Implement generalized btree push, note the boundary conditions when id=weight

This really just required care around calculating the expected B-tree id
and rbyd id (which are different!).

B-tree append, aka B-tree push with id=weight, is actually the outlier.
We need a B-tree id that can identify the rbyd we're appending to, but
this id itself doesn't exist in the tree yet, which can be a bit tricky.
This commit is contained in:
Christopher Haster
2023-03-07 13:04:57 -06:00
parent d6a2666614
commit 0a3c6b39c1
4 changed files with 276 additions and 36 deletions
+220 -2
View File
@@ -115,6 +115,51 @@ code = '''
buffer, 4) => LFS_ERR_NOENT;
'''
[cases.test_btree_two_backwards]
in = 'lfs.c'
code = '''
lfs_t lfs;
lfs_init(&lfs, cfg) => 0;
// create free lookahead
memset(lfs.free.buffer, 0, lfs.cfg->lookahead_size);
lfs.free.off = 0;
lfs.free.size = lfs_min(8*lfs.cfg->lookahead_size,
lfs.cfg->block_count);
lfs.free.i = 0;
lfs_alloc_ack(&lfs);
// create a two-entry tree
lfsr_btree_t btree = LFSR_BTREE_NULL;
lfsr_btree_push(&lfs, &btree, 0, LFSR_TAG_INLINED, 1, "b", 1) => 0;
lfsr_btree_push(&lfs, &btree, 0, LFSR_TAG_INLINED, 1, "a", 1) => 0;
// try looking up tags
uint8_t buffer[4];
lfsr_tag_t tag_;
lfs_size_t id_;
lfs_size_t weight_;
lfsr_btree_get(&lfs, &btree, 0,
&tag_, &id_, &weight_,
buffer, 4) => 1;
assert(tag_ == LFSR_TAG_INLINED);
assert(id_ == 0);
assert(weight_ == 1);
assert(memcmp(buffer, "a", 1) == 0);
lfsr_btree_get(&lfs, &btree, 1,
&tag_, &id_, &weight_,
buffer, 4) => 1;
assert(tag_ == LFSR_TAG_INLINED);
assert(id_ == 1);
assert(weight_ == 1);
assert(memcmp(buffer, "b", 1) == 0);
lfsr_btree_get(&lfs, &btree, 2,
&tag_, &id_, &weight_,
buffer, 4) => LFS_ERR_NOENT;
'''
# still a single-rbyd tree, just making sure it works
[cases.test_btree_three]
in = 'lfs.c'
@@ -170,10 +215,64 @@ code = '''
buffer, 4) => LFS_ERR_NOENT;
'''
[cases.test_btree_three_backwards]
in = 'lfs.c'
code = '''
lfs_t lfs;
lfs_init(&lfs, cfg) => 0;
// create free lookahead
memset(lfs.free.buffer, 0, lfs.cfg->lookahead_size);
lfs.free.off = 0;
lfs.free.size = lfs_min(8*lfs.cfg->lookahead_size,
lfs.cfg->block_count);
lfs.free.i = 0;
lfs_alloc_ack(&lfs);
// create a two-entry tree
lfsr_btree_t btree = LFSR_BTREE_NULL;
lfsr_btree_push(&lfs, &btree, 0, LFSR_TAG_INLINED, 1, "c", 1) => 0;
lfsr_btree_push(&lfs, &btree, 0, LFSR_TAG_INLINED, 1, "b", 1) => 0;
lfsr_btree_push(&lfs, &btree, 0, LFSR_TAG_INLINED, 1, "a", 1) => 0;
// try looking up tags
uint8_t buffer[4];
lfsr_tag_t tag_;
lfs_size_t id_;
lfs_size_t weight_;
lfsr_btree_get(&lfs, &btree, 0,
&tag_, &id_, &weight_,
buffer, 4) => 1;
assert(tag_ == LFSR_TAG_INLINED);
assert(id_ == 0);
assert(weight_ == 1);
assert(memcmp(buffer, "a", 1) == 0);
lfsr_btree_get(&lfs, &btree, 1,
&tag_, &id_, &weight_,
buffer, 4) => 1;
assert(tag_ == LFSR_TAG_INLINED);
assert(id_ == 1);
assert(weight_ == 1);
assert(memcmp(buffer, "b", 1) == 0);
lfsr_btree_get(&lfs, &btree, 2,
&tag_, &id_, &weight_,
buffer, 4) => 1;
assert(tag_ == LFSR_TAG_INLINED);
assert(id_ == 2);
assert(weight_ == 1);
assert(memcmp(buffer, "c", 1) == 0);
lfsr_btree_get(&lfs, &btree, 3,
&tag_, &id_, &weight_,
buffer, 4) => LFS_ERR_NOENT;
'''
# try larger trees, when exactly a tree splits depends on the disk geometry, so
# we don't really have a better way of testing multi-rbyd trees
[cases.test_btree_more]
defines.N = 'range(10, 1001, 50)'
[cases.test_btree_push]
defines.N = [4, 8, 16, 32, 64, 128, 256, 512, 1024]
in = 'lfs.c'
code = '''
lfs_t lfs;
@@ -215,3 +314,122 @@ code = '''
&tag_, &id_, &weight_,
buffer, 4) => LFS_ERR_NOENT;
'''
[cases.test_btree_push_backwards]
defines.N = [4, 8, 16, 32, 64, 128, 256, 512, 1024]
in = 'lfs.c'
code = '''
lfs_t lfs;
lfs_init(&lfs, cfg) => 0;
// create free lookahead
memset(lfs.free.buffer, 0, lfs.cfg->lookahead_size);
lfs.free.off = 0;
lfs.free.size = lfs_min(8*lfs.cfg->lookahead_size,
lfs.cfg->block_count);
lfs.free.i = 0;
lfs_alloc_ack(&lfs);
// create a tree with N elements
lfsr_btree_t btree = LFSR_BTREE_NULL;
const char *alphas = "abcdefghijklmnopqrstuvwxyz";
for (lfs_size_t i = 0; i < N; i++) {
lfsr_btree_push(&lfs, &btree, 0, LFSR_TAG_INLINED, 1,
&alphas[(N-1-i) % 26], 1) => 0;
}
// check that the elements are in the tree
uint8_t buffer[4];
lfsr_tag_t tag_;
lfs_size_t id_;
lfs_size_t weight_;
for (lfs_size_t i = 0; i < N; i++) {
lfsr_btree_get(&lfs, &btree, i,
&tag_, &id_, &weight_,
buffer, 4) => 1;
assert(tag_ == LFSR_TAG_INLINED);
assert(id_ == i);
assert(weight_ == 1);
assert(memcmp(buffer, &alphas[i % 26], 1) == 0);
}
// and check that we can't lookup elements that aren't in the tree
lfsr_btree_get(&lfs, &btree, N,
&tag_, &id_, &weight_,
buffer, 4) => LFS_ERR_NOENT;
'''
[cases.test_btree_push_fuzz]
defines.N = [1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024]
defines.ITER = 10
in = 'lfs.c'
code = '''
const char *alphas = "abcdefghijklmnopqrstuvwxyz";
// iterate through severals seeds that we can reproduce easily
for (uint32_t seed = 1; seed < ITER+1; seed++) {
// create lfs here since we need to reset each iteration, we're
// space constrained and we can't expect gc to work at this point
lfs_t lfs;
lfs_init(&lfs, cfg) => 0;
// create free lookahead
memset(lfs.free.buffer, 0, lfs.cfg->lookahead_size);
lfs.free.off = 0;
lfs.free.size = lfs_min(8*lfs.cfg->lookahead_size,
lfs.cfg->block_count);
lfs.free.i = 0;
lfs_alloc_ack(&lfs);
// create a btree
lfsr_btree_t btree = LFSR_BTREE_NULL;
// set up a simulation to compare against
//
// fun fact this is slower than our actual tree! unfun fact this is
// starting to be a problem...
char *sim = malloc(N);
lfs_size_t sim_size = 0;
memset(sim, 0, N);
uint32_t prng = seed;
for (lfs_size_t i = 0; i < N; i++) {
// choose a pseudo-random id
lfs_size_t id = TEST_PRNG(&prng) % (sim_size+1);
// add to btree
lfsr_btree_push(&lfs, &btree, id, LFSR_TAG_INLINED, 1,
&alphas[i], 1) => 0;
// add to sim
memcpy(&sim[id+1], &sim[id], sim_size-id);
sim[id] = alphas[i];
sim_size += 1;
}
// check that btree matches sim
assert(btree.weight == N);
uint8_t buffer[4];
lfsr_tag_t tag_;
lfs_size_t id_;
lfs_size_t weight_;
for (lfs_size_t i = 0; i < N; i++) {
lfsr_btree_get(&lfs, &btree, i,
&tag_, &id_, &weight_,
buffer, 4) => 1;
assert(tag_ == LFSR_TAG_INLINED);
assert(id_ == i);
assert(weight_ == 1);
assert(memcmp(buffer, &sim[i], 1) == 0);
}
// and no extra elements
lfsr_btree_get(&lfs, &btree, N,
&tag_, &id_, &weight_,
buffer, 4) => LFS_ERR_NOENT;
// clean up sim
free(sim);
}
'''