecx
+ movl ORIG_EAX(%esp), %esi # get the error code
+ movl GS(%esp), %edi # get the function address
+ movl %eax, ORIG_EAX(%esp)
+ movl %ecx, GS(%esp)
+ movl $(__HYPERVISOR_DS),%edx
+ movl %edx,%ds
+ movl %edx,%es
+ movl %edx,%fs
+ movl %edx,%gs
+ movl %esp,%edx
+ pushl %esi # push the error code
+ pushl %edx # push the pt_regs pointer
+ GET_CURRENT(%ebx)
+ call *%edi
+ addl $8,%esp
+ movb CS(%esp),%al
+ testb $3,%al
+ je restore_all_xen
+ jmp process_guest_exception_and_events
+
+ENTRY(coprocessor_error)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_coprocessor_error)
+ jmp error_code
+
+ENTRY(simd_coprocessor_error)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_simd_coprocessor_error)
+ jmp error_code
+
+ENTRY(device_not_available)
+ pushl $0
+ pushl $SYMBOL_NAME(math_state_restore)
+ jmp error_code
+
+ENTRY(debug)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_debug)
+ jmp error_code
+
+ENTRY(int3)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_int3)
+ jmp error_code
+
+ENTRY(overflow)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_overflow)
+ jmp error_code
+
+ENTRY(bounds)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_bounds)
+ jmp error_code
+
+ENTRY(invalid_op)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_invalid_op)
+ jmp error_code
+
+ENTRY(coprocessor_segment_overrun)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_coprocessor_segment_overrun)
+ jmp error_code
+
+ENTRY(invalid_TSS)
+ pushl $ SYMBOL_NAME(do_invalid_TSS)
+ jmp error_code
+
+ENTRY(segment_not_present)
+ pushl $ SYMBOL_NAME(do_segment_not_present)
+ jmp error_code
+
+ENTRY(stack_segment)
+ pushl $ SYMBOL_NAME(do_stack_segment)
+ jmp error_code
+
+ENTRY(general_protection)
+ pushl $ SYMBOL_NAME(do_general_protection)
+ jmp error_code
+
+ENTRY(alignment_check)
+ pushl $ SYMBOL_NAME(do_alignment_check)
+ jmp error_code
+
+ENTRY(page_fault)
+ pushl $ SYMBOL_NAME(do_page_fault)
+ jmp error_code
+
+ENTRY(machine_check)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_machine_check)
+ jmp error_code
+
+ENTRY(spurious_interrupt_bug)
+ pushl $0
+ pushl $ SYMBOL_NAME(do_spurious_interrupt_bug)
+ jmp error_code
+
+ENTRY(nmi)
+ # Save state but do not trash the segment registers!
+ # We may otherwise be unable to reload them or copy them to ring 1.
+ pushl %eax
+ SAVE_ALL_NOSEGREGS
+
+ # Check for hardware problems. These are always fatal so we can
+ # reload DS and ES when handling them.
+ inb $0x61,%al
+ testb $0x80,%al
+ jne nmi_parity_err
+ testb $0x40,%al
+ jne nmi_io_err
+ movl %eax,%ebx
+
+ # Okay, its almost a normal NMI tick. We can only process it if:
+ # A. We are the outermost Xen activation (in which case we have
+ # the selectors safely saved on our stack)
+ # B. DS-GS all contain sane Xen values.
+ # In all other cases we bail without touching DS-GS, as we have
+ # interrupted an enclosing Xen activation in tricky prologue or
+ # epilogue code.
+ movb CS(%esp),%al
+ testb $3,%al
+ jne do_watchdog_tick
+ movl DS(%esp),%eax
+ cmpw $(__HYPERVISOR_DS),%ax
+ jne nmi_badseg
+ movl ES(%esp),%eax
+ cmpw $(__HYPERVISOR_DS),%ax
+ jne nmi_badseg
+ movl FS(%esp),%eax
+ cmpw $(__HYPERVISOR_DS),%ax
+ jne nmi_badseg
+ movl GS(%esp),%eax
+ cmpw $(__HYPERVISOR_DS),%ax
+ jne nmi_badseg
+
+do_watchdog_tick:
+ movl $(__HYPERVISOR_DS),%edx
+ movl %edx,%ds
+ movl %edx,%es
+ movl %esp,%edx
+ pushl %ebx # reason
+ pushl %edx # regs
+ call SYMBOL_NAME(do_nmi)
+ addl $8,%esp
+ movb CS(%esp),%al
+ testb $3,%al
+ je restore_all_xen
+ GET_CURRENT(%ebx)
+ jmp restore_all_guest
+
+nmi_badseg:
+ popl %ebx
+ popl %ecx
+ popl %edx
+ popl %esi
+ popl %edi
+ popl %ebp
+ popl %eax
+ addl $20,%esp
+ iret
+
+nmi_parity_err:
+ movl $(__HYPERVISOR_DS),%edx
+ movl %edx,%ds
+ movl %edx,%es
+ jmp SYMBOL_NAME(mem_parity_error)
+
+nmi_io_err:
+ movl $(__HYPERVISOR_DS),%edx
+ movl %edx,%ds
+ movl %edx,%es
+ jmp SYMBOL_NAME(io_check_error)
+
+.data
+ENTRY(hypercall_table)
+ .long SYMBOL_NAME(do_set_trap_table) /* 0 */
+ .long SYMBOL_NAME(do_mmu_update)
+ .long SYMBOL_NAME(do_set_gdt)
+ .long SYMBOL_NAME(do_stack_switch)
+ .long SYMBOL_NAME(do_set_callbacks)
+ .long SYMBOL_NAME(do_fpu_taskswitch) /* 5 */
+ .long SYMBOL_NAME(do_sched_op)
+ .long SYMBOL_NAME(do_dom0_op)
+ .long SYMBOL_NAME(do_set_debugreg)
+ .long SYMBOL_NAME(do_get_debugreg)
+ .long SYMBOL_NAME(do_update_descriptor) /* 10 */
+ .long SYMBOL_NAME(do_set_fast_trap)
+ .long SYMBOL_NAME(do_dom_mem_op)
+ .long SYMBOL_NAME(do_multicall)
+ .long SYMBOL_NAME(do_update_va_mapping)
+ .long SYMBOL_NAME(do_set_timer_op) /* 15 */
+ .long SYMBOL_NAME(do_event_channel_op)
+ .long SYMBOL_NAME(do_xen_version)
+ .long SYMBOL_NAME(do_console_io)
+ .long SYMBOL_NAME(do_physdev_op)
+ .long SYMBOL_NAME(do_update_va_mapping_otherdomain) /* 20 */
+ .rept NR_hypercalls-((.-hypercall_table)/4)
+ .long SYMBOL_NAME(do_ni_hypercall)
+ .endr
--- /dev/null
+/******************************************************************************
+ * arch/i386/mm.c
+ *
+ * Modifications to Linux original are copyright (c) 2002-2003, K A Fraser
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ */
+
+#include <xen/config.h>
+#include <xen/lib.h>
+#include <xen/init.h>
+#include <xen/mm.h>
+#include <asm/page.h>
+#include <asm/flushtlb.h>
+#include <asm/fixmap.h>
+#include <asm/domain_page.h>
+
+static inline void set_pte_phys(unsigned long vaddr,
+ l1_pgentry_t entry)
+{
+ l2_pgentry_t *l2ent;
+ l1_pgentry_t *l1ent;
+
+ l2ent = &idle_pg_table[l2_table_offset(vaddr)];
+ l1ent = l2_pgentry_to_l1(*l2ent) + l1_table_offset(vaddr);
+ *l1ent = entry;
+
+ /* It's enough to flush this one mapping. */
+ __flush_tlb_one(vaddr);
+}
+
+
+void __set_fixmap(enum fixed_addresses idx,
+ l1_pgentry_t entry)
+{
+ unsigned long address = fix_to_virt(idx);
+
+ if ( likely(idx < __end_of_fixed_addresses) )
+ set_pte_phys(address, entry);
+ else
+ printk("Invalid __set_fixmap\n");
+}
+
+
+static void __init fixrange_init(unsigned long start,
+ unsigned long end,
+ l2_pgentry_t *pg_base)
+{
+ l2_pgentry_t *l2e;
+ int i;
+ unsigned long vaddr, page;
+
+ vaddr = start;
+ i = l2_table_offset(vaddr);
+ l2e = pg_base + i;
+
+ for ( ; (i < ENTRIES_PER_L2_PAGETABLE) && (vaddr != end); l2e++, i++ )
+ {
+ if ( !l2_pgentry_empty(*l2e) )
+ continue;
+ page = (unsigned long)get_free_page();
+ clear_page(page);
+ *l2e = mk_l2_pgentry(__pa(page) | __PAGE_HYPERVISOR);
+ vaddr += 1 << L2_PAGETABLE_SHIFT;
+ }
+}
+
+void __init paging_init(void)
+{
+ unsigned long addr;
+ void *ioremap_pt;
+ int i;
+
+ /* Idle page table 1:1 maps the first part of physical memory. */
+ for ( i = 0; i < DOMAIN_ENTRIES_PER_L2_PAGETABLE; i++ )
+ idle_pg_table[i] =
+ mk_l2_pgentry((i << L2_PAGETABLE_SHIFT) |
+ __PAGE_HYPERVISOR | _PAGE_PSE);
+
+ /*
+ * Fixed mappings, only the page table structure has to be
+ * created - mappings will be set by set_fixmap():
+ */
+ addr = FIXADDR_START & ~((1<<L2_PAGETABLE_SHIFT)-1);
+ fixrange_init(addr, 0, idle_pg_table);
+
+ /* Create page table for ioremap(). */
+ ioremap_pt = (void *)get_free_page();
+ clear_page(ioremap_pt);
+ idle_pg_table[IOREMAP_VIRT_START >> L2_PAGETABLE_SHIFT] =
+ mk_l2_pgentry(__pa(ioremap_pt) | __PAGE_HYPERVISOR);
+
+ /* Create read-only mapping of MPT for guest-OS use. */
+ idle_pg_table[RO_MPT_VIRT_START >> L2_PAGETABLE_SHIFT] =
+ idle_pg_table[RDWR_MPT_VIRT_START >> L2_PAGETABLE_SHIFT];
+ mk_l2_readonly(idle_pg_table +
+ (RO_MPT_VIRT_START >> L2_PAGETABLE_SHIFT));
+
+ /* Set up mapping cache for domain pages. */
+ mapcache = (unsigned long *)get_free_page();
+ clear_page(mapcache);
+ idle_pg_table[MAPCACHE_VIRT_START >> L2_PAGETABLE_SHIFT] =
+ mk_l2_pgentry(__pa(mapcache) | __PAGE_HYPERVISOR);
+
+ /* Set up linear page table mapping. */
+ idle_pg_table[LINEAR_PT_VIRT_START >> L2_PAGETABLE_SHIFT] =
+ mk_l2_pgentry(__pa(idle_pg_table) | __PAGE_HYPERVISOR);
+
+}
+
+void __init zap_low_mappings(void)
+{
+ int i;
+ for ( i = 0; i < DOMAIN_ENTRIES_PER_L2_PAGETABLE; i++ )
+ idle_pg_table[i] = mk_l2_pgentry(0);
+ flush_tlb_all_pge();
+}
+
+
+long do_stack_switch(unsigned long ss, unsigned long esp)
+{
+ int nr = smp_processor_id();
+ struct tss_struct *t = &init_tss[nr];
+
+ /* We need to do this check as we load and use SS on guest's behalf. */
+ if ( (ss & 3) == 0 )
+ return -EPERM;
+
+ current->thread.guestos_ss = ss;
+ current->thread.guestos_sp = esp;
+ t->ss1 = ss;
+ t->esp1 = esp;
+
+ return 0;
+}
+
+
+/* Returns TRUE if given descriptor is valid for GDT or LDT. */
+int check_descriptor(unsigned long a, unsigned long b)
+{
+ unsigned long base, limit;
+
+ /* A not-present descriptor will always fault, so is safe. */
+ if ( !(b & _SEGMENT_P) )
+ goto good;
+
+ /*
+ * We don't allow a DPL of zero. There is no legitimate reason for
+ * specifying DPL==0, and it gets rather dangerous if we also accept call
+ * gates (consider a call gate pointing at another guestos descriptor with
+ * DPL 0 -- this would get the OS ring-0 privileges).
+ */
+ if ( (b & _SEGMENT_DPL) == 0 )
+ goto bad;
+
+ if ( !(b & _SEGMENT_S) )
+ {
+ /*
+ * System segment:
+ * 1. Don't allow interrupt or trap gates as they belong in the IDT.
+ * 2. Don't allow TSS descriptors or task gates as we don't
+ * virtualise x86 tasks.
+ * 3. Don't allow LDT descriptors because they're unnecessary and
+ * I'm uneasy about allowing an LDT page to contain LDT
+ * descriptors. In any case, Xen automatically creates the
+ * required descriptor when reloading the LDT register.
+ * 4. We allow call gates but they must not jump to a private segment.
+ */
+
+ /* Disallow everything but call gates. */
+ if ( (b & _SEGMENT_TYPE) != 0xc00 )
+ goto bad;
+
+ /* Can't allow far jump to a Xen-private segment. */
+ if ( !VALID_CODESEL(a>>16) )
+ goto bad;
+
+ /* Reserved bits must be zero. */
+ if ( (b & 0xe0) != 0 )
+ goto bad;
+
+ /* No base/limit check is needed for a call gate. */
+ goto good;
+ }
+
+ /* Check that base/limit do not overlap Xen-private space. */
+ base = (b&(0xff<<24)) | ((b&0xff)<<16) | (a>>16);
+ limit = (b&0xf0000) | (a&0xffff);
+ limit++; /* We add one because limit is inclusive. */
+ if ( (b & _SEGMENT_G) )
+ limit <<= 12;
+ if ( ((base + limit) <= base) ||
+ ((base + limit) > PAGE_OFFSET) )
+ goto bad;
+
+ good:
+ return 1;
+ bad:
+ return 0;
+}
+
+
+long set_gdt(struct domain *p,
+ unsigned long *frames,
+ unsigned int entries)
+{
+ /* NB. There are 512 8-byte entries per GDT page. */
+ int i, nr_pages = (entries + 511) / 512;
+ unsigned long pfn;
+ struct desc_struct *vgdt;
+
+ /* Check the new GDT. */
+ for ( i = 0; i < nr_pages; i++ )
+ {
+ if ( unlikely(frames[i] >= max_page) ||
+ unlikely(!get_page_and_type(&frame_table[frames[i]],
+ p, PGT_gdt_page)) )
+ goto fail;
+ }
+
+ /* Copy reserved GDT entries to the new GDT. */
+ vgdt = map_domain_mem(frames[0] << PAGE_SHIFT);
+ memcpy(vgdt + FIRST_RESERVED_GDT_ENTRY,
+ gdt_table + FIRST_RESERVED_GDT_ENTRY,
+ NR_RESERVED_GDT_ENTRIES*8);
+ unmap_domain_mem(vgdt);
+
+ /* Tear down the old GDT. */
+ for ( i = 0; i < 16; i++ )
+ {
+ if ( (pfn = l1_pgentry_to_pagenr(p->mm.perdomain_pt[i])) != 0 )
+ put_page_and_type(&frame_table[pfn]);
+ p->mm.perdomain_pt[i] = mk_l1_pgentry(0);
+ }
+
+ /* Install the new GDT. */
+ for ( i = 0; i < nr_pages; i++ )
+ p->mm.perdomain_pt[i] =
+ mk_l1_pgentry((frames[i] << PAGE_SHIFT) | __PAGE_HYPERVISOR);
+
+ SET_GDT_ADDRESS(p, GDT_VIRT_START);
+ SET_GDT_ENTRIES(p, (entries*8)-1);
+
+ return 0;
+
+ fail:
+ while ( i-- > 0 )
+ put_page_and_type(&frame_table[frames[i]]);
+ return -EINVAL;
+}
+
+
+long do_set_gdt(unsigned long *frame_list, unsigned int entries)
+{
+ int nr_pages = (entries + 511) / 512;
+ unsigned long frames[16];
+ long ret;
+
+ if ( (entries <= LAST_RESERVED_GDT_ENTRY) || (entries > 8192) )
+ return -EINVAL;
+
+ if ( copy_from_user(frames, frame_list, nr_pages * sizeof(unsigned long)) )
+ return -EFAULT;
+
+ if ( (ret = set_gdt(current, frames, entries)) == 0 )
+ {
+ local_flush_tlb();
+ __asm__ __volatile__ ("lgdt %0" : "=m" (*current->mm.gdt));
+ }
+
+ return ret;
+}
+
+
+long do_update_descriptor(
+ unsigned long pa, unsigned long word1, unsigned long word2)
+{
+ unsigned long *gdt_pent, pfn = pa >> PAGE_SHIFT;
+ struct pfn_info *page;
+ long ret = -EINVAL;
+
+ if ( (pa & 7) || (pfn >= max_page) || !check_descriptor(word1, word2) )
+ return -EINVAL;
+
+ page = &frame_table[pfn];
+ if ( unlikely(!get_page(page, current)) )
+ goto out;
+
+ /* Check if the given frame is in use in an unsafe context. */
+ switch ( page->type_and_flags & PGT_type_mask )
+ {
+ case PGT_gdt_page:
+ /* Disallow updates of Xen-reserved descriptors in the current GDT. */
+ if ( (l1_pgentry_to_pagenr(current->mm.perdomain_pt[0]) == pfn) &&
+ (((pa&(PAGE_SIZE-1))>>3) >= FIRST_RESERVED_GDT_ENTRY) &&
+ (((pa&(PAGE_SIZE-1))>>3) <= LAST_RESERVED_GDT_ENTRY) )
+ goto out;
+ if ( unlikely(!get_page_type(page, PGT_gdt_page)) )
+ goto out;
+ break;
+ case PGT_ldt_page:
+ if ( unlikely(!get_page_type(page, PGT_ldt_page)) )
+ goto out;
+ break;
+ default:
+ if ( unlikely(!get_page_type(page, PGT_writeable_page)) )
+ goto out;
+ break;
+ }
+
+ /* All is good so make the update. */
+ gdt_pent = map_domain_mem(pa);
+ gdt_pent[0] = word1;
+ gdt_pent[1] = word2;
+ unmap_domain_mem(gdt_pent);
+
+ put_page_type(page);
+
+ ret = 0; /* success */
+
+ out:
+ put_page(page);
+ return ret;
+}
+
+#ifdef MEMORY_GUARD
+
+void *memguard_init(void *heap_start)
+{
+ l1_pgentry_t *l1;
+ int i, j;
+
+ /* Round the allocation pointer up to a page boundary. */
+ heap_start = (void *)(((unsigned long)heap_start + (PAGE_SIZE-1)) &
+ PAGE_MASK);
+
+ /* Memory guarding is incompatible with super pages. */
+ for ( i = 0; i < (xenheap_phys_end >> L2_PAGETABLE_SHIFT); i++ )
+ {
+ l1 = (l1_pgentry_t *)heap_start;
+ heap_start = (void *)((unsigned long)heap_start + PAGE_SIZE);
+ for ( j = 0; j < ENTRIES_PER_L1_PAGETABLE; j++ )
+ l1[j] = mk_l1_pgentry((i << L2_PAGETABLE_SHIFT) |
+ (j << L1_PAGETABLE_SHIFT) |
+ __PAGE_HYPERVISOR);
+ idle_pg_table[i] = idle_pg_table[i + l2_table_offset(PAGE_OFFSET)] =
+ mk_l2_pgentry(virt_to_phys(l1) | __PAGE_HYPERVISOR);
+ }
+
+ return heap_start;
+}
+
+static void __memguard_change_range(void *p, unsigned long l, int guard)
+{
+ l1_pgentry_t *l1;
+ l2_pgentry_t *l2;
+ unsigned long _p = (unsigned long)p;
+ unsigned long _l = (unsigned long)l;
+
+ /* Ensure we are dealing with a page-aligned whole number of pages. */
+ ASSERT((_p&PAGE_MASK) != 0);
+ ASSERT((_l&PAGE_MASK) != 0);
+ ASSERT((_p&~PAGE_MASK) == 0);
+ ASSERT((_l&~PAGE_MASK) == 0);
+
+ while ( _l != 0 )
+ {
+ l2 = &idle_pg_table[l2_table_offset(_p)];
+ l1 = l2_pgentry_to_l1(*l2) + l1_table_offset(_p);
+ if ( guard )
+ *l1 = mk_l1_pgentry(l1_pgentry_val(*l1) & ~_PAGE_PRESENT);
+ else
+ *l1 = mk_l1_pgentry(l1_pgentry_val(*l1) | _PAGE_PRESENT);
+ _p += PAGE_SIZE;
+ _l -= PAGE_SIZE;
+ }
+}
+
+void memguard_guard_range(void *p, unsigned long l)
+{
+ __memguard_change_range(p, l, 1);
+ local_flush_tlb();
+}
+
+void memguard_unguard_range(void *p, unsigned long l)
+{
+ __memguard_change_range(p, l, 0);
+}
+
+int memguard_is_guarded(void *p)
+{
+ l1_pgentry_t *l1;
+ l2_pgentry_t *l2;
+ unsigned long _p = (unsigned long)p;
+ l2 = &idle_pg_table[l2_table_offset(_p)];
+ l1 = l2_pgentry_to_l1(*l2) + l1_table_offset(_p);
+ return !(l1_pgentry_val(*l1) & _PAGE_PRESENT);
+}
+
+#endif
--- /dev/null
+/*
+ * User address space access functions.
+ * The non inlined parts of asm-i386/uaccess.h are here.
+ *
+ * Copyright 1997 Andi Kleen <ak@muc.de>
+ * Copyright 1997 Linus Torvalds
+ */
+#include <xen/config.h>
+#include <asm/uaccess.h>
+//#include <asm/mmx.h>
+
+#ifdef CONFIG_X86_USE_3DNOW_AND_WORKS
+
+unsigned long
+__generic_copy_to_user(void *to, const void *from, unsigned long n)
+{
+ if (access_ok(VERIFY_WRITE, to, n))
+ {
+ if(n<512)
+ __copy_user(to,from,n);
+ else
+ mmx_copy_user(to,from,n);
+ }
+ return n;
+}
+
+unsigned long
+__generic_copy_from_user(void *to, const void *from, unsigned long n)
+{
+ if (access_ok(VERIFY_READ, from, n))
+ {
+ if(n<512)
+ __copy_user_zeroing(to,from,n);
+ else
+ mmx_copy_user_zeroing(to, from, n);
+ }
+ else
+ memset(to, 0, n);
+ return n;
+}
+
+#else
+
+unsigned long
+__generic_copy_to_user(void *to, const void *from, unsigned long n)
+{
+ prefetch(from);
+ if (access_ok(VERIFY_WRITE, to, n))
+ __copy_user(to,from,n);
+ return n;
+}
+
+unsigned long
+__generic_copy_from_user(void *to, const void *from, unsigned long n)
+{
+ prefetchw(to);
+ if (access_ok(VERIFY_READ, from, n))
+ __copy_user_zeroing(to,from,n);
+ else
+ memset(to, 0, n);
+ return n;
+}
+
+#endif
+
+/*
+ * Copy a null terminated string from userspace.
+ */
+
+#define __do_strncpy_from_user(dst,src,count,res) \
+do { \
+ int __d0, __d1, __d2; \
+ __asm__ __volatile__( \
+ " testl %1,%1\n" \
+ " jz 2f\n" \
+ "0: lodsb\n" \
+ " stosb\n" \
+ " testb %%al,%%al\n" \
+ " jz 1f\n" \
+ " decl %1\n" \
+ " jnz 0b\n" \
+ "1: subl %1,%0\n" \
+ "2:\n" \
+ ".section .fixup,\"ax\"\n" \
+ "3: movl %5,%0\n" \
+ " jmp 2b\n" \
+ ".previous\n" \
+ ".section __ex_table,\"a\"\n" \
+ " .align 4\n" \
+ " .long 0b,3b\n" \
+ ".previous" \
+ : "=d"(res), "=c"(count), "=&a" (__d0), "=&S" (__d1), \
+ "=&D" (__d2) \
+ : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \
+ : "memory"); \
+} while (0)
+
+long
+__strncpy_from_user(char *dst, const char *src, long count)
+{
+ long res;
+ __do_strncpy_from_user(dst, src, count, res);
+ return res;
+}
+
+long
+strncpy_from_user(char *dst, const char *src, long count)
+{
+ long res = -EFAULT;
+ if (access_ok(VERIFY_READ, src, 1))
+ __do_strncpy_from_user(dst, src, count, res);
+ return res;
+}
+
+
+/*
+ * Zero Userspace
+ */
+
+#define __do_clear_user(addr,size) \
+do { \
+ int __d0; \
+ __asm__ __volatile__( \
+ "0: rep; stosl\n" \
+ " movl %2,%0\n" \
+ "1: rep; stosb\n" \
+ "2:\n" \
+ ".section .fixup,\"ax\"\n" \
+ "3: lea 0(%2,%0,4),%0\n" \
+ " jmp 2b\n" \
+ ".previous\n" \
+ ".section __ex_table,\"a\"\n" \
+ " .align 4\n" \
+ " .long 0b,3b\n" \
+ " .long 1b,2b\n" \
+ ".previous" \
+ : "=&c"(size), "=&D" (__d0) \
+ : "r"(size & 3), "0"(size / 4), "1"(addr), "a"(0)); \
+} while (0)
+
+unsigned long
+clear_user(void *to, unsigned long n)
+{
+ if (access_ok(VERIFY_WRITE, to, n))
+ __do_clear_user(to, n);
+ return n;
+}
+
+unsigned long
+__clear_user(void *to, unsigned long n)
+{
+ __do_clear_user(to, n);
+ return n;
+}
+
+/*
+ * Return the size of a string (including the ending 0)
+ *
+ * Return 0 on exception, a value greater than N if too long
+ */
+
+long strnlen_user(const char *s, long n)
+{
+ unsigned long mask = -__addr_ok(s);
+ unsigned long res, tmp;
+
+ __asm__ __volatile__(
+ " testl %0, %0\n"
+ " jz 3f\n"
+ " andl %0,%%ecx\n"
+ "0: repne; scasb\n"
+ " setne %%al\n"
+ " subl %%ecx,%0\n"
+ " addl %0,%%eax\n"
+ "1:\n"
+ ".section .fixup,\"ax\"\n"
+ "2: xorl %%eax,%%eax\n"
+ " jmp 1b\n"
+ "3: movb $1,%%al\n"
+ " jmp 1b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 4\n"
+ " .long 0b,2b\n"
+ ".previous"
+ :"=r" (n), "=D" (s), "=a" (res), "=c" (tmp)
+ :"0" (n), "1" (s), "2" (0), "3" (mask)
+ :"cc");
+ return res & mask;
+}
--- /dev/null
+/* ld script to make i386 Linux kernel
+ * Written by Martin Mares <mj@atrey.karlin.mff.cuni.cz>;
+ */
+OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386")
+OUTPUT_ARCH(i386)
+ENTRY(start)
+SECTIONS
+{
+ . = 0xFC400000 + 0x100000;
+ _text = .; /* Text and read-only data */
+ .text : {
+ *(.text)
+ *(.fixup)
+ *(.gnu.warning)
+ } = 0x9090
+ .text.lock : { *(.text.lock) } /* out-of-line lock text */
+
+ _etext = .; /* End of text section */
+
+ .rodata : { *(.rodata) *(.rodata.*) }
+ .kstrtab : { *(.kstrtab) }
+
+ . = ALIGN(16); /* Exception table */
+ __start___ex_table = .;
+ __ex_table : { *(__ex_table) }
+ __stop___ex_table = .;
+
+ __start___ksymtab = .; /* Kernel symbol table */
+ __ksymtab : { *(__ksymtab) }
+ __stop___ksymtab = .;
+
+ __start___kallsyms = .; /* All kernel symbols */
+ __kallsyms : { *(__kallsyms) }
+ __stop___kallsyms = .;
+
+ .data : { /* Data */
+ *(.data)
+ CONSTRUCTORS
+ }
+
+ _edata = .; /* End of data section */
+
+ . = ALIGN(8192); /* init_task */
+ .data.init_task : { *(.data.init_task) }
+
+ . = ALIGN(4096); /* Init code and data */
+ __init_begin = .;
+ .text.init : { *(.text.init) }
+ .data.init : { *(.data.init) }
+ . = ALIGN(16);
+ __setup_start = .;
+ .setup.init : { *(.setup.init) }
+ __setup_end = .;
+ __initcall_start = .;
+ .initcall.init : { *(.initcall.init) }
+ __initcall_end = .;
+ . = ALIGN(4096);
+ __init_end = .;
+
+ . = ALIGN(4096);
+ .data.page_aligned : { *(.data.idt) }
+
+ . = ALIGN(32);
+ .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+
+ __bss_start = .; /* BSS */
+ .bss : {
+ *(.bss)
+ }
+ _end = . ;
+
+ /* Sections to be discarded */
+ /DISCARD/ : {
+ *(.text.exit)
+ *(.data.exit)
+ *(.exitcall.exit)
+ }
+
+ /* Stabs debugging sections. */
+ .stab 0 : { *(.stab) }
+ .stabstr 0 : { *(.stabstr) }
+ .stab.excl 0 : { *(.stab.excl) }
+ .stab.exclstr 0 : { *(.stab.exclstr) }
+ .stab.index 0 : { *(.stab.index) }
+ .stab.indexstr 0 : { *(.stab.indexstr) }
+ .comment 0 : { *(.comment) }
+}
--- /dev/null
+/*
+ * User address space access functions.
+ *
+ * Copyright 1997 Andi Kleen <ak@muc.de>
+ * Copyright 1997 Linus Torvalds
+ * Copyright 2002 Andi Kleen <ak@suse.de>
+ */
+#include <asm/uaccess.h>
+
+/*
+ * Copy a null terminated string from userspace.
+ */
+
+#define __do_strncpy_from_user(dst,src,count,res) \
+do { \
+ long __d0, __d1, __d2; \
+ __asm__ __volatile__( \
+ " testq %1,%1\n" \
+ " jz 2f\n" \
+ "0: lodsb\n" \
+ " stosb\n" \
+ " testb %%al,%%al\n" \
+ " jz 1f\n" \
+ " decq %1\n" \
+ " jnz 0b\n" \
+ "1: subq %1,%0\n" \
+ "2:\n" \
+ ".section .fixup,\"ax\"\n" \
+ "3: movq %5,%0\n" \
+ " jmp 2b\n" \
+ ".previous\n" \
+ ".section __ex_table,\"a\"\n" \
+ " .align 8\n" \
+ " .quad 0b,3b\n" \
+ ".previous" \
+ : "=r"(res), "=c"(count), "=&a" (__d0), "=&S" (__d1), \
+ "=&D" (__d2) \
+ : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \
+ : "memory"); \
+} while (0)
+
+long
+__strncpy_from_user(char *dst, const char *src, long count)
+{
+ long res;
+ __do_strncpy_from_user(dst, src, count, res);
+ return res;
+}
+
+long
+strncpy_from_user(char *dst, const char *src, long count)
+{
+ long res = -EFAULT;
+ if (access_ok(VERIFY_READ, src, 1))
+ __do_strncpy_from_user(dst, src, count, res);
+ return res;
+}
+
+/*
+ * Zero Userspace
+ */
+
+unsigned long __clear_user(void *addr, unsigned long size)
+{
+ long __d0;
+ /* no memory constraint because it doesn't change any memory gcc knows
+ about */
+ asm volatile(
+ " testq %[size8],%[size8]\n"
+ " jz 4f\n"
+ "0: movq %[zero],(%[dst])\n"
+ " addq %[eight],%[dst]\n"
+ " decl %%ecx ; jnz 0b\n"
+ "4: movq %[size1],%%rcx\n"
+ " testl %%ecx,%%ecx\n"
+ " jz 2f\n"
+ "1: movb %b[zero],(%[dst])\n"
+ " incq %[dst]\n"
+ " decl %%ecx ; jnz 1b\n"
+ "2:\n"
+ ".section .fixup,\"ax\"\n"
+ "3: lea 0(%[size1],%[size8],8),%[size8]\n"
+ " jmp 2b\n"
+ ".previous\n"
+ ".section __ex_table,\"a\"\n"
+ " .align 8\n"
+ " .quad 0b,3b\n"
+ " .quad 1b,2b\n"
+ ".previous"
+ : [size8] "=c"(size), [dst] "=&D" (__d0)
+ : [size1] "r"(size & 7), "[size8]" (size / 8), "[dst] "(addr),
+ [zero] "r" (0UL), [eight] "r" (8UL));
+ return size;
+}
+
+
+unsigned long clear_user(void *to, unsigned long n)
+{
+ if (access_ok(VERIFY_WRITE, to, n))
+ return __clear_user(to, n);
+ return n;
+}
+
+/*
+ * Return the size of a string (including the ending 0)
+ *
+ * Return 0 on exception, a value greater than N if too long
+ */
+
+long strnlen_user(const char *s, long n)
+{
+ unsigned long res = 0;
+ char c;
+
+ if (!access_ok(VERIFY_READ, s, n))
+ return 0;
+
+ while (1) {
+ if (get_user(c, s))
+ return 0;
+ if (!c)
+ return res+1;
+ if (res>n)
+ return n+1;
+ res++;
+ s++;
+ }
+}
+
+unsigned long copy_in_user(void *to, const void *from, unsigned len)
+{
+ if (access_ok(VERIFY_WRITE, to, len) && access_ok(VERIFY_READ, from, len)) {
+ return copy_user_generic(to, from, len);
+ }
+ return len;
+}
--- /dev/null
+/* Excerpts written by Martin Mares <mj@atrey.karlin.mff.cuni.cz> */
+OUTPUT_FORMAT("elf64-x86-64", "elf64-x86-64", "elf64-x86-64")
+OUTPUT_ARCH(i386:x86-64)
+ENTRY(start)
+SECTIONS
+{
+ . = 0xFFFF840000100000;
+ _text = .; /* Text and read-only data */
+ .text : {
+ *(.text)
+ *(.fixup)
+ *(.gnu.warning)
+ } = 0x9090
+ .text.lock : { *(.text.lock) } /* out-of-line lock text */
+
+ _etext = .; /* End of text section */
+
+ .rodata : { *(.rodata) *(.rodata.*) }
+ .kstrtab : { *(.kstrtab) }
+
+ . = ALIGN(16); /* Exception table */
+ __start___ex_table = .;
+ __ex_table : { *(__ex_table) }
+ __stop___ex_table = .;
+
+ __start___ksymtab = .; /* Kernel symbol table */
+ __ksymtab : { *(__ksymtab) }
+ __stop___ksymtab = .;
+
+ __start___kallsyms = .; /* All kernel symbols */
+ __kallsyms : { *(__kallsyms) }
+ __stop___kallsyms = .;
+
+ .data : { /* Data */
+ *(.data)
+ CONSTRUCTORS
+ }
+
+ _edata = .; /* End of data section */
+
+ . = ALIGN(8192); /* init_task */
+ .data.init_task : { *(.data.init_task) }
+
+ . = ALIGN(4096); /* Init code and data */
+ __init_begin = .;
+ .text.init : { *(.text.init) }
+ .data.init : { *(.data.init) }
+ . = ALIGN(16);
+ __setup_start = .;
+ .setup.init : { *(.setup.init) }
+ __setup_end = .;
+ __initcall_start = .;
+ .initcall.init : { *(.initcall.init) }
+ __initcall_end = .;
+ . = ALIGN(4096);
+ __init_end = .;
+
+ . = ALIGN(4096);
+ .data.page_aligned : { *(.data.idt) }
+
+ . = ALIGN(32);
+ .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+
+ __bss_start = .; /* BSS */
+ .bss : {
+ *(.bss)
+ }
+ _end = . ;
+
+ /* Sections to be discarded */
+ /DISCARD/ : {
+ *(.text.exit)
+ *(.data.exit)
+ *(.exitcall.exit)
+ }
+
+ /* Stabs debugging sections. */
+ .stab 0 : { *(.stab) }
+ .stabstr 0 : { *(.stabstr) }
+ .stab.excl 0 : { *(.stab.excl) }
+ .stab.exclstr 0 : { *(.stab.exclstr) }
+ .stab.index 0 : { *(.stab.index) }
+ .stab.indexstr 0 : { *(.stab.indexstr) }
+ .comment 0 : { *(.comment) }
+}
+++ /dev/null
-/* ld script to make i386 Linux kernel
- * Written by Martin Mares <mj@atrey.karlin.mff.cuni.cz>;
- */
-OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386")
-OUTPUT_ARCH(i386)
-ENTRY(start)
-SECTIONS
-{
- . = 0xFC400000 + 0x100000;
- _text = .; /* Text and read-only data */
- .text : {
- *(.text)
- *(.fixup)
- *(.gnu.warning)
- } = 0x9090
- .text.lock : { *(.text.lock) } /* out-of-line lock text */
-
- _etext = .; /* End of text section */
-
- .rodata : { *(.rodata) *(.rodata.*) }
- .kstrtab : { *(.kstrtab) }
-
- . = ALIGN(16); /* Exception table */
- __start___ex_table = .;
- __ex_table : { *(__ex_table) }
- __stop___ex_table = .;
-
- __start___ksymtab = .; /* Kernel symbol table */
- __ksymtab : { *(__ksymtab) }
- __stop___ksymtab = .;
-
- __start___kallsyms = .; /* All kernel symbols */
- __kallsyms : { *(__kallsyms) }
- __stop___kallsyms = .;
-
- .data : { /* Data */
- *(.data)
- CONSTRUCTORS
- }
-
- _edata = .; /* End of data section */
-
- . = ALIGN(8192); /* init_task */
- .data.init_task : { *(.data.init_task) }
-
- . = ALIGN(4096); /* Init code and data */
- __init_begin = .;
- .text.init : { *(.text.init) }
- .data.init : { *(.data.init) }
- . = ALIGN(16);
- __setup_start = .;
- .setup.init : { *(.setup.init) }
- __setup_end = .;
- __initcall_start = .;
- .initcall.init : { *(.initcall.init) }
- __initcall_end = .;
- . = ALIGN(4096);
- __init_end = .;
-
- . = ALIGN(4096);
- .data.page_aligned : { *(.data.idt) }
-
- . = ALIGN(32);
- .data.cacheline_aligned : { *(.data.cacheline_aligned) }
-
- __bss_start = .; /* BSS */
- .bss : {
- *(.bss)
- }
- _end = . ;
-
- /* Sections to be discarded */
- /DISCARD/ : {
- *(.text.exit)
- *(.data.exit)
- *(.exitcall.exit)
- }
-
- /* Stabs debugging sections. */
- .stab 0 : { *(.stab) }
- .stabstr 0 : { *(.stabstr) }
- .stab.excl 0 : { *(.stab.excl) }
- .stab.exclstr 0 : { *(.stab.exclstr) }
- .stab.index 0 : { *(.stab.index) }
- .stab.indexstr 0 : { *(.stab.indexstr) }
- .comment 0 : { *(.comment) }
-}
#define CONFIG_ACPI_BOOT 1
#define CONFIG_PCI 1
-#define CONFIG_PCI_BIOS 1
#define CONFIG_PCI_DIRECT 1
+#if defined(__i386__)
+#define CONFIG_PCI_BIOS 1
+#endif
#define CONFIG_IDE 1
#define CONFIG_BLK_DEV_IDE 1
#define __STR(x) #x
#define STR(x) __STR(x)
+#if defined(__i386__)
+
#define SAVE_ALL \
"cld\n\t" \
"pushl %gs\n\t" \
"movl %edx,%fs\n\t" \
"movl %edx,%gs\n\t"
-#define IRQ_NAME2(nr) nr##_interrupt(void)
-#define IRQ_NAME(nr) IRQ_NAME2(IRQ##nr)
+#else
+
+#define SAVE_ALL
+
+#endif
#define BUILD_SMP_INTERRUPT(x,v) XBUILD_SMP_INTERRUPT(x,v)
#define XBUILD_SMP_INTERRUPT(x,v)\
__asm__( \
"\n"__ALIGN_STR"\n" \
SYMBOL_NAME_STR(x) ":\n\t" \
- "pushl $"#v"\n\t" \
+ "push"__OS" $"#v"\n\t" \
SAVE_ALL \
SYMBOL_NAME_STR(call_##x)":\n\t" \
"call "SYMBOL_NAME_STR(smp_##x)"\n\t" \
__asm__( \
"\n"__ALIGN_STR"\n" \
SYMBOL_NAME_STR(x) ":\n\t" \
- "pushl $"#v"\n\t" \
+ "push"__OS" $"#v"\n\t" \
SAVE_ALL \
- "movl %esp,%eax\n\t" \
- "pushl %eax\n\t" \
+ "mov %"__OP"sp,%"__OP"ax\n\t" \
+ "push %"__OP"ax\n\t" \
SYMBOL_NAME_STR(call_##x)":\n\t" \
"call "SYMBOL_NAME_STR(smp_##x)"\n\t" \
- "addl $4,%esp\n\t" \
+ "add $4,%"__OP"sp\n\t" \
"jmp ret_from_intr\n");
#define BUILD_COMMON_IRQ() \
"call " SYMBOL_NAME_STR(do_IRQ) "\n\t" \
"jmp ret_from_intr\n");
+#define IRQ_NAME2(nr) nr##_interrupt(void)
+#define IRQ_NAME(nr) IRQ_NAME2(IRQ##nr)
+
#define BUILD_IRQ(nr) \
asmlinkage void IRQ_NAME(nr); \
__asm__( \
"\n"__ALIGN_STR"\n" \
SYMBOL_NAME_STR(IRQ) #nr "_interrupt:\n\t" \
- "pushl $"#nr"\n\t" \
+ "push"__OS" $"#nr"\n\t" \
"jmp common_interrupt");
extern unsigned long prof_cpu_mask;
projects / xen.git / commitdiff