linux-kernel - Re: [PATCH 11/18] spi: qup: properly detect extra interrupts

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]

Message-ID: <CAEsOVNfKcc7BxRpU7S8m30kuHhh9j4MUw5pY50aWWaQnUDfvng@mail.gmail.com>
Date:   Wed, 14 Jun 2017 16:51:07 -0500
From:   Matthew McClintock <mmcclint@...eaurora.org>
To:     Andy Gross <andy.gross@...aro.org>
Cc:     Sricharan R <sricharan@...eaurora.org>,
        Varadarajan Narayanan <varada@...eaurora.org>,
        broonie@...nel.org, robh+dt@...nel.org, mark.rutland@....com,
        david.brown@...aro.org, linux-spi@...r.kernel.org,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-msm@...r.kernel.org, linux-soc@...r.kernel.org
Subject: Re: [PATCH 11/18] spi: qup: properly detect extra interrupts

On Wed, Jun 14, 2017 at 2:59 PM, Andy Gross <andy.gross@...aro.org> wrote:
> On Wed, Jun 14, 2017 at 12:57:25PM +0530, Sricharan R wrote:
>> Hi Varada,
>>
>> On 6/14/2017 11:22 AM, Varadarajan Narayanan wrote:
>> > It's possible for a SPI transaction to complete and get another
>> > interrupt and have it processed on the same spi_transfer before the
>> > transfer_one can set it to NULL.
>> >
>> > This masks unexpected interrupts, so let's set the spi_transfer to
>> > NULL in the interrupt once the transaction is done. So we can
>> > properly detect these bad interrupts and print warning messages.
>> >
>> > Signed-off-by: Matthew McClintock <mmcclint@...eaurora.org>
>> > Signed-off-by: Varadarajan Narayanan <varada@...eaurora.org>
>> > ---
>> >  drivers/spi/spi-qup.c | 20 +++++++++++---------
>> >  1 file changed, 11 insertions(+), 9 deletions(-)
>> >
>> > diff --git a/drivers/spi/spi-qup.c b/drivers/spi/spi-qup.c
>> > index bd53e82..1a2a9d9 100644
>> > --- a/drivers/spi/spi-qup.c
>> > +++ b/drivers/spi/spi-qup.c
>> > @@ -496,13 +496,13 @@ static irqreturn_t spi_qup_qup_irq(int irq, void *dev_id)
>> >     struct spi_qup *controller = dev_id;
>> >     struct spi_transfer *xfer;
>> >     u32 opflags, qup_err, spi_err;
>> > -   unsigned long flags;
>> >     int error = 0;
>> > +   bool done = 0;
>> >
>> > -   spin_lock_irqsave(&controller->lock, flags);
>> > +   spin_lock(&controller->lock);
>> >     xfer = controller->xfer;
>> >     controller->xfer = NULL;
>> > -   spin_unlock_irqrestore(&controller->lock, flags);
>> > +   spin_unlock(&controller->lock);
>>
>>  Why change the locking here ?
>>
>> >
>> >     qup_err = readl_relaxed(controller->base + QUP_ERROR_FLAGS);
>> >     spi_err = readl_relaxed(controller->base + SPI_ERROR_FLAGS);
>> > @@ -556,16 +556,19 @@ static irqreturn_t spi_qup_qup_irq(int irq, void *dev_id)
>> >                     spi_qup_write(controller, xfer);
>> >     }
>> >
>> > -   spin_lock_irqsave(&controller->lock, flags);
>> > -   controller->error = error;
>> > -   controller->xfer = xfer;
>> > -   spin_unlock_irqrestore(&controller->lock, flags);
>> > -
>> >     /* re-read opflags as flags may have changed due to actions above */
>> >     opflags = readl_relaxed(controller->base + QUP_OPERATIONAL);
>> >
>> >     if ((controller->rx_bytes == xfer->len &&
>> >             (opflags & QUP_OP_MAX_INPUT_DONE_FLAG)) ||  error)
>> > +           done = true;
>> > +
>> > +   spin_lock(&controller->lock);
>> > +   controller->error = error;
>> > +   controller->xfer = done ? NULL : xfer;
>> > +   spin_unlock(&controller->lock);
>> > +
>> > +   if (done)
>> >             complete(&controller->done);
>> >
>>   Its not clear, why the driver is setting the controller->xfer = NULL
>>   and restoring it inside the irq. This patch seems to fix things on
>>   top of that.
>
> I think the original intent was to make sure that the irqhandler knew that there
> was no outstanding transaction.  This begs the question of why that would ever
> be necessary.  I think it would suffice to rework all of that to remove that
> behavior and perhaps enable/disable the irq as we need to during transactions.
>
> I've never been a fan of the controller->xfer being set to NULL.

Also, this patch presumably fixes an issue seen on Dakota SoCs,
doesn't add or remote the NULL bits.

-M