diff options
Diffstat (limited to 'drivers/xen/gntdev.c')
| -rw-r--r-- | drivers/xen/gntdev.c | 320 |
1 files changed, 203 insertions, 117 deletions
diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c index 59ffea800079..2c960f187f7c 100644 --- a/drivers/xen/gntdev.c +++ b/drivers/xen/gntdev.c @@ -35,6 +35,7 @@ #include <linux/slab.h> #include <linux/highmem.h> #include <linux/refcount.h> +#include <linux/workqueue.h> #include <xen/xen.h> #include <xen/grant_table.h> @@ -55,15 +56,25 @@ MODULE_AUTHOR("Derek G. Murray <Derek.Murray@cl.cam.ac.uk>, " "Gerd Hoffmann <kraxel@redhat.com>"); MODULE_DESCRIPTION("User-space granted page access driver"); +#define GNTDEV_COPY_BATCH 16 + +struct gntdev_copy_batch { + struct gnttab_copy ops[GNTDEV_COPY_BATCH]; + struct page *pages[GNTDEV_COPY_BATCH]; + s16 __user *status[GNTDEV_COPY_BATCH]; + unsigned int nr_ops; + unsigned int nr_pages; + bool writeable; + struct gntdev_copy_batch *next; +}; + static unsigned int limit = 64*1024; module_param(limit, uint, 0644); MODULE_PARM_DESC(limit, "Maximum number of grants that may be mapped by one mapping request"); -static int use_ptemod; - -static int unmap_grant_pages(struct gntdev_grant_map *map, - int offset, int pages); +static void unmap_grant_pages(struct gntdev_grant_map *map, + int offset, int pages); static struct miscdevice gntdev_miscdev; @@ -120,6 +131,7 @@ static void gntdev_free_map(struct gntdev_grant_map *map) kvfree(map->unmap_ops); kvfree(map->kmap_ops); kvfree(map->kunmap_ops); + kvfree(map->being_removed); kfree(map); } @@ -140,12 +152,15 @@ struct gntdev_grant_map *gntdev_alloc_map(struct gntdev_priv *priv, int count, add->unmap_ops = kvmalloc_array(count, sizeof(add->unmap_ops[0]), GFP_KERNEL); add->pages = kvcalloc(count, sizeof(add->pages[0]), GFP_KERNEL); + add->being_removed = + kvcalloc(count, sizeof(add->being_removed[0]), GFP_KERNEL); if (NULL == add->grants || NULL == add->map_ops || NULL == add->unmap_ops || - NULL == add->pages) + NULL == add->pages || + NULL == add->being_removed) goto err; - if (use_ptemod) { + if (xen_pv_domain()) { add->kmap_ops = kvmalloc_array(count, sizeof(add->kmap_ops[0]), GFP_KERNEL); add->kunmap_ops = kvmalloc_array(count, sizeof(add->kunmap_ops[0]), @@ -193,7 +208,7 @@ struct gntdev_grant_map *gntdev_alloc_map(struct gntdev_priv *priv, int count, add->grants[i].ref = INVALID_GRANT_REF; add->map_ops[i].handle = INVALID_GRANT_HANDLE; add->unmap_ops[i].handle = INVALID_GRANT_HANDLE; - if (use_ptemod) { + if (xen_pv_domain()) { add->kmap_ops[i].handle = INVALID_GRANT_HANDLE; add->kunmap_ops[i].handle = INVALID_GRANT_HANDLE; } @@ -250,9 +265,39 @@ void gntdev_put_map(struct gntdev_priv *priv, struct gntdev_grant_map *map) if (!refcount_dec_and_test(&map->users)) return; - if (map->pages && !use_ptemod) + if (map->pages && !xen_pv_domain()) { + /* + * Increment the reference count. This ensures that the + * subsequent call to unmap_grant_pages() will not wind up + * re-entering itself. It *can* wind up calling + * gntdev_put_map() recursively, but such calls will be with a + * reference count greater than 1, so they will return before + * this code is reached. The recursion depth is thus limited to + * 1. Do NOT use refcount_inc() here, as it will detect that + * the reference count is zero and WARN(). + */ + refcount_set(&map->users, 1); + + /* + * Unmap the grants. This may or may not be asynchronous, so it + * is possible that the reference count is 1 on return, but it + * could also be greater than 1. + */ unmap_grant_pages(map, 0, map->count); + /* Check if the memory now needs to be freed */ + if (!refcount_dec_and_test(&map->users)) + return; + + /* + * All pages have been returned to the hypervisor, so free the + * map. + */ + } + + if (xen_pv_domain() && map->notifier_init) + mmu_interval_notifier_remove(&map->notifier); + if (map->notify.flags & UNMAP_NOTIFY_SEND_EVENT) { notify_remote_via_evtchn(map->notify.event); evtchn_put(map->notify.event); @@ -265,7 +310,7 @@ void gntdev_put_map(struct gntdev_priv *priv, struct gntdev_grant_map *map) static int find_grant_ptes(pte_t *pte, unsigned long addr, void *data) { struct gntdev_grant_map *map = data; - unsigned int pgnr = (addr - map->vma->vm_start) >> PAGE_SHIFT; + unsigned int pgnr = (addr - map->pages_vm_start) >> PAGE_SHIFT; int flags = map->flags | GNTMAP_application_map | GNTMAP_contains_pte | (1 << _GNTMAP_guest_avail0); u64 pte_maddr; @@ -273,6 +318,7 @@ static int find_grant_ptes(pte_t *pte, unsigned long addr, void *data) BUG_ON(pgnr >= map->count); pte_maddr = arbitrary_virt_to_machine(pte).maddr; + /* Note: this will perform a pte_mkspecial() through the hypercall. */ gnttab_set_map_op(&map->map_ops[pgnr], pte_maddr, flags, map->grants[pgnr].ref, map->grants[pgnr].domid); @@ -283,9 +329,10 @@ static int find_grant_ptes(pte_t *pte, unsigned long addr, void *data) int gntdev_map_grant_pages(struct gntdev_grant_map *map) { + size_t alloced = 0; int i, err = 0; - if (!use_ptemod) { + if (!xen_pv_domain()) { /* Note: it could already be mapped */ if (map->map_ops[0].handle != INVALID_GRANT_HANDLE) return 0; @@ -331,97 +378,130 @@ int gntdev_map_grant_pages(struct gntdev_grant_map *map) map->count); for (i = 0; i < map->count; i++) { - if (map->map_ops[i].status == GNTST_okay) + if (map->map_ops[i].status == GNTST_okay) { map->unmap_ops[i].handle = map->map_ops[i].handle; - else if (!err) + alloced++; + } else if (!err) err = -EINVAL; if (map->flags & GNTMAP_device_map) map->unmap_ops[i].dev_bus_addr = map->map_ops[i].dev_bus_addr; - if (use_ptemod) { - if (map->kmap_ops[i].status == GNTST_okay) + if (xen_pv_domain()) { + if (map->kmap_ops[i].status == GNTST_okay) { + alloced++; map->kunmap_ops[i].handle = map->kmap_ops[i].handle; - else if (!err) + } else if (!err) err = -EINVAL; } } + atomic_add(alloced, &map->live_grants); return err; } -static int __unmap_grant_pages(struct gntdev_grant_map *map, int offset, - int pages) +static void __unmap_grant_pages_done(int result, + struct gntab_unmap_queue_data *data) { - int i, err = 0; - struct gntab_unmap_queue_data unmap_data; + unsigned int i; + struct gntdev_grant_map *map = data->data; + unsigned int offset = data->unmap_ops - map->unmap_ops; + int successful_unmaps = 0; + int live_grants; + + for (i = 0; i < data->count; i++) { + if (map->unmap_ops[offset + i].status == GNTST_okay && + map->unmap_ops[offset + i].handle != INVALID_GRANT_HANDLE) + successful_unmaps++; + + WARN_ON(map->unmap_ops[offset + i].status != GNTST_okay && + map->unmap_ops[offset + i].handle != INVALID_GRANT_HANDLE); + pr_debug("unmap handle=%d st=%d\n", + map->unmap_ops[offset+i].handle, + map->unmap_ops[offset+i].status); + map->unmap_ops[offset+i].handle = INVALID_GRANT_HANDLE; + if (xen_pv_domain()) { + if (map->kunmap_ops[offset + i].status == GNTST_okay && + map->kunmap_ops[offset + i].handle != INVALID_GRANT_HANDLE) + successful_unmaps++; + + WARN_ON(map->kunmap_ops[offset + i].status != GNTST_okay && + map->kunmap_ops[offset + i].handle != INVALID_GRANT_HANDLE); + pr_debug("kunmap handle=%u st=%d\n", + map->kunmap_ops[offset+i].handle, + map->kunmap_ops[offset+i].status); + map->kunmap_ops[offset+i].handle = INVALID_GRANT_HANDLE; + } + } + + /* + * Decrease the live-grant counter. This must happen after the loop to + * prevent premature reuse of the grants by gnttab_mmap(). + */ + live_grants = atomic_sub_return(successful_unmaps, &map->live_grants); + if (WARN_ON(live_grants < 0)) + pr_err("%s: live_grants became negative (%d) after unmapping %d pages!\n", + __func__, live_grants, successful_unmaps); + + /* Release reference taken by __unmap_grant_pages */ + gntdev_put_map(NULL, map); +} +static void __unmap_grant_pages(struct gntdev_grant_map *map, int offset, + int pages) +{ if (map->notify.flags & UNMAP_NOTIFY_CLEAR_BYTE) { int pgno = (map->notify.addr >> PAGE_SHIFT); + if (pgno >= offset && pgno < offset + pages) { /* No need for kmap, pages are in lowmem */ uint8_t *tmp = pfn_to_kaddr(page_to_pfn(map->pages[pgno])); + tmp[map->notify.addr & (PAGE_SIZE-1)] = 0; map->notify.flags &= ~UNMAP_NOTIFY_CLEAR_BYTE; } } - unmap_data.unmap_ops = map->unmap_ops + offset; - unmap_data.kunmap_ops = use_ptemod ? map->kunmap_ops + offset : NULL; - unmap_data.pages = map->pages + offset; - unmap_data.count = pages; - - err = gnttab_unmap_refs_sync(&unmap_data); - if (err) - return err; + map->unmap_data.unmap_ops = map->unmap_ops + offset; + map->unmap_data.kunmap_ops = xen_pv_domain() ? map->kunmap_ops + offset : NULL; + map->unmap_data.pages = map->pages + offset; + map->unmap_data.count = pages; + map->unmap_data.done = __unmap_grant_pages_done; + map->unmap_data.data = map; + refcount_inc(&map->users); /* to keep map alive during async call below */ - for (i = 0; i < pages; i++) { - if (map->unmap_ops[offset+i].status) - err = -EINVAL; - pr_debug("unmap handle=%d st=%d\n", - map->unmap_ops[offset+i].handle, - map->unmap_ops[offset+i].status); - map->unmap_ops[offset+i].handle = INVALID_GRANT_HANDLE; - if (use_ptemod) { - if (map->kunmap_ops[offset+i].status) - err = -EINVAL; - pr_debug("kunmap handle=%u st=%d\n", - map->kunmap_ops[offset+i].handle, - map->kunmap_ops[offset+i].status); - map->kunmap_ops[offset+i].handle = INVALID_GRANT_HANDLE; - } - } - return err; + gnttab_unmap_refs_async(&map->unmap_data); } -static int unmap_grant_pages(struct gntdev_grant_map *map, int offset, - int pages) +static void unmap_grant_pages(struct gntdev_grant_map *map, int offset, + int pages) { - int range, err = 0; + int range; + + if (atomic_read(&map->live_grants) == 0) + return; /* Nothing to do */ pr_debug("unmap %d+%d [%d+%d]\n", map->index, map->count, offset, pages); /* It is possible the requested range will have a "hole" where we * already unmapped some of the grants. Only unmap valid ranges. */ - while (pages && !err) { - while (pages && - map->unmap_ops[offset].handle == INVALID_GRANT_HANDLE) { + while (pages) { + while (pages && map->being_removed[offset]) { offset++; pages--; } range = 0; while (range < pages) { - if (map->unmap_ops[offset + range].handle == - INVALID_GRANT_HANDLE) + if (map->being_removed[offset + range]) break; + map->being_removed[offset + range] = true; range++; } - err = __unmap_grant_pages(map, offset, range); + if (range) + __unmap_grant_pages(map, offset, range); offset += range; pages -= range; } - - return err; } /* ------------------------------------------------------------------ */ @@ -441,16 +521,12 @@ static void gntdev_vma_close(struct vm_area_struct *vma) struct gntdev_priv *priv = file->private_data; pr_debug("gntdev_vma_close %p\n", vma); - if (use_ptemod) { - WARN_ON(map->vma != vma); - mmu_interval_notifier_remove(&map->notifier); - map->vma = NULL; - } + vma->vm_private_data = NULL; gntdev_put_map(priv, map); } -static struct page *gntdev_vma_find_special_page(struct vm_area_struct *vma, +static struct page *gntdev_vma_find_normal_page(struct vm_area_struct *vma, unsigned long addr) { struct gntdev_grant_map *map = vma->vm_private_data; @@ -461,7 +537,7 @@ static struct page *gntdev_vma_find_special_page(struct vm_area_struct *vma, static const struct vm_operations_struct gntdev_vmops = { .open = gntdev_vma_open, .close = gntdev_vma_close, - .find_special_page = gntdev_vma_find_special_page, + .find_normal_page = gntdev_vma_find_normal_page, }; /* ------------------------------------------------------------------ */ @@ -473,31 +549,30 @@ static bool gntdev_invalidate(struct mmu_interval_notifier *mn, struct gntdev_grant_map *map = container_of(mn, struct gntdev_grant_map, notifier); unsigned long mstart, mend; - int err; + unsigned long map_start, map_end; if (!mmu_notifier_range_blockable(range)) return false; + map_start = map->pages_vm_start; + map_end = map->pages_vm_start + (map->count << PAGE_SHIFT); + /* * If the VMA is split or otherwise changed the notifier is not * updated, but we don't want to process VA's outside the modified * VMA. FIXME: It would be much more understandable to just prevent * modifying the VMA in the first place. */ - if (map->vma->vm_start >= range->end || - map->vma->vm_end <= range->start) + if (map_start >= range->end || map_end <= range->start) return true; - mstart = max(range->start, map->vma->vm_start); - mend = min(range->end, map->vma->vm_end); + mstart = max(range->start, map_start); + mend = min(range->end, map_end); pr_debug("map %d+%d (%lx %lx), range %lx %lx, mrange %lx %lx\n", - map->index, map->count, - map->vma->vm_start, map->vma->vm_end, - range->start, range->end, mstart, mend); - err = unmap_grant_pages(map, - (mstart - map->vma->vm_start) >> PAGE_SHIFT, - (mend - mstart) >> PAGE_SHIFT); - WARN_ON(err); + map->index, map->count, map_start, map_end, + range->start, range->end, mstart, mend); + unmap_grant_pages(map, (mstart - map_start) >> PAGE_SHIFT, + (mend - mstart) >> PAGE_SHIFT); return true; } @@ -519,6 +594,8 @@ static int gntdev_open(struct inode *inode, struct file *flip) INIT_LIST_HEAD(&priv->maps); mutex_init(&priv->lock); + mutex_init(&priv->batch_lock); + #ifdef CONFIG_XEN_GNTDEV_DMABUF priv->dmabuf_priv = gntdev_dmabuf_init(flip); if (IS_ERR(priv->dmabuf_priv)) { @@ -543,6 +620,7 @@ static int gntdev_release(struct inode *inode, struct file *flip) { struct gntdev_priv *priv = flip->private_data; struct gntdev_grant_map *map; + struct gntdev_copy_batch *batch; pr_debug("priv %p\n", priv); @@ -555,6 +633,14 @@ static int gntdev_release(struct inode *inode, struct file *flip) } mutex_unlock(&priv->lock); + mutex_lock(&priv->batch_lock); + while (priv->batch) { + batch = priv->batch; + priv->batch = batch->next; + kfree(batch); + } + mutex_unlock(&priv->batch_lock); + #ifdef CONFIG_XEN_GNTDEV_DMABUF gntdev_dmabuf_fini(priv->dmabuf_priv); #endif @@ -720,17 +806,6 @@ static long gntdev_ioctl_notify(struct gntdev_priv *priv, void __user *u) return rc; } -#define GNTDEV_COPY_BATCH 16 - -struct gntdev_copy_batch { - struct gnttab_copy ops[GNTDEV_COPY_BATCH]; - struct page *pages[GNTDEV_COPY_BATCH]; - s16 __user *status[GNTDEV_COPY_BATCH]; - unsigned int nr_ops; - unsigned int nr_pages; - bool writeable; -}; - static int gntdev_get_page(struct gntdev_copy_batch *batch, void __user *virt, unsigned long *gfn) { @@ -888,36 +963,53 @@ static int gntdev_grant_copy_seg(struct gntdev_copy_batch *batch, static long gntdev_ioctl_grant_copy(struct gntdev_priv *priv, void __user *u) { struct ioctl_gntdev_grant_copy copy; - struct gntdev_copy_batch batch; + struct gntdev_copy_batch *batch; unsigned int i; int ret = 0; if (copy_from_user(©, u, sizeof(copy))) return -EFAULT; - batch.nr_ops = 0; - batch.nr_pages = 0; + mutex_lock(&priv->batch_lock); + if (!priv->batch) { + batch = kmalloc(sizeof(*batch), GFP_KERNEL); + } else { + batch = priv->batch; + priv->batch = batch->next; + } + mutex_unlock(&priv->batch_lock); + if (!batch) + return -ENOMEM; + + batch->nr_ops = 0; + batch->nr_pages = 0; for (i = 0; i < copy.count; i++) { struct gntdev_grant_copy_segment seg; if (copy_from_user(&seg, ©.segments[i], sizeof(seg))) { ret = -EFAULT; + gntdev_put_pages(batch); goto out; } - ret = gntdev_grant_copy_seg(&batch, &seg, ©.segments[i].status); - if (ret < 0) + ret = gntdev_grant_copy_seg(batch, &seg, ©.segments[i].status); + if (ret < 0) { + gntdev_put_pages(batch); goto out; + } cond_resched(); } - if (batch.nr_ops) - ret = gntdev_copy(&batch); - return ret; + if (batch->nr_ops) + ret = gntdev_copy(batch); + + out: + mutex_lock(&priv->batch_lock); + batch->next = priv->batch; + priv->batch = batch; + mutex_unlock(&priv->batch_lock); - out: - gntdev_put_pages(&batch); return ret; } @@ -945,7 +1037,7 @@ static long gntdev_ioctl(struct file *flip, #ifdef CONFIG_XEN_GNTDEV_DMABUF case IOCTL_GNTDEV_DMABUF_EXP_FROM_REFS: - return gntdev_ioctl_dmabuf_exp_from_refs(priv, use_ptemod, ptr); + return gntdev_ioctl_dmabuf_exp_from_refs(priv, ptr); case IOCTL_GNTDEV_DMABUF_EXP_WAIT_RELEASED: return gntdev_ioctl_dmabuf_exp_wait_released(priv, ptr); @@ -977,22 +1069,23 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma) return -EINVAL; pr_debug("map %d+%d at %lx (pgoff %lx)\n", - index, count, vma->vm_start, vma->vm_pgoff); + index, count, vma->vm_start, vma->vm_pgoff); mutex_lock(&priv->lock); map = gntdev_find_map_index(priv, index, count); if (!map) goto unlock_out; - if (use_ptemod && map->vma) + if (!atomic_add_unless(&map->in_use, 1, 1)) goto unlock_out; + refcount_inc(&map->users); vma->vm_ops = &gntdev_vmops; - vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP | VM_MIXEDMAP; + vm_flags_set(vma, VM_DONTEXPAND | VM_DONTDUMP | VM_MIXEDMAP); - if (use_ptemod) - vma->vm_flags |= VM_DONTCOPY; + if (xen_pv_domain()) + vm_flags_set(vma, VM_DONTCOPY); vma->vm_private_data = map; if (map->flags) { @@ -1005,19 +1098,20 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma) map->flags |= GNTMAP_readonly; } - if (use_ptemod) { - map->vma = vma; + map->pages_vm_start = vma->vm_start; + + if (xen_pv_domain()) { err = mmu_interval_notifier_insert_locked( &map->notifier, vma->vm_mm, vma->vm_start, vma->vm_end - vma->vm_start, &gntdev_mmu_ops); - if (err) { - map->vma = NULL; + if (err) goto out_unlock_put; - } + + map->notifier_init = true; } mutex_unlock(&priv->lock); - if (use_ptemod) { + if (xen_pv_domain()) { /* * gntdev takes the address of the PTE in find_grant_ptes() and * passes it to the hypervisor in gntdev_map_grant_pages(). The @@ -1030,7 +1124,6 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma) */ mmu_interval_read_begin(&map->notifier); - map->pages_vm_start = vma->vm_start; err = apply_to_page_range(vma->vm_mm, vma->vm_start, vma->vm_end - vma->vm_start, find_grant_ptes, map); @@ -1044,7 +1137,7 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma) if (err) goto out_put_map; - if (!use_ptemod) { + if (!xen_pv_domain()) { err = vm_map_pages_zero(vma, map->pages, map->count); if (err) goto out_put_map; @@ -1059,13 +1152,8 @@ unlock_out: out_unlock_put: mutex_unlock(&priv->lock); out_put_map: - if (use_ptemod) { + if (xen_pv_domain()) unmap_grant_pages(map, 0, map->count); - if (map->vma) { - mmu_interval_notifier_remove(&map->notifier); - map->vma = NULL; - } - } gntdev_put_map(priv, map); return err; } @@ -1093,8 +1181,6 @@ static int __init gntdev_init(void) if (!xen_domain()) return -ENODEV; - use_ptemod = !xen_feature(XENFEAT_auto_translated_physmap); - err = misc_register(&gntdev_miscdev); if (err != 0) { pr_err("Could not register gntdev device\n"); |
