github_mirrors/vectorscan
3
0
Fork 1
mirror of https://github.com/VectorCamp/vectorscan.git synced 2025-11-16 09:21:52 +03:00
Code Issues Packages Projects Releases Wiki Activity

rose: simplify long lit table, add bloom filter

Browse Source 
Replaces the original long lit hash table (used in streaming mode) with a
smaller, simpler linear probing approach. Adds a bloom filter in front
of it to reduce time spent on false positives.

Sizing of both the hash table and bloom filter are done based on max
load.
This commit is contained in:
Justin Viiret
2016-09-22 13:58:42 +10:00
committed by Matthew Barr
parent 68bf473e2e
commit 8869dee643
8 changed files with 520 additions and 435 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
src/rose/stream.c
View File

@@ -551,6 +551,11 @@ void roseStreamExec(const struct RoseEngine *t, struct hs_scratch *scratch) {
tctxt->minMatchOffset = offset;
tctxt->minNonMpvMatchOffset = offset;
tctxt->next_mpv_offset = 0;
tctxt->ll_buf = scratch->core_info.hbuf;
tctxt->ll_len = scratch->core_info.hlen;
tctxt->ll_buf_nocase = scratch->core_info.hbuf;
tctxt->ll_len_nocase = scratch->core_info.hlen;
DEBUG_PRINTF("BEGIN: history len=%zu, buffer len=%zu groups=%016llx\n",
scratch->core_info.hlen, scratch->core_info.len, tctxt->groups);
@@ -590,18 +595,14 @@ void roseStreamExec(const struct RoseEngine *t, struct hs_scratch *scratch) {
MIN(t->floatingDistance, length + offset) - offset : 0;
}
loadLongLiteralState(t, state, scratch);
size_t hlength = scratch->core_info.hlen;
char rebuild = 0;
if (hlength) {
// Can only have long literal state or rebuild if this is not the
// first write to this stream.
loadLongLiteralState(t, state, scratch);
rebuild = (scratch->core_info.status & STATUS_DELAY_DIRTY) &&
(t->maxFloatingDelayedMatch == ROSE_BOUND_INF ||
offset < t->maxFloatingDelayedMatch);
}
char rebuild = hlength &&
(scratch->core_info.status & STATUS_DELAY_DIRTY) &&
(t->maxFloatingDelayedMatch == ROSE_BOUND_INF ||
offset < t->maxFloatingDelayedMatch);
DEBUG_PRINTF("**rebuild %hhd status %hhu mfdm %u, offset %llu\n",
rebuild, scratch->core_info.status,
t->maxFloatingDelayedMatch, offset);