| blob | 107395c042ae0736941ea4f5d887b3b0eacfdae6 |
1 =============================
2 BTT - Block Translation Table
3 =============================
6 1. Introduction
7 ===============
9 Persistent memory based storage is able to perform IO at byte (or more
10 accurately, cache line) granularity. However, we often want to expose such
11 storage as traditional block devices. The block drivers for persistent memory
12 will do exactly this. However, they do not provide any atomicity guarantees.
13 Traditional SSDs typically provide protection against torn sectors in hardware,
14 using stored energy in capacitors to complete in-flight block writes, or perhaps
15 in firmware. We don't have this luxury with persistent memory - if a write is in
16 progress, and we experience a power failure, the block will contain a mix of old
17 and new data. Applications may not be prepared to handle such a scenario.
19 The Block Translation Table (BTT) provides atomic sector update semantics for
20 persistent memory devices, so that applications that rely on sector writes not
21 being torn can continue to do so. The BTT manifests itself as a stacked block
22 device, and reserves a portion of the underlying storage for its metadata. At
23 the heart of it, is an indirection table that re-maps all the blocks on the
24 volume. It can be thought of as an extremely simple file system that only
25 provides atomic sector updates.
28 2. Static Layout
29 ================
31 The underlying storage on which a BTT can be laid out is not limited in any way.
32 The BTT, however, splits the available space into chunks of up to 512 GiB,
33 called "Arenas".
35 Each arena follows the same layout for its metadata, and all references in an
36 arena are internal to it (with the exception of one field that points to the
37 next arena). The following depicts the "On-disk" metadata layout::
40 Backing Store +-------> Arena
41 +---------------+ | +------------------+
42 | | | | Arena info block |
43 | Arena 0 +---+ | 4K |
44 | 512G | +------------------+
45 | | | |
46 +---------------+ | |
47 | | | |
48 | Arena 1 | | Data Blocks |
49 | 512G | | |
50 | | | |
51 +---------------+ | |
52 | . | | |
53 | . | | |
54 | . | | |
55 | | | |
56 | | | |
57 +---------------+ +------------------+
58 | |
59 | BTT Map |
60 | |
61 | |
62 +------------------+
63 | |
64 | BTT Flog |
65 | |
66 +------------------+
67 | Info block copy |
68 | 4K |
69 +------------------+
72 3. Theory of Operation
73 ======================
76 a. The BTT Map
77 --------------
79 The map is a simple lookup/indirection table that maps an LBA to an internal
80 block. Each map entry is 32 bits. The two most significant bits are special
81 flags, and the remaining form the internal block number.
83 ======== =============================================================
84 Bit Description
85 ======== =============================================================
86 31 - 30 Error and Zero flags - Used in the following way::
88 == == ====================================================
89 31 30 Description
90 == == ====================================================
91 0 0 Initial state. Reads return zeroes; Premap = Postmap
92 0 1 Zero state: Reads return zeroes
93 1 0 Error state: Reads fail; Writes clear 'E' bit
94 1 1 Normal Block – has valid postmap
95 == == ====================================================
97 29 - 0 Mappings to internal 'postmap' blocks
98 ======== =============================================================
101 Some of the terminology that will be subsequently used:
103 ============ ================================================================
104 External LBA LBA as made visible to upper layers.
105 ABA Arena Block Address - Block offset/number within an arena
106 Premap ABA The block offset into an arena, which was decided upon by range
107 checking the External LBA
108 Postmap ABA The block number in the "Data Blocks" area obtained after
109 indirection from the map
110 nfree The number of free blocks that are maintained at any given time.
111 This is the number of concurrent writes that can happen to the
112 arena.
113 ============ ================================================================
116 For example, after adding a BTT, we surface a disk of 1024G. We get a read for
117 the external LBA at 768G. This falls into the second arena, and of the 512G
118 worth of blocks that this arena contributes, this block is at 256G. Thus, the
119 premap ABA is 256G. We now refer to the map, and find out the mapping for block
120 'X' (256G) points to block 'Y', say '64'. Thus the postmap ABA is 64.
123 b. The BTT Flog
124 ---------------
126 The BTT provides sector atomicity by making every write an "allocating write",
127 i.e. Every write goes to a "free" block. A running list of free blocks is
128 maintained in the form of the BTT flog. 'Flog' is a combination of the words
129 "free list" and "log". The flog contains 'nfree' entries, and an entry contains:
131 ======== =====================================================================
132 lba The premap ABA that is being written to
133 old_map The old postmap ABA - after 'this' write completes, this will be a
134 free block.
135 new_map The new postmap ABA. The map will up updated to reflect this
136 lba->postmap_aba mapping, but we log it here in case we have to
137 recover.
138 seq Sequence number to mark which of the 2 sections of this flog entry is
139 valid/newest. It cycles between 01->10->11->01 (binary) under normal
140 operation, with 00 indicating an uninitialized state.
141 lba' alternate lba entry
142 old_map' alternate old postmap entry
143 new_map' alternate new postmap entry
144 seq' alternate sequence number.
145 ======== =====================================================================
147 Each of the above fields is 32-bit, making one entry 32 bytes. Entries are also
148 padded to 64 bytes to avoid cache line sharing or aliasing. Flog updates are
149 done such that for any entry being written, it:
150 a. overwrites the 'old' section in the entry based on sequence numbers
151 b. writes the 'new' section such that the sequence number is written last.
154 c. The concept of lanes
155 -----------------------
157 While 'nfree' describes the number of concurrent IOs an arena can process
158 concurrently, 'nlanes' is the number of IOs the BTT device as a whole can
159 process::
161 nlanes = min(nfree, num_cpus)
163 A lane number is obtained at the start of any IO, and is used for indexing into
164 all the on-disk and in-memory data structures for the duration of the IO. If
165 there are more CPUs than the max number of available lanes, than lanes are
166 protected by spinlocks.
169 d. In-memory data structure: Read Tracking Table (RTT)
170 ------------------------------------------------------
172 Consider a case where we have two threads, one doing reads and the other,
173 writes. We can hit a condition where the writer thread grabs a free block to do
174 a new IO, but the (slow) reader thread is still reading from it. In other words,
175 the reader consulted a map entry, and started reading the corresponding block. A
176 writer started writing to the same external LBA, and finished the write updating
177 the map for that external LBA to point to its new postmap ABA. At this point the
178 internal, postmap block that the reader is (still) reading has been inserted
179 into the list of free blocks. If another write comes in for the same LBA, it can
180 grab this free block, and start writing to it, causing the reader to read
181 incorrect data. To prevent this, we introduce the RTT.
183 The RTT is a simple, per arena table with 'nfree' entries. Every reader inserts
184 into rtt[lane_number], the postmap ABA it is reading, and clears it after the
185 read is complete. Every writer thread, after grabbing a free block, checks the
186 RTT for its presence. If the postmap free block is in the RTT, it waits till the
187 reader clears the RTT entry, and only then starts writing to it.
190 e. In-memory data structure: map locks
191 --------------------------------------
193 Consider a case where two writer threads are writing to the same LBA. There can
194 be a race in the following sequence of steps::
196 free[lane] = map[premap_aba]
197 map[premap_aba] = postmap_aba
199 Both threads can update their respective free[lane] with the same old, freed
200 postmap_aba. This has made the layout inconsistent by losing a free entry, and
201 at the same time, duplicating another free entry for two lanes.
203 To solve this, we could have a single map lock (per arena) that has to be taken
204 before performing the above sequence, but we feel that could be too contentious.
205 Instead we use an array of (nfree) map_locks that is indexed by
206 (premap_aba modulo nfree).
209 f. Reconstruction from the Flog
210 -------------------------------
212 On startup, we analyze the BTT flog to create our list of free blocks. We walk
213 through all the entries, and for each lane, of the set of two possible
214 'sections', we always look at the most recent one only (based on the sequence
215 number). The reconstruction rules/steps are simple:
217 - Read map[log_entry.lba].
218 - If log_entry.new matches the map entry, then log_entry.old is free.
219 - If log_entry.new does not match the map entry, then log_entry.new is free.
220 (This case can only be caused by power-fails/unsafe shutdowns)
223 g. Summarizing - Read and Write flows
224 -------------------------------------
226 Read:
228 1. Convert external LBA to arena number + pre-map ABA
229 2. Get a lane (and take lane_lock)
230 3. Read map to get the entry for this pre-map ABA
231 4. Enter post-map ABA into RTT[lane]
232 5. If TRIM flag set in map, return zeroes, and end IO (go to step 8)
233 6. If ERROR flag set in map, end IO with EIO (go to step 8)
234 7. Read data from this block
235 8. Remove post-map ABA entry from RTT[lane]
236 9. Release lane (and lane_lock)
238 Write:
240 1. Convert external LBA to Arena number + pre-map ABA
241 2. Get a lane (and take lane_lock)
242 3. Use lane to index into in-memory free list and obtain a new block, next flog
243 index, next sequence number
244 4. Scan the RTT to check if free block is present, and spin/wait if it is.
245 5. Write data to this free block
246 6. Read map to get the existing post-map ABA entry for this pre-map ABA
247 7. Write flog entry: [premap_aba / old postmap_aba / new postmap_aba / seq_num]
248 8. Write new post-map ABA into map.
249 9. Write old post-map entry into the free list
250 10. Calculate next sequence number and write into the free list entry
251 11. Release lane (and lane_lock)
254 4. Error Handling
255 =================
257 An arena would be in an error state if any of the metadata is corrupted
258 irrecoverably, either due to a bug or a media error. The following conditions
259 indicate an error:
261 - Info block checksum does not match (and recovering from the copy also fails)
262 - All internal available blocks are not uniquely and entirely addressed by the
263 sum of mapped blocks and free blocks (from the BTT flog).
264 - Rebuilding free list from the flog reveals missing/duplicate/impossible
265 entries
266 - A map entry is out of bounds
268 If any of these error conditions are encountered, the arena is put into a read
269 only state using a flag in the info block.
272 5. Usage
273 ========
275 The BTT can be set up on any disk (namespace) exposed by the libnvdimm subsystem
276 (pmem, or blk mode). The easiest way to set up such a namespace is using the
277 'ndctl' utility [1]:
279 For example, the ndctl command line to setup a btt with a 4k sector size is::
281 ndctl create-namespace -f -e namespace0.0 -m sector -l 4k
283 See ndctl create-namespace --help for more options.
285 [1]: https://github.com/pmem/ndctl