[PATCH V2 2/2] scripts: dtc: fix memory leak after realloc

[PATCH V2 2/2] scripts: dtc: fix memory leak after realloc

[Maxim Zhukov]

This commit fixed memory leak after errors realloc.

Signed-off-by: Maxim Zhukov <mussitantesmortem@gmail.com>
---
 scripts/dtc/fdtput.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/scripts/dtc/fdtput.c b/scripts/dtc/fdtput.c
index f2197f5..1042319 100644
--- a/scripts/dtc/fdtput.c
+++ b/scripts/dtc/fdtput.c
@@ -75,8 +75,9 @@ static int encode_value(struct display_info *disp, char **arg, int arg_count,
 	char *ptr;		/* pointer to current value position */
 	int len;		/* length of this cell/string/byte */
 	int ival;
-	int upto;	/* the number of bytes we have written to buf */
+	int upto;		/* the number of bytes we have written to buf */
 	char fmt[3];
+	void *save_ptr = NULL;	/* save pointer to realloc */
 
 	upto = 0;
 
@@ -96,12 +97,15 @@ static int encode_value(struct display_info *disp, char **arg, int arg_count,
 		/* enlarge our value buffer by a suitable margin if needed */
 		if (upto + len > value_size) {
 			value_size = (upto + len) + 500;
-			value = realloc(value, value_size);
-			if (!value) {
+			void *save_ptr = realloc(value, value_size);
+
+			if (!save_ptr) {
+				free(value);
 				fprintf(stderr, "Out of mmory: cannot alloc "
 					"%d bytes\n", value_size);
 				return -1;
 			}
+			value = save_ptr;
 		}
 
 		ptr = value + upto;
-- 
2.7.1.1.g3617aa0

Re: [PATCH V2 2/2] scripts: dtc: fix memory leak after realloc

[Rob Herring]

On Tue, Apr 12, 2016 at 3:55 PM, Maxim Zhukov
<mussitantesmortem-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> This commit fixed memory leak after errors realloc.
>
> Signed-off-by: Maxim Zhukov <mussitantesmortem-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> ---
>  scripts/dtc/fdtput.c | 10 +++++++---

Again, your patch is against the kernel copy of dtc. It needs to be
against upstream dtc and then I can update the kernel copy once this
lands.

Rob

(unknown),

[Maxim Zhukov]

Subject: [PATCH RESEND 2/2] scripts: dtc: fix memory leak after realloc

This commit fixed memory leak after errors realloc.

Signed-off-by: Maxim Zhukov <mussitantesmortem-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
---
 scripts/dtc/fdtput.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/scripts/dtc/fdtput.c b/scripts/dtc/fdtput.c
index f2197f5..1042319 100644
--- a/scripts/dtc/fdtput.c
+++ b/scripts/dtc/fdtput.c
@@ -75,8 +75,9 @@ static int encode_value(struct display_info *disp, char **arg, int arg_count,
 	char *ptr;		/* pointer to current value position */
 	int len;		/* length of this cell/string/byte */
 	int ival;
-	int upto;	/* the number of bytes we have written to buf */
+	int upto;		/* the number of bytes we have written to buf */
 	char fmt[3];
+	void *save_ptr = NULL;	/* save pointer to realloc */
 
 	upto = 0;
 
@@ -96,12 +97,15 @@ static int encode_value(struct display_info *disp, char **arg, int arg_count,
 		/* enlarge our value buffer by a suitable margin if needed */
 		if (upto + len > value_size) {
 			value_size = (upto + len) + 500;
-			value = realloc(value, value_size);
-			if (!value) {
+			void *save_ptr = realloc(value, value_size);
+
+			if (!save_ptr) {
+				free(value);
 				fprintf(stderr, "Out of mmory: cannot alloc "
 					"%d bytes\n", value_size);
 				return -1;
 			}
+			value = save_ptr;
 		}
 
 		ptr = value + upto;
-- 
2.7.1.1.g3617aa0

Re: [PATCH V2 2/2] scripts: dtc: fix memory leak after realloc

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 902 bytes --]

On Tue, Apr 12, 2016 at 04:11:17PM -0500, Rob Herring wrote:
> On Tue, Apr 12, 2016 at 3:55 PM, Maxim Zhukov
> <mussitantesmortem-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > This commit fixed memory leak after errors realloc.
> >
> > Signed-off-by: Maxim Zhukov <mussitantesmortem-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> > ---
> >  scripts/dtc/fdtput.c | 10 +++++++---
> 
> Again, your patch is against the kernel copy of dtc. It needs to be
> against upstream dtc and then I can update the kernel copy once this
> lands.

Right.  TBH, I'm not particularly concerned about the leak anyway.
The memory will be freed when fdtput exits moments later, so it's
really not a big deal.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]