Netdev List
 help / color / mirror / Atom feed
From: Simon Horman <horms@kernel.org>
To: Pengpeng Hou <pengpeng@iscas.ac.cn>
Cc: Cai Huoqing <cai.huoqing@linux.dev>,
	Andrew Lunn <andrew+netdev@lunn.ch>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] net: hinic: validate firmware image section bounds
Date: Sat, 11 Jul 2026 17:21:00 +0100	[thread overview]
Message-ID: <20260711162100.GB1364329@horms.kernel.org> (raw)
In-Reply-To: <20260706093455.81168-1-pengpeng@iscas.ac.cn>

On Mon, Jul 06, 2026 at 05:34:55PM +0800, Pengpeng Hou wrote:
> The firmware update path copies each section from the image with an offset
> and length supplied by the firmware header. The image validator only
> checked the sum of section lengths and used fw_len + header_size to check
> the file size, but it did not prove that each section offset and length
> fits in the firmware payload.
> 
> Reject images with a truncated header, avoid the fw_len plus header-size
> addition, validate each section type before it is used as a bit index, and
> prove every section range with offset <= fw_len and len <= fw_len - offset.
> 
> Signed-off-by: Pengpeng Hou <pengpeng@iscas.ac.cn>
> ---
>  .../net/ethernet/huawei/hinic/hinic_devlink.c | 57 ++++++++++++++++---
>  1 file changed, 49 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/net/ethernet/huawei/hinic/hinic_devlink.c b/drivers/net/ethernet/huawei/hinic/hinic_devlink.c
> index c977c43e5d5a..0c06fbdadad3 100644
> --- a/drivers/net/ethernet/huawei/hinic/hinic_devlink.c
> +++ b/drivers/net/ethernet/huawei/hinic/hinic_devlink.c
> @@ -20,14 +20,25 @@
>  #include "hinic_devlink.h"
>  #include "hinic_hw_dev.h"
>  
> +static bool hinic_fw_section_valid(u32 fw_len, u32 offset, u32 len)
> +{
> +	return offset <= fw_len && len <= fw_len - offset;
> +}
> +
>  static bool check_image_valid(struct hinic_devlink_priv *priv, const u8 *buf,
> -			      u32 image_size, struct host_image_st *host_image)
> +			      size_t image_size, struct host_image_st *host_image)
>  {
> -	struct fw_image_st *fw_image = NULL;
> +	const struct fw_image_st *fw_image;

I don't think the const update is strictly related to this patch,
and thus doesn't belong in this patch.

>  	u32 len = 0;
>  	u32 i;
>  
> -	fw_image = (struct fw_image_st *)buf;
> +	if (image_size < UPDATEFW_IMAGE_HEAD_SIZE) {
> +		dev_err(&priv->hwdev->hwif->pdev->dev,
> +			"Wrong image size read from file\n");
> +		return false;
> +	}
> +
> +	fw_image = (const struct fw_image_st *)buf;
>  
>  	if (fw_image->fw_magic != HINIC_MAGIC_NUM) {
>  		dev_err(&priv->hwdev->hwif->pdev->dev, "Wrong fw_magic read from file, fw_magic: 0x%x\n",
> @@ -41,14 +52,44 @@ static bool check_image_valid(struct hinic_devlink_priv *priv, const u8 *buf,
>  		return false;
>  	}
>  
> +	if (fw_image->fw_len != image_size - UPDATEFW_IMAGE_HEAD_SIZE) {
> +		dev_err(&priv->hwdev->hwif->pdev->dev,
> +			"Wrong data size read from file\n");
> +		return false;
> +	}
> +
>  	for (i = 0; i < fw_image->fw_info.fw_section_cnt; i++) {
> -		len += fw_image->fw_section_info[i].fw_section_len;
> -		host_image->image_section_info[i] = fw_image->fw_section_info[i];
> +		const struct fw_section_info_st *section =
> +			&fw_image->fw_section_info[i];
> +
> +		if (section->fw_section_type >= FILE_TYPE_TOTAL_NUM) {
> +			dev_err(&priv->hwdev->hwif->pdev->dev,
> +				"Wrong section type read from file: %u\n",
> +				section->fw_section_type);
> +			return false;
> +		}
> +
> +		if (!hinic_fw_section_valid(fw_image->fw_len,
> +					    section->fw_section_offset,
> +					    section->fw_section_len)) {
> +			dev_err(&priv->hwdev->hwif->pdev->dev,
> +				"Wrong section size read from file\n");
> +			return false;
> +		}
> +
> +		if (section->fw_section_len > fw_image->fw_len - len) {
> +			dev_err(&priv->hwdev->hwif->pdev->dev,
> +				"Wrong data size read from file\n");
> +			return false;
> +		}
> +
> +		len += section->fw_section_len;
> +		host_image->image_section_info[i] = *section;
>  	}
>  
> -	if (len != fw_image->fw_len ||
> -	    (fw_image->fw_len + UPDATEFW_IMAGE_HEAD_SIZE) != image_size) {
> -		dev_err(&priv->hwdev->hwif->pdev->dev, "Wrong data size read from file\n");
> +	if (len != fw_image->fw_len) {
> +		dev_err(&priv->hwdev->hwif->pdev->dev,
> +			"Wrong data size read from file\n");

Likewise, the update to the dev_err() call, as distinct from the update
to the condition that precedes it, appears to be a whitespace cleanup.
If so, it doesn't belong in this patch.

>  		return false;
>  	}
>  

-- 
pw-bot: changes-requested

      reply	other threads:[~2026-07-11 16:21 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-06  9:34 [PATCH] net: hinic: validate firmware image section bounds Pengpeng Hou
2026-07-11 16:21 ` Simon Horman [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260711162100.GB1364329@horms.kernel.org \
    --to=horms@kernel.org \
    --cc=andrew+netdev@lunn.ch \
    --cc=cai.huoqing@linux.dev \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=pengpeng@iscas.ac.cn \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox