netdev - Re: [PATCH] RDMA/cxgb3: test before subtraction on unsigned

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] [day] [month] [year] [list]

Date:	Tue, 03 Mar 2009 09:41:56 -0600
From:	Steve Wise <swise@...ngridcomputing.com>
To:	Roel Kluin <roel.kluin@...il.com>
CC:	swise@...lsio.com, general@...ts.openfabrics.org,
	netdev@...r.kernel.org, Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH] RDMA/cxgb3: test before subtraction on unsigned


Roel Kluin wrote:
> I think there's someting wrong in iwch_post_send():
>
> // vi drivers/infiniband/hw/cxgb3/iwch_qp.c +353
> int iwch_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
>                       struct ib_send_wr **bad_wr)
> {
> 	...
>         u32 num_wrs;
> 	...
>         num_wrs = Q_FREECNT(qhp->wq.sq_rptr, qhp->wq.sq_wptr,
>                   qhp->wq.sq_size_log2);
>         if (num_wrs <= 0) {
>                 spin_unlock_irqrestore(&qhp->lock, flag);
>                 return -ENOMEM;
>         }
>
> // vi drivers/infiniband/hw/cxgb3/cxio_wr.h +50
> #define Q_FREECNT(rptr,wptr,size_log2) ((1UL<<size_log2)-((wptr)-(rptr)))
>
> Since num_wrs is unsigned, I think the test should occur before the subtraction,
> right? please review.
> ------------------------------>8-------------8<---------------------------------
> num_wrs is unsigned so test before the subtraction.
>
> Signed-off-by: Roel Kluin <roel.kluin@...il.com>
> ---
> diff --git a/drivers/infiniband/hw/cxgb3/iwch_qp.c b/drivers/infiniband/hw/cxgb3/iwch_qp.c
> index 19661b2..1c6ebaf 100644
> --- a/drivers/infiniband/hw/cxgb3/iwch_qp.c
> +++ b/drivers/infiniband/hw/cxgb3/iwch_qp.c
> @@ -369,12 +369,13 @@ int iwch_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
>  		spin_unlock_irqrestore(&qhp->lock, flag);
>  		return -EINVAL;
>  	}
> -	num_wrs = Q_FREECNT(qhp->wq.sq_rptr, qhp->wq.sq_wptr,
> -		  qhp->wq.sq_size_log2);
> -	if (num_wrs <= 0) {
> +	if ((1UL << qhp->wq.sq_size_log2) + qhp->wq.sq_rptr <=
> +			qhp->wq.sq_wptr) {
>  		spin_unlock_irqrestore(&qhp->lock, flag);
>  		return -ENOMEM;
>  	}
> +	num_wrs = Q_FREECNT(qhp->wq.sq_rptr, qhp->wq.sq_wptr,
> +		  qhp->wq.sq_size_log2);
>  	while (wr) {
>  		if (num_wrs == 0) {
>  			err = -ENOMEM;
>
>   

I think the test is just wrong.  It should be:

if (num_wrs == 0).

Q_FREECNT() will not return < 0.  The free count is either zero if it is 
full,  or non zero count of the number of free elements.
 
Steve.






--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html