forked from ~ljy/RK356X_SDK_RELEASE
blame | history | patch | commit | commitdiff
hc
2023-12-09 b22da3d8526a935aa31e086e63f60ff3246cb61c 
 kernel/drivers/md/dm-raid.c
....@@ -129,7 +129,9 @@
129129 				  CTR_FLAG_RAID10_COPIES | \
130130 				  CTR_FLAG_RAID10_FORMAT | \
131131 				  CTR_FLAG_DELTA_DISKS | \
132
-				  CTR_FLAG_DATA_OFFSET)
132
+				  CTR_FLAG_DATA_OFFSET | \
133
+				  CTR_FLAG_JOURNAL_DEV | \
134
+				  CTR_FLAG_JOURNAL_MODE)
133135 
134136 /* Valid options definitions per raid level... */
135137 
....@@ -209,6 +211,7 @@
209211 #define RT_FLAG_RS_SUSPENDED		5
210212 #define RT_FLAG_RS_IN_SYNC		6
211213 #define RT_FLAG_RS_RESYNCING		7
214
+#define RT_FLAG_RS_GROW			8
212215 
213216 /* Array elements of 64 bit needed for rebuild/failed disk bits */
214217 #define DISKS_ARRAY_ELEMS ((MAX_RAID_DEVICES + (sizeof(uint64_t) * 8 - 1)) / sizeof(uint64_t) / 8)
....@@ -239,7 +242,9 @@
239242 
240243 	struct mddev md;
241244 	struct raid_type *raid_type;
242
-	struct dm_target_callbacks callbacks;
245
+
246
+	sector_t array_sectors;
247
+	sector_t dev_sectors;
243248 
244249 	/* Optional raid4/5/6 journal device */
245250 	struct journal_dev {
....@@ -248,7 +253,7 @@
248253 		int mode;
249254 	} journal_dev;
250255 
251
-	struct raid_dev dev[0];
256
+	struct raid_dev dev[];
252257 };
253258 
254259 static void rs_config_backup(struct raid_set *rs, struct rs_layout *l)
....@@ -616,7 +621,6 @@
616621 
617622 	} else if (algorithm == ALGORITHM_RAID10_FAR) {
618623 		f = copies;
619
-		r = !RAID10_OFFSET;
620624 		if (!test_bit(__CTR_FLAG_RAID10_USE_NEAR_SETS, &rs->ctr_flags))
621625 			r |= RAID10_USE_FAR_SETS;
622626 
....@@ -697,7 +701,7 @@
697701 	struct gendisk *gendisk = dm_disk(dm_table_get_md(rs->ti->table));
698702 
699703 	set_capacity(gendisk, rs->md.array_sectors);
700
-	revalidate_disk(gendisk);
704
+	revalidate_disk_size(gendisk, true);
701705 }
702706 
703707 /*
....@@ -998,12 +1002,13 @@
9981002 static int validate_raid_redundancy(struct raid_set *rs)
9991003 {
10001004 	unsigned int i, rebuild_cnt = 0;
1001
-	unsigned int rebuilds_per_group = 0, copies;
1005
+	unsigned int rebuilds_per_group = 0, copies, raid_disks;
10021006 	unsigned int group_size, last_group_start;
10031007 
1004
-	for (i = 0; i < rs->md.raid_disks; i++)
1005
-		if (!test_bit(In_sync, &rs->dev[i].rdev.flags) ||
1006
-		    !rs->dev[i].rdev.sb_page)
1008
+	for (i = 0; i < rs->raid_disks; i++)
1009
+		if (!test_bit(FirstUse, &rs->dev[i].rdev.flags) &&
1010
+		    ((!test_bit(In_sync, &rs->dev[i].rdev.flags) ||
1011
+		      !rs->dev[i].rdev.sb_page)))
10071012 			rebuild_cnt++;
10081013 
10091014 	switch (rs->md.level) {
....@@ -1043,8 +1048,9 @@
10431048 		 *	    A	 A    B	   B	C
10441049 		 *	    C	 D    D	   E	E
10451050 		 */
1051
+		raid_disks = min(rs->raid_disks, rs->md.raid_disks);
10461052 		if (__is_raid10_near(rs->md.new_layout)) {
1047
-			for (i = 0; i < rs->md.raid_disks; i++) {
1053
+			for (i = 0; i < raid_disks; i++) {
10481054 				if (!(i % copies))
10491055 					rebuilds_per_group = 0;
10501056 				if ((!rs->dev[i].rdev.sb_page ||
....@@ -1067,10 +1073,10 @@
10671073 		 * results in the need to treat the last (potentially larger)
10681074 		 * set differently.
10691075 		 */
1070
-		group_size = (rs->md.raid_disks / copies);
1071
-		last_group_start = (rs->md.raid_disks / group_size) - 1;
1076
+		group_size = (raid_disks / copies);
1077
+		last_group_start = (raid_disks / group_size) - 1;
10721078 		last_group_start *= group_size;
1073
-		for (i = 0; i < rs->md.raid_disks; i++) {
1079
+		for (i = 0; i < raid_disks; i++) {
10741080 			if (!(i % copies) && !(i > last_group_start))
10751081 				rebuilds_per_group = 0;
10761082 			if ((!rs->dev[i].rdev.sb_page ||
....@@ -1585,7 +1591,7 @@
15851591 {
15861592 	int i;
15871593 
1588
-	for (i = 0; i < rs->md.raid_disks; i++) {
1594
+	for (i = 0; i < rs->raid_disks; i++) {
15891595 		struct md_rdev *rdev = &rs->dev[i].rdev;
15901596 
15911597 		if (!test_bit(Journal, &rdev->flags) &&
....@@ -1615,13 +1621,12 @@
16151621 }
16161622 
16171623 /* Calculate the sectors per device and per array used for @rs */
1618
-static int rs_set_dev_and_array_sectors(struct raid_set *rs, bool use_mddev)
1624
+static int rs_set_dev_and_array_sectors(struct raid_set *rs, sector_t sectors, bool use_mddev)
16191625 {
16201626 	int delta_disks;
16211627 	unsigned int data_stripes;
1628
+	sector_t array_sectors = sectors, dev_sectors = sectors;
16221629 	struct mddev *mddev = &rs->md;
1623
-	struct md_rdev *rdev;
1624
-	sector_t array_sectors = rs->ti->len, dev_sectors = rs->ti->len;
16251630 
16261631 	if (use_mddev) {
16271632 		delta_disks = mddev->delta_disks;
....@@ -1656,12 +1661,9 @@
16561661 		/* Striped layouts */
16571662 		array_sectors = (data_stripes + delta_disks) * dev_sectors;
16581663 
1659
-	rdev_for_each(rdev, mddev)
1660
-		if (!test_bit(Journal, &rdev->flags))
1661
-			rdev->sectors = dev_sectors;
1662
-
16631664 	mddev->array_sectors = array_sectors;
16641665 	mddev->dev_sectors = dev_sectors;
1666
+	rs_set_rdev_sectors(rs);
16651667 
16661668 	return _check_data_dev_sectors(rs);
16671669 bad:
....@@ -1670,7 +1672,7 @@
16701672 }
16711673 
16721674 /* Setup recovery on @rs */
1673
-static void __rs_setup_recovery(struct raid_set *rs, sector_t dev_sectors)
1675
+static void rs_setup_recovery(struct raid_set *rs, sector_t dev_sectors)
16741676 {
16751677 	/* raid0 does not recover */
16761678 	if (rs_is_raid0(rs))
....@@ -1691,22 +1693,6 @@
16911693 				     ? MaxSector : dev_sectors;
16921694 }
16931695 
1694
-/* Setup recovery on @rs based on raid type, device size and 'nosync' flag */
1695
-static void rs_setup_recovery(struct raid_set *rs, sector_t dev_sectors)
1696
-{
1697
-	if (!dev_sectors)
1698
-		/* New raid set or 'sync' flag provided */
1699
-		__rs_setup_recovery(rs, 0);
1700
-	else if (dev_sectors == MaxSector)
1701
-		/* Prevent recovery */
1702
-		__rs_setup_recovery(rs, MaxSector);
1703
-	else if (__rdev_sectors(rs) < dev_sectors)
1704
-		/* Grown raid set */
1705
-		__rs_setup_recovery(rs, __rdev_sectors(rs));
1706
-	else
1707
-		__rs_setup_recovery(rs, MaxSector);
1708
-}
1709
-
17101696 static void do_table_event(struct work_struct *ws)
17111697 {
17121698 	struct raid_set *rs = container_of(ws, struct raid_set, md.event_work);
....@@ -1718,13 +1704,6 @@
17181704 		rs_set_capacity(rs);
17191705 	}
17201706 	dm_table_event(rs->ti->table);
1721
-}
1722
-
1723
-static int raid_is_congested(struct dm_target_callbacks *cb, int bits)
1724
-{
1725
-	struct raid_set *rs = container_of(cb, struct raid_set, callbacks);
1726
-
1727
-	return mddev_congested(&rs->md, bits);
17281707 }
17291708 
17301709 /*
....@@ -2366,8 +2345,6 @@
23662345 
23672346 	if (new_devs == rs->raid_disks || !rebuilds) {
23682347 		/* Replace a broken device */
2369
-		if (new_devs == 1 && !rs->delta_disks)
2370
-			;
23712348 		if (new_devs == rs->raid_disks) {
23722349 			DMINFO("Superblocks created for new raid set");
23732350 			set_bit(MD_ARRAY_FIRST_USE, &mddev->flags);
....@@ -2480,7 +2457,7 @@
24802457 		return -EINVAL;
24812458 	}
24822459 
2483
-	/* Enable bitmap creation for RAID levels != 0 */
2460
+	/* Enable bitmap creation on @rs unless no metadevs or raid0 or journaled raid4/5/6 set. */
24842461 	mddev->bitmap_info.offset = (rt_is_raid0(rs->raid_type) || rs->journal_dev.dev) ? 0 : to_sector(4096);
24852462 	mddev->bitmap_info.default_offset = mddev->bitmap_info.offset;
24862463 
....@@ -2917,7 +2894,7 @@
29172894 
29182895 	/* Remove disk(s) */
29192896 	} else if (rs->delta_disks < 0) {
2920
-		r = rs_set_dev_and_array_sectors(rs, true);
2897
+		r = rs_set_dev_and_array_sectors(rs, rs->ti->len, true);
29212898 		mddev->reshape_backwards = 1; /* removing disk(s) -> backward reshape */
29222899 
29232900 	/* Change layout and/or chunk size */
....@@ -3006,11 +2983,6 @@
30062983 		}
30072984 	}
30082985 
3009
-	/*
3010
-	 * RAID1 and RAID10 personalities require bio splitting,
3011
-	 * RAID0/4/5/6 don't and process large discard bios properly.
3012
-	 */
3013
-	ti->split_discard_bios = !!(rs_is_raid1(rs) || rs_is_raid10(rs));
30142986 	ti->num_discard_bios = 1;
30152987 }
30162988 
....@@ -3033,7 +3005,7 @@
30333005 	bool resize = false;
30343006 	struct raid_type *rt;
30353007 	unsigned int num_raid_params, num_raid_devs;
3036
-	sector_t calculated_dev_sectors, rdev_sectors, reshape_sectors;
3008
+	sector_t sb_array_sectors, rdev_sectors, reshape_sectors;
30373009 	struct raid_set *rs = NULL;
30383010 	const char *arg;
30393011 	struct rs_layout rs_layout;
....@@ -3043,7 +3015,6 @@
30433015 		{ 1, 254, "Cannot understand number of raid devices parameters" }
30443016 	};
30453017 
3046
-	/* Must have <raid_type> */
30473018 	arg = dm_shift_arg(&as);
30483019 	if (!arg) {
30493020 		ti->error = "No arguments";
....@@ -3092,11 +3063,13 @@
30923063 	 *
30933064 	 * Any existing superblock will overwrite the array and device sizes
30943065 	 */
3095
-	r = rs_set_dev_and_array_sectors(rs, false);
3066
+	r = rs_set_dev_and_array_sectors(rs, rs->ti->len, false);
30963067 	if (r)
30973068 		goto bad;
30983069 
3099
-	calculated_dev_sectors = rs->md.dev_sectors;
3070
+	/* Memorize just calculated, potentially larger sizes to grow the raid set in preresume */
3071
+	rs->array_sectors = rs->md.array_sectors;
3072
+	rs->dev_sectors = rs->md.dev_sectors;
31003073 
31013074 	/*
31023075 	 * Backup any new raid set level, layout, ...
....@@ -3109,6 +3082,8 @@
31093082 	if (r)
31103083 		goto bad;
31113084 
3085
+	/* All in-core metadata now as of current superblocks after calling analyse_superblocks() */
3086
+	sb_array_sectors = rs->md.array_sectors;
31123087 	rdev_sectors = __rdev_sectors(rs);
31133088 	if (!rdev_sectors) {
31143089 		ti->error = "Invalid rdev size";
....@@ -3118,8 +3093,11 @@
31183093 
31193094 
31203095 	reshape_sectors = _get_reshape_sectors(rs);
3121
-	if (calculated_dev_sectors != rdev_sectors)
3122
-		resize = calculated_dev_sectors != (reshape_sectors ? rdev_sectors - reshape_sectors : rdev_sectors);
3096
+	if (rs->dev_sectors != rdev_sectors) {
3097
+		resize = (rs->dev_sectors != rdev_sectors - reshape_sectors);
3098
+		if (rs->dev_sectors > rdev_sectors - reshape_sectors)
3099
+			set_bit(RT_FLAG_RS_GROW, &rs->runtime_flags);
3100
+	}
31233101 
31243102 	INIT_WORK(&rs->md.event_work, do_table_event);
31253103 	ti->private = rs;
....@@ -3146,13 +3124,8 @@
31463124 		set_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags);
31473125 		rs_set_new(rs);
31483126 	} else if (rs_is_recovering(rs)) {
3149
-		/* Rebuild particular devices */
3150
-		if (test_bit(__CTR_FLAG_REBUILD, &rs->ctr_flags)) {
3151
-			set_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags);
3152
-			rs_setup_recovery(rs, MaxSector);
3153
-		}
31543127 		/* A recovering raid set may be resized */
3155
-		; /* skip setup rs */
3128
+		goto size_check;
31563129 	} else if (rs_is_reshaping(rs)) {
31573130 		/* Have to reject size change request during reshape */
31583131 		if (resize) {
....@@ -3196,6 +3169,9 @@
31963169 		rs_setup_recovery(rs, MaxSector);
31973170 		rs_set_new(rs);
31983171 	} else if (rs_reshape_requested(rs)) {
3172
+		/* Only request grow on raid set size extensions, not on reshapes. */
3173
+		clear_bit(RT_FLAG_RS_GROW, &rs->runtime_flags);
3174
+
31993175 		/*
32003176 		 * No need to check for 'ongoing' takeover here, because takeover
32013177 		 * is an instant operation as oposed to an ongoing reshape.
....@@ -3226,13 +3202,31 @@
32263202 		}
32273203 		rs_set_cur(rs);
32283204 	} else {
3205
+size_check:
32293206 		/* May not set recovery when a device rebuild is requested */
32303207 		if (test_bit(__CTR_FLAG_REBUILD, &rs->ctr_flags)) {
3231
-			rs_setup_recovery(rs, MaxSector);
3208
+			clear_bit(RT_FLAG_RS_GROW, &rs->runtime_flags);
32323209 			set_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags);
3233
-		} else
3234
-			rs_setup_recovery(rs, test_bit(__CTR_FLAG_SYNC, &rs->ctr_flags) ?
3235
-					      0 : (resize ? calculated_dev_sectors : MaxSector));
3210
+			rs_setup_recovery(rs, MaxSector);
3211
+		} else if (test_bit(RT_FLAG_RS_GROW, &rs->runtime_flags)) {
3212
+			/*
3213
+			 * Set raid set to current size, i.e. size as of
3214
+			 * superblocks to grow to larger size in preresume.
3215
+			 */
3216
+			r = rs_set_dev_and_array_sectors(rs, sb_array_sectors, false);
3217
+			if (r)
3218
+				goto bad;
3219
+
3220
+			rs_setup_recovery(rs, rs->md.recovery_cp < rs->md.dev_sectors ? rs->md.recovery_cp : rs->md.dev_sectors);
3221
+		} else {
3222
+			/* This is no size change or it is shrinking, update size and record in superblocks */
3223
+			r = rs_set_dev_and_array_sectors(rs, rs->ti->len, false);
3224
+			if (r)
3225
+				goto bad;
3226
+
3227
+			if (sb_array_sectors > rs->array_sectors)
3228
+				set_bit(RT_FLAG_UPDATE_SBS, &rs->runtime_flags);
3229
+		}
32363230 		rs_set_cur(rs);
32373231 	}
32383232 
....@@ -3267,9 +3261,6 @@
32673261 		mddev_unlock(&rs->md);
32683262 		goto bad_md_start;
32693263 	}
3270
-
3271
-	rs->callbacks.congested_fn = raid_is_congested;
3272
-	dm_table_add_target_callbacks(ti->table, &rs->callbacks);
32733264 
32743265 	/* If raid4/5/6 journal mode explicitly requested (only possible with journal dev) -> set it */
32753266 	if (test_bit(__CTR_FLAG_JOURNAL_MODE, &rs->ctr_flags)) {
....@@ -3330,7 +3321,6 @@
33303321 {
33313322 	struct raid_set *rs = ti->private;
33323323 
3333
-	list_del_init(&rs->callbacks.list);
33343324 	md_stop(&rs->md);
33353325 	raid_set_free(rs);
33363326 }
....@@ -3433,10 +3423,9 @@
34333423 
34343424 /* Helper to return resync/reshape progress for @rs and runtime flags for raid set in sync / resynching */
34353425 static sector_t rs_get_progress(struct raid_set *rs, unsigned long recovery,
3436
-				sector_t resync_max_sectors)
3426
+				enum sync_state state, sector_t resync_max_sectors)
34373427 {
34383428 	sector_t r;
3439
-	enum sync_state state;
34403429 	struct mddev *mddev = &rs->md;
34413430 
34423431 	clear_bit(RT_FLAG_RS_IN_SYNC, &rs->runtime_flags);
....@@ -3447,8 +3436,6 @@
34473436 		set_bit(RT_FLAG_RS_IN_SYNC, &rs->runtime_flags);
34483437 
34493438 	} else {
3450
-		state = decipher_sync_action(mddev, recovery);
3451
-
34523439 		if (state == st_idle && !test_bit(MD_RECOVERY_INTR, &recovery))
34533440 			r = mddev->recovery_cp;
34543441 		else
....@@ -3466,18 +3453,14 @@
34663453 			/*
34673454 			 * In case we are recovering, the array is not in sync
34683455 			 * and health chars should show the recovering legs.
3456
+			 *
3457
+			 * Already retrieved recovery offset from curr_resync_completed above.
34693458 			 */
34703459 			;
3471
-		else if (state == st_resync)
3460
+
3461
+		else if (state == st_resync || state == st_reshape)
34723462 			/*
3473
-			 * If "resync" is occurring, the raid set
3474
-			 * is or may be out of sync hence the health
3475
-			 * characters shall be 'a'.
3476
-			 */
3477
-			set_bit(RT_FLAG_RS_RESYNCING, &rs->runtime_flags);
3478
-		else if (state == st_reshape)
3479
-			/*
3480
-			 * If "reshape" is occurring, the raid set
3463
+			 * If "resync/reshape" is occurring, the raid set
34813464 			 * is or may be out of sync hence the health
34823465 			 * characters shall be 'a'.
34833466 			 */
....@@ -3491,22 +3474,22 @@
34913474 			 */
34923475 			set_bit(RT_FLAG_RS_IN_SYNC, &rs->runtime_flags);
34933476 
3494
-		else {
3495
-			struct md_rdev *rdev;
3496
-
3477
+		else if (test_bit(MD_RECOVERY_NEEDED, &recovery))
34973478 			/*
34983479 			 * We are idle and recovery is needed, prevent 'A' chars race
34993480 			 * caused by components still set to in-sync by constructor.
35003481 			 */
3501
-			if (test_bit(MD_RECOVERY_NEEDED, &recovery))
3502
-				set_bit(RT_FLAG_RS_RESYNCING, &rs->runtime_flags);
3482
+			set_bit(RT_FLAG_RS_RESYNCING, &rs->runtime_flags);
35033483 
3484
+		else {
35043485 			/*
3505
-			 * The raid set may be doing an initial sync, or it may
3506
-			 * be rebuilding individual components.	 If all the
3507
-			 * devices are In_sync, then it is the raid set that is
3508
-			 * being initialized.
3486
+			 * We are idle and the raid set may be doing an initial
3487
+			 * sync, or it may be rebuilding individual components.
3488
+			 * If all the devices are In_sync, then it is the raid set
3489
+			 * that is being initialized.
35093490 			 */
3491
+			struct md_rdev *rdev;
3492
+
35103493 			set_bit(RT_FLAG_RS_IN_SYNC, &rs->runtime_flags);
35113494 			rdev_for_each(rdev, mddev)
35123495 				if (!test_bit(Journal, &rdev->flags) &&
....@@ -3531,15 +3514,14 @@
35313514 {
35323515 	struct raid_set *rs = ti->private;
35333516 	struct mddev *mddev = &rs->md;
3534
-	struct r5conf *conf = mddev->private;
3517
+	struct r5conf *conf = rs_is_raid456(rs) ? mddev->private : NULL;
35353518 	int i, max_nr_stripes = conf ? conf->max_nr_stripes : 0;
35363519 	unsigned long recovery;
35373520 	unsigned int raid_param_cnt = 1; /* at least 1 for chunksize */
35383521 	unsigned int sz = 0;
3539
-	unsigned int rebuild_disks;
3540
-	unsigned int write_mostly_params = 0;
3522
+	unsigned int rebuild_writemostly_count = 0;
35413523 	sector_t progress, resync_max_sectors, resync_mismatches;
3542
-	const char *sync_action;
3524
+	enum sync_state state;
35433525 	struct raid_type *rt;
35443526 
35453527 	switch (type) {
....@@ -3553,14 +3535,14 @@
35533535 
35543536 		/* Access most recent mddev properties for status output */
35553537 		smp_rmb();
3556
-		recovery = rs->md.recovery;
35573538 		/* Get sensible max sectors even if raid set not yet started */
35583539 		resync_max_sectors = test_bit(RT_FLAG_RS_PRERESUMED, &rs->runtime_flags) ?
35593540 				      mddev->resync_max_sectors : mddev->dev_sectors;
3560
-		progress = rs_get_progress(rs, recovery, resync_max_sectors);
3541
+		recovery = rs->md.recovery;
3542
+		state = decipher_sync_action(mddev, recovery);
3543
+		progress = rs_get_progress(rs, recovery, state, resync_max_sectors);
35613544 		resync_mismatches = (mddev->last_sync_action && !strcasecmp(mddev->last_sync_action, "check")) ?
35623545 				    atomic64_read(&mddev->resync_mismatches) : 0;
3563
-		sync_action = sync_str(decipher_sync_action(&rs->md, recovery));
35643546 
35653547 		/* HM FIXME: do we want another state char for raid0? It shows 'D'/'A'/'-' now */
35663548 		for (i = 0; i < rs->raid_disks; i++)
....@@ -3585,10 +3567,10 @@
35853567 		 * v1.5.0+:
35863568 		 *
35873569 		 * Sync action:
3588
-		 *   See Documentation/device-mapper/dm-raid.txt for
3570
+		 *   See Documentation/admin-guide/device-mapper/dm-raid.rst for
35893571 		 *   information on each of these states.
35903572 		 */
3591
-		DMEMIT(" %s", sync_action);
3573
+		DMEMIT(" %s", sync_str(state));
35923574 
35933575 		/*
35943576 		 * v1.5.0+:
....@@ -3621,18 +3603,20 @@
36213603 	case STATUSTYPE_TABLE:
36223604 		/* Report the table line string you would use to construct this raid set */
36233605 
3624
-		/* Calculate raid parameter count */
3625
-		for (i = 0; i < rs->raid_disks; i++)
3626
-			if (test_bit(WriteMostly, &rs->dev[i].rdev.flags))
3627
-				write_mostly_params += 2;
3628
-		rebuild_disks = memweight(rs->rebuild_disks, DISKS_ARRAY_ELEMS * sizeof(*rs->rebuild_disks));
3629
-		raid_param_cnt += rebuild_disks * 2 +
3630
-				  write_mostly_params +
3606
+		/*
3607
+		 * Count any rebuild or writemostly argument pairs and subtract the
3608
+		 * hweight count being added below of any rebuild and writemostly ctr flags.
3609
+		 */
3610
+		for (i = 0; i < rs->raid_disks; i++) {
3611
+			rebuild_writemostly_count += (test_bit(i, (void *) rs->rebuild_disks) ? 2 : 0) +
3612
+						     (test_bit(WriteMostly, &rs->dev[i].rdev.flags) ? 2 : 0);
3613
+		}
3614
+		rebuild_writemostly_count -= (test_bit(__CTR_FLAG_REBUILD, &rs->ctr_flags) ? 2 : 0) +
3615
+					     (test_bit(__CTR_FLAG_WRITE_MOSTLY, &rs->ctr_flags) ? 2 : 0);
3616
+		/* Calculate raid parameter count based on ^ rebuild/writemostly argument counts and ctr flags set. */
3617
+		raid_param_cnt += rebuild_writemostly_count +
36313618 				  hweight32(rs->ctr_flags & CTR_FLAG_OPTIONS_NO_ARGS) +
3632
-				  hweight32(rs->ctr_flags & CTR_FLAG_OPTIONS_ONE_ARG) * 2 +
3633
-				  (test_bit(__CTR_FLAG_JOURNAL_DEV, &rs->ctr_flags) ? 2 : 0) +
3634
-				  (test_bit(__CTR_FLAG_JOURNAL_MODE, &rs->ctr_flags) ? 2 : 0);
3635
-
3619
+				  hweight32(rs->ctr_flags & CTR_FLAG_OPTIONS_ONE_ARG) * 2;
36363620 		/* Emit table line */
36373621 		/* This has to be in the documented order for userspace! */
36383622 		DMEMIT("%s %u %u", rs->raid_type->name, raid_param_cnt, mddev->new_chunk_sectors);
....@@ -3640,11 +3624,10 @@
36403624 			DMEMIT(" %s", dm_raid_arg_name_by_flag(CTR_FLAG_SYNC));
36413625 		if (test_bit(__CTR_FLAG_NOSYNC, &rs->ctr_flags))
36423626 			DMEMIT(" %s", dm_raid_arg_name_by_flag(CTR_FLAG_NOSYNC));
3643
-		if (rebuild_disks)
3627
+		if (test_bit(__CTR_FLAG_REBUILD, &rs->ctr_flags))
36443628 			for (i = 0; i < rs->raid_disks; i++)
3645
-				if (test_bit(rs->dev[i].rdev.raid_disk, (void *) rs->rebuild_disks))
3646
-					DMEMIT(" %s %u", dm_raid_arg_name_by_flag(CTR_FLAG_REBUILD),
3647
-							 rs->dev[i].rdev.raid_disk);
3629
+				if (test_bit(i, (void *) rs->rebuild_disks))
3630
+					DMEMIT(" %s %u", dm_raid_arg_name_by_flag(CTR_FLAG_REBUILD), i);
36483631 		if (test_bit(__CTR_FLAG_DAEMON_SLEEP, &rs->ctr_flags))
36493632 			DMEMIT(" %s %lu", dm_raid_arg_name_by_flag(CTR_FLAG_DAEMON_SLEEP),
36503633 					  mddev->bitmap_info.daemon_sleep);
....@@ -3654,7 +3637,7 @@
36543637 		if (test_bit(__CTR_FLAG_MAX_RECOVERY_RATE, &rs->ctr_flags))
36553638 			DMEMIT(" %s %d", dm_raid_arg_name_by_flag(CTR_FLAG_MAX_RECOVERY_RATE),
36563639 					 mddev->sync_speed_max);
3657
-		if (write_mostly_params)
3640
+		if (test_bit(__CTR_FLAG_WRITE_MOSTLY, &rs->ctr_flags))
36583641 			for (i = 0; i < rs->raid_disks; i++)
36593642 				if (test_bit(WriteMostly, &rs->dev[i].rdev.flags))
36603643 					DMEMIT(" %s %d", dm_raid_arg_name_by_flag(CTR_FLAG_WRITE_MOSTLY),
....@@ -3751,13 +3734,13 @@
37513734 	unsigned int i;
37523735 	int r = 0;
37533736 
3754
-	for (i = 0; !r && i < rs->md.raid_disks; i++)
3755
-		if (rs->dev[i].data_dev)
3756
-			r = fn(ti,
3757
-				 rs->dev[i].data_dev,
3758
-				 0, /* No offset on data devs */
3759
-				 rs->md.dev_sectors,
3760
-				 data);
3737
+	for (i = 0; !r && i < rs->raid_disks; i++) {
3738
+		if (rs->dev[i].data_dev) {
3739
+			r = fn(ti, rs->dev[i].data_dev,
3740
+			       0, /* No offset on data devs */
3741
+			       rs->md.dev_sectors, data);
3742
+		}
3743
+	}
37613744 
37623745 	return r;
37633746 }
....@@ -3765,10 +3748,19 @@
37653748 static void raid_io_hints(struct dm_target *ti, struct queue_limits *limits)
37663749 {
37673750 	struct raid_set *rs = ti->private;
3768
-	unsigned int chunk_size = to_bytes(rs->md.chunk_sectors);
3751
+	unsigned int chunk_size_bytes = to_bytes(rs->md.chunk_sectors);
37693752 
3770
-	blk_limits_io_min(limits, chunk_size);
3771
-	blk_limits_io_opt(limits, chunk_size * mddev_data_stripes(rs));
3753
+	blk_limits_io_min(limits, chunk_size_bytes);
3754
+	blk_limits_io_opt(limits, chunk_size_bytes * mddev_data_stripes(rs));
3755
+
3756
+	/*
3757
+	 * RAID0 and RAID10 personalities require bio splitting,
3758
+	 * RAID1/4/5/6 don't and process large discard bios properly.
3759
+	 */
3760
+	if (rs_is_raid0(rs) || rs_is_raid10(rs)) {
3761
+		limits->discard_granularity = chunk_size_bytes;
3762
+		limits->max_discard_sectors = rs->md.chunk_sectors;
3763
+	}
37723764 }
37733765 
37743766 static void raid_postsuspend(struct dm_target *ti)
....@@ -3802,7 +3794,7 @@
38023794 
38033795 	memset(cleared_failed_devices, 0, sizeof(cleared_failed_devices));
38043796 
3805
-	for (i = 0; i < mddev->raid_disks; i++) {
3797
+	for (i = 0; i < rs->raid_disks; i++) {
38063798 		r = &rs->dev[i].rdev;
38073799 		/* HM FIXME: enhance journal device recovery processing */
38083800 		if (test_bit(Journal, &r->flags))
....@@ -3973,11 +3965,22 @@
39733965 	if (r)
39743966 		return r;
39753967 
3976
-	/* Resize bitmap to adjust to changed region size (aka MD bitmap chunksize) */
3977
-	if (test_bit(RT_FLAG_RS_BITMAP_LOADED, &rs->runtime_flags) && mddev->bitmap &&
3978
-	    mddev->bitmap_info.chunksize != to_bytes(rs->requested_bitmap_chunk_sectors)) {
3979
-		r = md_bitmap_resize(mddev->bitmap, mddev->dev_sectors,
3980
-				     to_bytes(rs->requested_bitmap_chunk_sectors), 0);
3968
+	/* We are extending the raid set size, adjust mddev/md_rdev sizes and set capacity. */
3969
+	if (test_bit(RT_FLAG_RS_GROW, &rs->runtime_flags)) {
3970
+		mddev->array_sectors = rs->array_sectors;
3971
+		mddev->dev_sectors = rs->dev_sectors;
3972
+		rs_set_rdev_sectors(rs);
3973
+		rs_set_capacity(rs);
3974
+	}
3975
+
3976
+	/* Resize bitmap to adjust to changed region size (aka MD bitmap chunksize) or grown device size */
3977
+        if (test_bit(RT_FLAG_RS_BITMAP_LOADED, &rs->runtime_flags) && mddev->bitmap &&
3978
+	    (test_bit(RT_FLAG_RS_GROW, &rs->runtime_flags) ||
3979
+	     (rs->requested_bitmap_chunk_sectors &&
3980
+	       mddev->bitmap_info.chunksize != to_bytes(rs->requested_bitmap_chunk_sectors)))) {
3981
+		int chunksize = to_bytes(rs->requested_bitmap_chunk_sectors) ?: mddev->bitmap_info.chunksize;
3982
+
3983
+		r = md_bitmap_resize(mddev->bitmap, mddev->dev_sectors, chunksize, 0);
39813984 		if (r)
39823985 			DMERR("Failed to resize bitmap");
39833986 	}
....@@ -3986,8 +3989,10 @@
39863989 	/* Be prepared for mddev_resume() in raid_resume() */
39873990 	set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
39883991 	if (mddev->recovery_cp && mddev->recovery_cp < MaxSector) {
3989
-		set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
3992
+		set_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
39903993 		mddev->resync_min = mddev->recovery_cp;
3994
+		if (test_bit(RT_FLAG_RS_GROW, &rs->runtime_flags))
3995
+			mddev->resync_max_sectors = mddev->dev_sectors;
39913996 	}
39923997 
39933998 	/* Check for any reshape request unless new raid set */
....@@ -4035,7 +4040,7 @@
40354040 
40364041 static struct target_type raid_target = {
40374042 	.name = "raid",
4038
-	.version = {1, 14, 0},
4043
+	.version = {1, 15, 1},
40394044 	.module = THIS_MODULE,
40404045 	.ctr = raid_ctr,
40414046 	.dtr = raid_dtr,