aboutsummaryrefslogtreecommitdiffstats
path: root/docs
diff options
context:
space:
mode:
Diffstat (limited to 'docs')
-rw-r--r--docs/gettingstarted/developers/bihash.md38
1 files changed, 19 insertions, 19 deletions
diff --git a/docs/gettingstarted/developers/bihash.md b/docs/gettingstarted/developers/bihash.md
index 3754d5f9fa3..b0f98869887 100644
--- a/docs/gettingstarted/developers/bihash.md
+++ b/docs/gettingstarted/developers/bihash.md
@@ -11,17 +11,17 @@ implementation:
* Template implementation, it's easy to support arbitrary (key,value) types
Bounded-index extensible hashing has been widely used in databases for
-decades.
+decades.
Bihash uses a two-level data structure:
```
- +-----------------+
- | bucket-0 |
- | log2_size |
- | backing store |
- +-----------------+
- | bucket-1 |
+ +-----------------+
+ | bucket-0 |
+ | log2_size |
+ | backing store |
+ +-----------------+
+ | bucket-1 |
| log2_size | +--------------------------------+
| backing store | --------> | KVP_PER_PAGE * key-value-pairs |
+-----------------+ | page 0 |
@@ -29,12 +29,12 @@ Bihash uses a two-level data structure:
+-----------------+ | KVP_PER_PAGE * key-value-pairs |
| bucket-2**N-1 | | page 1 |
| log2_size | +--------------------------------+
- | backing store | ---
+ | backing store | ---
+-----------------+ +--------------------------------+
| KVP_PER_PAGE * key-value-pairs |
| page 2**(log2(size)) - 1 |
+--------------------------------+
-```
+```
Discussion of the algorithm
---------------------------
@@ -105,7 +105,7 @@ often in a subsystem header file:
If you're building a standalone application, you'll need to define the
various functions by #including the method implementation file in a C
-source file.
+source file.
The core vpp engine currently uses most if not all of the known bihash
types, so you probably won't need to #include the method
@@ -123,9 +123,9 @@ Add an instance of the selected bihash data structure to e.g. a
typedef struct
{
...
- BVT (clib_bihash) hash;
+ BVT (clib_bihash) hash_table;
or
- clib_bihash_8_8_t hash;
+ clib_bihash_8_8_t hash_table;
...
} my_main_t;
```
@@ -147,7 +147,7 @@ stand no chance of working.
Call the init function as shown. As a rough guide, pick a number of
buckets which is approximately
number_of_expected_records/BIHASH_KVP_PER_PAGE from the relevant
-template instance header-file. See previous discussion.
+template instance header-file. See previous discussion.
The amount of memory selected should easily contain all of the
records, with a generous allowance for hash collisions. Bihash memory
@@ -160,10 +160,10 @@ For example:
```c
my_main_t *mm = &my_main;
clib_bihash_8_8_t *h;
-
+
h = &mm->hash_table;
- clib_bihash_init_8_8 (h, "test", (u32) number_of_buckets,
+ clib_bihash_init_8_8 (h, "test", (u32) number_of_buckets,
(uword) memory_size);
```
@@ -176,7 +176,7 @@ Use BV(clib_bihash_add_del), or the explicit type variant:
clib_bihash_8_8_t * h;
my_main_t *mm = &my_main;
clib_bihash_8_8_t *h;
-
+
h = &mm->hash_table;
kv.key = key_to_add_or_delete;
kv.value = value_to_add_or_delete;
@@ -196,7 +196,7 @@ The simplest possible (key, value) search goes like so:
clib_bihash_8_8_t * h;
my_main_t *mm = &my_main;
clib_bihash_8_8_t *h;
-
+
h = &mm->hash_table;
search_kv.key = key_to_add_or_delete;
@@ -228,7 +228,7 @@ Dual-loop:
* 4 packets ahead, form 2x record keys and call BV(clib_bihash_hash)
or the explicit hash function to calculate the record hashes.
Call 2x BV(clib_bihash_prefetch_bucket) to prefetch the buckets
-* 2 packets ahead, call 2x BV(clib_bihash_prefetch_data) to prefetch
+* 2 packets ahead, call 2x BV(clib_bihash_prefetch_data) to prefetch
2x (key,value) data pages.
* In the processing section, call 2x BV(clib_bihash_search_inline_with_hash)
to perform the search
@@ -270,7 +270,7 @@ For another, the iterator template is not safe under all conditions:
iterator checks whether the current bucket has been freed after each
_callback_fn(...)_ invocation.
-* It is __not OK to add__ entries during a table-walk.
+* It is __not OK to add__ entries during a table-walk.
The add-during-walk case involves a jackpot: while processing a
key-value-pair in a particular bucket, add a certain number of