diff options
Diffstat (limited to 'drivers/xen/xenbus/xenbus_client.c')
| -rw-r--r-- | drivers/xen/xenbus/xenbus_client.c | 737 |
1 files changed, 475 insertions, 262 deletions
diff --git a/drivers/xen/xenbus/xenbus_client.c b/drivers/xen/xenbus/xenbus_client.c index ec097d6f964d..2dc874fb5506 100644 --- a/drivers/xen/xenbus/xenbus_client.c +++ b/drivers/xen/xenbus/xenbus_client.c @@ -37,7 +37,7 @@ #include <linux/vmalloc.h> #include <linux/export.h> #include <asm/xen/hypervisor.h> -#include <asm/xen/page.h> +#include <xen/page.h> #include <xen/interface/xen.h> #include <xen/interface/event_channel.h> #include <xen/balloon.h> @@ -45,23 +45,50 @@ #include <xen/grant_table.h> #include <xen/xenbus.h> #include <xen/xen.h> +#include <xen/features.h> -#include "xenbus_probe.h" +#include "xenbus.h" + +#define XENBUS_PAGES(_grants) (DIV_ROUND_UP(_grants, XEN_PFN_PER_PAGE)) + +#define XENBUS_MAX_RING_PAGES (XENBUS_PAGES(XENBUS_MAX_RING_GRANTS)) struct xenbus_map_node { struct list_head next; union { - struct vm_struct *area; /* PV */ - struct page *page; /* HVM */ + struct { + struct vm_struct *area; + } pv; + struct { + struct page *pages[XENBUS_MAX_RING_PAGES]; + unsigned long addrs[XENBUS_MAX_RING_GRANTS]; + void *addr; + } hvm; }; - grant_handle_t handle; + grant_handle_t handles[XENBUS_MAX_RING_GRANTS]; + unsigned int nr_handles; +}; + +struct map_ring_valloc { + struct xenbus_map_node *node; + + /* Why do we need two arrays? See comment of __xenbus_map_ring */ + unsigned long addrs[XENBUS_MAX_RING_GRANTS]; + phys_addr_t phys_addrs[XENBUS_MAX_RING_GRANTS]; + + struct gnttab_map_grant_ref map[XENBUS_MAX_RING_GRANTS]; + struct gnttab_unmap_grant_ref unmap[XENBUS_MAX_RING_GRANTS]; + + unsigned int idx; }; static DEFINE_SPINLOCK(xenbus_valloc_lock); static LIST_HEAD(xenbus_valloc_pages); struct xenbus_ring_ops { - int (*map)(struct xenbus_device *dev, int gnt, void **vaddr); + int (*map)(struct xenbus_device *dev, struct map_ring_valloc *info, + grant_ref_t *gnt_refs, unsigned int nr_grefs, + void **vaddr); int (*unmap)(struct xenbus_device *dev, void *vaddr); }; @@ -89,29 +116,36 @@ EXPORT_SYMBOL_GPL(xenbus_strstate); * @dev: xenbus device * @path: path to watch * @watch: watch to register + * @will_handle: events queuing determine callback * @callback: callback to register * * Register a @watch on the given path, using the given xenbus_watch structure - * for storage, and the given @callback function as the callback. Return 0 on - * success, or -errno on error. On success, the given @path will be saved as - * @watch->node, and remains the caller's to free. On error, @watch->node will - * be NULL, the device will switch to %XenbusStateClosing, and the error will - * be saved in the store. + * for storage, @will_handle function as the callback to determine if each + * event need to be queued, and the given @callback function as the callback. + * On success, the given @path will be saved as @watch->node, and remains the + * caller's to free. On error, @watch->node will be NULL, the device will + * switch to %XenbusStateClosing, and the error will be saved in the store. + * + * Returns: %0 on success or -errno on error */ int xenbus_watch_path(struct xenbus_device *dev, const char *path, struct xenbus_watch *watch, + bool (*will_handle)(struct xenbus_watch *, + const char *, const char *), void (*callback)(struct xenbus_watch *, - const char **, unsigned int)) + const char *, const char *)) { int err; watch->node = path; + watch->will_handle = will_handle; watch->callback = callback; err = register_xenbus_watch(watch); if (err) { watch->node = NULL; + watch->will_handle = NULL; watch->callback = NULL; xenbus_dev_fatal(dev, err, "adding watch on %s", path); } @@ -125,21 +159,27 @@ EXPORT_SYMBOL_GPL(xenbus_watch_path); * xenbus_watch_pathfmt - register a watch on a sprintf-formatted path * @dev: xenbus device * @watch: watch to register + * @will_handle: events queuing determine callback * @callback: callback to register * @pathfmt: format of path to watch * * Register a watch on the given @path, using the given xenbus_watch - * structure for storage, and the given @callback function as the callback. - * Return 0 on success, or -errno on error. On success, the watched path - * (@path/@path2) will be saved as @watch->node, and becomes the caller's to - * kfree(). On error, watch->node will be NULL, so the caller has nothing to + * structure for storage, @will_handle function as the callback to determine if + * each event need to be queued, and the given @callback function as the + * callback. On success, the watched path (@path/@path2) will be saved + * as @watch->node, and becomes the caller's to kfree(). + * On error, watch->node will be NULL, so the caller has nothing to * free, the device will switch to %XenbusStateClosing, and the error will be * saved in the store. + * + * Returns: %0 on success or -errno on error */ int xenbus_watch_pathfmt(struct xenbus_device *dev, struct xenbus_watch *watch, + bool (*will_handle)(struct xenbus_watch *, + const char *, const char *), void (*callback)(struct xenbus_watch *, - const char **, unsigned int), + const char *, const char *), const char *pathfmt, ...) { int err; @@ -154,7 +194,7 @@ int xenbus_watch_pathfmt(struct xenbus_device *dev, xenbus_dev_fatal(dev, -ENOMEM, "allocating path for watch"); return -ENOMEM; } - err = xenbus_watch_path(dev, path, watch, callback); + err = xenbus_watch_path(dev, path, watch, will_handle, callback); if (err) kfree(path); @@ -162,6 +202,7 @@ int xenbus_watch_pathfmt(struct xenbus_device *dev, } EXPORT_SYMBOL_GPL(xenbus_watch_pathfmt); +__printf(4, 5) static void xenbus_switch_fatal(struct xenbus_device *, int, int, const char *, ...); @@ -222,13 +263,15 @@ abort: } /** - * xenbus_switch_state + * xenbus_switch_state - save the new state of a driver * @dev: xenbus device * @state: new state * * Advertise in the store a change of the given driver to the given new_state. - * Return 0 on success, or -errno on error. On error, the device will switch - * to XenbusStateClosing, and the error will be saved in the store. + * On error, the device will switch to XenbusStateClosing, and the error + * will be saved in the store. + * + * Returns: %0 on success or -errno on error */ int xenbus_switch_state(struct xenbus_device *dev, enum xenbus_state state) { @@ -245,58 +288,35 @@ int xenbus_frontend_closed(struct xenbus_device *dev) } EXPORT_SYMBOL_GPL(xenbus_frontend_closed); -/** - * Return the path to the error node for the given device, or NULL on failure. - * If the value returned is non-NULL, then it is the caller's to kfree. - */ -static char *error_path(struct xenbus_device *dev) -{ - return kasprintf(GFP_KERNEL, "error/%s", dev->nodename); -} - - +__printf(3, 0) static void xenbus_va_dev_error(struct xenbus_device *dev, int err, const char *fmt, va_list ap) { - int ret; unsigned int len; - char *printf_buffer = NULL; - char *path_buffer = NULL; + char *printf_buffer; + char *path_buffer; #define PRINTF_BUFFER_SIZE 4096 + printf_buffer = kmalloc(PRINTF_BUFFER_SIZE, GFP_KERNEL); - if (printf_buffer == NULL) - goto fail; + if (!printf_buffer) + return; len = sprintf(printf_buffer, "%i ", -err); - ret = vsnprintf(printf_buffer+len, PRINTF_BUFFER_SIZE-len, fmt, ap); - - BUG_ON(len + ret > PRINTF_BUFFER_SIZE-1); + vsnprintf(printf_buffer + len, PRINTF_BUFFER_SIZE - len, fmt, ap); dev_err(&dev->dev, "%s\n", printf_buffer); - path_buffer = error_path(dev); + path_buffer = kasprintf(GFP_KERNEL, "error/%s", dev->nodename); + if (path_buffer) + xenbus_write(XBT_NIL, path_buffer, "error", printf_buffer); - if (path_buffer == NULL) { - dev_err(&dev->dev, "failed to write error node for %s (%s)\n", - dev->nodename, printf_buffer); - goto fail; - } - - if (xenbus_write(XBT_NIL, path_buffer, "error", printf_buffer) != 0) { - dev_err(&dev->dev, "failed to write error node for %s (%s)\n", - dev->nodename, printf_buffer); - goto fail; - } - -fail: kfree(printf_buffer); kfree(path_buffer); } - /** - * xenbus_dev_error + * xenbus_dev_error - place an error message into the store * @dev: xenbus device * @err: error to report * @fmt: error message format @@ -315,7 +335,7 @@ void xenbus_dev_error(struct xenbus_device *dev, int err, const char *fmt, ...) EXPORT_SYMBOL_GPL(xenbus_dev_error); /** - * xenbus_dev_fatal + * xenbus_dev_fatal - put an error messages into the store and then shutdown * @dev: xenbus device * @err: error to report * @fmt: error message format @@ -337,7 +357,7 @@ void xenbus_dev_fatal(struct xenbus_device *dev, int err, const char *fmt, ...) } EXPORT_SYMBOL_GPL(xenbus_dev_fatal); -/** +/* * Equivalent to xenbus_dev_fatal(dev, err, fmt, args), but helps * avoiding recursion within xenbus_switch_state. */ @@ -354,32 +374,103 @@ static void xenbus_switch_fatal(struct xenbus_device *dev, int depth, int err, __xenbus_switch_state(dev, XenbusStateClosing, 1); } -/** - * xenbus_grant_ring +/* + * xenbus_setup_ring * @dev: xenbus device - * @ring_mfn: mfn of ring to grant - - * Grant access to the given @ring_mfn to the peer of the given device. Return - * 0 on success, or -errno on error. On error, the device will switch to - * XenbusStateClosing, and the error will be saved in the store. + * @vaddr: pointer to starting virtual address of the ring + * @nr_pages: number of pages to be granted + * @grefs: grant reference array to be filled in + * + * Allocate physically contiguous pages for a shared ring buffer and grant it + * to the peer of the given device. The ring buffer is initially filled with + * zeroes. The virtual address of the ring is stored at @vaddr and the + * grant references are stored in the @grefs array. In case of error @vaddr + * will be set to NULL and @grefs will be filled with INVALID_GRANT_REF. */ -int xenbus_grant_ring(struct xenbus_device *dev, unsigned long ring_mfn) +int xenbus_setup_ring(struct xenbus_device *dev, gfp_t gfp, void **vaddr, + unsigned int nr_pages, grant_ref_t *grefs) { - int err = gnttab_grant_foreign_access(dev->otherend_id, ring_mfn, 0); - if (err < 0) - xenbus_dev_fatal(dev, err, "granting access to ring page"); - return err; + unsigned long ring_size = nr_pages * XEN_PAGE_SIZE; + grant_ref_t gref_head; + unsigned int i; + void *addr; + int ret; + + addr = *vaddr = alloc_pages_exact(ring_size, gfp | __GFP_ZERO); + if (!*vaddr) { + ret = -ENOMEM; + goto err; + } + + ret = gnttab_alloc_grant_references(nr_pages, &gref_head); + if (ret) { + xenbus_dev_fatal(dev, ret, "granting access to %u ring pages", + nr_pages); + goto err; + } + + for (i = 0; i < nr_pages; i++) { + unsigned long gfn; + + if (is_vmalloc_addr(*vaddr)) + gfn = pfn_to_gfn(vmalloc_to_pfn(addr)); + else + gfn = virt_to_gfn(addr); + + grefs[i] = gnttab_claim_grant_reference(&gref_head); + gnttab_grant_foreign_access_ref(grefs[i], dev->otherend_id, + gfn, 0); + + addr += XEN_PAGE_SIZE; + } + + return 0; + + err: + if (*vaddr) + free_pages_exact(*vaddr, ring_size); + for (i = 0; i < nr_pages; i++) + grefs[i] = INVALID_GRANT_REF; + *vaddr = NULL; + + return ret; } -EXPORT_SYMBOL_GPL(xenbus_grant_ring); +EXPORT_SYMBOL_GPL(xenbus_setup_ring); +/* + * xenbus_teardown_ring + * @vaddr: starting virtual address of the ring + * @nr_pages: number of pages + * @grefs: grant reference array + * + * Remove grants for the shared ring buffer and free the associated memory. + * On return the grant reference array is filled with INVALID_GRANT_REF. + */ +void xenbus_teardown_ring(void **vaddr, unsigned int nr_pages, + grant_ref_t *grefs) +{ + unsigned int i; -/** + for (i = 0; i < nr_pages; i++) { + if (grefs[i] != INVALID_GRANT_REF) { + gnttab_end_foreign_access(grefs[i], NULL); + grefs[i] = INVALID_GRANT_REF; + } + } + + if (*vaddr) + free_pages_exact(*vaddr, nr_pages * XEN_PAGE_SIZE); + *vaddr = NULL; +} +EXPORT_SYMBOL_GPL(xenbus_teardown_ring); + +/* * Allocate an event channel for the given xenbus_device, assigning the newly * created local port to *port. Return 0 on success, or -errno on error. On * error, the device will switch to XenbusStateClosing, and the error will be * saved in the store. */ -int xenbus_alloc_evtchn(struct xenbus_device *dev, int *port) +int xenbus_alloc_evtchn(struct xenbus_device *dev, evtchn_port_t *port) { struct evtchn_alloc_unbound alloc_unbound; int err; @@ -399,37 +490,10 @@ int xenbus_alloc_evtchn(struct xenbus_device *dev, int *port) EXPORT_SYMBOL_GPL(xenbus_alloc_evtchn); -/** - * Bind to an existing interdomain event channel in another domain. Returns 0 - * on success and stores the local port in *port. On error, returns -errno, - * switches the device to XenbusStateClosing, and saves the error in XenStore. - */ -int xenbus_bind_evtchn(struct xenbus_device *dev, int remote_port, int *port) -{ - struct evtchn_bind_interdomain bind_interdomain; - int err; - - bind_interdomain.remote_dom = dev->otherend_id; - bind_interdomain.remote_port = remote_port; - - err = HYPERVISOR_event_channel_op(EVTCHNOP_bind_interdomain, - &bind_interdomain); - if (err) - xenbus_dev_fatal(dev, err, - "binding to event channel %d from domain %d", - remote_port, dev->otherend_id); - else - *port = bind_interdomain.local_port; - - return err; -} -EXPORT_SYMBOL_GPL(xenbus_bind_evtchn); - - -/** +/* * Free an existing event channel. Returns 0 on success or -errno on error. */ -int xenbus_free_evtchn(struct xenbus_device *dev, int port) +int xenbus_free_evtchn(struct xenbus_device *dev, evtchn_port_t port) { struct evtchn_close close; int err; @@ -438,7 +502,7 @@ int xenbus_free_evtchn(struct xenbus_device *dev, int port) err = HYPERVISOR_event_channel_op(EVTCHNOP_close, &close); if (err) - xenbus_dev_error(dev, err, "freeing event channel %d", port); + xenbus_dev_error(dev, err, "freeing event channel %u", port); return err; } @@ -446,149 +510,222 @@ EXPORT_SYMBOL_GPL(xenbus_free_evtchn); /** - * xenbus_map_ring_valloc + * xenbus_map_ring_valloc - allocate & map pages of VA space * @dev: xenbus device - * @gnt_ref: grant reference + * @gnt_refs: grant reference array + * @nr_grefs: number of grant references * @vaddr: pointer to address to be filled out by mapping * - * Based on Rusty Russell's skeleton driver's map_page. - * Map a page of memory into this domain from another domain's grant table. - * xenbus_map_ring_valloc allocates a page of virtual address space, maps the - * page to that address, and sets *vaddr to that address. - * Returns 0 on success, and GNTST_* (see xen/include/interface/grant_table.h) - * or -ENOMEM on error. If an error is returned, device will switch to + * Map @nr_grefs pages of memory into this domain from another + * domain's grant table. xenbus_map_ring_valloc allocates @nr_grefs + * pages of virtual address space, maps the pages to that address, and sets + * *vaddr to that address. If an error is returned, device will switch to * XenbusStateClosing and the error message will be saved in XenStore. + * + * Returns: %0 on success or -errno on error */ -int xenbus_map_ring_valloc(struct xenbus_device *dev, int gnt_ref, void **vaddr) +int xenbus_map_ring_valloc(struct xenbus_device *dev, grant_ref_t *gnt_refs, + unsigned int nr_grefs, void **vaddr) { - return ring_ops->map(dev, gnt_ref, vaddr); + int err; + struct map_ring_valloc *info; + + *vaddr = NULL; + + if (nr_grefs > XENBUS_MAX_RING_GRANTS) + return -EINVAL; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return -ENOMEM; + + info->node = kzalloc(sizeof(*info->node), GFP_KERNEL); + if (!info->node) + err = -ENOMEM; + else + err = ring_ops->map(dev, info, gnt_refs, nr_grefs, vaddr); + + kfree(info->node); + kfree(info); + return err; } EXPORT_SYMBOL_GPL(xenbus_map_ring_valloc); -static int xenbus_map_ring_valloc_pv(struct xenbus_device *dev, - int gnt_ref, void **vaddr) +/* N.B. sizeof(phys_addr_t) doesn't always equal to sizeof(unsigned + * long), e.g. 32-on-64. Caller is responsible for preparing the + * right array to feed into this function */ +static int __xenbus_map_ring(struct xenbus_device *dev, + grant_ref_t *gnt_refs, + unsigned int nr_grefs, + grant_handle_t *handles, + struct map_ring_valloc *info, + unsigned int flags, + bool *leaked) { - struct gnttab_map_grant_ref op = { - .flags = GNTMAP_host_map | GNTMAP_contains_pte, - .ref = gnt_ref, - .dom = dev->otherend_id, - }; - struct xenbus_map_node *node; - struct vm_struct *area; - pte_t *pte; + int i, j; - *vaddr = NULL; + if (nr_grefs > XENBUS_MAX_RING_GRANTS) + return -EINVAL; - node = kzalloc(sizeof(*node), GFP_KERNEL); - if (!node) - return -ENOMEM; - - area = alloc_vm_area(PAGE_SIZE, &pte); - if (!area) { - kfree(node); - return -ENOMEM; + for (i = 0; i < nr_grefs; i++) { + gnttab_set_map_op(&info->map[i], info->phys_addrs[i], flags, + gnt_refs[i], dev->otherend_id); + handles[i] = INVALID_GRANT_HANDLE; } - op.host_addr = arbitrary_virt_to_machine(pte).maddr; + gnttab_batch_map(info->map, i); - gnttab_batch_map(&op, 1); + for (i = 0; i < nr_grefs; i++) { + if (info->map[i].status != GNTST_okay) { + xenbus_dev_fatal(dev, info->map[i].status, + "mapping in shared page %d from domain %d", + gnt_refs[i], dev->otherend_id); + goto fail; + } else + handles[i] = info->map[i].handle; + } - if (op.status != GNTST_okay) { - free_vm_area(area); - kfree(node); - xenbus_dev_fatal(dev, op.status, - "mapping in shared page %d from domain %d", - gnt_ref, dev->otherend_id); - return op.status; + return 0; + + fail: + for (i = j = 0; i < nr_grefs; i++) { + if (handles[i] != INVALID_GRANT_HANDLE) { + gnttab_set_unmap_op(&info->unmap[j], + info->phys_addrs[i], + GNTMAP_host_map, handles[i]); + j++; + } } - node->handle = op.handle; - node->area = area; + BUG_ON(HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, info->unmap, j)); - spin_lock(&xenbus_valloc_lock); - list_add(&node->next, &xenbus_valloc_pages); - spin_unlock(&xenbus_valloc_lock); + *leaked = false; + for (i = 0; i < j; i++) { + if (info->unmap[i].status != GNTST_okay) { + *leaked = true; + break; + } + } - *vaddr = area->addr; - return 0; + return -ENOENT; } -static int xenbus_map_ring_valloc_hvm(struct xenbus_device *dev, - int gnt_ref, void **vaddr) +/** + * xenbus_unmap_ring - unmap memory from another domain + * @dev: xenbus device + * @handles: grant handle array + * @nr_handles: number of handles in the array + * @vaddrs: addresses to unmap + * + * Unmap memory in this domain that was imported from another domain. + * + * Returns: %0 on success or GNTST_* on error + * (see xen/include/interface/grant_table.h). + */ +static int xenbus_unmap_ring(struct xenbus_device *dev, grant_handle_t *handles, + unsigned int nr_handles, unsigned long *vaddrs) { - struct xenbus_map_node *node; + struct gnttab_unmap_grant_ref unmap[XENBUS_MAX_RING_GRANTS]; + int i; int err; - void *addr; - *vaddr = NULL; + if (nr_handles > XENBUS_MAX_RING_GRANTS) + return -EINVAL; - node = kzalloc(sizeof(*node), GFP_KERNEL); - if (!node) - return -ENOMEM; + for (i = 0; i < nr_handles; i++) + gnttab_set_unmap_op(&unmap[i], vaddrs[i], + GNTMAP_host_map, handles[i]); + + BUG_ON(HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, unmap, i)); + + err = GNTST_okay; + for (i = 0; i < nr_handles; i++) { + if (unmap[i].status != GNTST_okay) { + xenbus_dev_error(dev, unmap[i].status, + "unmapping page at handle %d error %d", + handles[i], unmap[i].status); + err = unmap[i].status; + break; + } + } - err = alloc_xenballooned_pages(1, &node->page, false /* lowmem */); + return err; +} + +static void xenbus_map_ring_setup_grant_hvm(unsigned long gfn, + unsigned int goffset, + unsigned int len, + void *data) +{ + struct map_ring_valloc *info = data; + unsigned long vaddr = (unsigned long)gfn_to_virt(gfn); + + info->phys_addrs[info->idx] = vaddr; + info->addrs[info->idx] = vaddr; + + info->idx++; +} + +static int xenbus_map_ring_hvm(struct xenbus_device *dev, + struct map_ring_valloc *info, + grant_ref_t *gnt_ref, + unsigned int nr_grefs, + void **vaddr) +{ + struct xenbus_map_node *node = info->node; + int err; + void *addr; + bool leaked = false; + unsigned int nr_pages = XENBUS_PAGES(nr_grefs); + + err = xen_alloc_unpopulated_pages(nr_pages, node->hvm.pages); if (err) goto out_err; - addr = pfn_to_kaddr(page_to_pfn(node->page)); + gnttab_foreach_grant(node->hvm.pages, nr_grefs, + xenbus_map_ring_setup_grant_hvm, + info); + + err = __xenbus_map_ring(dev, gnt_ref, nr_grefs, node->handles, + info, GNTMAP_host_map, &leaked); + node->nr_handles = nr_grefs; - err = xenbus_map_ring(dev, gnt_ref, &node->handle, addr); if (err) - goto out_err_free_ballooned_pages; + goto out_free_ballooned_pages; + + addr = vmap(node->hvm.pages, nr_pages, VM_MAP | VM_IOREMAP, + PAGE_KERNEL); + if (!addr) { + err = -ENOMEM; + goto out_xenbus_unmap_ring; + } + + node->hvm.addr = addr; spin_lock(&xenbus_valloc_lock); list_add(&node->next, &xenbus_valloc_pages); spin_unlock(&xenbus_valloc_lock); *vaddr = addr; + info->node = NULL; + return 0; - out_err_free_ballooned_pages: - free_xenballooned_pages(1, &node->page); + out_xenbus_unmap_ring: + if (!leaked) + xenbus_unmap_ring(dev, node->handles, nr_grefs, info->addrs); + else + pr_alert("leaking %p size %u page(s)", + addr, nr_pages); + out_free_ballooned_pages: + if (!leaked) + xen_free_unpopulated_pages(nr_pages, node->hvm.pages); out_err: - kfree(node); return err; } - -/** - * xenbus_map_ring - * @dev: xenbus device - * @gnt_ref: grant reference - * @handle: pointer to grant handle to be filled - * @vaddr: address to be mapped to - * - * Map a page of memory into this domain from another domain's grant table. - * xenbus_map_ring does not allocate the virtual address space (you must do - * this yourself!). It only maps in the page to the specified address. - * Returns 0 on success, and GNTST_* (see xen/include/interface/grant_table.h) - * or -ENOMEM on error. If an error is returned, device will switch to - * XenbusStateClosing and the error message will be saved in XenStore. - */ -int xenbus_map_ring(struct xenbus_device *dev, int gnt_ref, - grant_handle_t *handle, void *vaddr) -{ - struct gnttab_map_grant_ref op; - - gnttab_set_map_op(&op, (unsigned long)vaddr, GNTMAP_host_map, gnt_ref, - dev->otherend_id); - - gnttab_batch_map(&op, 1); - - if (op.status != GNTST_okay) { - xenbus_dev_fatal(dev, op.status, - "mapping in shared page %d from domain %d", - gnt_ref, dev->otherend_id); - } else - *handle = op.handle; - - return op.status; -} -EXPORT_SYMBOL_GPL(xenbus_map_ring); - - /** - * xenbus_unmap_ring_vfree + * xenbus_unmap_ring_vfree - unmap a page of memory from another domain * @dev: xenbus device * @vaddr: addr to unmap * @@ -596,7 +733,8 @@ EXPORT_SYMBOL_GPL(xenbus_map_ring); * Unmap a page of memory in this domain that was imported from another domain. * Use xenbus_unmap_ring_vfree if you mapped in your memory with * xenbus_map_ring_valloc (it will free the virtual address space). - * Returns 0 on success and returns GNTST_* on error + * + * Returns: %0 on success or GNTST_* on error * (see xen/include/interface/grant_table.h). */ int xenbus_unmap_ring_vfree(struct xenbus_device *dev, void *vaddr) @@ -605,17 +743,71 @@ int xenbus_unmap_ring_vfree(struct xenbus_device *dev, void *vaddr) } EXPORT_SYMBOL_GPL(xenbus_unmap_ring_vfree); -static int xenbus_unmap_ring_vfree_pv(struct xenbus_device *dev, void *vaddr) +#ifdef CONFIG_XEN_PV +static int map_ring_apply(pte_t *pte, unsigned long addr, void *data) +{ + struct map_ring_valloc *info = data; + + info->phys_addrs[info->idx++] = arbitrary_virt_to_machine(pte).maddr; + return 0; +} + +static int xenbus_map_ring_pv(struct xenbus_device *dev, + struct map_ring_valloc *info, + grant_ref_t *gnt_refs, + unsigned int nr_grefs, + void **vaddr) +{ + struct xenbus_map_node *node = info->node; + struct vm_struct *area; + bool leaked = false; + int err = -ENOMEM; + + area = get_vm_area(XEN_PAGE_SIZE * nr_grefs, VM_IOREMAP); + if (!area) + return -ENOMEM; + if (apply_to_page_range(&init_mm, (unsigned long)area->addr, + XEN_PAGE_SIZE * nr_grefs, map_ring_apply, info)) + goto failed; + err = __xenbus_map_ring(dev, gnt_refs, nr_grefs, node->handles, + info, GNTMAP_host_map | GNTMAP_contains_pte, + &leaked); + if (err) + goto failed; + + node->nr_handles = nr_grefs; + node->pv.area = area; + + spin_lock(&xenbus_valloc_lock); + list_add(&node->next, &xenbus_valloc_pages); + spin_unlock(&xenbus_valloc_lock); + + *vaddr = area->addr; + info->node = NULL; + + return 0; + +failed: + if (!leaked) + free_vm_area(area); + else + pr_alert("leaking VM area %p size %u page(s)", area, nr_grefs); + + return err; +} + +static int xenbus_unmap_ring_pv(struct xenbus_device *dev, void *vaddr) { struct xenbus_map_node *node; - struct gnttab_unmap_grant_ref op = { - .host_addr = (unsigned long)vaddr, - }; + struct gnttab_unmap_grant_ref unmap[XENBUS_MAX_RING_GRANTS]; unsigned int level; + int i; + bool leaked = false; + int err; spin_lock(&xenbus_valloc_lock); list_for_each_entry(node, &xenbus_valloc_pages, next) { - if (node->area->addr == vaddr) { + if (node->pv.area->addr == vaddr) { list_del(&node->next); goto found; } @@ -630,33 +822,79 @@ static int xenbus_unmap_ring_vfree_pv(struct xenbus_device *dev, void *vaddr) return GNTST_bad_virt_addr; } - op.handle = node->handle; - op.host_addr = arbitrary_virt_to_machine( - lookup_address((unsigned long)vaddr, &level)).maddr; + for (i = 0; i < node->nr_handles; i++) { + unsigned long addr; - if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1)) - BUG(); + memset(&unmap[i], 0, sizeof(unmap[i])); + addr = (unsigned long)vaddr + (XEN_PAGE_SIZE * i); + unmap[i].host_addr = arbitrary_virt_to_machine( + lookup_address(addr, &level)).maddr; + unmap[i].dev_bus_addr = 0; + unmap[i].handle = node->handles[i]; + } - if (op.status == GNTST_okay) - free_vm_area(node->area); + BUG_ON(HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, unmap, i)); + + err = GNTST_okay; + leaked = false; + for (i = 0; i < node->nr_handles; i++) { + if (unmap[i].status != GNTST_okay) { + leaked = true; + xenbus_dev_error(dev, unmap[i].status, + "unmapping page at handle %d error %d", + node->handles[i], unmap[i].status); + err = unmap[i].status; + break; + } + } + + if (!leaked) + free_vm_area(node->pv.area); else - xenbus_dev_error(dev, op.status, - "unmapping page at handle %d error %d", - node->handle, op.status); + pr_alert("leaking VM area %p size %u page(s)", + node->pv.area, node->nr_handles); kfree(node); - return op.status; + return err; } -static int xenbus_unmap_ring_vfree_hvm(struct xenbus_device *dev, void *vaddr) +static const struct xenbus_ring_ops ring_ops_pv = { + .map = xenbus_map_ring_pv, + .unmap = xenbus_unmap_ring_pv, +}; +#endif + +struct unmap_ring_hvm +{ + unsigned int idx; + unsigned long addrs[XENBUS_MAX_RING_GRANTS]; +}; + +static void xenbus_unmap_ring_setup_grant_hvm(unsigned long gfn, + unsigned int goffset, + unsigned int len, + void *data) +{ + struct unmap_ring_hvm *info = data; + + info->addrs[info->idx] = (unsigned long)gfn_to_virt(gfn); + + info->idx++; +} + +static int xenbus_unmap_ring_hvm(struct xenbus_device *dev, void *vaddr) { int rv; struct xenbus_map_node *node; void *addr; + struct unmap_ring_hvm info = { + .idx = 0, + }; + unsigned int nr_pages; spin_lock(&xenbus_valloc_lock); list_for_each_entry(node, &xenbus_valloc_pages, next) { - addr = pfn_to_kaddr(page_to_pfn(node->page)); + addr = node->hvm.addr; if (addr == vaddr) { list_del(&node->next); goto found; @@ -672,52 +910,30 @@ static int xenbus_unmap_ring_vfree_hvm(struct xenbus_device *dev, void *vaddr) return GNTST_bad_virt_addr; } - rv = xenbus_unmap_ring(dev, node->handle, addr); + nr_pages = XENBUS_PAGES(node->nr_handles); + + gnttab_foreach_grant(node->hvm.pages, node->nr_handles, + xenbus_unmap_ring_setup_grant_hvm, + &info); - if (!rv) - free_xenballooned_pages(1, &node->page); + rv = xenbus_unmap_ring(dev, node->handles, node->nr_handles, + info.addrs); + if (!rv) { + vunmap(vaddr); + xen_free_unpopulated_pages(nr_pages, node->hvm.pages); + } else - WARN(1, "Leaking %p\n", vaddr); + WARN(1, "Leaking %p, size %u page(s)\n", vaddr, nr_pages); kfree(node); return rv; } /** - * xenbus_unmap_ring - * @dev: xenbus device - * @handle: grant handle - * @vaddr: addr to unmap - * - * Unmap a page of memory in this domain that was imported from another domain. - * Returns 0 on success and returns GNTST_* on error - * (see xen/include/interface/grant_table.h). - */ -int xenbus_unmap_ring(struct xenbus_device *dev, - grant_handle_t handle, void *vaddr) -{ - struct gnttab_unmap_grant_ref op; - - gnttab_set_unmap_op(&op, (unsigned long)vaddr, GNTMAP_host_map, handle); - - if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1)) - BUG(); - - if (op.status != GNTST_okay) - xenbus_dev_error(dev, op.status, - "unmapping page at handle %d error %d", - handle, op.status); - - return op.status; -} -EXPORT_SYMBOL_GPL(xenbus_unmap_ring); - - -/** - * xenbus_read_driver_state + * xenbus_read_driver_state - read state from a store path * @path: path for driver * - * Return the state of the driver rooted at the given store path, or + * Returns: the state of the driver rooted at the given store path, or * XenbusStateUnknown if no state can be read. */ enum xenbus_state xenbus_read_driver_state(const char *path) @@ -731,20 +947,17 @@ enum xenbus_state xenbus_read_driver_state(const char *path) } EXPORT_SYMBOL_GPL(xenbus_read_driver_state); -static const struct xenbus_ring_ops ring_ops_pv = { - .map = xenbus_map_ring_valloc_pv, - .unmap = xenbus_unmap_ring_vfree_pv, -}; - static const struct xenbus_ring_ops ring_ops_hvm = { - .map = xenbus_map_ring_valloc_hvm, - .unmap = xenbus_unmap_ring_vfree_hvm, + .map = xenbus_map_ring_hvm, + .unmap = xenbus_unmap_ring_hvm, }; void __init xenbus_ring_ops_init(void) { +#ifdef CONFIG_XEN_PV if (xen_pv_domain()) ring_ops = &ring_ops_pv; else +#endif ring_ops = &ring_ops_hvm; } |