IronPort-SDR: pLIhsEu0JnAFXLFpYWRcOwanhtYwhwu+Bl3lMmZs9zlbNjLfJzXrcpbrJ/KVbZq9nuk+BxZxJr
 a1UoKyi3LlLIee2/jLSFavaq0xSKoPYMLSseg7Oaw7mlJQI60im+3BUSanxcRHUTYWHFs831sj
 Y5SOHDkAGvME+kHxSAQFH68hdDDGNVKN3Ng8aLMT2rkkLT40DvbQe4JjU86usBOUmOufDeMsf/
 bQBQ60A561bSywVp1rxrTFZnDlvZKTrqjmrHZEXcqHt5CzRubdOsbU3VaHB2nuW95PhpGtpNbf
 GQc=
IronPort-SDR: SyAzyvghB5ka6v0bsH7tABPrTVYEL5O34yFeCoswPf/7Ups1ObzLEt9HvBVgVxdy7o2hWprrSD
 FChbBAZVZEaaqEzVFv0VmErB1R4KDvkqhFMv7cueTFsO5Snk45uKmdAvLk1QTNjIRRd47A6/6/
 lBrH4YqcrjAh2qBGhfIsF4dtSQSwQNbmmXcMC1Kcnq0Oilz2JWHuI7HNAB/UIL+D4mvDr9/ujM
 vDegAdGvPHek5w1gqe3yagW2UBZPZIYhS91diCSUza0Tz3+470b26MD0d3Ux9TunoAR8HYeaC1
 f1c=
WDCIronportException: Internal
Message-ID: <9b335f88-a20a-83da-3ce1-df409e05004a@opensource.wdc.com>
Date:   Tue, 14 Feb 2023 16:27:50 +0900
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.7.1
Subject: Re: [PATCH] scsi: ipr: work around fortify-string warning
Content-Language: en-US
To:     Arnd Bergmann <arnd@arndb.de>, Arnd Bergmann <arnd@kernel.org>,
        Brian King <brking@us.ibm.com>,
        "James E.J. Bottomley" <jejb@linux.ibm.com>,
        "Martin K. Petersen" <martin.petersen@oracle.com>
Cc:     Kees Cook <keescook@chromium.org>,
        Nathan Chancellor <nathan@kernel.org>,
        Nick Desaulniers <ndesaulniers@google.com>,
        Tom Rix <trix@redhat.com>, linux-scsi@vger.kernel.org,
        linux-kernel@vger.kernel.org, llvm@lists.linux.dev
References: <20230213101143.3821483-1-arnd@kernel.org>
 <3d94fc3f-fbe5-a56e-3713-4f0788ebc7f9@opensource.wdc.com>
 <94b362b4-a30d-4b69-8126-a64f5d025740@app.fastmail.com>
From:   Damien Le Moal <damien.lemoal@opensource.wdc.com>
Organization: Western Digital Research
In-Reply-To: <94b362b4-a30d-4b69-8126-a64f5d025740@app.fastmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2/14/23 16:14, Arnd Bergmann wrote:
> On Tue, Feb 14, 2023, at 04:59, Damien Le Moal wrote:
>> On 2/13/23 19:10, Arnd Bergmann wrote:
>>> From: Arnd Bergmann <arnd@arndb.de>
> 
>>>   **/
>>> -static int strip_and_pad_whitespace(int i, char *buf)
>>> +static void strip_whitespace(int i, char *buf)
>>>  {
>>>  	while (i && buf[i] == ' ')
>>>  		i--;
>>> -	buf[i+1] = ' ';
>>> -	buf[i+2] = '\0';
>>> -	return i + 2;
>>> +	buf[i+1] = '\0';
>>
>> If i is now the size of the buffer, this is a buffer overflow, no ? And
>> the initial loop should start at "i - 1" I think...
> 
> Right, I definitely have the wrong length here.
> 
>>>  }
>>>  
>>>  /**
>>> @@ -1547,19 +1543,21 @@ static int strip_and_pad_whitespace(int i, char *buf)
>>>  static void ipr_log_vpd_compact(char *prefix, struct ipr_hostrcb *hostrcb,
>>>  				struct ipr_vpd *vpd)
>>>  {
>>> -	char buffer[IPR_VENDOR_ID_LEN + IPR_PROD_ID_LEN + IPR_SERIAL_NUM_LEN + 3];
>>> -	int i = 0;
>>> +	char vendor_id[IPR_VENDOR_ID_LEN + 1];
>>
>> ...but the size is in fact "i + 1"... So in strip_whitespace(), i is the
>> index of the last possible character in the string, and given that the
>> string may be much shorter, that function may not actually strip
>> whitespaces after the string...
> 
> I think aside from the off-by-one bug I introduced, this is not
> a (new) problem as the old code already assumed that the input
> is padded with spaces rather than nul-terminated.

Yeah. The HW probably always give the same amount of chars with short
strings completed by spaces...

> 
>>> +	char product_id[IPR_PROD_ID_LEN + 1];
>>> +	char sn[IPR_SERIAL_NUM_LEN + 1];
>>>  
>>> -	memcpy(buffer, vpd->vpids.vendor_id, IPR_VENDOR_ID_LEN);
>>> -	i = strip_and_pad_whitespace(IPR_VENDOR_ID_LEN - 1, buffer);
>>> +	memcpy(vendor_id, vpd->vpids.vendor_id, IPR_VENDOR_ID_LEN);
>>> +	strip_whitespace(IPR_VENDOR_ID_LEN, vendor_id);
>>
>> So this call should really be:
>>
>> 	strip_whitespace(strlen(vendor_id) - 1, vendor_id);
>>
>> Which means that this helper can be turned into:
>>
>> static void strip_whitespace(char *buf)
>> {
>> 	int i = strlen(buf) - 1;
>>
>> 	while (i > 0 && buf[i] == ' ')
>> 		i--;
>> 	buf[i+1] = '\0';
>> }
>>
>> Unless I am missing something :)
> 
> The strlen() here requires the input to be a properly terminated string,
> so this would at least require adding a \0.
> 
> Also, if the input is a short nul-terminated string instead of
> a space padded array, we probably don't need to strip the trailing
> whitespace, or at least the original version didn't either.
> 
> Let me try to respin my patch with just the off-by-one error
> fixed but otherwise keeping the output of ipr_log_vpd_compact()
> unchanged.
> 
>       Arnd

-- 
Damien Le Moal
Western Digital Research