[Qemu-devel] [PATCH 2 of 5] [UPDATE] implementing qemu_reallocz

[Qemu-devel] [PATCH 2 of 5] [UPDATE] implementing qemu_reallocz

[Stefano Stabellini]

Adding a qemu_reallocz implementation.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

---

diff -r aceb9c1e17b1 qemu-common.h
--- a/qemu-common.h	Fri Nov 21 12:41:26 2008 +0000
+++ b/qemu-common.h	Fri Nov 21 14:36:45 2008 +0000
@@ -96,6 +96,7 @@
 
 void *qemu_malloc(size_t size);
 void *qemu_realloc(void *ptr, size_t size);
+void *qemu_reallocz(void *ptr, size_t size);
 void *qemu_mallocz(size_t size);
 void qemu_free(void *ptr);
 char *qemu_strdup(const char *str);
diff -r aceb9c1e17b1 qemu-malloc.c
--- a/qemu-malloc.c	Fri Nov 21 12:41:26 2008 +0000
+++ b/qemu-malloc.c	Fri Nov 21 14:36:45 2008 +0000
@@ -53,6 +53,16 @@
     return ptr;
 }
 
+void *qemu_reallocz(void *ptr, size_t size)
+{
+    void *res;
+    res = realloc(ptr, size);
+    if (!res)
+        return NULL;
+    memset(res, 0, size);
+    return res;
+}
+
 char *qemu_strdup(const char *str)
 {
     char *ptr;

Re: [Qemu-devel] [PATCH 2 of 5] [UPDATE] implementing qemu_reallocz

[Anthony Liguori]

Stefano Stabellini wrote:
> Adding a qemu_reallocz implementation.
>
> +void *qemu_reallocz(void *ptr, size_t size)
> +{
> +    void *res;
> +    res = realloc(ptr, size);
> +    if (!res)
> +        return NULL;
> +    memset(res, 0, size);
> +    return res;
> +}
> +
>   

I think this interface is pretty non-intuitive.  My first expectation is 
that only the grown region would be zero'd, not the whole thing.

In general, if you don't care about the old memory, free()/malloc() 
would be a better combination.  You're just doing extra work by using 
realloc() here.

Regards,

Anthony Liguori

>  char *qemu_strdup(const char *str)
>  {
>      char *ptr;
>
>
>