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: <cac2eaf0-2a56-d750-3508-7b5ed03e1eb5@rasmusvillemoes.dk>
Date:   Tue, 30 Aug 2022 09:32:19 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.11.0
Subject: Re: [PATCH 1/2] lib/test_printf.c: Add ip6 tests
Content-Language: en-US
To:     Sergey Senozhatsky <senozhatsky@chromium.org>,
        Kent Overstreet <kent.overstreet@linux.dev>
Cc:     linux-kernel@vger.kernel.org, Petr Mladek <pmladek@suse.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Rasmus Villemoes <linux@rasmusvillemoes.dk>
References: <20220830003119.1793219-1-kent.overstreet@linux.dev>
 <20220830003119.1793219-2-kent.overstreet@linux.dev>
 <Yw1sKTOEs1Nrdb2S@google.com>
From:   Rasmus Villemoes <linux@rasmusvillemoes.dk>
In-Reply-To: <Yw1sKTOEs1Nrdb2S@google.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 30/08/2022 03.47, Sergey Senozhatsky wrote:
> On (22/08/29 20:31), Kent Overstreet wrote:
>> diff --git a/lib/test_printf.c b/lib/test_printf.c
>> index 4bd15a593f..6a56dbf076 100644
>> --- a/lib/test_printf.c
>> +++ b/lib/test_printf.c
>> @@ -18,6 +18,7 @@
>>  #include <linux/dcache.h>
>>  #include <linux/socket.h>
>>  #include <linux/in.h>
>> +#include <linux/in6.h>
>>  
>>  #include <linux/gfp.h>
>>  #include <linux/mm.h>
>> @@ -61,6 +62,9 @@ do_test(int bufsize, const char *expect, int elen,
>>  		pr_warn("vsnprintf(buf, %d, \"%s\", ...) returned %d, expected %d\n",
>>  			bufsize, fmt, ret, elen);
>>  		return 1;
>> +		pr_warn("vsnprintf(buf, %d, \"%s\", ...) returned %d, expected %d (%s != %s)\n",
>> +			bufsize, fmt, ret, elen, test_buffer, expect);
>> +		return 1;
>>  	}
> 
> I assume you intended to replace first pr_warn() with the second one?

Probably, but that's not ok. The test framework does not trust
vsnprintf(), especially not when it does not behave as expected. So I
very much carefully do not treat the buffer as a nul-terminated string
until I have verified that it does have a nul character (that's tested a
few lines below), and then when I compare the buffer contents can I pass
it as a %s argument. Also note how that test takes the 'we may be
testing a truncated write' into consideration, by printing the expect
string via %.*s.

tl;dr, please just remove that hunk.

Rasmus