Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20200802120958.19ba400a@inria.fr>
Date: Sun, 2 Aug 2020 12:09:58 +0200
From: Jens Gustedt <jens.gustedt@...ia.fr>
To: Ariadne Conill <ariadne@...eferenced.org>
Cc: musl@...ts.openwall.com
Subject: Re: [PATCH v3] implement recallocarray(3)

Hello,

on Sat,  1 Aug 2020 15:42:16 -0600 you (Ariadne Conill
<ariadne@...eferenced.org>) wrote:

> This OpenBSD extension is similar to reallocarray(3), but
> zero-initializes the new memory area.
> 
> This extension is placed in _BSD_SOURCE, like
> reallocarray(3).
> 
> Changes from v2:
> - drop overflow checking for old size
> 
> Changes from v1:
> - use realloc() instead of reallocarray()
> ---
>  include/stdlib.h           |  1 +
>  src/malloc/recallocarray.c | 27 +++++++++++++++++++++++++++
>  2 files changed, 28 insertions(+)
>  create mode 100644 src/malloc/recallocarray.c
> 
> diff --git a/include/stdlib.h b/include/stdlib.h
> index b54a051f..a0412ad4 100644
> --- a/include/stdlib.h
> +++ b/include/stdlib.h
> @@ -146,6 +146,7 @@ int clearenv(void);
>  #define WCOREDUMP(s) ((s) & 0x80)
>  #define WIFCONTINUED(s) ((s) == 0xffff)
>  void *reallocarray (void *, size_t, size_t);
> +void *recallocarray (void *, size_t, size_t, size_t);
>  #endif
>  
>  #ifdef _GNU_SOURCE
> diff --git a/src/malloc/recallocarray.c b/src/malloc/recallocarray.c
> new file mode 100644
> index 00000000..a7827604
> --- /dev/null
> +++ b/src/malloc/recallocarray.c
> @@ -0,0 +1,27 @@
> +#define _BSD_SOURCE
> +#include <errno.h>
> +#include <stdlib.h>
> +#include <string.h>
> +
> +void *recallocarray(void *ptr, size_t om, size_t m, size_t n)
> +{
> +	void *newptr;
> +	size_t old_size = om * n, new_size;
> +
> +	if (n && m > -1 / n) {
> +		errno = ENOMEM;
> +		return 0;
> +	}
> +	new_size = m * n;
> +
> +	if (new_size <= old_size) {
> +		memset((char *) ptr + new_size, 0, old_size -
> new_size);
> +	}
> +
> +	newptr = realloc(ptr, m * n);

I think, this would better be

	newptr = realloc(ptr, new_size);

> +	if (new_size > old_size) {
> +		memset((char *) ptr + old_size, 0, new_size - old_size);
> +     }

Generally, if `realloc` succeeds, access to the object behind `ptr` is
invalid, even if `ptr == newptr`.

Also `newptr` may be null if `realloc` fails, so this should read

	if (newptr && new_size > old_size) {
		memset((char *)newptr + old_size, 0, new_size - old_size);
        }


Thanks
Jens

-- 
:: INRIA Nancy Grand Est ::: Camus ::::::: ICube/ICPS :::
:: ::::::::::::::: office Strasbourg : +33 368854536   ::
:: :::::::::::::::::::::: gsm France : +33 651400183   ::
:: ::::::::::::::: gsm international : +49 15737185122 ::
:: http://icube-icps.unistra.fr/index.php/Jens_Gustedt ::

Content of type "application/pgp-signature" skipped

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.