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]
Date: Fri, 10 May 2024 13:19:48 +0100
From: Simon Horman <horms@...nel.org>
To: Xiaolei Wang <xiaolei.wang@...driver.com>
Cc: jhs@...atatu.com, xiyou.wangcong@...il.com, jiri@...nulli.us,
	davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
	pabeni@...hat.com, netdev@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [net PATCH] net/sched: Get stab before calling ops->change()

On Thu, May 09, 2024 at 10:40:43AM +0800, Xiaolei Wang wrote:
> ops->change() depends on stab, there is such a situation
> When no parameters are passed in for the first time, stab
> is omitted, as in configuration 1 below. At this time, a
> warning "Warning: sch_taprio: Size table not specified, frame
> length estimates may be inaccurate" will be received. When
> stab is added for the second time, parameters, like configuration
> 2 below, because the stab is still empty when ops->change()
> is running, you will also receive the above warning.
> 
> 1. tc qdisc replace dev eth1 parent root handle 100 taprio \
>   num_tc 5 map 0 1 2 3 4 queues 1@0 1@1 1@2 1@3 1@4 base-time 0 \
>   sched-entry S 1 100000 \
>   sched-entry S 2 100000 \
>   sched-entry S 4 100000 \
>   max-sdu 0 0 0 0 0 0 0 200 \
>   flags 2
> 
>   2. tc qdisc replace dev eth1 parent root overhead 24 handle 100 taprio \
>   num_tc 5 map 0 1 2 3 4 queues 1@0 1@1 1@2 1@3 1@4 base-time 0 \
>   sched-entry S 1 100000 \
>   sched-entry S 2 100000 \
>   sched-entry S 4 100000 \
>   max-sdu 0 0 0 0 0 0 0 200 \
>   flags 2
> 

Hi Xiaolei Wang,

If this is a fix, targeted at the net tree, then it should probably have
a Fixes tag here (no blank line between it and other tags).

> Signed-off-by: Xiaolei Wang <xiaolei.wang@...driver.com>
> ---
>  net/sched/sch_api.c | 20 ++++++++++----------
>  1 file changed, 10 insertions(+), 10 deletions(-)
> 
> diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
> index 60239378d43f..fec358f497d5 100644
> --- a/net/sched/sch_api.c
> +++ b/net/sched/sch_api.c
> @@ -1404,6 +1404,16 @@ static int qdisc_change(struct Qdisc *sch, struct nlattr **tca,
>  	struct qdisc_size_table *ostab, *stab = NULL;
>  	int err = 0;
>  
> +	if (tca[TCA_STAB]) {
> +		stab = qdisc_get_stab(tca[TCA_STAB], extack);
> +		if (IS_ERR(stab))
> +			return PTR_ERR(stab);
> +	}
> +
> +	ostab = rtnl_dereference(sch->stab);
> +	rcu_assign_pointer(sch->stab, stab);
> +	qdisc_put_stab(ostab);
> +
>  	if (tca[TCA_OPTIONS]) {
>  		if (!sch->ops->change) {
>  			NL_SET_ERR_MSG(extack, "Change operation not supported by specified qdisc");

I am concerned that in this case the stab will be updated even if the
change operation is rejected by the following code in the if
(tca[TCA_OPTIONS]) block, just below the above hunk.

		if (tca[TCA_INGRESS_BLOCK] || tca[TCA_EGRESS_BLOCK]) {
			NL_SET_ERR_MSG(extack, "Change of blocks is not supported");
			return -EOPNOTSUPP;
		}
..

-- 
pw-bot: under-review

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ