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: <baf121ae-a5a4-47a3-bc3a-9255708009b9@selasky.org>
Date:   Mon, 12 Sep 2022 14:36:12 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:91.0) Gecko/20100101
 Thunderbird/91.13.0
Subject: Re: [PATCH] media: dvb_ringbuffer : Fix a bug in dvb_ringbuffer.c
Content-Language: en-US
From:   Hans Petter Selasky <hps@selasky.org>
To:     =?UTF-8?B?7Jyg7Jqp7IiY?= <yongsuyoo0215@gmail.com>,
        Kieran Bingham <kieran.bingham@ideasonboard.com>,
        linux-media@vger.kernel.org, linux-kernel@vger.kernel.org,
        mchehab@kernel.org
Cc:     0215yys@hanmail.net
References: <CANXPkT6mYusYe8O0dbq3vW+24SsUZ19PqhOL+wLFRnbFXwu0Zg@mail.gmail.com>
 <CANXPkT7nOhH+5bD0ycyRBT9FKQBBszCVuWkqp4tFtVRf2+8DFg@mail.gmail.com>
 <CANXPkT5k9Pw4ka6CihyCg0oTd-32Te-ox=f3=9rtCphVgrdctA@mail.gmail.com>
 <165590120140.1149771.2257818527859865760@Monstersaurus>
 <4883f0a7-6a1b-31bd-33fe-db8f6dcf73fa@selasky.org>
 <CANXPkT73ssg6RRyfDtp7c_8sO60a-UT0-Y4S1_=D=M_mcLNN9g@mail.gmail.com>
 <CANXPkT4qYOYPL+F=-Pi_NbQErq9WwrR-M-BHe=gP9Ay4bSs+=w@mail.gmail.com>
 <CANXPkT5=ryAFvb1cO+Wb0CQYmytwedS2dqVTYqt2Km1fkK4w9Q@mail.gmail.com>
 <CANXPkT7vt8gq5UO4OXK2pTUyyB102ANJ5i9s92AW+a3rAioMog@mail.gmail.com>
 <085b9025-bc23-37d4-d430-afc432b4d783@selasky.org>
In-Reply-To: <085b9025-bc23-37d4-d430-afc432b4d783@selasky.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi Mauro and YongSu,

Answering my own question: The reason nobody has triggered this yet, is 
because the buffer size used is power of two. Because unsigned modulus 
is used, the result becomes correct. See below. But if non-power of two 
ring-buffer is used, then the result becomes incorrect. There is no 
block for non-power of two sized buffers. See:

https://github.com/search?q=dvb_set_pesfilter&type=code

cat << EOF > testX.c
#include <stdio.h>

int
main()
{
int consumed_old;
int consumed_fix;
size_t idx = 3;
ssize_t pread = 15;
ssize_t size = 256;

consumed_old = (idx - pread) % size;

consumed_fix = (idx - pread);
if (consumed_fix < 0)
consumed_fix += size;

printf("old=%d new=%d size=%zd\n", consumed_old, consumed_fix, size);

size = 254;

consumed_old = (idx - pread) % size;

consumed_fix = (idx - pread);
if (consumed_fix < 0)
consumed_fix += size;

printf("old=%d new=%d size=%zd\n", consumed_old, consumed_fix, size);

return (0);
}
EOF

cc testX.c && ./a.out
old=244 new=244 size=256
old=244 new=242 size=254

So either push the suggested fix, or block non-power of two buffer sizes!

Best regards,
--HPS