/*P:010 * A hypervisor allows multiple Operating Systems to run on a single machine. * To quote David Wheeler: "Any problem in computer science can be solved with * another layer of indirection." * * We keep things simple in two ways. First, we start with a normal Linux * kernel and insert a module (lg.ko) which allows us to run other Linux * kernels the same way we'd run processes. We call the first kernel the Host, * and the others the Guests. The program which sets up and configures Guests * (such as the example in Documentation/virtual/lguest/lguest.c) is called the * Launcher. * * Secondly, we only run specially modified Guests, not normal kernels: setting * CONFIG_LGUEST_GUEST to "y" compiles this file into the kernel so it knows * how to be a Guest at boot time. This means that you can use the same kernel * you boot normally (ie. as a Host) as a Guest. * * These Guests know that they cannot do privileged operations, such as disable * interrupts, and that they have to ask the Host to do such things explicitly. * This file consists of all the replacements for such low-level native * hardware operations: these special Guest versions call the Host. * * So how does the kernel know it's a Guest? We'll see that later, but let's * just say that we end up here where we replace the native functions various * "paravirt" structures with our Guest versions, then boot like normal. :*/ /* * Copyright (C) 2006, Rusty Russell IBM Corporation. * * 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, GOOD TITLE or * NON INFRINGEMENT. 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., 675 Mass Ave, Cambridge, MA 02139, USA. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* for struct machine_ops */ /*G:010 * Welcome to the Guest! * * The Guest in our tale is a simple creature: identical to the Host but * behaving in simplified but equivalent ways. In particular, the Guest is the * same kernel as the Host (or at least, built from the same source code). :*/ struct lguest_data lguest_data = { .hcall_status = { [0 ... LHCALL_RING_SIZE-1] = 0xFF }, .noirq_start = (u32)lguest_noirq_start, .noirq_end = (u32)lguest_noirq_end, .kernel_address = PAGE_OFFSET, .blocked_interrupts = { 1 }, /* Block timer interrupts */ .syscall_vec = SYSCALL_VECTOR, }; /*G:037 * async_hcall() is pretty simple: I'm quite proud of it really. We have a * ring buffer of stored hypercalls which the Host will run though next time we * do a normal hypercall. Each entry in the ring has 5 slots for the hypercall * arguments, and a "hcall_status" word which is 0 if the call is ready to go, * and 255 once the Host has finished with it. * * If we come around to a slot which hasn't been finished, then the table is * full and we just make the hypercall directly. This has the nice side * effect of causing the Host to run all the stored calls in the ring buffer * which empties it for next time! */ static void async_hcall(unsigned long call, unsigned long arg1, unsigned long arg2, unsigned long arg3, unsigned long arg4) { /* Note: This code assumes we're uniprocessor. */ static unsigned int next_call; unsigned long flags; /* * Disable interrupts if not already disabled: we don't want an * interrupt handler making a hypercall while we're already doing * one! */ local_irq_save(flags); if (lguest_data.hcall_status[next_call] != 0xFF) { /* Table full, so do normal hcall which will flush table. */ hcall(call, arg1, arg2, arg3, arg4); } else { lguest_data.hcalls[next_call].arg0 = call; lguest_data.hcalls[next_call].arg1 = arg1; lguest_data.hcalls[next_call].arg2 = arg2; lguest_data.hcalls[next_call].arg3 = arg3; lguest_data.hcalls[next_call].arg4 = arg4; /* Arguments must all be written before we mark it to go */ wmb(); lguest_data.hcall_status[next_call] = 0; if (++next_call == LHCALL_RING_SIZE) next_call = 0; } local_irq_restore(flags); } /*G:035 * Notice the lazy_hcall() above, rather than hcall(). This is our first real * optimization trick! * * When lazy_mode is set, it means we're allowed to defer all hypercalls and do * them as a batch when lazy_mode is eventually turned off. Because hypercalls * are reasonably expensive, batching them up makes sense. For example, a * large munmap might update dozens of page table entries: that code calls * paravirt_enter_lazy_mmu(), does the dozen updates, then calls * lguest_leave_lazy_mode(). * * So, when we're in lazy mode, we call async_hcall() to store the call for * future processing: */ static void lazy_hcall1(unsigned long call, unsigned long arg1) { if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) hcall(call, arg1, 0, 0, 0); else async_hcall(call, arg1, 0, 0, 0); } /* You can imagine what lazy_hcall2, 3 and 4 look like. :*/ static void lazy_hcall2(unsigned long call, unsigned long arg1, unsigned long arg2) { if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) hcall(call, arg1, arg2, 0, 0); else async_hcall(call, arg1, arg2, 0, 0); } static void lazy_hcall3(unsigned long call, unsigned long arg1, unsigned long arg2, unsigned long arg3) { if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) hcall(call, arg1, arg2, arg3, 0); else async_hcall(call, arg1, arg2, arg3, 0); } #ifdef CONFIG_X86_PAE static void lazy_hcall4(unsigned long call, unsigned long arg1, unsigned long arg2, unsigned long arg3, unsigned long arg4) { if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) hcall(call, arg1, arg2, arg3, arg4); else async_hcall(call, arg1, arg2, arg3, arg4); } #endif /*G:036 * When lazy mode is turned off, we issue the do-nothing hypercall to * flush any stored calls, and call the generic helper to reset the * per-cpu lazy mode variable. */ static void lguest_leave_lazy_mmu_mode(void) { hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0, 0); paravirt_leave_lazy_mmu(); } /* * We also catch the end of context switch; we enter lazy mode for much of * that too, so again we need to flush here. * * (Technically, this is lazy CPU mode, and normally we're in lazy MMU * mode, but unlike Xen, lguest doesn't care about the difference). */ static void lguest_end_context_switch(struct task_struct *next) { hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0, 0); paravirt_end_context_switch(next); } /*G:032 * After that diversion we return to our first native-instruction * replacements: four functions for interrupt control. * * The simplest way of implementing these would be to have "turn interrupts * off" and "turn interrupts on" hypercalls. Unfortunately, this is too slow: * these are by far the most commonly called functions of those we override. * * So instead we keep an "irq_enabled" field inside our "struct lguest_data", * which the Guest can update with a single instruction. The Host knows to * check there before it tries to deliver an interrupt. */ /* * save_flags() is expected to return the processor state (ie. "flags"). The * flags word contains all kind of stuff, but in practice Linux only cares * about the interrupt flag. Our "save_flags()" just returns that. */ static unsigned long save_fl(void) { return lguest_data.irq_enabled; } /* Interrupts go off... */ static void irq_disable(void) { lguest_data.irq_enabled = 0; } /* * Let's pause a moment. Remember how I said these are called so often? * Jeremy Fitzhardinge optimized them so hard early in 2009 that he had to * break some rules. In particular, these functions are assumed to save their * own registers if they need to: normal C functions assume they can trash the * eax register. To use normal C functions, we use * PV_CALLEE_SAVE_REGS_THUNK(), which pushes %eax onto the stack, calls the * C function, then restores it. */ PV_CALLEE_SAVE_REGS_THUNK(save_fl); PV_CALLEE_SAVE_REGS_THUNK(irq_disable); /*:*/ /* These are in i386_head.S */ extern void lg_irq_enable(void); extern void lg_restore_fl(unsigned long flags); /*M:003 * We could be more efficient in our checking of outstanding interrupts, rather * than using a branch. One way would be to put the "irq_enabled" field in a * page by itself, and have the Host write-protect it when an interrupt comes * in when irqs are disabled. There will then be a page fault as soon as * interrupts are re-enabled. * * A better method is to implement soft interrupt disable generally for x86: * instead of disabling interrupts, we set a flag. If an interrupt does come * in, we then disable them for real. This is uncommon, so we could simply use * a hypercall for interrupt control and not worry about efficiency. :*/ /*G:034 * The Interrupt Descriptor Table (IDT). * * The IDT tells the processor what to do when an interrupt comes in. Each * entry in the table is a 64-bit descriptor: this holds the privilege level, * address of the handler, and... well, who cares? The Guest just asks the * Host to make the change anyway, because the Host controls the real IDT. */ static void lguest_write_idt_entry(gate_desc *dt, int entrynum, const gate_desc *g) { /* * The gate_desc structure is 8 bytes long: we hand it to the Host in * two 32-bit chunks. The whole 32-bit kernel used to hand descriptors * around like this; typesafety wasn't a big concern in Linux's early * years. */ u32 *desc = (u32 *)g; /* Keep the local copy up to date. */ native_write_idt_entry(dt, entrynum, g); /* Tell Host about this new entry. */ hcall(LHCALL_LOAD_IDT_ENTRY, entrynum, desc[0], desc[1], 0); } /* * Changing to a different IDT is very rare: we keep the IDT up-to-date every * time it is written, so we can simply loop through all entries and tell the * Host about them. */ static void lguest_load_idt(const struct desc_ptr *desc) { unsigned int i; struct desc_struct *idt = (void *)desc->address; for (i = 0; i < (desc->size+1)/8; i++) hcall(LHCALL_LOAD_IDT_ENTRY, i, idt[i].a, idt[i].b, 0); } /* * The Global Descriptor Table. * * The Intel architecture defines another table, called the Global Descriptor * Table (GDT). You tell the CPU where it is (and its size) using the "lgdt" * instruction, and then several other instructions refer to entries in the * table. There are three entries which the Switcher needs, so the Host simply * controls the entire thing and the Guest asks it to make changes using the * LOAD_GDT hypercall. * * This is the exactly like the IDT code. */ static void lguest_load_gdt(const struct desc_ptr *desc) { unsigned int i; struct desc_struct *gdt = (void *)desc->address; for (i = 0; i < (desc->size+1)/8; i++) hcall(LHCALL_LOAD_GDT_ENTRY, i, gdt[i].a, gdt[i].b, 0); } /* * For a single GDT entry which changes, we simply change our copy and * then tell the host about it. */ static void lguest_write_gdt_entry(struct desc_struct *dt, int entrynum, const void *desc, int type) { native_write_gdt_entry(dt, entrynum, desc, type); /* Tell Host about this new entry. */ hcall(LHCALL_LOAD_GDT_ENTRY, entrynum, dt[entrynum].a, dt[entrynum].b, 0); } /* * There are three "thread local storage" GDT entries which change * on every context switch (these three entries are how glibc implements * __thread variables). As an optimization, we have a hypercall * specifically for this case. * * Wouldn't it be nicer to have a general LOAD_GDT_ENTRIES hypercall * which took a range of entries? */ static void lguest_load_tls(struct thread_struct *t, unsigned int cpu) { /* * There's one problem which normal hardware doesn't have: the Host * can't handle us removing entries we're currently using. So we clear * the GS register here: if it's needed it'll be reloaded anyway. */ lazy_load_gs(0); lazy_hcall2(LHCALL_LOAD_TLS, __pa(&t->tls_array), cpu); } /*G:038 * That's enough excitement for now, back to ploughing through each of the * different pv_ops structures (we're about 1/3 of the way through). * * This is the Local Descriptor Table, another weird Intel thingy. Linux only * uses this for some strange applications like Wine. We don't do anything * here, so they'll get an informative and friendly Segmentation Fault. */ static void lguest_set_ldt(const void *addr, unsigned entries) { } /* * This loads a GDT entry into the "Task Register": that entry points to a * structure called the Task State Segment. Some comments scattered though the * kernel code indicate that this used for task switching in ages past, along * with blood sacrifice and astrology. * * Now there's nothing interesting in here that we don't get told elsewhere. * But the native version uses the "ltr" instruction, which makes the Host * complain to the Guest about a Segmentation Fault and it'll oops. So we * override the native version with a do-nothing version. */ static void lguest_load_tr_desc(void) { } /* * The "cpuid" instruction is a way of querying both the CPU identity * (manufacturer, model, etc) and its features. It was introduced before the * Pentium in 1993 and keeps getting extended by both Intel, AMD and others. * As you might imagine, after a decade and a half this treatment, it is now a * giant ball of hair. Its entry in the current Intel manual runs to 28 pages. * * This instruction even it has its own Wikipedia entry. The Wikipedia entry * has been translated into 5 languages. I am not making this up! * * We could get funky here and identify ourselves as "GenuineLguest", but * instead we just use the real "cpuid" instruction. Then I pretty much turned * off feature bits until the Guest booted. (Don't say that: you'll damage * lguest sales!) Shut up, inner voice! (Hey, just pointing out that this is * hardly future proof.) No one's listening! They don't like you anyway, * parenthetic weirdo! * * Replacing the cpuid so we can turn features off is great for the kernel, but * anyone (including userspace) can just use the raw "cpuid" instruction and * the Host won't even notice since it isn't privileged. So we try not to get * too worked up about it. */ static void lguest_cpuid(unsigned int *ax, unsigned int *bx, unsigned int *cx, unsigned int *dx) { int function = *ax; native_cpuid(ax, bx, cx, dx); switch (function) { /* * CPUID 0 gives the highest legal CPUID number (and the ID string). * We futureproof our code a little by sticking to known CPUID values. */ case 0: if (*ax > 5) *ax = 5; break; /* * CPUID 1 is a basic feature request. * * CX: we only allow kernel to see SSE3, CMPXCHG16B and SSSE3 * DX: SSE, SSE2, FXSR, MMX, CMOV, CMPXCHG8B, TSC, FPU and PAE. */ case 1: *cx &= 0x00002201; *dx &= 0x07808151; /* * The Host can do a nice optimization if it knows that the * kernel mappings (addresses above 0xC0000000 or whatever * PAGE_OFFSET is set to) haven't changed. But Linux calls * flush_tlb_user() for both user and kernel mappings unless * the Page Global Enable (PGE) feature bit is set. */ *dx |= 0x00002000; /* * We also lie, and say we're family id 5. 6 or greater * leads to a rdmsr in early_init_intel which we can't handle. * Family ID is returned as bits 8-12 in ax. */ *ax &= 0xFFFFF0FF; *ax |= 0x00000500; break; /* * 0x80000000 returns the highest Extended Function, so we futureproof * like we do above by limiting it to known fields. */ case 0x80000000: if (*ax > 0x80000008) *ax = 0x80000008; break; /* * PAE systems can mark pages as non-executable. Linux calls this the * NX bit. Intel calls it XD (eXecute Disable), AMD EVP (Enhanced * Virus Protection). We just switch it off here, since we don't * support it. */ case 0x80000001: *dx &= ~(1 << 20); break; } } /* * Intel has four control registers, imaginatively named cr0, cr2, cr3 and cr4. * I assume there's a cr1, but it hasn't bothered us yet, so we'll not bother * it. The Host needs to know when the Guest wants to change them, so we have * a whole series of functions like read_cr0() and write_cr0(). * * We start with cr0. cr0 allows you to turn on and off all kinds of basic * features, but Linux only really cares about one: the horrifically-named Task * Switched (TS) bit at bit 3 (ie. 8) * * What does the TS bit do? Well, it causes the CPU to trap (interrupt 7) if * the floating point unit is used. Which allows us to restore FPU state * lazily after a task switch, and Linux uses that gratefully, but wouldn't a * name like "FPUTRAP bit" be a little less cryptic? * * We store cr0 locally because the Host never changes it. The Guest sometimes * wants to read it and we'd prefer not to bother the Host unnecessarily. */ static unsigned long current_cr0; static void lguest_write_cr0(unsigned long val) { lazy_hcall1(LHCALL_TS, val & X86_CR0_TS); current_cr0 = val; } static unsigned long lguest_read_cr0(void) { return current_cr0; } /* * Intel provided a special instruction to clear the TS bit for people too cool * to use write_cr0() to do it. This "clts" instruction is faster, because all * the vowels have been optimized out. */ static void lguest_clts(void) { lazy_hcall1(LHCALL_TS, 0); current_cr0 &= ~X86_CR0_TS; } /* * cr2 is the virtual address of the last page fault, which the Guest only ever * reads. The Host kindly writes this into our "struct lguest_data", so we * just read it out of there. */ static unsigned long lguest_read_cr2(void) { return lguest_data.cr2; } /* See lguest_set_pte() below. */ static bool cr3_changed = false; static unsigned long current_cr3; /* * cr3 is the current toplevel pagetable page: the principle is the same as * cr0. Keep a local copy, and tell the Host when it changes. */ static void lguest_write_cr3(unsigned long cr3) { lazy_hcall1(LHCALL_NEW_PGTABLE, cr3); current_cr3 = cr3; /* These two page tables are simple, linear, and used during boot */ if (cr3 != __pa(swapper_pg_dir) && cr3 != __pa(initial_page_table)) cr3_changed = true; } static unsigned long lguest_read_cr3(void) { return current_cr3; } /* cr4 is used to enable and disable PGE, but we don't care. */ static unsigned long lguest_read_cr4(void) { return 0; } static void lguest_write_cr4(unsigned long val) { } /* * Page Table Handling. * * Now would be a good time to take a rest and grab a coffee or similarly * relaxing stimulant. The easy parts are behind us, and the trek gradually * winds uphill from here. * * Quick refresher: memory is divided into "pages" of 4096 bytes each. The CPU * maps virtual addresses to physical addresses using "page tables". We could * use one huge index of 1 million entries: each address is 4 bytes, so that's * 1024 pages just to hold the page tables. But since most virtual addresses * are unused, we use a two level index which saves space. The cr3 register * contains the physical address of the top level "page directory" page, which * contains physical addresses of up to 1024 second-level pages. Each of these * second level pages contains up to 1024 physical addresses of actual pages, * or Page Table Entries (PTEs). * * Here's a diagram, where arrows indicate physical addresses: * * cr3 ---> +---------+ * | --------->+---------+ * | | | PADDR1 | * Mid-level | | PADDR2 | * (PMD) page | | | * | | Lower-level | * | | (PTE) page | * | | | | * .... .... * * So to convert a virtual address to a physical address, we look up the top * level, which points us to the second level, which gives us the physical * address of that page. If the top level entry was not present, or the second * level entry was not present, then the virtual address is invalid (we * say "the page was not mapped"). * * Put another way, a 32-bit virtual address is divided up like so: * * 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 * |<---- 10 bits ---->|<---- 10 bits ---->|<------ 12 bits ------>| * Index into top Index into second Offset within page * page directory page pagetable page * * Now, unfortunately, this isn't the whole story: Intel added Physical Address * Extension (PAE) to allow 32 bit systems to use 64GB of memory (ie. 36 bits). * These are held in 64-bit page table entries, so we can now only fit 512 * entries in a page, and the neat three-level tree breaks down. * * The result is a four level page table: * * cr3 --> [ 4 Upper ] * [ Level ] * [ Entries ] * [(PUD Page)]---> +---------+ * | --------->+---------+ * | | | PADDR1 | * Mid-level | | PADDR2 | * (PMD) page | | | * | | Lower-level | * | | (PTE) page | * | | | | * .... .... * * * And the virtual address is decoded as: * * 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 * |<-2->|<--- 9 bits ---->|<---- 9 bits --->|<------ 12 bits ------>| * Index into Index into mid Index into lower Offset within page * top entries directory page pagetable page * * It's too hard to switch between these two formats at runtime, so Linux only * supports one or the other depending on whether CONFIG_X86_PAE is set. Many * distributions turn it on, and not just for people with silly amounts of * memory: the larger PTE entries allow room for the NX bit, which lets the * kernel disable execution of pages and increase security. * * This was a problem for lguest, which couldn't run on these distributions; * then Matias Zabaljauregui figured it all out and implemented it, and only a * handful of puppies were crushed in the process! * * Back to our point: the kernel spends a lot of time changing both the * top-level page directory and lower-level pagetable pages. The Guest doesn't * know physical addresses, so while it maintains these page tables exactly * like normal, it also needs to keep the Host informed whenever it makes a * change: the Host will create the real page tables based on the Guests'. */ /* * The Guest calls this after it has set a second-level entry (pte), ie. to map * a page into a process' address space. We tell the Host the toplevel and * address this corresponds to. The Guest uses one pagetable per process, so * we need to tell the Host which one we're changing (mm->pgd). */ static void lguest_pte_update(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { #ifdef CONFIG_X86_PAE /* PAE needs to hand a 64 bit page table entry, so it uses two args. */ lazy_hcall4(LHCALL_SET_PTE, __pa(mm->pgd), addr, ptep->pte_low, ptep->pte_high); #else lazy_hcall3(LHCALL_SET_PTE, __pa(mm->pgd), addr, ptep->pte_low); #endif } /* This is the "set and update" combo-meal-deal version. */ static void lguest_set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pteval) { native_set_pte(ptep, pteval); lguest_pte_update(mm, addr, ptep); } /* * The Guest calls lguest_set_pud to set a top-level entry and lguest_set_pmd * to set a middle-level entry when PAE is activated. * * Again, we set the entry then tell the Host which page we changed, * and the index of the entry we changed. */ #ifdef CONFIG_X86_PAE static void lguest_set_pud(pud_t *pudp, pud_t pudval) { native_set_pud(pudp, pudval); /* 32 bytes aligned pdpt address and the index. */ lazy_hcall2(LHCALL_SET_PGD, __pa(pudp) & 0xFFFFFFE0, (__pa(pudp) & 0x1F) / sizeof(pud_t)); } static void lguest_set_pmd(pmd_t *pmdp, pmd_t pmdval) { native_set_pmd(pmdp, pmdval); lazy_hcall2(LHCALL_SET_PMD, __pa(pmdp) & PAGE_MASK, (__pa(pmdp) & (PAGE_SIZE - 1)) / sizeof(pmd_t)); } #else /* The Guest calls lguest_set_pmd to set a top-level entry when !PAE. */ static void lguest_set_pmd(pmd_t *pmdp, pmd_t pmdval) { native_set_pmd(pmdp, pmdval); lazy_hcall2(LHCALL_SET_PGD, __pa(pmdp) & PAGE_MASK, (__pa(pmdp) & (PAGE_SIZE - 1)) / sizeof(pmd_t)); } #endif /* * There are a couple of legacy places where the kernel sets a PTE, but we * don't know the top level any more. This is useless for us, since we don't * know which pagetable is changing or what address, so we just tell the Host * to forget all of them. Fortunately, this is very rare. * * ... except in early boot when the kernel sets up the initial pagetables, * which makes booting astonishingly slow: 48 seconds! So we don't even tell * the Host anything changed until we've done the first real page table switch, * which brings boot back to 4.3 seconds. */ static void lguest_set_pte(pte_t *ptep, pte_t pteval) { native_set_pte(ptep, pteval); if (cr3_changed) lazy_hcall1(LHCALL_FLUSH_TLB, 1); } #ifdef CONFIG_X86_PAE /* * With 64-bit PTE values, we need to be careful setting them: if we set 32 * bits at a time, the hardware could see a weird half-set entry. These * versions ensure we update all 64 bits at once. */ static void lguest_set_pte_atomic(pte_t *ptep, pte_t pte) { native_set_pte_atomic(ptep, pte); if (cr3_changed) lazy_hcall1(LHCALL_FLUSH_TLB, 1); } static void lguest_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { native_pte_clear(mm, addr, ptep); lguest_pte_update(mm, addr, ptep); } static void lguest_pmd_clear(pmd_t *pmdp) { lguest_set_pmd(pmdp, __pmd(0)); } #endif /* * Unfortunately for Lguest, the pv_mmu_ops for page tables were based on * native page table operations. On native hardware you can set a new page * table entry whenever you want, but if you want to remove one you have to do * a TLB flush (a TLB is a little cache of page table entries kept by the CPU). * * So the lguest_set_pte_at() and lguest_set_pmd() functions above are only * called when a valid entry is written, not when it's removed (ie. marked not * present). Instead, this is where we come when the Guest wants to remove a * page table entry: we tell the Host to set that entry to 0 (ie. the present * bit is zero). */ static void lguest_flush_tlb_single(unsigned long addr) { /* Simply set it to zero: if it was not, it will fault back in. */ lazy_hcall3(LHCALL_SET_PTE, current_cr3, addr, 0); } /* * This is what happens after the Guest has removed a large number of entries. * This tells the Host that any of the page table entries for userspace might * have changed, ie. virtual addresses below PAGE_OFFSET. */ static void lguest_flush_tlb_user(void) { lazy_hcall1(LHCALL_FLUSH_TLB, 0); } /* * This is called when the kernel page tables have changed. That's not very * common (unless the Guest is using highmem, which makes the Guest extremely * slow), so it's worth separating this from the user flushing above. */ static void lguest_flush_tlb_kernel(void) { lazy_hcall1(LHCALL_FLUSH_TLB, 1); } /* * The Unadvanced Programmable Interrupt Controller. * * This is an attempt to implement the simplest possible interrupt controller. * I spent some time looking though routines like set_irq_chip_and_handler, * set_irq_chip_and_handler_name, set_irq_chip_data and set_phasers_to_stun and * I *think* this is as simple as it gets. * * We can tell the Host what interrupts we want blocked ready for using the * lguest_data.interrupts bitmap, so disabling (aka "masking") them is as * simple as setting a bit. We don't actually "ack" interrupts as such, we * just mask and unmask them. I wonder if we should be cleverer? */ static void disable_lguest_irq(struct irq_data *data) { set_bit(data->irq, lguest_data.blocked_interrupts); } static void enable_lguest_irq(struct irq_data *data) { clear_bit(data->irq, lguest_data.blocked_interrupts); } /* This structure describes the lguest IRQ controller. */ static struct irq_chip lguest_irq_controller = { .name = "lguest", .irq_mask = disable_lguest_irq, .irq_mask_ack = disable_lguest_irq, .irq_unmask = enable_lguest_irq, }; /* * This sets up the Interrupt Descriptor Table (IDT) entry for each hardware * interrupt (except 128, which is used for system calls), and then tells the * Linux infrastructure that each interrupt is controlled by our level-based * lguest interrupt controller. */ static void __init lguest_init_IRQ(void) { unsigned int i; for (i = FIRST_EXTERNAL_VECTOR; i < NR_VECTORS; i++) { /* Some systems map "vectors" to interrupts weirdly. Not us! */ __this_cpu_write(vector_irq[i], i - FIRST_EXTERNAL_VECTOR); if (i != SYSCALL_VECTOR) set_intr_gate(i, interrupt[i - FIRST_EXTERNAL_VECTOR]); } /* * This call is required to set up for 4k stacks, where we have * separate stacks for hard and soft interrupts. */ irq_ctx_init(smp_processor_id()); } /* * With CONFIG_SPARSE_IRQ, interrupt descriptors are allocated as-needed, so * rather than set them in lguest_init_IRQ we are called here every time an * lguest device needs an interrupt. * * FIXME: irq_alloc_desc_at() can fail due to lack of memory, we should * pass that up! */ void lguest_setup_irq(unsigned int irq) { irq_alloc_desc_at(irq, 0); irq_set_chip_and_handler_name(irq, &lguest_irq_controller, handle_level_irq, "level"); } /* * Time. * * It would be far better for everyone if the Guest had its own clock, but * until then the Host gives us the time on every interrupt. */ static unsigned long lguest_get_wallclock(void) { return lguest_data.time.tv_sec; } /* * The TSC is an Intel thing called the Time Stamp Counter. The Host tells us * what speed it runs at, or 0 if it's unusable as a reliable clock source. * This matches what we want here: if we return 0 from this function, the x86 * TSC clock will give up and not register itself. */ static unsigned long lguest_tsc_khz(void) { return lguest_data.tsc_khz; } /* * If we can't use the TSC, the kernel falls back to our lower-priority * "lguest_clock", where we read the time value given to us by the Host. */ static cycle_t lguest_clock_read(struct clocksource *cs) { unsigned long sec, nsec; /* * Since the time is in two parts (seconds and nanoseconds), we risk * reading it just as it's changing from 99 & 0.999999999 to 100 and 0, * and getting 99 and 0. As Linux tends to come apart under the stress * of time travel, we must be careful: */ do { /* First we read the seconds part. */ sec = lguest_data.time.tv_sec; /* * This read memory barrier tells the compiler and the CPU that * this can't be reordered: we have to complete the above * before going on. */ rmb(); /* Now we read the nanoseconds part. */ nsec = lguest_data.time.tv_nsec; /* Make sure we've done that. */ rmb(); /* Now if the seconds part has changed, try again. */ } while (unlikely(lguest_data.time.tv_sec != sec)); /* Our lguest clock is in real nanoseconds. */ return sec*1000000000ULL + nsec; } /* This is the fallback clocksource: lower priority than the TSC clocksource. */ static struct clocksource lguest_clock = { .name = "lguest", .rating = 200, .read = lguest_clock_read, .mask = CLOCKSOURCE_MASK(64), .flags = CLOCK_SOURCE_IS_CONTINUOUS, }; /* * We also need a "struct clock_event_device": Linux asks us to set it to go * off some time in the future. Actually, James Morris figured all this out, I * just applied the patch. */ static int lguest_clockevent_set_next_event(unsigned long delta, struct clock_event_device *evt) { /* FIXME: I don't think this can ever happen, but James tells me he had * to put this code in. Maybe we should remove it now. Anyone? */ if (delta < LG_CLOCK_MIN_DELTA) { if (printk_ratelimit()) printk(KERN_DEBUG "%s: small delta %lu ns\n", __func__, delta); return -ETIME; } /* Please wake us this far in the future. */ hcall(LHCALL_SET_CLOCKEVENT, delta, 0, 0, 0); return 0; } static void lguest_clockevent_set_mode(enum clock_event_mode mode, struct clock_event_device *evt) { switch (mode) { case CLOCK_EVT_MODE_UNUSED: case CLOCK_EVT_MODE_SHUTDOWN: /* A 0 argument shuts the clock down. */ hcall(LHCALL_SET_CLOCKEVENT, 0, 0, 0, 0); break; case CLOCK_EVT_MODE_ONESHOT: /* This is what we expect. */ break; case CLOCK_EVT_MODE_PERIODIC: BUG(); case CLOCK_EVT_MODE_RESUME: break; } } /* This describes our primitive timer chip. */ static struct clock_event_device lguest_clockevent = { .name = "lguest", .features = CLOCK_EVT_FEAT_ONESHOT, .set_next_event = lguest_clockevent_set_next_event, .set_mode = lguest_clockevent_set_mode, .rating = INT_MAX, .mult = 1, .shift = 0, .min_delta_ns = LG_CLOCK_MIN_DELTA, .max_delta_ns = LG_CLOCK_MAX_DELTA, }; /* * This is the Guest timer interrupt handler (hardware interrupt 0). We just * call the clockevent infrastructure and it does whatever needs doing. */ static void lguest_time_irq(unsigned int irq, struct irq_desc *desc) { unsigned long flags; /* Don't interrupt us while this is running. */ local_irq_save(flags); lguest_clockevent.event_handler(&lguest_clockevent); local_irq_restore(flags); } /* * At some point in the boot process, we get asked to set up our timing * infrastructure. The kernel doesn't expect timer interrupts before this, but * we cleverly initialized the "blocked_interrupts" field of "struct * lguest_data" so that timer interrupts were blocked until now. */ static void lguest_time_init(void) { /* Set up the timer interrupt (0) to go to our simple timer routine */ lguest_setup_irq(0); irq_set_handler(0, lguest_time_irq); clocksource_register_hz(&lguest_clock, NSEC_PER_SEC); /* We can't set cpumask in the initializer: damn C limitations! Set it * here and register our timer device. */ lguest_clockevent.cpumask = cpumask_of(0); clockevents_register_device(&lguest_clockevent); /* Finally, we unblock the timer interrupt. */ clear_bit(0, lguest_data.blocked_interrupts); } /* * Miscellaneous bits and pieces. * * Here is an oddball collection of functions which the Guest needs for things * to work. They're pretty simple. */ /* * The Guest needs to tell the Host what stack it expects traps to use. For * native hardware, this is part of the Task State Segment mentioned above in * lguest_load_tr_desc(), but to help hypervisors there's this special call. * * We tell the Host the segment we want to use (__KERNEL_DS is the kernel data * segment), the privilege level (we're privilege level 1, the Host is 0 and * will not tolerate us trying to use that), the stack pointer, and the number * of pages in the stack. */ static void lguest_load_sp0(struct tss_struct *tss, struct thread_struct *thread) { lazy_hcall3(LHCALL_SET_STACK, __KERNEL_DS | 0x1, thread->sp0, THREAD_SIZE / PAGE_SIZE); } /* Let's just say, I wouldn't do debugging under a Guest. */ static void lguest_set_debugreg(int regno, unsigned long value) { /* FIXME: Implement */ } /* * There are times when the kernel wants to make sure that no memory writes are * caught in the cache (that they've all reached real hardware devices). This * doesn't matter for the Guest which has virtual hardware. * * On the Pentium 4 and above, cpuid() indicates that the Cache Line Flush * (clflush) instruction is available and the kernel uses that. Otherwise, it * uses the older "Write Back and Invalidate Cache" (wbinvd) instruction. * Unlike clflush, wbinvd can only be run at privilege level 0. So we can * ignore clflush, but replace wbinvd. */ static void lguest_wbinvd(void) { } /* * If the Guest expects to have an Advanced Programmable Interrupt Controller, * we play dumb by ignoring writes and returning 0 for reads. So it's no * longer Programmable nor Controlling anything, and I don't think 8 lines of * code qualifies for Advanced. It will also never interrupt anything. It * does, however, allow us to get through the Linux boot code. */ #ifdef CONFIG_X86_LOCAL_APIC static void lguest_apic_write(u32 reg, u32 v) { } static u32 lguest_apic_read(u32 reg) { return 0; } static u64 lguest_apic_icr_read(void) { return 0; } static void lguest_apic_icr_write(u32 low, u32 id) { /* Warn to see if there's any stray references */ WARN_ON(1); } static void lguest_apic_wait_icr_idle(void) { return; } static u32 lguest_apic_safe_wait_icr_idle(void) { return 0; } static void set_lguest_basic_apic_ops(void) { apic->read = lguest_apic_read; apic->write = lguest_apic_write; apic->icr_read = lguest_apic_icr_read; apic->icr_write = lguest_apic_icr_write; apic->wait_icr_idle = lguest_apic_wait_icr_idle; apic->safe_wait_icr_idle = lguest_apic_safe_wait_icr_idle; }; #endif /* STOP! Until an interrupt comes in. */ static void lguest_safe_halt(void) { hcall(LHCALL_HALT, 0, 0, 0, 0); } /* * The SHUTDOWN hypercall takes a string to describe what's happening, and * an argument which says whether this to restart (reboot) the Guest or not. * * Note that the Host always prefers that the Guest speak in physical addresses * rather than virtual addresses, so we use __pa() here. */ static void lguest_power_off(void) { hcall(LHCALL_SHUTDOWN, __pa("Power down"), LGUEST_SHUTDOWN_POWEROFF, 0, 0); } /* * Panicing. * * Don't. But if you did, this is what happens. */ static int lguest_panic(struct notifier_block *nb, unsigned long l, void *p) { hcall(LHCALL_SHUTDOWN, __pa(p), LGUEST_SHUTDOWN_POWEROFF, 0, 0); /* The hcall won't return, but to keep gcc happy, we're "done". */ return NOTIFY_DONE; } static struct notifier_block paniced = { .notifier_call = lguest_panic }; /* Setting up memory is fairly easy. */ static __init char *lguest_memory_setup(void) { /* * The Linux bootloader header contains an "e820" memory map: the * Launcher populated the first entry with our memory limit. */ e820_add_region(boot_params.e820_map[0].addr, boot_params.e820_map[0].size, boot_params.e820_map[0].type); /* This string is for the boot messages. */ return "LGUEST"; } /* * We will eventually use the virtio console device to produce console output, * but before that is set up we use LHCALL_NOTIFY on normal memory to produce * console output. */ static __init int early_put_chars(u32 vtermno, const char *buf, int count) { char scratch[17]; unsigned int len = count; /* We use a nul-terminated string, so we make a copy. Icky, huh? */ if (len > sizeof(scratch) - 1) len = sizeof(scratch) - 1; scratch[len] = '\0'; memcpy(scratch, buf, len); hcall(LHCALL_NOTIFY, __pa(scratch), 0, 0, 0); /* This routine returns the number of bytes actually written. */ return len; } /* * Rebooting also tells the Host we're finished, but the RESTART flag tells the * Launcher to reboot us. */ static void lguest_restart(char *reason) { hcall(LHCALL_SHUTDOWN, __pa(reason), LGUEST_SHUTDOWN_RESTART, 0, 0); } /*G:050 * Patching (Powerfully Placating Performance Pedants) * * We have already seen that pv_ops structures let us replace simple native * instructions with calls to the appropriate back end all throughout the * kernel. This allows the same kernel to run as a Guest and as a native * kernel, but it's slow because of all the indirect branches. * * Remember that David Wheeler quote about "Any problem in computer science can * be solved with another layer of indirection"? The rest of that quote is * "... But that usually will create another problem." This is the first of * those problems. * * Our current solution is to allow the paravirt back end to optionally patch * over the indirect calls to replace them with something more efficient. We * patch two of the simplest of the most commonly called functions: disable * interrupts and save interrupts. We usually have 6 or 10 bytes to patch * into: the Guest versions of these operations are small enough that we can * fit comfortably. * * First we need assembly templates of each of the patchable Guest operations, * and these are in i386_head.S. */ /*G:060 We construct a table from the assembler templates: */ static const struct lguest_insns { const char *start, *end; } lguest_insns[] = { [PARAVIRT_PATCH(pv_irq_ops.irq_disable)] = { lgstart_cli, lgend_cli }, [PARAVIRT_PATCH(pv_irq_ops.save_fl)] = { lgstart_pushf, lgend_pushf }, }; /* * Now our patch routine is fairly simple (based on the native one in * paravirt.c). If we have a replacement, we copy it in and return how much of * the available space we used. */ static unsigned lguest_patch(u8 type, u16 clobber, void *ibuf, unsigned long addr, unsigned len) { unsigned int insn_len; /* Don't do anything special if we don't have a replacement */ if (type >= ARRAY_SIZE(lguest_insns) || !lguest_insns[type].start) return paravirt_patch_default(type, clobber, ibuf, addr, len); insn_len = lguest_insns[type].end - lguest_insns[type].start; /* Similarly if it can't fit (doesn't happen, but let's be thorough). */ if (len < insn_len) return paravirt_patch_default(type, clobber, ibuf, addr, len); /* Copy in our instructions. */ memcpy(ibuf, lguest_insns[type].start, insn_len); return insn_len; } /*G:029 * Once we get to lguest_init(), we know we're a Guest. The various * pv_ops structures in the kernel provide points for (almost) every routine we * have to override to avoid privileged instructions. */ __init void lguest_init(void) { /* We're under lguest. */ pv_info.name = "lguest"; /* Paravirt is enabled. */ pv_info.paravirt_enabled = 1; /* We're running at privilege level 1, not 0 as normal. */ pv_info.kernel_rpl = 1; /* Everyone except Xen runs with this set. */ pv_info.shared_kernel_pmd = 1; /* * We set up all the lguest overrides for sensitive operations. These * are detailed with the operations themselves. */ /* Interrupt-related operations */ pv_irq_ops.save_fl = PV_CALLEE_SAVE(save_fl); pv_irq_ops.restore_fl = __PV_IS_CALLEE_SAVE(lg_restore_fl); pv_irq_ops.irq_disable = PV_CALLEE_SAVE(irq_disable); pv_irq_ops.irq_enable = __PV_IS_CALLEE_SAVE(lg_irq_enable); pv_irq_ops.safe_halt = lguest_safe_halt; /* Setup operations */ pv_init_ops.patch = lguest_patch; /* Intercepts of various CPU instructions */ pv_cpu_ops.load_gdt = lguest_load_gdt; pv_cpu_ops.cpuid = lguest_cpuid; pv_cpu_ops.load_idt = lguest_load_idt; pv_cpu_ops.iret = lguest_iret; pv_cpu_ops.load_sp0 = lguest_load_sp0; pv_cpu_ops.load_tr_desc = lguest_load_tr_desc; pv_cpu_ops.set_ldt = lguest_set_ldt; pv_cpu_ops.load_tls = lguest_load_tls; pv_cpu_ops.set_debugreg = lguest_set_debugreg; pv_cpu_ops.clts = lguest_clts; pv_cpu_ops.read_cr0 = lguest_read_cr0; pv_cpu_ops.write_cr0 = lguest_write_cr0; pv_cpu_ops.read_cr4 = lguest_read_cr4; pv_cpu_ops.write_cr4 = lguest_write_cr4; pv_cpu_ops.write_gdt_entry = lguest_write_gdt_entry; pv_cpu_ops.write_idt_entry = lguest_write_idt_entry; pv_cpu_ops.wbinvd = lguest_wbinvd; pv_cpu_ops.start_context_switch = paravirt_start_context_switch; pv_cpu_ops.end_context_switch = lguest_end_context_switch; /* Pagetable management */ pv_mmu_ops.write_cr3 = lguest_write_cr3; pv_mmu_ops.flush_tlb_user = lguest_flush_tlb_user; pv_mmu_ops.flush_tlb_single = lguest_flush_tlb_single; pv_mmu_ops.flush_tlb_kernel = lguest_flush_tlb_kernel; pv_mmu_ops.set_pte = lguest_set_pte; pv_mmu_ops.set_pte_at = lguest_set_pte_at; pv_mmu_ops.set_pmd = lguest_set_pmd; #ifdef CONFIG_X86_PAE pv_mmu_ops.set_pte_atomic = lguest_set_pte_atomic; pv_mmu_ops.pte_clear = lguest_pte_clear; pv_mmu_ops.pmd_clear = lguest_pmd_clear; pv_mmu_ops.set_pud = lguest_set_pud; #endif pv_mmu_ops.read_cr2 = lguest_read_cr2; pv_mmu_ops.read_cr3 = lguest_read_cr3; pv_mmu_ops.lazy_mode.enter = paravirt_enter_lazy_mmu; pv_mmu_ops.lazy_mode.leave = lguest_leave_lazy_mmu_mode; pv_mmu_ops.pte_update = lguest_pte_update; pv_mmu_ops.pte_update_defer = lguest_pte_update; #ifdef CONFIG_X86_LOCAL_APIC /* APIC read/write intercepts */ set_lguest_basic_apic_ops(); #endif x86_init.resources.memory_setup = lguest_memory_setup; x86_init.irqs.intr_init = lguest_init_IRQ; x86_init.timers.timer_init = lguest_time_init; x86_platform.calibrate_tsc = lguest_tsc_khz; x86_platform.get_wallclock = lguest_get_wallclock; /* * Now is a good time to look at the implementations of these functions * before returning to the rest of lguest_init(). */ /*G:070 * Now we've seen all the paravirt_ops, we return to * lguest_init() where the rest of the fairly chaotic boot setup * occurs. */ /* * The stack protector is a weird thing where gcc places a canary * value on the stack and then checks it on return. This file is * compiled with -fno-stack-protector it, so we got this far without * problems. The value of the canary is kept at offset 20 from the * %gs register, so we need to set that up before calling C functions * in other files. */ setup_stack_canary_segment(0); /* * We could just call load_stack_canary_segment(), but we might as well * call switch_to_new_gdt() which loads the whole table and sets up the * per-cpu segment descriptor register %fs as well. */ switch_to_new_gdt(0); /* * The Host<->Guest Switcher lives at the top of our address space, and * the Host told us how big it is when we made LGUEST_INIT hypercall: * it put the answer in lguest_data.reserve_mem */ reserve_top_address(lguest_data.reserve_mem); /* * If we don't initialize the lock dependency checker now, it crashes * atomic_notifier_chain_register, then paravirt_disable_iospace. */ lockdep_init(); /* Hook in our special panic hypercall code. */ atomic_notifier_chain_register(&panic_notifier_list, &paniced); /* * The IDE code spends about 3 seconds probing for disks: if we reserve * all the I/O ports up front it can't get them and so doesn't probe. * Other device drivers are similar (but less severe). This cuts the * kernel boot time on my machine from 4.1 seconds to 0.45 seconds. */ paravirt_disable_iospace(); /* * This is messy CPU setup stuff which the native boot code does before * start_kernel, so we have to do, too: */ cpu_detect(&new_cpu_data); /* head.S usually sets up the first capability word, so do it here. */ new_cpu_data.x86_capability[0] = cpuid_edx(1); /* Math is always hard! */ new_cpu_data.hard_math = 1; /* We don't have features. We have puppies! Puppies! */ #ifdef CONFIG_X86_MCE mce_disabled = 1; #endif #ifdef CONFIG_ACPI acpi_disabled = 1; #endif /* * We set the preferred console to "hvc". This is the "hypervisor * virtual console" driver written by the PowerPC people, which we also * adapted for lguest's use. */ add_preferred_console("hvc", 0, NULL); /* Register our very early console. */ virtio_cons_early_init(early_put_chars); /* * Last of all, we set the power management poweroff hook to point to * the Guest routine to power off, and the reboot hook to our restart * routine. */ pm_power_off = lguest_power_off; machine_ops.restart = lguest_restart; /* * Now we're set up, call i386_start_kernel() in head32.c and we proceed * to boot as normal. It never returns. */ i386_start_kernel(); } /* * This marks the end of stage II of our journey, The Guest. * * It is now time for us to explore the layer of virtual drivers and complete * our understanding of the Guest in "make Drivers". */