izzy/xemu
0
0
Fork 0
mirror of https://github.com/izzy2lost/xemu.git synced 2026-07-06 00:20:22 -07:00
Code Issues Packages Projects Releases Wiki Activity

Merge remote branch 'kwolf/for-anthony' into staging

Browse Source
This commit is contained in:
Anthony Liguori
2011-01-24 15:16:56 -06:00
parent 4a2ba23284 a5c062edd2
commit b22b7b729d
25 changed files with 767 additions and 455 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Makefile.objs
+1 -1
View File
@@ -19,7 +19,7 @@ block-obj-$(CONFIG_POSIX) += posix-aio-compat.o
block-obj-$(CONFIG_LINUX_AIO) += linux-aio.o
block-nested-y += raw.o cow.o qcow.o vdi.o vmdk.o cloop.o dmg.o bochs.o vpc.o vvfat.o
block-nested-y += qcow2.o qcow2-refcount.o qcow2-cluster.o qcow2-snapshot.o
block-nested-y += qcow2.o qcow2-refcount.o qcow2-cluster.o qcow2-snapshot.o qcow2-cache.o
block-nested-y += qed.o qed-gencb.o qed-l2-cache.o qed-table.o qed-cluster.o
block-nested-y += qed-check.o
block-nested-y += parallels.o nbd.o blkdebug.o sheepdog.o blkverify.o
block-migration.c
+26 -3
View File
@@ -350,7 +350,12 @@ static int blk_mig_save_bulked_block(Monitor *mon, QEMUFile *f)
}
}
progress = completed_sector_sum * 100 / block_mig_state.total_sector_sum;
if (block_mig_state.total_sector_sum != 0) {
progress = completed_sector_sum * 100 /
block_mig_state.total_sector_sum;
} else {
progress = 100;
}
if (progress != block_mig_state.prev_progress) {
block_mig_state.prev_progress = progress;
qemu_put_be64(f, (progress << BDRV_SECTOR_BITS)
@@ -633,8 +638,10 @@ static int block_load(QEMUFile *f, void *opaque, int version_id)
int len, flags;
char device_name[256];
int64_t addr;
BlockDriverState *bs;
BlockDriverState *bs, *bs_prev = NULL;
uint8_t *buf;
int64_t total_sectors = 0;
int nr_sectors;
do {
addr = qemu_get_be64(f);
@@ -656,10 +663,26 @@ static int block_load(QEMUFile *f, void *opaque, int version_id)
return -EINVAL;
}
if (bs != bs_prev) {
bs_prev = bs;
total_sectors = bdrv_getlength(bs) >> BDRV_SECTOR_BITS;
if (total_sectors <= 0) {
error_report("Error getting length of block device %s\n",
device_name);
return -EINVAL;
}
}
if (total_sectors - addr < BDRV_SECTORS_PER_DIRTY_CHUNK) {
nr_sectors = total_sectors - addr;
} else {
nr_sectors = BDRV_SECTORS_PER_DIRTY_CHUNK;
}
buf = qemu_malloc(BLOCK_SIZE);
qemu_get_buffer(f, buf, BLOCK_SIZE);
ret = bdrv_write(bs, addr, buf, BDRV_SECTORS_PER_DIRTY_CHUNK);
ret = bdrv_write(bs, addr, buf, nr_sectors);
qemu_free(buf);
if (ret < 0) {
block.c
+5 -3
View File
@@ -2778,6 +2778,7 @@ int bdrv_img_create(const char *filename, const char *fmt,
QEMUOptionParameter *backing_fmt, *backing_file;
BlockDriverState *bs = NULL;
BlockDriver *drv, *proto_drv;
BlockDriver *backing_drv = NULL;
int ret = 0;
/* Find driver and parse its options */
@@ -2846,7 +2847,8 @@ int bdrv_img_create(const char *filename, const char *fmt,
backing_fmt = get_option_parameter(param, BLOCK_OPT_BACKING_FMT);
if (backing_fmt && backing_fmt->value.s) {
if (!bdrv_find_format(backing_fmt->value.s)) {
backing_drv = bdrv_find_format(backing_fmt->value.s);
if (!backing_drv) {
error_report("Unknown backing file format '%s'",
backing_fmt->value.s);
ret = -EINVAL;
@@ -2863,9 +2865,9 @@ int bdrv_img_create(const char *filename, const char *fmt,
bs = bdrv_new("");
ret = bdrv_open(bs, backing_file->value.s, flags, drv);
ret = bdrv_open(bs, backing_file->value.s, flags, backing_drv);
if (ret < 0) {
error_report("Could not open '%s'", filename);
error_report("Could not open '%s'", backing_file->value.s);
goto out;
}
bdrv_get_geometry(bs, &size);
block/qcow2-cache.c
+314
View File
@@ -0,0 +1,314 @@
/*
* L2/refcount table cache for the QCOW2 format
*
* Copyright (c) 2010 Kevin Wolf <kwolf@redhat.com>
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "block_int.h"
#include "qemu-common.h"
#include "qcow2.h"
typedef struct Qcow2CachedTable {
void* table;
int64_t offset;
bool dirty;
int cache_hits;
int ref;
} Qcow2CachedTable;
struct Qcow2Cache {
int size;
Qcow2CachedTable* entries;
struct Qcow2Cache* depends;
bool depends_on_flush;
bool writethrough;
};
Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables,
bool writethrough)
{
BDRVQcowState *s = bs->opaque;
Qcow2Cache *c;
int i;
c = qemu_mallocz(sizeof(*c));
c->size = num_tables;
c->entries = qemu_mallocz(sizeof(*c->entries) * num_tables);
c->writethrough = writethrough;
for (i = 0; i < c->size; i++) {
c->entries[i].table = qemu_blockalign(bs, s->cluster_size);
}
return c;
}
int qcow2_cache_destroy(BlockDriverState* bs, Qcow2Cache *c)
{
int i;
for (i = 0; i < c->size; i++) {
assert(c->entries[i].ref == 0);
qemu_vfree(c->entries[i].table);
}
qemu_free(c->entries);
qemu_free(c);
return 0;
}
static int qcow2_cache_flush_dependency(BlockDriverState *bs, Qcow2Cache *c)
{
int ret;
ret = qcow2_cache_flush(bs, c->depends);
if (ret < 0) {
return ret;
}
c->depends = NULL;
c->depends_on_flush = false;
return 0;
}
static int qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i)
{
BDRVQcowState *s = bs->opaque;
int ret = 0;
if (!c->entries[i].dirty || !c->entries[i].offset) {
return 0;
}
if (c->depends) {
ret = qcow2_cache_flush_dependency(bs, c);
} else if (c->depends_on_flush) {
ret = bdrv_flush(bs->file);
if (ret >= 0) {
c->depends_on_flush = false;
}
}
if (ret < 0) {
return ret;
}
if (c == s->refcount_block_cache) {
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
} else if (c == s->l2_table_cache) {
BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
}
ret = bdrv_pwrite(bs->file, c->entries[i].offset, c->entries[i].table,
s->cluster_size);
if (ret < 0) {
return ret;
}
c->entries[i].dirty = false;
return 0;
}
int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c)
{
int result = 0;
int ret;
int i;
for (i = 0; i < c->size; i++) {
ret = qcow2_cache_entry_flush(bs, c, i);
if (ret < 0 && result != -ENOSPC) {
result = ret;
}
}
if (result == 0) {
ret = bdrv_flush(bs->file);
if (ret < 0) {
result = ret;
}
}
return result;
}
int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c,
Qcow2Cache *dependency)
{
int ret;
if (dependency->depends) {
ret = qcow2_cache_flush_dependency(bs, dependency);
if (ret < 0) {
return ret;
}
}
if (c->depends && (c->depends != dependency)) {
ret = qcow2_cache_flush_dependency(bs, c);
if (ret < 0) {
return ret;
}
}
c->depends = dependency;
return 0;
}
void qcow2_cache_depends_on_flush(Qcow2Cache *c)
{
c->depends_on_flush = true;
}
static int qcow2_cache_find_entry_to_replace(Qcow2Cache *c)
{
int i;
int min_count = INT_MAX;
int min_index = -1;
for (i = 0; i < c->size; i++) {
if (c->entries[i].ref) {
continue;
}
if (c->entries[i].cache_hits < min_count) {
min_index = i;
min_count = c->entries[i].cache_hits;
}
/* Give newer hits priority */
/* TODO Check how to optimize the replacement strategy */
c->entries[i].cache_hits /= 2;
}
if (min_index == -1) {
/* This can't happen in current synchronous code, but leave the check
* here as a reminder for whoever starts using AIO with the cache */
abort();
}
return min_index;
}
static int qcow2_cache_do_get(BlockDriverState *bs, Qcow2Cache *c,
uint64_t offset, void **table, bool read_from_disk)
{
BDRVQcowState *s = bs->opaque;
int i;
int ret;
/* Check if the table is already cached */
for (i = 0; i < c->size; i++) {
if (c->entries[i].offset == offset) {
goto found;
}
}
/* If not, write a table back and replace it */
i = qcow2_cache_find_entry_to_replace(c);
if (i < 0) {
return i;
}
ret = qcow2_cache_entry_flush(bs, c, i);
if (ret < 0) {
return ret;
}
c->entries[i].offset = 0;
if (read_from_disk) {
if (c == s->l2_table_cache) {
BLKDBG_EVENT(bs->file, BLKDBG_L2_LOAD);
}
ret = bdrv_pread(bs->file, offset, c->entries[i].table, s->cluster_size);
if (ret < 0) {
return ret;
}
}
/* Give the table some hits for the start so that it won't be replaced
* immediately. The number 32 is completely arbitrary. */
c->entries[i].cache_hits = 32;
c->entries[i].offset = offset;
/* And return the right table */
found:
c->entries[i].cache_hits++;
c->entries[i].ref++;
*table = c->entries[i].table;
return 0;
}
int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
void **table)
{
return qcow2_cache_do_get(bs, c, offset, table, true);
}
int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
void **table)
{
return qcow2_cache_do_get(bs, c, offset, table, false);
}
int qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, void **table)
{
int i;
for (i = 0; i < c->size; i++) {
if (c->entries[i].table == *table) {
goto found;
}
}
return -ENOENT;
found:
c->entries[i].ref--;
*table = NULL;
assert(c->entries[i].ref >= 0);
if (c->writethrough) {
return qcow2_cache_entry_flush(bs, c, i);
} else {
return 0;
}
}
void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table)
{
int i;
for (i = 0; i < c->size; i++) {
if (c->entries[i].table == table) {
goto found;
}
}
abort();
found:
c->entries[i].dirty = true;
}
block/qcow2-cluster.c
+73 -137
View File
@@ -67,7 +67,11 @@ int qcow2_grow_l1_table(BlockDriverState *bs, int min_size, bool exact_size)
qemu_free(new_l1_table);
return new_l1_table_offset;
}
bdrv_flush(bs->file);
ret = qcow2_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
return ret;
}
BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_WRITE_TABLE);
for(i = 0; i < s->l1_size; i++)
@@ -81,7 +85,7 @@ int qcow2_grow_l1_table(BlockDriverState *bs, int min_size, bool exact_size)
/* set new table */
BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_ACTIVATE_TABLE);
cpu_to_be32w((uint32_t*)data, new_l1_size);
cpu_to_be64w((uint64_t*)(data + 4), new_l1_table_offset);
cpu_to_be64wu((uint64_t*)(data + 4), new_l1_table_offset);
ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, l1_size), data,sizeof(data));
if (ret < 0) {
goto fail;
@@ -98,63 +102,6 @@ int qcow2_grow_l1_table(BlockDriverState *bs, int min_size, bool exact_size)
return ret;
}
void qcow2_l2_cache_reset(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
memset(s->l2_cache, 0, s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
memset(s->l2_cache_offsets, 0, L2_CACHE_SIZE * sizeof(uint64_t));
memset(s->l2_cache_counts, 0, L2_CACHE_SIZE * sizeof(uint32_t));
}
static inline int l2_cache_new_entry(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
uint32_t min_count;
int min_index, i;
/* find a new entry in the least used one */
min_index = 0;
min_count = 0xffffffff;
for(i = 0; i < L2_CACHE_SIZE; i++) {
if (s->l2_cache_counts[i] < min_count) {
min_count = s->l2_cache_counts[i];
min_index = i;
}
}
return min_index;
}
/*
* seek_l2_table
*
* seek l2_offset in the l2_cache table
* if not found, return NULL,
* if found,
* increments the l2 cache hit count of the entry,
* if counter overflow, divide by two all counters
* return the pointer to the l2 cache entry
*
*/
static uint64_t *seek_l2_table(BDRVQcowState *s, uint64_t l2_offset)
{
int i, j;
for(i = 0; i < L2_CACHE_SIZE; i++) {
if (l2_offset == s->l2_cache_offsets[i]) {
/* increment the hit count */
if (++s->l2_cache_counts[i] == 0xffffffff) {
for(j = 0; j < L2_CACHE_SIZE; j++) {
s->l2_cache_counts[j] >>= 1;
}
}
return s->l2_cache + (i << s->l2_bits);
}
}
return NULL;
}
/*
* l2_load
*
@@ -169,33 +116,11 @@ static int l2_load(BlockDriverState *bs, uint64_t l2_offset,
uint64_t **l2_table)
{
BDRVQcowState *s = bs->opaque;
int min_index;
int ret;
/* seek if the table for the given offset is in the cache */
ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset, (void**) l2_table);
*l2_table = seek_l2_table(s, l2_offset);
if (*l2_table != NULL) {
return 0;
}
/* not found: load a new entry in the least used one */
min_index = l2_cache_new_entry(bs);
*l2_table = s->l2_cache + (min_index << s->l2_bits);
BLKDBG_EVENT(bs->file, BLKDBG_L2_LOAD);
ret = bdrv_pread(bs->file, l2_offset, *l2_table,
s->l2_size * sizeof(uint64_t));
if (ret < 0) {
qcow2_l2_cache_reset(bs);
return ret;
}
s->l2_cache_offsets[min_index] = l2_offset;
s->l2_cache_counts[min_index] = 1;
return 0;
return ret;
}
/*
@@ -238,7 +163,6 @@ static int write_l1_entry(BlockDriverState *bs, int l1_index)
static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
{
BDRVQcowState *s = bs->opaque;
int min_index;
uint64_t old_l2_offset;
uint64_t *l2_table;
int64_t l2_offset;
@@ -252,29 +176,48 @@ static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
if (l2_offset < 0) {
return l2_offset;
}
bdrv_flush(bs->file);
ret = qcow2_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
goto fail;
}
/* allocate a new entry in the l2 cache */
min_index = l2_cache_new_entry(bs);
l2_table = s->l2_cache + (min_index << s->l2_bits);
ret = qcow2_cache_get_empty(bs, s->l2_table_cache, l2_offset, (void**) table);
if (ret < 0) {
return ret;
}
l2_table = *table;
if (old_l2_offset == 0) {
/* if there was no old l2 table, clear the new table */
memset(l2_table, 0, s->l2_size * sizeof(uint64_t));
} else {
uint64_t* old_table;
/* if there was an old l2 table, read it from the disk */
BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_COW_READ);
ret = bdrv_pread(bs->file, old_l2_offset, l2_table,
s->l2_size * sizeof(uint64_t));
ret = qcow2_cache_get(bs, s->l2_table_cache, old_l2_offset,
(void**) &old_table);
if (ret < 0) {
goto fail;
}
memcpy(l2_table, old_table, s->cluster_size);
ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &old_table);
if (ret < 0) {
goto fail;
}
}
/* write the l2 table to the file */
BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_WRITE);
ret = bdrv_pwrite_sync(bs->file, l2_offset, l2_table,
s->l2_size * sizeof(uint64_t));
qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
ret = qcow2_cache_flush(bs, s->l2_table_cache);
if (ret < 0) {
goto fail;
}
@@ -286,17 +229,12 @@ static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
goto fail;
}
/* update the l2 cache entry */
s->l2_cache_offsets[min_index] = l2_offset;
s->l2_cache_counts[min_index] = 1;
*table = l2_table;
return 0;
fail:
qcow2_cache_put(bs, s->l2_table_cache, (void**) table);
s->l1_table[l1_index] = old_l2_offset;
qcow2_l2_cache_reset(bs);
return ret;
}
@@ -521,6 +459,8 @@ int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
&l2_table[l2_index], 0, QCOW_OFLAG_COPIED);
}
qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
nb_available = (c * s->cluster_sectors);
out:
if (nb_available > nb_needed)
@@ -575,6 +515,7 @@ static int get_cluster_table(BlockDriverState *bs, uint64_t offset,
return ret;
}
} else {
/* FIXME Order */
if (l2_offset)
qcow2_free_clusters(bs, l2_offset, s->l2_size * sizeof(uint64_t));
ret = l2_allocate(bs, l1_index, &l2_table);
@@ -632,6 +573,7 @@ uint64_t qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
cluster_offset = qcow2_alloc_bytes(bs, compressed_size);
if (cluster_offset < 0) {
qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
return 0;
}
@@ -646,38 +588,14 @@ uint64_t qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
/* compressed clusters never have the copied flag */
BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE_COMPRESSED);
qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
l2_table[l2_index] = cpu_to_be64(cluster_offset);
if (bdrv_pwrite_sync(bs->file,
l2_offset + l2_index * sizeof(uint64_t),
l2_table + l2_index,
sizeof(uint64_t)) < 0)
return 0;
return cluster_offset;
}
/*
* Write L2 table updates to disk, writing whole sectors to avoid a
* read-modify-write in bdrv_pwrite
*/
#define L2_ENTRIES_PER_SECTOR (512 / 8)
static int write_l2_entries(BlockDriverState *bs, uint64_t *l2_table,
uint64_t l2_offset, int l2_index, int num)
{
int l2_start_index = l2_index & ~(L1_ENTRIES_PER_SECTOR - 1);
int start_offset = (8 * l2_index) & ~511;
int end_offset = (8 * (l2_index + num) + 511) & ~511;
size_t len = end_offset - start_offset;
int ret;
BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
ret = bdrv_pwrite(bs->file, l2_offset + start_offset,
&l2_table[l2_start_index], len);
ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
if (ret < 0) {
return ret;
return 0;
}
return 0;
return cluster_offset;
}
int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
@@ -686,6 +604,7 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
int i, j = 0, l2_index, ret;
uint64_t *old_cluster, start_sect, l2_offset, *l2_table;
uint64_t cluster_offset = m->cluster_offset;
bool cow = false;
if (m->nb_clusters == 0)
return 0;
@@ -695,6 +614,7 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
/* copy content of unmodified sectors */
start_sect = (m->offset & ~(s->cluster_size - 1)) >> 9;
if (m->n_start) {
cow = true;
ret = copy_sectors(bs, start_sect, cluster_offset, 0, m->n_start);
if (ret < 0)
goto err;
@@ -702,17 +622,30 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
if (m->nb_available & (s->cluster_sectors - 1)) {
uint64_t end = m->nb_available & ~(uint64_t)(s->cluster_sectors - 1);
cow = true;
ret = copy_sectors(bs, start_sect + end, cluster_offset + (end << 9),
m->nb_available - end, s->cluster_sectors);
if (ret < 0)
goto err;
}
/* update L2 table */
/*
* Update L2 table.
*
* Before we update the L2 table to actually point to the new cluster, we
* need to be sure that the refcounts have been increased and COW was
* handled.
*/
if (cow) {
qcow2_cache_depends_on_flush(s->l2_table_cache);
}
qcow2_cache_set_dependency(bs, s->l2_table_cache, s->refcount_block_cache);
ret = get_cluster_table(bs, m->offset, &l2_table, &l2_offset, &l2_index);
if (ret < 0) {
goto err;
}
qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
for (i = 0; i < m->nb_clusters; i++) {
/* if two concurrent writes happen to the same unallocated cluster
@@ -728,16 +661,9 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
(i << s->cluster_bits)) | QCOW_OFLAG_COPIED);
}
/*
* Before we update the L2 table to actually point to the new cluster, we
* need to be sure that the refcounts have been increased and COW was
* handled.
*/
bdrv_flush(bs->file);
ret = write_l2_entries(bs, l2_table, l2_offset, l2_index, m->nb_clusters);
ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
if (ret < 0) {
qcow2_l2_cache_reset(bs);
goto err;
}
@@ -746,7 +672,6 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
* Also flush bs->file to get the right order for L2 and refcount update.
*/
if (j != 0) {
bdrv_flush(bs->file);
for (i = 0; i < j; i++) {
qcow2_free_any_clusters(bs,
be64_to_cpu(old_cluster[i]) & ~QCOW_OFLAG_COPIED, 1);
@@ -868,7 +793,8 @@ int qcow2_alloc_cluster_offset(BlockDriverState *bs, uint64_t offset,
m->depends_on = old_alloc;
m->nb_clusters = 0;
*num = 0;
return 0;
ret = 0;
goto fail;
}
}
}
@@ -884,7 +810,8 @@ int qcow2_alloc_cluster_offset(BlockDriverState *bs, uint64_t offset,
cluster_offset = qcow2_alloc_clusters(bs, nb_clusters * s->cluster_size);
if (cluster_offset < 0) {
QLIST_REMOVE(m, next_in_flight);
return cluster_offset;
ret = cluster_offset;
goto fail;
}
/* save info needed for meta data update */
@@ -893,12 +820,21 @@ int qcow2_alloc_cluster_offset(BlockDriverState *bs, uint64_t offset,
m->nb_clusters = nb_clusters;
out:
ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
if (ret < 0) {
return ret;
}
m->nb_available = MIN(nb_clusters << (s->cluster_bits - 9), n_end);
m->cluster_offset = cluster_offset;
*num = m->nb_available - n_start;
return 0;
fail:
qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
return ret;
}
static int decompress_buffer(uint8_t *out_buf, int out_buf_size,
block/qcow2-refcount.c
+110 -156
View File
@@ -32,27 +32,6 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
int addend);
static int cache_refcount_updates = 0;
static int write_refcount_block(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
size_t size = s->cluster_size;
if (s->refcount_block_cache_offset == 0) {
return 0;
}
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE);
if (bdrv_pwrite_sync(bs->file, s->refcount_block_cache_offset,
s->refcount_block_cache, size) < 0)
{
return -EIO;
}
return 0;
}
/*********************************************************/
/* refcount handling */
@@ -61,7 +40,6 @@ int qcow2_refcount_init(BlockDriverState *bs)
BDRVQcowState *s = bs->opaque;
int ret, refcount_table_size2, i;
s->refcount_block_cache = qemu_malloc(s->cluster_size);
refcount_table_size2 = s->refcount_table_size * sizeof(uint64_t);
s->refcount_table = qemu_malloc(refcount_table_size2);
if (s->refcount_table_size > 0) {
@@ -81,34 +59,22 @@ int qcow2_refcount_init(BlockDriverState *bs)
void qcow2_refcount_close(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
qemu_free(s->refcount_block_cache);
qemu_free(s->refcount_table);
}
static int load_refcount_block(BlockDriverState *bs,
int64_t refcount_block_offset)
int64_t refcount_block_offset,
void **refcount_block)
{
BDRVQcowState *s = bs->opaque;
int ret;
if (cache_refcount_updates) {
ret = write_refcount_block(bs);
if (ret < 0) {
return ret;
}
}
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_LOAD);
ret = bdrv_pread(bs->file, refcount_block_offset, s->refcount_block_cache,
s->cluster_size);
if (ret < 0) {
s->refcount_block_cache_offset = 0;
return ret;
}
ret = qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
refcount_block);
s->refcount_block_cache_offset = refcount_block_offset;
return 0;
return ret;
}
/*
@@ -122,6 +88,8 @@ static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
int refcount_table_index, block_index;
int64_t refcount_block_offset;
int ret;
uint16_t *refcount_block;
uint16_t refcount;
refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
if (refcount_table_index >= s->refcount_table_size)
@@ -129,16 +97,24 @@ static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
refcount_block_offset = s->refcount_table[refcount_table_index];
if (!refcount_block_offset)
return 0;
if (refcount_block_offset != s->refcount_block_cache_offset) {
/* better than nothing: return allocated if read error */
ret = load_refcount_block(bs, refcount_block_offset);
if (ret < 0) {
return ret;
}
ret = qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
(void**) &refcount_block);
if (ret < 0) {
return ret;
}
block_index = cluster_index &
((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
return be16_to_cpu(s->refcount_block_cache[block_index]);
refcount = be16_to_cpu(refcount_block[block_index]);
ret = qcow2_cache_put(bs, s->refcount_block_cache,
(void**) &refcount_block);
if (ret < 0) {
return ret;
}
return refcount;
}
/*
@@ -174,9 +150,10 @@ static int in_same_refcount_block(BDRVQcowState *s, uint64_t offset_a,
* Loads a refcount block. If it doesn't exist yet, it is allocated first
* (including growing the refcount table if needed).
*
* Returns the offset of the refcount block on success or -errno in error case
* Returns 0 on success or -errno in error case
*/
static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
static int alloc_refcount_block(BlockDriverState *bs,
int64_t cluster_index, uint16_t **refcount_block)
{
BDRVQcowState *s = bs->opaque;
unsigned int refcount_table_index;
@@ -194,13 +171,8 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
/* If it's already there, we're done */
if (refcount_block_offset) {
if (refcount_block_offset != s->refcount_block_cache_offset) {
ret = load_refcount_block(bs, refcount_block_offset);
if (ret < 0) {
return ret;
}
}
return refcount_block_offset;
return load_refcount_block(bs, refcount_block_offset,
(void**) refcount_block);
}
}
@@ -226,12 +198,10 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
* refcount block into the cache
*/
if (cache_refcount_updates) {
ret = write_refcount_block(bs);
if (ret < 0) {
return ret;
}
}
*refcount_block = NULL;
/* We write to the refcount table, so we might depend on L2 tables */
qcow2_cache_flush(bs, s->l2_table_cache);
/* Allocate the refcount block itself and mark it as used */
int64_t new_block = alloc_clusters_noref(bs, s->cluster_size);
@@ -247,13 +217,18 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
if (in_same_refcount_block(s, new_block, cluster_index << s->cluster_bits)) {
/* Zero the new refcount block before updating it */
memset(s->refcount_block_cache, 0, s->cluster_size);
s->refcount_block_cache_offset = new_block;
ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
(void**) refcount_block);
if (ret < 0) {
goto fail_block;
}
memset(*refcount_block, 0, s->cluster_size);
/* The block describes itself, need to update the cache */
int block_index = (new_block >> s->cluster_bits) &
((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
s->refcount_block_cache[block_index] = cpu_to_be16(1);
(*refcount_block)[block_index] = cpu_to_be16(1);
} else {
/* Described somewhere else. This can recurse at most twice before we
* arrive at a block that describes itself. */
@@ -266,14 +241,19 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
/* Initialize the new refcount block only after updating its refcount,
* update_refcount uses the refcount cache itself */
memset(s->refcount_block_cache, 0, s->cluster_size);
s->refcount_block_cache_offset = new_block;
ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
(void**) refcount_block);
if (ret < 0) {
goto fail_block;
}
memset(*refcount_block, 0, s->cluster_size);
}
/* Now the new refcount block needs to be written to disk */
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE);
ret = bdrv_pwrite_sync(bs->file, new_block, s->refcount_block_cache,
s->cluster_size);
qcow2_cache_entry_mark_dirty(s->refcount_block_cache, *refcount_block);
ret = qcow2_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
goto fail_block;
}
@@ -290,7 +270,12 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
}
s->refcount_table[refcount_table_index] = new_block;
return new_block;
return 0;
}
ret = qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block);
if (ret < 0) {
goto fail_block;
}
/*
@@ -410,9 +395,9 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
qcow2_free_clusters(bs, old_table_offset, old_table_size * sizeof(uint64_t));
s->free_cluster_index = old_free_cluster_index;
ret = load_refcount_block(bs, new_block);
ret = load_refcount_block(bs, new_block, (void**) refcount_block);
if (ret < 0) {
goto fail_block;
return ret;
}
return new_block;
@@ -420,52 +405,20 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
fail_table:
qemu_free(new_table);
fail_block:
s->refcount_block_cache_offset = 0;
if (*refcount_block != NULL) {
qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block);
}
return ret;
}
#define REFCOUNTS_PER_SECTOR (512 >> REFCOUNT_SHIFT)
static int write_refcount_block_entries(BlockDriverState *bs,
int64_t refcount_block_offset, int first_index, int last_index)
{
BDRVQcowState *s = bs->opaque;
size_t size;
int ret;
if (cache_refcount_updates) {
return 0;
}
if (first_index < 0) {
return 0;
}
first_index &= ~(REFCOUNTS_PER_SECTOR - 1);
last_index = (last_index + REFCOUNTS_PER_SECTOR)
& ~(REFCOUNTS_PER_SECTOR - 1);
size = (last_index - first_index) << REFCOUNT_SHIFT;
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
ret = bdrv_pwrite(bs->file,
refcount_block_offset + (first_index << REFCOUNT_SHIFT),
&s->refcount_block_cache[first_index], size);
if (ret < 0) {
return ret;
}
return 0;
}
/* XXX: cache several refcount block clusters ? */
static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
int64_t offset, int64_t length, int addend)
{
BDRVQcowState *s = bs->opaque;
int64_t start, last, cluster_offset;
int64_t refcount_block_offset = 0;
int64_t table_index = -1, old_table_index;
int first_index = -1, last_index = -1;
uint16_t *refcount_block = NULL;
int64_t old_table_index = -1;
int ret;
#ifdef DEBUG_ALLOC2
@@ -478,6 +431,11 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
return 0;
}
if (addend < 0) {
qcow2_cache_set_dependency(bs, s->refcount_block_cache,
s->l2_table_cache);
}
start = offset & ~(s->cluster_size - 1);
last = (offset + length - 1) & ~(s->cluster_size - 1);
for(cluster_offset = start; cluster_offset <= last;
@@ -485,42 +443,33 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
{
int block_index, refcount;
int64_t cluster_index = cluster_offset >> s->cluster_bits;
int64_t new_block;
/* Only write refcount block to disk when we are done with it */
old_table_index = table_index;
table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
if ((old_table_index >= 0) && (table_index != old_table_index)) {
ret = write_refcount_block_entries(bs, refcount_block_offset,
first_index, last_index);
if (ret < 0) {
return ret;
}
first_index = -1;
last_index = -1;
}
int64_t table_index =
cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
/* Load the refcount block and allocate it if needed */
new_block = alloc_refcount_block(bs, cluster_index);
if (new_block < 0) {
ret = new_block;
goto fail;
if (table_index != old_table_index) {
if (refcount_block) {
ret = qcow2_cache_put(bs, s->refcount_block_cache,
(void**) &refcount_block);
if (ret < 0) {
goto fail;
}
}
ret = alloc_refcount_block(bs, cluster_index, &refcount_block);
if (ret < 0) {
goto fail;
}
}
refcount_block_offset = new_block;
old_table_index = table_index;
qcow2_cache_entry_mark_dirty(s->refcount_block_cache, refcount_block);
/* we can update the count and save it */
block_index = cluster_index &
((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
if (first_index == -1 || block_index < first_index) {
first_index = block_index;
}
if (block_index > last_index) {
last_index = block_index;
}
refcount = be16_to_cpu(s->refcount_block_cache[block_index]);
refcount = be16_to_cpu(refcount_block[block_index]);
refcount += addend;
if (refcount < 0 || refcount > 0xffff) {
ret = -EINVAL;
@@ -529,17 +478,16 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
if (refcount == 0 && cluster_index < s->free_cluster_index) {
s->free_cluster_index = cluster_index;
}
s->refcount_block_cache[block_index] = cpu_to_be16(refcount);
refcount_block[block_index] = cpu_to_be16(refcount);
}
ret = 0;
fail:
/* Write last changed block to disk */
if (refcount_block_offset != 0) {
if (refcount_block) {
int wret;
wret = write_refcount_block_entries(bs, refcount_block_offset,
first_index, last_index);
wret = qcow2_cache_put(bs, s->refcount_block_cache,
(void**) &refcount_block);
if (wret < 0) {
return ret < 0 ? ret : wret;
}
@@ -758,9 +706,7 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
uint64_t *l1_table, *l2_table, l2_offset, offset, l1_size2, l1_allocated;
int64_t old_offset, old_l2_offset;
int l2_size, i, j, l1_modified, l2_modified, nb_csectors, refcount;
qcow2_l2_cache_reset(bs);
cache_refcount_updates = 1;
int ret;
l2_table = NULL;
l1_table = NULL;
@@ -784,7 +730,6 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
}
l2_size = s->l2_size * sizeof(uint64_t);
l2_table = qemu_malloc(l2_size);
l1_modified = 0;
for(i = 0; i < l1_size; i++) {
l2_offset = l1_table[i];
@@ -792,8 +737,13 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
old_l2_offset = l2_offset;
l2_offset &= ~QCOW_OFLAG_COPIED;
l2_modified = 0;
if (bdrv_pread(bs->file, l2_offset, l2_table, l2_size) != l2_size)
ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset,
(void**) &l2_table);
if (ret < 0) {
goto fail;
}
for(j = 0; j < s->l2_size; j++) {
offset = be64_to_cpu(l2_table[j]);
if (offset != 0) {
@@ -833,17 +783,23 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
offset |= QCOW_OFLAG_COPIED;
}
if (offset != old_offset) {
if (addend > 0) {
qcow2_cache_set_dependency(bs, s->l2_table_cache,
s->refcount_block_cache);
}
l2_table[j] = cpu_to_be64(offset);
l2_modified = 1;
qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
}
}
}
if (l2_modified) {
if (bdrv_pwrite_sync(bs->file,
l2_offset, l2_table, l2_size) < 0)
goto fail;
ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
if (ret < 0) {
goto fail;
}
if (addend != 0) {
refcount = update_cluster_refcount(bs, l2_offset >> s->cluster_bits, addend);
} else {
@@ -871,16 +827,14 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
}
if (l1_allocated)
qemu_free(l1_table);
qemu_free(l2_table);
cache_refcount_updates = 0;
write_refcount_block(bs);
return 0;
fail:
if (l2_table) {
qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
}
if (l1_allocated)
qemu_free(l1_table);
qemu_free(l2_table);
cache_refcount_updates = 0;
write_refcount_block(bs);
return -EIO;
}
block/qcow2.c
+44 -4
View File
@@ -143,6 +143,7 @@ static int qcow2_open(BlockDriverState *bs, int flags)
int len, i, ret = 0;
QCowHeader header;
uint64_t ext_end;
bool writethrough;
ret = bdrv_pread(bs->file, 0, &header, sizeof(header));
if (ret < 0) {
@@ -217,8 +218,13 @@ static int qcow2_open(BlockDriverState *bs, int flags)
be64_to_cpus(&s->l1_table[i]);
}
}
/* alloc L2 cache */
s->l2_cache = qemu_malloc(s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
/* alloc L2 table/refcount block cache */
writethrough = ((flags & BDRV_O_CACHE_MASK) == 0);
s->l2_table_cache = qcow2_cache_create(bs, L2_CACHE_SIZE, writethrough);
s->refcount_block_cache = qcow2_cache_create(bs, REFCOUNT_CACHE_SIZE,
writethrough);
s->cluster_cache = qemu_malloc(s->cluster_size);
/* one more sector for decompressed data alignment */
s->cluster_data = qemu_malloc(QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size
@@ -270,7 +276,9 @@ static int qcow2_open(BlockDriverState *bs, int flags)
qcow2_free_snapshots(bs);
qcow2_refcount_close(bs);
qemu_free(s->l1_table);
qemu_free(s->l2_cache);
if (s->l2_table_cache) {
qcow2_cache_destroy(bs, s->l2_table_cache);
}
qemu_free(s->cluster_cache);
qemu_free(s->cluster_data);
return ret;
@@ -719,7 +727,13 @@ static void qcow2_close(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
qemu_free(s->l1_table);
qemu_free(s->l2_cache);
qcow2_cache_flush(bs, s->l2_table_cache);
qcow2_cache_flush(bs, s->refcount_block_cache);
qcow2_cache_destroy(bs, s->l2_table_cache);
qcow2_cache_destroy(bs, s->refcount_block_cache);
qemu_free(s->cluster_cache);
qemu_free(s->cluster_data);
qcow2_refcount_close(bs);
@@ -1179,6 +1193,19 @@ static int qcow2_write_compressed(BlockDriverState *bs, int64_t sector_num,
static int qcow2_flush(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
int ret;
ret = qcow2_cache_flush(bs, s->l2_table_cache);
if (ret < 0) {
return ret;
}
ret = qcow2_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
return ret;
}
return bdrv_flush(bs->file);
}
@@ -1186,6 +1213,19 @@ static BlockDriverAIOCB *qcow2_aio_flush(BlockDriverState *bs,
BlockDriverCompletionFunc *cb,
void *opaque)
{
BDRVQcowState *s = bs->opaque;
int ret;
ret = qcow2_cache_flush(bs, s->l2_table_cache);
if (ret < 0) {
return NULL;
}
ret = qcow2_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
return NULL;
}
return bdrv_aio_flush(bs->file, cb, opaque);
}
block/qcow2.h
+27 -5
View File
@@ -51,6 +51,9 @@
#define L2_CACHE_SIZE 16
/* Must be at least 4 to cover all cases of refcount table growth */
#define REFCOUNT_CACHE_SIZE 4
typedef struct QCowHeader {
uint32_t magic;
uint32_t version;
@@ -78,6 +81,9 @@ typedef struct QCowSnapshot {
uint64_t vm_clock_nsec;
} QCowSnapshot;
struct Qcow2Cache;
typedef struct Qcow2Cache Qcow2Cache;
typedef struct BDRVQcowState {
int cluster_bits;
int cluster_size;
@@ -91,9 +97,10 @@ typedef struct BDRVQcowState {
uint64_t cluster_offset_mask;
uint64_t l1_table_offset;
uint64_t *l1_table;
uint64_t *l2_cache;
uint64_t l2_cache_offsets[L2_CACHE_SIZE];
uint32_t l2_cache_counts[L2_CACHE_SIZE];
Qcow2Cache* l2_table_cache;
Qcow2Cache* refcount_block_cache;
uint8_t *cluster_cache;
uint8_t *cluster_data;
uint64_t cluster_cache_offset;
@@ -102,8 +109,6 @@ typedef struct BDRVQcowState {
uint64_t *refcount_table;
uint64_t refcount_table_offset;
uint32_t refcount_table_size;
uint64_t refcount_block_cache_offset;
uint16_t *refcount_block_cache;
int64_t free_cluster_index;
int64_t free_byte_offset;
@@ -215,4 +220,21 @@ int qcow2_snapshot_load_tmp(BlockDriverState *bs, const char *snapshot_name);
void qcow2_free_snapshots(BlockDriverState *bs);
int qcow2_read_snapshots(BlockDriverState *bs);
/* qcow2-cache.c functions */
Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables,
bool writethrough);
int qcow2_cache_destroy(BlockDriverState* bs, Qcow2Cache *c);
void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table);
int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c);
int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c,
Qcow2Cache *dependency);
void qcow2_cache_depends_on_flush(Qcow2Cache *c);
int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
void **table);
int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
void **table);
int qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, void **table);
#endif
block/qed.c
+6
View File
@@ -469,6 +469,12 @@ static int qed_create(const char *filename, uint32_t cluster_size,
return ret;
}
/* File must start empty and grow, check truncate is supported */
ret = bdrv_truncate(bs, 0);
if (ret < 0) {
goto out;
}
if (backing_file) {
header.features |= QED_F_BACKING_FILE;
header.backing_filename_offset = sizeof(le_header);
blockdev.c
+43 -38
View File
@@ -107,7 +107,7 @@ DriveInfo *drive_get_by_blockdev(BlockDriverState *bs)
static void bdrv_format_print(void *opaque, const char *name)
{
fprintf(stderr, " %s", name);
error_printf(" %s", name);
}
void drive_uninit(DriveInfo *dinfo)
@@ -129,8 +129,8 @@ static int parse_block_error_action(const char *buf, int is_read)
} else if (!strcmp(buf, "report")) {
return BLOCK_ERR_REPORT;
} else {
fprintf(stderr, "qemu: '%s' invalid %s error action\n",
buf, is_read ? "read" : "write");
error_report("'%s' invalid %s error action",
buf, is_read ? "read" : "write");
return -1;
}
}
@@ -217,31 +217,30 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
type = IF_NONE;
max_devs = 0;
} else {
fprintf(stderr, "qemu: unsupported bus type '%s'\n", buf);
error_report("unsupported bus type '%s'", buf);
return NULL;
}
}
if (cyls || heads || secs) {
if (cyls < 1 || (type == IF_IDE && cyls > 16383)) {
fprintf(stderr, "qemu: '%s' invalid physical cyls number\n", buf);
error_report("invalid physical cyls number");
return NULL;
}
if (heads < 1 || (type == IF_IDE && heads > 16)) {
fprintf(stderr, "qemu: '%s' invalid physical heads number\n", buf);
error_report("invalid physical heads number");
return NULL;
}
if (secs < 1 || (type == IF_IDE && secs > 63)) {
fprintf(stderr, "qemu: '%s' invalid physical secs number\n", buf);
error_report("invalid physical secs number");
return NULL;
}
}
if ((buf = qemu_opt_get(opts, "trans")) != NULL) {
if (!cyls) {
fprintf(stderr,
"qemu: '%s' trans must be used with cyls,heads and secs\n",
buf);
error_report("'%s' trans must be used with cyls,heads and secs",
buf);
return NULL;
}
if (!strcmp(buf, "none"))
@@ -251,7 +250,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
else if (!strcmp(buf, "auto"))
translation = BIOS_ATA_TRANSLATION_AUTO;
else {
fprintf(stderr, "qemu: '%s' invalid translation type\n", buf);
error_report("'%s' invalid translation type", buf);
return NULL;
}
}
@@ -261,13 +260,12 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
media = MEDIA_DISK;
} else if (!strcmp(buf, "cdrom")) {
if (cyls || secs || heads) {
fprintf(stderr,
"qemu: '%s' invalid physical CHS format\n", buf);
error_report("'%s' invalid physical CHS format", buf);
return NULL;
}
media = MEDIA_CDROM;
} else {
fprintf(stderr, "qemu: '%s' invalid media\n", buf);
error_report("'%s' invalid media", buf);
return NULL;
}
}
@@ -283,7 +281,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
} else if (!strcmp(buf, "writethrough")) {
/* this is the default */
} else {
fprintf(stderr, "qemu: invalid cache option\n");
error_report("invalid cache option");
return NULL;
}
}
@@ -295,7 +293,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
} else if (!strcmp(buf, "threads")) {
/* this is the default */
} else {
fprintf(stderr, "qemu: invalid aio option\n");
error_report("invalid aio option");
return NULL;
}
}
@@ -303,14 +301,14 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
if ((buf = qemu_opt_get(opts, "format")) != NULL) {
if (strcmp(buf, "?") == 0) {
fprintf(stderr, "qemu: Supported formats:");
bdrv_iterate_format(bdrv_format_print, NULL);
fprintf(stderr, "\n");
return NULL;
error_printf("Supported formats:");
bdrv_iterate_format(bdrv_format_print, NULL);
error_printf("\n");
return NULL;
}
drv = bdrv_find_whitelisted_format(buf);
if (!drv) {
fprintf(stderr, "qemu: '%s' invalid format\n", buf);
error_report("'%s' invalid format", buf);
return NULL;
}
}
@@ -318,7 +316,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
on_write_error = BLOCK_ERR_STOP_ENOSPC;
if ((buf = qemu_opt_get(opts, "werror")) != NULL) {
if (type != IF_IDE && type != IF_SCSI && type != IF_VIRTIO && type != IF_NONE) {
fprintf(stderr, "werror is not supported by this format\n");
error_report("werror is not supported by this bus type");
return NULL;
}
@@ -331,7 +329,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
on_read_error = BLOCK_ERR_REPORT;
if ((buf = qemu_opt_get(opts, "rerror")) != NULL) {
if (type != IF_IDE && type != IF_VIRTIO && type != IF_SCSI && type != IF_NONE) {
fprintf(stderr, "rerror is not supported by this format\n");
error_report("rerror is not supported by this bus type");
return NULL;
}
@@ -343,7 +341,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
if ((devaddr = qemu_opt_get(opts, "addr")) != NULL) {
if (type != IF_VIRTIO) {
fprintf(stderr, "addr is not supported\n");
error_report("addr is not supported by this bus type");
return NULL;
}
}
@@ -352,8 +350,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
if (index != -1) {
if (bus_id != 0 || unit_id != -1) {
fprintf(stderr,
"qemu: index cannot be used with bus and unit\n");
error_report("index cannot be used with bus and unit");
return NULL;
}
if (max_devs == 0)
@@ -384,8 +381,8 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
/* check unit id */
if (max_devs && unit_id >= max_devs) {
fprintf(stderr, "qemu: unit %d too big (max is %d)\n",
unit_id, max_devs - 1);
error_report("unit %d too big (max is %d)",
unit_id, max_devs - 1);
return NULL;
}
@@ -479,7 +476,7 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
ro = 1;
} else if (ro == 1) {
if (type != IF_SCSI && type != IF_VIRTIO && type != IF_FLOPPY && type != IF_NONE) {
fprintf(stderr, "qemu: readonly flag not supported for drive with this interface\n");
error_report("readonly not supported by this bus type");
return NULL;
}
}
@@ -488,8 +485,8 @@ DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi, int *fatal_error)
ret = bdrv_open(dinfo->bdrv, file, bdrv_flags, drv);
if (ret < 0) {
fprintf(stderr, "qemu: could not open disk image %s: %s\n",
file, strerror(-ret));
error_report("could not open disk image %s: %s",
file, strerror(-ret));
return NULL;
}
@@ -526,6 +523,12 @@ int do_snapshot_blkdev(Monitor *mon, const QDict *qdict, QObject **ret_data)
int ret = 0;
int flags;
if (!filename) {
qerror_report(QERR_MISSING_PARAMETER, "snapshot_file");
ret = -1;
goto out;
}
bs = bdrv_find(device);
if (!bs) {
qerror_report(QERR_DEVICE_NOT_FOUND, device);
@@ -684,13 +687,15 @@ int do_drive_del(Monitor *mon, const QDict *qdict, QObject **ret_data)
/* clean up guest state from pointing to host resource by
* finding and removing DeviceState "drive" property */
for (prop = bs->peer->info->props; prop && prop->name; prop++) {
if (prop->info->type == PROP_TYPE_DRIVE) {
ptr = qdev_get_prop_ptr(bs->peer, prop);
if ((*ptr) == bs) {
bdrv_detach(bs, bs->peer);
*ptr = NULL;
break;
if (bs->peer) {
for (prop = bs->peer->info->props; prop && prop->name; prop++) {
if (prop->info->type == PROP_TYPE_DRIVE) {
ptr = qdev_get_prop_ptr(bs->peer, prop);
if (*ptr == bs) {
bdrv_detach(bs, bs->peer);
*ptr = NULL;
break;
}
}
}
}
cutils.c
+4 -4
View File
@@ -291,9 +291,9 @@ int fcntl_setfl(int fd, int flag)
* value must be terminated by whitespace, ',' or '\0'. Return -1 on
* error.
*/
ssize_t strtosz_suffix(const char *nptr, char **end, const char default_suffix)
int64_t strtosz_suffix(const char *nptr, char **end, const char default_suffix)
{
ssize_t retval = -1;
int64_t retval = -1;
char *endptr, c, d;
int mul_required = 0;
double val, mul, integral, fraction;
@@ -365,7 +365,7 @@ ssize_t strtosz_suffix(const char *nptr, char **end, const char default_suffix)
goto fail;
}
}
if ((val * mul >= ~(size_t)0) || val < 0) {
if ((val * mul >= INT64_MAX) || val < 0) {
goto fail;
}
retval = val * mul;
@@ -378,7 +378,7 @@ fail:
return retval;
}
ssize_t strtosz(const char *nptr, char **end)
int64_t strtosz(const char *nptr, char **end)
{
return strtosz_suffix(nptr, end, STRTOSZ_DEFSUFFIX_MB);
}
docs/qdev-device-use.txt
+11 -2
View File
@@ -80,7 +80,11 @@ The -device argument differs in detail for each kind of drive:
This SCSI controller a single SCSI bus, named ID.0. Put a disk on
it:
-device scsi-disk,drive=DRIVE-ID,bus=ID.0,scsi-id=SCSI-ID
-device scsi-disk,drive=DRIVE-ID,bus=ID.0,scsi-id=SCSI-ID,removable=RMB
The (optional) removable parameter lets you override the SCSI INQUIRY
removable (RMB) bit for non CD-ROM devices. It is ignored for CD-ROM devices
which are always removable. RMB is "on" or "off".
* if=floppy
@@ -116,7 +120,12 @@ For USB devices, the old way is actually different:
Provides much less control than -drive's HOST-OPTS... The new way
fixes that:
-device usb-storage,drive=DRIVE-ID
-device usb-storage,drive=DRIVE-ID,removable=RMB
The removable parameter gives control over the SCSI INQUIRY removable (RMB)
bit. USB thumbdrives usually set removable=on, while USB hard disks set
removable=off. See the if=scsi description above for details on the removable
parameter, which applies only to scsi-disk devices and not to scsi-generic.
=== Character Devices ===
hw/ide/core.c
+36 -77
View File
@@ -321,14 +321,6 @@ static inline void ide_abort_command(IDEState *s)
s->error = ABRT_ERR;
}
static inline void ide_dma_submit_check(IDEState *s,
BlockDriverCompletionFunc *dma_cb)
{
if (s->bus->dma->aiocb)
return;
dma_cb(s, -1);
}
/* prepare data transfer and tell what to do after */
static void ide_transfer_start(IDEState *s, uint8_t *buf, int size,
EndTransferFunc *end_transfer_func)
@@ -487,16 +479,19 @@ static int ide_handle_rw_error(IDEState *s, int error, int op)
return 1;
}
void ide_read_dma_cb(void *opaque, int ret)
void ide_dma_cb(void *opaque, int ret)
{
IDEState *s = opaque;
int n;
int64_t sector_num;
handle_rw_error:
if (ret < 0) {
if (ide_handle_rw_error(s, -ret,
BM_STATUS_DMA_RETRY | BM_STATUS_RETRY_READ))
{
int op = BM_STATUS_DMA_RETRY;
if (s->is_read)
op |= BM_STATUS_RETRY_READ;
if (ide_handle_rw_error(s, -ret, op)) {
return;
}
}
@@ -504,7 +499,7 @@ void ide_read_dma_cb(void *opaque, int ret)
n = s->io_buffer_size >> 9;
sector_num = ide_get_sector(s);
if (n > 0) {
dma_buf_commit(s, 1);
dma_buf_commit(s, s->is_read);
sector_num += n;
ide_set_sector(s, sector_num);
s->nsector -= n;
@@ -514,32 +509,47 @@ void ide_read_dma_cb(void *opaque, int ret)
if (s->nsector == 0) {
s->status = READY_STAT | SEEK_STAT;
ide_set_irq(s->bus);
eot:
s->bus->dma->ops->add_status(s->bus->dma, BM_STATUS_INT);
ide_set_inactive(s);
return;
goto eot;
}
/* launch next transfer */
n = s->nsector;
s->io_buffer_index = 0;
s->io_buffer_size = n * 512;
if (s->bus->dma->ops->prepare_buf(s->bus->dma, 1) == 0)
if (s->bus->dma->ops->prepare_buf(s->bus->dma, s->is_read) == 0)
goto eot;
#ifdef DEBUG_AIO
printf("aio_read: sector_num=%" PRId64 " n=%d\n", sector_num, n);
printf("ide_dma_cb: sector_num=%" PRId64 " n=%d, is_read=%d\n",
sector_num, n, s->is_read);
#endif
s->bus->dma->aiocb = dma_bdrv_read(s->bs, &s->sg, sector_num, ide_read_dma_cb, s);
ide_dma_submit_check(s, ide_read_dma_cb);
if (s->is_read) {
s->bus->dma->aiocb = dma_bdrv_read(s->bs, &s->sg, sector_num,
ide_dma_cb, s);
} else {
s->bus->dma->aiocb = dma_bdrv_write(s->bs, &s->sg, sector_num,
ide_dma_cb, s);
}
if (!s->bus->dma->aiocb) {
ret = -1;
goto handle_rw_error;
}
return;
eot:
s->bus->dma->ops->add_status(s->bus->dma, BM_STATUS_INT);
ide_set_inactive(s);
}
static void ide_sector_read_dma(IDEState *s)
static void ide_sector_start_dma(IDEState *s, int is_read)
{
s->status = READY_STAT | SEEK_STAT | DRQ_STAT | BUSY_STAT;
s->io_buffer_index = 0;
s->io_buffer_size = 0;
s->is_read = 1;
s->bus->dma->ops->start_dma(s->bus->dma, s, ide_read_dma_cb);
s->is_read = is_read;
s->bus->dma->ops->start_dma(s->bus->dma, s, ide_dma_cb);
}
static void ide_sector_write_timer_cb(void *opaque)
@@ -594,57 +604,6 @@ void ide_sector_write(IDEState *s)
}
}
void ide_write_dma_cb(void *opaque, int ret)
{
IDEState *s = opaque;
int n;
int64_t sector_num;
if (ret < 0) {
if (ide_handle_rw_error(s, -ret, BM_STATUS_DMA_RETRY))
return;
}
n = s->io_buffer_size >> 9;
sector_num = ide_get_sector(s);
if (n > 0) {
dma_buf_commit(s, 0);
sector_num += n;
ide_set_sector(s, sector_num);
s->nsector -= n;
}
/* end of transfer ? */
if (s->nsector == 0) {
s->status = READY_STAT | SEEK_STAT;
ide_set_irq(s->bus);
eot:
s->bus->dma->ops->add_status(s->bus->dma, BM_STATUS_INT);
ide_set_inactive(s);
return;
}
n = s->nsector;
s->io_buffer_size = n * 512;
/* launch next transfer */
if (s->bus->dma->ops->prepare_buf(s->bus->dma, 0) == 0)
goto eot;
#ifdef DEBUG_AIO
printf("aio_write: sector_num=%" PRId64 " n=%d\n", sector_num, n);
#endif
s->bus->dma->aiocb = dma_bdrv_write(s->bs, &s->sg, sector_num, ide_write_dma_cb, s);
ide_dma_submit_check(s, ide_write_dma_cb);
}
static void ide_sector_write_dma(IDEState *s)
{
s->status = READY_STAT | SEEK_STAT | DRQ_STAT | BUSY_STAT;
s->io_buffer_index = 0;
s->io_buffer_size = 0;
s->is_read = 0;
s->bus->dma->ops->start_dma(s->bus->dma, s, ide_write_dma_cb);
}
void ide_atapi_cmd_ok(IDEState *s)
{
s->error = 0;
@@ -1858,7 +1817,7 @@ void ide_exec_cmd(IDEBus *bus, uint32_t val)
if (!s->bs)
goto abort_cmd;
ide_cmd_lba48_transform(s, lba48);
ide_sector_read_dma(s);
ide_sector_start_dma(s, 1);
break;
case WIN_WRITEDMA_EXT:
lba48 = 1;
@@ -1867,7 +1826,7 @@ void ide_exec_cmd(IDEBus *bus, uint32_t val)
if (!s->bs)
goto abort_cmd;
ide_cmd_lba48_transform(s, lba48);
ide_sector_write_dma(s);
ide_sector_start_dma(s, 0);
s->media_changed = 1;
break;
case WIN_READ_NATIVE_MAX_EXT:
hw/ide/internal.h
+1 -3
View File
@@ -439,7 +439,6 @@ struct IDEState {
uint32_t mdata_size;
uint8_t *mdata_storage;
int media_changed;
/* for pmac */
int is_read;
/* SMART */
uint8_t smart_enabled;
@@ -560,8 +559,7 @@ void ide_init2_with_non_qdev_drives(IDEBus *bus, DriveInfo *hd0,
void ide_init_ioport(IDEBus *bus, int iobase, int iobase2);
void ide_exec_cmd(IDEBus *bus, uint32_t val);
void ide_read_dma_cb(void *opaque, int ret);
void ide_write_dma_cb(void *opaque, int ret);
void ide_dma_cb(void *opaque, int ret);
void ide_sector_write(IDEState *s);
void ide_sector_read(IDEState *s);
void ide_flush_cache(IDEState *s);
hw/ide/pci.c
+3 -10
View File
@@ -178,14 +178,9 @@ static void bmdma_restart_dma(BMDMAState *bm, int is_read)
s->io_buffer_index = 0;
s->io_buffer_size = 0;
s->nsector = bm->nsector;
s->is_read = is_read;
bm->cur_addr = bm->addr;
if (is_read) {
bm->dma_cb = ide_read_dma_cb;
} else {
bm->dma_cb = ide_write_dma_cb;
}
bm->dma_cb = ide_dma_cb;
bmdma_start_dma(&bm->dma, s, bm->dma_cb);
}
@@ -272,9 +267,7 @@ static void bmdma_irq(void *opaque, int n, int level)
return;
}
if (bm) {
bm->status |= BM_STATUS_INT;
}
bm->status |= BM_STATUS_INT;
/* trigger the real irq */
qemu_set_irq(bm->irq, level);
hw/pci-hotplug.c
+1 -1
View File
@@ -90,7 +90,7 @@ static int scsi_hot_add(Monitor *mon, DeviceState *adapter,
* specified).
*/
dinfo->unit = qemu_opt_get_number(dinfo->opts, "unit", -1);
scsidev = scsi_bus_legacy_add_drive(scsibus, dinfo->bdrv, dinfo->unit);
scsidev = scsi_bus_legacy_add_drive(scsibus, dinfo->bdrv, dinfo->unit, false);
if (!scsidev) {
return -1;
}
hw/scsi-bus.c
+6 -2
View File
@@ -87,7 +87,8 @@ void scsi_qdev_register(SCSIDeviceInfo *info)
}
/* handle legacy '-drive if=scsi,...' cmd line args */
SCSIDevice *scsi_bus_legacy_add_drive(SCSIBus *bus, BlockDriverState *bdrv, int unit)
SCSIDevice *scsi_bus_legacy_add_drive(SCSIBus *bus, BlockDriverState *bdrv,
int unit, bool removable)
{
const char *driver;
DeviceState *dev;
@@ -95,6 +96,9 @@ SCSIDevice *scsi_bus_legacy_add_drive(SCSIBus *bus, BlockDriverState *bdrv, int
driver = bdrv_is_sg(bdrv) ? "scsi-generic" : "scsi-disk";
dev = qdev_create(&bus->qbus, driver);
qdev_prop_set_uint32(dev, "scsi-id", unit);
if (qdev_prop_exists(dev, "removable")) {
qdev_prop_set_bit(dev, "removable", removable);
}
if (qdev_prop_set_drive(dev, "drive", bdrv) < 0) {
qdev_free(dev);
return NULL;
@@ -117,7 +121,7 @@ int scsi_bus_legacy_handle_cmdline(SCSIBus *bus)
continue;
}
qemu_opts_loc_restore(dinfo->opts);
if (!scsi_bus_legacy_add_drive(bus, dinfo->bdrv, unit)) {
if (!scsi_bus_legacy_add_drive(bus, dinfo->bdrv, unit, false)) {
res = -1;
break;
}
hw/scsi-disk.c
+3
View File
@@ -72,6 +72,7 @@ struct SCSIDiskState
/* The qemu block layer uses a fixed 512 byte sector size.
This is the number of 512 byte blocks in a single scsi sector. */
int cluster_size;
uint32_t removable;
uint64_t max_lba;
QEMUBH *bh;
char *version;
@@ -552,6 +553,7 @@ static int scsi_disk_emulate_inquiry(SCSIRequest *req, uint8_t *outbuf)
memcpy(&outbuf[16], "QEMU CD-ROM ", 16);
} else {
outbuf[0] = 0;
outbuf[1] = s->removable ? 0x80 : 0;
memcpy(&outbuf[16], "QEMU HARDDISK ", 16);
}
memcpy(&outbuf[8], "QEMU ", 8);
@@ -1295,6 +1297,7 @@ static SCSIDeviceInfo scsi_disk_info = {
DEFINE_BLOCK_PROPERTIES(SCSIDiskState, qdev.conf),
DEFINE_PROP_STRING("ver", SCSIDiskState, version),
DEFINE_PROP_STRING("serial", SCSIDiskState, serial),
DEFINE_PROP_BIT("removable", SCSIDiskState, removable, 0, false),
DEFINE_PROP_END_OF_LIST(),
},
};
hw/scsi.h
+2 -1
View File
@@ -94,7 +94,8 @@ static inline SCSIBus *scsi_bus_from_device(SCSIDevice *d)
return DO_UPCAST(SCSIBus, qbus, d->qdev.parent_bus);
}
SCSIDevice *scsi_bus_legacy_add_drive(SCSIBus *bus, BlockDriverState *bdrv, int unit);
SCSIDevice *scsi_bus_legacy_add_drive(SCSIBus *bus, BlockDriverState *bdrv,
int unit, bool removable);
int scsi_bus_legacy_handle_cmdline(SCSIBus *bus);
SCSIRequest *scsi_req_alloc(size_t size, SCSIDevice *d, uint32_t tag, uint32_t lun);
hw/usb-msd.c
+3 -1
View File
@@ -51,6 +51,7 @@ typedef struct {
SCSIBus bus;
BlockConf conf;
SCSIDevice *scsi_dev;
uint32_t removable;
int result;
/* For async completion. */
USBPacket *packet;
@@ -515,7 +516,7 @@ static int usb_msd_initfn(USBDevice *dev)
usb_desc_init(dev);
scsi_bus_new(&s->bus, &s->dev.qdev, 0, 1, usb_msd_command_complete);
s->scsi_dev = scsi_bus_legacy_add_drive(&s->bus, bs, 0);
s->scsi_dev = scsi_bus_legacy_add_drive(&s->bus, bs, 0, !!s->removable);
if (!s->scsi_dev) {
return -1;
}
@@ -607,6 +608,7 @@ static struct USBDeviceInfo msd_info = {
.usbdevice_init = usb_msd_init,
.qdev.props = (Property[]) {
DEFINE_BLOCK_PROPERTIES(MSDState, conf),
DEFINE_PROP_BIT("removable", MSDState, removable, 0, false),
DEFINE_PROP_END_OF_LIST(),
},
};

Some files were not shown because too many files have changed in this diff Show More