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] [day] [month] [year] [list]
Message-ID: <CAFbo-qVpgf-OMhwXjLvXPL9u3M_XCYoBaiogw-K+as_FX8t9qQ@mail.gmail.com>
Date:	Mon, 11 Jul 2016 22:38:50 +0530
From:	prahlad venkata <prahlad.eee@...il.com>
To:	Vignesh R <vigneshr@...com>
Cc:	"broonie@...nel.org" <broonie@...nel.org>,
	"linux-spi@...r.kernel.org" <linux-spi@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] spi: spi-ti-qspi: clear wlen field while setting word length.

On Mon, Jul 11, 2016 at 3:53 PM, Vignesh R <vigneshr@...com> wrote:
>
>
> On Monday 11 July 2016 02:49 PM, prahlad venkata wrote:
>> On Mon, Jul 11, 2016 at 2:45 PM, Vignesh R <vigneshr@...com> wrote:
> [...]
>>>>>>>>> diff --git a/drivers/spi/spi-ti-qspi.c b/drivers/spi/spi-ti-qspi.c
>>>>>>>>> index 29ea8d2..6c61f54 100644
>>>>>>>>> --- a/drivers/spi/spi-ti-qspi.c
>>>>>>>>> +++ b/drivers/spi/spi-ti-qspi.c
>>>>>>>>> @@ -276,9 +276,9 @@ static int qspi_write_msg(struct ti_qspi *qspi, struct spi_transfer *t,
>>>>>>>>>                               cmd |= QSPI_WLEN(QSPI_WLEN_MAX_BITS);
>>>>>>>>>                       } else {
>>>>>>>>>                               writeb(*txbuf, qspi->base + QSPI_SPI_DATA_REG);
>>>>>>>>> -                             cmd = qspi->cmd | QSPI_WR_SNGL;
>>>>>>
>>>>>> This is wrong. Deleting this line means QSPI_WR_SNGL is not set and no
>>>>>> data is sent out on the wire.
>>>>> QSPI_WR_SNGL is already set as soon as we enter the function.
>>>>>>
>>>>>>>>
>>>>>>>> qspi->cmd always has WLEN field cleared and set to WLEN = 1 byte (see
>>>>>>>> ti_qspi_start_transfer_one()). And hence variable 'cmd' will also have
>>>>>>>> WLEN set to 1 byte.
>>>>>>> Even though WLEN=1 is set in the ti_qspi_transfer_one, if we ask for a
>>>>>>> transfer of large data,
>>>>>>> say 300 bytes in length, for attaining faster data rate WLEN 128 is
>>>>>>> selected for the first two
>>>>>>> transactions and remaining 44 bytes will be transmitted with WLEN 1.
>>>>>>> During that case,
>>>>>>> WLEN will be changed inside qspi_write_msg function itself and the
>>>>>>> field should be cleared
>>>>>>> first while doing that.
>>>>>>
>>>>>> In qspi_write_msg(), qspi->cmd will always have WLEN set to
>>>>>> QSPI_WLEN(t->bits_per_word) and qspi->cmd is never changed within this
>>>>>> function.
>>>>>> It is the value of local variable 'cmd' that is changed to appropriate
>>>>>> WLEN (128bit or 8bit) as necessary.
>>>
>>>> 'cmd' is written back to qspi->cmd for every transaction.
>>>
>>> You mean qspi->cmd = cmd ?
>>> I don't see this happening anywhere in the driver. Can you point me to
>>> that line of code?
>> line 296:
>> ti_qspi_write(qspi, cmd, QSPI_SPI_CMD_REG);
>
> Sorry, I don't understand. QSPI_SPI_CMD_REG and qspi->cmd are different.
> qspi->cmd does not represent the QSPI_SPI_CMD_REG register. qspi->cmd is
> just local driver data for book-keeping.
>
> Please add some prints in driver to see how 'cmd' (and qspi->cmd)
> variable changes in case of 128bit mode and 8bit mode.
I don't have hardware setup to verify this. Is there anyway to verify
this without hardware?
>
> Regards
> Vignesh
>
>
> --
> Regards
> Vignesh



-- 
Regards,
Prahlad.

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ