All of lore.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Nikos Tsironis <ntsironis@arrikto.com>,
	Mike Snitzer <snitzer@redhat.com>
Subject: [PATCH 5.4 53/80] dm clone metadata: Track exact changes per transaction
Date: Thu, 19 Dec 2019 19:34:45 +0100	[thread overview]
Message-ID: <20191219183124.248718718@linuxfoundation.org> (raw)
In-Reply-To: <20191219183031.278083125@linuxfoundation.org>

From: Nikos Tsironis <ntsironis@arrikto.com>

commit e6a505f3f9fae572fb3ab3bc486e755ac9cef32c upstream.

Extend struct dirty_map with a second bitmap which tracks the exact
regions that were hydrated during the current metadata transaction.

Moreover, fix __flush_dmap() to only commit the metadata of the regions
that were hydrated during the current transaction.

This is required by the following commits to fix a data corruption bug.

Fixes: 7431b7835f55 ("dm: add clone target")
Cc: stable@vger.kernel.org # v5.4+
Signed-off-by: Nikos Tsironis <ntsironis@arrikto.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/md/dm-clone-metadata.c |   90 ++++++++++++++++++++++++++++-------------
 1 file changed, 62 insertions(+), 28 deletions(-)

--- a/drivers/md/dm-clone-metadata.c
+++ b/drivers/md/dm-clone-metadata.c
@@ -67,23 +67,34 @@ struct superblock_disk {
  * To save constantly doing look ups on disk we keep an in core copy of the
  * on-disk bitmap, the region_map.
  *
- * To further reduce metadata I/O overhead we use a second bitmap, the dmap
- * (dirty bitmap), which tracks the dirty words, i.e. longs, of the region_map.
+ * In order to track which regions are hydrated during a metadata transaction,
+ * we use a second set of bitmaps, the dmap (dirty bitmap), which includes two
+ * bitmaps, namely dirty_regions and dirty_words. The dirty_regions bitmap
+ * tracks the regions that got hydrated during the current metadata
+ * transaction. The dirty_words bitmap tracks the dirty words, i.e. longs, of
+ * the dirty_regions bitmap.
+ *
+ * This allows us to precisely track the regions that were hydrated during the
+ * current metadata transaction and update the metadata accordingly, when we
+ * commit the current transaction. This is important because dm-clone should
+ * only commit the metadata of regions that were properly flushed to the
+ * destination device beforehand. Otherwise, in case of a crash, we could end
+ * up with a corrupted dm-clone device.
  *
  * When a region finishes hydrating dm-clone calls
  * dm_clone_set_region_hydrated(), or for discard requests
  * dm_clone_cond_set_range(), which sets the corresponding bits in region_map
  * and dmap.
  *
- * During a metadata commit we scan the dmap for dirty region_map words (longs)
- * and update accordingly the on-disk metadata. Thus, we don't have to flush to
- * disk the whole region_map. We can just flush the dirty region_map words.
+ * During a metadata commit we scan dmap->dirty_words and dmap->dirty_regions
+ * and update the on-disk metadata accordingly. Thus, we don't have to flush to
+ * disk the whole region_map. We can just flush the dirty region_map bits.
  *
- * We use a dirty bitmap, which is smaller than the original region_map, to
- * reduce the amount of memory accesses during a metadata commit. As dm-bitset
- * accesses the on-disk bitmap in 64-bit word granularity, there is no
- * significant benefit in tracking the dirty region_map bits with a smaller
- * granularity.
+ * We use the helper dmap->dirty_words bitmap, which is smaller than the
+ * original region_map, to reduce the amount of memory accesses during a
+ * metadata commit. Moreover, as dm-bitset also accesses the on-disk bitmap in
+ * 64-bit word granularity, the dirty_words bitmap helps us avoid useless disk
+ * accesses.
  *
  * We could update directly the on-disk bitmap, when dm-clone calls either
  * dm_clone_set_region_hydrated() or dm_clone_cond_set_range(), buts this
@@ -92,12 +103,13 @@ struct superblock_disk {
  * e.g., in a hooked overwrite bio's completion routine, and further reduce the
  * I/O completion latency.
  *
- * We maintain two dirty bitmaps. During a metadata commit we atomically swap
- * the currently used dmap with the unused one. This allows the metadata update
- * functions to run concurrently with an ongoing commit.
+ * We maintain two dirty bitmap sets. During a metadata commit we atomically
+ * swap the currently used dmap with the unused one. This allows the metadata
+ * update functions to run concurrently with an ongoing commit.
  */
 struct dirty_map {
 	unsigned long *dirty_words;
+	unsigned long *dirty_regions;
 	unsigned int changed;
 };
 
@@ -461,22 +473,40 @@ static size_t bitmap_size(unsigned long
 	return BITS_TO_LONGS(nr_bits) * sizeof(long);
 }
 
-static int dirty_map_init(struct dm_clone_metadata *cmd)
+static int __dirty_map_init(struct dirty_map *dmap, unsigned long nr_words,
+			    unsigned long nr_regions)
 {
-	cmd->dmap[0].changed = 0;
-	cmd->dmap[0].dirty_words = kvzalloc(bitmap_size(cmd->nr_words), GFP_KERNEL);
+	dmap->changed = 0;
 
-	if (!cmd->dmap[0].dirty_words) {
-		DMERR("Failed to allocate dirty bitmap");
+	dmap->dirty_words = kvzalloc(bitmap_size(nr_words), GFP_KERNEL);
+	if (!dmap->dirty_words)
+		return -ENOMEM;
+
+	dmap->dirty_regions = kvzalloc(bitmap_size(nr_regions), GFP_KERNEL);
+	if (!dmap->dirty_regions) {
+		kvfree(dmap->dirty_words);
 		return -ENOMEM;
 	}
 
-	cmd->dmap[1].changed = 0;
-	cmd->dmap[1].dirty_words = kvzalloc(bitmap_size(cmd->nr_words), GFP_KERNEL);
+	return 0;
+}
 
-	if (!cmd->dmap[1].dirty_words) {
+static void __dirty_map_exit(struct dirty_map *dmap)
+{
+	kvfree(dmap->dirty_words);
+	kvfree(dmap->dirty_regions);
+}
+
+static int dirty_map_init(struct dm_clone_metadata *cmd)
+{
+	if (__dirty_map_init(&cmd->dmap[0], cmd->nr_words, cmd->nr_regions)) {
 		DMERR("Failed to allocate dirty bitmap");
-		kvfree(cmd->dmap[0].dirty_words);
+		return -ENOMEM;
+	}
+
+	if (__dirty_map_init(&cmd->dmap[1], cmd->nr_words, cmd->nr_regions)) {
+		DMERR("Failed to allocate dirty bitmap");
+		__dirty_map_exit(&cmd->dmap[0]);
 		return -ENOMEM;
 	}
 
@@ -487,8 +517,8 @@ static int dirty_map_init(struct dm_clon
 
 static void dirty_map_exit(struct dm_clone_metadata *cmd)
 {
-	kvfree(cmd->dmap[0].dirty_words);
-	kvfree(cmd->dmap[1].dirty_words);
+	__dirty_map_exit(&cmd->dmap[0]);
+	__dirty_map_exit(&cmd->dmap[1]);
 }
 
 static int __load_bitset_in_core(struct dm_clone_metadata *cmd)
@@ -633,21 +663,23 @@ unsigned long dm_clone_find_next_unhydra
 	return find_next_zero_bit(cmd->region_map, cmd->nr_regions, start);
 }
 
-static int __update_metadata_word(struct dm_clone_metadata *cmd, unsigned long word)
+static int __update_metadata_word(struct dm_clone_metadata *cmd,
+				  unsigned long *dirty_regions,
+				  unsigned long word)
 {
 	int r;
 	unsigned long index = word * BITS_PER_LONG;
 	unsigned long max_index = min(cmd->nr_regions, (word + 1) * BITS_PER_LONG);
 
 	while (index < max_index) {
-		if (test_bit(index, cmd->region_map)) {
+		if (test_bit(index, dirty_regions)) {
 			r = dm_bitset_set_bit(&cmd->bitset_info, cmd->bitset_root,
 					      index, &cmd->bitset_root);
-
 			if (r) {
 				DMERR("dm_bitset_set_bit failed");
 				return r;
 			}
+			__clear_bit(index, dirty_regions);
 		}
 		index++;
 	}
@@ -721,7 +753,7 @@ static int __flush_dmap(struct dm_clone_
 		if (word == cmd->nr_words)
 			break;
 
-		r = __update_metadata_word(cmd, word);
+		r = __update_metadata_word(cmd, dmap->dirty_regions, word);
 
 		if (r)
 			return r;
@@ -803,6 +835,7 @@ int dm_clone_set_region_hydrated(struct
 	dmap = cmd->current_dmap;
 
 	__set_bit(word, dmap->dirty_words);
+	__set_bit(region_nr, dmap->dirty_regions);
 	__set_bit(region_nr, cmd->region_map);
 	dmap->changed = 1;
 
@@ -831,6 +864,7 @@ int dm_clone_cond_set_range(struct dm_cl
 		if (!test_bit(region_nr, cmd->region_map)) {
 			word = region_nr / BITS_PER_LONG;
 			__set_bit(word, dmap->dirty_words);
+			__set_bit(region_nr, dmap->dirty_regions);
 			__set_bit(region_nr, cmd->region_map);
 			dmap->changed = 1;
 		}



  parent reply	other threads:[~2019-12-19 18:55 UTC|newest]

Thread overview: 96+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-19 18:33 [PATCH 5.4 00/80] 5.4.6-stable review Greg Kroah-Hartman
2019-12-19 18:33 ` [PATCH 5.4 01/80] USB: Fix incorrect DMA allocations for local memory pool drivers Greg Kroah-Hartman
2019-12-19 18:33 ` [PATCH 5.4 02/80] mmc: block: Make card_busy_detect() a bit more generic Greg Kroah-Hartman
2019-12-19 18:33 ` [PATCH 5.4 03/80] mmc: block: Add CMD13 polling for MMC IOCTLS with R1B response Greg Kroah-Hartman
2019-12-19 18:33 ` [PATCH 5.4 04/80] mmc: core: Drop check for mmc_card_is_removable() in mmc_rescan() Greg Kroah-Hartman
2019-12-19 18:33 ` [PATCH 5.4 05/80] mmc: core: Re-work HW reset for SDIO cards Greg Kroah-Hartman
2019-12-19 18:33 ` [PATCH 5.4 06/80] PCI/switchtec: Read all 64 bits of part_event_bitmap Greg Kroah-Hartman
2019-12-19 18:33 ` [PATCH 5.4 07/80] PCI/PM: Always return devices to D0 when thawing Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 08/80] PCI: pciehp: Avoid returning prematurely from sysfs requests Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 09/80] PCI: Fix Intel ACS quirk UPDCR register address Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 10/80] PCI/MSI: Fix incorrect MSI-X masking on resume Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 11/80] PCI: Do not use bus number zero from EA capability Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 12/80] PCI: rcar: Fix missing MACCTLR register setting in initialization sequence Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 13/80] PCI: Apply Cavium ACS quirk to ThunderX2 and ThunderX3 Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 14/80] PM / QoS: Redefine FREQ_QOS_MAX_DEFAULT_VALUE to S32_MAX Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 15/80] block: fix "check bi_size overflow before merge" Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 16/80] xtensa: use MEMBLOCK_ALLOC_ANYWHERE for KASAN shadow map Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 17/80] gfs2: Multi-block allocations in gfs2_page_mkwrite Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 18/80] gfs2: fix glock reference problem in gfs2_trans_remove_revoke Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 19/80] xtensa: fix TLB sanity checker Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 20/80] xtensa: fix syscall_set_return_value Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 21/80] rpmsg: glink: Set tail pointer to 0 at end of FIFO Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 22/80] rpmsg: glink: Fix reuse intents memory leak issue Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 23/80] rpmsg: glink: Fix use after free in open_ack TIMEOUT case Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 24/80] rpmsg: glink: Put an extra reference during cleanup Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 25/80] rpmsg: glink: Fix rpmsg_register_device err handling Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 26/80] rpmsg: glink: Dont send pending rx_done during remove Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 27/80] rpmsg: glink: Free pending deferred work on remove Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 28/80] cifs: smbd: Return -EAGAIN when transport is reconnecting Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 29/80] cifs: smbd: Only queue work for error recovery on memory registration Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 30/80] cifs: smbd: Add messages on RDMA session destroy and reconnection Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 31/80] cifs: smbd: Return -EINVAL when the number of iovs exceeds SMBDIRECT_MAX_SGE Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 32/80] cifs: smbd: Return -ECONNABORTED when trasnport is not in connected state Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 33/80] cifs: Dont display RDMA transport on reconnect Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 34/80] CIFS: Respect O_SYNC and O_DIRECT flags during reconnect Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 35/80] CIFS: Close open handle after interrupted close Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 36/80] CIFS: Do not miss cancelled OPEN responses Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 37/80] CIFS: Fix NULL pointer dereference in mid callback Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 38/80] cifs: Fix retrieval of DFS referrals in cifs_mount() Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 39/80] ARM: dts: s3c64xx: Fix init order of clock providers Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 40/80] ARM: tegra: Fix FLOW_CTLR_HALT register clobbering by tegra_resume() Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 41/80] vfio/pci: call irq_bypass_unregister_producer() before freeing irq Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 42/80] dma-buf: Fix memory leak in sync_file_merge() Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 43/80] drm/panfrost: Fix a race in panfrost_ioctl_madvise() Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 44/80] drm/panfrost: Fix a BO leak in panfrost_ioctl_mmap_bo() Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 45/80] drm/panfrost: Fix a race in panfrost_gem_free_object() Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 46/80] drm/mgag200: Extract device type from flags Greg Kroah-Hartman
2019-12-19 18:34   ` Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 47/80] drm/mgag200: Store flags from PCI driver data in device structure Greg Kroah-Hartman
2019-12-19 18:34   ` Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 48/80] drm/mgag200: Add workaround for HW that does not support startadd Greg Kroah-Hartman
2019-12-19 18:34   ` Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 49/80] drm/mgag200: Flag all G200 SE A machines as broken wrt <startadd> Greg Kroah-Hartman
2019-12-19 18:34   ` Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 50/80] drm: meson: venc: cvbs: fix CVBS mode matching Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 51/80] dm mpath: remove harmful bio-based optimization Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 52/80] dm btree: increase rebalance threshold in __rebalance2() Greg Kroah-Hartman
2019-12-19 18:34 ` Greg Kroah-Hartman [this message]
2019-12-19 18:34 ` [PATCH 5.4 54/80] dm clone metadata: Use a two phase commit Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 55/80] dm clone: Flush destination device before committing metadata Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 56/80] dm thin metadata: Add support for a pre-commit callback Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 57/80] dm thin: Flush data device before committing metadata Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 58/80] scsi: ufs: Disable autohibern8 feature in Cadence UFS Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 59/80] scsi: iscsi: Fix a potential deadlock in the timeout handler Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 60/80] scsi: qla2xxx: Ignore NULL pointer in tcm_qla2xxx_free_mcmd Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 61/80] scsi: qla2xxx: Initialize free_work before flushing it Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 62/80] scsi: qla2xxx: Added support for MPI and PEP regions for ISP28XX Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 63/80] scsi: qla2xxx: Change discovery state before PLOGI Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 64/80] scsi: qla2xxx: Correctly retrieve and interpret active flash region Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 65/80] scsi: qla2xxx: Fix incorrect SFUB length used for Secure Flash Update MB Cmd Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 66/80] drm/nouveau/kms/nv50-: Call outp_atomic_check_view() before handling PBN Greg Kroah-Hartman
2019-12-19 18:34 ` [PATCH 5.4 67/80] drm/nouveau/kms/nv50-: Store the bpc were using in nv50_head_atom Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 68/80] drm/nouveau/kms/nv50-: Limit MST BPC to 8 Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 69/80] drm/i915/fbc: Disable fbc by default on all glk+ Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 70/80] drm/radeon: fix r1xx/r2xx register checker for POT textures Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 71/80] drm/dp_mst: Correct the bug in drm_dp_update_payload_part1() Greg Kroah-Hartman
2019-12-19 18:35   ` Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 72/80] drm/amd/display: re-enable wait in pipelock, but add timeout Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 73/80] drm/amd/display: add default clocks if not able to fetch them Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 74/80] drm/amdgpu: initialize vm_inv_eng0_sem for gfxhub and mmhub Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 75/80] drm/amdgpu: invalidate mmhub semaphore workaround in gmc9/gmc10 Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 76/80] drm/amdgpu/gfx10: explicitly wait for cp idle after halt/unhalt Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 77/80] drm/amdgpu/gfx10: re-init clear state buffer after gpu reset Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 78/80] drm/i915/gvt: Fix cmd length check for MI_ATOMIC Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 79/80] drm/amdgpu: avoid using invalidate semaphore for picasso Greg Kroah-Hartman
2019-12-19 18:35 ` [PATCH 5.4 80/80] drm/amdgpu: add invalidate semaphore limit for SRIOV and picasso in gmc9 Greg Kroah-Hartman
2019-12-20  4:30 ` [PATCH 5.4 00/80] 5.4.6-stable review shuah
2019-12-20  6:29   ` Greg Kroah-Hartman
2019-12-20  7:22 ` Naresh Kamboju
2019-12-20  8:20   ` Greg Kroah-Hartman
2019-12-20 10:30 ` Jon Hunter
2019-12-20 10:30   ` Jon Hunter
2019-12-20 13:50   ` Greg Kroah-Hartman
2019-12-20 18:50 ` Guenter Roeck
2019-12-20 21:21   ` Greg Kroah-Hartman
2019-12-21  9:36 ` Jeffrin Jose

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191219183124.248718718@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=ntsironis@arrikto.com \
    --cc=snitzer@redhat.com \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.