ShuriZma
/
xemu
mirror of https://github.com/xemu-project/xemu.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Block layer patches: 

- qcow2: Decompression worker threads
 - dmg: lzfse compression support
 - file-posix: Simplify delegation to worker thread
 - Don't pass flags to bdrv_reopen_queue()
 - iotests: make 235 work on s390 (and others)
 -----BEGIN PGP SIGNATURE-----
 
 iQIcBAABAgAGBQJcE4wNAAoJEH8JsnLIjy/WpJQP/39XmFQr/UO/Z7fsQNJD7Kbn
 yUzAunMt7r7nfyuC5CP7a57apjKzbLHIbKDKrI8v2/SHysZ2zvjGx9QFCYNM44P7
 XRmwd/fJJUqcyaDZDjiIHZtfSvVQB09xOjl62K9b6tVYCTztBwqVzY9uE4oA0coh
 tAofAwG8vHYYxhjkPxKaftBv/GO/a9jB1Dk6DG7cX4FUm0lwEnGcT3ZmRNUBRAQ4
 F0HfG+OubqljHOSR3VN3PPoienDwQOTsroqhIL4R0Jeb6I/1IVyeO56C4WYrfn9L
 Tjgsu1v/te4F+7/BBICQKp5y9nNYrg6uPlC4cD/st/xZQe0oMUHEGcSESm61wOc5
 bP8A5D7iiCn1c3kZXrPVyuvUQBn3fIJUOgVHQ7Oa4x2i9VcjpzQKAL2Wuu9NEgwc
 Acn9lj9ey3rZwcJisCyOchn5sG/M4dYstHP8aAUafeSpAvsXje+hPKnWe0+SqxZx
 btmVt6Suh205fP86w9POeNzy1la69FzF/xqe3Eohl5mEZsylL5jT0w9CfAzJSJrz
 dDhgnelgQZ0/YcoEc1pqqQ8EP+9EJuIzjB7mEaCfZUmylq7mL/QvWgtjSbIr1yFG
 RFvg6wTqcnrtOKoLvLSfw64QJXgDFwQ3cZ7Wl8XakZNPMfffndk9AThQxBBgofqg
 XOyuW5gg3g3xzZrQswsf
 =XKq9
 -----END PGP SIGNATURE-----

Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches:

- qcow2: Decompression worker threads
- dmg: lzfse compression support
- file-posix: Simplify delegation to worker thread
- Don't pass flags to bdrv_reopen_queue()
- iotests: make 235 work on s390 (and others)

# gpg: Signature made Fri 14 Dec 2018 10:55:09 GMT
# gpg:                using RSA key 7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kevin/tags/for-upstream: (42 commits)
  block/mirror: add missing coroutine_fn annotations
  iotests: make 235 work on s390 (and others)
  block: Assert that flags are up-to-date in bdrv_reopen_prepare()
  block: Remove assertions from update_flags_from_options()
  block: Stop passing flags to bdrv_reopen_queue_child()
  block: Remove flags parameter from bdrv_reopen_queue()
  block: Clean up reopen_backing_file() in block/replication.c
  qemu-io: Put flag changes in the options QDict in reopen_f()
  block: Drop bdrv_reopen()
  block: Use bdrv_reopen_set_read_only() in the mirror driver
  block: Use bdrv_reopen_set_read_only() in external_snapshot_commit()
  block: Use bdrv_reopen_set_read_only() in qmp_change_backing_file()
  block: Use bdrv_reopen_set_read_only() in stream_start/complete()
  block: Use bdrv_reopen_set_read_only() in bdrv_commit()
  block: Use bdrv_reopen_set_read_only() in commit_start/complete()
  block: Use bdrv_reopen_set_read_only() in bdrv_backing_update_filename()
  block: Add bdrv_reopen_set_read_only()
  file-posix: Avoid aio_worker() for QEMU_AIO_IOCTL
  file-posix: Switch to .bdrv_co_ioctl
  file-posix: Remove paio_submit_co()
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This commit is contained in:
Peter Maydell 2018-12-16 12:49:06 +00:00
parent 3866e6bebd 537c3d4f64
commit 58b1f0f21e
25 changed files with 674 additions and 510 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

89
block.c
View File

@ -1079,11 +1079,11 @@ static int bdrv_backing_update_filename(BdrvChild *c, BlockDriverState *base,
const char *filename, Error **errp)
{
BlockDriverState *parent = c->opaque;
int orig_flags = bdrv_get_flags(parent);
bool read_only = bdrv_is_read_only(parent);
int ret;
if (!(orig_flags & BDRV_O_RDWR)) {
ret = bdrv_reopen(parent, orig_flags | BDRV_O_RDWR, errp);
if (read_only) {
ret = bdrv_reopen_set_read_only(parent, false, errp);
if (ret < 0) {
return ret;
}
@ -1095,8 +1095,8 @@ static int bdrv_backing_update_filename(BdrvChild *c, BlockDriverState *base,
error_setg_errno(errp, -ret, "Could not update backing file link");
}
if (!(orig_flags & BDRV_O_RDWR)) {
bdrv_reopen(parent, orig_flags, NULL);
if (read_only) {
bdrv_reopen_set_read_only(parent, true, NULL);
}
return ret;
@ -1139,22 +1139,18 @@ static void update_flags_from_options(int *flags, QemuOpts *opts)
{
*flags &= ~(BDRV_O_CACHE_MASK | BDRV_O_RDWR | BDRV_O_AUTO_RDONLY);
assert(qemu_opt_find(opts, BDRV_OPT_CACHE_NO_FLUSH));
if (qemu_opt_get_bool_del(opts, BDRV_OPT_CACHE_NO_FLUSH, false)) {
*flags |= BDRV_O_NO_FLUSH;
}
assert(qemu_opt_find(opts, BDRV_OPT_CACHE_DIRECT));
if (qemu_opt_get_bool_del(opts, BDRV_OPT_CACHE_DIRECT, false)) {
*flags |= BDRV_O_NOCACHE;
}
assert(qemu_opt_find(opts, BDRV_OPT_READ_ONLY));
if (!qemu_opt_get_bool_del(opts, BDRV_OPT_READ_ONLY, false)) {
*flags |= BDRV_O_RDWR;
}
assert(qemu_opt_find(opts, BDRV_OPT_AUTO_READ_ONLY));
if (qemu_opt_get_bool_del(opts, BDRV_OPT_AUTO_READ_ONLY, false)) {
*flags |= BDRV_O_AUTO_RDONLY;
}
@ -2931,7 +2927,6 @@ BlockDriverState *bdrv_open(const char *filename, const char *reference,
static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue,
BlockDriverState *bs,
QDict *options,
int flags,
const BdrvChildRole *role,
QDict *parent_options,
int parent_flags)
@ -2940,7 +2935,9 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue,
BlockReopenQueueEntry *bs_entry;
BdrvChild *child;
QDict *old_options, *explicit_options;
QDict *old_options, *explicit_options, *options_copy;
int flags;
QemuOpts *opts;
/* Make sure that the caller remembered to use a drained section. This is
* important to avoid graph changes between the recursive queuing here and
@ -2966,22 +2963,11 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue,
/*
* Precedence of options:
* 1. Explicitly passed in options (highest)
* 2. Set in flags (only for top level)
* 3. Retained from explicitly set options of bs
* 4. Inherited from parent node
* 5. Retained from effective options of bs
* 2. Retained from explicitly set options of bs
* 3. Inherited from parent node
* 4. Retained from effective options of bs
*/
if (!parent_options) {
/*
* Any setting represented by flags is always updated. If the
* corresponding QDict option is set, it takes precedence. Otherwise
* the flag is translated into a QDict option. The old setting of bs is
* not considered.
*/
update_options_from_flags(options, flags);
}
/* Old explicitly set values (don't overwrite by inherited value) */
if (bs_entry) {
old_options = qdict_clone_shallow(bs_entry->state.explicit_options);
@ -2995,16 +2981,10 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue,
/* Inherit from parent node */
if (parent_options) {
QemuOpts *opts;
QDict *options_copy;
assert(!flags);
flags = 0;
role->inherit_options(&flags, options, parent_flags, parent_options);
options_copy = qdict_clone_shallow(options);
opts = qemu_opts_create(&bdrv_runtime_opts, NULL, 0, &error_abort);
qemu_opts_absorb_qdict(opts, options_copy, NULL);
update_flags_from_options(&flags, opts);
qemu_opts_del(opts);
qobject_unref(options_copy);
} else {
flags = bdrv_get_flags(bs);
}
/* Old values are used for options that aren't set yet */
@ -3012,6 +2992,14 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue,
bdrv_join_options(bs, options, old_options);
qobject_unref(old_options);
/* We have the final set of options so let's update the flags */
options_copy = qdict_clone_shallow(options);
opts = qemu_opts_create(&bdrv_runtime_opts, NULL, 0, &error_abort);
qemu_opts_absorb_qdict(opts, options_copy, NULL);
update_flags_from_options(&flags, opts);
qemu_opts_del(opts);
qobject_unref(options_copy);
/* bdrv_open_inherit() sets and clears some additional flags internally */
flags &= ~BDRV_O_PROTOCOL;
if (flags & BDRV_O_RDWR) {
@ -3051,7 +3039,7 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue,
qdict_extract_subqdict(options, &new_child_options, child_key_dot);
g_free(child_key_dot);
bdrv_reopen_queue_child(bs_queue, child->bs, new_child_options, 0,
bdrv_reopen_queue_child(bs_queue, child->bs, new_child_options,
child->role, options, flags);
}
@ -3060,10 +3048,9 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue,
BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue,
BlockDriverState *bs,
QDict *options, int flags)
QDict *options)
{
return bdrv_reopen_queue_child(bs_queue, bs, options, flags,
NULL, NULL, 0);
return bdrv_reopen_queue_child(bs_queue, bs, options, NULL, NULL, 0);
}
/*
@ -3125,22 +3112,18 @@ cleanup:
return ret;
}
/* Reopen a single BlockDriverState with the specified flags. */
int bdrv_reopen(BlockDriverState *bs, int bdrv_flags, Error **errp)
int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only,
Error **errp)
{
int ret = -1;
Error *local_err = NULL;
int ret;
BlockReopenQueue *queue;
QDict *opts = qdict_new();
qdict_put_bool(opts, BDRV_OPT_READ_ONLY, read_only);
bdrv_subtree_drained_begin(bs);
queue = bdrv_reopen_queue(NULL, bs, NULL, bdrv_flags);
ret = bdrv_reopen_multiple(bdrv_get_aio_context(bs), queue, &local_err);
if (local_err != NULL) {
error_propagate(errp, local_err);
}
queue = bdrv_reopen_queue(NULL, bs, opts);
ret = bdrv_reopen_multiple(bdrv_get_aio_context(bs), queue, errp);
bdrv_subtree_drained_end(bs);
return ret;
@ -3214,6 +3197,7 @@ int bdrv_reopen_prepare(BDRVReopenState *reopen_state, BlockReopenQueue *queue,
Error **errp)
{
int ret = -1;
int old_flags;
Error *local_err = NULL;
BlockDriver *drv;
QemuOpts *opts;
@ -3240,7 +3224,12 @@ int bdrv_reopen_prepare(BDRVReopenState *reopen_state, BlockReopenQueue *queue,
goto error;
}
/* This was already called in bdrv_reopen_queue_child() so the flags
* are up-to-date. This time we simply want to remove the options from
* QemuOpts in order to indicate that they have been processed. */
old_flags = reopen_state->flags;
update_flags_from_options(&reopen_state->flags, opts);
assert(old_flags == reopen_state->flags);
discard = qemu_opt_get_del(opts, BDRV_OPT_DISCARD);
if (discard != NULL) {

2
block/Makefile.objs
View File

@ -57,6 +57,8 @@ ssh.o-libs := $(LIBSSH2_LIBS)
block-obj-dmg-bz2-$(CONFIG_BZIP2) += dmg-bz2.o
block-obj-$(if $(CONFIG_DMG),m,n) += $(block-obj-dmg-bz2-y)
dmg-bz2.o-libs := $(BZIP2_LIBS)
block-obj-$(if $(CONFIG_LZFSE),m,n) += dmg-lzfse.o
dmg-lzfse.o-libs := $(LZFSE_LIBS)
qcow.o-libs := -lz
linux-aio.o-libs := -laio
parallels.o-cflags := $(LIBXML2_CFLAGS)

38
block/backup.c
View File

@ -28,6 +28,13 @@
#define BACKUP_CLUSTER_SIZE_DEFAULT (1 << 16)
typedef struct CowRequest {
int64_t start_byte;
int64_t end_byte;
QLIST_ENTRY(CowRequest) list;
CoQueue wait_queue; /* coroutines blocked on this request */
} CowRequest;
typedef struct BackupBlockJob {
BlockJob common;
BlockBackend *target;
@ -322,37 +329,6 @@ void backup_do_checkpoint(BlockJob *job, Error **errp)
hbitmap_set(backup_job->copy_bitmap, 0, len);
}
void backup_wait_for_overlapping_requests(BlockJob *job, int64_t offset,
uint64_t bytes)
{
BackupBlockJob *backup_job = container_of(job, BackupBlockJob, common);
int64_t start, end;
assert(block_job_driver(job) == &backup_job_driver);
start = QEMU_ALIGN_DOWN(offset, backup_job->cluster_size);
end = QEMU_ALIGN_UP(offset + bytes, backup_job->cluster_size);
wait_for_overlapping_requests(backup_job, start, end);
}
void backup_cow_request_begin(CowRequest *req, BlockJob *job,
int64_t offset, uint64_t bytes)
{
BackupBlockJob *backup_job = container_of(job, BackupBlockJob, common);
int64_t start, end;
assert(block_job_driver(job) == &backup_job_driver);
start = QEMU_ALIGN_DOWN(offset, backup_job->cluster_size);
end = QEMU_ALIGN_UP(offset + bytes, backup_job->cluster_size);
cow_request_begin(req, backup_job, start, end);
}
void backup_cow_request_end(CowRequest *req)
{
cow_request_end(req);
}
static void backup_drain(BlockJob *job)
{
BackupBlockJob *s = container_of(job, BackupBlockJob, common);

23
block/commit.c
View File

@ -38,7 +38,7 @@ typedef struct CommitBlockJob {
BlockBackend *base;
BlockDriverState *base_bs;
BlockdevOnError on_error;
int base_flags;
bool base_read_only;
char *backing_file_str;
} CommitBlockJob;
@ -124,8 +124,8 @@ static void commit_clean(Job *job)
/* restore base open flags here if appropriate (e.g., change the base back
* to r/o). These reopens do not need to be atomic, since we won't abort
* even on failure here */
if (s->base_flags != bdrv_get_flags(s->base_bs)) {
bdrv_reopen(s->base_bs, s->base_flags, NULL);
if (s->base_read_only) {
bdrv_reopen_set_read_only(s->base_bs, true, NULL);
}
g_free(s->backing_file_str);
@ -264,7 +264,6 @@ void commit_start(const char *job_id, BlockDriverState *bs,
const char *filter_node_name, Error **errp)
{
CommitBlockJob *s;
int orig_base_flags;
BlockDriverState *iter;
BlockDriverState *commit_top_bs = NULL;
Error *local_err = NULL;
@ -283,11 +282,9 @@ void commit_start(const char *job_id, BlockDriverState *bs,
}
/* convert base to r/w, if necessary */
orig_base_flags = bdrv_get_flags(base);
if (!(orig_base_flags & BDRV_O_RDWR)) {
bdrv_reopen(base, orig_base_flags | BDRV_O_RDWR, &local_err);
if (local_err != NULL) {
error_propagate(errp, local_err);
s->base_read_only = bdrv_is_read_only(base);
if (s->base_read_only) {
if (bdrv_reopen_set_read_only(base, false, errp) != 0) {
goto fail;
}
}
@ -363,7 +360,6 @@ void commit_start(const char *job_id, BlockDriverState *bs,
goto fail;
}
s->base_flags = orig_base_flags;
s->backing_file_str = g_strdup(backing_file_str);
s->on_error = on_error;
@ -395,7 +391,7 @@ int bdrv_commit(BlockDriverState *bs)
BlockDriverState *commit_top_bs = NULL;
BlockDriver *drv = bs->drv;
int64_t offset, length, backing_length;
int ro, open_flags;
int ro;
int64_t n;
int ret = 0;
uint8_t *buf = NULL;
@ -414,10 +410,9 @@ int bdrv_commit(BlockDriverState *bs)
}
ro = bs->backing->bs->read_only;
open_flags = bs->backing->bs->open_flags;
if (ro) {
if (bdrv_reopen(bs->backing->bs, open_flags | BDRV_O_RDWR, NULL)) {
if (bdrv_reopen_set_read_only(bs->backing->bs, false, NULL)) {
return -EACCES;
}
}
@ -527,7 +522,7 @@ ro_cleanup:
if (ro) {
/* ignoring error return here */
bdrv_reopen(bs->backing->bs, open_flags & ~BDRV_O_RDWR, NULL);
bdrv_reopen_set_read_only(bs->backing->bs, true, NULL);
}
return ret;

49
block/dmg-lzfse.c Normal file
View File

@ -0,0 +1,49 @@
/*
* DMG lzfse uncompression
*
* Copyright (c) 2018 Julio Cesar Faracco
*
* 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 "qemu/osdep.h"
#include "qemu-common.h"
#include "dmg.h"
#include <lzfse.h>
static int dmg_uncompress_lzfse_do(char *next_in, unsigned int avail_in,
char *next_out, unsigned int avail_out)
{
size_t out_size = lzfse_decode_buffer((uint8_t *) next_out, avail_out,
(uint8_t *) next_in, avail_in,
NULL);
/* We need to decode the single chunk only. */
/* So, out_size == avail_out is not an error here. */
if (out_size > 0) {
return out_size;
}
return -1;
}
__attribute__((constructor))
static void dmg_lzfse_init(void)
{
assert(!dmg_uncompress_lzfse);
dmg_uncompress_lzfse = dmg_uncompress_lzfse_do;
}

65
block/dmg.c
View File

@ -33,6 +33,9 @@
int (*dmg_uncompress_bz2)(char *next_in, unsigned int avail_in,
char *next_out, unsigned int avail_out);
int (*dmg_uncompress_lzfse)(char *next_in, unsigned int avail_in,
char *next_out, unsigned int avail_out);
enum {
/* Limit chunk sizes to prevent unreasonable amounts of memory being used
* or truncating when converting to 32-bit types
@ -41,6 +44,19 @@ enum {
DMG_SECTORCOUNTS_MAX = DMG_LENGTHS_MAX / 512,
};
enum {
/* DMG Block Type */
UDZE = 0, /* Zeroes */
UDRW, /* RAW type */
UDIG, /* Ignore */
UDCO = 0x80000004,
UDZO,
UDBZ,
ULFO,
UDCM = 0x7ffffffe, /* Comments */
UDLE /* Last Entry */
};
static int dmg_probe(const uint8_t *buf, int buf_size, const char *filename)
{
int len;
@ -105,15 +121,16 @@ static void update_max_chunk_size(BDRVDMGState *s, uint32_t chunk,
uint32_t uncompressed_sectors = 0;
switch (s->types[chunk]) {
case 0x80000005: /* zlib compressed */
case 0x80000006: /* bzip2 compressed */
case UDZO: /* zlib compressed */
case UDBZ: /* bzip2 compressed */
case ULFO: /* lzfse compressed */
compressed_size = s->lengths[chunk];
uncompressed_sectors = s->sectorcounts[chunk];
break;
case 1: /* copy */
case UDRW: /* copy */
uncompressed_sectors = DIV_ROUND_UP(s->lengths[chunk], 512);
break;
case 2: /* zero */
case UDIG: /* zero */
/* as the all-zeroes block may be large, it is treated specially: the
* sector is not copied from a large buffer, a simple memset is used
* instead. Therefore uncompressed_sectors does not need to be set. */
@ -182,12 +199,14 @@ typedef struct DmgHeaderState {
static bool dmg_is_known_block_type(uint32_t entry_type)
{
switch (entry_type) {
case 0x00000001: /* uncompressed */
case 0x00000002: /* zeroes */
case 0x80000005: /* zlib */
case UDRW: /* uncompressed */
case UDIG: /* zeroes */
case UDZO: /* zlib */
return true;
case 0x80000006: /* bzip2 */
case UDBZ: /* bzip2 */
return !!dmg_uncompress_bz2;
case ULFO: /* lzfse */
return !!dmg_uncompress_lzfse;
default:
return false;
}
@ -425,6 +444,7 @@ static int dmg_open(BlockDriverState *bs, QDict *options, int flags,
}
block_module_load_one("dmg-bz2");
block_module_load_one("dmg-lzfse");
s->n_chunks = 0;
s->offsets = s->lengths = s->sectors = s->sectorcounts = NULL;
@ -579,7 +599,7 @@ static inline int dmg_read_chunk(BlockDriverState *bs, uint64_t sector_num)
s->current_chunk = s->n_chunks;
switch (s->types[chunk]) { /* block entry type */
case 0x80000005: { /* zlib compressed */
case UDZO: { /* zlib compressed */
/* we need to buffer, because only the chunk as whole can be
* inflated. */
ret = bdrv_pread(bs->file, s->offsets[chunk],
@ -602,7 +622,7 @@ static inline int dmg_read_chunk(BlockDriverState *bs, uint64_t sector_num)
return -1;
}
break; }
case 0x80000006: /* bzip2 compressed */
case UDBZ: /* bzip2 compressed */
if (!dmg_uncompress_bz2) {
break;
}
@ -623,14 +643,35 @@ static inline int dmg_read_chunk(BlockDriverState *bs, uint64_t sector_num)
return ret;
}
break;
case 1: /* copy */
case ULFO:
if (!dmg_uncompress_lzfse) {
break;
}
/* we need to buffer, because only the chunk as whole can be
* inflated. */
ret = bdrv_pread(bs->file, s->offsets[chunk],
s->compressed_chunk, s->lengths[chunk]);
if (ret != s->lengths[chunk]) {
return -1;
}
ret = dmg_uncompress_lzfse((char *)s->compressed_chunk,
(unsigned int) s->lengths[chunk],
(char *)s->uncompressed_chunk,
(unsigned int)
(512 * s->sectorcounts[chunk]));
if (ret < 0) {
return ret;
}
break;
case UDRW: /* copy */
ret = bdrv_pread(bs->file, s->offsets[chunk],
s->uncompressed_chunk, s->lengths[chunk]);
if (ret != s->lengths[chunk]) {
return -1;
}
break;
case 2: /* zero */
case UDIG: /* zero */
/* see dmg_read, it is treated specially. No buffer needs to be
* pre-filled, the zeroes can be set directly. */
break;

3
block/dmg.h
View File

@ -55,4 +55,7 @@ typedef struct BDRVDMGState {
extern int (*dmg_uncompress_bz2)(char *next_in, unsigned int avail_in,
char *next_out, unsigned int avail_out);
extern int (*dmg_uncompress_lzfse)(char *next_in, unsigned int avail_in,
char *next_out, unsigned int avail_out);
#endif

388
block/file-posix.c
View File

@ -182,25 +182,29 @@ static int64_t raw_getlength(BlockDriverState *bs);
typedef struct RawPosixAIOData {
BlockDriverState *bs;
int aio_fildes;
union {
struct iovec *aio_iov;
void *aio_ioctl_buf;
};
int aio_niov;
uint64_t aio_nbytes;
#define aio_ioctl_cmd aio_nbytes /* for QEMU_AIO_IOCTL */
off_t aio_offset;
int aio_type;
int aio_fildes;
off_t aio_offset;
uint64_t aio_nbytes;
union {
struct {
struct iovec *iov;
int niov;
} io;
struct {
uint64_t cmd;
void *buf;
} ioctl;
struct {
int aio_fd2;
off_t aio_offset2;
};
} copy_range;
struct {
PreallocMode prealloc;
Error **errp;
};
} truncate;
};
} RawPosixAIOData;
@ -1148,20 +1152,24 @@ static int hdev_probe_geometry(BlockDriverState *bs, HDGeometry *geo)
}
#endif
static ssize_t handle_aiocb_ioctl(RawPosixAIOData *aiocb)
#if defined(__linux__)
static int handle_aiocb_ioctl(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
int ret;
ret = ioctl(aiocb->aio_fildes, aiocb->aio_ioctl_cmd, aiocb->aio_ioctl_buf);
ret = ioctl(aiocb->aio_fildes, aiocb->ioctl.cmd, aiocb->ioctl.buf);
if (ret == -1) {
return -errno;
}
return 0;
}
#endif /* linux */
static ssize_t handle_aiocb_flush(RawPosixAIOData *aiocb)
static int handle_aiocb_flush(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
BDRVRawState *s = aiocb->bs->opaque;
int ret;
@ -1233,13 +1241,13 @@ static ssize_t handle_aiocb_rw_vector(RawPosixAIOData *aiocb)
do {
if (aiocb->aio_type & QEMU_AIO_WRITE)
len = qemu_pwritev(aiocb->aio_fildes,
aiocb->aio_iov,
aiocb->aio_niov,
aiocb->io.iov,
aiocb->io.niov,
aiocb->aio_offset);
else
len = qemu_preadv(aiocb->aio_fildes,
aiocb->aio_iov,
aiocb->aio_niov,
aiocb->io.iov,
aiocb->io.niov,
aiocb->aio_offset);
} while (len == -1 && errno == EINTR);
@ -1295,8 +1303,9 @@ static ssize_t handle_aiocb_rw_linear(RawPosixAIOData *aiocb, char *buf)
return offset;
}
static ssize_t handle_aiocb_rw(RawPosixAIOData *aiocb)
static int handle_aiocb_rw(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
ssize_t nbytes;
char *buf;
@ -1305,8 +1314,9 @@ static ssize_t handle_aiocb_rw(RawPosixAIOData *aiocb)
* If there is just a single buffer, and it is properly aligned
* we can just use plain pread/pwrite without any problems.
*/
if (aiocb->aio_niov == 1) {
return handle_aiocb_rw_linear(aiocb, aiocb->aio_iov->iov_base);
if (aiocb->io.niov == 1) {
nbytes = handle_aiocb_rw_linear(aiocb, aiocb->io.iov->iov_base);
goto out;
}
/*
* We have more than one iovec, and all are properly aligned.
@ -1318,7 +1328,7 @@ static ssize_t handle_aiocb_rw(RawPosixAIOData *aiocb)
nbytes = handle_aiocb_rw_vector(aiocb);
if (nbytes == aiocb->aio_nbytes ||
(nbytes < 0 && nbytes != -ENOSYS)) {
return nbytes;
goto out;
}
preadv_present = false;
}
@ -1336,16 +1346,17 @@ static ssize_t handle_aiocb_rw(RawPosixAIOData *aiocb)
*/
buf = qemu_try_blockalign(aiocb->bs, aiocb->aio_nbytes);
if (buf == NULL) {
return -ENOMEM;
nbytes = -ENOMEM;
goto out;
}
if (aiocb->aio_type & QEMU_AIO_WRITE) {
char *p = buf;
int i;
for (i = 0; i < aiocb->aio_niov; ++i) {
memcpy(p, aiocb->aio_iov[i].iov_base, aiocb->aio_iov[i].iov_len);
p += aiocb->aio_iov[i].iov_len;
for (i = 0; i < aiocb->io.niov; ++i) {
memcpy(p, aiocb->io.iov[i].iov_base, aiocb->io.iov[i].iov_len);
p += aiocb->io.iov[i].iov_len;
}
assert(p - buf == aiocb->aio_nbytes);
}
@ -1356,12 +1367,12 @@ static ssize_t handle_aiocb_rw(RawPosixAIOData *aiocb)
size_t count = aiocb->aio_nbytes, copy;
int i;
for (i = 0; i < aiocb->aio_niov && count; ++i) {
for (i = 0; i < aiocb->io.niov && count; ++i) {
copy = count;
if (copy > aiocb->aio_iov[i].iov_len) {
copy = aiocb->aio_iov[i].iov_len;
if (copy > aiocb->io.iov[i].iov_len) {
copy = aiocb->io.iov[i].iov_len;
}
memcpy(aiocb->aio_iov[i].iov_base, p, copy);
memcpy(aiocb->io.iov[i].iov_base, p, copy);
assert(count >= copy);
p += copy;
count -= copy;
@ -1370,7 +1381,21 @@ static ssize_t handle_aiocb_rw(RawPosixAIOData *aiocb)
}
qemu_vfree(buf);
return nbytes;
out:
if (nbytes == aiocb->aio_nbytes) {
return 0;
} else if (nbytes >= 0 && nbytes < aiocb->aio_nbytes) {
if (aiocb->aio_type & QEMU_AIO_WRITE) {
return -EINVAL;
} else {
iov_memset(aiocb->io.iov, aiocb->io.niov, nbytes,
0, aiocb->aio_nbytes - nbytes);
return 0;
}
} else {
assert(nbytes < 0);
return nbytes;
}
}
#ifdef CONFIG_XFS
@ -1460,8 +1485,9 @@ static ssize_t handle_aiocb_write_zeroes_block(RawPosixAIOData *aiocb)
return ret;
}
static ssize_t handle_aiocb_write_zeroes(RawPosixAIOData *aiocb)
static int handle_aiocb_write_zeroes(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
#if defined(CONFIG_FALLOCATE) || defined(CONFIG_XFS)
BDRVRawState *s = aiocb->bs->opaque;
#endif
@ -1525,8 +1551,9 @@ static ssize_t handle_aiocb_write_zeroes(RawPosixAIOData *aiocb)
return -ENOTSUP;
}
static ssize_t handle_aiocb_write_zeroes_unmap(RawPosixAIOData *aiocb)
static int handle_aiocb_write_zeroes_unmap(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
BDRVRawState *s G_GNUC_UNUSED = aiocb->bs->opaque;
int ret;
@ -1568,18 +1595,20 @@ static off_t copy_file_range(int in_fd, off_t *in_off, int out_fd,
}
#endif
static ssize_t handle_aiocb_copy_range(RawPosixAIOData *aiocb)
static int handle_aiocb_copy_range(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
uint64_t bytes = aiocb->aio_nbytes;
off_t in_off = aiocb->aio_offset;
off_t out_off = aiocb->aio_offset2;
off_t out_off = aiocb->copy_range.aio_offset2;
while (bytes) {
ssize_t ret = copy_file_range(aiocb->aio_fildes, &in_off,
aiocb->aio_fd2, &out_off,
aiocb->copy_range.aio_fd2, &out_off,
bytes, 0);
trace_file_copy_file_range(aiocb->bs, aiocb->aio_fildes, in_off,
aiocb->aio_fd2, out_off, bytes, 0, ret);
aiocb->copy_range.aio_fd2, out_off, bytes,
0, ret);
if (ret == 0) {
/* No progress (e.g. when beyond EOF), let the caller fall back to
* buffer I/O. */
@ -1600,8 +1629,9 @@ static ssize_t handle_aiocb_copy_range(RawPosixAIOData *aiocb)
return 0;
}
static ssize_t handle_aiocb_discard(RawPosixAIOData *aiocb)
static int handle_aiocb_discard(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
int ret = -EOPNOTSUPP;
BDRVRawState *s = aiocb->bs->opaque;
@ -1640,15 +1670,17 @@ static ssize_t handle_aiocb_discard(RawPosixAIOData *aiocb)
return ret;
}
static int handle_aiocb_truncate(RawPosixAIOData *aiocb)
static int handle_aiocb_truncate(void *opaque)
{
RawPosixAIOData *aiocb = opaque;
int result = 0;
int64_t current_length = 0;
char *buf = NULL;
struct stat st;
int fd = aiocb->aio_fildes;
int64_t offset = aiocb->aio_offset;
Error **errp = aiocb->errp;
PreallocMode prealloc = aiocb->truncate.prealloc;
Error **errp = aiocb->truncate.errp;
if (fstat(fd, &st) < 0) {
result = -errno;
@ -1657,12 +1689,12 @@ static int handle_aiocb_truncate(RawPosixAIOData *aiocb)
}
current_length = st.st_size;
if (current_length > offset && aiocb->prealloc != PREALLOC_MODE_OFF) {
if (current_length > offset && prealloc != PREALLOC_MODE_OFF) {
error_setg(errp, "Cannot use preallocation for shrinking files");
return -ENOTSUP;
}
switch (aiocb->prealloc) {
switch (prealloc) {
#ifdef CONFIG_POSIX_FALLOCATE
case PREALLOC_MODE_FALLOC:
/*
@ -1743,7 +1775,7 @@ static int handle_aiocb_truncate(RawPosixAIOData *aiocb)
default:
result = -ENOTSUP;
error_setg(errp, "Unsupported preallocation mode: %s",
PreallocMode_str(aiocb->prealloc));
PreallocMode_str(prealloc));
return result;
}
@ -1759,104 +1791,19 @@ out:
return result;
}
static int aio_worker(void *arg)
static int coroutine_fn raw_thread_pool_submit(BlockDriverState *bs,
ThreadPoolFunc func, void *arg)
{
RawPosixAIOData *aiocb = arg;
ssize_t ret = 0;
switch (aiocb->aio_type & QEMU_AIO_TYPE_MASK) {
case QEMU_AIO_READ:
ret = handle_aiocb_rw(aiocb);
if (ret >= 0 && ret < aiocb->aio_nbytes) {
iov_memset(aiocb->aio_iov, aiocb->aio_niov, ret,
0, aiocb->aio_nbytes - ret);
ret = aiocb->aio_nbytes;
}
if (ret == aiocb->aio_nbytes) {
ret = 0;
} else if (ret >= 0 && ret < aiocb->aio_nbytes) {
ret = -EINVAL;
}
break;
case QEMU_AIO_WRITE:
ret = handle_aiocb_rw(aiocb);
if (ret == aiocb->aio_nbytes) {
ret = 0;
} else if (ret >= 0 && ret < aiocb->aio_nbytes) {
ret = -EINVAL;
}
break;
case QEMU_AIO_FLUSH:
ret = handle_aiocb_flush(aiocb);
break;
case QEMU_AIO_IOCTL:
ret = handle_aiocb_ioctl(aiocb);
break;
case QEMU_AIO_DISCARD:
ret = handle_aiocb_discard(aiocb);
break;
case QEMU_AIO_WRITE_ZEROES:
ret = handle_aiocb_write_zeroes(aiocb);
break;
case QEMU_AIO_WRITE_ZEROES | QEMU_AIO_DISCARD:
ret = handle_aiocb_write_zeroes_unmap(aiocb);
break;
case QEMU_AIO_COPY_RANGE:
ret = handle_aiocb_copy_range(aiocb);
break;
case QEMU_AIO_TRUNCATE:
ret = handle_aiocb_truncate(aiocb);
break;
default:
error_report("invalid aio request (0x%x)", aiocb->aio_type);
ret = -EINVAL;
break;
}
g_free(aiocb);
return ret;
}
static int paio_submit_co_full(BlockDriverState *bs, int fd,
int64_t offset, int fd2, int64_t offset2,
QEMUIOVector *qiov,
int bytes, int type)
{
RawPosixAIOData *acb = g_new(RawPosixAIOData, 1);
ThreadPool *pool;
acb->bs = bs;
acb->aio_type = type;
acb->aio_fildes = fd;
acb->aio_fd2 = fd2;
acb->aio_offset2 = offset2;
acb->aio_nbytes = bytes;
acb->aio_offset = offset;
if (qiov) {
acb->aio_iov = qiov->iov;
acb->aio_niov = qiov->niov;
assert(qiov->size == bytes);
}
trace_file_paio_submit_co(offset, bytes, type);
pool = aio_get_thread_pool(bdrv_get_aio_context(bs));
return thread_pool_submit_co(pool, aio_worker, acb);
}
static inline int paio_submit_co(BlockDriverState *bs, int fd,
int64_t offset, QEMUIOVector *qiov,
int bytes, int type)
{
return paio_submit_co_full(bs, fd, offset, -1, 0, qiov, bytes, type);
/* @bs can be NULL, bdrv_get_aio_context() returns the main context then */
ThreadPool *pool = aio_get_thread_pool(bdrv_get_aio_context(bs));
return thread_pool_submit_co(pool, func, arg);
}
static int coroutine_fn raw_co_prw(BlockDriverState *bs, uint64_t offset,
uint64_t bytes, QEMUIOVector *qiov, int type)
{
BDRVRawState *s = bs->opaque;
RawPosixAIOData acb;
if (fd_open(bs) < 0)
return -EIO;
@ -1879,7 +1826,20 @@ static int coroutine_fn raw_co_prw(BlockDriverState *bs, uint64_t offset,
}
}
return paio_submit_co(bs, s->fd, offset, qiov, bytes, type);
acb = (RawPosixAIOData) {
.bs = bs,
.aio_fildes = s->fd,
.aio_type = type,
.aio_offset = offset,
.aio_nbytes = bytes,
.io = {
.iov = qiov->iov,
.niov = qiov->niov,
},
};
assert(qiov->size == bytes);
return raw_thread_pool_submit(bs, handle_aiocb_rw, &acb);
}
static int coroutine_fn raw_co_preadv(BlockDriverState *bs, uint64_t offset,
@ -1922,6 +1882,7 @@ static void raw_aio_unplug(BlockDriverState *bs)
static int raw_co_flush_to_disk(BlockDriverState *bs)
{
BDRVRawState *s = bs->opaque;
RawPosixAIOData acb;
int ret;
ret = fd_open(bs);
@ -1929,7 +1890,13 @@ static int raw_co_flush_to_disk(BlockDriverState *bs)
return ret;
}
return paio_submit_co(bs, s->fd, 0, NULL, 0, QEMU_AIO_FLUSH);
acb = (RawPosixAIOData) {
.bs = bs,
.aio_fildes = s->fd,
.aio_type = QEMU_AIO_FLUSH,
};
return raw_thread_pool_submit(bs, handle_aiocb_flush, &acb);
}
static void raw_aio_attach_aio_context(BlockDriverState *bs,
@ -1968,21 +1935,20 @@ static int coroutine_fn
raw_regular_truncate(BlockDriverState *bs, int fd, int64_t offset,
PreallocMode prealloc, Error **errp)
{
RawPosixAIOData *acb = g_new(RawPosixAIOData, 1);
ThreadPool *pool;
RawPosixAIOData acb;
*acb = (RawPosixAIOData) {
acb = (RawPosixAIOData) {
.bs = bs,
.aio_fildes = fd,
.aio_type = QEMU_AIO_TRUNCATE,
.aio_offset = offset,
.prealloc = prealloc,
.errp = errp,
.truncate = {
.prealloc = prealloc,
.errp = errp,
},
};
/* @bs can be NULL, bdrv_get_aio_context() returns the main context then */
pool = aio_get_thread_pool(bdrv_get_aio_context(bs));
return thread_pool_submit_co(pool, aio_worker, acb);
return raw_thread_pool_submit(bs, handle_aiocb_truncate, &acb);
}
static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,
@ -2613,25 +2579,67 @@ static void coroutine_fn raw_co_invalidate_cache(BlockDriverState *bs,
}
static coroutine_fn int
raw_co_pdiscard(BlockDriverState *bs, int64_t offset, int bytes)
raw_do_pdiscard(BlockDriverState *bs, int64_t offset, int bytes, bool blkdev)
{
BDRVRawState *s = bs->opaque;
RawPosixAIOData acb;
return paio_submit_co(bs, s->fd, offset, NULL, bytes, QEMU_AIO_DISCARD);
acb = (RawPosixAIOData) {
.bs = bs,
.aio_fildes = s->fd,
.aio_type = QEMU_AIO_DISCARD,
.aio_offset = offset,
.aio_nbytes = bytes,
};
if (blkdev) {
acb.aio_type |= QEMU_AIO_BLKDEV;
}
return raw_thread_pool_submit(bs, handle_aiocb_discard, &acb);
}
static coroutine_fn int
raw_co_pdiscard(BlockDriverState *bs, int64_t offset, int bytes)
{
return raw_do_pdiscard(bs, offset, bytes, false);
}
static int coroutine_fn
raw_do_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int bytes,
BdrvRequestFlags flags, bool blkdev)
{
BDRVRawState *s = bs->opaque;
RawPosixAIOData acb;
ThreadPoolFunc *handler;
acb = (RawPosixAIOData) {
.bs = bs,
.aio_fildes = s->fd,
.aio_type = QEMU_AIO_WRITE_ZEROES,
.aio_offset = offset,
.aio_nbytes = bytes,
};
if (blkdev) {
acb.aio_type |= QEMU_AIO_BLKDEV;
}
if (flags & BDRV_REQ_MAY_UNMAP) {
acb.aio_type |= QEMU_AIO_DISCARD;
handler = handle_aiocb_write_zeroes_unmap;
} else {
handler = handle_aiocb_write_zeroes;
}
return raw_thread_pool_submit(bs, handler, &acb);
}
static int coroutine_fn raw_co_pwrite_zeroes(
BlockDriverState *bs, int64_t offset,
int bytes, BdrvRequestFlags flags)
{
BDRVRawState *s = bs->opaque;
int operation = QEMU_AIO_WRITE_ZEROES;
if (flags & BDRV_REQ_MAY_UNMAP) {
operation |= QEMU_AIO_DISCARD;
}
return paio_submit_co(bs, s->fd, offset, NULL, bytes, operation);
return raw_do_pwrite_zeroes(bs, offset, bytes, flags, false);
}
static int raw_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
@ -2702,6 +2710,7 @@ static int coroutine_fn raw_co_copy_range_to(BlockDriverState *bs,
BdrvRequestFlags read_flags,
BdrvRequestFlags write_flags)
{
RawPosixAIOData acb;
BDRVRawState *s = bs->opaque;
BDRVRawState *src_s;
@ -2714,8 +2723,20 @@ static int coroutine_fn raw_co_copy_range_to(BlockDriverState *bs,
if (fd_open(src->bs) < 0 || fd_open(dst->bs) < 0) {
return -EIO;
}
return paio_submit_co_full(bs, src_s->fd, src_offset, s->fd, dst_offset,
NULL, bytes, QEMU_AIO_COPY_RANGE);
acb = (RawPosixAIOData) {
.bs = bs,
.aio_type = QEMU_AIO_COPY_RANGE,
.aio_fildes = src_s->fd,
.aio_offset = src_offset,
.aio_nbytes = bytes,
.copy_range = {
.aio_fd2 = s->fd,
.aio_offset2 = dst_offset,
},
};
return raw_thread_pool_submit(bs, handle_aiocb_copy_range, &acb);
}
BlockDriver bdrv_file = {
@ -3063,36 +3084,39 @@ hdev_open_Mac_error:
}
#if defined(__linux__)
static BlockAIOCB *hdev_aio_ioctl(BlockDriverState *bs,
unsigned long int req, void *buf,
BlockCompletionFunc *cb, void *opaque)
static int coroutine_fn
hdev_co_ioctl(BlockDriverState *bs, unsigned long int req, void *buf)
{
BDRVRawState *s = bs->opaque;
RawPosixAIOData *acb;
ThreadPool *pool;
RawPosixAIOData acb;
int ret;
if (fd_open(bs) < 0)
return NULL;
ret = fd_open(bs);
if (ret < 0) {
return ret;
}
if (req == SG_IO && s->pr_mgr) {
struct sg_io_hdr *io_hdr = buf;
if (io_hdr->cmdp[0] == PERSISTENT_RESERVE_OUT ||
io_hdr->cmdp[0] == PERSISTENT_RESERVE_IN) {
return pr_manager_execute(s->pr_mgr, bdrv_get_aio_context(bs),
s->fd, io_hdr, cb, opaque);
s->fd, io_hdr);
}
}
acb = g_new(RawPosixAIOData, 1);
acb->bs = bs;
acb->aio_type = QEMU_AIO_IOCTL;
acb->aio_fildes = s->fd;
acb->aio_offset = 0;
acb->aio_ioctl_buf = buf;
acb->aio_ioctl_cmd = req;
pool = aio_get_thread_pool(bdrv_get_aio_context(bs));
return thread_pool_submit_aio(pool, aio_worker, acb, cb, opaque);
acb = (RawPosixAIOData) {
.bs = bs,
.aio_type = QEMU_AIO_IOCTL,
.aio_fildes = s->fd,
.aio_offset = 0,
.ioctl = {
.buf = buf,
.cmd = req,
},
};
return raw_thread_pool_submit(bs, handle_aiocb_ioctl, &acb);
}
#endif /* linux */
@ -3109,22 +3133,18 @@ static int fd_open(BlockDriverState *bs)
static coroutine_fn int
hdev_co_pdiscard(BlockDriverState *bs, int64_t offset, int bytes)
{
BDRVRawState *s = bs->opaque;
int ret;
ret = fd_open(bs);
if (ret < 0) {
return ret;
}
return paio_submit_co(bs, s->fd, offset, NULL, bytes,
QEMU_AIO_DISCARD | QEMU_AIO_BLKDEV);
return raw_do_pdiscard(bs, offset, bytes, true);
}
static coroutine_fn int hdev_co_pwrite_zeroes(BlockDriverState *bs,
int64_t offset, int bytes, BdrvRequestFlags flags)
{
BDRVRawState *s = bs->opaque;
int operation = QEMU_AIO_WRITE_ZEROES | QEMU_AIO_BLKDEV;
int rc;
rc = fd_open(bs);
@ -3132,11 +3152,7 @@ static coroutine_fn int hdev_co_pwrite_zeroes(BlockDriverState *bs,
return rc;
}
if (flags & BDRV_REQ_MAY_UNMAP) {
operation |= QEMU_AIO_DISCARD;
}
return paio_submit_co(bs, s->fd, offset, NULL, bytes, operation);
return raw_do_pwrite_zeroes(bs, offset, bytes, flags, true);
}
static int coroutine_fn hdev_co_create_opts(const char *filename, QemuOpts *opts,
@ -3241,7 +3257,7 @@ static BlockDriver bdrv_host_device = {
/* generic scsi device */
#ifdef __linux__
.bdrv_aio_ioctl = hdev_aio_ioctl,
.bdrv_co_ioctl = hdev_co_ioctl,
#endif
};
@ -3363,7 +3379,7 @@ static BlockDriver bdrv_host_cdrom = {
.bdrv_lock_medium = cdrom_lock_medium,
/* generic scsi device */
.bdrv_aio_ioctl = hdev_aio_ioctl,
.bdrv_co_ioctl = hdev_co_ioctl,
};
#endif /* __linux__ */

38
block/mirror.c
View File

@ -277,7 +277,8 @@ static int mirror_cow_align(MirrorBlockJob *s, int64_t *offset,
return ret;
}
static inline void mirror_wait_for_any_operation(MirrorBlockJob *s, bool active)
static inline void coroutine_fn
mirror_wait_for_any_operation(MirrorBlockJob *s, bool active)
{
MirrorOp *op;
@ -295,7 +296,8 @@ static inline void mirror_wait_for_any_operation(MirrorBlockJob *s, bool active)
abort();
}
static inline void mirror_wait_for_free_in_flight_slot(MirrorBlockJob *s)
static inline void coroutine_fn
mirror_wait_for_free_in_flight_slot(MirrorBlockJob *s)
{
/* Only non-active operations use up in-flight slots */
mirror_wait_for_any_operation(s, false);
@ -598,7 +600,7 @@ static void mirror_free_init(MirrorBlockJob *s)
* mirror_resume() because mirror_run() will begin iterating again
* when the job is resumed.
*/
static void mirror_wait_for_all_io(MirrorBlockJob *s)
static void coroutine_fn mirror_wait_for_all_io(MirrorBlockJob *s)
{
while (s->in_flight > 0) {
mirror_wait_for_free_in_flight_slot(s);
@ -669,9 +671,10 @@ static int mirror_exit_common(Job *job)
if (s->should_complete && !abort) {
BlockDriverState *to_replace = s->to_replace ?: src;
bool ro = bdrv_is_read_only(to_replace);
if (bdrv_get_flags(target_bs) != bdrv_get_flags(to_replace)) {
bdrv_reopen(target_bs, bdrv_get_flags(to_replace), NULL);
if (ro != bdrv_is_read_only(target_bs)) {
bdrv_reopen_set_read_only(target_bs, ro, NULL);
}
/* The mirror job has no requests in flight any more, but we need to
@ -731,7 +734,7 @@ static void mirror_abort(Job *job)
assert(ret == 0);
}
static void mirror_throttle(MirrorBlockJob *s)
static void coroutine_fn mirror_throttle(MirrorBlockJob *s)
{
int64_t now = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
@ -1106,7 +1109,7 @@ static void mirror_complete(Job *job, Error **errp)
job_enter(job);
}
static void mirror_pause(Job *job)
static void coroutine_fn mirror_pause(Job *job)
{
MirrorBlockJob *s = container_of(job, MirrorBlockJob, common.job);
@ -1177,9 +1180,10 @@ static const BlockJobDriver commit_active_job_driver = {
.drain = mirror_drain,
};
static void do_sync_target_write(MirrorBlockJob *job, MirrorMethod method,
uint64_t offset, uint64_t bytes,
QEMUIOVector *qiov, int flags)
static void coroutine_fn
do_sync_target_write(MirrorBlockJob *job, MirrorMethod method,
uint64_t offset, uint64_t bytes,
QEMUIOVector *qiov, int flags)
{
BdrvDirtyBitmapIter *iter;
QEMUIOVector target_qiov;
@ -1689,13 +1693,15 @@ void commit_active_start(const char *job_id, BlockDriverState *bs,
BlockCompletionFunc *cb, void *opaque,
bool auto_complete, Error **errp)
{
int orig_base_flags;
bool base_read_only;
Error *local_err = NULL;
orig_base_flags = bdrv_get_flags(base);
base_read_only = bdrv_is_read_only(base);
if (bdrv_reopen(base, bs->open_flags, errp)) {
return;
if (base_read_only) {
if (bdrv_reopen_set_read_only(base, false, errp) < 0) {
return;
}
}
mirror_start_job(job_id, bs, creation_flags, base, NULL, speed, 0, 0,
@ -1714,6 +1720,8 @@ void commit_active_start(const char *job_id, BlockDriverState *bs,
error_restore_flags:
/* ignore error and errp for bdrv_reopen, because we want to propagate
* the original error */
bdrv_reopen(base, orig_base_flags, NULL);
if (base_read_only) {
bdrv_reopen_set_read_only(base, true, NULL);
}
return;
}

70
block/qcow2-cluster.c
View File

@ -1571,76 +1571,6 @@ again:
return 0;
}
static int decompress_buffer(uint8_t *out_buf, int out_buf_size,
const uint8_t *buf, int buf_size)
{
z_stream strm1, *strm = &strm1;
int ret, out_len;
memset(strm, 0, sizeof(*strm));
strm->next_in = (uint8_t *)buf;
strm->avail_in = buf_size;
strm->next_out = out_buf;
strm->avail_out = out_buf_size;
ret = inflateInit2(strm, -12);
if (ret != Z_OK)
return -1;
ret = inflate(strm, Z_FINISH);
out_len = strm->next_out - out_buf;
if ((ret != Z_STREAM_END && ret != Z_BUF_ERROR) ||
out_len != out_buf_size) {
inflateEnd(strm);
return -1;
}
inflateEnd(strm);
return 0;
}
int qcow2_decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset)
{
BDRVQcow2State *s = bs->opaque;
int ret, csize, nb_csectors, sector_offset;
uint64_t coffset;
coffset = cluster_offset & s->cluster_offset_mask;
if (s->cluster_cache_offset != coffset) {
nb_csectors = ((cluster_offset >> s->csize_shift) & s->csize_mask) + 1;
sector_offset = coffset & 511;
csize = nb_csectors * 512 - sector_offset;
/* Allocate buffers on first decompress operation, most images are
* uncompressed and the memory overhead can be avoided. The buffers
* are freed in .bdrv_close().
*/
if (!s->cluster_data) {
/* one more sector for decompressed data alignment */
s->cluster_data = qemu_try_blockalign(bs->file->bs,
QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size + 512);
if (!s->cluster_data) {
return -ENOMEM;
}
}
if (!s->cluster_cache) {
s->cluster_cache = g_malloc(s->cluster_size);
}
BLKDBG_EVENT(bs->file, BLKDBG_READ_COMPRESSED);
ret = bdrv_read(bs->file, coffset >> 9, s->cluster_data,
nb_csectors);
if (ret < 0) {
return ret;
}
if (decompress_buffer(s->cluster_cache, s->cluster_size,
s->cluster_data + sector_offset, csize) < 0) {
return -EIO;
}
s->cluster_cache_offset = coffset;
}
return 0;
}
/*
* This discards as many clusters of nb_clusters as possible at once (i.e.
* all clusters in the same L2 slice) and returns the number of discarded

170
block/qcow2.c
View File

@ -74,6 +74,13 @@ typedef struct {
#define QCOW2_EXT_MAGIC_CRYPTO_HEADER 0x0537be77
#define QCOW2_EXT_MAGIC_BITMAPS 0x23852875
static int coroutine_fn
qcow2_co_preadv_compressed(BlockDriverState *bs,
uint64_t file_cluster_offset,
uint64_t offset,
uint64_t bytes,
QEMUIOVector *qiov);
static int qcow2_probe(const uint8_t *buf, int buf_size, const char *filename)
{
const QCowHeader *cow_header = (const void *)buf;
@ -1414,7 +1421,6 @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
goto fail;
}
s->cluster_cache_offset = -1;
s->flags = flags;
ret = qcow2_refcount_init(bs);
@ -1914,15 +1920,15 @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset,
break;
case QCOW2_CLUSTER_COMPRESSED:
/* add AIO support for compressed blocks ? */
ret = qcow2_decompress_cluster(bs, cluster_offset);
qemu_co_mutex_unlock(&s->lock);
ret = qcow2_co_preadv_compressed(bs, cluster_offset,
offset, cur_bytes,
&hd_qiov);
qemu_co_mutex_lock(&s->lock);
if (ret < 0) {
goto fail;
}
qemu_iovec_from_buf(&hd_qiov, 0,
s->cluster_cache + offset_in_cluster,
cur_bytes);
break;
case QCOW2_CLUSTER_NORMAL:
@ -2058,8 +2064,6 @@ static coroutine_fn int qcow2_co_pwritev(BlockDriverState *bs, uint64_t offset,
qemu_iovec_init(&hd_qiov, qiov->niov);
s->cluster_cache_offset = -1; /* disable compressed cache */
qemu_co_mutex_lock(&s->lock);
while (bytes != 0) {
@ -2223,8 +2227,6 @@ static void qcow2_close(BlockDriverState *bs)
g_free(s->image_backing_file);
g_free(s->image_backing_format);
g_free(s->cluster_cache);
qemu_vfree(s->cluster_data);
qcow2_refcount_close(bs);
qcow2_free_snapshots(bs);
}
@ -3401,7 +3403,6 @@ qcow2_co_copy_range_to(BlockDriverState *bs,
QCowL2Meta *l2meta = NULL;
assert(!bs->encrypted);
s->cluster_cache_offset = -1; /* disable compressed cache */
qemu_co_mutex_lock(&s->lock);
@ -3722,14 +3723,15 @@ fail:
/*
* qcow2_compress()
*
* @dest - destination buffer, at least of @size-1 bytes
* @src - source buffer, @size bytes
* @dest - destination buffer, @dest_size bytes
* @src - source buffer, @src_size bytes
*
* Returns: compressed size on success
* -1 if compression is inefficient
* -1 destination buffer is not enough to store compressed data
* -2 on any other error
*/
static ssize_t qcow2_compress(void *dest, const void *src, size_t size)
static ssize_t qcow2_compress(void *dest, size_t dest_size,
const void *src, size_t src_size)
{
ssize_t ret;
z_stream strm;
@ -3738,20 +3740,20 @@ static ssize_t qcow2_compress(void *dest, const void *src, size_t size)
memset(&strm, 0, sizeof(strm));
ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION, Z_DEFLATED,
-12, 9, Z_DEFAULT_STRATEGY);
if (ret != 0) {
if (ret != Z_OK) {
return -2;
}
/* strm.next_in is not const in old zlib versions, such as those used on
* OpenBSD/NetBSD, so cast the const away */
strm.avail_in = size;
strm.avail_in = src_size;
strm.next_in = (void *) src;
strm.avail_out = size - 1;
strm.avail_out = dest_size;
strm.next_out = dest;
ret = deflate(&strm, Z_FINISH);
if (ret == Z_STREAM_END) {
ret = size - 1 - strm.avail_out;
ret = dest_size - strm.avail_out;
} else {
ret = (ret == Z_OK ? -1 : -2);
}
@ -3761,20 +3763,68 @@ static ssize_t qcow2_compress(void *dest, const void *src, size_t size)
return ret;
}
/*
* qcow2_decompress()
*
* Decompress some data (not more than @src_size bytes) to produce exactly
* @dest_size bytes.
*
* @dest - destination buffer, @dest_size bytes
* @src - source buffer, @src_size bytes
*
* Returns: 0 on success
* -1 on fail
*/
static ssize_t qcow2_decompress(void *dest, size_t dest_size,
const void *src, size_t src_size)
{
int ret = 0;
z_stream strm;
memset(&strm, 0, sizeof(strm));
strm.avail_in = src_size;
strm.next_in = (void *) src;
strm.avail_out = dest_size;
strm.next_out = dest;
ret = inflateInit2(&strm, -12);
if (ret != Z_OK) {
return -1;
}
ret = inflate(&strm, Z_FINISH);
if ((ret != Z_STREAM_END && ret != Z_BUF_ERROR) || strm.avail_out != 0) {
/* We approve Z_BUF_ERROR because we need @dest buffer to be filled, but
* @src buffer may be processed partly (because in qcow2 we know size of
* compressed data with precision of one sector) */
ret = -1;
}
inflateEnd(&strm);
return ret;
}
#define MAX_COMPRESS_THREADS 4
typedef ssize_t (*Qcow2CompressFunc)(void *dest, size_t dest_size,
const void *src, size_t src_size);
typedef struct Qcow2CompressData {
void *dest;
size_t dest_size;
const void *src;
size_t size;
size_t src_size;
ssize_t ret;
Qcow2CompressFunc func;
} Qcow2CompressData;
static int qcow2_compress_pool_func(void *opaque)
{
Qcow2CompressData *data = opaque;
data->ret = qcow2_compress(data->dest, data->src, data->size);
data->ret = data->func(data->dest, data->dest_size,
data->src, data->src_size);
return 0;
}
@ -3784,17 +3834,19 @@ static void qcow2_compress_complete(void *opaque, int ret)
qemu_coroutine_enter(opaque);
}
/* See qcow2_compress definition for parameters description */
static ssize_t qcow2_co_compress(BlockDriverState *bs,
void *dest, const void *src, size_t size)
static ssize_t coroutine_fn
qcow2_co_do_compress(BlockDriverState *bs, void *dest, size_t dest_size,
const void *src, size_t src_size, Qcow2CompressFunc func)
{
BDRVQcow2State *s = bs->opaque;
BlockAIOCB *acb;
ThreadPool *pool = aio_get_thread_pool(bdrv_get_aio_context(bs));
Qcow2CompressData arg = {
.dest = dest,
.dest_size = dest_size,
.src = src,
.size = size,
.src_size = src_size,
.func = func,
};
while (s->nb_compress_threads >= MAX_COMPRESS_THREADS) {
@ -3817,6 +3869,22 @@ static ssize_t qcow2_co_compress(BlockDriverState *bs,
return arg.ret;
}
static ssize_t coroutine_fn
qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,
const void *src, size_t src_size)
{
return qcow2_co_do_compress(bs, dest, dest_size, src, src_size,
qcow2_compress);
}
static ssize_t coroutine_fn
qcow2_co_decompress(BlockDriverState *bs, void *dest, size_t dest_size,
const void *src, size_t src_size)
{
return qcow2_co_do_compress(bs, dest, dest_size, src, src_size,
qcow2_decompress);
}
/* XXX: put compressed sectors first, then all the cluster aligned
tables to avoid losing bytes in alignment */
static coroutine_fn int
@ -3861,7 +3929,8 @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset,
out_buf = g_malloc(s->cluster_size);
out_len = qcow2_co_compress(bs, out_buf, buf, s->cluster_size);
out_len = qcow2_co_compress(bs, out_buf, s->cluster_size - 1,
buf, s->cluster_size);
if (out_len == -2) {
ret = -EINVAL;
goto fail;
@ -3909,6 +3978,55 @@ fail:
return ret;
}
static int coroutine_fn
qcow2_co_preadv_compressed(BlockDriverState *bs,
uint64_t file_cluster_offset,
uint64_t offset,
uint64_t bytes,
QEMUIOVector *qiov)
{
BDRVQcow2State *s = bs->opaque;
int ret = 0, csize, nb_csectors;
uint64_t coffset;
uint8_t *buf, *out_buf;
struct iovec iov;
QEMUIOVector local_qiov;
int offset_in_cluster = offset_into_cluster(s, offset);
coffset = file_cluster_offset & s->cluster_offset_mask;
nb_csectors = ((file_cluster_offset >> s->csize_shift) & s->csize_mask) + 1;
csize = nb_csectors * 512 - (coffset & 511);
buf = g_try_malloc(csize);
if (!buf) {
return -ENOMEM;
}
iov.iov_base = buf;
iov.iov_len = csize;
qemu_iovec_init_external(&local_qiov, &iov, 1);
out_buf = qemu_blockalign(bs, s->cluster_size);
BLKDBG_EVENT(bs->file, BLKDBG_READ_COMPRESSED);
ret = bdrv_co_preadv(bs->file, coffset, csize, &local_qiov, 0);
if (ret < 0) {
goto fail;
}
if (qcow2_co_decompress(bs, out_buf, s->cluster_size, buf, csize) < 0) {
ret = -EIO;
goto fail;
}
qemu_iovec_from_buf(qiov, 0, out_buf + offset_in_cluster, bytes);
fail:
qemu_vfree(out_buf);
g_free(buf);
return ret;
}
static int make_completely_empty(BlockDriverState *bs)
{
BDRVQcow2State *s = bs->opaque;

4
block/qcow2.h
View File

@ -278,9 +278,6 @@ typedef struct BDRVQcow2State {
QEMUTimer *cache_clean_timer;
unsigned cache_clean_interval;
uint8_t *cluster_cache;
uint8_t *cluster_data;
uint64_t cluster_cache_offset;
QLIST_HEAD(QCowClusterAlloc, QCowL2Meta) cluster_allocs;
uint64_t *refcount_table;
@ -616,7 +613,6 @@ int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size,
bool exact_size);
int qcow2_shrink_l1_table(BlockDriverState *bs, uint64_t max_size);
int qcow2_write_l1_entry(BlockDriverState *bs, int l1_index);
int qcow2_decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset);
int qcow2_encrypt_sectors(BDRVQcow2State *s, int64_t sector_num,
uint8_t *buf, int nb_sectors, bool enc, Error **errp);

67
block/replication.c
View File

@ -20,6 +20,7 @@
#include "block/block_backup.h"
#include "sysemu/block-backend.h"
#include "qapi/error.h"
#include "qapi/qmp/qdict.h"
#include "replication.h"
typedef enum {
@ -39,8 +40,8 @@ typedef struct BDRVReplicationState {
char *top_id;
ReplicationState *rs;
Error *blocker;
int orig_hidden_flags;
int orig_secondary_flags;
bool orig_hidden_read_only;
bool orig_secondary_read_only;
int error;
} BDRVReplicationState;
@ -218,9 +219,6 @@ static coroutine_fn int replication_co_readv(BlockDriverState *bs,
QEMUIOVector *qiov)
{
BDRVReplicationState *s = bs->opaque;
BdrvChild *child = s->secondary_disk;
BlockJob *job = NULL;
CowRequest req;
int ret;
if (s->mode == REPLICATION_MODE_PRIMARY) {
@ -233,28 +231,9 @@ static coroutine_fn int replication_co_readv(BlockDriverState *bs,
return ret;
}
if (child && child->bs) {
job = child->bs->job;
}
if (job) {
uint64_t remaining_bytes = remaining_sectors * BDRV_SECTOR_SIZE;
backup_wait_for_overlapping_requests(child->bs->job,
sector_num * BDRV_SECTOR_SIZE,
remaining_bytes);
backup_cow_request_begin(&req, child->bs->job,
sector_num * BDRV_SECTOR_SIZE,
remaining_bytes);
ret = bdrv_co_preadv(bs->file, sector_num * BDRV_SECTOR_SIZE,
remaining_bytes, qiov, 0);
backup_cow_request_end(&req);
goto out;
}
ret = bdrv_co_preadv(bs->file, sector_num * BDRV_SECTOR_SIZE,
remaining_sectors * BDRV_SECTOR_SIZE, qiov, 0);
out:
return replication_return_value(s, ret);
}
@ -371,44 +350,38 @@ static void secondary_do_checkpoint(BDRVReplicationState *s, Error **errp)
}
}
/* This function is supposed to be called twice:
* first with writable = true, then with writable = false.
* The first call puts s->hidden_disk and s->secondary_disk in
* r/w mode, and the second puts them back in their original state.
*/
static void reopen_backing_file(BlockDriverState *bs, bool writable,
Error **errp)
{
BDRVReplicationState *s = bs->opaque;
BlockReopenQueue *reopen_queue = NULL;
int orig_hidden_flags, orig_secondary_flags;
int new_hidden_flags, new_secondary_flags;
Error *local_err = NULL;
if (writable) {
orig_hidden_flags = s->orig_hidden_flags =
bdrv_get_flags(s->hidden_disk->bs);
new_hidden_flags = (orig_hidden_flags | BDRV_O_RDWR) &
~BDRV_O_INACTIVE;
orig_secondary_flags = s->orig_secondary_flags =
bdrv_get_flags(s->secondary_disk->bs);
new_secondary_flags = (orig_secondary_flags | BDRV_O_RDWR) &
~BDRV_O_INACTIVE;
} else {
orig_hidden_flags = (s->orig_hidden_flags | BDRV_O_RDWR) &
~BDRV_O_INACTIVE;
new_hidden_flags = s->orig_hidden_flags;
orig_secondary_flags = (s->orig_secondary_flags | BDRV_O_RDWR) &
~BDRV_O_INACTIVE;
new_secondary_flags = s->orig_secondary_flags;
s->orig_hidden_read_only = bdrv_is_read_only(s->hidden_disk->bs);
s->orig_secondary_read_only = bdrv_is_read_only(s->secondary_disk->bs);
}
bdrv_subtree_drained_begin(s->hidden_disk->bs);
bdrv_subtree_drained_begin(s->secondary_disk->bs);
if (orig_hidden_flags != new_hidden_flags) {
reopen_queue = bdrv_reopen_queue(reopen_queue, s->hidden_disk->bs, NULL,
new_hidden_flags);
if (s->orig_hidden_read_only) {
QDict *opts = qdict_new();
qdict_put_bool(opts, BDRV_OPT_READ_ONLY, !writable);
reopen_queue = bdrv_reopen_queue(reopen_queue, s->hidden_disk->bs,
opts);
}
if (!(orig_secondary_flags & BDRV_O_RDWR)) {
if (s->orig_secondary_read_only) {
QDict *opts = qdict_new();
qdict_put_bool(opts, BDRV_OPT_READ_ONLY, !writable);
reopen_queue = bdrv_reopen_queue(reopen_queue, s->secondary_disk->bs,
NULL, new_secondary_flags);
opts);
}
if (reopen_queue) {

20
block/stream.c
View File

@ -34,7 +34,7 @@ typedef struct StreamBlockJob {
BlockDriverState *base;
BlockdevOnError on_error;
char *backing_file_str;
int bs_flags;
bool bs_read_only;
} StreamBlockJob;
static int coroutine_fn stream_populate(BlockBackend *blk,
@ -89,10 +89,10 @@ static void stream_clean(Job *job)
BlockDriverState *bs = blk_bs(bjob->blk);
/* Reopen the image back in read-only mode if necessary */
if (s->bs_flags != bdrv_get_flags(bs)) {
if (s->bs_read_only) {
/* Give up write permissions before making it read-only */
blk_set_perm(bjob->blk, 0, BLK_PERM_ALL, &error_abort);
bdrv_reopen(bs, s->bs_flags, NULL);
bdrv_reopen_set_read_only(bs, true, NULL);
}
g_free(s->backing_file_str);
@ -226,12 +226,12 @@ void stream_start(const char *job_id, BlockDriverState *bs,
{
StreamBlockJob *s;
BlockDriverState *iter;
int orig_bs_flags;
bool bs_read_only;
/* Make sure that the image is opened in read-write mode */
orig_bs_flags = bdrv_get_flags(bs);
if (!(orig_bs_flags & BDRV_O_RDWR)) {
if (bdrv_reopen(bs, orig_bs_flags | BDRV_O_RDWR, errp) != 0) {
bs_read_only = bdrv_is_read_only(bs);
if (bs_read_only) {
if (bdrv_reopen_set_read_only(bs, false, errp) != 0) {
return;
}
}
@ -261,7 +261,7 @@ void stream_start(const char *job_id, BlockDriverState *bs,
s->base = base;
s->backing_file_str = g_strdup(backing_file_str);
s->bs_flags = orig_bs_flags;
s->bs_read_only = bs_read_only;
s->on_error = on_error;
trace_stream_start(bs, base, s);
@ -269,7 +269,7 @@ void stream_start(const char *job_id, BlockDriverState *bs,
return;
fail:
if (orig_bs_flags != bdrv_get_flags(bs)) {
bdrv_reopen(bs, orig_bs_flags, NULL);
if (bs_read_only) {
bdrv_reopen_set_read_only(bs, true, NULL);
}
}

11
blockdev.c
View File

@ -1703,8 +1703,7 @@ static void external_snapshot_commit(BlkActionState *common)
* bdrv_reopen_multiple() across all the entries at once, because we
* don't want to abort all of them if one of them fails the reopen */
if (!atomic_read(&state->old_bs->copy_on_read)) {
bdrv_reopen(state->old_bs, state->old_bs->open_flags & ~BDRV_O_RDWR,
NULL);
bdrv_reopen_set_read_only(state->old_bs, true, NULL);
}
aio_context_release(aio_context);
@ -4100,7 +4099,6 @@ void qmp_change_backing_file(const char *device,
BlockDriverState *image_bs = NULL;
Error *local_err = NULL;
bool ro;
int open_flags;
int ret;
bs = qmp_get_root_bs(device, errp);
@ -4142,13 +4140,10 @@ void qmp_change_backing_file(const char *device,
}
/* if not r/w, reopen to make r/w */
open_flags = image_bs->open_flags;
ro = bdrv_is_read_only(image_bs);
if (ro) {
bdrv_reopen(image_bs, open_flags | BDRV_O_RDWR, &local_err);
if (local_err) {
error_propagate(errp, local_err);
if (bdrv_reopen_set_read_only(image_bs, false, errp) != 0) {
goto out;
}
}
@ -4164,7 +4159,7 @@ void qmp_change_backing_file(const char *device,
}
if (ro) {
bdrv_reopen(image_bs, open_flags, &local_err);
bdrv_reopen_set_read_only(image_bs, true, &local_err);
error_propagate(errp, local_err);
}

31
configure vendored
View File

@ -434,6 +434,7 @@ capstone=""
lzo=""
snappy=""
bzip2=""
lzfse=""
guest_agent=""
guest_agent_with_vss="no"
guest_agent_ntddscsi="no"
@ -1310,6 +1311,10 @@ for opt do
;;
--enable-bzip2) bzip2="yes"
;;
--enable-lzfse) lzfse="yes"
;;
--disable-lzfse) lzfse="no"
;;
--enable-guest-agent) guest_agent="yes"
;;
--disable-guest-agent) guest_agent="no"
@ -1745,6 +1750,8 @@ disabled with --disable-FEATURE, default is enabled if available:
snappy support of snappy compression library
bzip2 support of bzip2 compression library
(for reading bzip2-compressed dmg images)
lzfse support of lzfse compression library
(for reading lzfse-compressed dmg images)
seccomp seccomp support
coroutine-pool coroutine freelist (better performance)
glusterfs GlusterFS backend
@ -2274,6 +2281,24 @@ EOF
fi
fi
##########################################
# lzfse check
if test "$lzfse" != "no" ; then
cat > $TMPC << EOF
#include <lzfse.h>
int main(void) { lzfse_decode_scratch_size(); return 0; }
EOF
if compile_prog "" "-llzfse" ; then
lzfse="yes"
else
if test "$lzfse" = "yes"; then
feature_not_found "lzfse" "Install lzfse devel"
fi
lzfse="no"
fi
fi
##########################################
# libseccomp check
@ -6096,6 +6121,7 @@ echo "Live block migration $live_block_migration"
echo "lzo support $lzo"
echo "snappy support $snappy"
echo "bzip2 support $bzip2"
echo "lzfse support $lzfse"
echo "NUMA host support $numa"
echo "libxml2 $libxml2"
echo "tcmalloc support $tcmalloc"
@ -6618,6 +6644,11 @@ if test "$bzip2" = "yes" ; then
echo "BZIP2_LIBS=-lbz2" >> $config_host_mak
fi
if test "$lzfse" = "yes" ; then
echo "CONFIG_LZFSE=y" >> $config_host_mak
echo "LZFSE_LIBS=-llzfse" >> $config_host_mak
fi
if test "$libiscsi" = "yes" ; then
echo "CONFIG_LIBISCSI=m" >> $config_host_mak
echo "LIBISCSI_CFLAGS=$libiscsi_cflags" >> $config_host_mak

6
include/block/block.h
View File

@ -299,10 +299,10 @@ BlockDriverState *bdrv_open(const char *filename, const char *reference,
BlockDriverState *bdrv_new_open_driver(BlockDriver *drv, const char *node_name,
int flags, Error **errp);
BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue,
BlockDriverState *bs,
QDict *options, int flags);
BlockDriverState *bs, QDict *options);
int bdrv_reopen_multiple(AioContext *ctx, BlockReopenQueue *bs_queue, Error **errp);
int bdrv_reopen(BlockDriverState *bs, int bdrv_flags, Error **errp);
int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only,
Error **errp);
int bdrv_reopen_prepare(BDRVReopenState *reopen_state,
BlockReopenQueue *queue, Error **errp);
void bdrv_reopen_commit(BDRVReopenState *reopen_state);

13
include/block/block_backup.h
View File

@ -20,19 +20,6 @@
#include "block/block_int.h"
typedef struct CowRequest {
int64_t start_byte;
int64_t end_byte;
QLIST_ENTRY(CowRequest) list;
CoQueue wait_queue; /* coroutines blocked on this request */
} CowRequest;
void backup_wait_for_overlapping_requests(BlockJob *job, int64_t offset,
uint64_t bytes);
void backup_cow_request_begin(CowRequest *req, BlockJob *job,
int64_t offset, uint64_t bytes);
void backup_cow_request_end(CowRequest *req);
void backup_do_checkpoint(BlockJob *job, Error **errp);
#endif

8
include/scsi/pr-manager.h
View File

@ -5,6 +5,7 @@
#include "qapi/visitor.h"
#include "qom/object_interfaces.h"
#include "block/aio.h"
#include "qemu/coroutine.h"
#define TYPE_PR_MANAGER "pr-manager"
@ -37,11 +38,8 @@ typedef struct PRManagerClass {
} PRManagerClass;
bool pr_manager_is_connected(PRManager *pr_mgr);
BlockAIOCB *pr_manager_execute(PRManager *pr_mgr,
AioContext *ctx, int fd,
struct sg_io_hdr *hdr,
BlockCompletionFunc *complete,
void *opaque);
int coroutine_fn pr_manager_execute(PRManager *pr_mgr, AioContext *ctx, int fd,
struct sg_io_hdr *hdr);
PRManager *pr_manager_lookup(const char *id, Error **errp);

29
qemu-io-cmds.c
View File

@ -10,6 +10,7 @@
#include "qemu/osdep.h"
#include "qapi/error.h"
#include "qapi/qmp/qdict.h"
#include "qemu-io.h"
#include "sysemu/block-backend.h"
#include "block/block.h"
@ -1978,6 +1979,7 @@ static int reopen_f(BlockBackend *blk, int argc, char **argv)
int flags = bs->open_flags;
bool writethrough = !blk_enable_write_cache(blk);
bool has_rw_option = false;
bool has_cache_option = false;
BlockReopenQueue *brq;
Error *local_err = NULL;
@ -1989,6 +1991,7 @@ static int reopen_f(BlockBackend *blk, int argc, char **argv)
error_report("Invalid cache option: %s", optarg);
return -EINVAL;
}
has_cache_option = true;
break;
case 'o':
if (!qemu_opts_parse_noisily(&reopen_opts, optarg, 0)) {
@ -2046,11 +2049,33 @@ static int reopen_f(BlockBackend *blk, int argc, char **argv)
}
qopts = qemu_opts_find(&reopen_opts, NULL);
opts = qopts ? qemu_opts_to_qdict(qopts, NULL) : NULL;
opts = qopts ? qemu_opts_to_qdict(qopts, NULL) : qdict_new();
qemu_opts_reset(&reopen_opts);
if (qdict_haskey(opts, BDRV_OPT_READ_ONLY)) {
if (has_rw_option) {
error_report("Cannot set both -r/-w and '" BDRV_OPT_READ_ONLY "'");
qobject_unref(opts);
return -EINVAL;
}
} else {
qdict_put_bool(opts, BDRV_OPT_READ_ONLY, !(flags & BDRV_O_RDWR));
}
if (qdict_haskey(opts, BDRV_OPT_CACHE_DIRECT) ||
qdict_haskey(opts, BDRV_OPT_CACHE_NO_FLUSH)) {
if (has_cache_option) {
error_report("Cannot set both -c and the cache options");
qobject_unref(opts);
return -EINVAL;
}
} else {
qdict_put_bool(opts, BDRV_OPT_CACHE_DIRECT, flags & BDRV_O_NOCACHE);
qdict_put_bool(opts, BDRV_OPT_CACHE_NO_FLUSH, flags & BDRV_O_NO_FLUSH);
}
bdrv_subtree_drained_begin(bs);
brq = bdrv_reopen_queue(NULL, bs, opts, flags);
brq = bdrv_reopen_queue(NULL, bs, opts);
bdrv_reopen_multiple(bdrv_get_aio_context(bs), brq, &local_err);
bdrv_subtree_drained_end(bs);

21
scsi/pr-manager.c
View File

@ -48,24 +48,21 @@ static int pr_manager_worker(void *opaque)
}
BlockAIOCB *pr_manager_execute(PRManager *pr_mgr,
AioContext *ctx, int fd,
struct sg_io_hdr *hdr,
BlockCompletionFunc *complete,
void *opaque)
int coroutine_fn pr_manager_execute(PRManager *pr_mgr, AioContext *ctx, int fd,
struct sg_io_hdr *hdr)
{
PRManagerData *data = g_new(PRManagerData, 1);
ThreadPool *pool = aio_get_thread_pool(ctx);
PRManagerData data = {
.pr_mgr = pr_mgr,
.fd = fd,
.hdr = hdr,
};
trace_pr_manager_execute(fd, hdr->cmdp[0], hdr->cmdp[1], opaque);
data->pr_mgr = pr_mgr;
data->fd = fd;
data->hdr = hdr;
trace_pr_manager_execute(fd, hdr->cmdp[0], hdr->cmdp[1]);
/* The matching object_unref is in pr_manager_worker. */
object_ref(OBJECT(pr_mgr));
return thread_pool_submit_aio(pool, pr_manager_worker,
data, complete, opaque);
return thread_pool_submit_co(pool, pr_manager_worker, &data);
}
bool pr_manager_is_connected(PRManager *pr_mgr)

2
scsi/trace-events
View File

@ -1,3 +1,3 @@
# scsi/pr-manager.c
pr_manager_execute(int fd, int cmd, int sa, void *opaque) "fd=%d cmd=0x%02x service action=0x%02x opaque=%p"
pr_manager_execute(int fd, int cmd, int sa) "fd=%d cmd=0x%02x service action=0x%02x"
pr_manager_run(int fd, int cmd, int sa) "fd=%d cmd=0x%02x service action=0x%02x"

18
tests/qemu-iotests/133
View File

@ -91,6 +91,24 @@ echo
IMGOPTSSYNTAX=false $QEMU_IO -f null-co -c 'reopen' -c 'info' \
"json:{'driver': 'null-co', 'size': 65536}"
echo
echo "=== Check that mixing -c/-r/-w and their corresponding options is forbidden ==="
echo
$QEMU_IO -c 'reopen -r -o read-only=on' $TEST_IMG
$QEMU_IO -c 'reopen -w -o read-only=on' $TEST_IMG
$QEMU_IO -c 'reopen -c none -o cache.direct=on' $TEST_IMG
$QEMU_IO -c 'reopen -c writeback -o cache.direct=on' $TEST_IMG
$QEMU_IO -c 'reopen -c directsync -o cache.no-flush=on' $TEST_IMG
echo
echo "=== Check that invalid options are handled correctly ==="
echo
$QEMU_IO -c 'reopen -o read-only=foo' $TEST_IMG
$QEMU_IO -c 'reopen -o cache.no-flush=bar' $TEST_IMG
$QEMU_IO -c 'reopen -o cache.direct=baz' $TEST_IMG
$QEMU_IO -c 'reopen -o auto-read-only=qux' $TEST_IMG
# success, all done
echo "*** done"
rm -f $seq.full

15
tests/qemu-iotests/133.out
View File

@ -24,4 +24,19 @@ Cannot change the option 'driver'
format name: null-co
format name: null-co
=== Check that mixing -c/-r/-w and their corresponding options is forbidden ===
Cannot set both -r/-w and 'read-only'
Cannot set both -r/-w and 'read-only'
Cannot set both -c and the cache options
Cannot set both -c and the cache options
Cannot set both -c and the cache options
=== Check that invalid options are handled correctly ===
Parameter 'read-only' expects 'on' or 'off'
Parameter 'cache.no-flush' expects 'on' or 'off'
Parameter 'cache.direct' expects 'on' or 'off'
Parameter 'auto-read-only' expects 'on' or 'off'
*** done

4
tests/qemu-iotests/235
View File

@ -49,7 +49,9 @@ qemu_img_create('-f', iotests.imgfmt, '-o', 'preallocation=metadata', disk,
str(size))
vm = QEMUMachine(iotests.qemu_prog)
vm.add_args('-machine', 'pc,accel=kvm')
vm.add_args('-machine', 'accel=kvm')
if iotests.qemu_default_machine == 's390-ccw-virtio':
vm.add_args('-no-shutdown')
vm.add_args('-drive', 'id=src,file=' + disk)
vm.launch()