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

Re: [PATCH 06/10] commit-graph: delay base_graph assignment in add_graph_to_chain()

From
Taylor Blau <me@ttaylorr.com>
Date
Oct 5, 2023, 17:44 UTC
Message-ID
<ZR71+Ht3G/xLqJ++@nand.local>
In-Reply-To
<20231003203004.GF7812@coredump.intra.peff.net>
On Tue, Oct 03, 2023 at 04:30:04PM -0400, Jeff King wrote:
Show 36 quoted lines
> When adding a graph to a chain, we do some consistency checks and then
> if everything looks good, set g->base_graph to add a link to the chain.
> But when we added a new consistency check in 209250ef38 (commit-graph.c:
> prevent overflow in add_graph_to_chain(), 2023-07-12), it comes _after_
> we've already set g->base_graph. So we might return failure, even though
> we actually added to the chain.
>
> This hasn't caused a bug yet, because after failing to add to the chain,
> we discard the failed graph struct completely, leaking it. But in order
> to fix that, it's important that the struct be in a consistent and
> predictable state after the failure.
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
>  commit-graph.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/commit-graph.c b/commit-graph.c
> index 2f75ecd9ae..2c72a554c2 100644
> --- a/commit-graph.c
> +++ b/commit-graph.c
> @@ -498,8 +498,6 @@ static int add_graph_to_chain(struct commit_graph *g,
>  		cur_g = cur_g->base_graph;
>  	}
>
> -	g->base_graph = chain;
> -
>  	if (chain) {
>  		if (unsigned_add_overflows(chain->num_commits,
>  					   chain->num_commits_in_base)) {
> @@ -510,6 +508,8 @@ static int add_graph_to_chain(struct commit_graph *g,
>  		g->num_commits_in_base = chain->num_commits + chain->num_commits_in_base;
>  	}
>
> +	g->base_graph = chain;
> +

Oops. That looks like my fault. Thanks for catching, this switch makes sense to me.

Thanks, Taylor

Previous: Jeff KingNext: Jeff King
Message 11 of 24 in “some commit-graph leak fixes”
  1. 0/10 some commit-graph leak fixesJeff King, Oct 3, 2023
  2. 01/10 t6700: mark test as leak-freeJeff King, Oct 3, 2023
  3. Taylor BlauOct 5, 2023
  4. 02/10 commit-reach: free temporary list in get_octopus_merge_bases()Jeff King, Oct 3, 2023
  5. 03/10 merge: free result of repo_get_merge_bases()Jeff King, Oct 3, 2023
  6. Taylor BlauOct 5, 2023
  7. 04/10 commit-graph: move slab-clearing to close_commit_graph()Jeff King, Oct 3, 2023
  8. Taylor BlauOct 5, 2023
  9. 05/10 commit-graph: free all elements of graph chainJeff King, Oct 3, 2023
  10. 06/10 commit-graph: delay base_graph assignment in add_graph_to_chain()Jeff King, Oct 3, 2023
  11. Taylor BlauOct 5, 2023
  12. 07/10 commit-graph: free graph struct that was not added to chainJeff King, Oct 3, 2023
  13. 08/10 commit-graph: free write-context entries before overwritingJeff King, Oct 3, 2023
  14. Taylor BlauOct 5, 2023
  15. Jeff KingOct 5, 2023
  16. 09/10 commit-graph: free write-context base_graph_name during cleanupJeff King, Oct 3, 2023
  17. 10/10 commit-graph: clear oidset after finishing writeJeff King, Oct 3, 2023
  18. Is SANITIZE=leak make test unreliable for anyone else?Eric W. Biederman, Oct 4, 2023
  19. Jeff KingOct 4, 2023
  20. Eric W. BiedermanOct 4, 2023
  21. Jeff KingOct 4, 2023
  22. Eric W. BiedermanOct 4, 2023
  23. Taylor BlauOct 5, 2023
  24. Junio C HamanoOct 6, 2023

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.