Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH V6] i2c: tegra: remove BUG, BUG_ON
To:     Bitan Biswas <bbiswas@nvidia.com>,
        Laxman Dewangan <ldewangan@nvidia.com>,
        Thierry Reding <treding@nvidia.com>,
        Jonathan Hunter <jonathanh@nvidia.com>,
        linux-i2c@vger.kernel.org, linux-tegra@vger.kernel.org,
        linux-kernel@vger.kernel.org, Peter Rosin <peda@axentia.se>,
        Wolfram Sang <wsa@the-dreams.de>
Cc:     Shardar Mohammed <smohammed@nvidia.com>,
        Sowjanya Komatineni <skomatineni@nvidia.com>,
        Mantravadi Karthik <mkarthik@nvidia.com>
References: <1560527438-30150-1-git-send-email-bbiswas@nvidia.com>
 <504e19d0-05dd-dc80-3aaf-cdab8f88002d@gmail.com>
 <fee98b4f-630e-1685-05e3-e8dae0f412bd@nvidia.com>
From:   Dmitry Osipenko <digetx@gmail.com>
Message-ID: <cadd6687-7b1b-ccce-b34b-a175adafa675@gmail.com>
Date:   Mon, 17 Jun 2019 22:32:14 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.7.1
MIME-Version: 1.0
In-Reply-To: <fee98b4f-630e-1685-05e3-e8dae0f412bd@nvidia.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

17.06.2019 22:08, Bitan Biswas пишет:
> 
> 
> On 6/14/19 10:51 AM, Dmitry Osipenko wrote:
>> 14.06.2019 18:50, Bitan Biswas пишет:
>>> Remove redundant BUG_ON calls or replace with WARN_ON_ONCE
>>> as needed. Remove BUG() and mask Rx interrupt similar as Tx
>>> for message fully sent case. Add WARN_ON_ONCE check
>>> for non-zero rx_fifo_avail in tegra_i2c_empty_rx_fifo()
>>> after all processing. Error handling in tegra_i2c_empty_rx_fifo
>>> caller is also added.
>>>
>>> Signed-off-by: Bitan Biswas <bbiswas@nvidia.com>
>>> ---
>>>   drivers/i2c/busses/i2c-tegra.c | 46 ++++++++++++++++++++++++++++++++++--------
>>>   1 file changed, 38 insertions(+), 8 deletions(-)
>>>
>>> diff --git a/drivers/i2c/busses/i2c-tegra.c b/drivers/i2c/busses/i2c-tegra.c
>>> index 4dfb4c1..26a7c8c 100644
>>> --- a/drivers/i2c/busses/i2c-tegra.c
>>> +++ b/drivers/i2c/busses/i2c-tegra.c
>>> @@ -73,6 +73,7 @@
>>>   #define I2C_ERR_NO_ACK                BIT(0)
>>>   #define I2C_ERR_ARBITRATION_LOST        BIT(1)
>>>   #define I2C_ERR_UNKNOWN_INTERRUPT        BIT(2)
>>> +#define I2C_ERR_UNEXPECTED_STATUS        BIT(3)
>>
>> What about I2C_ERR_RX_BUFFER_OVERFLOW?
> OK.
> 
>>
>>>   #define PACKET_HEADER0_HEADER_SIZE_SHIFT    28
>>>   #define PACKET_HEADER0_PACKET_ID_SHIFT        16
>>> @@ -515,15 +516,23 @@ static int tegra_i2c_empty_rx_fifo(struct tegra_i2c_dev
>>> *i2c_dev)
>>>        * prevent overwriting past the end of buf
>>>        */
>>>       if (rx_fifo_avail > 0 && buf_remaining > 0) {
>>> -        BUG_ON(buf_remaining > 3);
>>> +        /* buf_remaining > 3 check not needed as rx_fifo_avail == 0
>>> +         * when (words_to_transfer was > rx_fifo_avail) earlier
>>> +         * in this function
>>> +         */
>>
>> Please start all multiline comments with an empty "/*", it should be the correct
>> style. There are some places in the kernel where style like yours is used, but I
>> assume they are not very correct. Besides, yours variant is not consistent with the
>> style of the rest of comments in this source file. And put a dot in the end for
>> completeness. Same for the other comments in this patch.
>>
> OK
> 
>>>           val = i2c_readl(i2c_dev, I2C_RX_FIFO);
>>>           val = cpu_to_le32(val);
>>>           memcpy(buf, &val, buf_remaining);
>>>           buf_remaining = 0;
>>>           rx_fifo_avail--;
>>>       }
>>
>> Please add a newline here. All logical parts of the code should be separated to ease
>> reading and following.
> OK
> 
>>
>>> +    if (WARN_ON_ONCE(rx_fifo_avail))
>>> +        return -EINVAL;
>>>   -    BUG_ON(rx_fifo_avail > 0 && buf_remaining > 0);
>>> +    /* buf_remaining > 0 at this point can only have rx_fifo_avail == 0
>>> +     * as this corresponds to (words_to_transfer was > rx_fifo_avail)
>>> +     * case earlier in this function
>>> +     */
>>>       i2c_dev->msg_buf_remaining = buf_remaining;
>>>       i2c_dev->msg_buf = buf;
>>>   @@ -581,7 +590,10 @@ static int tegra_i2c_fill_tx_fifo(struct tegra_i2c_dev
>>> *i2c_dev)
>>>        * boundary and fault.
>>>        */
>>>       if (tx_fifo_avail > 0 && buf_remaining > 0) {
>>> -        BUG_ON(buf_remaining > 3);
>>> +        /* buf_remaining > 3 check not needed as tx_fifo_avail == 0
>>> +         * when (words_to_transfer was > tx_fifo_avail) earlier
>>> +         * in this function for non-zero words_to_transfer
>>> +         */
>>>           memcpy(&val, buf, buf_remaining);
>>>           val = le32_to_cpu(val);
>>>   @@ -811,6 +823,7 @@ static irqreturn_t tegra_i2c_isr(int irq, void *dev_id)
>>>       u32 status;
>>>       const u32 status_err = I2C_INT_NO_ACK | I2C_INT_ARBITRATION_LOST;
>>>       struct tegra_i2c_dev *i2c_dev = dev_id;
>>> +    int err_val;
>>>         status = i2c_readl(i2c_dev, I2C_INT_STATUS);
>>>   @@ -847,10 +860,21 @@ static irqreturn_t tegra_i2c_isr(int irq, void *dev_id)
>>>         if (!i2c_dev->is_curr_dma_xfer) {
>>>           if (i2c_dev->msg_read && (status & I2C_INT_RX_FIFO_DATA_REQ)) {
>>> -            if (i2c_dev->msg_buf_remaining)
>>> -                tegra_i2c_empty_rx_fifo(i2c_dev);
>>> -            else
>>> -                BUG();
>>> +            err_val = tegra_i2c_empty_rx_fifo(i2c_dev);
>>> +            if ((!(i2c_dev->msg_buf_remaining)) &&
>>
>> Let's move this check into tegra_i2c_empty_rx_fifo() and return -EINVAL for that case.
>> This will make code to look cleaner.
> OK.
> 
> 
>>
>>> +                (!(status & I2C_INT_PACKET_XFER_COMPLETE)) &&
>>
>> It shouldn't matter that XFER_COMPLETE is set if RX FIFO isn't fully emptied because
>> it always shall be emptied. Hence this check is not needed and we should error out
>> regardless.
> OK
> 
>>
>>> +                err_val) {
>>> +                /*
>>> +                 * Overflow error condition: message fully sent,
>>> +                 * with no XFER_COMPLETE interrupt but hardware
>>> +                 * asks to transfer more.
>>> +                 */
>>> +                tegra_i2c_mask_irq(i2c_dev,
>>> +                           I2C_INT_RX_FIFO_DATA_REQ);
>>
>> No need to mask RX_FIFO_DATA_REQ here because all interrupts are masked on "goto
>> err:", hence just remove the tegra_i2c_mask_irq().
>>
>>> +                i2c_dev->msg_err |=
>>> +                    I2C_ERR_UNEXPECTED_STATUS;
>>
>> No need to split this into two lines because it's less than 80 chars, write this in a
>> single line.
> OK
> 
>>
>>> +                goto err;
>>> +            }
>>>           }
>>>             if (!i2c_dev->msg_read && (status & I2C_INT_TX_FIFO_DATA_REQ)) {
>>> @@ -876,7 +900,13 @@ static irqreturn_t tegra_i2c_isr(int irq, void *dev_id)
>>>       if (status & I2C_INT_PACKET_XFER_COMPLETE) {
>>>           if (i2c_dev->is_curr_dma_xfer)
>>>               i2c_dev->msg_buf_remaining = 0;
>>> -        BUG_ON(i2c_dev->msg_buf_remaining);
>>> +        /* Underflow error condition: XFER_COMPLETE before message
>>> +         * fully sent.
>>> +         */
>>> +        if (WARN_ON_ONCE(i2c_dev->msg_buf_remaining)) {
>>> +            i2c_dev->msg_err |= I2C_ERR_UNKNOWN_INTERRUPT;
>>> +            goto err;
>>> +        }
>>>           complete(&i2c_dev->msg_complete);
>>>       }
>>>       goto done;
>>>
>>
>> Please address comments in the next revision.
>>
> 
> Sorry for the delayed reply. I shared Patch V7 with above changes earlier today.

No problems, please take your time. The V7 is almost good, looking forward to v8!