Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <80429172-37c6-c9ce-4df7-259bb90338a8@roeck-us.net>
Date:   Sat, 26 Mar 2022 22:38:05 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.5.0
Subject: Re: [PATCH] device: fix missing check on list iterator
Content-Language: en-US
To:     Xiaomeng Tong <xiam0nd.tong@gmail.com>, bskeggs@redhat.com
Cc:     kherbst@redhat.com, lyude@redhat.com, airlied@linux.ie,
        daniel@ffwll.ch, dri-devel@lists.freedesktop.org,
        nouveau@lists.freedesktop.org, linux-kernel@vger.kernel.org,
        stable@vger.kernel.org
References: <20220327053139.2572-1-xiam0nd.tong@gmail.com>
From:   Guenter Roeck <linux@roeck-us.net>
In-Reply-To: <20220327053139.2572-1-xiam0nd.tong@gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 3/26/22 22:31, Xiaomeng Tong wrote:
> The bug is here:
> 	lo = pstate->base.domain[domain->name];
> 
> The list iterator 'pstate' will point to a bogus position containing
> HEAD if the list is empty or no element is found. This case should
> be checked before any use of the iterator, otherwise it will lead
> to a invalid memory access.
> 
> To fix this bug, add an check. Use a new value 'iter' as the list
> iterator, while use the old value 'pstate' as a dedicated variable
> to point to the found element.
> 
> Cc: stable@vger.kernel.org
> Fixes: 9838366c1597d ("drm/nouveau/device: initial control object class, with pstate control methods")
> Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com>
> ---
>   drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c | 11 ++++++++---
>   1 file changed, 8 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c
> index ce774579c89d..6b768635e8ba 100644
> --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c
> +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c
> @@ -72,7 +72,7 @@ nvkm_control_mthd_pstate_attr(struct nvkm_control *ctrl, void *data, u32 size)
>   	} *args = data;
>   	struct nvkm_clk *clk = ctrl->device->clk;
>   	const struct nvkm_domain *domain;
> -	struct nvkm_pstate *pstate;
> +	struct nvkm_pstate *pstate = NULL, *iter;
>   	struct nvkm_cstate *cstate;
>   	int i = 0, j = -1;
>   	u32 lo, hi;
> @@ -103,11 +103,16 @@ nvkm_control_mthd_pstate_attr(struct nvkm_control *ctrl, void *data, u32 size)
>   		return -EINVAL;
>   
>   	if (args->v0.state != NVIF_CONTROL_PSTATE_ATTR_V0_STATE_CURRENT) {
> -		list_for_each_entry(pstate, &clk->states, head) {
> -			if (i++ == args->v0.state)
> +		list_for_each_entry(iter, &clk->states, head) {
> +			if (i++ == args->v0.state) {
> +				pstate = iter;

Is iter and the assignment really necessary ? Unless I am missing something,
list_for_each_entry() always assigns pos (pstate/iter), even if the list is
empty. If nothing is found, pstate would be NULL at the end, so

>   				break;
> +			}
>   		}
>   
> +		if (!pstate)
> +			return -EINVAL;
> +
... just this check should do to cover both the "not found" and "list empty"
cases.

Thanks,
Guenter

>   		lo = pstate->base.domain[domain->name];
>   		hi = lo;
>   		list_for_each_entry(cstate, &pstate->list, head) {