e778d07c4c
Resolves: RHEL-86676, RHEL-72803, RHEL-88793, RHEL-88791 Signed-off-by: Xiao Ni <xni@redhat.com>
207 lines
8.1 KiB
Diff
207 lines
8.1 KiB
Diff
From 17fed47a64e1890df9820b93548c396b7de54e31 Mon Sep 17 00:00:00 2001
|
|
From: Mariusz Tkaczyk <mtkaczyk@kernel.org>
|
|
Date: Wed, 5 Feb 2025 11:34:45 +0100
|
|
Subject: [PATCH 21/37] bitmap.h: Minor fixes
|
|
|
|
Move documentation to documentation/bitmap.md. Add Neil's copyrights,
|
|
add missing license. Remove unused macros.
|
|
|
|
Signed-off-by: Mariusz Tkaczyk <mtkaczyk@kernel.org>
|
|
---
|
|
bitmap.h | 94 ++++-------------------------------------
|
|
documentation/bitmap.md | 67 +++++++++++++++++++++++++++++
|
|
2 files changed, 75 insertions(+), 86 deletions(-)
|
|
create mode 100644 documentation/bitmap.md
|
|
|
|
diff --git a/bitmap.h b/bitmap.h
|
|
index 2614a14e..9f3d4f3e 100644
|
|
--- a/bitmap.h
|
|
+++ b/bitmap.h
|
|
@@ -1,83 +1,19 @@
|
|
+// SPDX-License-Identifier: GPL-2.0-only
|
|
+
|
|
/*
|
|
- * bitmap.h: Copyright (C) Peter T. Breuer (ptb@ot.uc3m.es) 2003
|
|
- *
|
|
- * additions: Copyright (C) 2003-2004, Paul Clements, SteelEye Technology, Inc.
|
|
+ * Copyright (C) Peter T. Breuer (ptb@ot.uc3m.es) 2003
|
|
+ * Copyright (C) 2003-2004, Paul Clements, SteelEye Technology, Inc.
|
|
+ * Copyright (C) 2005 Neil Brown <neilb@suse.com>
|
|
*/
|
|
+
|
|
+/* See documentation/bitmap.md */
|
|
+
|
|
#ifndef BITMAP_H
|
|
#define BITMAP_H 1
|
|
|
|
#define BITMAP_MAJOR_LO 3
|
|
-/* version 4 insists the bitmap is in little-endian order
|
|
- * with version 3, it is host-endian which is non-portable
|
|
- */
|
|
#define BITMAP_MAJOR_HI 4
|
|
-#define BITMAP_MAJOR_HOSTENDIAN 3
|
|
#define BITMAP_MAJOR_CLUSTERED 5
|
|
-
|
|
-#define BITMAP_MINOR 39
|
|
-
|
|
-/*
|
|
- * in-memory bitmap:
|
|
- *
|
|
- * Use 16 bit block counters to track pending writes to each "chunk".
|
|
- * The 2 high order bits are special-purpose, the first is a flag indicating
|
|
- * whether a resync is needed. The second is a flag indicating whether a
|
|
- * resync is active.
|
|
- * This means that the counter is actually 14 bits:
|
|
- *
|
|
- * +--------+--------+------------------------------------------------+
|
|
- * | resync | resync | counter |
|
|
- * | needed | active | |
|
|
- * | (0-1) | (0-1) | (0-16383) |
|
|
- * +--------+--------+------------------------------------------------+
|
|
- *
|
|
- * The "resync needed" bit is set when:
|
|
- * a '1' bit is read from storage at startup.
|
|
- * a write request fails on some drives
|
|
- * a resync is aborted on a chunk with 'resync active' set
|
|
- * It is cleared (and resync-active set) when a resync starts across all drives
|
|
- * of the chunk.
|
|
- *
|
|
- *
|
|
- * The "resync active" bit is set when:
|
|
- * a resync is started on all drives, and resync_needed is set.
|
|
- * resync_needed will be cleared (as long as resync_active wasn't already set).
|
|
- * It is cleared when a resync completes.
|
|
- *
|
|
- * The counter counts pending write requests, plus the on-disk bit.
|
|
- * When the counter is '1' and the resync bits are clear, the on-disk
|
|
- * bit can be cleared as well, thus setting the counter to 0.
|
|
- * When we set a bit, or in the counter (to start a write), if the fields is
|
|
- * 0, we first set the disk bit and set the counter to 1.
|
|
- *
|
|
- * If the counter is 0, the on-disk bit is clear and the stipe is clean
|
|
- * Anything that dirties the stipe pushes the counter to 2 (at least)
|
|
- * and sets the on-disk bit (lazily).
|
|
- * If a periodic sweep find the counter at 2, it is decremented to 1.
|
|
- * If the sweep find the counter at 1, the on-disk bit is cleared and the
|
|
- * counter goes to zero.
|
|
- *
|
|
- * Also, we'll hijack the "map" pointer itself and use it as two 16 bit block
|
|
- * counters as a fallback when "page" memory cannot be allocated:
|
|
- *
|
|
- * Normal case (page memory allocated):
|
|
- *
|
|
- * page pointer (32-bit)
|
|
- *
|
|
- * [ ] ------+
|
|
- * |
|
|
- * +-------> [ ][ ]..[ ] (4096 byte page == 2048 counters)
|
|
- * c1 c2 c2048
|
|
- *
|
|
- * Hijacked case (page memory allocation failed):
|
|
- *
|
|
- * hijacked page pointer (32-bit)
|
|
- *
|
|
- * [ ][ ] (no page memory allocated)
|
|
- * counter #1 (16-bit) counter #2 (16-bit)
|
|
- *
|
|
- */
|
|
-
|
|
#define BITMAP_MAGIC 0x6d746962
|
|
|
|
/* use these for bitmap->flags and bitmap->sb->state bit-fields */
|
|
@@ -105,18 +41,4 @@ typedef struct bitmap_super_s {
|
|
__u8 pad[256 - 136]; /* set to zero */
|
|
} bitmap_super_t;
|
|
|
|
-/* notes:
|
|
- * (1) This event counter is updated before the eventcounter in the md superblock
|
|
- * When a bitmap is loaded, it is only accepted if this event counter is equal
|
|
- * to, or one greater than, the event counter in the superblock.
|
|
- * (2) This event counter is updated when the other one is *if*and*only*if* the
|
|
- * array is not degraded. As bits are not cleared when the array is degraded,
|
|
- * this represents the last time that any bits were cleared.
|
|
- * If a device is being added that has an event count with this value or
|
|
- * higher, it is accepted as conforming to the bitmap.
|
|
- * (3)This is the number of sectors represented by the bitmap, and is the range that
|
|
- * resync happens across. For raid1 and raid5/6 it is the size of individual
|
|
- * devices. For raid10 it is the size of the array.
|
|
- */
|
|
-
|
|
#endif
|
|
diff --git a/documentation/bitmap.md b/documentation/bitmap.md
|
|
new file mode 100644
|
|
index 00000000..bdc146e5
|
|
--- /dev/null
|
|
+++ b/documentation/bitmap.md
|
|
@@ -0,0 +1,67 @@
|
|
+# Internal bitmap block design
|
|
+
|
|
+Use 16 bit block counters to track pending writes to each "chunk".
|
|
+The 2 high order bits are special-purpose, the first is a flag indicating
|
|
+whether a resync is needed. The second is a flag indicating whether a
|
|
+resync is active. This means that the counter is actually 14 bits:
|
|
+
|
|
+| resync_needed | resync_active | counter |
|
|
+| :----: | :----: | :----: |
|
|
+| (0-1) | (0-1) | (0-16383) |
|
|
+
|
|
+The `resync_needed` bit is set when:
|
|
+- a `1` bit is read from storage at startup;
|
|
+- a write request fails on some drives;
|
|
+- a resync is aborted on a chunk with `resync_active` set;
|
|
+- It is cleared (and `resync_active` set) when a resync starts across all drives of the chunk.
|
|
+
|
|
+The `resync_active` bit is set when:
|
|
+- a resync is started on all drives, and `resync_needed` is set.
|
|
+- `resync_needed` will be cleared (as long as `resync_active` wasn't already set).
|
|
+- It is cleared when a resync completes.
|
|
+
|
|
+The counter counts pending write requests, plus the on-disk bit.
|
|
+When the counter is `1` and the resync bits are clear, the on-disk
|
|
+bit can be cleared as well, thus setting the counter to `0`.
|
|
+When we set a bit, or in the counter (to start a write), if the fields is
|
|
+`0`, we first set the disk bit and set the counter to `1`.
|
|
+
|
|
+If the counter is `0`, the on-disk bit is clear and the stipe is clean
|
|
+Anything that dirties the stipe pushes the counter to `2` (at least)
|
|
+and sets the on-disk bit (lazily).
|
|
+If a periodic sweep find the counter at `2`, it is decremented to `1`.
|
|
+If the sweep find the counter at `1`, the on-disk bit is cleared and the
|
|
+counter goes to `0`.
|
|
+
|
|
+Also, we'll hijack the "map" pointer itself and use it as two 16 bit block
|
|
+counters as a fallback when "page" memory cannot be allocated:
|
|
+
|
|
+Normal case (page memory allocated):
|
|
+
|
|
+page pointer (32-bit)
|
|
+
|
|
+ [ ] ------+
|
|
+ |
|
|
+ +-------> [ ][ ]..[ ] (4096 byte page == 2048 counters)
|
|
+ c1 c2 c2048
|
|
+
|
|
+ Hijacked case (page memory allocation failed):
|
|
+
|
|
+ hijacked page pointer (32-bit)
|
|
+
|
|
+ [ ][ ] (no page memory allocated)
|
|
+ counter #1 (16-bit) counter #2 (16-bit)
|
|
+
|
|
+
|
|
+## Notes:
|
|
+1. bitmap_super_s->events counter is updated before the event counter in the md superblock;
|
|
+ When a bitmap is loaded, it is only accepted if this event counter is equal
|
|
+ to, or one greater than, the event counter in the superblock.
|
|
+2. bitmap_super_s->events is updated when the other one is `if` and `only if` the
|
|
+ array is not degraded. As bits are not cleared when the array is degraded,
|
|
+ this represents the last time that any bits were cleared. If a device is being
|
|
+ added that has an event count with this value or higher, it is accepted
|
|
+ as conforming to the bitmap.
|
|
+3. bitmap_super_s->chunksize is the number of sectors represented by the bitmap,
|
|
+ and is the range that resync happens across. For raid1 and raid5/6 it is the
|
|
+ size of individual devices. For raid10 it is the size of the array.
|
|
--
|
|
2.41.0
|
|
|