* [PATCH] clone: plug a miniscule leak
@ 2022-05-01 5:17 Junio C Hamano
2022-05-02 13:43 ` Derrick Stolee
0 siblings, 1 reply; 4+ messages in thread
From: Junio C Hamano @ 2022-05-01 5:17 UTC (permalink / raw)
To: git
The remote_name variable is first assigned a copy of the value of
the "clone.defaultremotename" configuration variable and then by the
value of the "--origin" command line option. The former is prepared
to see multiple instances of the variable by freeing the current
value of the variable before a copy of the newly discovered value
gets assigned to it. The latter blindly assigned a copy of the new
value to it, thereby leaking the value read from the configuration.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
This was discovered by a recently removed bogus coccinelle
rewrite rule---if you apply an incorrect change suggested by the
bogus rewrite rule to unconditionally assign to remote_name a
copy of option_origin, or NULL, the value read from the
configuration would be lost and never be used, thereby breaking
a test to ensure the configuration is used, instead of the
default remote nickname "origin".
Perhaps a Coccinelle rule like this might have caught similar
leaks:
@@
expression E;
expression V;
@@
- if (E)
- V = xstrdup(E);
+ if (E) {
+ free(V);
+ V = xstrdup(E);
+ }
The fact that the result of xstrdup() is assigned to V is that V
is meant to hold a pointer to an allocated piece of memory.
With the preimage of the above semantic patch, it is reasonable
to expect that V may be initialized to NULL or may be holding a
pointer to a piece of allocated memory when the control reaches
here, because otherwise, V will be either need to be freed (when
E was not NULL, in which case we assigned the result of
xstrdup() to it) or V has garbage that cannot be freed later.
builtin/clone.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git c/builtin/clone.c w/builtin/clone.c
index 5231656379..194d50f75f 100644
--- c/builtin/clone.c
+++ w/builtin/clone.c
@@ -1106,8 +1106,10 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
* apply the remote name provided by --origin only after this second
* call to git_config, to ensure it overrides all config-based values.
*/
- if (option_origin != NULL)
+ if (option_origin != NULL) {
+ free(remote_name);
remote_name = xstrdup(option_origin);
+ }
if (remote_name == NULL)
remote_name = xstrdup("origin");
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] clone: plug a miniscule leak
2022-05-01 5:17 [PATCH] clone: plug a miniscule leak Junio C Hamano
@ 2022-05-02 13:43 ` Derrick Stolee
2022-05-02 17:12 ` Junio C Hamano
0 siblings, 1 reply; 4+ messages in thread
From: Derrick Stolee @ 2022-05-02 13:43 UTC (permalink / raw)
To: Junio C Hamano, git
On 5/1/2022 1:17 AM, Junio C Hamano wrote:
> Perhaps a Coccinelle rule like this might have caught similar
> leaks:
>
> @@
> expression E;
> expression V;
> @@
> - if (E)
> - V = xstrdup(E);
> + if (E) {
> + free(V);
> + V = xstrdup(E);
> + }
>
> The fact that the result of xstrdup() is assigned to V is that V
> is meant to hold a pointer to an allocated piece of memory.
>
> With the preimage of the above semantic patch, it is reasonable
> to expect that V may be initialized to NULL or may be holding a
> pointer to a piece of allocated memory when the control reaches
> here, because otherwise, V will be either need to be freed (when
> E was not NULL, in which case we assigned the result of
> xstrdup() to it) or V has garbage that cannot be freed later.
Initially, I did think "what if V is not initialized to NULL?" but
you are right that the code would already be broken in that case.
> - if (option_origin != NULL)
This technically wouldn't hit your rule, since "E" isn't just the
variable name, as we typically do with our style. Is that something
that Coccinelle automatically simplifies?
> + if (option_origin != NULL) {
Do you want to take this opportunity to drop the "!= NULL" here?
> + free(remote_name);
> remote_name = xstrdup(option_origin);
> + }
> > if (remote_name == NULL)
Or do you want to keep similar style from the surrounding code?
Either way, looks good to me.
Thanks,
-Stolee
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] clone: plug a miniscule leak
2022-05-02 13:43 ` Derrick Stolee
@ 2022-05-02 17:12 ` Junio C Hamano
2022-05-02 20:39 ` Philip Oakley
0 siblings, 1 reply; 4+ messages in thread
From: Junio C Hamano @ 2022-05-02 17:12 UTC (permalink / raw)
To: Derrick Stolee; +Cc: git
Derrick Stolee <derrickstolee@github.com> writes:
>> - if (option_origin != NULL)
>
> This technically wouldn't hit your rule, since "E" isn't just the
> variable name, as we typically do with our style. Is that something
> that Coccinelle automatically simplifies?
>
>> + if (option_origin != NULL) {
>
> Do you want to take this opportunity to drop the "!= NULL" here?
>
>> + free(remote_name);
>> remote_name = xstrdup(option_origin);
>> + }
>> > if (remote_name == NULL)
>
> Or do you want to keep similar style from the surrounding code?
I think that it is better to leave that particular clean-up to
the equals-null.cocci topic started by Elia; I know having them
separate would cause a merge conflict, but even if I change them
here, it will result in the same merge conflict anyway ;-)
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] clone: plug a miniscule leak
2022-05-02 17:12 ` Junio C Hamano
@ 2022-05-02 20:39 ` Philip Oakley
0 siblings, 0 replies; 4+ messages in thread
From: Philip Oakley @ 2022-05-02 20:39 UTC (permalink / raw)
To: Junio C Hamano, Derrick Stolee; +Cc: git, Elia Pinto
CC'ing Elia as per the mention. Thread
https://lore.kernel.org/git/xmqqzgjzzwnd.fsf@gitster.g/t/#u.
On 02/05/2022 18:12, Junio C Hamano wrote: <xmqqzgjzzwnd.fsf@gitster.g>
> Derrick Stolee <derrickstolee@github.com> writes:
>
>>> - if (option_origin != NULL)
>> This technically wouldn't hit your rule, since "E" isn't just the
>> variable name, as we typically do with our style. Is that something
>> that Coccinelle automatically simplifies?
>>
>>> + if (option_origin != NULL) {
>> Do you want to take this opportunity to drop the "!= NULL" here?
>>
>>> + free(remote_name);
>>> remote_name = xstrdup(option_origin);
>>> + }
>>>> if (remote_name == NULL)
>> Or do you want to keep similar style from the surrounding code?
> I think that it is better to leave that particular clean-up to
> the equals-null.cocci topic started by Elia; I know having them
> separate would cause a merge conflict, but even if I change them
> here, it will result in the same merge conflict anyway ;-)
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2022-05-02 20:41 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-05-01 5:17 [PATCH] clone: plug a miniscule leak Junio C Hamano
2022-05-02 13:43 ` Derrick Stolee
2022-05-02 17:12 ` Junio C Hamano
2022-05-02 20:39 ` Philip Oakley
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).