diff options
Diffstat (limited to 'drivers/lguest')
| -rw-r--r-- | drivers/lguest/core.c | 2 | ||||
| -rw-r--r-- | drivers/lguest/interrupts_and_traps.c | 10 | ||||
| -rw-r--r-- | drivers/lguest/lg.h | 2 | ||||
| -rw-r--r-- | drivers/lguest/lguest_device.c | 37 | ||||
| -rw-r--r-- | drivers/lguest/lguest_user.c | 17 | ||||
| -rw-r--r-- | drivers/lguest/page_tables.c | 282 | ||||
| -rw-r--r-- | drivers/lguest/x86/core.c | 107 |
7 files changed, 140 insertions, 317 deletions
diff --git a/drivers/lguest/core.c b/drivers/lguest/core.c index efa202499e37..2535933c49f8 100644 --- a/drivers/lguest/core.c +++ b/drivers/lguest/core.c @@ -117,7 +117,7 @@ static __init int map_switcher(void) /* * Now the Switcher is mapped at the right address, we can't fail! - * Copy in the compiled-in Switcher code (from <arch>_switcher.S). + * Copy in the compiled-in Switcher code (from x86/switcher_32.S). */ memcpy(switcher_vma->addr, start_switcher_text, end_switcher_text - start_switcher_text); diff --git a/drivers/lguest/interrupts_and_traps.c b/drivers/lguest/interrupts_and_traps.c index daaf86631647..28433a155d67 100644 --- a/drivers/lguest/interrupts_and_traps.c +++ b/drivers/lguest/interrupts_and_traps.c @@ -375,11 +375,9 @@ static bool direct_trap(unsigned int num) /* * The Host needs to see page faults (for shadow paging and to save the * fault address), general protection faults (in/out emulation) and - * device not available (TS handling), invalid opcode fault (kvm hcall), - * and of course, the hypercall trap. + * device not available (TS handling) and of course, the hypercall trap. */ - return num != 14 && num != 13 && num != 7 && - num != 6 && num != LGUEST_TRAP_ENTRY; + return num != 14 && num != 13 && num != 7 && num != LGUEST_TRAP_ENTRY; } /*:*/ @@ -429,8 +427,8 @@ void pin_stack_pages(struct lg_cpu *cpu) /* * Direct traps also mean that we need to know whenever the Guest wants to use - * a different kernel stack, so we can change the IDT entries to use that - * stack. The IDT entries expect a virtual address, so unlike most addresses + * a different kernel stack, so we can change the guest TSS to use that + * stack. The TSS entries expect a virtual address, so unlike most addresses * the Guest gives us, the "esp" (stack pointer) value here is virtual, not * physical. * diff --git a/drivers/lguest/lg.h b/drivers/lguest/lg.h index 9136411fadd5..295df06e6590 100644 --- a/drivers/lguest/lg.h +++ b/drivers/lguest/lg.h @@ -59,6 +59,8 @@ struct lg_cpu { struct lguest_pages *last_pages; + /* Initialization mode: linear map everything. */ + bool linear_pages; int cpu_pgd; /* Which pgd this cpu is currently using */ /* If a hypercall was asked for, this points to the arguments. */ diff --git a/drivers/lguest/lguest_device.c b/drivers/lguest/lguest_device.c index 69c84a1d88ea..5289ffa2e500 100644 --- a/drivers/lguest/lguest_device.c +++ b/drivers/lguest/lguest_device.c @@ -109,6 +109,17 @@ static u32 lg_get_features(struct virtio_device *vdev) } /* + * To notify on reset or feature finalization, we (ab)use the NOTIFY + * hypercall, with the descriptor address of the device. + */ +static void status_notify(struct virtio_device *vdev) +{ + unsigned long offset = (void *)to_lgdev(vdev)->desc - lguest_devices; + + hcall(LHCALL_NOTIFY, (max_pfn << PAGE_SHIFT) + offset, 0, 0, 0); +} + +/* * The virtio core takes the features the Host offers, and copies the ones * supported by the driver into the vdev->features array. Once that's all * sorted out, this routine is called so we can tell the Host which features we @@ -135,6 +146,9 @@ static void lg_finalize_features(struct virtio_device *vdev) if (test_bit(i, vdev->features)) out_features[i / 8] |= (1 << (i % 8)); } + + /* Tell Host we've finished with this device's feature negotiation */ + status_notify(vdev); } /* Once they've found a field, getting a copy of it is easy. */ @@ -168,28 +182,21 @@ static u8 lg_get_status(struct virtio_device *vdev) return to_lgdev(vdev)->desc->status; } -/* - * To notify on status updates, we (ab)use the NOTIFY hypercall, with the - * descriptor address of the device. A zero status means "reset". - */ -static void set_status(struct virtio_device *vdev, u8 status) -{ - unsigned long offset = (void *)to_lgdev(vdev)->desc - lguest_devices; - - /* We set the status. */ - to_lgdev(vdev)->desc->status = status; - hcall(LHCALL_NOTIFY, (max_pfn << PAGE_SHIFT) + offset, 0, 0, 0); -} - static void lg_set_status(struct virtio_device *vdev, u8 status) { BUG_ON(!status); - set_status(vdev, status); + to_lgdev(vdev)->desc->status = status; + + /* Tell Host immediately if we failed. */ + if (status & VIRTIO_CONFIG_S_FAILED) + status_notify(vdev); } static void lg_reset(struct virtio_device *vdev) { - set_status(vdev, 0); + /* 0 status means "reset" */ + to_lgdev(vdev)->desc->status = 0; + status_notify(vdev); } /* diff --git a/drivers/lguest/lguest_user.c b/drivers/lguest/lguest_user.c index 948c547b8e9e..f97e625241ad 100644 --- a/drivers/lguest/lguest_user.c +++ b/drivers/lguest/lguest_user.c @@ -1,8 +1,10 @@ -/*P:200 This contains all the /dev/lguest code, whereby the userspace launcher - * controls and communicates with the Guest. For example, the first write will - * tell us the Guest's memory layout and entry point. A read will run the - * Guest until something happens, such as a signal or the Guest doing a NOTIFY - * out to the Launcher. +/*P:200 This contains all the /dev/lguest code, whereby the userspace + * launcher controls and communicates with the Guest. For example, + * the first write will tell us the Guest's memory layout and entry + * point. A read will run the Guest until something happens, such as + * a signal or the Guest doing a NOTIFY out to the Launcher. There is + * also a way for the Launcher to attach eventfds to particular NOTIFY + * values instead of returning from the read() call. :*/ #include <linux/uaccess.h> #include <linux/miscdevice.h> @@ -357,8 +359,8 @@ static int initialize(struct file *file, const unsigned long __user *input) goto free_eventfds; /* - * Initialize the Guest's shadow page tables, using the toplevel - * address the Launcher gave us. This allocates memory, so can fail. + * Initialize the Guest's shadow page tables. This allocates + * memory, so can fail. */ err = init_guest_pagetable(lg); if (err) @@ -516,6 +518,7 @@ static const struct file_operations lguest_fops = { .read = read, .llseek = default_llseek, }; +/*:*/ /* * This is a textbook example of a "misc" character device. Populate a "struct diff --git a/drivers/lguest/page_tables.c b/drivers/lguest/page_tables.c index d21578ee95de..3b62be160a6e 100644 --- a/drivers/lguest/page_tables.c +++ b/drivers/lguest/page_tables.c @@ -17,7 +17,6 @@ #include <linux/percpu.h> #include <asm/tlbflush.h> #include <asm/uaccess.h> -#include <asm/bootparam.h> #include "lg.h" /*M:008 @@ -156,7 +155,7 @@ static pte_t *spte_addr(struct lg_cpu *cpu, pgd_t spgd, unsigned long vaddr) } /* - * These functions are just like the above two, except they access the Guest + * These functions are just like the above, except they access the Guest * page tables. Hence they return a Guest address. */ static unsigned long gpgd_addr(struct lg_cpu *cpu, unsigned long vaddr) @@ -196,7 +195,7 @@ static unsigned long gpte_addr(struct lg_cpu *cpu, #endif /*:*/ -/*M:014 +/*M:007 * get_pfn is slow: we could probably try to grab batches of pages here as * an optimization (ie. pre-faulting). :*/ @@ -325,10 +324,15 @@ bool demand_page(struct lg_cpu *cpu, unsigned long vaddr, int errcode) #endif /* First step: get the top-level Guest page table entry. */ - gpgd = lgread(cpu, gpgd_addr(cpu, vaddr), pgd_t); - /* Toplevel not present? We can't map it in. */ - if (!(pgd_flags(gpgd) & _PAGE_PRESENT)) - return false; + if (unlikely(cpu->linear_pages)) { + /* Faking up a linear mapping. */ + gpgd = __pgd(CHECK_GPGD_MASK); + } else { + gpgd = lgread(cpu, gpgd_addr(cpu, vaddr), pgd_t); + /* Toplevel not present? We can't map it in. */ + if (!(pgd_flags(gpgd) & _PAGE_PRESENT)) + return false; + } /* Now look at the matching shadow entry. */ spgd = spgd_addr(cpu, cpu->cpu_pgd, vaddr); @@ -353,10 +357,15 @@ bool demand_page(struct lg_cpu *cpu, unsigned long vaddr, int errcode) } #ifdef CONFIG_X86_PAE - gpmd = lgread(cpu, gpmd_addr(gpgd, vaddr), pmd_t); - /* Middle level not present? We can't map it in. */ - if (!(pmd_flags(gpmd) & _PAGE_PRESENT)) - return false; + if (unlikely(cpu->linear_pages)) { + /* Faking up a linear mapping. */ + gpmd = __pmd(_PAGE_TABLE); + } else { + gpmd = lgread(cpu, gpmd_addr(gpgd, vaddr), pmd_t); + /* Middle level not present? We can't map it in. */ + if (!(pmd_flags(gpmd) & _PAGE_PRESENT)) + return false; + } /* Now look at the matching shadow entry. */ spmd = spmd_addr(cpu, *spgd, vaddr); @@ -397,8 +406,13 @@ bool demand_page(struct lg_cpu *cpu, unsigned long vaddr, int errcode) gpte_ptr = gpte_addr(cpu, gpgd, vaddr); #endif - /* Read the actual PTE value. */ - gpte = lgread(cpu, gpte_ptr, pte_t); + if (unlikely(cpu->linear_pages)) { + /* Linear? Make up a PTE which points to same page. */ + gpte = __pte((vaddr & PAGE_MASK) | _PAGE_RW | _PAGE_PRESENT); + } else { + /* Read the actual PTE value. */ + gpte = lgread(cpu, gpte_ptr, pte_t); + } /* If this page isn't in the Guest page tables, we can't page it in. */ if (!(pte_flags(gpte) & _PAGE_PRESENT)) @@ -454,7 +468,8 @@ bool demand_page(struct lg_cpu *cpu, unsigned long vaddr, int errcode) * Finally, we write the Guest PTE entry back: we've set the * _PAGE_ACCESSED and maybe the _PAGE_DIRTY flags. */ - lgwrite(cpu, gpte_ptr, pte_t, gpte); + if (likely(!cpu->linear_pages)) + lgwrite(cpu, gpte_ptr, pte_t, gpte); /* * The fault is fixed, the page table is populated, the mapping @@ -612,6 +627,11 @@ unsigned long guest_pa(struct lg_cpu *cpu, unsigned long vaddr) #ifdef CONFIG_X86_PAE pmd_t gpmd; #endif + + /* Still not set up? Just map 1:1. */ + if (unlikely(cpu->linear_pages)) + return vaddr; + /* First step: get the top-level Guest page table entry. */ gpgd = lgread(cpu, gpgd_addr(cpu, vaddr), pgd_t); /* Toplevel not present? We can't map it in. */ @@ -708,32 +728,6 @@ static unsigned int new_pgdir(struct lg_cpu *cpu, return next; } -/*H:430 - * (iv) Switching page tables - * - * Now we've seen all the page table setting and manipulation, let's see - * what happens when the Guest changes page tables (ie. changes the top-level - * pgdir). This occurs on almost every context switch. - */ -void guest_new_pagetable(struct lg_cpu *cpu, unsigned long pgtable) -{ - int newpgdir, repin = 0; - - /* Look to see if we have this one already. */ - newpgdir = find_pgdir(cpu->lg, pgtable); - /* - * If not, we allocate or mug an existing one: if it's a fresh one, - * repin gets set to 1. - */ - if (newpgdir == ARRAY_SIZE(cpu->lg->pgdirs)) - newpgdir = new_pgdir(cpu, pgtable, &repin); - /* Change the current pgd index to the new one. */ - cpu->cpu_pgd = newpgdir; - /* If it was completely blank, we map in the Guest kernel stack */ - if (repin) - pin_stack_pages(cpu); -} - /*H:470 * Finally, a routine which throws away everything: all PGD entries in all * the shadow page tables, including the Guest's kernel mappings. This is used @@ -780,6 +774,44 @@ void guest_pagetable_clear_all(struct lg_cpu *cpu) /* We need the Guest kernel stack mapped again. */ pin_stack_pages(cpu); } + +/*H:430 + * (iv) Switching page tables + * + * Now we've seen all the page table setting and manipulation, let's see + * what happens when the Guest changes page tables (ie. changes the top-level + * pgdir). This occurs on almost every context switch. + */ +void guest_new_pagetable(struct lg_cpu *cpu, unsigned long pgtable) +{ + int newpgdir, repin = 0; + + /* + * The very first time they call this, we're actually running without + * any page tables; we've been making it up. Throw them away now. + */ + if (unlikely(cpu->linear_pages)) { + release_all_pagetables(cpu->lg); + cpu->linear_pages = false; + /* Force allocation of a new pgdir. */ + newpgdir = ARRAY_SIZE(cpu->lg->pgdirs); + } else { + /* Look to see if we have this one already. */ + newpgdir = find_pgdir(cpu->lg, pgtable); + } + + /* + * If not, we allocate or mug an existing one: if it's a fresh one, + * repin gets set to 1. + */ + if (newpgdir == ARRAY_SIZE(cpu->lg->pgdirs)) + newpgdir = new_pgdir(cpu, pgtable, &repin); + /* Change the current pgd index to the new one. */ + cpu->cpu_pgd = newpgdir; + /* If it was completely blank, we map in the Guest kernel stack */ + if (repin) + pin_stack_pages(cpu); +} /*:*/ /*M:009 @@ -919,168 +951,26 @@ void guest_set_pmd(struct lguest *lg, unsigned long pmdp, u32 idx) } #endif -/*H:505 - * To get through boot, we construct simple identity page mappings (which - * set virtual == physical) and linear mappings which will get the Guest far - * enough into the boot to create its own. The linear mapping means we - * simplify the Guest boot, but it makes assumptions about their PAGE_OFFSET, - * as you'll see. - * - * We lay them out of the way, just below the initrd (which is why we need to - * know its size here). - */ -static unsigned long setup_pagetables(struct lguest *lg, - unsigned long mem, - unsigned long initrd_size) -{ - pgd_t __user *pgdir; - pte_t __user *linear; - unsigned long mem_base = (unsigned long)lg->mem_base; - unsigned int mapped_pages, i, linear_pages; -#ifdef CONFIG_X86_PAE - pmd_t __user *pmds; - unsigned int j; - pgd_t pgd; - pmd_t pmd; -#else - unsigned int phys_linear; -#endif - - /* - * We have mapped_pages frames to map, so we need linear_pages page - * tables to map them. - */ - mapped_pages = mem / PAGE_SIZE; - linear_pages = (mapped_pages + PTRS_PER_PTE - 1) / PTRS_PER_PTE; - - /* We put the toplevel page directory page at the top of memory. */ - pgdir = (pgd_t *)(mem + mem_base - initrd_size - PAGE_SIZE); - - /* Now we use the next linear_pages pages as pte pages */ - linear = (void *)pgdir - linear_pages * PAGE_SIZE; - -#ifdef CONFIG_X86_PAE - /* - * And the single mid page goes below that. We only use one, but - * that's enough to map 1G, which definitely gets us through boot. - */ - pmds = (void *)linear - PAGE_SIZE; -#endif - /* - * Linear mapping is easy: put every page's address into the - * mapping in order. - */ - for (i = 0; i < mapped_pages; i++) { - pte_t pte; - pte = pfn_pte(i, __pgprot(_PAGE_PRESENT|_PAGE_RW|_PAGE_USER)); - if (copy_to_user(&linear[i], &pte, sizeof(pte)) != 0) - return -EFAULT; - } - -#ifdef CONFIG_X86_PAE - /* - * Make the Guest PMD entries point to the corresponding place in the - * linear mapping (up to one page worth of PMD). - */ - for (i = j = 0; i < mapped_pages && j < PTRS_PER_PMD; - i += PTRS_PER_PTE, j++) { - pmd = pfn_pmd(((unsigned long)&linear[i] - mem_base)/PAGE_SIZE, - __pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER)); - - if (copy_to_user(&pmds[j], &pmd, sizeof(pmd)) != 0) - return -EFAULT; - } - - /* One PGD entry, pointing to that PMD page. */ - pgd = __pgd(((unsigned long)pmds - mem_base) | _PAGE_PRESENT); - /* Copy it in as the first PGD entry (ie. addresses 0-1G). */ - if (copy_to_user(&pgdir[0], &pgd, sizeof(pgd)) != 0) - return -EFAULT; - /* - * And the other PGD entry to make the linear mapping at PAGE_OFFSET - */ - if (copy_to_user(&pgdir[KERNEL_PGD_BOUNDARY], &pgd, sizeof(pgd))) - return -EFAULT; -#else - /* - * The top level points to the linear page table pages above. - * We setup the identity and linear mappings here. - */ - phys_linear = (unsigned long)linear - mem_base; - for (i = 0; i < mapped_pages; i += PTRS_PER_PTE) { - pgd_t pgd; - /* - * Create a PGD entry which points to the right part of the - * linear PTE pages. - */ - pgd = __pgd((phys_linear + i * sizeof(pte_t)) | - (_PAGE_PRESENT | _PAGE_RW | _PAGE_USER)); - - /* - * Copy it into the PGD page at 0 and PAGE_OFFSET. - */ - if (copy_to_user(&pgdir[i / PTRS_PER_PTE], &pgd, sizeof(pgd)) - || copy_to_user(&pgdir[pgd_index(PAGE_OFFSET) - + i / PTRS_PER_PTE], - &pgd, sizeof(pgd))) - return -EFAULT; - } -#endif - - /* - * We return the top level (guest-physical) address: we remember where - * this is to write it into lguest_data when the Guest initializes. - */ - return (unsigned long)pgdir - mem_base; -} - /*H:500 * (vii) Setting up the page tables initially. * - * When a Guest is first created, the Launcher tells us where the toplevel of - * its first page table is. We set some things up here: + * When a Guest is first created, set initialize a shadow page table which + * we will populate on future faults. The Guest doesn't have any actual + * pagetables yet, so we set linear_pages to tell demand_page() to fake it + * for the moment. */ int init_guest_pagetable(struct lguest *lg) { - u64 mem; - u32 initrd_size; - struct boot_params __user *boot = (struct boot_params *)lg->mem_base; -#ifdef CONFIG_X86_PAE - pgd_t *pgd; - pmd_t *pmd_table; -#endif - /* - * Get the Guest memory size and the ramdisk size from the boot header - * located at lg->mem_base (Guest address 0). - */ - if (copy_from_user(&mem, &boot->e820_map[0].size, sizeof(mem)) - || get_user(initrd_size, &boot->hdr.ramdisk_size)) - return -EFAULT; + struct lg_cpu *cpu = &lg->cpus[0]; + int allocated = 0; - /* - * We start on the first shadow page table, and give it a blank PGD - * page. - */ - lg->pgdirs[0].gpgdir = setup_pagetables(lg, mem, initrd_size); - if (IS_ERR_VALUE(lg->pgdirs[0].gpgdir)) - return lg->pgdirs[0].gpgdir; - lg->pgdirs[0].pgdir = (pgd_t *)get_zeroed_page(GFP_KERNEL); - if (!lg->pgdirs[0].pgdir) + /* lg (and lg->cpus[]) starts zeroed: this allocates a new pgdir */ + cpu->cpu_pgd = new_pgdir(cpu, 0, &allocated); + if (!allocated) return -ENOMEM; -#ifdef CONFIG_X86_PAE - /* For PAE, we also create the initial mid-level. */ - pgd = lg->pgdirs[0].pgdir; - pmd_table = (pmd_t *) get_zeroed_page(GFP_KERNEL); - if (!pmd_table) - return -ENOMEM; - - set_pgd(pgd + SWITCHER_PGD_INDEX, - __pgd(__pa(pmd_table) | _PAGE_PRESENT)); -#endif - - /* This is the current page table. */ - lg->cpus[0].cpu_pgd = 0; + /* We start with a linear mapping until the initialize. */ + cpu->linear_pages = true; return 0; } @@ -1095,10 +985,10 @@ void page_table_guest_data_init(struct lg_cpu *cpu) * of virtual addresses used by the Switcher. */ || put_user(RESERVE_MEM * 1024 * 1024, - &cpu->lg->lguest_data->reserve_mem) - || put_user(cpu->lg->pgdirs[0].gpgdir, - &cpu->lg->lguest_data->pgdir)) + &cpu->lg->lguest_data->reserve_mem)) { kill_guest(cpu, "bad guest page %p", cpu->lg->lguest_data); + return; + } /* * In flush_user_mappings() we loop from 0 to diff --git a/drivers/lguest/x86/core.c b/drivers/lguest/x86/core.c index 9f1659c3d1f3..65af42f2d593 100644 --- a/drivers/lguest/x86/core.c +++ b/drivers/lguest/x86/core.c @@ -269,10 +269,10 @@ void lguest_arch_run_guest(struct lg_cpu *cpu) static int emulate_insn(struct lg_cpu *cpu) { u8 insn; - unsigned int insnlen = 0, in = 0, shift = 0; + unsigned int insnlen = 0, in = 0, small_operand = 0; /* * The eip contains the *virtual* address of the Guest's instruction: - * guest_pa just subtracts the Guest's page_offset. + * walk the Guest's page tables to find the "physical" address. */ unsigned long physaddr = guest_pa(cpu, cpu->regs->eip); @@ -300,11 +300,10 @@ static int emulate_insn(struct lg_cpu *cpu) } /* - * 0x66 is an "operand prefix". It means it's using the upper 16 bits - * of the eax register. + * 0x66 is an "operand prefix". It means a 16, not 32 bit in/out. */ if (insn == 0x66) { - shift = 16; + small_operand = 1; /* The instruction is 1 byte so far, read the next byte. */ insnlen = 1; insn = lgread(cpu, physaddr + insnlen, u8); @@ -340,11 +339,14 @@ static int emulate_insn(struct lg_cpu *cpu) * traditionally means "there's nothing there". */ if (in) { - /* Lower bit tells is whether it's a 16 or 32 bit access */ - if (insn & 0x1) - cpu->regs->eax = 0xFFFFFFFF; - else - cpu->regs->eax |= (0xFFFF << shift); + /* Lower bit tells means it's a 32/16 bit access */ + if (insn & 0x1) { + if (small_operand) + cpu->regs->eax |= 0xFFFF; + else + cpu->regs->eax = 0xFFFFFFFF; + } else + cpu->regs->eax |= 0xFF; } /* Finally, we've "done" the instruction, so move past it. */ cpu->regs->eip += insnlen; @@ -352,69 +354,6 @@ static int emulate_insn(struct lg_cpu *cpu) return 1; } -/* - * Our hypercalls mechanism used to be based on direct software interrupts. - * After Anthony's "Refactor hypercall infrastructure" kvm patch, we decided to - * change over to using kvm hypercalls. - * - * KVM_HYPERCALL is actually a "vmcall" instruction, which generates an invalid - * opcode fault (fault 6) on non-VT cpus, so the easiest solution seemed to be - * an *emulation approach*: if the fault was really produced by an hypercall - * (is_hypercall() does exactly this check), we can just call the corresponding - * hypercall host implementation function. - * - * But these invalid opcode faults are notably slower than software interrupts. - * So we implemented the *patching (or rewriting) approach*: every time we hit - * the KVM_HYPERCALL opcode in Guest code, we patch it to the old "int 0x1f" - * opcode, so next time the Guest calls this hypercall it will use the - * faster trap mechanism. - * - * Matias even benchmarked it to convince you: this shows the average cycle - * cost of a hypercall. For each alternative solution mentioned above we've - * made 5 runs of the benchmark: - * - * 1) direct software interrupt: 2915, 2789, 2764, 2721, 2898 - * 2) emulation technique: 3410, 3681, 3466, 3392, 3780 - * 3) patching (rewrite) technique: 2977, 2975, 2891, 2637, 2884 - * - * One two-line function is worth a 20% hypercall speed boost! - */ -static void rewrite_hypercall(struct lg_cpu *cpu) -{ - /* - * This are the opcodes we use to patch the Guest. The opcode for "int - * $0x1f" is "0xcd 0x1f" but vmcall instruction is 3 bytes long, so we - * complete the sequence with a NOP (0x90). - */ - u8 insn[3] = {0xcd, 0x1f, 0x90}; - - __lgwrite(cpu, guest_pa(cpu, cpu->regs->eip), insn, sizeof(insn)); - /* - * The above write might have caused a copy of that page to be made - * (if it was read-only). We need to make sure the Guest has - * up-to-date pagetables. As this doesn't happen often, we can just - * drop them all. - */ - guest_pagetable_clear_all(cpu); -} - -static bool is_hypercall(struct lg_cpu *cpu) -{ - u8 insn[3]; - - /* - * This must be the Guest kernel trying to do something. - * The bottom two bits of the CS segment register are the privilege - * level. - */ - if ((cpu->regs->cs & 3) != GUEST_PL) - return false; - - /* Is it a vmcall? */ - __lgread(cpu, insn, guest_pa(cpu, cpu->regs->eip), sizeof(insn)); - return insn[0] == 0x0f && insn[1] == 0x01 && insn[2] == 0xc1; -} - /*H:050 Once we've re-enabled interrupts, we look at why the Guest exited. */ void lguest_arch_handle_trap(struct lg_cpu *cpu) { @@ -429,20 +368,6 @@ void lguest_arch_handle_trap(struct lg_cpu *cpu) if (emulate_insn(cpu)) return; } - /* - * If KVM is active, the vmcall instruction triggers a General - * Protection Fault. Normally it triggers an invalid opcode - * fault (6): - */ - case 6: - /* - * We need to check if ring == GUEST_PL and faulting - * instruction == vmcall. - */ - if (is_hypercall(cpu)) { - rewrite_hypercall(cpu); - return; - } break; case 14: /* We've intercepted a Page Fault. */ /* @@ -486,7 +411,7 @@ void lguest_arch_handle_trap(struct lg_cpu *cpu) * These values mean a real interrupt occurred, in which case * the Host handler has already been run. We just do a * friendly check if another process should now be run, then - * return to run the Guest again + * return to run the Guest again. */ cond_resched(); return; @@ -536,7 +461,7 @@ void __init lguest_arch_host_init(void) int i; /* - * Most of the i386/switcher.S doesn't care that it's been moved; on + * Most of the x86/switcher_32.S doesn't care that it's been moved; on * Intel, jumps are relative, and it doesn't access any references to * external code or data. * @@ -664,7 +589,7 @@ void __init lguest_arch_host_init(void) clear_cpu_cap(&boot_cpu_data, X86_FEATURE_PGE); } put_online_cpus(); -}; +} /*:*/ void __exit lguest_arch_host_fini(void) @@ -747,8 +672,6 @@ int lguest_arch_init_hypercalls(struct lg_cpu *cpu) /*:*/ /*L:030 - * lguest_arch_setup_regs() - * * Most of the Guest's registers are left alone: we used get_zeroed_page() to * allocate the structure, so they will be 0. */ |