Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1633219 > unrolled thread

[PATCH] staging: comedi: use memdup_user

Started byGeliang Tang <geliangtang@gmail.com>
First post2017-04-29 03:50 +0200
Last post2017-04-29 07:20 +0200
Articles 3 — 3 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  [PATCH] staging: comedi: use memdup_user Geliang Tang <geliangtang@gmail.com> - 2017-04-29 03:50 +0200
    Re: [PATCH] staging: comedi: use memdup_user Joe Perches <joe@perches.com> - 2017-04-29 04:00 +0200
      Re: [PATCH] staging: comedi: use memdup_user Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-04-29 07:20 +0200

#1633219 — [PATCH] staging: comedi: use memdup_user

FromGeliang Tang <geliangtang@gmail.com>
Date2017-04-29 03:50 +0200
Subject[PATCH] staging: comedi: use memdup_user
Message-ID<tBpGh-7H8-3@gated-at.bofh.it>
Use memdup_user() helper instead of open-coding to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
---
 drivers/staging/comedi/comedi_fops.c | 22 +++++++---------------
 1 file changed, 7 insertions(+), 15 deletions(-)

diff --git a/drivers/staging/comedi/comedi_fops.c b/drivers/staging/comedi/comedi_fops.c
index f191c2a..4797c8f 100644
--- a/drivers/staging/comedi/comedi_fops.c
+++ b/drivers/staging/comedi/comedi_fops.c
@@ -1450,22 +1450,14 @@ static int do_insnlist_ioctl(struct comedi_device *dev,
 		return -EFAULT;
 
 	data = kmalloc_array(MAX_SAMPLES, sizeof(unsigned int), GFP_KERNEL);
-	if (!data) {
-		ret = -ENOMEM;
-		goto error;
-	}
-
-	insns = kcalloc(insnlist.n_insns, sizeof(*insns), GFP_KERNEL);
-	if (!insns) {
-		ret = -ENOMEM;
-		goto error;
-	}
+	if (!data)
+		return -ENOMEM;
 
-	if (copy_from_user(insns, insnlist.insns,
-			   sizeof(*insns) * insnlist.n_insns)) {
-		dev_dbg(dev->class_dev, "copy_from_user failed\n");
-		ret = -EFAULT;
-		goto error;
+	insns = memdup_user(insnlist.insns, sizeof(*insns) * insnlist.n_insns);
+	if (IS_ERR(insns)) {
+		dev_dbg(dev->class_dev, "memdup_user failed\n");
+		kfree(data);
+		return PTR_ERR(insns);
 	}
 
 	for (i = 0; i < insnlist.n_insns; i++) {
-- 
2.9.3

[toc] | [next] | [standalone]


#1633229

FromJoe Perches <joe@perches.com>
Date2017-04-29 04:00 +0200
Message-ID<tBpPX-7Lk-3@gated-at.bofh.it>
In reply to#1633219
On Sat, 2017-04-29 at 09:45 +0800, Geliang Tang wrote:
> Use memdup_user() helper instead of open-coding to simplify the code.

While I doubt this is a problem, this loses
the multiplication overflow check for
sizeof(*insns) * insnlist.n_isns

> diff --git a/drivers/staging/comedi/comedi_fops.c b/drivers/staging/comedi/comedi_fops.c
[]
> @@ -1450,22 +1450,14 @@ static int do_insnlist_ioctl(struct comedi_device *dev,
>  		return -EFAULT;
>  
>  	data = kmalloc_array(MAX_SAMPLES, sizeof(unsigned int), GFP_KERNEL);
> -	if (!data) {
> -		ret = -ENOMEM;
> -		goto error;
> -	}
> -
> -	insns = kcalloc(insnlist.n_insns, sizeof(*insns), GFP_KERNEL);
> -	if (!insns) {
> -		ret = -ENOMEM;
> -		goto error;
> -	}
> +	if (!data)
> +		return -ENOMEM;
>  
> -	if (copy_from_user(insns, insnlist.insns,
> -			   sizeof(*insns) * insnlist.n_insns)) {
> -		dev_dbg(dev->class_dev, "copy_from_user failed\n");
> -		ret = -EFAULT;
> -		goto error;
> +	insns = memdup_user(insnlist.insns, sizeof(*insns) * insnlist.n_insns);
> +	if (IS_ERR(insns)) {
> +		dev_dbg(dev->class_dev, "memdup_user failed\n");
> +		kfree(data);
> +		return PTR_ERR(insns);
>  	}
>  
>  	for (i = 0; i < insnlist.n_insns; i++) {

[toc] | [prev] | [next] | [standalone]


#1633241

FromGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Date2017-04-29 07:20 +0200
Message-ID<tBsXv-1PB-1@gated-at.bofh.it>
In reply to#1633229
On Fri, Apr 28, 2017 at 06:53:32PM -0700, Joe Perches wrote:
> On Sat, 2017-04-29 at 09:45 +0800, Geliang Tang wrote:
> > Use memdup_user() helper instead of open-coding to simplify the code.
> 
> While I doubt this is a problem, this loses
> the multiplication overflow check for
> sizeof(*insns) * insnlist.n_isns

Yes it is a problem, we need that check.

thanks,

greg k-h

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web