mirrors/qemu
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'sstabellini/xen-2013-01-14' into staging

Browse Source 
* sstabellini/xen-2013-01-14:
  xen_disk: implement BLKIF_OP_FLUSH_DISKCACHE, remove BLKIF_OP_WRITE_BARRIER
  xen_disk: add persistent grant support to xen_disk backend
  xen_disk: fix memory leak

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This commit is contained in:
Anthony Liguori 2013-01-15 16:49:18 -06:00
commit 5e72179b8f
1 changed files with 184 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

208
hw/xen_disk.c
View File

@ -51,6 +51,13 @@ static int max_requests = 32;
#define BLOCK_SIZE 512 #define BLOCK_SIZE 512
#define IOCB_COUNT (BLKIF_MAX_SEGMENTS_PER_REQUEST + 2) #define IOCB_COUNT (BLKIF_MAX_SEGMENTS_PER_REQUEST + 2)
struct PersistentGrant {
void *page;
struct XenBlkDev *blkdev;
};
typedef struct PersistentGrant PersistentGrant;
struct ioreq { struct ioreq {
blkif_request_t req; blkif_request_t req;
int16_t status; int16_t status;
@ -68,6 +75,7 @@ struct ioreq {
int prot; int prot;
void *page[BLKIF_MAX_SEGMENTS_PER_REQUEST]; void *page[BLKIF_MAX_SEGMENTS_PER_REQUEST];
void *pages; void *pages;
int num_unmap;
/* aio status */ /* aio status */
int aio_inflight; int aio_inflight;
@ -104,6 +112,12 @@ struct XenBlkDev {
int requests_inflight; int requests_inflight;
int requests_finished; int requests_finished;
/* Persistent grants extension */
gboolean feature_persistent;
GTree *persistent_gnts;
unsigned int persistent_gnt_count;
unsigned int max_grants;
/* qemu block driver */ /* qemu block driver */
DriveInfo *dinfo; DriveInfo *dinfo;
BlockDriverState *bs; BlockDriverState *bs;
@ -112,6 +126,54 @@ struct XenBlkDev {
/* ------------------------------------------------------------- */ /* ------------------------------------------------------------- */
static void ioreq_reset(struct ioreq *ioreq)
{
memset(&ioreq->req, 0, sizeof(ioreq->req));
ioreq->status = 0;
ioreq->start = 0;
ioreq->presync = 0;
ioreq->postsync = 0;
ioreq->mapped = 0;
memset(ioreq->domids, 0, sizeof(ioreq->domids));
memset(ioreq->refs, 0, sizeof(ioreq->refs));
ioreq->prot = 0;
memset(ioreq->page, 0, sizeof(ioreq->page));
ioreq->pages = NULL;
ioreq->aio_inflight = 0;
ioreq->aio_errors = 0;
ioreq->blkdev = NULL;
memset(&ioreq->list, 0, sizeof(ioreq->list));
memset(&ioreq->acct, 0, sizeof(ioreq->acct));
qemu_iovec_reset(&ioreq->v);
}
static gint int_cmp(gconstpointer a, gconstpointer b, gpointer user_data)
{
uint ua = GPOINTER_TO_UINT(a);
uint ub = GPOINTER_TO_UINT(b);
return (ua > ub) - (ua < ub);
}
static void destroy_grant(gpointer pgnt)
{
PersistentGrant *grant = pgnt;
XenGnttab gnt = grant->blkdev->xendev.gnttabdev;
if (xc_gnttab_munmap(gnt, grant->page, 1) != 0) {
xen_be_printf(&grant->blkdev->xendev, 0,
"xc_gnttab_munmap failed: %s\n",
strerror(errno));
}
grant->blkdev->persistent_gnt_count--;
xen_be_printf(&grant->blkdev->xendev, 3,
"unmapped grant %p\n", grant->page);
g_free(grant);
}
static struct ioreq *ioreq_start(struct XenBlkDev *blkdev) static struct ioreq *ioreq_start(struct XenBlkDev *blkdev)
{ {
struct ioreq *ioreq = NULL; struct ioreq *ioreq = NULL;
@ -129,7 +191,6 @@ static struct ioreq *ioreq_start(struct XenBlkDev *blkdev)
/* get one from freelist */ /* get one from freelist */
ioreq = QLIST_FIRST(&blkdev->freelist); ioreq = QLIST_FIRST(&blkdev->freelist);
QLIST_REMOVE(ioreq, list); QLIST_REMOVE(ioreq, list);
qemu_iovec_reset(&ioreq->v);
} }
QLIST_INSERT_HEAD(&blkdev->inflight, ioreq, list); QLIST_INSERT_HEAD(&blkdev->inflight, ioreq, list);
blkdev->requests_inflight++; blkdev->requests_inflight++;
@ -153,7 +214,7 @@ static void ioreq_release(struct ioreq *ioreq, bool finish)
struct XenBlkDev *blkdev = ioreq->blkdev; struct XenBlkDev *blkdev = ioreq->blkdev;
QLIST_REMOVE(ioreq, list); QLIST_REMOVE(ioreq, list);
memset(ioreq, 0, sizeof(*ioreq)); ioreq_reset(ioreq);
ioreq->blkdev = blkdev; ioreq->blkdev = blkdev;
QLIST_INSERT_HEAD(&blkdev->freelist, ioreq, list); QLIST_INSERT_HEAD(&blkdev->freelist, ioreq, list);
if (finish) { if (finish) {
@ -182,12 +243,11 @@ static int ioreq_parse(struct ioreq *ioreq)
case BLKIF_OP_READ: case BLKIF_OP_READ:
ioreq->prot = PROT_WRITE; /* to memory */ ioreq->prot = PROT_WRITE; /* to memory */
break; break;
case BLKIF_OP_WRITE_BARRIER: case BLKIF_OP_FLUSH_DISKCACHE:
ioreq->presync = 1;
if (!ioreq->req.nr_segments) { if (!ioreq->req.nr_segments) {
ioreq->presync = 1;
return 0; return 0;
} }
ioreq->presync = ioreq->postsync = 1;
/* fall through */ /* fall through */
case BLKIF_OP_WRITE: case BLKIF_OP_WRITE:
ioreq->prot = PROT_READ; /* from memory */ ioreq->prot = PROT_READ; /* from memory */
@ -241,21 +301,21 @@ static void ioreq_unmap(struct ioreq *ioreq)
XenGnttab gnt = ioreq->blkdev->xendev.gnttabdev; XenGnttab gnt = ioreq->blkdev->xendev.gnttabdev;
int i; int i;
if (ioreq->v.niov == 0 || ioreq->mapped == 0) { if (ioreq->num_unmap == 0 || ioreq->mapped == 0) {
return; return;
} }
if (batch_maps) { if (batch_maps) {
if (!ioreq->pages) { if (!ioreq->pages) {
return; return;
} }
if (xc_gnttab_munmap(gnt, ioreq->pages, ioreq->v.niov) != 0) { if (xc_gnttab_munmap(gnt, ioreq->pages, ioreq->num_unmap) != 0) {
xen_be_printf(&ioreq->blkdev->xendev, 0, "xc_gnttab_munmap failed: %s\n", xen_be_printf(&ioreq->blkdev->xendev, 0, "xc_gnttab_munmap failed: %s\n",
strerror(errno)); strerror(errno));
} }
ioreq->blkdev->cnt_map -= ioreq->v.niov; ioreq->blkdev->cnt_map -= ioreq->num_unmap;
ioreq->pages = NULL; ioreq->pages = NULL;
} else { } else {
for (i = 0; i < ioreq->v.niov; i++) { for (i = 0; i < ioreq->num_unmap; i++) {
if (!ioreq->page[i]) { if (!ioreq->page[i]) {
continue; continue;
} }
@ -273,41 +333,120 @@ static void ioreq_unmap(struct ioreq *ioreq)
static int ioreq_map(struct ioreq *ioreq) static int ioreq_map(struct ioreq *ioreq)
{ {
XenGnttab gnt = ioreq->blkdev->xendev.gnttabdev; XenGnttab gnt = ioreq->blkdev->xendev.gnttabdev;
int i; uint32_t domids[BLKIF_MAX_SEGMENTS_PER_REQUEST];
uint32_t refs[BLKIF_MAX_SEGMENTS_PER_REQUEST];
void *page[BLKIF_MAX_SEGMENTS_PER_REQUEST];
int i, j, new_maps = 0;
PersistentGrant *grant;
/* domids and refs variables will contain the information necessary
* to map the grants that are needed to fulfill this request.
*
* After mapping the needed grants, the page array will contain the
* memory address of each granted page in the order specified in ioreq
* (disregarding if it's a persistent grant or not).
*/
if (ioreq->v.niov == 0 || ioreq->mapped == 1) { if (ioreq->v.niov == 0 || ioreq->mapped == 1) {
return 0; return 0;
} }
if (batch_maps) { if (ioreq->blkdev->feature_persistent) {
for (i = 0; i < ioreq->v.niov; i++) {
grant = g_tree_lookup(ioreq->blkdev->persistent_gnts,
GUINT_TO_POINTER(ioreq->refs[i]));
if (grant != NULL) {
page[i] = grant->page;
xen_be_printf(&ioreq->blkdev->xendev, 3,
"using persistent-grant %" PRIu32 "\n",
ioreq->refs[i]);
} else {
/* Add the grant to the list of grants that
* should be mapped
*/
domids[new_maps] = ioreq->domids[i];
refs[new_maps] = ioreq->refs[i];
page[i] = NULL;
new_maps++;
}
}
/* Set the protection to RW, since grants may be reused later
* with a different protection than the one needed for this request
*/
ioreq->prot = PROT_WRITE | PROT_READ;
} else {
/* All grants in the request should be mapped */
memcpy(refs, ioreq->refs, sizeof(refs));
memcpy(domids, ioreq->domids, sizeof(domids));
memset(page, 0, sizeof(page));
new_maps = ioreq->v.niov;
}
if (batch_maps && new_maps) {
ioreq->pages = xc_gnttab_map_grant_refs ioreq->pages = xc_gnttab_map_grant_refs
(gnt, ioreq->v.niov, ioreq->domids, ioreq->refs, ioreq->prot); (gnt, new_maps, domids, refs, ioreq->prot);
if (ioreq->pages == NULL) { if (ioreq->pages == NULL) {
xen_be_printf(&ioreq->blkdev->xendev, 0, xen_be_printf(&ioreq->blkdev->xendev, 0,
"can't map %d grant refs (%s, %d maps)\n", "can't map %d grant refs (%s, %d maps)\n",
ioreq->v.niov, strerror(errno), ioreq->blkdev->cnt_map); new_maps, strerror(errno), ioreq->blkdev->cnt_map);
return -1; return -1;
} }
for (i = 0; i < ioreq->v.niov; i++) { for (i = 0, j = 0; i < ioreq->v.niov; i++) {
ioreq->v.iov[i].iov_base = ioreq->pages + i * XC_PAGE_SIZE + if (page[i] == NULL) {
(uintptr_t)ioreq->v.iov[i].iov_base; page[i] = ioreq->pages + (j++) * XC_PAGE_SIZE;
}
} }
ioreq->blkdev->cnt_map += ioreq->v.niov; ioreq->blkdev->cnt_map += new_maps;
} else { } else if (new_maps) {
for (i = 0; i < ioreq->v.niov; i++) { for (i = 0; i < new_maps; i++) {
ioreq->page[i] = xc_gnttab_map_grant_ref ioreq->page[i] = xc_gnttab_map_grant_ref
(gnt, ioreq->domids[i], ioreq->refs[i], ioreq->prot); (gnt, domids[i], refs[i], ioreq->prot);
if (ioreq->page[i] == NULL) { if (ioreq->page[i] == NULL) {
xen_be_printf(&ioreq->blkdev->xendev, 0, xen_be_printf(&ioreq->blkdev->xendev, 0,
"can't map grant ref %d (%s, %d maps)\n", "can't map grant ref %d (%s, %d maps)\n",
ioreq->refs[i], strerror(errno), ioreq->blkdev->cnt_map); refs[i], strerror(errno), ioreq->blkdev->cnt_map);
ioreq_unmap(ioreq); ioreq_unmap(ioreq);
return -1; return -1;
} }
ioreq->v.iov[i].iov_base = ioreq->page[i] + (uintptr_t)ioreq->v.iov[i].iov_base;
ioreq->blkdev->cnt_map++; ioreq->blkdev->cnt_map++;
} }
for (i = 0, j = 0; i < ioreq->v.niov; i++) {
if (page[i] == NULL) {
page[i] = ioreq->page[j++];
}
}
}
if (ioreq->blkdev->feature_persistent) {
while ((ioreq->blkdev->persistent_gnt_count < ioreq->blkdev->max_grants)
&& new_maps) {
/* Go through the list of newly mapped grants and add as many
* as possible to the list of persistently mapped grants.
*
* Since we start at the end of ioreq->page(s), we only need
* to decrease new_maps to prevent this granted pages from
* being unmapped in ioreq_unmap.
*/
grant = g_malloc0(sizeof(*grant));
new_maps--;
if (batch_maps) {
grant->page = ioreq->pages + (new_maps) * XC_PAGE_SIZE;
} else {
grant->page = ioreq->page[new_maps];
}
grant->blkdev = ioreq->blkdev;
xen_be_printf(&ioreq->blkdev->xendev, 3,
"adding grant %" PRIu32 " page: %p\n",
refs[new_maps], grant->page);
g_tree_insert(ioreq->blkdev->persistent_gnts,
GUINT_TO_POINTER(refs[new_maps]),
grant);
ioreq->blkdev->persistent_gnt_count++;
}
}
for (i = 0; i < ioreq->v.niov; i++) {
ioreq->v.iov[i].iov_base += (uintptr_t)page[i];
} }
ioreq->mapped = 1; ioreq->mapped = 1;
ioreq->num_unmap = new_maps;
return 0; return 0;
} }
@ -369,7 +508,7 @@ static int ioreq_runio_qemu_aio(struct ioreq *ioreq)
qemu_aio_complete, ioreq); qemu_aio_complete, ioreq);
break; break;
case BLKIF_OP_WRITE: case BLKIF_OP_WRITE:
case BLKIF_OP_WRITE_BARRIER: case BLKIF_OP_FLUSH_DISKCACHE:
if (!ioreq->req.nr_segments) { if (!ioreq->req.nr_segments) {
break; break;
} }
@ -654,7 +793,8 @@ static int blk_init(struct XenDevice *xendev)
blkdev->file_size, blkdev->file_size >> 20); blkdev->file_size, blkdev->file_size >> 20);
/* fill info */ /* fill info */
xenstore_write_be_int(&blkdev->xendev, "feature-barrier", 1); xenstore_write_be_int(&blkdev->xendev, "feature-flush-cache", 1);
xenstore_write_be_int(&blkdev->xendev, "feature-persistent", 1);
xenstore_write_be_int(&blkdev->xendev, "info", info); xenstore_write_be_int(&blkdev->xendev, "info", info);
xenstore_write_be_int(&blkdev->xendev, "sector-size", blkdev->file_blk); xenstore_write_be_int(&blkdev->xendev, "sector-size", blkdev->file_blk);
xenstore_write_be_int(&blkdev->xendev, "sectors", xenstore_write_be_int(&blkdev->xendev, "sectors",
@ -678,6 +818,7 @@ out_error:
static int blk_connect(struct XenDevice *xendev) static int blk_connect(struct XenDevice *xendev)
{ {
struct XenBlkDev *blkdev = container_of(xendev, struct XenBlkDev, xendev); struct XenBlkDev *blkdev = container_of(xendev, struct XenBlkDev, xendev);
int pers;
if (xenstore_read_fe_int(&blkdev->xendev, "ring-ref", &blkdev->ring_ref) == -1) { if (xenstore_read_fe_int(&blkdev->xendev, "ring-ref", &blkdev->ring_ref) == -1) {
return -1; return -1;
@ -686,6 +827,11 @@ static int blk_connect(struct XenDevice *xendev)
&blkdev->xendev.remote_port) == -1) { &blkdev->xendev.remote_port) == -1) {
return -1; return -1;
} }
if (xenstore_read_fe_int(&blkdev->xendev, "feature-persistent", &pers)) {
blkdev->feature_persistent = FALSE;
} else {
blkdev->feature_persistent = !!pers;
}
blkdev->protocol = BLKIF_PROTOCOL_NATIVE; blkdev->protocol = BLKIF_PROTOCOL_NATIVE;
if (blkdev->xendev.protocol) { if (blkdev->xendev.protocol) {
@ -729,6 +875,15 @@ static int blk_connect(struct XenDevice *xendev)
} }
} }
if (blkdev->feature_persistent) {
/* Init persistent grants */
blkdev->max_grants = max_requests * BLKIF_MAX_SEGMENTS_PER_REQUEST;
blkdev->persistent_gnts = g_tree_new_full((GCompareDataFunc)int_cmp,
NULL, NULL,
(GDestroyNotify)destroy_grant);
blkdev->persistent_gnt_count = 0;
}
xen_be_bind_evtchn(&blkdev->xendev); xen_be_bind_evtchn(&blkdev->xendev);
xen_be_printf(&blkdev->xendev, 1, "ok: proto %s, ring-ref %d, " xen_be_printf(&blkdev->xendev, 1, "ok: proto %s, ring-ref %d, "
@ -769,6 +924,11 @@ static int blk_free(struct XenDevice *xendev)
blk_disconnect(xendev); blk_disconnect(xendev);
} }
/* Free persistent grants */
if (blkdev->feature_persistent) {
g_tree_destroy(blkdev->persistent_gnts);
}
while (!QLIST_EMPTY(&blkdev->freelist)) { while (!QLIST_EMPTY(&blkdev->freelist)) {
ioreq = QLIST_FIRST(&blkdev->freelist); ioreq = QLIST_FIRST(&blkdev->freelist);
QLIST_REMOVE(ioreq, list); QLIST_REMOVE(ioreq, list);