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

Re: [PATCH] revision: fix missing null for freed memory

From
Junio C Hamano <gitster@pobox.com>
Date
Feb 10, 2025, 16:02 UTC
Message-ID
<xmqqtt91dbzt.fsf@gitster.g>
In-Reply-To
<20250208061702.88469-1-forivall@gmail.com>
Emily M Klassen <forivall@gmail.com> writes:
Show 6 quoted lines
> Subject: Re: [PATCH] revision: fix missing null for freed memory
>
> "git log --graph --no-graph" missed cleaning up the output_prefix and
> output_prefix_data pointers. This resulted in a segfault when using "--patch",
> "--name-status" or "--name-only", as the output_prefix_data continued to be in
> use after free()

Rereading the title, I cannot make sense out of "fix missing null" and guess what it wants to say. Is "null" here used as a verb to mean "to assign a NULL to a variable that points at ..."?

    revision: clear graph callback upon "--no-graph"
    "git log --graph --no-graph" first populates the .output_prefix
    member of diffopt, which is a callback function, to compute
    "--graph" header, and then discards the data the callback needs
    to compute the graph header but forgets to clear .output_prefix
    pointer in response to "--no-graph".  At runtime, we end up
    calling the function that we should not.
    Clear the member to stop making callback, and for a better
    hyginene, also clear the pointer pointing at a freed memory.
or something?
Other than that, as I said earlier, the patch looks good.
Thanks.
Show 21 quoted lines
> Signed-off-by: Emily M Klassen <forivall@gmail.com>
> ---
> I previously reported this a few hours ago, and ended up digging in and figuring
> it out. I'll make sure to bottom reply in the follow ups to this patch.
>
>  revision.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/revision.c b/revision.c
> index 474fa1e767..84cb028e11 100644
> --- a/revision.c
> +++ b/revision.c
> @@ -2615,6 +2615,8 @@ static int handle_revision_opt(struct rev_info *revs, int argc, const char **arg
>  		graph_clear(revs->graph);
>  		revs->graph = graph_init(revs);
>  	} else if (!strcmp(arg, "--no-graph")) {
> +		revs->diffopt.output_prefix = NULL;
> +		revs->diffopt.output_prefix_data = NULL;
>  		graph_clear(revs->graph);
>  		revs->graph = NULL;
>  	} else if (!strcmp(arg, "--encode-email-headers")) {
Previous: Junio C HamanoNext: Emily Klassen
Message 3 of 12 in “revision: fix missing null for freed memory”
  1. revision: fix missing null for freed memoryEmily M Klassen, Feb 8, 2025
  2. Junio C HamanoFeb 8, 2025
  3. Junio C HamanoFeb 10, 2025
  4. Emily KlassenFeb 10, 2025
  5. Junio C HamanoFeb 13, 2025
  6. Patrick SteinhardtFeb 11, 2025
  7. D. Ben KnobleFeb 11, 2025
  8. D. Ben KnobleFeb 11, 2025
  9. Jeff KingFeb 11, 2025
  10. Junio C HamanoFeb 11, 2025
  11. Patrick SteinhardtFeb 12, 2025
  12. Ben KnobleFeb 13, 2025

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.