linux-kernel - Re: [PATCH] i2c: ocores: generate stop condition after timeout in polling mode

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

Date:   Wed, 12 Apr 2023 13:50:56 +0200
From:   Federico Vaga <federico.vaga@...n.ch>
To:     Matthias Schiffer <matthias.schiffer@...tq-group.com>
Cc:     Peter Korsgaard <peter@...sgaard.com>,
        Andrew Lunn <andrew@...n.ch>, Wolfram Sang <wsa@...nel.org>,
        "linux-i2c@...r.kernel.org" <linux-i2c@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux@...tq-group.com" <linux@...tq-group.com>
Subject: Re: [PATCH] i2c: ocores: generate stop condition after timeout in
 polling mode

On Wed, Apr 12, 2023 at 11:49:32AM +0200, Matthias Schiffer wrote:
>On Mon, 2023-02-20 at 16:17 +0000, Matthias Schiffer wrote:
>> From: Gregor Herburger <gregor.herburger@...group.com>
>>
>> In polling mode, no stop condition is generated after a timeout. This
>> causes SCL to remain low and thereby block the bus. If this happens
>> during a transfer it can cause slaves to misinterpret the subsequent
>> transfer and return wrong values.
>>
>> To solve this, pass the ETIMEDOUT error up from ocores_process_polling()
>> instead of setting STATE_ERROR directly. The caller is adjusted to call
>> ocores_process_timeout() on error both in polling and in IRQ mode, which
>> will set STATE_ERROR and generate a stop condition.

I can see and confirm the bug.

The patch seems fine to me, but I do not have the hardware at hand right now for a
real test.

Following a few cosmetic comments.

>> Fixes: 69c8c0c0efa8 ("i2c: ocores: add polling interface")
>> Signed-off-by: Gregor Herburger <gregor.herburger@...group.com>
>> Signed-off-by: Matthias Schiffer <matthias.schiffer@...tq-group.com>
>
>
>Any chance we can get someone to have a look at this patch?
>
>Regards,
>Matthias
>
>
>
>> ---
>>  drivers/i2c/busses/i2c-ocores.c | 28 +++++++++++++++-------------
>>  1 file changed, 15 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/i2c/busses/i2c-ocores.c b/drivers/i2c/busses/i2c-ocores.c
>> index a0af027db04c1..28bcda3f7040a 100644
>> --- a/drivers/i2c/busses/i2c-ocores.c
>> +++ b/drivers/i2c/busses/i2c-ocores.c
>> @@ -342,18 +342,18 @@ static int ocores_poll_wait(struct ocores_i2c *i2c)
>>   * ocores_isr(), we just add our polling code around it.
>>   *
>>   * It can run in atomic context
>> + *
>> + * Return: 0 on success, -ETIMEDOUT on timeout
>>   */
>> -static void ocores_process_polling(struct ocores_i2c *i2c)
>> +static int ocores_process_polling(struct ocores_i2c *i2c)
>>  {
>>  	while (1) {
>>  		irqreturn_t ret;
>>  		int err;
>>
>>  		err = ocores_poll_wait(i2c);
>> -		if (err) {
>> -			i2c->state = STATE_ERROR;
>> -			break; /* timeout */
>> -		}
>> +		if (err)
>> +			return err;
>>
>>  		ret = ocores_isr(-1, i2c);
>>  		if (ret == IRQ_NONE)
>> @@ -364,6 +364,8 @@ static void ocores_process_polling(struct ocores_i2c *i2c)
>>  					break;
>>  		}
>>  	}
>> +
>> +	return 0;
>>  }

I would try to have a single return point at the end `return err` and inside the
loop just a break.

>>
>>  static int ocores_xfer_core(struct ocores_i2c *i2c,
>> @@ -387,16 +389,16 @@ static int ocores_xfer_core(struct ocores_i2c *i2c,
>>  	oc_setreg(i2c, OCI2C_DATA, i2c_8bit_addr_from_msg(i2c->msg));
>>  	oc_setreg(i2c, OCI2C_CMD, OCI2C_CMD_START);
>>
>> -	if (polling) {
>> -		ocores_process_polling(i2c);
>> -	} else {
>> +	if (polling)
>> +		ret = ocores_process_polling(i2c);
>> +	else
>>  		ret = wait_event_timeout(i2c->wait,
>>  					 (i2c->state == STATE_ERROR) ||
>> -					 (i2c->state == STATE_DONE), HZ);
>> -		if (ret == 0) {
>> -			ocores_process_timeout(i2c);
>> -			return -ETIMEDOUT;
>> -		}
>> +					 (i2c->state == STATE_DONE), HZ) ?
>> +						0 : -ETIMEDOUT;

For readability, I would avoid the ternary operator here and do the assignement
with an `if` as it was before.

if (ret == 0)
     ret = -ETIMEDOUT;

>> +	if (ret) {
>> +		ocores_process_timeout(i2c);
>> +		return ret;
>>  	}
>>
>>  	return (i2c->state == STATE_DONE) ? num : -EIO;
>
>-- 
>TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany
>Amtsgericht München, HRB 105018
>Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider
>http://www.tq-group.com/
>

-- 
Federico Vaga - CERN BE-CEM-EDL