threads / patch / 63420

patchREFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed

Subject: [PATCH] REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed

## tl;dr

10 messages between May 8, 2025 and May 9, 2025. Diffs are folded; open one to read it.

replies: 9people: 5as markdown or json

Lidong Yan via GitGitGadget· May 8, 2025, 13:39 UTC · lore
From: Lidong Yan <502024330056@smail.nju.edu.cn>

REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc failed. This leak can be fixed by add a free(x) before set x to NULL.

Signed-off-by: Lidong Yan <502024330056@smail.nju.edu.cn>
---
    REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed
    
    REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc
    failed. This leak can be fixed by add a free(x) before set x to NULL.
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1955%2Fbrandb97%2Ffix-REFTABLE-REALLOC-leak-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1955/brandb97/fix-REFTABLE-REALLOC-leak-v1
Pull-Request: https://github.com/git/git/pull/1955
 reftable/basics.h | 28 ++++++++++++++++++++--------
 1 file changed, 20 insertions(+), 8 deletions(-)
Show changes to reftable/basics.h +20 −8
diff --git a/reftable/basics.h b/reftable/basics.h
index d8888c12629..c7651f0cda8 100644
--- a/reftable/basics.h
+++ b/reftable/basics.h
@@ -200,14 +200,26 @@ static inline int reftable_alloc_size(size_t nelem, size_t elsize, size_t *out)
 		} \
 	} while (0)
 #define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
-#define REFTABLE_REALLOC_ARRAY(x, alloc) do { \
-		size_t alloc_size; \
-		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < 0) { \
-			errno = ENOMEM; \
-			(x) = NULL; \
-		} else { \
-			(x) = reftable_realloc((x), alloc_size); \
-		} \
+#define REFTABLE_REALLOC_ARRAY(x, alloc)                                      \
+	do {                                                                  \
+		size_t alloc_size;                                            \
+		void *new_p;                                                  \
+		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < \
+		    0) {                                                      \
+			goto cleanup;                                         \
+		} else {                                                      \
+			new_p = reftable_realloc((x), alloc_size);            \
+			if (!new_p) {                                         \
+				goto cleanup;                                 \
+			}                                                     \
+			(x) = new_p;                                          \
+		}                                                             \
+		break;                                                        \
+	cleanup:                                                              \
+		if (x)                                                        \
+			free(x);                                              \
+		errno = ENOMEM;                                               \
+		(x) = NULL;                                                   \
 	} while (0)
 
 static inline void *reftable_alloc_grow(void *p, size_t nelem, size_t elsize,

base-commit: 6f84262c44a89851c3ae5a6e4c1a9d06b2068d75
-- 
gitgitgadget
René Scharfe· May 8, 2025, 21:42 UTC · re: Lidong Yan via GitGitGadget · lore

Re: [PATCH] REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed

Am 08.05.25 um 15:39 schrieb Lidong Yan via GitGitGadget:
> From: Lidong Yan <502024330056@smail.nju.edu.cn>
> 
> REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc
> failed. This leak can be fixed by add a free(x) before set x to NULL.
Hmm, this macro is unused.  Perhaps remove it?
René
lidongyan· May 9, 2025, 01:54 UTC · re: René Scharfe · lore

Re: [PATCH] REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed

Ok, I will remove it in the next patch.
Show 13 quoted lines
> 2025年5月9日 05:42,René Scharfe <l.s.r@web.de> 写道:
> 
> Am 08.05.25 um 15:39 schrieb Lidong Yan via GitGitGadget:
>> From: Lidong Yan <502024330056@smail.nju.edu.cn>
>> 
>> REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc
>> failed. This leak can be fixed by add a free(x) before set x to NULL.
> 
> Hmm, this macro is unused.  Perhaps remove it?
> 
> René
> 
> 
Junio C Hamano· May 9, 2025, 03:43 UTC · re: René Scharfe · lore

Re: [PATCH] REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed

René Scharfe <l.s.r@web.de> writes:
Show 7 quoted lines
> Am 08.05.25 um 15:39 schrieb Lidong Yan via GitGitGadget:
>> From: Lidong Yan <502024330056@smail.nju.edu.cn>
>> 
>> REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc
>> failed. This leak can be fixed by add a free(x) before set x to NULL.
>
> Hmm, this macro is unused.  Perhaps remove it?
Or let the sleeping dog alone, perhaps?
Lidong Yan via GitGitGadget· May 9, 2025, 02:04 UTC · re: Lidong Yan via GitGitGadget · lore

[PATCH v2] REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro

From: Lidong Yan <502024330056@smail.nju.edu.cn>

REFTABLE_REALLOC_ARRAY will cause memory leak if realloc failed. Since it is unused, remove this unsafe macro.

Signed-off-by: Lidong Yan <502024330056@smail.nju.edu.cn>
---
    REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed
    
    REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc
    failed. This leak can be fixed by add a free(x) before set x to NULL.
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1955%2Fbrandb97%2Ffix-REFTABLE-REALLOC-leak-v2
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1955/brandb97/fix-REFTABLE-REALLOC-leak-v2
Pull-Request: https://github.com/git/git/pull/1955
Range-diff vs v1:
 1:  4d786a0ec17 ! 1:  6cc191f9db8 REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed
     @@ Metadata
      Author: Lidong Yan <502024330056@smail.nju.edu.cn>
      
       ## Commit message ##
     -    REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed
     +    REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro
      
     -    REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc
     -    failed. This leak can be fixed by add a free(x) before set x to NULL.
     +    REFTABLE_REALLOC_ARRAY will cause memory leak if realloc failed.
     +    Since it is unused, remove this unsafe macro.
      
          Signed-off-by: Lidong Yan <502024330056@smail.nju.edu.cn>
      
       ## reftable/basics.h ##
      @@ reftable/basics.h: static inline int reftable_alloc_size(size_t nelem, size_t elsize, size_t *out)
     + 			(x) = reftable_malloc(alloc_size); \
       		} \
       	} while (0)
     - #define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
     +-#define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
      -#define REFTABLE_REALLOC_ARRAY(x, alloc) do { \
      -		size_t alloc_size; \
      -		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < 0) { \
     @@ reftable/basics.h: static inline int reftable_alloc_size(size_t nelem, size_t el
      -		} else { \
      -			(x) = reftable_realloc((x), alloc_size); \
      -		} \
     -+#define REFTABLE_REALLOC_ARRAY(x, alloc)                                      \
     -+	do {                                                                  \
     -+		size_t alloc_size;                                            \
     -+		void *new_p;                                                  \
     -+		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < \
     -+		    0) {                                                      \
     -+			goto cleanup;                                         \
     -+		} else {                                                      \
     -+			new_p = reftable_realloc((x), alloc_size);            \
     -+			if (!new_p) {                                         \
     -+				goto cleanup;                                 \
     -+			}                                                     \
     -+			(x) = new_p;                                          \
     -+		}                                                             \
     -+		break;                                                        \
     -+	cleanup:                                                              \
     -+		if (x)                                                        \
     -+			free(x);                                              \
     -+		errno = ENOMEM;                                               \
     -+		(x) = NULL;                                                   \
     - 	} while (0)
     +-	} while (0)
     ++#define REFTABLE_CALLOC_ARRAY(x, alloc) \
     ++	(x) = reftable_calloc((alloc), sizeof(*(x)))
       
       static inline void *reftable_alloc_grow(void *p, size_t nelem, size_t elsize,
     + 					size_t *allocp)
 reftable/basics.h | 12 ++----------
 1 file changed, 2 insertions(+), 10 deletions(-)
Show changes to reftable/basics.h +2 −10
diff --git a/reftable/basics.h b/reftable/basics.h
index d8888c12629..667feffd935 100644
--- a/reftable/basics.h
+++ b/reftable/basics.h
@@ -199,16 +199,8 @@ static inline int reftable_alloc_size(size_t nelem, size_t elsize, size_t *out)
 			(x) = reftable_malloc(alloc_size); \
 		} \
 	} while (0)
-#define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
-#define REFTABLE_REALLOC_ARRAY(x, alloc) do { \
-		size_t alloc_size; \
-		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < 0) { \
-			errno = ENOMEM; \
-			(x) = NULL; \
-		} else { \
-			(x) = reftable_realloc((x), alloc_size); \
-		} \
-	} while (0)
+#define REFTABLE_CALLOC_ARRAY(x, alloc) \
+	(x) = reftable_calloc((alloc), sizeof(*(x)))
 
 static inline void *reftable_alloc_grow(void *p, size_t nelem, size_t elsize,
 					size_t *allocp)

base-commit: 6f84262c44a89851c3ae5a6e4c1a9d06b2068d75
-- 
gitgitgadget
Patrick Steinhardt· May 9, 2025, 06:51 UTC · re: Lidong Yan via GitGitGadget · lore

Re: [PATCH v2] REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro

On Fri, May 09, 2025 at 02:04:22AM +0000, Lidong Yan via GitGitGadget wrote:
Show 20 quoted lines
> diff --git a/reftable/basics.h b/reftable/basics.h
> index d8888c12629..667feffd935 100644
> --- a/reftable/basics.h
> +++ b/reftable/basics.h
> @@ -199,16 +199,8 @@ static inline int reftable_alloc_size(size_t nelem, size_t elsize, size_t *out)
>  			(x) = reftable_malloc(alloc_size); \
>  		} \
>  	} while (0)
> -#define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
> -#define REFTABLE_REALLOC_ARRAY(x, alloc) do { \
> -		size_t alloc_size; \
> -		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < 0) { \
> -			errno = ENOMEM; \
> -			(x) = NULL; \
> -		} else { \
> -			(x) = reftable_realloc((x), alloc_size); \
> -		} \
> -	} while (0)
> +#define REFTABLE_CALLOC_ARRAY(x, alloc) \
> +	(x) = reftable_calloc((alloc), sizeof(*(x)))

Let's avoid reformatting unrelated macros. But other than that I fully agree -- we should remove stuff that we don't use in the first place.

Thanks!
Patrick
lidongyan· May 9, 2025, 07:38 UTC · re: Patrick Steinhardt · lore

Re: [PATCH v2] REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro

2025年5月9日 14:51,Patrick Steinhardt <ps@pks.im> 写道:
Show 30 quoted lines
> 
> On Fri, May 09, 2025 at 02:04:22AM +0000, Lidong Yan via GitGitGadget wrote:
>> diff --git a/reftable/basics.h b/reftable/basics.h
>> index d8888c12629..667feffd935 100644
>> --- a/reftable/basics.h
>> +++ b/reftable/basics.h
>> @@ -199,16 +199,8 @@ static inline int reftable_alloc_size(size_t nelem, size_t elsize, size_t *out)
>> (x) = reftable_malloc(alloc_size); \
>> } \
>> } while (0)
>> -#define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
>> -#define REFTABLE_REALLOC_ARRAY(x, alloc) do { \
>> - size_t alloc_size; \
>> - if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < 0) { \
>> - errno = ENOMEM; \
>> - (x) = NULL; \
>> - } else { \
>> - (x) = reftable_realloc((x), alloc_size); \
>> - } \
>> - } while (0)
>> +#define REFTABLE_CALLOC_ARRAY(x, alloc) \
>> + (x) = reftable_calloc((alloc), sizeof(*(x)))
> 
> Let's avoid reformatting unrelated macros. But other than that I fully
> agree -- we should remove stuff that we don't use in the first place.
> 
> Thanks!
> 
> Patrick
> 
Ok, I will restore REFTABLE_CALLOC_ARRAY in the next patch.
Lidong Yan via GitGitGadget· May 9, 2025, 07:44 UTC · re: Lidong Yan via GitGitGadget · lore

[PATCH v3] REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro

From: Lidong Yan <502024330056@smail.nju.edu.cn>

REFTABLE_REALLOC_ARRAY will cause memory leak if realloc failed. Since it is unused, remove this unsafe macro.

Signed-off-by: Lidong Yan <502024330056@smail.nju.edu.cn>
---
    REFTABLE_REALLOC_ARRAY: fix potential memory leak if realloc failed
    
    REFTABLE_REALLOC_ARRAY doesn't free origin pointer when reftable_realloc
    failed. This leak can be fixed by add a free(x) before set x to NULL.
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1955%2Fbrandb97%2Ffix-REFTABLE-REALLOC-leak-v3
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1955/brandb97/fix-REFTABLE-REALLOC-leak-v3
Pull-Request: https://github.com/git/git/pull/1955
Range-diff vs v2:
 1:  6cc191f9db8 ! 1:  107f9ce3bd0 REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro
     @@ Commit message
      
       ## reftable/basics.h ##
      @@ reftable/basics.h: static inline int reftable_alloc_size(size_t nelem, size_t elsize, size_t *out)
     - 			(x) = reftable_malloc(alloc_size); \
       		} \
       	} while (0)
     --#define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
     + #define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
      -#define REFTABLE_REALLOC_ARRAY(x, alloc) do { \
      -		size_t alloc_size; \
      -		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < 0) { \
     @@ reftable/basics.h: static inline int reftable_alloc_size(size_t nelem, size_t el
      -			(x) = reftable_realloc((x), alloc_size); \
      -		} \
      -	} while (0)
     -+#define REFTABLE_CALLOC_ARRAY(x, alloc) \
     -+	(x) = reftable_calloc((alloc), sizeof(*(x)))
       
       static inline void *reftable_alloc_grow(void *p, size_t nelem, size_t elsize,
       					size_t *allocp)
 reftable/basics.h | 9 ---------
 1 file changed, 9 deletions(-)
Show changes to reftable/basics.h +0 −9
diff --git a/reftable/basics.h b/reftable/basics.h
index d8888c12629..1674a6f8b80 100644
--- a/reftable/basics.h
+++ b/reftable/basics.h
@@ -200,15 +200,6 @@ static inline int reftable_alloc_size(size_t nelem, size_t elsize, size_t *out)
 		} \
 	} while (0)
 #define REFTABLE_CALLOC_ARRAY(x, alloc) (x) = reftable_calloc((alloc), sizeof(*(x)))
-#define REFTABLE_REALLOC_ARRAY(x, alloc) do { \
-		size_t alloc_size; \
-		if (reftable_alloc_size(sizeof(*(x)), (alloc), &alloc_size) < 0) { \
-			errno = ENOMEM; \
-			(x) = NULL; \
-		} else { \
-			(x) = reftable_realloc((x), alloc_size); \
-		} \
-	} while (0)
 
 static inline void *reftable_alloc_grow(void *p, size_t nelem, size_t elsize,
 					size_t *allocp)

base-commit: 6f84262c44a89851c3ae5a6e4c1a9d06b2068d75
-- 
gitgitgadget
Patrick Steinhardt· May 9, 2025, 08:23 UTC · re: Lidong Yan via GitGitGadget · lore

Re: [PATCH v3] REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro

On Fri, May 09, 2025 at 07:44:46AM +0000, Lidong Yan via GitGitGadget wrote:
> From: Lidong Yan <502024330056@smail.nju.edu.cn>
> 
> REFTABLE_REALLOC_ARRAY will cause memory leak if realloc failed.
> Since it is unused, remove this unsafe macro.
This looks good to me, thanks!
Patrick
Junio C Hamano· May 9, 2025, 14:25 UTC · re: Patrick Steinhardt · lore

Re: [PATCH v3] REFTABLE_REALLOC_ARRAY: remove this unsafe yet unused macro

Patrick Steinhardt <ps@pks.im> writes:
Show 7 quoted lines
> On Fri, May 09, 2025 at 07:44:46AM +0000, Lidong Yan via GitGitGadget wrote:
>> From: Lidong Yan <502024330056@smail.nju.edu.cn>
>> 
>> REFTABLE_REALLOC_ARRAY will cause memory leak if realloc failed.
>> Since it is unused, remove this unsafe macro.
>
> This looks good to me, thanks!

Alright. As long as we do not foresee adding a new caller of this in the near future, removal is always preferred ;-)

Thanks.

← back to recent threads