linux-kernel - Re: [PATCH v2 3/6] soc: hisilicon: kunpeng_hccs: Add the check for base address and size of shared memory

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: <20240823093858.00007047@Huawei.com>
Date: Fri, 23 Aug 2024 09:38:58 +0100
From: Jonathan Cameron <Jonathan.Cameron@...wei.com>
To: Huisong Li <lihuisong@...wei.com>
CC: <xuwei5@...ilicon.com>, <linux-kernel@...r.kernel.org>, <soc@...nel.org>,
	<linux-arm-kernel@...ts.infradead.org>, <krzk@...nel.org>,
	<wanghuiqiang@...wei.com>, <liuyonglong@...wei.com>
Subject: Re: [PATCH v2 3/6] soc: hisilicon: kunpeng_hccs: Add the check for
 base address and size of shared memory

On Fri, 23 Aug 2024 11:10:56 +0800
Huisong Li <lihuisong@...wei.com> wrote:

> If the shmem_base_addr from PCCT is zero, hccs_register_pcc_channel will
> return success. And then driver will access to illegal address when send
> PCC command. In addition, the size of shared memory used for communication
> between driver and platform is fixed, namely 64 Bytes which is
> unchangeable. So add the verification for them.
> 
As with previous, make it clear if this hardening or fix
fix to catch a problem on shipping hardware (I assume not, but you never
know!)

A comment on existing code inline.  Not a suggestion for a change
in this series, but maybe for the future.  There are a lot
of goto err_mbx_channel_Free in here already and this patch adds
another.  The cleanup there is trivial so DEFINE_FREE() and __free
usage will make this code quite a bit nicer to read.

> Signed-off-by: Huisong Li <lihuisong@...wei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@...wei.com>

> ---
>  drivers/soc/hisilicon/kunpeng_hccs.c | 24 +++++++++++++++---------
>  1 file changed, 15 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/soc/hisilicon/kunpeng_hccs.c b/drivers/soc/hisilicon/kunpeng_hccs.c
> index 6e88f597f267..6055e5091cbd 100644
> --- a/drivers/soc/hisilicon/kunpeng_hccs.c
> +++ b/drivers/soc/hisilicon/kunpeng_hccs.c
> @@ -170,15 +170,21 @@ static int hccs_register_pcc_channel(struct hccs_dev *hdev)
>  		goto err_mbx_channel_free;
>  	}
>  
> -	if (pcc_chan->shmem_base_addr) {
> -		cl_info->pcc_comm_addr = ioremap(pcc_chan->shmem_base_addr,
> -						 pcc_chan->shmem_size);
> -		if (!cl_info->pcc_comm_addr) {
> -			dev_err(dev, "Failed to ioremap PCC communication region for channel-%u.\n",
> -				hdev->chan_id);
> -			rc = -ENOMEM;
> -			goto err_mbx_channel_free;
> -		}
> +	if (!pcc_chan->shmem_base_addr ||
> +	    pcc_chan->shmem_size != HCCS_PCC_SHARE_MEM_BYTES) {
> +		dev_err(dev, "The base address or size (%llu) of PCC communication region is invalid.\n",
> +			pcc_chan->shmem_size);
> +		rc = -EINVAL;
> +		goto err_mbx_channel_free;
Worth considering for the future: Maybe a DEFINE_FREE for pcc_mbox_free_channel) makes sense,
though if you do you should only assign cl_info->pcc_chan after all possible error paths.

> +	}
> +
> +	cl_info->pcc_comm_addr = ioremap(pcc_chan->shmem_base_addr,
> +					 pcc_chan->shmem_size);
> +	if (!cl_info->pcc_comm_addr) {
> +		dev_err(dev, "Failed to ioremap PCC communication region for channel-%u.\n",
> +			hdev->chan_id);
> +		rc = -ENOMEM;
> +		goto err_mbx_channel_free;
>  	}
>  
>  	return 0;