From: Phillip Wood <phillip.wood123@gmail.com>
To: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>, git@vger.kernel.org
Cc: "Johannes Schindelin" <Johannes.Schindelin@gmx.de>,
"Junio C Hamano" <gitster@pobox.com>,
"Taylor Blau" <me@ttaylorr.com>, "René Scharfe" <l.s.r@web.de>
Subject: Re: [PATCH 03/10] rebase & sequencer API: fix get_replay_opts() leak in "rebase"
Date: Sat, 31 Dec 2022 14:54:52 +0000 [thread overview]
Message-ID: <0fd32979-9beb-93e8-8a03-253a08588c42@dunelm.org.uk> (raw)
In-Reply-To: <patch-03.10-3e9c4df61fe-20221230T071741Z-avarab@gmail.com>
Hi Ævar
On 30/12/2022 07:28, Ævar Arnfjörð Bjarmason wrote:
> Make the recently added replay_opts_release() function non-static and
> use it for freeing the "struct replay_opts" constructed by the
> get_replay_opts() function in "builtin/rebase.c". See [1] for the
> initial addition of get_replay_opts().
>
> To safely call our new replay_opts_release() we'll need to change all
> the free() to a FREE_AND_NULL(), and set "xopts_nr" to "0" after we
> loop over it and free() it (the free() in the loop doesn't need to be
> a FREE_AND_NULL()).
>
> This is because in e.g. do_interactive_rebase() we construct a "struct
> replay_opts" with "get_replay_opts()", and then call
> "complete_action()". If we get far enough in that function without
> encountering errors we'll call "pick_commits()" which (indirectly)
> calls sequencer_remove_state() at the end.
The way to fix this is to stop calling replay_opts_release() from
sequencer_remove_state(), then you can be sure to call
replay_opts_release() only once by adding calls to the functions that
initialize a "struct replay_opts". That would be much cleaner than this
FREE_AND_NULL() hackery.
Best Wishes
Phillip
> But if we encounter errors anywhere along the way we'd punt out early,
> and not free() the memory we allocated. Remembering whether we
> previously called sequencer_remove_state() would be a hassle, so let's
> make it safe to re-invoke replay_opts_release() instead.
>
> I experimented with a change to be more paranoid instead, i.e. to
> exhaustively check our state via an enum. We could make sure that we:
>
> - Only allow calling "replay_opts_release()" after
> "sequencer_remove_state()", but not the other way around.
>
> - Forbid invoking either function twice in a row.
>
> But such paranoia isn't warranted here, let's instead take the easy
> way out and FREE_AND_NULL() this.
>
> See [2] for the initial implementation of "sequencer_remove_state()",
> which assumed that it should be removing the full (including on-disk)
> rebase state as a one-off.
>
> 1. 73fdc535d26 (rebase -i: use struct rebase_options to parse args,
> 2019-04-17)
> 2. 26ae337be11 (revert: Introduce --reset to remove sequencer state,
> 2011-08-04)
>
> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
> ---
>
> Re the "I experimented with...", the earlier version can be seen at
> https://github.com/avar/git/commit/7a150d1b7e2; It's the commit I
> linked to from
> https://lore.kernel.org/git/221108.864jv9sc9r.gmgdl@evledraar.gmail.com/
> (which is mentioned in the CL).
>
> builtin/rebase.c | 2 ++
> sequencer.c | 13 +++++++------
> sequencer.h | 1 +
> t/t3405-rebase-malformed.sh | 1 +
> t/t3412-rebase-root.sh | 1 +
> t/t3423-rebase-reword.sh | 1 +
> t/t3437-rebase-fixup-options.sh | 1 +
> t/t3438-rebase-broken-files.sh | 2 ++
> t/t7402-submodule-rebase.sh | 1 +
> t/t9106-git-svn-commit-diff-clobber.sh | 1 -
> t/t9164-git-svn-dcommit-concurrent.sh | 1 -
> 11 files changed, 17 insertions(+), 8 deletions(-)
>
> diff --git a/builtin/rebase.c b/builtin/rebase.c
> index 7141fd5e0c1..91bf55be6e6 100644
> --- a/builtin/rebase.c
> +++ b/builtin/rebase.c
> @@ -301,6 +301,7 @@ static int do_interactive_rebase(struct rebase_options *opts, unsigned flags)
> }
>
> cleanup:
> + replay_opts_release(&replay);
> string_list_clear(&commands, 0);
> free(revisions);
> free(shortrevisions);
> @@ -343,6 +344,7 @@ static int run_sequencer_rebase(struct rebase_options *opts)
> struct replay_opts replay_opts = get_replay_opts(opts);
>
> ret = sequencer_continue(the_repository, &replay_opts);
> + replay_opts_release(&replay_opts);
> break;
> }
> case ACTION_EDIT_TODO:
> diff --git a/sequencer.c b/sequencer.c
> index 655ae9f1a72..e29a97b6caa 100644
> --- a/sequencer.c
> +++ b/sequencer.c
> @@ -351,15 +351,16 @@ static const char *gpg_sign_opt_quoted(struct replay_opts *opts)
> return buf.buf;
> }
>
> -static void replay_opts_release(struct replay_opts *opts)
> +void replay_opts_release(struct replay_opts *opts)
> {
> - free(opts->gpg_sign);
> - free(opts->reflog_action);
> - free(opts->default_strategy);
> - free(opts->strategy);
> + FREE_AND_NULL(opts->gpg_sign);
> + FREE_AND_NULL(opts->reflog_action);
> + FREE_AND_NULL(opts->default_strategy);
> + FREE_AND_NULL(opts->strategy);
> for (size_t i = 0; i < opts->xopts_nr; i++)
> free(opts->xopts[i]);
> - free(opts->xopts);
> + opts->xopts_nr = 0;
> + FREE_AND_NULL(opts->xopts);
> strbuf_release(&opts->current_fixups);
> }
>
> diff --git a/sequencer.h b/sequencer.h
> index 888c18aad71..3bcdfa1b586 100644
> --- a/sequencer.h
> +++ b/sequencer.h
> @@ -158,6 +158,7 @@ int sequencer_pick_revisions(struct repository *repo,
> int sequencer_continue(struct repository *repo, struct replay_opts *opts);
> int sequencer_rollback(struct repository *repo, struct replay_opts *opts);
> int sequencer_skip(struct repository *repo, struct replay_opts *opts);
> +void replay_opts_release(struct replay_opts *opts);
> int sequencer_remove_state(struct replay_opts *opts);
>
> #define TODO_LIST_KEEP_EMPTY (1U << 0)
> diff --git a/t/t3405-rebase-malformed.sh b/t/t3405-rebase-malformed.sh
> index 25243318618..8979bc34073 100755
> --- a/t/t3405-rebase-malformed.sh
> +++ b/t/t3405-rebase-malformed.sh
> @@ -5,6 +5,7 @@ test_description='rebase should handle arbitrary git message'
> GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=main
> export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME
>
> +TEST_PASSES_SANITIZE_LEAK=true
> . ./test-lib.sh
> . "$TEST_DIRECTORY"/lib-rebase.sh
>
> diff --git a/t/t3412-rebase-root.sh b/t/t3412-rebase-root.sh
> index 58371d8a547..e75b3d0e07c 100755
> --- a/t/t3412-rebase-root.sh
> +++ b/t/t3412-rebase-root.sh
> @@ -7,6 +7,7 @@ Tests if git rebase --root --onto <newparent> can rebase the root commit.
> GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=main
> export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME
>
> +TEST_PASSES_SANITIZE_LEAK=true
> . ./test-lib.sh
>
> log_with_names () {
> diff --git a/t/t3423-rebase-reword.sh b/t/t3423-rebase-reword.sh
> index 4859bb8f722..2fab703d615 100755
> --- a/t/t3423-rebase-reword.sh
> +++ b/t/t3423-rebase-reword.sh
> @@ -2,6 +2,7 @@
>
> test_description='git rebase interactive with rewording'
>
> +TEST_PASSES_SANITIZE_LEAK=true
> . ./test-lib.sh
>
> . "$TEST_DIRECTORY"/lib-rebase.sh
> diff --git a/t/t3437-rebase-fixup-options.sh b/t/t3437-rebase-fixup-options.sh
> index c023fefd681..274699dadb8 100755
> --- a/t/t3437-rebase-fixup-options.sh
> +++ b/t/t3437-rebase-fixup-options.sh
> @@ -14,6 +14,7 @@ to the "fixup" command that works with "fixup!", "fixup -C" works with
> "amend!" upon --autosquash.
> '
>
> +TEST_PASSES_SANITIZE_LEAK=true
> . ./test-lib.sh
>
> . "$TEST_DIRECTORY"/lib-rebase.sh
> diff --git a/t/t3438-rebase-broken-files.sh b/t/t3438-rebase-broken-files.sh
> index b92a3ce46b8..c614c4f2e4b 100755
> --- a/t/t3438-rebase-broken-files.sh
> +++ b/t/t3438-rebase-broken-files.sh
> @@ -1,6 +1,8 @@
> #!/bin/sh
>
> test_description='rebase behavior when on-disk files are broken'
> +
> +TEST_PASSES_SANITIZE_LEAK=true
> . ./test-lib.sh
>
> test_expect_success 'set up conflicting branches' '
> diff --git a/t/t7402-submodule-rebase.sh b/t/t7402-submodule-rebase.sh
> index ebeca12a711..b19792b3269 100755
> --- a/t/t7402-submodule-rebase.sh
> +++ b/t/t7402-submodule-rebase.sh
> @@ -5,6 +5,7 @@
>
> test_description='Test rebasing, stashing, etc. with submodules'
>
> +TEST_PASSES_SANITIZE_LEAK=true
> . ./test-lib.sh
>
> test_expect_success setup '
> diff --git a/t/t9106-git-svn-commit-diff-clobber.sh b/t/t9106-git-svn-commit-diff-clobber.sh
> index 3cab0b9720a..bca496c40e0 100755
> --- a/t/t9106-git-svn-commit-diff-clobber.sh
> +++ b/t/t9106-git-svn-commit-diff-clobber.sh
> @@ -3,7 +3,6 @@
> # Copyright (c) 2006 Eric Wong
> test_description='git svn commit-diff clobber'
>
> -TEST_FAILS_SANITIZE_LEAK=true
> . ./lib-git-svn.sh
>
> test_expect_success 'initialize repo' '
> diff --git a/t/t9164-git-svn-dcommit-concurrent.sh b/t/t9164-git-svn-dcommit-concurrent.sh
> index 1465156072e..c8e6c0733f4 100755
> --- a/t/t9164-git-svn-dcommit-concurrent.sh
> +++ b/t/t9164-git-svn-dcommit-concurrent.sh
> @@ -5,7 +5,6 @@
>
> test_description='concurrent git svn dcommit'
>
> -TEST_FAILS_SANITIZE_LEAK=true
> . ./lib-git-svn.sh
>
>
next prev parent reply other threads:[~2022-12-31 14:54 UTC|newest]
Thread overview: 60+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-30 7:28 [PATCH 00/10] sequencer API & users: fix widespread leaks Ævar Arnfjörð Bjarmason
2022-12-30 7:28 ` [PATCH 01/10] rebase: use "cleanup" pattern in do_interactive_rebase() Ævar Arnfjörð Bjarmason
2022-12-31 14:50 ` Phillip Wood
2022-12-30 7:28 ` [PATCH 02/10] sequencer.c: split up sequencer_remove_state() Ævar Arnfjörð Bjarmason
2022-12-30 17:35 ` René Scharfe
2022-12-31 14:51 ` Phillip Wood
2022-12-30 7:28 ` [PATCH 03/10] rebase & sequencer API: fix get_replay_opts() leak in "rebase" Ævar Arnfjörð Bjarmason
2022-12-31 14:54 ` Phillip Wood [this message]
2022-12-30 7:28 ` [PATCH 04/10] builtin/revert.c: refactor run_sequencer() return pattern Ævar Arnfjörð Bjarmason
2023-01-01 4:25 ` Junio C Hamano
2022-12-30 7:28 ` [PATCH 05/10] builtin/revert.c: fix common leak by using replay_opts_release() Ævar Arnfjörð Bjarmason
2022-12-30 23:37 ` René Scharfe
2022-12-31 14:55 ` Phillip Wood
2022-12-30 7:28 ` [PATCH 06/10] builtin/revert.c: move free-ing of "revs" to replay_opts_release() Ævar Arnfjörð Bjarmason
2022-12-30 7:28 ` [PATCH 07/10] builtin/rebase.c: fix "options.onto_name" leak Ævar Arnfjörð Bjarmason
2022-12-31 14:59 ` Phillip Wood
2022-12-30 7:28 ` [PATCH 08/10] sequencer.c: always free() the "msgbuf" in do_pick_commit() Ævar Arnfjörð Bjarmason
2022-12-31 15:03 ` Phillip Wood
2022-12-30 7:28 ` [PATCH 09/10] builtin/rebase.c: free() "options.strategy_opts" Ævar Arnfjörð Bjarmason
2022-12-30 7:28 ` [PATCH 10/10] commit.c: free() revs.commit in get_fork_point() Ævar Arnfjörð Bjarmason
2022-12-31 15:06 ` [PATCH 00/10] sequencer API & users: fix widespread leaks Phillip Wood
2023-01-01 4:27 ` Junio C Hamano
2023-01-12 12:45 ` [PATCH v2 0/9] " Ævar Arnfjörð Bjarmason
2023-01-12 12:45 ` [PATCH v2 1/9] rebase: use "cleanup" pattern in do_interactive_rebase() Ævar Arnfjörð Bjarmason
2023-01-12 12:45 ` [PATCH v2 2/9] sequencer.c: split up sequencer_remove_state() Ævar Arnfjörð Bjarmason
2023-01-12 12:45 ` [PATCH v2 3/9] rebase & sequencer API: fix get_replay_opts() leak in "rebase" Ævar Arnfjörð Bjarmason
2023-01-13 10:34 ` Phillip Wood
2023-01-12 12:45 ` [PATCH v2 4/9] builtin/revert.c: move free-ing of "revs" to replay_opts_release() Ævar Arnfjörð Bjarmason
2023-01-13 10:36 ` Phillip Wood
2023-01-12 12:45 ` [PATCH v2 5/9] builtin/rebase.c: rename "squash_onto_name" to "to_free" Ævar Arnfjörð Bjarmason
2023-01-13 10:37 ` Phillip Wood
2023-01-12 12:45 ` [PATCH v2 6/9] builtin/rebase.c: fix "options.onto_name" leak Ævar Arnfjörð Bjarmason
2023-01-13 10:41 ` Phillip Wood
2023-01-12 12:45 ` [PATCH v2 7/9] sequencer.c: always free() the "msgbuf" in do_pick_commit() Ævar Arnfjörð Bjarmason
2023-01-12 12:46 ` [PATCH v2 8/9] builtin/rebase.c: free() "options.strategy_opts" Ævar Arnfjörð Bjarmason
2023-01-12 12:46 ` [PATCH v2 9/9] commit.c: free() revs.commit in get_fork_point() Ævar Arnfjörð Bjarmason
2023-01-13 10:45 ` [PATCH v2 0/9] sequencer API & users: fix widespread leaks Phillip Wood
2023-01-13 20:47 ` Junio C Hamano
2023-01-18 16:09 ` [PATCH v3 0/8] " Ævar Arnfjörð Bjarmason
2023-01-18 16:09 ` [PATCH v3 1/8] rebase: use "cleanup" pattern in do_interactive_rebase() Ævar Arnfjörð Bjarmason
2023-01-18 16:09 ` [PATCH v3 2/8] sequencer.c: split up sequencer_remove_state() Ævar Arnfjörð Bjarmason
2023-01-18 16:09 ` [PATCH v3 3/8] rebase & sequencer API: fix get_replay_opts() leak in "rebase" Ævar Arnfjörð Bjarmason
2023-01-24 14:36 ` Phillip Wood
2023-01-18 16:09 ` [PATCH v3 4/8] builtin/revert.c: move free-ing of "revs" to replay_opts_release() Ævar Arnfjörð Bjarmason
2023-01-18 16:09 ` [PATCH v3 5/8] builtin/rebase.c: fix "options.onto_name" leak Ævar Arnfjörð Bjarmason
2023-01-24 14:40 ` Phillip Wood
2023-01-18 16:09 ` [PATCH v3 6/8] sequencer.c: always free() the "msgbuf" in do_pick_commit() Ævar Arnfjörð Bjarmason
2023-01-18 16:09 ` [PATCH v3 7/8] builtin/rebase.c: free() "options.strategy_opts" Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 0/8] sequencer API & users: fix widespread leaks Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 1/8] rebase: use "cleanup" pattern in do_interactive_rebase() Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 2/8] sequencer.c: split up sequencer_remove_state() Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 3/8] sequencer API users: fix get_replay_opts() leaks Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 4/8] builtin/revert.c: move free-ing of "revs" to replay_opts_release() Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 5/8] builtin/rebase.c: fix "options.onto_name" leak Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 6/8] sequencer.c: always free() the "msgbuf" in do_pick_commit() Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 7/8] builtin/rebase.c: free() "options.strategy_opts" Ævar Arnfjörð Bjarmason
2023-02-06 19:08 ` [PATCH v4 8/8] commit.c: free() revs.commit in get_fork_point() Ævar Arnfjörð Bjarmason
2023-02-07 10:19 ` [PATCH v4 0/8] sequencer API & users: fix widespread leaks Phillip Wood
2023-01-18 16:09 ` [PATCH v3 8/8] commit.c: free() revs.commit in get_fork_point() Ævar Arnfjörð Bjarmason
2023-01-24 14:41 ` [PATCH v3 0/8] sequencer API & users: fix widespread leaks Phillip Wood
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
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=0fd32979-9beb-93e8-8a03-253a08588c42@dunelm.org.uk \
--to=phillip.wood123@gmail.com \
--cc=Johannes.Schindelin@gmx.de \
--cc=avarab@gmail.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=l.s.r@web.de \
--cc=me@ttaylorr.com \
--cc=phillip.wood@dunelm.org.uk \
/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.
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 NNTP newsgroup(s).