r15693@tombo: nickm | 2007-12-25 19:11:29 -0500

 Here, have some terribly clever new buffer code.  It uses a mbuf-like strategy rather than a ring buffer strategy, so it should require far far less extra memory to hold any given amount of data.  Also, it avoids access patterns like x=malloc(1024);x=realloc(x,1048576);x=realloc(x,1024);append_to_freelist(x) that might have been contributing to memory fragmentation.  I've  tested it out a little on peacetime, and it seems to work so far.  If you want to benchmark it for speed, make sure to remove the #define PARANOIA; #define NOINLINE macros at the head of the module.


svn:r12983

1 files changed, 2 insertions, 1 deletions

diff --git a/src/or/or.h b/src/or/or.h
index d0aa9902bd..1979d7638f 100644
--- a/src/or/or.h
+++ b/src/or/or.h
@@ -2439,7 +2439,8 @@ void buf_shrink_freelists(int free_all);
 void buf_dump_freelist_sizes(int severity);
 
 size_t buf_datalen(const buf_t *buf);
-size_t buf_capacity(const buf_t *buf);
+size_t buf_allocation(const buf_t *buf);
+size_t buf_slack(const buf_t *buf);
 const char *_buf_peek_raw_buffer(const buf_t *buf);
 
 int read_to_buf(int s, size_t at_most, buf_t *buf, int *reached_eof);