linux-kernel - Re: [PATCH v2] mtd: Add check and kfree() for kcalloc()

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <e41d9378-e5e5-478d-bead-aa50a9f79d4d@wanadoo.fr>
Date: Tue, 4 Feb 2025 07:17:55 +0100
From: Christophe JAILLET <christophe.jaillet@...adoo.fr>
To: Jiasheng Jiang <jiashengjiangcool@...il.com>
Cc: gmpy.liaowx@...il.com, kees@...nel.org, linux-kernel@...r.kernel.org,
 linux-mtd@...ts.infradead.org, miquel.raynal@...tlin.com, richard@....at,
 vigneshr@...com, stable@...r.kernel.org
Subject: Re: [PATCH v2] mtd: Add check and kfree() for kcalloc()

Le 04/02/2025 à 03:33, Jiasheng Jiang a écrit :
> Add a check for kcalloc() to ensure successful allocation.
> Moreover, add kfree() in the error-handling path to prevent memory leaks.
> 
> Fixes: 78c08247b9d3 ("mtd: Support kmsg dumper based on pstore/blk")
> Cc: <stable@...r.kernel.org> # v5.10+
> Signed-off-by: Jiasheng Jiang <jiashengjiangcool@...il.com>
> ---
> Changelog:
> 
> v1 -> v2:
> 
> 1. Remove redundant logging.
> 2. Add kfree() in the error-handling path.
> ---
>   drivers/mtd/mtdpstore.c | 19 ++++++++++++++++++-
>   1 file changed, 18 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mtd/mtdpstore.c b/drivers/mtd/mtdpstore.c
> index 7ac8ac901306..2d8e330dd215 100644
> --- a/drivers/mtd/mtdpstore.c
> +++ b/drivers/mtd/mtdpstore.c
> @@ -418,10 +418,17 @@ static void mtdpstore_notify_add(struct mtd_info *mtd)
>   
>   	longcnt = BITS_TO_LONGS(div_u64(mtd->size, info->kmsg_size));
>   	cxt->rmmap = kcalloc(longcnt, sizeof(long), GFP_KERNEL);
> +	if (!cxt->rmmap)
> +		goto end;

Nitpick: Could be a direct return.

> +
>   	cxt->usedmap = kcalloc(longcnt, sizeof(long), GFP_KERNEL);
> +	if (!cxt->usedmap)
> +		goto free_rmmap;
>   
>   	longcnt = BITS_TO_LONGS(div_u64(mtd->size, mtd->erasesize));
>   	cxt->badmap = kcalloc(longcnt, sizeof(long), GFP_KERNEL);
> +	if (!cxt->badmap)
> +		goto free_usedmap;
>   
>   	/* just support dmesg right now */
>   	cxt->dev.flags = PSTORE_FLAGS_DMESG;
> @@ -435,10 +442,20 @@ static void mtdpstore_notify_add(struct mtd_info *mtd)
>   	if (ret) {
>   		dev_err(&mtd->dev, "mtd%d register to psblk failed\n",
>   				mtd->index);
> -		return;
> +		goto free_badmap;
>   	}
>   	cxt->mtd = mtd;
>   	dev_info(&mtd->dev, "Attached to MTD device %d\n", mtd->index);
> +	goto end;

Mater of taste, but I think that having an explicit return here would be 
clearer that a goto end;

> +
> +free_badmap:
> +	kfree(cxt->badmap);
> +free_usedmap:
> +	kfree(cxt->usedmap);
> +free_rmmap:
> +	kfree(cxt->rmmap);

I think that in all these paths, you should also have
	cxt->XXXmap = NULL;
after the kfree().

otherwise when mtdpstore_notify_remove() is called, you could have a 
double free.

CJ

> +end:
> +	return;
>   }
>   
>   static int mtdpstore_flush_removed_do(struct mtdpstore_context *cxt,