git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH v2 1/5] upload-pack: look up "want" lines via commit-graph

From
Patrick Steinhardt <ps@pks.im>
Date
Mar 1, 2022, 09:33 UTC
Message-ID
<b5c696bd8e4b33e782e4837ad5051268e35f9828.1646127015.git.ps@pks.im>
In-Reply-To
<cover.1646127015.git.ps@pks.im>

During packfile negotiation the client will send "want" and "want-ref" lines to the server to tell it which objects it is interested in. The server-side parses each of those and looks them up to see whether it actually has requested objects. This lookup is performed by calling `parse_object()` directly, which thus hits the object database. In the general case though most of the objects the client requests will be commits. We can thus try to look up the object via the commit-graph opportunistically, which is much faster than doing the same via the object database.

Refactor parsing of both "want" and "want-ref" lines to do so.

The following benchmark is executed in a repository with a huge number of references. It uses cached request from git-fetch(1) as input to git-upload-pack(1) that contains about 876,000 "want" lines:

    Benchmark 1: HEAD~
      Time (mean ± σ):      7.113 s ±  0.028 s    [User: 6.900 s, System: 0.662 s]
      Range (min … max):    7.072 s …  7.168 s    10 runs
    Benchmark 2: HEAD
      Time (mean ± σ):      6.622 s ±  0.061 s    [User: 6.452 s, System: 0.650 s]
      Range (min … max):    6.535 s …  6.727 s    10 runs
    Summary
      'HEAD' ran
        1.07 ± 0.01 times faster than 'HEAD~'
Signed-off-by: Patrick Steinhardt <ps@pks.im>
---
 upload-pack.c | 20 +++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)
diff --git a/upload-pack.c b/upload-pack.c
index 8acc98741b..3a851b3606 100644
--- a/upload-pack.c
+++ b/upload-pack.c
@@ -1400,13 +1400,19 @@ static int parse_want(struct packet_writer *writer, const char *line,
 	const char *arg;
 	if (skip_prefix(line, "want ", &arg)) {
 		struct object_id oid;
+		struct commit *commit;
 		struct object *o;
 
 		if (get_oid_hex(arg, &oid))
 			die("git upload-pack: protocol error, "
 			    "expected to get oid, not '%s'", line);
 
-		o = parse_object(the_repository, &oid);
+		commit = lookup_commit_in_graph(the_repository, &oid);
+		if (commit)
+			o = &commit->object;
+		else
+			o = parse_object(the_repository, &oid);
+
 		if (!o) {
 			packet_writer_error(writer,
 					    "upload-pack: not our ref %s",
@@ -1434,7 +1440,7 @@ static int parse_want_ref(struct packet_writer *writer, const char *line,
 	if (skip_prefix(line, "want-ref ", &refname_nons)) {
 		struct object_id oid;
 		struct string_list_item *item;
-		struct object *o;
+		struct object *o = NULL;
 		struct strbuf refname = STRBUF_INIT;
 
 		strbuf_addf(&refname, "%s%s", get_git_namespace(), refname_nons);
@@ -1448,7 +1454,15 @@ static int parse_want_ref(struct packet_writer *writer, const char *line,
 		item = string_list_append(wanted_refs, refname_nons);
 		item->util = oiddup(&oid);
 
-		o = parse_object_or_die(&oid, refname_nons);
+		if (!starts_with(refname_nons, "refs/tags/")) {
+			struct commit *commit = lookup_commit_in_graph(the_repository, &oid);
+			if (commit)
+				o = &commit->object;
+		}
+
+		if (!o)
+			o = parse_object_or_die(&oid, refname_nons);
+
 		if (!(o->flags & WANTED)) {
 			o->flags |= WANTED;
 			add_object_array(o, NULL, want_obj);
-- 
2.35.1
Previous: Patrick SteinhardtNext: Patrick Steinhardt
Message 14 of 20 in “fetch: more optimizations for mirror fetches”
  1. 0/5 fetch: more optimizations for mirror fetchesPatrick Steinhardt, Feb 23, 2022
  2. 2/5 fetch: avoid lookup of commits when not appending to FETCH_HEADPatrick Steinhardt, Feb 23, 2022
  3. Derrick StoleeFeb 23, 2022
  4. Patrick SteinhardtMar 1, 2022
  5. 1/5 upload-pack: look up "want" lines via commit-graphPatrick Steinhardt, Feb 23, 2022
  6. Derrick StoleeFeb 23, 2022
  7. Patrick SteinhardtMar 1, 2022
  8. Patrick SteinhardtMar 1, 2022
  9. Derrick StoleeMar 2, 2022
  10. 3/5 refs: add ability for backends to special-case reading of symbolic refsPatrick Steinhardt, Feb 23, 2022
  11. 4/5 remote: read symbolic refs via `refs_read_symbolic_ref()`Patrick Steinhardt, Feb 23, 2022
  12. 5/5 refs/files-backend: optimize reading of symbolic refsPatrick Steinhardt, Feb 23, 2022
  13. 0/5 fetch: more optimizations for mirror fetchesPatrick Steinhardt, Mar 1, 2022
  14. 1/5 upload-pack: look up "want" lines via commit-graphPatrick Steinhardt, Mar 1, 2022
  15. 2/5 fetch: avoid lookup of commits when not appending to FETCH_HEADPatrick Steinhardt, Mar 1, 2022
  16. 3/5 refs: add ability for backends to special-case reading of symbolic refsPatrick Steinhardt, Mar 1, 2022
  17. 4/5 remote: read symbolic refs via `refs_read_symbolic_ref()`Patrick Steinhardt, Mar 1, 2022
  18. 5/5 refs/files-backend: optimize reading of symbolic refsPatrick Steinhardt, Mar 1, 2022
  19. Junio C HamanoMar 1, 2022
  20. Derrick StoleeMar 2, 2022

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.