diff options
Diffstat (limited to 'drivers/lguest/core.c')
| -rw-r--r-- | drivers/lguest/core.c | 217 |
1 files changed, 132 insertions, 85 deletions
diff --git a/drivers/lguest/core.c b/drivers/lguest/core.c index 5eea4356d70..0bf1e4edf04 100644 --- a/drivers/lguest/core.c +++ b/drivers/lguest/core.c @@ -1,6 +1,8 @@ -/*P:400 This contains run_guest() which actually calls into the Host<->Guest +/*P:400 + * This contains run_guest() which actually calls into the Host<->Guest * Switcher and analyzes the return, such as determining if the Guest wants the - * Host to do something. This file also contains useful helper routines. :*/ + * Host to do something. This file also contains useful helper routines. +:*/ #include <linux/module.h> #include <linux/stringify.h> #include <linux/stddef.h> @@ -10,6 +12,7 @@ #include <linux/cpu.h> #include <linux/freezer.h> #include <linux/highmem.h> +#include <linux/slab.h> #include <asm/paravirt.h> #include <asm/pgtable.h> #include <asm/uaccess.h> @@ -17,14 +20,15 @@ #include <asm/asm-offsets.h> #include "lg.h" - +unsigned long switcher_addr; +struct page **lg_switcher_pages; static struct vm_struct *switcher_vma; -static struct page **switcher_page; /* This One Big lock protects all inter-guest data structures. */ DEFINE_MUTEX(lguest_lock); -/*H:010 We need to set up the Switcher at a high virtual address. Remember the +/*H:010 + * We need to set up the Switcher at a high virtual address. Remember the * Switcher is a few hundred bytes of assembler code which actually changes the * CPU to run the Guest, and then changes back to the Host when a trap or * interrupt happens. @@ -33,7 +37,8 @@ DEFINE_MUTEX(lguest_lock); * Host since it will be running as the switchover occurs. * * Trying to map memory at a particular address is an unusual thing to do, so - * it's not a simple one-liner. */ + * it's not a simple one-liner. + */ static __init int map_switcher(void) { int i, err; @@ -47,41 +52,53 @@ static __init int map_switcher(void) * easy. */ - /* We allocate an array of struct page pointers. map_vm_area() wants - * this, rather than just an array of pages. */ - switcher_page = kmalloc(sizeof(switcher_page[0])*TOTAL_SWITCHER_PAGES, - GFP_KERNEL); - if (!switcher_page) { + /* We assume Switcher text fits into a single page. */ + if (end_switcher_text - start_switcher_text > PAGE_SIZE) { + printk(KERN_ERR "lguest: switcher text too large (%zu)\n", + end_switcher_text - start_switcher_text); + return -EINVAL; + } + + /* + * We allocate an array of struct page pointers. map_vm_area() wants + * this, rather than just an array of pages. + */ + lg_switcher_pages = kmalloc(sizeof(lg_switcher_pages[0]) + * TOTAL_SWITCHER_PAGES, + GFP_KERNEL); + if (!lg_switcher_pages) { err = -ENOMEM; goto out; } - /* Now we actually allocate the pages. The Guest will see these pages, - * so we make sure they're zeroed. */ + /* + * Now we actually allocate the pages. The Guest will see these pages, + * so we make sure they're zeroed. + */ for (i = 0; i < TOTAL_SWITCHER_PAGES; i++) { - unsigned long addr = get_zeroed_page(GFP_KERNEL); - if (!addr) { + lg_switcher_pages[i] = alloc_page(GFP_KERNEL|__GFP_ZERO); + if (!lg_switcher_pages[i]) { err = -ENOMEM; goto free_some_pages; } - switcher_page[i] = virt_to_page(addr); } - /* First we check that the Switcher won't overlap the fixmap area at - * the top of memory. It's currently nowhere near, but it could have - * very strange effects if it ever happened. */ - if (SWITCHER_ADDR + (TOTAL_SWITCHER_PAGES+1)*PAGE_SIZE > FIXADDR_START){ - err = -ENOMEM; - printk("lguest: mapping switcher would thwack fixmap\n"); - goto free_pages; - } + /* + * We place the Switcher underneath the fixmap area, which is the + * highest virtual address we can get. This is important, since we + * tell the Guest it can't access this memory, so we want its ceiling + * as high as possible. + */ + switcher_addr = FIXADDR_START - (TOTAL_SWITCHER_PAGES+1)*PAGE_SIZE; - /* Now we reserve the "virtual memory area" we want: 0xFFC00000 - * (SWITCHER_ADDR). We might not get it in theory, but in practice - * it's worked so far. The end address needs +1 because __get_vm_area - * allocates an extra guard page, so we need space for that. */ + /* + * Now we reserve the "virtual memory area" we want. We might + * not get it in theory, but in practice it's worked so far. + * The end address needs +1 because __get_vm_area allocates an + * extra guard page, so we need space for that. + */ switcher_vma = __get_vm_area(TOTAL_SWITCHER_PAGES * PAGE_SIZE, - VM_ALLOC, SWITCHER_ADDR, SWITCHER_ADDR + VM_ALLOC, switcher_addr, switcher_addr + (TOTAL_SWITCHER_PAGES+1) * PAGE_SIZE); if (!switcher_vma) { err = -ENOMEM; @@ -89,20 +106,24 @@ static __init int map_switcher(void) goto free_pages; } - /* This code actually sets up the pages we've allocated to appear at - * SWITCHER_ADDR. map_vm_area() takes the vma we allocated above, the + /* + * This code actually sets up the pages we've allocated to appear at + * switcher_addr. map_vm_area() takes the vma we allocated above, the * kind of pages we're mapping (kernel pages), and a pointer to our * array of struct pages. It increments that pointer, but we don't - * care. */ - pagep = switcher_page; - err = map_vm_area(switcher_vma, PAGE_KERNEL, &pagep); + * care. + */ + pagep = lg_switcher_pages; + err = map_vm_area(switcher_vma, PAGE_KERNEL_EXEC, &pagep); if (err) { printk("lguest: map_vm_area failed: %i\n", err); goto free_vma; } - /* Now the Switcher is mapped at the right address, we can't fail! - * Copy in the compiled-in Switcher code (from <arch>_switcher.S). */ + /* + * Now the Switcher is mapped at the right address, we can't fail! + * 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); @@ -117,15 +138,14 @@ free_pages: i = TOTAL_SWITCHER_PAGES; free_some_pages: for (--i; i >= 0; i--) - __free_pages(switcher_page[i], 0); - kfree(switcher_page); + __free_pages(lg_switcher_pages[i], 0); + kfree(lg_switcher_pages); out: return err; } /*:*/ -/* Cleaning up the mapping when the module is unloaded is almost... - * too easy. */ +/* Cleaning up the mapping when the module is unloaded is almost... too easy. */ static void unmap_switcher(void) { unsigned int i; @@ -134,7 +154,8 @@ static void unmap_switcher(void) vunmap(switcher_vma->addr); /* Now we just need to free the pages we copied the switcher into */ for (i = 0; i < TOTAL_SWITCHER_PAGES; i++) - __free_pages(switcher_page[i], 0); + __free_pages(lg_switcher_pages[i], 0); + kfree(lg_switcher_pages); } /*H:032 @@ -150,16 +171,19 @@ static void unmap_switcher(void) * But we can't trust the Guest: it might be trying to access the Launcher * code. We have to check that the range is below the pfn_limit the Launcher * gave us. We have to make sure that addr + len doesn't give us a false - * positive by overflowing, too. */ -int lguest_address_ok(const struct lguest *lg, - unsigned long addr, unsigned long len) + * positive by overflowing, too. + */ +bool lguest_address_ok(const struct lguest *lg, + unsigned long addr, unsigned long len) { return (addr+len) / PAGE_SIZE < lg->pfn_limit && (addr+len >= addr); } -/* This routine copies memory from the Guest. Here we can see how useful the +/* + * This routine copies memory from the Guest. Here we can see how useful the * kill_lguest() routine we met in the Launcher can be: we return a random - * value (all zeroes) instead of needing to return an error. */ + * value (all zeroes) instead of needing to return an error. + */ void __lgread(struct lg_cpu *cpu, void *b, unsigned long addr, unsigned bytes) { if (!lguest_address_ok(cpu->lg, addr, bytes) @@ -180,58 +204,87 @@ void __lgwrite(struct lg_cpu *cpu, unsigned long addr, const void *b, } /*:*/ -/*H:030 Let's jump straight to the the main loop which runs the Guest. +/*H:030 + * Let's jump straight to the the main loop which runs the Guest. * Remember, this is called by the Launcher reading /dev/lguest, and we keep - * going around and around until something interesting happens. */ + * going around and around until something interesting happens. + */ int run_guest(struct lg_cpu *cpu, unsigned long __user *user) { /* We stop running once the Guest is dead. */ while (!cpu->lg->dead) { + unsigned int irq; + bool more; + /* First we run any hypercalls the Guest wants done. */ if (cpu->hcall) do_hypercalls(cpu); - /* It's possible the Guest did a NOTIFY hypercall to the - * Launcher, in which case we return from the read() now. */ + /* + * It's possible the Guest did a NOTIFY hypercall to the + * Launcher. + */ if (cpu->pending_notify) { - if (put_user(cpu->pending_notify, user)) - return -EFAULT; - return sizeof(cpu->pending_notify); + /* + * Does it just needs to write to a registered + * eventfd (ie. the appropriate virtqueue thread)? + */ + if (!send_notify_to_eventfd(cpu)) { + /* OK, we tell the main Launcher. */ + if (put_user(cpu->pending_notify, user)) + return -EFAULT; + return sizeof(cpu->pending_notify); + } } + /* + * All long-lived kernel loops need to check with this horrible + * thing called the freezer. If the Host is trying to suspend, + * it stops us. + */ + try_to_freeze(); + /* Check for signals */ if (signal_pending(current)) return -ERESTARTSYS; - /* If Waker set break_out, return to Launcher. */ - if (cpu->break_out) - return -EAGAIN; - - /* Check if there are any interrupts which can be delivered now: + /* + * Check if there are any interrupts which can be delivered now: * if so, this sets up the hander to be executed when we next - * run the Guest. */ - maybe_do_interrupt(cpu); - - /* All long-lived kernel loops need to check with this horrible - * thing called the freezer. If the Host is trying to suspend, - * it stops us. */ - try_to_freeze(); - - /* Just make absolutely sure the Guest is still alive. One of - * those hypercalls could have been fatal, for example. */ + * run the Guest. + */ + irq = interrupt_pending(cpu, &more); + if (irq < LGUEST_IRQS) + try_deliver_interrupt(cpu, irq, more); + + /* + * Just make absolutely sure the Guest is still alive. One of + * those hypercalls could have been fatal, for example. + */ if (cpu->lg->dead) break; - /* If the Guest asked to be stopped, we sleep. The Guest's - * clock timer or LHCALL_BREAK from the Waker will wake us. */ + /* + * If the Guest asked to be stopped, we sleep. The Guest's + * clock timer will wake us. + */ if (cpu->halted) { set_current_state(TASK_INTERRUPTIBLE); - schedule(); + /* + * Just before we sleep, make sure no interrupt snuck in + * which we should be doing. + */ + if (interrupt_pending(cpu, &more) < LGUEST_IRQS) + set_current_state(TASK_RUNNING); + else + schedule(); continue; } - /* OK, now we're ready to jump into the Guest. First we put up - * the "Do Not Disturb" sign: */ + /* + * OK, now we're ready to jump into the Guest. First we put up + * the "Do Not Disturb" sign: + */ local_irq_disable(); /* Actually run the Guest until something happens. */ @@ -265,7 +318,7 @@ static int __init init(void) int err; /* Lguest can't run under Xen, VMI or itself. It does Tricky Stuff. */ - if (paravirt_enabled()) { + if (get_kernel_rpl() != 0) { printk("lguest is afraid of being a guest\n"); return -EPERM; } @@ -275,15 +328,10 @@ static int __init init(void) if (err) goto out; - /* Now we set up the pagetable implementation for the Guests. */ - err = init_pagetables(switcher_page, SHARED_SWITCHER_PAGES); - if (err) - goto unmap; - /* We might need to reserve an interrupt vector. */ err = init_interrupts(); if (err) - goto free_pgtables; + goto unmap; /* /dev/lguest needs to be registered. */ err = lguest_device_init(); @@ -298,8 +346,6 @@ static int __init init(void) free_interrupts: free_interrupts(); -free_pgtables: - free_pagetables(); unmap: unmap_switcher(); out: @@ -311,15 +357,16 @@ static void __exit fini(void) { lguest_device_remove(); free_interrupts(); - free_pagetables(); unmap_switcher(); lguest_arch_host_fini(); } /*:*/ -/* The Host side of lguest can be a module. This is a nice way for people to - * play with it. */ +/* + * The Host side of lguest can be a module. This is a nice way for people to + * play with it. + */ module_init(init); module_exit(fini); MODULE_LICENSE("GPL"); |