linux-kernel - Re: [PATCH] drivers: ipmi: Support for both IPMB Req and Resp

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-next>] [day] [month] [year] [list]

Message-ID: <20191106005332.GA2754@minyard.net>
Date:   Tue, 5 Nov 2019 18:53:32 -0600
From:   Corey Minyard <minyard@....org>
To:     Vijay Khemka <vijaykhemka@...com>
Cc:     Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        openipmi-developer@...ts.sourceforge.net,
        linux-kernel@...r.kernel.org, cminyard@...sta.com,
        asmaa@...lanox.com, joel@....id.au, linux-aspeed@...ts.ozlabs.org,
        sdasari@...com
Subject: Re: [PATCH] drivers: ipmi: Support for both IPMB Req and Resp

On Tue, Nov 05, 2019 at 11:47:31AM -0800, Vijay Khemka wrote:
> Removed check for request or response in IPMB packets coming from
> device as well as from host. Now it supports both way communication
> to device via IPMB. Both request and response will be passed to
> application.
> 
> Signed-off-by: Vijay Khemka <vijaykhemka@...com>
> ---
>  drivers/char/ipmi/ipmb_dev_int.c | 29 +----------------------------
>  1 file changed, 1 insertion(+), 28 deletions(-)
> 
> diff --git a/drivers/char/ipmi/ipmb_dev_int.c b/drivers/char/ipmi/ipmb_dev_int.c
> index 285e0b8f9a97..7201fdb533d8 100644
> --- a/drivers/char/ipmi/ipmb_dev_int.c
> +++ b/drivers/char/ipmi/ipmb_dev_int.c
> @@ -133,9 +133,6 @@ static ssize_t ipmb_write(struct file *file, const char __user *buf,
>  	rq_sa = GET_7BIT_ADDR(msg[RQ_SA_8BIT_IDX]);
>  	netf_rq_lun = msg[NETFN_LUN_IDX];
>  
> -	if (!(netf_rq_lun & NETFN_RSP_BIT_MASK))
> -		return -EINVAL;
> -
>  	/*
>  	 * subtract rq_sa and netf_rq_lun from the length of the msg passed to
>  	 * i2c_smbus_xfer
> @@ -203,28 +200,6 @@ static u8 ipmb_verify_checksum1(struct ipmb_dev *ipmb_dev, u8 rs_sa)
>  		ipmb_dev->request.checksum1);
>  }
>  
> -static bool is_ipmb_request(struct ipmb_dev *ipmb_dev, u8 rs_sa)
> -{
> -	if (ipmb_dev->msg_idx >= IPMB_REQUEST_LEN_MIN) {
> -		if (ipmb_verify_checksum1(ipmb_dev, rs_sa))
> -			return false;

You still need to check the message length and checksum, you just need
to ignore the req/resp bit.

-corey

> -
> -		/*
> -		 * Check whether this is an IPMB request or
> -		 * response.
> -		 * The 6 MSB of netfn_rs_lun are dedicated to the netfn
> -		 * while the remaining bits are dedicated to the lun.
> -		 * If the LSB of the netfn is cleared, it is associated
> -		 * with an IPMB request.
> -		 * If the LSB of the netfn is set, it is associated with
> -		 * an IPMB response.
> -		 */
> -		if (!(ipmb_dev->request.netfn_rs_lun & NETFN_RSP_BIT_MASK))
> -			return true;
> -	}
> -	return false;
> -}
> -
>  /*
>   * The IPMB protocol only supports I2C Writes so there is no need
>   * to support I2C_SLAVE_READ* events.
> @@ -273,9 +248,7 @@ static int ipmb_slave_cb(struct i2c_client *client,
>  
>  	case I2C_SLAVE_STOP:
>  		ipmb_dev->request.len = ipmb_dev->msg_idx;
> -
> -		if (is_ipmb_request(ipmb_dev, GET_8BIT_ADDR(client->addr)))
> -			ipmb_handle_request(ipmb_dev);
> +		ipmb_handle_request(ipmb_dev);
>  		break;
>  
>  	default:
> -- 
> 2.17.1
>