2010-02-18 19:48:12 +03:00
|
|
|
/*
|
|
|
|
* Block protocol for I/O error injection
|
|
|
|
*
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
* Copyright (C) 2016-2017 Red Hat, Inc.
|
2010-02-18 19:48:12 +03:00
|
|
|
* 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.
|
|
|
|
*/
|
|
|
|
|
2016-01-18 21:01:42 +03:00
|
|
|
#include "qemu/osdep.h"
|
include/qemu/osdep.h: Don't include qapi/error.h
Commit 57cb38b included qapi/error.h into qemu/osdep.h to get the
Error typedef. Since then, we've moved to include qemu/osdep.h
everywhere. Its file comment explains: "To avoid getting into
possible circular include dependencies, this file should not include
any other QEMU headers, with the exceptions of config-host.h,
compiler.h, os-posix.h and os-win32.h, all of which are doing a
similar job to this file and are under similar constraints."
qapi/error.h doesn't do a similar job, and it doesn't adhere to
similar constraints: it includes qapi-types.h. That's in excess of
100KiB of crap most .c files don't actually need.
Add the typedef to qemu/typedefs.h, and include that instead of
qapi/error.h. Include qapi/error.h in .c files that need it and don't
get it now. Include qapi-types.h in qom/object.h for uint16List.
Update scripts/clean-includes accordingly. Update it further to match
reality: replace config.h by config-target.h, add sysemu/os-posix.h,
sysemu/os-win32.h. Update the list of includes in the qemu/osdep.h
comment quoted above similarly.
This reduces the number of objects depending on qapi/error.h from "all
of them" to less than a third. Unfortunately, the number depending on
qapi-types.h shrinks only a little. More work is needed for that one.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
[Fix compilation without the spice devel packages. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2016-03-14 11:01:28 +03:00
|
|
|
#include "qapi/error.h"
|
2016-03-20 20:16:19 +03:00
|
|
|
#include "qemu/cutils.h"
|
2012-12-17 21:20:00 +04:00
|
|
|
#include "qemu/config-file.h"
|
2012-12-17 21:19:44 +04:00
|
|
|
#include "block/block_int.h"
|
2019-11-08 15:34:53 +03:00
|
|
|
#include "block/qdict.h"
|
2012-12-17 21:20:00 +04:00
|
|
|
#include "qemu/module.h"
|
2018-02-01 14:18:46 +03:00
|
|
|
#include "qemu/option.h"
|
2019-11-08 15:34:53 +03:00
|
|
|
#include "qapi/qapi-visit-block-core.h"
|
2014-07-18 22:24:57 +04:00
|
|
|
#include "qapi/qmp/qdict.h"
|
2019-11-08 15:34:53 +03:00
|
|
|
#include "qapi/qmp/qlist.h"
|
2014-07-18 22:24:57 +04:00
|
|
|
#include "qapi/qmp/qstring.h"
|
2019-11-08 15:34:53 +03:00
|
|
|
#include "qapi/qobject-input-visitor.h"
|
2015-11-30 14:44:44 +03:00
|
|
|
#include "sysemu/qtest.h"
|
2010-02-18 19:48:12 +03:00
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
/* All APIs are thread-safe */
|
|
|
|
|
2010-02-18 19:48:12 +03:00
|
|
|
typedef struct BDRVBlkdebugState {
|
2021-06-14 11:29:31 +03:00
|
|
|
/* IN: initialized in blkdebug_open() and never changed */
|
2017-04-29 22:14:17 +03:00
|
|
|
uint64_t align;
|
2017-04-29 22:14:18 +03:00
|
|
|
uint64_t max_transfer;
|
|
|
|
uint64_t opt_write_zero;
|
|
|
|
uint64_t max_write_zero;
|
|
|
|
uint64_t opt_discard;
|
|
|
|
uint64_t max_discard;
|
2021-06-14 11:29:31 +03:00
|
|
|
char *config_file; /* For blkdebug_refresh_filename() */
|
|
|
|
/* initialized in blkdebug_parse_perms() */
|
2019-11-08 15:34:53 +03:00
|
|
|
uint64_t take_child_perms;
|
|
|
|
uint64_t unshare_child_perms;
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
/* State. Protected by lock */
|
|
|
|
int state;
|
qapi: Don't let implicit enum MAX member collide
Now that we guarantee the user doesn't have any enum values
beginning with a single underscore, we can use that for our
own purposes. Renaming ENUM_MAX to ENUM__MAX makes it obvious
that the sentinel is generated.
This patch was mostly generated by applying a temporary patch:
|diff --git a/scripts/qapi.py b/scripts/qapi.py
|index e6d014b..b862ec9 100644
|--- a/scripts/qapi.py
|+++ b/scripts/qapi.py
|@@ -1570,6 +1570,7 @@ const char *const %(c_name)s_lookup[] = {
| max_index = c_enum_const(name, 'MAX', prefix)
| ret += mcgen('''
| [%(max_index)s] = NULL,
|+// %(max_index)s
| };
| ''',
| max_index=max_index)
then running:
$ cat qapi-{types,event}.c tests/test-qapi-types.c |
sed -n 's,^// \(.*\)MAX,s|\1MAX|\1_MAX|g,p' > list
$ git grep -l _MAX | xargs sed -i -f list
The only things not generated are the changes in scripts/qapi.py.
Rejecting enum members named 'MAX' is now useless, and will be dropped
in the next patch.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <1447836791-369-23-git-send-email-eblake@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
[Rebased to current master, commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
2015-11-18 11:52:57 +03:00
|
|
|
QLIST_HEAD(, BlkdebugRule) rules[BLKDBG__MAX];
|
2012-06-06 10:10:42 +04:00
|
|
|
QSIMPLEQ_HEAD(, BlkdebugRule) active_rules;
|
2012-12-06 17:32:57 +04:00
|
|
|
QLIST_HEAD(, BlkdebugSuspendedReq) suspended_reqs;
|
2021-06-14 11:29:31 +03:00
|
|
|
QemuMutex lock;
|
2010-02-18 19:48:12 +03:00
|
|
|
} BDRVBlkdebugState;
|
|
|
|
|
2010-02-19 18:24:35 +03:00
|
|
|
typedef struct BlkdebugAIOCB {
|
2014-10-07 15:59:14 +04:00
|
|
|
BlockAIOCB common;
|
2010-02-19 18:24:35 +03:00
|
|
|
int ret;
|
|
|
|
} BlkdebugAIOCB;
|
|
|
|
|
2012-12-06 17:32:57 +04:00
|
|
|
typedef struct BlkdebugSuspendedReq {
|
2021-06-14 11:29:31 +03:00
|
|
|
/* IN: initialized in suspend_request() */
|
2012-12-06 17:32:57 +04:00
|
|
|
Coroutine *co;
|
|
|
|
char *tag;
|
2021-06-14 11:29:31 +03:00
|
|
|
|
|
|
|
/* List entry protected BDRVBlkdebugState's lock */
|
2012-12-06 17:32:57 +04:00
|
|
|
QLIST_ENTRY(BlkdebugSuspendedReq) next;
|
|
|
|
} BlkdebugSuspendedReq;
|
|
|
|
|
2010-03-15 19:27:00 +03:00
|
|
|
enum {
|
|
|
|
ACTION_INJECT_ERROR,
|
|
|
|
ACTION_SET_STATE,
|
2012-12-06 17:32:57 +04:00
|
|
|
ACTION_SUSPEND,
|
2021-06-14 11:29:28 +03:00
|
|
|
ACTION__MAX,
|
2010-03-15 19:27:00 +03:00
|
|
|
};
|
|
|
|
|
|
|
|
typedef struct BlkdebugRule {
|
2021-06-14 11:29:31 +03:00
|
|
|
/* IN: initialized in add_rule() or blkdebug_debug_breakpoint() */
|
2015-11-18 11:52:54 +03:00
|
|
|
BlkdebugEvent event;
|
2010-03-15 19:27:00 +03:00
|
|
|
int action;
|
|
|
|
int state;
|
|
|
|
union {
|
|
|
|
struct {
|
2019-05-07 23:35:04 +03:00
|
|
|
uint64_t iotype_mask;
|
2010-03-15 19:27:00 +03:00
|
|
|
int error;
|
|
|
|
int immediately;
|
|
|
|
int once;
|
2016-11-04 23:13:45 +03:00
|
|
|
int64_t offset;
|
2010-03-15 19:27:00 +03:00
|
|
|
} inject;
|
|
|
|
struct {
|
|
|
|
int new_state;
|
|
|
|
} set_state;
|
2012-12-06 17:32:57 +04:00
|
|
|
struct {
|
|
|
|
char *tag;
|
|
|
|
} suspend;
|
2010-03-15 19:27:00 +03:00
|
|
|
} options;
|
2021-06-14 11:29:31 +03:00
|
|
|
|
|
|
|
/* List entries protected BDRVBlkdebugState's lock */
|
2010-03-15 19:27:00 +03:00
|
|
|
QLIST_ENTRY(BlkdebugRule) next;
|
2012-06-06 10:10:42 +04:00
|
|
|
QSIMPLEQ_ENTRY(BlkdebugRule) active_next;
|
2010-03-15 19:27:00 +03:00
|
|
|
} BlkdebugRule;
|
|
|
|
|
2019-05-07 23:35:04 +03:00
|
|
|
QEMU_BUILD_BUG_MSG(BLKDEBUG_IO_TYPE__MAX > 64,
|
|
|
|
"BlkdebugIOType mask does not fit into an uint64_t");
|
|
|
|
|
2010-03-15 19:27:00 +03:00
|
|
|
static QemuOptsList inject_error_opts = {
|
|
|
|
.name = "inject-error",
|
|
|
|
.head = QTAILQ_HEAD_INITIALIZER(inject_error_opts.head),
|
|
|
|
.desc = {
|
|
|
|
{
|
|
|
|
.name = "event",
|
|
|
|
.type = QEMU_OPT_STRING,
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "state",
|
|
|
|
.type = QEMU_OPT_NUMBER,
|
|
|
|
},
|
2019-05-07 23:35:04 +03:00
|
|
|
{
|
|
|
|
.name = "iotype",
|
|
|
|
.type = QEMU_OPT_STRING,
|
|
|
|
},
|
2010-03-15 19:27:00 +03:00
|
|
|
{
|
|
|
|
.name = "errno",
|
|
|
|
.type = QEMU_OPT_NUMBER,
|
|
|
|
},
|
2012-06-06 10:10:43 +04:00
|
|
|
{
|
|
|
|
.name = "sector",
|
|
|
|
.type = QEMU_OPT_NUMBER,
|
|
|
|
},
|
2010-03-15 19:27:00 +03:00
|
|
|
{
|
|
|
|
.name = "once",
|
|
|
|
.type = QEMU_OPT_BOOL,
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "immediately",
|
|
|
|
.type = QEMU_OPT_BOOL,
|
|
|
|
},
|
|
|
|
{ /* end of list */ }
|
|
|
|
},
|
|
|
|
};
|
|
|
|
|
|
|
|
static QemuOptsList set_state_opts = {
|
|
|
|
.name = "set-state",
|
2010-06-30 19:40:42 +04:00
|
|
|
.head = QTAILQ_HEAD_INITIALIZER(set_state_opts.head),
|
2010-03-15 19:27:00 +03:00
|
|
|
.desc = {
|
|
|
|
{
|
|
|
|
.name = "event",
|
|
|
|
.type = QEMU_OPT_STRING,
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "state",
|
|
|
|
.type = QEMU_OPT_NUMBER,
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "new_state",
|
|
|
|
.type = QEMU_OPT_NUMBER,
|
|
|
|
},
|
|
|
|
{ /* end of list */ }
|
|
|
|
},
|
|
|
|
};
|
|
|
|
|
|
|
|
static QemuOptsList *config_groups[] = {
|
|
|
|
&inject_error_opts,
|
|
|
|
&set_state_opts,
|
|
|
|
NULL
|
|
|
|
};
|
|
|
|
|
|
|
|
struct add_rule_data {
|
|
|
|
BDRVBlkdebugState *s;
|
|
|
|
int action;
|
|
|
|
};
|
|
|
|
|
2015-03-13 15:35:14 +03:00
|
|
|
static int add_rule(void *opaque, QemuOpts *opts, Error **errp)
|
2010-03-15 19:27:00 +03:00
|
|
|
{
|
|
|
|
struct add_rule_data *d = opaque;
|
|
|
|
BDRVBlkdebugState *s = d->s;
|
2020-10-30 06:35:12 +03:00
|
|
|
const char *event_name;
|
2017-08-24 11:46:02 +03:00
|
|
|
int event;
|
2010-03-15 19:27:00 +03:00
|
|
|
struct BlkdebugRule *rule;
|
2016-11-04 23:13:45 +03:00
|
|
|
int64_t sector;
|
2019-05-07 23:35:04 +03:00
|
|
|
BlkdebugIOType iotype;
|
|
|
|
Error *local_error = NULL;
|
2010-03-15 19:27:00 +03:00
|
|
|
|
|
|
|
/* Find the right event for the rule */
|
|
|
|
event_name = qemu_opt_get(opts, "event");
|
2014-09-20 12:55:52 +04:00
|
|
|
if (!event_name) {
|
2015-03-13 15:38:42 +03:00
|
|
|
error_setg(errp, "Missing event name for rule");
|
2014-09-20 12:55:52 +04:00
|
|
|
return -1;
|
2017-08-24 11:46:02 +03:00
|
|
|
}
|
2017-08-24 11:46:10 +03:00
|
|
|
event = qapi_enum_parse(&BlkdebugEvent_lookup, event_name, -1, errp);
|
2017-08-24 11:46:02 +03:00
|
|
|
if (event < 0) {
|
2010-03-15 19:27:00 +03:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Set attributes common for all actions */
|
2011-08-21 07:09:37 +04:00
|
|
|
rule = g_malloc0(sizeof(*rule));
|
2010-03-15 19:27:00 +03:00
|
|
|
*rule = (struct BlkdebugRule) {
|
|
|
|
.event = event,
|
|
|
|
.action = d->action,
|
|
|
|
.state = qemu_opt_get_number(opts, "state", 0),
|
|
|
|
};
|
|
|
|
|
|
|
|
/* Parse action-specific options */
|
|
|
|
switch (d->action) {
|
|
|
|
case ACTION_INJECT_ERROR:
|
|
|
|
rule->options.inject.error = qemu_opt_get_number(opts, "errno", EIO);
|
|
|
|
rule->options.inject.once = qemu_opt_get_bool(opts, "once", 0);
|
|
|
|
rule->options.inject.immediately =
|
|
|
|
qemu_opt_get_bool(opts, "immediately", 0);
|
2016-11-04 23:13:45 +03:00
|
|
|
sector = qemu_opt_get_number(opts, "sector", -1);
|
|
|
|
rule->options.inject.offset =
|
|
|
|
sector == -1 ? -1 : sector * BDRV_SECTOR_SIZE;
|
2019-05-07 23:35:04 +03:00
|
|
|
|
|
|
|
iotype = qapi_enum_parse(&BlkdebugIOType_lookup,
|
|
|
|
qemu_opt_get(opts, "iotype"),
|
|
|
|
BLKDEBUG_IO_TYPE__MAX, &local_error);
|
|
|
|
if (local_error) {
|
|
|
|
error_propagate(errp, local_error);
|
2020-10-09 22:09:59 +03:00
|
|
|
g_free(rule);
|
2019-05-07 23:35:04 +03:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
if (iotype != BLKDEBUG_IO_TYPE__MAX) {
|
|
|
|
rule->options.inject.iotype_mask = (1ull << iotype);
|
|
|
|
} else {
|
|
|
|
/* Apply the default */
|
|
|
|
rule->options.inject.iotype_mask =
|
|
|
|
(1ull << BLKDEBUG_IO_TYPE_READ)
|
|
|
|
| (1ull << BLKDEBUG_IO_TYPE_WRITE)
|
|
|
|
| (1ull << BLKDEBUG_IO_TYPE_WRITE_ZEROES)
|
|
|
|
| (1ull << BLKDEBUG_IO_TYPE_DISCARD)
|
|
|
|
| (1ull << BLKDEBUG_IO_TYPE_FLUSH);
|
|
|
|
}
|
|
|
|
|
2010-03-15 19:27:00 +03:00
|
|
|
break;
|
|
|
|
|
|
|
|
case ACTION_SET_STATE:
|
|
|
|
rule->options.set_state.new_state =
|
|
|
|
qemu_opt_get_number(opts, "new_state", 0);
|
|
|
|
break;
|
2012-12-06 17:32:57 +04:00
|
|
|
|
|
|
|
case ACTION_SUSPEND:
|
|
|
|
rule->options.suspend.tag =
|
|
|
|
g_strdup(qemu_opt_get(opts, "tag"));
|
|
|
|
break;
|
2010-03-15 19:27:00 +03:00
|
|
|
};
|
|
|
|
|
|
|
|
/* Add the rule */
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_lock(&s->lock);
|
2010-03-15 19:27:00 +03:00
|
|
|
QLIST_INSERT_HEAD(&s->rules[event], rule, next);
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_unlock(&s->lock);
|
2010-03-15 19:27:00 +03:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
/* Called with lock held or from .bdrv_close */
|
2012-12-06 17:32:56 +04:00
|
|
|
static void remove_rule(BlkdebugRule *rule)
|
|
|
|
{
|
|
|
|
switch (rule->action) {
|
|
|
|
case ACTION_INJECT_ERROR:
|
|
|
|
case ACTION_SET_STATE:
|
|
|
|
break;
|
2012-12-06 17:32:57 +04:00
|
|
|
case ACTION_SUSPEND:
|
|
|
|
g_free(rule->options.suspend.tag);
|
|
|
|
break;
|
2012-12-06 17:32:56 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
QLIST_REMOVE(rule, next);
|
|
|
|
g_free(rule);
|
|
|
|
}
|
|
|
|
|
2013-12-20 22:28:07 +04:00
|
|
|
static int read_config(BDRVBlkdebugState *s, const char *filename,
|
|
|
|
QDict *options, Error **errp)
|
2010-03-15 19:27:00 +03:00
|
|
|
{
|
2013-12-20 22:28:06 +04:00
|
|
|
FILE *f = NULL;
|
2010-03-15 19:27:00 +03:00
|
|
|
int ret;
|
|
|
|
struct add_rule_data d;
|
2013-12-20 22:28:07 +04:00
|
|
|
Error *local_err = NULL;
|
2010-03-15 19:27:00 +03:00
|
|
|
|
2013-12-20 22:28:06 +04:00
|
|
|
if (filename) {
|
|
|
|
f = fopen(filename, "r");
|
|
|
|
if (f == NULL) {
|
|
|
|
error_setg_errno(errp, errno, "Could not read blkdebug config file");
|
|
|
|
return -errno;
|
|
|
|
}
|
2010-03-15 19:27:00 +03:00
|
|
|
|
2021-02-26 20:08:16 +03:00
|
|
|
ret = qemu_config_parse(f, config_groups, filename, errp);
|
2013-12-20 22:28:06 +04:00
|
|
|
if (ret < 0) {
|
|
|
|
goto fail;
|
|
|
|
}
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
|
|
|
|
2013-12-20 22:28:07 +04:00
|
|
|
qemu_config_parse_qdict(options, config_groups, &local_err);
|
2014-01-30 18:07:28 +04:00
|
|
|
if (local_err) {
|
2013-12-20 22:28:07 +04:00
|
|
|
error_propagate(errp, local_err);
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
|
2010-03-15 19:27:00 +03:00
|
|
|
d.s = s;
|
|
|
|
d.action = ACTION_INJECT_ERROR;
|
2015-03-13 15:38:42 +03:00
|
|
|
qemu_opts_foreach(&inject_error_opts, add_rule, &d, &local_err);
|
2014-09-20 12:55:52 +04:00
|
|
|
if (local_err) {
|
|
|
|
error_propagate(errp, local_err);
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto fail;
|
|
|
|
}
|
2010-03-15 19:27:00 +03:00
|
|
|
|
|
|
|
d.action = ACTION_SET_STATE;
|
2015-03-13 15:38:42 +03:00
|
|
|
qemu_opts_foreach(&set_state_opts, add_rule, &d, &local_err);
|
2014-09-20 12:55:52 +04:00
|
|
|
if (local_err) {
|
|
|
|
error_propagate(errp, local_err);
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto fail;
|
|
|
|
}
|
2010-03-15 19:27:00 +03:00
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
fail:
|
2010-06-30 19:42:23 +04:00
|
|
|
qemu_opts_reset(&inject_error_opts);
|
|
|
|
qemu_opts_reset(&set_state_opts);
|
2013-12-20 22:28:06 +04:00
|
|
|
if (f) {
|
|
|
|
fclose(f);
|
|
|
|
}
|
2010-03-15 19:27:00 +03:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Valid blkdebug filenames look like blkdebug:path/to/config:path/to/image */
|
2013-04-10 15:37:33 +04:00
|
|
|
static void blkdebug_parse_filename(const char *filename, QDict *options,
|
|
|
|
Error **errp)
|
2010-02-18 19:48:12 +03:00
|
|
|
{
|
2013-04-10 15:37:33 +04:00
|
|
|
const char *c;
|
2010-02-18 19:48:12 +03:00
|
|
|
|
2010-03-15 19:27:00 +03:00
|
|
|
/* Parse the blkdebug: prefix */
|
2013-04-10 15:37:33 +04:00
|
|
|
if (!strstart(filename, "blkdebug:", &filename)) {
|
2013-12-20 22:28:02 +04:00
|
|
|
/* There was no prefix; therefore, all options have to be already
|
|
|
|
present in the QDict (except for the filename) */
|
2017-04-28 00:58:17 +03:00
|
|
|
qdict_put_str(options, "x-image", filename);
|
2013-04-10 15:37:33 +04:00
|
|
|
return;
|
2010-02-18 19:48:12 +03:00
|
|
|
}
|
|
|
|
|
2013-04-10 15:37:33 +04:00
|
|
|
/* Parse config file path */
|
2010-03-15 19:27:00 +03:00
|
|
|
c = strchr(filename, ':');
|
|
|
|
if (c == NULL) {
|
2013-04-10 15:37:33 +04:00
|
|
|
error_setg(errp, "blkdebug requires both config file and image path");
|
|
|
|
return;
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
|
|
|
|
2013-04-10 15:37:33 +04:00
|
|
|
if (c != filename) {
|
|
|
|
QString *config_path;
|
2018-07-27 09:22:04 +03:00
|
|
|
config_path = qstring_from_substr(filename, 0, c - filename);
|
2013-04-10 15:37:33 +04:00
|
|
|
qdict_put(options, "config", config_path);
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
2013-04-10 15:37:33 +04:00
|
|
|
|
|
|
|
/* TODO Allow multi-level nesting and set file.filename here */
|
2010-03-15 19:27:00 +03:00
|
|
|
filename = c + 1;
|
2017-04-28 00:58:17 +03:00
|
|
|
qdict_put_str(options, "x-image", filename);
|
2013-04-10 15:37:33 +04:00
|
|
|
}
|
|
|
|
|
2019-11-08 15:34:53 +03:00
|
|
|
static int blkdebug_parse_perm_list(uint64_t *dest, QDict *options,
|
|
|
|
const char *prefix, Error **errp)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
QDict *subqdict = NULL;
|
|
|
|
QObject *crumpled_subqdict = NULL;
|
|
|
|
Visitor *v = NULL;
|
|
|
|
BlockPermissionList *perm_list = NULL, *element;
|
|
|
|
|
|
|
|
*dest = 0;
|
|
|
|
|
|
|
|
qdict_extract_subqdict(options, &subqdict, prefix);
|
|
|
|
if (!qdict_size(subqdict)) {
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
crumpled_subqdict = qdict_crumple(subqdict, errp);
|
|
|
|
if (!crumpled_subqdict) {
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
v = qobject_input_visitor_new(crumpled_subqdict);
|
2020-07-07 19:06:03 +03:00
|
|
|
if (!visit_type_BlockPermissionList(v, NULL, &perm_list, errp)) {
|
2019-11-08 15:34:53 +03:00
|
|
|
ret = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (element = perm_list; element; element = element->next) {
|
|
|
|
*dest |= bdrv_qapi_perm_to_blk_perm(element->value);
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
qapi_free_BlockPermissionList(perm_list);
|
|
|
|
visit_free(v);
|
|
|
|
qobject_unref(subqdict);
|
|
|
|
qobject_unref(crumpled_subqdict);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int blkdebug_parse_perms(BDRVBlkdebugState *s, QDict *options,
|
|
|
|
Error **errp)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = blkdebug_parse_perm_list(&s->take_child_perms, options,
|
|
|
|
"take-child-perms.", errp);
|
|
|
|
if (ret < 0) {
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = blkdebug_parse_perm_list(&s->unshare_child_perms, options,
|
|
|
|
"unshare-child-perms.", errp);
|
|
|
|
if (ret < 0) {
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-04-10 15:37:33 +04:00
|
|
|
static QemuOptsList runtime_opts = {
|
|
|
|
.name = "blkdebug",
|
|
|
|
.head = QTAILQ_HEAD_INITIALIZER(runtime_opts.head),
|
|
|
|
.desc = {
|
|
|
|
{
|
|
|
|
.name = "config",
|
|
|
|
.type = QEMU_OPT_STRING,
|
|
|
|
.help = "Path to the configuration file",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "x-image",
|
|
|
|
.type = QEMU_OPT_STRING,
|
|
|
|
.help = "[internal use only, will be removed]",
|
|
|
|
},
|
2014-01-14 16:44:35 +04:00
|
|
|
{
|
|
|
|
.name = "align",
|
|
|
|
.type = QEMU_OPT_SIZE,
|
|
|
|
.help = "Required alignment in bytes",
|
|
|
|
},
|
2017-04-29 22:14:18 +03:00
|
|
|
{
|
|
|
|
.name = "max-transfer",
|
|
|
|
.type = QEMU_OPT_SIZE,
|
|
|
|
.help = "Maximum transfer size in bytes",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "opt-write-zero",
|
|
|
|
.type = QEMU_OPT_SIZE,
|
|
|
|
.help = "Optimum write zero alignment in bytes",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "max-write-zero",
|
|
|
|
.type = QEMU_OPT_SIZE,
|
|
|
|
.help = "Maximum write zero size in bytes",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "opt-discard",
|
|
|
|
.type = QEMU_OPT_SIZE,
|
|
|
|
.help = "Optimum discard alignment in bytes",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = "max-discard",
|
|
|
|
.type = QEMU_OPT_SIZE,
|
|
|
|
.help = "Maximum discard size in bytes",
|
|
|
|
},
|
2013-04-10 15:37:33 +04:00
|
|
|
{ /* end of list */ }
|
|
|
|
},
|
|
|
|
};
|
|
|
|
|
2013-09-05 16:22:29 +04:00
|
|
|
static int blkdebug_open(BlockDriverState *bs, QDict *options, int flags,
|
|
|
|
Error **errp)
|
2013-04-10 15:37:33 +04:00
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
|
|
|
QemuOpts *opts;
|
|
|
|
int ret;
|
2017-04-29 22:14:18 +03:00
|
|
|
uint64_t align;
|
2013-04-10 15:37:33 +04:00
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_init(&s->lock);
|
2014-01-02 06:49:17 +04:00
|
|
|
opts = qemu_opts_create(&runtime_opts, NULL, 0, &error_abort);
|
2020-07-07 19:06:03 +03:00
|
|
|
if (!qemu_opts_absorb_qdict(opts, options, errp)) {
|
2013-04-10 15:37:33 +04:00
|
|
|
ret = -EINVAL;
|
2014-02-08 12:53:22 +04:00
|
|
|
goto out;
|
2013-04-10 15:37:33 +04:00
|
|
|
}
|
|
|
|
|
2013-12-20 22:28:07 +04:00
|
|
|
/* Read rules from config file or command line options */
|
2016-08-15 16:29:25 +03:00
|
|
|
s->config_file = g_strdup(qemu_opt_get(opts, "config"));
|
|
|
|
ret = read_config(s, s->config_file, options, errp);
|
2013-12-20 22:28:06 +04:00
|
|
|
if (ret) {
|
2014-02-08 12:53:22 +04:00
|
|
|
goto out;
|
2013-04-10 15:37:33 +04:00
|
|
|
}
|
2010-03-15 19:27:00 +03:00
|
|
|
|
2010-06-30 19:43:40 +04:00
|
|
|
/* Set initial state */
|
2012-06-06 10:10:42 +04:00
|
|
|
s->state = 1;
|
2010-06-30 19:43:40 +04:00
|
|
|
|
2019-11-08 15:34:53 +03:00
|
|
|
/* Parse permissions modifiers before opening the image file */
|
|
|
|
ret = blkdebug_parse_perms(s, options, errp);
|
|
|
|
if (ret < 0) {
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2015-10-16 13:46:04 +03:00
|
|
|
/* Open the image file */
|
2015-06-16 15:19:22 +03:00
|
|
|
bs->file = bdrv_open_child(qemu_opt_get(opts, "x-image"), options, "image",
|
2020-05-13 14:05:36 +03:00
|
|
|
bs, &child_of_bds,
|
|
|
|
BDRV_CHILD_FILTERED | BDRV_CHILD_PRIMARY,
|
2021-02-02 15:49:45 +03:00
|
|
|
false, errp);
|
|
|
|
if (!bs->file) {
|
2015-06-16 15:19:22 +03:00
|
|
|
ret = -EINVAL;
|
2014-02-08 12:53:22 +04:00
|
|
|
goto out;
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
|
|
|
|
2018-04-21 16:29:26 +03:00
|
|
|
bs->supported_write_flags = BDRV_REQ_WRITE_UNCHANGED |
|
|
|
|
(BDRV_REQ_FUA & bs->file->bs->supported_write_flags);
|
|
|
|
bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED |
|
2019-03-22 15:42:39 +03:00
|
|
|
((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK) &
|
2018-04-21 16:29:26 +03:00
|
|
|
bs->file->bs->supported_zero_flags);
|
2017-04-29 22:14:17 +03:00
|
|
|
ret = -EINVAL;
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
|
2017-04-29 22:14:18 +03:00
|
|
|
/* Set alignment overrides */
|
2017-04-29 22:14:17 +03:00
|
|
|
s->align = qemu_opt_get_size(opts, "align", 0);
|
|
|
|
if (s->align && (s->align >= INT_MAX || !is_power_of_2(s->align))) {
|
|
|
|
error_setg(errp, "Cannot meet constraints with align %" PRIu64,
|
|
|
|
s->align);
|
2017-04-13 18:43:34 +03:00
|
|
|
goto out;
|
2014-01-14 16:44:35 +04:00
|
|
|
}
|
2017-04-29 22:14:18 +03:00
|
|
|
align = MAX(s->align, bs->file->bs->bl.request_alignment);
|
|
|
|
|
|
|
|
s->max_transfer = qemu_opt_get_size(opts, "max-transfer", 0);
|
|
|
|
if (s->max_transfer &&
|
|
|
|
(s->max_transfer >= INT_MAX ||
|
|
|
|
!QEMU_IS_ALIGNED(s->max_transfer, align))) {
|
|
|
|
error_setg(errp, "Cannot meet constraints with max-transfer %" PRIu64,
|
|
|
|
s->max_transfer);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
s->opt_write_zero = qemu_opt_get_size(opts, "opt-write-zero", 0);
|
|
|
|
if (s->opt_write_zero &&
|
|
|
|
(s->opt_write_zero >= INT_MAX ||
|
|
|
|
!QEMU_IS_ALIGNED(s->opt_write_zero, align))) {
|
|
|
|
error_setg(errp, "Cannot meet constraints with opt-write-zero %" PRIu64,
|
|
|
|
s->opt_write_zero);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
s->max_write_zero = qemu_opt_get_size(opts, "max-write-zero", 0);
|
|
|
|
if (s->max_write_zero &&
|
|
|
|
(s->max_write_zero >= INT_MAX ||
|
|
|
|
!QEMU_IS_ALIGNED(s->max_write_zero,
|
|
|
|
MAX(s->opt_write_zero, align)))) {
|
|
|
|
error_setg(errp, "Cannot meet constraints with max-write-zero %" PRIu64,
|
|
|
|
s->max_write_zero);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
s->opt_discard = qemu_opt_get_size(opts, "opt-discard", 0);
|
|
|
|
if (s->opt_discard &&
|
|
|
|
(s->opt_discard >= INT_MAX ||
|
|
|
|
!QEMU_IS_ALIGNED(s->opt_discard, align))) {
|
|
|
|
error_setg(errp, "Cannot meet constraints with opt-discard %" PRIu64,
|
|
|
|
s->opt_discard);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
s->max_discard = qemu_opt_get_size(opts, "max-discard", 0);
|
|
|
|
if (s->max_discard &&
|
|
|
|
(s->max_discard >= INT_MAX ||
|
|
|
|
!QEMU_IS_ALIGNED(s->max_discard,
|
|
|
|
MAX(s->opt_discard, align)))) {
|
|
|
|
error_setg(errp, "Cannot meet constraints with max-discard %" PRIu64,
|
|
|
|
s->max_discard);
|
|
|
|
goto out;
|
|
|
|
}
|
2014-01-14 16:44:35 +04:00
|
|
|
|
2019-05-07 23:35:05 +03:00
|
|
|
bdrv_debug_event(bs, BLKDBG_NONE);
|
|
|
|
|
2013-04-10 15:37:33 +04:00
|
|
|
ret = 0;
|
2014-02-08 12:53:22 +04:00
|
|
|
out:
|
2016-08-15 16:29:25 +03:00
|
|
|
if (ret < 0) {
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_destroy(&s->lock);
|
2016-08-15 16:29:25 +03:00
|
|
|
g_free(s->config_file);
|
|
|
|
}
|
2013-04-10 15:37:33 +04:00
|
|
|
qemu_opts_del(opts);
|
|
|
|
return ret;
|
2010-02-18 19:48:12 +03:00
|
|
|
}
|
|
|
|
|
2019-05-07 23:35:04 +03:00
|
|
|
static int rule_check(BlockDriverState *bs, uint64_t offset, uint64_t bytes,
|
|
|
|
BlkdebugIOType iotype)
|
2010-02-19 18:24:35 +03:00
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
2017-04-29 22:14:15 +03:00
|
|
|
BlkdebugRule *rule = NULL;
|
|
|
|
int error;
|
|
|
|
bool immediately;
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_lock(&s->lock);
|
2017-04-29 22:14:15 +03:00
|
|
|
QSIMPLEQ_FOREACH(rule, &s->active_rules, active_next) {
|
|
|
|
uint64_t inject_offset = rule->options.inject.offset;
|
|
|
|
|
2019-05-07 23:35:04 +03:00
|
|
|
if ((inject_offset == -1 ||
|
|
|
|
(bytes && inject_offset >= offset &&
|
|
|
|
inject_offset < offset + bytes)) &&
|
|
|
|
(rule->options.inject.iotype_mask & (1ull << iotype)))
|
2017-04-29 22:14:15 +03:00
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!rule || !rule->options.inject.error) {
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_unlock(&s->lock);
|
2017-04-29 22:14:15 +03:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
immediately = rule->options.inject.immediately;
|
|
|
|
error = rule->options.inject.error;
|
2010-02-19 18:24:35 +03:00
|
|
|
|
2012-06-06 10:10:42 +04:00
|
|
|
if (rule->options.inject.once) {
|
blkdebug: fix "once" rule
Background:
The blkdebug scripts are currently engineered so that when a debug
event occurs, a prefilter browses a master list of parsed rules for a
certain event and adds them to an "active list" of rules to be used for
the forthcoming action, provided the events and state numbers match.
Then, once the request is received, the last active rule is used to
inject an error if certain parameters match.
This active list is cleared every time the prefilter injects a new
rule for the first time during a debug event.
The "once" rule currently causes the error injection, if it is
triggered, to only clear the active list. This is insufficient for
preventing future injections of the same rule.
Remedy:
This patch /deletes/ the rule from the list that the prefilter
browses, so it is gone for good. In V2, we remove only the rule of
interest from the active list instead of allowing the "once" rule to
clear the entire list of active rules.
Impact:
This affects iotests 026. Several ENOSPC tests that used "once" can
be seen to have output that shows multiple failure messages. After
this patch, the error messages tend to be smaller and less severe, but
the injection can still be seen to be working. I have patched the
expected output to expect the smaller error messages.
Signed-off-by: John Snow <jsnow@redhat.com>
Message-id: 1423257977-25630-1-git-send-email-jsnow@redhat.com
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2015-02-07 00:26:17 +03:00
|
|
|
QSIMPLEQ_REMOVE(&s->active_rules, rule, BlkdebugRule, active_next);
|
|
|
|
remove_rule(rule);
|
2010-02-19 18:24:35 +03:00
|
|
|
}
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_unlock(&s->lock);
|
2016-11-04 23:13:45 +03:00
|
|
|
if (!immediately) {
|
2017-02-13 16:52:26 +03:00
|
|
|
aio_co_schedule(qemu_get_current_aio_context(), qemu_coroutine_self());
|
2016-11-04 23:13:45 +03:00
|
|
|
qemu_coroutine_yield();
|
2010-02-19 18:24:35 +03:00
|
|
|
}
|
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
return -error;
|
2010-02-19 18:24:35 +03:00
|
|
|
}
|
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
static int coroutine_fn
|
block: use int64_t instead of uint64_t in driver read handlers
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver read handlers parameters which are already 64bit to
signed type.
While being here, convert also flags parameter to be BdrvRequestFlags.
Now let's consider all callers. Simple
git grep '\->bdrv_\(aio\|co\)_preadv\(_part\)\?'
shows that's there three callers of driver function:
bdrv_driver_preadv() in block/io.c, passes int64_t, checked by
bdrv_check_qiov_request() to be non-negative.
qcow2_load_vmstate() does bdrv_check_qiov_request().
do_perform_cow_read() has uint64_t argument. And a lot of things in
qcow2 driver are uint64_t, so converting it is big job. But we must
not work with requests that don't satisfy bdrv_check_qiov_request(),
so let's just assert it here.
Still, the functions may be called directly, not only by drv->...
Let's check:
git grep '\.bdrv_\(aio\|co\)_preadv\(_part\)\?\s*=' | \
awk '{print $4}' | sed 's/,//' | sed 's/&//' | sort | uniq | \
while read func; do git grep "$func(" | \
grep -v "$func(BlockDriverState"; done
The only one such caller:
QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, &data, 1);
...
ret = bdrv_replace_test_co_preadv(bs, 0, 1, &qiov, 0);
in tests/unit/test-bdrv-drain.c, and it's OK obviously.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-4-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: fix typos]
Signed-off-by: Eric Blake <eblake@redhat.com>
2021-09-03 13:27:59 +03:00
|
|
|
blkdebug_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes,
|
|
|
|
QEMUIOVector *qiov, BdrvRequestFlags flags)
|
2010-02-18 19:48:12 +03:00
|
|
|
{
|
2017-04-29 22:14:15 +03:00
|
|
|
int err;
|
2012-06-06 10:10:43 +04:00
|
|
|
|
2017-04-29 22:14:14 +03:00
|
|
|
/* Sanity check block layer guarantees */
|
|
|
|
assert(QEMU_IS_ALIGNED(offset, bs->bl.request_alignment));
|
|
|
|
assert(QEMU_IS_ALIGNED(bytes, bs->bl.request_alignment));
|
|
|
|
if (bs->bl.max_transfer) {
|
|
|
|
assert(bytes <= bs->bl.max_transfer);
|
|
|
|
}
|
|
|
|
|
2019-05-07 23:35:04 +03:00
|
|
|
err = rule_check(bs, offset, bytes, BLKDEBUG_IO_TYPE_READ);
|
2017-04-29 22:14:15 +03:00
|
|
|
if (err) {
|
|
|
|
return err;
|
2010-02-19 18:24:35 +03:00
|
|
|
}
|
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
return bdrv_co_preadv(bs->file, offset, bytes, qiov, flags);
|
2010-02-18 19:48:12 +03:00
|
|
|
}
|
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
static int coroutine_fn
|
block: use int64_t instead of uint64_t in driver write handlers
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver write handlers parameters which are already 64bit to
signed type.
While being here, convert also flags parameter to be BdrvRequestFlags.
Now let's consider all callers. Simple
git grep '\->bdrv_\(aio\|co\)_pwritev\(_part\)\?'
shows that's there three callers of driver function:
bdrv_driver_pwritev() and bdrv_driver_pwritev_compressed() in
block/io.c, both pass int64_t, checked by bdrv_check_qiov_request() to
be non-negative.
qcow2_save_vmstate() does bdrv_check_qiov_request().
Still, the functions may be called directly, not only by drv->...
Let's check:
git grep '\.bdrv_\(aio\|co\)_pwritev\(_part\)\?\s*=' | \
awk '{print $4}' | sed 's/,//' | sed 's/&//' | sort | uniq | \
while read func; do git grep "$func(" | \
grep -v "$func(BlockDriverState"; done
shows several callers:
qcow2:
qcow2_co_truncate() write at most up to @offset, which is checked in
generic qcow2_co_truncate() by bdrv_check_request().
qcow2_co_pwritev_compressed_task() pass the request (or part of the
request) that already went through normal write path, so it should
be OK
qcow:
qcow_co_pwritev_compressed() pass int64_t, it's updated by this patch
quorum:
quorum_co_pwrite_zeroes() pass int64_t and int - OK
throttle:
throttle_co_pwritev_compressed() pass int64_t, it's updated by this
patch
vmdk:
vmdk_co_pwritev_compressed() pass int64_t, it's updated by this
patch
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-5-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2021-09-03 13:28:00 +03:00
|
|
|
blkdebug_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes,
|
|
|
|
QEMUIOVector *qiov, BdrvRequestFlags flags)
|
2010-02-18 19:48:12 +03:00
|
|
|
{
|
2017-04-29 22:14:15 +03:00
|
|
|
int err;
|
2012-06-06 10:10:43 +04:00
|
|
|
|
2017-04-29 22:14:14 +03:00
|
|
|
/* Sanity check block layer guarantees */
|
|
|
|
assert(QEMU_IS_ALIGNED(offset, bs->bl.request_alignment));
|
|
|
|
assert(QEMU_IS_ALIGNED(bytes, bs->bl.request_alignment));
|
|
|
|
if (bs->bl.max_transfer) {
|
|
|
|
assert(bytes <= bs->bl.max_transfer);
|
|
|
|
}
|
|
|
|
|
2019-05-07 23:35:04 +03:00
|
|
|
err = rule_check(bs, offset, bytes, BLKDEBUG_IO_TYPE_WRITE);
|
2017-04-29 22:14:15 +03:00
|
|
|
if (err) {
|
|
|
|
return err;
|
2010-02-19 18:24:35 +03:00
|
|
|
}
|
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags);
|
2010-02-18 19:48:12 +03:00
|
|
|
}
|
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
static int blkdebug_co_flush(BlockDriverState *bs)
|
2014-08-05 01:11:02 +04:00
|
|
|
{
|
2019-05-07 23:35:04 +03:00
|
|
|
int err = rule_check(bs, 0, 0, BLKDEBUG_IO_TYPE_FLUSH);
|
2014-08-05 01:11:02 +04:00
|
|
|
|
2017-04-29 22:14:15 +03:00
|
|
|
if (err) {
|
|
|
|
return err;
|
2014-08-05 01:11:02 +04:00
|
|
|
}
|
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
return bdrv_co_flush(bs->file->bs);
|
2014-08-05 01:11:02 +04:00
|
|
|
}
|
|
|
|
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
static int coroutine_fn blkdebug_co_pwrite_zeroes(BlockDriverState *bs,
|
block: use int64_t instead of int in driver write_zeroes handlers
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver write_zeroes handlers bytes parameter to int64_t.
The only caller of all updated function is bdrv_co_do_pwrite_zeroes().
bdrv_co_do_pwrite_zeroes() itself is of course OK with widening of
callee parameter type. Also, bdrv_co_do_pwrite_zeroes()'s
max_write_zeroes is limited to INT_MAX. So, updated functions all are
safe, they will not get "bytes" larger than before.
Still, let's look through all updated functions, and add assertions to
the ones which are actually unprepared to values larger than INT_MAX.
For these drivers also set explicit max_pwrite_zeroes limit.
Let's go:
blkdebug: calculations can't overflow, thanks to
bdrv_check_qiov_request() in generic layer. rule_check() and
bdrv_co_pwrite_zeroes() both have 64bit argument.
blklogwrites: pass to blk_log_writes_co_log() with 64bit argument.
blkreplay, copy-on-read, filter-compress: pass to
bdrv_co_pwrite_zeroes() which is OK
copy-before-write: Calls cbw_do_copy_before_write() and
bdrv_co_pwrite_zeroes, both have 64bit argument.
file-posix: both handler calls raw_do_pwrite_zeroes, which is updated.
In raw_do_pwrite_zeroes() calculations are OK due to
bdrv_check_qiov_request(), bytes go to RawPosixAIOData::aio_nbytes
which is uint64_t.
Check also where that uint64_t gets handed:
handle_aiocb_write_zeroes_block() passes a uint64_t[2] to
ioctl(BLKZEROOUT), handle_aiocb_write_zeroes() calls do_fallocate()
which takes off_t (and we compile to always have 64-bit off_t), as
does handle_aiocb_write_zeroes_unmap. All look safe.
gluster: bytes go to GlusterAIOCB::size which is int64_t and to
glfs_zerofill_async works with off_t.
iscsi: Aha, here we deal with iscsi_writesame16_task() that has
uint32_t num_blocks argument and iscsi_writesame16_task() has
uint16_t argument. Make comments, add assertions and clarify
max_pwrite_zeroes calculation.
iscsi_allocmap_() functions already has int64_t argument
is_byte_request_lun_aligned is simple to update, do it.
mirror_top: pass to bdrv_mirror_top_do_write which has uint64_t
argument
nbd: Aha, here we have protocol limitation, and NBDRequest::len is
uint32_t. max_pwrite_zeroes is cleanly set to 32bit value, so we are
OK for now.
nvme: Again, protocol limitation. And no inherent limit for
write-zeroes at all. But from code that calculates cdw12 it's obvious
that we do have limit and alignment. Let's clarify it. Also,
obviously the code is not prepared to handle bytes=0. Let's handle
this case too.
trace events already 64bit
preallocate: pass to handle_write() and bdrv_co_pwrite_zeroes(), both
64bit.
rbd: pass to qemu_rbd_start_co() which is 64bit.
qcow2: offset + bytes and alignment still works good (thanks to
bdrv_check_qiov_request()), so tail calculation is OK
qcow2_subcluster_zeroize() has 64bit argument, should be OK
trace events updated
qed: qed_co_request wants int nb_sectors. Also in code we have size_t
used for request length which may be 32bit. So, let's just keep
INT_MAX as a limit (aligning it down to pwrite_zeroes_alignment) and
don't care.
raw-format: Is OK. raw_adjust_offset and bdrv_co_pwrite_zeroes are both
64bit.
throttle: Both throttle_group_co_io_limits_intercept() and
bdrv_co_pwrite_zeroes() are 64bit.
vmdk: pass to vmdk_pwritev which is 64bit
quorum: pass to quorum_co_pwritev() which is 64bit
Hooray!
At this point all block drivers are prepared to support 64bit
write-zero requests, or have explicitly set max_pwrite_zeroes.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-8-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: use <= rather than < in assertions relying on max_pwrite_zeroes]
Signed-off-by: Eric Blake <eblake@redhat.com>
2021-09-03 13:28:03 +03:00
|
|
|
int64_t offset, int64_t bytes,
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
BdrvRequestFlags flags)
|
|
|
|
{
|
|
|
|
uint32_t align = MAX(bs->bl.request_alignment,
|
|
|
|
bs->bl.pwrite_zeroes_alignment);
|
|
|
|
int err;
|
|
|
|
|
|
|
|
/* Only pass through requests that are larger than requested
|
|
|
|
* preferred alignment (so that we test the fallback to writes on
|
|
|
|
* unaligned portions), and check that the block layer never hands
|
|
|
|
* us anything unaligned that crosses an alignment boundary. */
|
2017-06-09 13:18:08 +03:00
|
|
|
if (bytes < align) {
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
assert(QEMU_IS_ALIGNED(offset, align) ||
|
2017-06-09 13:18:08 +03:00
|
|
|
QEMU_IS_ALIGNED(offset + bytes, align) ||
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
DIV_ROUND_UP(offset, align) ==
|
2017-06-09 13:18:08 +03:00
|
|
|
DIV_ROUND_UP(offset + bytes, align));
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
return -ENOTSUP;
|
|
|
|
}
|
|
|
|
assert(QEMU_IS_ALIGNED(offset, align));
|
2017-06-09 13:18:08 +03:00
|
|
|
assert(QEMU_IS_ALIGNED(bytes, align));
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
if (bs->bl.max_pwrite_zeroes) {
|
2017-06-09 13:18:08 +03:00
|
|
|
assert(bytes <= bs->bl.max_pwrite_zeroes);
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
}
|
|
|
|
|
2019-05-07 23:35:04 +03:00
|
|
|
err = rule_check(bs, offset, bytes, BLKDEBUG_IO_TYPE_WRITE_ZEROES);
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
if (err) {
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2017-06-09 13:18:08 +03:00
|
|
|
return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags);
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
static int coroutine_fn blkdebug_co_pdiscard(BlockDriverState *bs,
|
block: use int64_t instead of int in driver discard handlers
We are generally moving to int64_t for both offset and bytes parameters
on all io paths.
Main motivation is realization of 64-bit write_zeroes operation for
fast zeroing large disk chunks, up to the whole disk.
We chose signed type, to be consistent with off_t (which is signed) and
with possibility for signed return type (where negative value means
error).
So, convert driver discard handlers bytes parameter to int64_t.
The only caller of all updated function is bdrv_co_pdiscard in
block/io.c. It is already prepared to work with 64bit requests, but
pass at most max(bs->bl.max_pdiscard, INT_MAX) to the driver.
Let's look at all updated functions:
blkdebug: all calculations are still OK, thanks to
bdrv_check_qiov_request().
both rule_check and bdrv_co_pdiscard are 64bit
blklogwrites: pass to blk_loc_writes_co_log which is 64bit
blkreplay, copy-on-read, filter-compress: pass to bdrv_co_pdiscard, OK
copy-before-write: pass to bdrv_co_pdiscard which is 64bit and to
cbw_do_copy_before_write which is 64bit
file-posix: one handler calls raw_account_discard() is 64bit and both
handlers calls raw_do_pdiscard(). Update raw_do_pdiscard, which pass
to RawPosixAIOData::aio_nbytes, which is 64bit (and calls
raw_account_discard())
gluster: somehow, third argument of glfs_discard_async is size_t.
Let's set max_pdiscard accordingly.
iscsi: iscsi_allocmap_set_invalid is 64bit,
!is_byte_request_lun_aligned is 64bit.
list.num is uint32_t. Let's clarify max_pdiscard and
pdiscard_alignment.
mirror_top: pass to bdrv_mirror_top_do_write() which is
64bit
nbd: protocol limitation. max_pdiscard is alredy set strict enough,
keep it as is for now.
nvme: buf.nlb is uint32_t and we do shift. So, add corresponding limits
to nvme_refresh_limits().
preallocate: pass to bdrv_co_pdiscard() which is 64bit.
rbd: pass to qemu_rbd_start_co() which is 64bit.
qcow2: calculations are still OK, thanks to bdrv_check_qiov_request(),
qcow2_cluster_discard() is 64bit.
raw-format: raw_adjust_offset() is 64bit, bdrv_co_pdiscard too.
throttle: pass to bdrv_co_pdiscard() which is 64bit and to
throttle_group_co_io_limits_intercept() which is 64bit as well.
test-block-iothread: bytes argument is unused
Great! Now all drivers are prepared to handle 64bit discard requests,
or else have explicit max_pdiscard limits.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210903102807.27127-11-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2021-09-03 13:28:06 +03:00
|
|
|
int64_t offset, int64_t bytes)
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
{
|
|
|
|
uint32_t align = bs->bl.pdiscard_alignment;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
/* Only pass through requests that are larger than requested
|
|
|
|
* minimum alignment, and ensure that unaligned requests do not
|
|
|
|
* cross optimum discard boundaries. */
|
2017-06-09 13:18:08 +03:00
|
|
|
if (bytes < bs->bl.request_alignment) {
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
assert(QEMU_IS_ALIGNED(offset, align) ||
|
2017-06-09 13:18:08 +03:00
|
|
|
QEMU_IS_ALIGNED(offset + bytes, align) ||
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
DIV_ROUND_UP(offset, align) ==
|
2017-06-09 13:18:08 +03:00
|
|
|
DIV_ROUND_UP(offset + bytes, align));
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
return -ENOTSUP;
|
|
|
|
}
|
|
|
|
assert(QEMU_IS_ALIGNED(offset, bs->bl.request_alignment));
|
2017-06-09 13:18:08 +03:00
|
|
|
assert(QEMU_IS_ALIGNED(bytes, bs->bl.request_alignment));
|
|
|
|
if (align && bytes >= align) {
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
assert(QEMU_IS_ALIGNED(offset, align));
|
2017-06-09 13:18:08 +03:00
|
|
|
assert(QEMU_IS_ALIGNED(bytes, align));
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
}
|
|
|
|
if (bs->bl.max_pdiscard) {
|
2017-06-09 13:18:08 +03:00
|
|
|
assert(bytes <= bs->bl.max_pdiscard);
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
}
|
|
|
|
|
2019-05-07 23:35:04 +03:00
|
|
|
err = rule_check(bs, offset, bytes, BLKDEBUG_IO_TYPE_DISCARD);
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
if (err) {
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2018-07-10 09:31:17 +03:00
|
|
|
return bdrv_co_pdiscard(bs->file, offset, bytes);
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
}
|
2012-12-06 17:32:57 +04:00
|
|
|
|
2018-02-13 23:26:43 +03:00
|
|
|
static int coroutine_fn blkdebug_co_block_status(BlockDriverState *bs,
|
|
|
|
bool want_zero,
|
|
|
|
int64_t offset,
|
|
|
|
int64_t bytes,
|
|
|
|
int64_t *pnum,
|
|
|
|
int64_t *map,
|
|
|
|
BlockDriverState **file)
|
2017-10-12 06:47:17 +03:00
|
|
|
{
|
2019-05-07 23:35:06 +03:00
|
|
|
int err;
|
|
|
|
|
2018-02-13 23:26:43 +03:00
|
|
|
assert(QEMU_IS_ALIGNED(offset | bytes, bs->bl.request_alignment));
|
2019-05-07 23:35:06 +03:00
|
|
|
|
|
|
|
err = rule_check(bs, offset, bytes, BLKDEBUG_IO_TYPE_BLOCK_STATUS);
|
|
|
|
if (err) {
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2019-02-13 20:05:23 +03:00
|
|
|
assert(bs->file && bs->file->bs);
|
|
|
|
*pnum = bytes;
|
|
|
|
*map = offset;
|
|
|
|
*file = bs->file->bs;
|
|
|
|
return BDRV_BLOCK_RAW | BDRV_BLOCK_OFFSET_VALID;
|
2017-10-12 06:47:17 +03:00
|
|
|
}
|
|
|
|
|
2010-02-18 19:48:12 +03:00
|
|
|
static void blkdebug_close(BlockDriverState *bs)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
2010-03-15 19:27:00 +03:00
|
|
|
BlkdebugRule *rule, *next;
|
|
|
|
int i;
|
|
|
|
|
qapi: Don't let implicit enum MAX member collide
Now that we guarantee the user doesn't have any enum values
beginning with a single underscore, we can use that for our
own purposes. Renaming ENUM_MAX to ENUM__MAX makes it obvious
that the sentinel is generated.
This patch was mostly generated by applying a temporary patch:
|diff --git a/scripts/qapi.py b/scripts/qapi.py
|index e6d014b..b862ec9 100644
|--- a/scripts/qapi.py
|+++ b/scripts/qapi.py
|@@ -1570,6 +1570,7 @@ const char *const %(c_name)s_lookup[] = {
| max_index = c_enum_const(name, 'MAX', prefix)
| ret += mcgen('''
| [%(max_index)s] = NULL,
|+// %(max_index)s
| };
| ''',
| max_index=max_index)
then running:
$ cat qapi-{types,event}.c tests/test-qapi-types.c |
sed -n 's,^// \(.*\)MAX,s|\1MAX|\1_MAX|g,p' > list
$ git grep -l _MAX | xargs sed -i -f list
The only things not generated are the changes in scripts/qapi.py.
Rejecting enum members named 'MAX' is now useless, and will be dropped
in the next patch.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <1447836791-369-23-git-send-email-eblake@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
[Rebased to current master, commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
2015-11-18 11:52:57 +03:00
|
|
|
for (i = 0; i < BLKDBG__MAX; i++) {
|
2010-03-15 19:27:00 +03:00
|
|
|
QLIST_FOREACH_SAFE(rule, &s->rules[i], next, next) {
|
2012-12-06 17:32:56 +04:00
|
|
|
remove_rule(rule);
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
|
|
|
}
|
2016-08-15 16:29:25 +03:00
|
|
|
|
|
|
|
g_free(s->config_file);
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_destroy(&s->lock);
|
2010-02-18 19:48:12 +03:00
|
|
|
}
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
/* Called with lock held. */
|
2012-12-06 17:32:57 +04:00
|
|
|
static void suspend_request(BlockDriverState *bs, BlkdebugRule *rule)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
2021-06-14 11:29:27 +03:00
|
|
|
BlkdebugSuspendedReq *r;
|
2012-12-06 17:32:57 +04:00
|
|
|
|
2021-06-14 11:29:27 +03:00
|
|
|
r = g_new(BlkdebugSuspendedReq, 1);
|
|
|
|
|
|
|
|
r->co = qemu_coroutine_self();
|
|
|
|
r->tag = g_strdup(rule->options.suspend.tag);
|
2012-12-06 17:32:57 +04:00
|
|
|
|
|
|
|
remove_rule(rule);
|
2021-06-14 11:29:27 +03:00
|
|
|
QLIST_INSERT_HEAD(&s->suspended_reqs, r, next);
|
2012-12-06 17:32:57 +04:00
|
|
|
|
2015-11-30 14:44:44 +03:00
|
|
|
if (!qtest_enabled()) {
|
2021-06-14 11:29:27 +03:00
|
|
|
printf("blkdebug: Suspended request '%s'\n", r->tag);
|
2015-11-30 14:44:44 +03:00
|
|
|
}
|
2012-12-06 17:32:57 +04:00
|
|
|
}
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
/* Called with lock held. */
|
2021-06-14 11:29:28 +03:00
|
|
|
static void process_rule(BlockDriverState *bs, struct BlkdebugRule *rule,
|
2021-06-14 11:29:30 +03:00
|
|
|
int *action_count, int *new_state)
|
2010-03-15 19:27:00 +03:00
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
|
|
|
|
|
|
|
/* Only process rules for the current state */
|
2012-09-28 19:23:00 +04:00
|
|
|
if (rule->state && rule->state != s->state) {
|
2021-06-14 11:29:28 +03:00
|
|
|
return;
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Take the action */
|
2021-06-14 11:29:28 +03:00
|
|
|
action_count[rule->action]++;
|
2010-03-15 19:27:00 +03:00
|
|
|
switch (rule->action) {
|
|
|
|
case ACTION_INJECT_ERROR:
|
2021-06-14 11:29:28 +03:00
|
|
|
if (action_count[ACTION_INJECT_ERROR] == 1) {
|
2012-06-06 10:10:42 +04:00
|
|
|
QSIMPLEQ_INIT(&s->active_rules);
|
|
|
|
}
|
|
|
|
QSIMPLEQ_INSERT_HEAD(&s->active_rules, rule, active_next);
|
2010-03-15 19:27:00 +03:00
|
|
|
break;
|
|
|
|
|
|
|
|
case ACTION_SET_STATE:
|
2021-06-14 11:29:30 +03:00
|
|
|
*new_state = rule->options.set_state.new_state;
|
2010-03-15 19:27:00 +03:00
|
|
|
break;
|
2012-12-06 17:32:57 +04:00
|
|
|
|
|
|
|
case ACTION_SUSPEND:
|
|
|
|
suspend_request(bs, rule);
|
|
|
|
break;
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-11-18 11:52:54 +03:00
|
|
|
static void blkdebug_debug_event(BlockDriverState *bs, BlkdebugEvent event)
|
2010-03-15 19:27:00 +03:00
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
2012-12-06 17:32:57 +04:00
|
|
|
struct BlkdebugRule *rule, *next;
|
2021-06-14 11:29:30 +03:00
|
|
|
int new_state;
|
2021-06-14 11:29:28 +03:00
|
|
|
int actions_count[ACTION__MAX] = { 0 };
|
2010-03-15 19:27:00 +03:00
|
|
|
|
qapi: Don't let implicit enum MAX member collide
Now that we guarantee the user doesn't have any enum values
beginning with a single underscore, we can use that for our
own purposes. Renaming ENUM_MAX to ENUM__MAX makes it obvious
that the sentinel is generated.
This patch was mostly generated by applying a temporary patch:
|diff --git a/scripts/qapi.py b/scripts/qapi.py
|index e6d014b..b862ec9 100644
|--- a/scripts/qapi.py
|+++ b/scripts/qapi.py
|@@ -1570,6 +1570,7 @@ const char *const %(c_name)s_lookup[] = {
| max_index = c_enum_const(name, 'MAX', prefix)
| ret += mcgen('''
| [%(max_index)s] = NULL,
|+// %(max_index)s
| };
| ''',
| max_index=max_index)
then running:
$ cat qapi-{types,event}.c tests/test-qapi-types.c |
sed -n 's,^// \(.*\)MAX,s|\1MAX|\1_MAX|g,p' > list
$ git grep -l _MAX | xargs sed -i -f list
The only things not generated are the changes in scripts/qapi.py.
Rejecting enum members named 'MAX' is now useless, and will be dropped
in the next patch.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <1447836791-369-23-git-send-email-eblake@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
[Rebased to current master, commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
2015-11-18 11:52:57 +03:00
|
|
|
assert((int)event >= 0 && event < BLKDBG__MAX);
|
2010-03-15 19:27:00 +03:00
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
WITH_QEMU_LOCK_GUARD(&s->lock) {
|
|
|
|
new_state = s->state;
|
|
|
|
QLIST_FOREACH_SAFE(rule, &s->rules[event], next, next) {
|
|
|
|
process_rule(bs, rule, actions_count, &new_state);
|
|
|
|
}
|
|
|
|
s->state = new_state;
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
2021-06-14 11:29:29 +03:00
|
|
|
|
|
|
|
while (actions_count[ACTION_SUSPEND] > 0) {
|
|
|
|
qemu_coroutine_yield();
|
|
|
|
actions_count[ACTION_SUSPEND]--;
|
|
|
|
}
|
2010-03-15 19:27:00 +03:00
|
|
|
}
|
|
|
|
|
2012-12-06 17:32:57 +04:00
|
|
|
static int blkdebug_debug_breakpoint(BlockDriverState *bs, const char *event,
|
|
|
|
const char *tag)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
|
|
|
struct BlkdebugRule *rule;
|
2017-08-24 11:46:02 +03:00
|
|
|
int blkdebug_event;
|
2012-12-06 17:32:57 +04:00
|
|
|
|
2017-08-24 11:46:10 +03:00
|
|
|
blkdebug_event = qapi_enum_parse(&BlkdebugEvent_lookup, event, -1, NULL);
|
2017-08-24 11:46:02 +03:00
|
|
|
if (blkdebug_event < 0) {
|
2012-12-06 17:32:57 +04:00
|
|
|
return -ENOENT;
|
|
|
|
}
|
|
|
|
|
|
|
|
rule = g_malloc(sizeof(*rule));
|
|
|
|
*rule = (struct BlkdebugRule) {
|
|
|
|
.event = blkdebug_event,
|
|
|
|
.action = ACTION_SUSPEND,
|
|
|
|
.state = 0,
|
|
|
|
.options.suspend.tag = g_strdup(tag),
|
|
|
|
};
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_lock(&s->lock);
|
2012-12-06 17:32:57 +04:00
|
|
|
QLIST_INSERT_HEAD(&s->rules[blkdebug_event], rule, next);
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_unlock(&s->lock);
|
2012-12-06 17:32:57 +04:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
/* Called with lock held. May temporarily release lock. */
|
2021-06-14 11:29:26 +03:00
|
|
|
static int resume_req_by_tag(BDRVBlkdebugState *s, const char *tag, bool all)
|
2012-12-06 17:32:57 +04:00
|
|
|
{
|
2021-06-14 11:29:26 +03:00
|
|
|
BlkdebugSuspendedReq *r;
|
2012-12-06 17:32:57 +04:00
|
|
|
|
2021-06-14 11:29:26 +03:00
|
|
|
retry:
|
|
|
|
/*
|
|
|
|
* No need for _SAFE, since a different coroutine can remove another node
|
|
|
|
* (not the current one) in this list, and when the current one is removed
|
|
|
|
* the iteration starts back from beginning anyways.
|
|
|
|
*/
|
|
|
|
QLIST_FOREACH(r, &s->suspended_reqs, next) {
|
2012-12-06 17:32:57 +04:00
|
|
|
if (!strcmp(r->tag, tag)) {
|
2021-06-14 11:29:27 +03:00
|
|
|
Coroutine *co = r->co;
|
|
|
|
|
|
|
|
if (!qtest_enabled()) {
|
|
|
|
printf("blkdebug: Resuming request '%s'\n", r->tag);
|
|
|
|
}
|
|
|
|
|
2021-06-14 11:29:26 +03:00
|
|
|
QLIST_REMOVE(r, next);
|
2021-06-14 11:29:27 +03:00
|
|
|
g_free(r->tag);
|
|
|
|
g_free(r);
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_unlock(&s->lock);
|
2021-06-14 11:29:27 +03:00
|
|
|
qemu_coroutine_enter(co);
|
2021-06-14 11:29:31 +03:00
|
|
|
qemu_mutex_lock(&s->lock);
|
2021-06-14 11:29:27 +03:00
|
|
|
|
2021-06-14 11:29:26 +03:00
|
|
|
if (all) {
|
|
|
|
goto retry;
|
|
|
|
}
|
2012-12-06 17:32:57 +04:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return -ENOENT;
|
|
|
|
}
|
|
|
|
|
2021-06-14 11:29:26 +03:00
|
|
|
static int blkdebug_debug_resume(BlockDriverState *bs, const char *tag)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
2021-06-14 11:29:31 +03:00
|
|
|
QEMU_LOCK_GUARD(&s->lock);
|
2021-06-14 11:29:26 +03:00
|
|
|
return resume_req_by_tag(s, tag, false);
|
|
|
|
}
|
|
|
|
|
2013-11-20 06:01:54 +04:00
|
|
|
static int blkdebug_debug_remove_breakpoint(BlockDriverState *bs,
|
|
|
|
const char *tag)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
|
|
|
BlkdebugRule *rule, *next;
|
|
|
|
int i, ret = -ENOENT;
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
QEMU_LOCK_GUARD(&s->lock);
|
qapi: Don't let implicit enum MAX member collide
Now that we guarantee the user doesn't have any enum values
beginning with a single underscore, we can use that for our
own purposes. Renaming ENUM_MAX to ENUM__MAX makes it obvious
that the sentinel is generated.
This patch was mostly generated by applying a temporary patch:
|diff --git a/scripts/qapi.py b/scripts/qapi.py
|index e6d014b..b862ec9 100644
|--- a/scripts/qapi.py
|+++ b/scripts/qapi.py
|@@ -1570,6 +1570,7 @@ const char *const %(c_name)s_lookup[] = {
| max_index = c_enum_const(name, 'MAX', prefix)
| ret += mcgen('''
| [%(max_index)s] = NULL,
|+// %(max_index)s
| };
| ''',
| max_index=max_index)
then running:
$ cat qapi-{types,event}.c tests/test-qapi-types.c |
sed -n 's,^// \(.*\)MAX,s|\1MAX|\1_MAX|g,p' > list
$ git grep -l _MAX | xargs sed -i -f list
The only things not generated are the changes in scripts/qapi.py.
Rejecting enum members named 'MAX' is now useless, and will be dropped
in the next patch.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <1447836791-369-23-git-send-email-eblake@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
[Rebased to current master, commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
2015-11-18 11:52:57 +03:00
|
|
|
for (i = 0; i < BLKDBG__MAX; i++) {
|
2013-11-20 06:01:54 +04:00
|
|
|
QLIST_FOREACH_SAFE(rule, &s->rules[i], next, next) {
|
|
|
|
if (rule->action == ACTION_SUSPEND &&
|
|
|
|
!strcmp(rule->options.suspend.tag, tag)) {
|
|
|
|
remove_rule(rule);
|
|
|
|
ret = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2021-06-14 11:29:26 +03:00
|
|
|
if (resume_req_by_tag(s, tag, true) == 0) {
|
|
|
|
ret = 0;
|
2013-11-20 06:01:54 +04:00
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
2012-12-06 17:32:57 +04:00
|
|
|
|
|
|
|
static bool blkdebug_debug_is_suspended(BlockDriverState *bs, const char *tag)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
|
|
|
BlkdebugSuspendedReq *r;
|
|
|
|
|
2021-06-14 11:29:31 +03:00
|
|
|
QEMU_LOCK_GUARD(&s->lock);
|
2012-12-06 17:32:57 +04:00
|
|
|
QLIST_FOREACH(r, &s->suspended_reqs, next) {
|
|
|
|
if (!strcmp(r->tag, tag)) {
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-06-06 10:10:41 +04:00
|
|
|
static int64_t blkdebug_getlength(BlockDriverState *bs)
|
|
|
|
{
|
2015-06-16 15:19:22 +03:00
|
|
|
return bdrv_getlength(bs->file->bs);
|
2012-06-06 10:10:41 +04:00
|
|
|
}
|
|
|
|
|
2019-02-01 22:29:28 +03:00
|
|
|
static void blkdebug_refresh_filename(BlockDriverState *bs)
|
2014-07-18 22:24:57 +04:00
|
|
|
{
|
2016-08-15 16:29:25 +03:00
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
2014-11-11 12:23:44 +03:00
|
|
|
const QDictEntry *e;
|
2019-02-01 22:29:28 +03:00
|
|
|
int ret;
|
2014-07-18 22:24:57 +04:00
|
|
|
|
2019-02-01 22:29:28 +03:00
|
|
|
if (!bs->file->bs->exact_filename[0]) {
|
2014-07-18 22:24:57 +04:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2019-02-01 22:29:28 +03:00
|
|
|
for (e = qdict_first(bs->full_open_options); e;
|
|
|
|
e = qdict_next(bs->full_open_options, e))
|
|
|
|
{
|
|
|
|
/* Real child options are under "image", but "x-image" may
|
|
|
|
* contain a filename */
|
|
|
|
if (strcmp(qdict_entry_key(e), "config") &&
|
|
|
|
strcmp(qdict_entry_key(e), "image") &&
|
|
|
|
strcmp(qdict_entry_key(e), "x-image") &&
|
|
|
|
strcmp(qdict_entry_key(e), "driver"))
|
|
|
|
{
|
|
|
|
return;
|
2017-06-13 20:20:05 +03:00
|
|
|
}
|
2014-11-11 12:23:44 +03:00
|
|
|
}
|
|
|
|
|
2019-02-01 22:29:28 +03:00
|
|
|
ret = snprintf(bs->exact_filename, sizeof(bs->exact_filename),
|
|
|
|
"blkdebug:%s:%s",
|
|
|
|
s->config_file ?: "", bs->file->bs->exact_filename);
|
|
|
|
if (ret >= sizeof(bs->exact_filename)) {
|
|
|
|
/* An overflow makes the filename unusable, so do not report any */
|
|
|
|
bs->exact_filename[0] = 0;
|
2014-07-18 22:24:57 +04:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-06-24 01:37:13 +03:00
|
|
|
static void blkdebug_refresh_limits(BlockDriverState *bs, Error **errp)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
|
|
|
|
|
|
|
if (s->align) {
|
2016-06-24 01:37:24 +03:00
|
|
|
bs->bl.request_alignment = s->align;
|
2016-06-24 01:37:13 +03:00
|
|
|
}
|
2017-04-29 22:14:18 +03:00
|
|
|
if (s->max_transfer) {
|
|
|
|
bs->bl.max_transfer = s->max_transfer;
|
|
|
|
}
|
|
|
|
if (s->opt_write_zero) {
|
|
|
|
bs->bl.pwrite_zeroes_alignment = s->opt_write_zero;
|
|
|
|
}
|
|
|
|
if (s->max_write_zero) {
|
|
|
|
bs->bl.max_pwrite_zeroes = s->max_write_zero;
|
|
|
|
}
|
|
|
|
if (s->opt_discard) {
|
|
|
|
bs->bl.pdiscard_alignment = s->opt_discard;
|
|
|
|
}
|
|
|
|
if (s->max_discard) {
|
|
|
|
bs->bl.max_pdiscard = s->max_discard;
|
|
|
|
}
|
2016-06-24 01:37:13 +03:00
|
|
|
}
|
|
|
|
|
2015-10-29 17:22:27 +03:00
|
|
|
static int blkdebug_reopen_prepare(BDRVReopenState *reopen_state,
|
|
|
|
BlockReopenQueue *queue, Error **errp)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2019-11-08 15:34:53 +03:00
|
|
|
static void blkdebug_child_perm(BlockDriverState *bs, BdrvChild *c,
|
2020-05-13 14:05:16 +03:00
|
|
|
BdrvChildRole role,
|
2019-11-08 15:34:53 +03:00
|
|
|
BlockReopenQueue *reopen_queue,
|
|
|
|
uint64_t perm, uint64_t shared,
|
|
|
|
uint64_t *nperm, uint64_t *nshared)
|
|
|
|
{
|
|
|
|
BDRVBlkdebugState *s = bs->opaque;
|
|
|
|
|
2020-05-13 14:05:44 +03:00
|
|
|
bdrv_default_perms(bs, c, role, reopen_queue,
|
2020-05-13 14:05:39 +03:00
|
|
|
perm, shared, nperm, nshared);
|
2019-11-08 15:34:53 +03:00
|
|
|
|
|
|
|
*nperm |= s->take_child_perms;
|
|
|
|
*nshared &= ~s->unshare_child_perms;
|
|
|
|
}
|
|
|
|
|
2019-02-01 22:29:25 +03:00
|
|
|
static const char *const blkdebug_strong_runtime_opts[] = {
|
|
|
|
"config",
|
|
|
|
"inject-error.",
|
|
|
|
"set-state.",
|
|
|
|
"align",
|
|
|
|
"max-transfer",
|
|
|
|
"opt-write-zero",
|
|
|
|
"max-write-zero",
|
|
|
|
"opt-discard",
|
|
|
|
"max-discard",
|
|
|
|
|
|
|
|
NULL
|
|
|
|
};
|
|
|
|
|
2010-02-18 19:48:12 +03:00
|
|
|
static BlockDriver bdrv_blkdebug = {
|
2013-04-10 15:37:33 +04:00
|
|
|
.format_name = "blkdebug",
|
|
|
|
.protocol_name = "blkdebug",
|
|
|
|
.instance_size = sizeof(BDRVBlkdebugState),
|
2017-07-13 18:30:27 +03:00
|
|
|
.is_filter = true,
|
2010-02-18 19:48:12 +03:00
|
|
|
|
2013-04-10 15:37:33 +04:00
|
|
|
.bdrv_parse_filename = blkdebug_parse_filename,
|
|
|
|
.bdrv_file_open = blkdebug_open,
|
|
|
|
.bdrv_close = blkdebug_close,
|
2015-10-29 17:22:27 +03:00
|
|
|
.bdrv_reopen_prepare = blkdebug_reopen_prepare,
|
2019-11-08 15:34:53 +03:00
|
|
|
.bdrv_child_perm = blkdebug_child_perm,
|
2016-12-15 14:28:58 +03:00
|
|
|
|
2013-04-10 15:37:33 +04:00
|
|
|
.bdrv_getlength = blkdebug_getlength,
|
2014-07-18 22:24:57 +04:00
|
|
|
.bdrv_refresh_filename = blkdebug_refresh_filename,
|
2016-06-24 01:37:13 +03:00
|
|
|
.bdrv_refresh_limits = blkdebug_refresh_limits,
|
2010-02-18 19:48:12 +03:00
|
|
|
|
2016-11-04 23:13:45 +03:00
|
|
|
.bdrv_co_preadv = blkdebug_co_preadv,
|
|
|
|
.bdrv_co_pwritev = blkdebug_co_pwritev,
|
|
|
|
.bdrv_co_flush_to_disk = blkdebug_co_flush,
|
blkdebug: Add pass-through write_zero and discard support
In order to test the effects of artificial geometry constraints
on operations like write zero or discard, we first need blkdebug
to manage these actions. It also allows us to inject errors on
those operations, just like we can for read/write/flush.
We can also test the contract promised by the block layer; namely,
if a device has specified limits on alignment or maximum size,
then those limits must be obeyed (for now, the blkdebug driver
merely inherits limits from whatever it is wrapping, but the next
patch will further enhance it to allow specific limit overrides).
This patch intentionally refuses to service requests smaller than
the requested alignments; this is because an upcoming patch adds
a qemu-iotest to prove that the block layer is correctly handling
fragmentation, but the test only works if there is a way to tell
the difference at artificial alignment boundaries when blkdebug is
using a larger-than-default alignment. If we let the blkdebug
layer always defer to the underlying layer, which potentially has
a smaller granularity, the iotest will be thwarted.
Tested by setting up an NBD server with export 'foo', then invoking:
$ ./qemu-io
qemu-io> open -o driver=blkdebug blkdebug::nbd://localhost:10809/foo
qemu-io> d 0 15M
qemu-io> w -z 0 15M
Pre-patch, the server never sees the discard (it was silently
eaten by the block layer); post-patch it is passed across the
wire. Likewise, pre-patch the write is always passed with
NBD_WRITE (with 15M of zeroes on the wire), while post-patch
it can utilize NBD_WRITE_ZEROES (for less traffic).
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 20170429191419.30051-7-eblake@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>
2017-04-29 22:14:16 +03:00
|
|
|
.bdrv_co_pwrite_zeroes = blkdebug_co_pwrite_zeroes,
|
|
|
|
.bdrv_co_pdiscard = blkdebug_co_pdiscard,
|
2018-02-13 23:26:43 +03:00
|
|
|
.bdrv_co_block_status = blkdebug_co_block_status,
|
2010-03-15 19:27:00 +03:00
|
|
|
|
2012-12-06 17:32:57 +04:00
|
|
|
.bdrv_debug_event = blkdebug_debug_event,
|
|
|
|
.bdrv_debug_breakpoint = blkdebug_debug_breakpoint,
|
2013-11-20 06:01:54 +04:00
|
|
|
.bdrv_debug_remove_breakpoint
|
|
|
|
= blkdebug_debug_remove_breakpoint,
|
2012-12-06 17:32:57 +04:00
|
|
|
.bdrv_debug_resume = blkdebug_debug_resume,
|
|
|
|
.bdrv_debug_is_suspended = blkdebug_debug_is_suspended,
|
2019-02-01 22:29:25 +03:00
|
|
|
|
|
|
|
.strong_runtime_opts = blkdebug_strong_runtime_opts,
|
2010-02-18 19:48:12 +03:00
|
|
|
};
|
|
|
|
|
|
|
|
static void bdrv_blkdebug_init(void)
|
|
|
|
{
|
|
|
|
bdrv_register(&bdrv_blkdebug);
|
|
|
|
}
|
|
|
|
|
|
|
|
block_init(bdrv_blkdebug_init);
|