git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Elijah Newren <newren@gmail.com>
To: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
Cc: "Git Mailing List" <git@vger.kernel.org>,
	"Junio C Hamano" <gitster@pobox.com>,
	"Martin Ågren" <martin.agren@gmail.com>,
	"Andrzej Hunt" <ajrhunt@google.com>, "Jeff King" <peff@peff.net>
Subject: Re: [PATCH v2 2/3] sequencer: add a "goto cleanup" to do_reset()
Date: Thu, 7 Oct 2021 09:06:48 -0700	[thread overview]
Message-ID: <CABPp-BH=3DP-dXRCphY53-3eZd1TU8h5GY_M12nnbEGm-UYB9Q@mail.gmail.com> (raw)
In-Reply-To: <patch-v2-2.3-1d5f5e9fff0-20211007T094019Z-avarab@gmail.com>

On Thu, Oct 7, 2021 at 2:46 AM Ævar Arnfjörð Bjarmason <avarab@gmail.com> wrote:
>
> Restructure code that's mostly added in 9055e401dd6 (sequencer:
> introduce new commands to reset the revision, 2018-04-25) to avoid
> code duplication, and to make freeing other resources easier in a
> subsequent commit.

Since the first goto you add precedes the call to
setup_unpack_trees_porcelain(), the next commit will introduce a case
where some code calls clear_unpack_trees_porcelain() without having
first called setup_unpack_trees_porcelain().  With the current code
that is fine since you do initialize unpack_trees_opts to { 0 } first,
and clear_unpack_trees_porcelain() doesn't rely on any special setup,
but I wonder if the naming similarity of the two functions might lead
to future authors to presume they are always called in pairs.

I wonder if we should rename clear_unpack_trees_porcelain() to e.g.
unpack_trees_clear() just to avoid this problem, and then include the
other bits of your previous round to just make all callers of
unpack_trees() call unpack_trees_clear() when they are done.

Or, maybe we could just make unpack_trees() call unpack_trees_clear()
automatically and remove that responsibility from callers...except
that merge-recursive is weird and special and would need a new
function (unpack_trees_without_automatic_cleanup() or something like
that?) since it's the one caller that needs to use
unpack_trees_options data after calling unpack_trees().

So, multiple competing ideas here, and all for theoretical future code
safety.  As such, if any of it sounds like too much of a pain, I'm
fine with punting on it.

> It's safe to initialize "tree_desc" to be zero'd out in order to
> unconditionally free desc.buffer, it won't be initialized on the first
> couple of "goto"'s.
>
> There are three earlier "return"'s in this function that I'm not
> bothering to covert,

s/covert/convert/

> those don't need to rollback anything,

They aren't currently rolling back the lockfile right now, but I think
they should.  Those error cases are pretty unlikely to happen (hard
disk full?), but if they did, they'd leave the lock file in place.

Granted, that's a pre-existing problem rather than a problem
introduced by your patch, so it doesn't need to be part of this
series, but the commit message should at least be tweaked.

> or free any resources

True.

> , so let's leave, even though they could safely "goto
> cleanup" as well.



>
> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
> ---
>  sequencer.c | 32 +++++++++++++-------------------
>  1 file changed, 13 insertions(+), 19 deletions(-)
>
> diff --git a/sequencer.c b/sequencer.c
> index 6872b7b00a4..457eba4ab10 100644
> --- a/sequencer.c
> +++ b/sequencer.c
> @@ -3650,7 +3650,7 @@ static int do_reset(struct repository *r,
>         struct strbuf ref_name = STRBUF_INIT;
>         struct object_id oid;
>         struct lock_file lock = LOCK_INIT;
> -       struct tree_desc desc;
> +       struct tree_desc desc = { 0 };
>         struct tree *tree;
>         struct unpack_trees_options unpack_tree_opts;
>         int ret = 0;
> @@ -3684,10 +3684,8 @@ static int do_reset(struct repository *r,
>                 strbuf_addf(&ref_name, "refs/rewritten/%.*s", len, name);
>                 if (get_oid(ref_name.buf, &oid) &&
>                     get_oid(ref_name.buf + strlen("refs/rewritten/"), &oid)) {
> -                       error(_("could not read '%s'"), ref_name.buf);
> -                       rollback_lock_file(&lock);
> -                       strbuf_release(&ref_name);
> -                       return -1;
> +                       ret = error(_("could not read '%s'"), ref_name.buf);
> +                       goto cleanup;
>                 }
>         }
>
> @@ -3703,24 +3701,18 @@ static int do_reset(struct repository *r,
>         init_checkout_metadata(&unpack_tree_opts.meta, name, &oid, NULL);
>
>         if (repo_read_index_unmerged(r)) {
> -               rollback_lock_file(&lock);
> -               strbuf_release(&ref_name);
> -               return error_resolve_conflict(_(action_name(opts)));
> +               ret = error_resolve_conflict(_(action_name(opts)));
> +               goto cleanup;
>         }
>
>         if (!fill_tree_descriptor(r, &desc, &oid)) {
> -               error(_("failed to find tree of %s"), oid_to_hex(&oid));
> -               rollback_lock_file(&lock);
> -               free((void *)desc.buffer);
> -               strbuf_release(&ref_name);
> -               return -1;
> +               ret = error(_("failed to find tree of %s"), oid_to_hex(&oid));
> +               goto cleanup;
>         }
>
>         if (unpack_trees(1, &desc, &unpack_tree_opts)) {
> -               rollback_lock_file(&lock);
> -               free((void *)desc.buffer);
> -               strbuf_release(&ref_name);
> -               return -1;
> +               ret = -1;
> +               goto cleanup;
>         }
>
>         tree = parse_tree_indirect(&oid);
> @@ -3728,13 +3720,15 @@ static int do_reset(struct repository *r,
>
>         if (write_locked_index(r->index, &lock, COMMIT_LOCK) < 0)
>                 ret = error(_("could not write index"));
> -       free((void *)desc.buffer);
>
>         if (!ret)
>                 ret = update_ref(reflog_message(opts, "reset", "'%.*s'",
>                                                 len, name), "HEAD", &oid,
>                                  NULL, 0, UPDATE_REFS_MSG_ON_ERR);
> -
> +cleanup:
> +       free((void *)desc.buffer);
> +       if (ret < 0)
> +               rollback_lock_file(&lock);
>         strbuf_release(&ref_name);
>         return ret;
>  }
> --
> 2.33.0.1446.g6af949f83bd
>

  reply	other threads:[~2021-10-07 16:07 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-06  9:40 [PATCH 0/2] dir & unpak-trees: memory-leak fixes Ævar Arnfjörð Bjarmason
2021-10-06  9:40 ` [PATCH 1/2] unpack-trees: don't leak memory in verify_clean_subdirectory() Ævar Arnfjörð Bjarmason
2021-10-06 15:58   ` Elijah Newren
2021-10-06  9:40 ` [PATCH 2/2] built-ins & lib: plug memory leaks with unpack_trees_options_release() Ævar Arnfjörð Bjarmason
2021-10-06 16:12   ` Elijah Newren
2021-10-06 16:33 ` [PATCH 0/2] dir & unpak-trees: memory-leak fixes Elijah Newren
2021-10-07  9:46 ` [PATCH v2 0/3] unpack-trees: " Ævar Arnfjörð Bjarmason
2021-10-07  9:46   ` [PATCH v2 1/3] unpack-trees: don't leak memory in verify_clean_subdirectory() Ævar Arnfjörð Bjarmason
2021-10-07 22:27     ` Junio C Hamano
2021-10-07  9:46   ` [PATCH v2 2/3] sequencer: add a "goto cleanup" to do_reset() Ævar Arnfjörð Bjarmason
2021-10-07 16:06     ` Elijah Newren [this message]
2021-10-07  9:46   ` [PATCH v2 3/3] sequencer: fix a memory leak in do_reset() Ævar Arnfjörð Bjarmason
2021-10-07 16:10   ` [PATCH v2 0/3] unpack-trees: memory-leak fixes Elijah Newren
2021-10-13 13:23   ` [PATCH v3 " Ævar Arnfjörð Bjarmason
2021-10-13 13:23     ` [PATCH v3 1/3] unpack-trees: don't leak memory in verify_clean_subdirectory() Ævar Arnfjörð Bjarmason
2021-10-13 13:23     ` [PATCH v3 2/3] sequencer: add a "goto cleanup" to do_reset() Ævar Arnfjörð Bjarmason
2021-10-13 13:23     ` [PATCH v3 3/3] sequencer: fix a memory leak in do_reset() Ævar Arnfjörð Bjarmason

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: http://vger.kernel.org/majordomo-info.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CABPp-BH=3DP-dXRCphY53-3eZd1TU8h5GY_M12nnbEGm-UYB9Q@mail.gmail.com' \
    --to=newren@gmail.com \
    --cc=ajrhunt@google.com \
    --cc=avarab@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=martin.agren@gmail.com \
    --cc=peff@peff.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).