Commit 6e53636f authored by Heinz Mauelshagen's avatar Heinz Mauelshagen Committed by Mike Snitzer

dm raid: add raid4/5/6 journal write-back support via journal_mode option

Commit 63c32ed4 ("dm raid: add raid4/5/6 journaling support") added
journal support to close the raid4/5/6 "write hole" -- in terms of
writethrough caching.

Introduce a "journal_mode" feature and use the new
r5c_journal_mode_set() API to add support for switching the journal
device's cache mode between write-through (the current default) and
write-back.

NOTE: If the journal device is not layered on resilent storage and it
fails, write-through mode will cause the "write hole" to reoccur.  But
if the journal fails while in write-back mode it will cause data loss
for any dirty cache entries unless resilent storage is used for the
journal.
Signed-off-by: default avatarHeinz Mauelshagen <heinzm@redhat.com>
Signed-off-by: default avatarMike Snitzer <snitzer@redhat.com>
parent 4464e36e
...@@ -170,6 +170,13 @@ The target is named "raid" and it accepts the following parameters: ...@@ -170,6 +170,13 @@ The target is named "raid" and it accepts the following parameters:
Takeover/reshape is not possible with a raid4/5/6 journal device; Takeover/reshape is not possible with a raid4/5/6 journal device;
it has to be deconfigured before requesting these. it has to be deconfigured before requesting these.
[journal_mode <mode>]
This option sets the caching mode on journaled raid4/5/6 raid sets
(see 'journal_dev <dev>' above) to 'writethrough' or 'writeback'.
If 'writeback' is selected the journal device has to be resilient
and must not suffer from the 'write hole' problem itself (e.g. use
raid1 or raid10) to avoid a single point of failure.
<#raid_devs>: The number of devices composing the array. <#raid_devs>: The number of devices composing the array.
Each device consists of two entries. The first is the device Each device consists of two entries. The first is the device
containing the metadata (if any); the second is the one containing the containing the metadata (if any); the second is the one containing the
...@@ -254,7 +261,8 @@ recovery. Here is a fuller description of the individual fields: ...@@ -254,7 +261,8 @@ recovery. Here is a fuller description of the individual fields:
<data_offset> The current data offset to the start of the user data on <data_offset> The current data offset to the start of the user data on
each component device of a raid set (see the respective each component device of a raid set (see the respective
raid parameter to support out-of-place reshaping). raid parameter to support out-of-place reshaping).
<journal_char> 'A' - active raid4/5/6 journal device. <journal_char> 'A' - active write-through journal device.
'a' - active write-back journal device.
'D' - dead journal device. 'D' - dead journal device.
'-' - no journal device. '-' - no journal device.
...@@ -334,3 +342,4 @@ Version History ...@@ -334,3 +342,4 @@ Version History
1.10.1 Fix data corruption on reshape request 1.10.1 Fix data corruption on reshape request
1.11.0 Fix table line argument order 1.11.0 Fix table line argument order
(wrong raid10_copies/raid10_format sequence) (wrong raid10_copies/raid10_format sequence)
1.11.1 Add raid4/5/6 journal write-back support via journal_mode option
/* /*
* Copyright (C) 2010-2011 Neil Brown * Copyright (C) 2010-2011 Neil Brown
* Copyright (C) 2010-2016 Red Hat, Inc. All rights reserved. * Copyright (C) 2010-2017 Red Hat, Inc. All rights reserved.
* *
* This file is released under the GPL. * This file is released under the GPL.
*/ */
...@@ -79,7 +79,10 @@ struct raid_dev { ...@@ -79,7 +79,10 @@ struct raid_dev {
#define __CTR_FLAG_RAID10_USE_NEAR_SETS 14 /* 2 */ /* Only with raid10! */ #define __CTR_FLAG_RAID10_USE_NEAR_SETS 14 /* 2 */ /* Only with raid10! */
/* New for v1.10.0 */ /* New for v1.10.0 */
#define __CTR_FLAG_JOURNAL_DEV 15 /* 2 */ /* Only with raid4/5/6! */ #define __CTR_FLAG_JOURNAL_DEV 15 /* 2 */ /* Only with raid4/5/6 (journal device)! */
/* New for v1.11.1 */
#define __CTR_FLAG_JOURNAL_MODE 16 /* 2 */ /* Only with raid4/5/6 (journal mode)! */
/* /*
* Flags for rs->ctr_flags field. * Flags for rs->ctr_flags field.
...@@ -100,6 +103,7 @@ struct raid_dev { ...@@ -100,6 +103,7 @@ struct raid_dev {
#define CTR_FLAG_DATA_OFFSET (1 << __CTR_FLAG_DATA_OFFSET) #define CTR_FLAG_DATA_OFFSET (1 << __CTR_FLAG_DATA_OFFSET)
#define CTR_FLAG_RAID10_USE_NEAR_SETS (1 << __CTR_FLAG_RAID10_USE_NEAR_SETS) #define CTR_FLAG_RAID10_USE_NEAR_SETS (1 << __CTR_FLAG_RAID10_USE_NEAR_SETS)
#define CTR_FLAG_JOURNAL_DEV (1 << __CTR_FLAG_JOURNAL_DEV) #define CTR_FLAG_JOURNAL_DEV (1 << __CTR_FLAG_JOURNAL_DEV)
#define CTR_FLAG_JOURNAL_MODE (1 << __CTR_FLAG_JOURNAL_MODE)
#define RESUME_STAY_FROZEN_FLAGS (CTR_FLAG_DELTA_DISKS | CTR_FLAG_DATA_OFFSET) #define RESUME_STAY_FROZEN_FLAGS (CTR_FLAG_DELTA_DISKS | CTR_FLAG_DATA_OFFSET)
...@@ -175,7 +179,8 @@ struct raid_dev { ...@@ -175,7 +179,8 @@ struct raid_dev {
CTR_FLAG_REGION_SIZE | \ CTR_FLAG_REGION_SIZE | \
CTR_FLAG_DELTA_DISKS | \ CTR_FLAG_DELTA_DISKS | \
CTR_FLAG_DATA_OFFSET | \ CTR_FLAG_DATA_OFFSET | \
CTR_FLAG_JOURNAL_DEV) CTR_FLAG_JOURNAL_DEV | \
CTR_FLAG_JOURNAL_MODE)
#define RAID6_VALID_FLAGS (CTR_FLAG_SYNC | \ #define RAID6_VALID_FLAGS (CTR_FLAG_SYNC | \
CTR_FLAG_REBUILD | \ CTR_FLAG_REBUILD | \
...@@ -186,7 +191,8 @@ struct raid_dev { ...@@ -186,7 +191,8 @@ struct raid_dev {
CTR_FLAG_REGION_SIZE | \ CTR_FLAG_REGION_SIZE | \
CTR_FLAG_DELTA_DISKS | \ CTR_FLAG_DELTA_DISKS | \
CTR_FLAG_DATA_OFFSET | \ CTR_FLAG_DATA_OFFSET | \
CTR_FLAG_JOURNAL_DEV) CTR_FLAG_JOURNAL_DEV | \
CTR_FLAG_JOURNAL_MODE)
/* ...valid options definitions per raid level */ /* ...valid options definitions per raid level */
/* /*
...@@ -239,6 +245,7 @@ struct raid_set { ...@@ -239,6 +245,7 @@ struct raid_set {
struct journal_dev { struct journal_dev {
struct dm_dev *dev; struct dm_dev *dev;
struct md_rdev rdev; struct md_rdev rdev;
int mode;
} journal_dev; } journal_dev;
struct raid_dev dev[0]; struct raid_dev dev[0];
...@@ -326,6 +333,7 @@ static struct arg_name_flag { ...@@ -326,6 +333,7 @@ static struct arg_name_flag {
{ CTR_FLAG_DELTA_DISKS, "delta_disks"}, { CTR_FLAG_DELTA_DISKS, "delta_disks"},
{ CTR_FLAG_RAID10_USE_NEAR_SETS, "raid10_use_near_sets"}, { CTR_FLAG_RAID10_USE_NEAR_SETS, "raid10_use_near_sets"},
{ CTR_FLAG_JOURNAL_DEV, "journal_dev" }, { CTR_FLAG_JOURNAL_DEV, "journal_dev" },
{ CTR_FLAG_JOURNAL_MODE, "journal_mode" },
}; };
/* Return argument name string for given @flag */ /* Return argument name string for given @flag */
...@@ -344,6 +352,39 @@ static const char *dm_raid_arg_name_by_flag(const uint32_t flag) ...@@ -344,6 +352,39 @@ static const char *dm_raid_arg_name_by_flag(const uint32_t flag)
return NULL; return NULL;
} }
/* Define correlation of raid456 journal cache modes and dm-raid target line parameters */
static struct {
const int mode;
const char *param;
} _raid456_journal_mode[] = {
{ R5C_JOURNAL_MODE_WRITE_THROUGH , "writethrough" },
{ R5C_JOURNAL_MODE_WRITE_BACK , "writeback" }
};
/* Return MD raid4/5/6 journal mode for dm @journal_mode one */
static int dm_raid_journal_mode_to_md(const char *mode)
{
int m = ARRAY_SIZE(_raid456_journal_mode);
while (m--)
if (!strcasecmp(mode, _raid456_journal_mode[m].param))
return _raid456_journal_mode[m].mode;
return -EINVAL;
}
/* Return dm-raid raid4/5/6 journal mode string for @mode */
static const char *md_journal_mode_to_dm_raid(const int mode)
{
int m = ARRAY_SIZE(_raid456_journal_mode);
while (m--)
if (mode == _raid456_journal_mode[m].mode)
return _raid456_journal_mode[m].param;
return "unknown";
}
/* /*
* Bool helpers to test for various raid levels of a raid set. * Bool helpers to test for various raid levels of a raid set.
* It's level as reported by the superblock rather than * It's level as reported by the superblock rather than
...@@ -1183,7 +1224,7 @@ static int parse_raid_params(struct raid_set *rs, struct dm_arg_set *as, ...@@ -1183,7 +1224,7 @@ static int parse_raid_params(struct raid_set *rs, struct dm_arg_set *as,
continue; continue;
} }
/* "journal_dev dev" */ /* "journal_dev <dev>" */
if (!strcasecmp(key, dm_raid_arg_name_by_flag(CTR_FLAG_JOURNAL_DEV))) { if (!strcasecmp(key, dm_raid_arg_name_by_flag(CTR_FLAG_JOURNAL_DEV))) {
int r; int r;
struct md_rdev *jdev; struct md_rdev *jdev;
...@@ -1211,10 +1252,32 @@ static int parse_raid_params(struct raid_set *rs, struct dm_arg_set *as, ...@@ -1211,10 +1252,32 @@ static int parse_raid_params(struct raid_set *rs, struct dm_arg_set *as,
rs->ti->error = "No space for raid4/5/6 journal"; rs->ti->error = "No space for raid4/5/6 journal";
return -ENOSPC; return -ENOSPC;
} }
rs->journal_dev.mode = R5C_JOURNAL_MODE_WRITE_THROUGH;
set_bit(Journal, &jdev->flags); set_bit(Journal, &jdev->flags);
continue; continue;
} }
/* "journal_mode <mode>" ("journal_dev" mandatory!) */
if (!strcasecmp(key, dm_raid_arg_name_by_flag(CTR_FLAG_JOURNAL_MODE))) {
int r;
if (!test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags)) {
rs->ti->error = "raid4/5/6 'journal_mode' is invalid without 'journal_dev'";
return -EINVAL;
}
if (test_and_set_bit(__CTR_FLAG_JOURNAL_MODE, &rs->ctr_flags)) {
rs->ti->error = "Only one raid4/5/6 'journal_mode' argument allowed";
return -EINVAL;
}
r = dm_raid_journal_mode_to_md(arg);
if (r < 0) {
rs->ti->error = "Invalid 'journal_mode' argument";
return r;
}
rs->journal_dev.mode = r;
continue;
}
/* /*
* Parameters with number values from here on. * Parameters with number values from here on.
*/ */
...@@ -3076,6 +3139,16 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv) ...@@ -3076,6 +3139,16 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv)
rs->callbacks.congested_fn = raid_is_congested; rs->callbacks.congested_fn = raid_is_congested;
dm_table_add_target_callbacks(ti->table, &rs->callbacks); dm_table_add_target_callbacks(ti->table, &rs->callbacks);
/* If raid4/5/6 journal mode explictely requested (only possible with journal dev) -> set it */
if (test_bit(__CTR_FLAG_JOURNAL_MODE, &rs->ctr_flags)) {
r = r5c_journal_mode_set(&rs->md, rs->journal_dev.mode);
if (r) {
ti->error = "Failed to set raid4/5/6 journal mode";
mddev_unlock(&rs->md);
goto bad_journal_mode_set;
}
}
mddev_suspend(&rs->md); mddev_suspend(&rs->md);
/* Try to adjust the raid4/5/6 stripe cache size to the stripe size */ /* Try to adjust the raid4/5/6 stripe cache size to the stripe size */
...@@ -3109,6 +3182,7 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv) ...@@ -3109,6 +3182,7 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv)
mddev_unlock(&rs->md); mddev_unlock(&rs->md);
return 0; return 0;
bad_journal_mode_set:
bad_stripe_cache: bad_stripe_cache:
bad_check_reshape: bad_check_reshape:
md_stop(&rs->md); md_stop(&rs->md);
...@@ -3180,18 +3254,18 @@ static const char *decipher_sync_action(struct mddev *mddev) ...@@ -3180,18 +3254,18 @@ static const char *decipher_sync_action(struct mddev *mddev)
* Status characters: * Status characters:
* *
* 'D' = Dead/Failed raid set component or raid4/5/6 journal device * 'D' = Dead/Failed raid set component or raid4/5/6 journal device
* 'a' = Alive but not in-sync * 'a' = Alive but not in-sync raid set component _or_ alive raid4/5/6 'write_back' journal device
* 'A' = Alive and in-sync raid set component or alive raid4/5/6 journal device * 'A' = Alive and in-sync raid set component _or_ alive raid4/5/6 'write_through' journal device
* '-' = Non-existing device (i.e. uspace passed '- -' into the ctr) * '-' = Non-existing device (i.e. uspace passed '- -' into the ctr)
*/ */
static const char *__raid_dev_status(struct md_rdev *rdev, bool array_in_sync) static const char *__raid_dev_status(struct raid_set *rs, struct md_rdev *rdev, bool array_in_sync)
{ {
if (!rdev->bdev) if (!rdev->bdev)
return "-"; return "-";
else if (test_bit(Faulty, &rdev->flags)) else if (test_bit(Faulty, &rdev->flags))
return "D"; return "D";
else if (test_bit(Journal, &rdev->flags)) else if (test_bit(Journal, &rdev->flags))
return "A"; return (rs->journal_dev.mode == R5C_JOURNAL_MODE_WRITE_THROUGH) ? "A" : "a";
else if (!array_in_sync || !test_bit(In_sync, &rdev->flags)) else if (!array_in_sync || !test_bit(In_sync, &rdev->flags))
return "a"; return "a";
else else
...@@ -3315,7 +3389,7 @@ static void raid_status(struct dm_target *ti, status_type_t type, ...@@ -3315,7 +3389,7 @@ static void raid_status(struct dm_target *ti, status_type_t type,
/* HM FIXME: do we want another state char for raid0? It shows 'D'/'A'/'-' now */ /* HM FIXME: do we want another state char for raid0? It shows 'D'/'A'/'-' now */
for (i = 0; i < rs->raid_disks; i++) for (i = 0; i < rs->raid_disks; i++)
DMEMIT(__raid_dev_status(&rs->dev[i].rdev, array_in_sync)); DMEMIT(__raid_dev_status(rs, &rs->dev[i].rdev, array_in_sync));
/* /*
* In-sync/Reshape ratio: * In-sync/Reshape ratio:
...@@ -3366,7 +3440,7 @@ static void raid_status(struct dm_target *ti, status_type_t type, ...@@ -3366,7 +3440,7 @@ static void raid_status(struct dm_target *ti, status_type_t type,
* v1.10.0+: * v1.10.0+:
*/ */
DMEMIT(" %s", test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags) ? DMEMIT(" %s", test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags) ?
__raid_dev_status(&rs->journal_dev.rdev, 0) : "-"); __raid_dev_status(rs, &rs->journal_dev.rdev, 0) : "-");
break; break;
case STATUSTYPE_TABLE: case STATUSTYPE_TABLE:
...@@ -3381,7 +3455,8 @@ static void raid_status(struct dm_target *ti, status_type_t type, ...@@ -3381,7 +3455,8 @@ static void raid_status(struct dm_target *ti, status_type_t type,
write_mostly_params + write_mostly_params +
hweight32(rs->ctr_flags & CTR_FLAG_OPTIONS_NO_ARGS) + hweight32(rs->ctr_flags & CTR_FLAG_OPTIONS_NO_ARGS) +
hweight32(rs->ctr_flags & CTR_FLAG_OPTIONS_ONE_ARG) * 2 + hweight32(rs->ctr_flags & CTR_FLAG_OPTIONS_ONE_ARG) * 2 +
(test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags) ? 2 : 0); (test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags) ? 2 : 0) +
(test_bit(__CTR_FLAG_JOURNAL_MODE, &rs->ctr_flags) ? 2 : 0);
/* Emit table line */ /* Emit table line */
/* This has to be in the documented order for userspace! */ /* This has to be in the documented order for userspace! */
...@@ -3433,6 +3508,9 @@ static void raid_status(struct dm_target *ti, status_type_t type, ...@@ -3433,6 +3508,9 @@ static void raid_status(struct dm_target *ti, status_type_t type,
if (test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags)) if (test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags))
DMEMIT(" %s %s", dm_raid_arg_name_by_flag(CTR_FLAG_JOURNAL_DEV), DMEMIT(" %s %s", dm_raid_arg_name_by_flag(CTR_FLAG_JOURNAL_DEV),
__get_dev_name(rs->journal_dev.dev)); __get_dev_name(rs->journal_dev.dev));
if (test_bit(__CTR_FLAG_JOURNAL_MODE, &rs->ctr_flags))
DMEMIT(" %s %s", dm_raid_arg_name_by_flag(CTR_FLAG_JOURNAL_MODE),
md_journal_mode_to_dm_raid(rs->journal_dev.mode));
DMEMIT(" %d", rs->raid_disks); DMEMIT(" %d", rs->raid_disks);
for (i = 0; i < rs->raid_disks; i++) for (i = 0; i < rs->raid_disks; i++)
DMEMIT(" %s %s", __get_dev_name(rs->dev[i].meta_dev), DMEMIT(" %s %s", __get_dev_name(rs->dev[i].meta_dev),
...@@ -3793,7 +3871,7 @@ static void raid_resume(struct dm_target *ti) ...@@ -3793,7 +3871,7 @@ static void raid_resume(struct dm_target *ti)
static struct target_type raid_target = { static struct target_type raid_target = {
.name = "raid", .name = "raid",
.version = {1, 11, 0}, .version = {1, 11, 1},
.module = THIS_MODULE, .module = THIS_MODULE,
.ctr = raid_ctr, .ctr = raid_ctr,
.dtr = raid_dtr, .dtr = raid_dtr,
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment