From: Jarkko Sakkinen <jarkko@kernel.org>
To: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Cc: peterhuewe@gmx.de, jgg@ziepe.ca, stefanb@linux.vnet.ibm.com,
linux@mniewoehner.de, linux-integrity@vger.kernel.org,
linux-kernel@vger.kernel.org, lukas@wunner.de,
p.rosenberger@kunbus.com,
Lino Sanfilippo <l.sanfilippo@kunbus.com>
Subject: Re: [PATCH v4 4/6] tpm, tpm_tis: avoid CPU cache incoherency in irq test
Date: Wed, 11 May 2022 18:06:51 +0300 [thread overview]
Message-ID: <YnvRC3FI7bcXj7zL@kernel.org> (raw)
In-Reply-To: <20220509080559.4381-5-LinoSanfilippo@gmx.de>
On Mon, May 09, 2022 at 10:05:57AM +0200, Lino Sanfilippo wrote:
> From: Lino Sanfilippo <l.sanfilippo@kunbus.com>
>
> The interrupt handler that sets irq_tested to indicate that interrupts are
> working may run on another CPU than the thread that checks this variable in
> tmp_tis_send().
>
> Since no synchronization is used to access irq_tested, there is no
> guarantee for cache coherency between the CPUs, so that the value set by
> the interrupt handler might not be visible to the testing thread.
>
> Avoid this issue by using a bitfield instead of a boolean variable and by
> accessing this field with bit manipulating functions that guarantee cache
> coherency.
>
> Signed-off-by: Lino Sanfilippo <l.sanfilippo@kunbus.com>
> ---
> drivers/char/tpm/tpm_tis_core.c | 13 +++++++------
> drivers/char/tpm/tpm_tis_core.h | 6 +++++-
> 2 files changed, 12 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/char/tpm/tpm_tis_core.c b/drivers/char/tpm/tpm_tis_core.c
> index 4f3b82c3f205..bdfde1cd71fe 100644
> --- a/drivers/char/tpm/tpm_tis_core.c
> +++ b/drivers/char/tpm/tpm_tis_core.c
> @@ -470,7 +470,8 @@ static int tpm_tis_send(struct tpm_chip *chip, u8 *buf, size_t len)
> int rc, irq;
> struct tpm_tis_data *priv = dev_get_drvdata(&chip->dev);
>
> - if (!(chip->flags & TPM_CHIP_FLAG_IRQ) || priv->irq_tested)
> + if (!(chip->flags & TPM_CHIP_FLAG_IRQ) ||
> + test_bit(TPM_TIS_IRQTEST_OK, &priv->irqtest_flags))
> return tpm_tis_send_main(chip, buf, len);
>
> /* Verify receipt of the expected IRQ */
> @@ -480,11 +481,11 @@ static int tpm_tis_send(struct tpm_chip *chip, u8 *buf, size_t len)
> rc = tpm_tis_send_main(chip, buf, len);
> priv->irq = irq;
> chip->flags |= TPM_CHIP_FLAG_IRQ;
> - if (!priv->irq_tested)
> + if (!test_bit(TPM_TIS_IRQTEST_OK, &priv->irqtest_flags))
> tpm_msleep(1);
> - if (!priv->irq_tested)
> + if (!test_bit(TPM_TIS_IRQTEST_OK, &priv->irqtest_flags))
> disable_interrupts(chip);
> - priv->irq_tested = true;
> + set_bit(TPM_TIS_IRQTEST_OK, &priv->irqtest_flags);
> return rc;
> }
>
> @@ -689,7 +690,7 @@ static irqreturn_t tis_int_handler(int dummy, void *dev_id)
> if (interrupt == 0)
> return IRQ_NONE;
>
> - priv->irq_tested = true;
> + set_bit(TPM_TIS_IRQTEST_OK, &priv->irqtest_flags);
> if (interrupt & TPM_INTF_DATA_AVAIL_INT)
> wake_up_interruptible(&priv->read_queue);
> if (interrupt & TPM_INTF_LOCALITY_CHANGE_INT)
> @@ -780,7 +781,7 @@ static int tpm_tis_probe_irq_single(struct tpm_chip *chip, u32 intmask,
> if (rc < 0)
> return rc;
>
> - priv->irq_tested = false;
> + clear_bit(TPM_TIS_IRQTEST_OK, &priv->irqtest_flags);
> chip->flags |= TPM_CHIP_FLAG_IRQ;
>
> /* Generate an interrupt by having the core call through to
> diff --git a/drivers/char/tpm/tpm_tis_core.h b/drivers/char/tpm/tpm_tis_core.h
> index 43b724e55192..c8972ea8e13e 100644
> --- a/drivers/char/tpm/tpm_tis_core.h
> +++ b/drivers/char/tpm/tpm_tis_core.h
> @@ -89,11 +89,15 @@ enum tpm_tis_flags {
> TPM_TIS_USE_THREADED_IRQ = BIT(2),
> };
>
> +enum tpm_tis_irqtest_flags {
> + TPM_TIS_IRQTEST_OK = BIT(0),
> +};
> +
> struct tpm_tis_data {
> u16 manufacturer_id;
> int locality;
> int irq;
> - bool irq_tested;
> + unsigned long irqtest_flags;
> unsigned long flags;
> void __iomem *ilb_base_addr;
> u16 clkrun_enabled;
> --
> 2.36.0
>
So, this would caused by changing to threaded IRQs?
BR, Jarkko
next prev parent reply other threads:[~2022-05-11 15:08 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-09 8:05 [PATCH v4 0/6] TPM irq fixes Lino Sanfilippo
2022-05-09 8:05 ` [PATCH v4 1/6] tpm, tpm_tis_spi: Request threaded irq Lino Sanfilippo
2022-05-11 11:22 ` Jarkko Sakkinen
2022-05-11 19:18 ` Lino Sanfilippo
2022-05-13 18:08 ` Jarkko Sakkinen
2022-05-16 19:52 ` Lino Sanfilippo
2022-05-17 18:23 ` Jarkko Sakkinen
2022-05-09 8:05 ` [PATCH v4 2/6] tpm, tpm_tis: Claim and release locality only once Lino Sanfilippo
2022-05-11 11:27 ` Jarkko Sakkinen
2022-05-11 19:29 ` Lino Sanfilippo
2022-05-13 17:59 ` Jarkko Sakkinen
2022-05-16 20:23 ` Lino Sanfilippo
2022-05-09 8:05 ` [PATCH v4 3/6] tpm, tpm_tis: enable irq test Lino Sanfilippo
2022-05-09 8:05 ` [PATCH v4 4/6] tpm, tpm_tis: avoid CPU cache incoherency in " Lino Sanfilippo
2022-05-11 15:06 ` Jarkko Sakkinen [this message]
2022-05-11 19:35 ` Lino Sanfilippo
2022-05-09 8:05 ` [PATCH v4 5/6] tpm, tpm_tis: Move irq test from tpm_tis_send() to tpm_tis_probe_irq_single() Lino Sanfilippo
2022-05-11 15:09 ` Jarkko Sakkinen
2022-05-11 19:56 ` Lino Sanfilippo
2022-05-16 17:51 ` Jarkko Sakkinen
2022-05-16 20:25 ` Lino Sanfilippo
2022-05-17 18:19 ` Michael Niewöhner
2022-05-18 1:26 ` Jarkko Sakkinen
2022-05-18 16:08 ` Michael Niewöhner
2022-05-09 8:05 ` [PATCH v4 6/6] tpm, tpm_tis: Only enable supported IRQs Lino Sanfilippo
2022-05-11 15:08 ` Jarkko Sakkinen
2022-05-11 19:58 ` Lino Sanfilippo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YnvRC3FI7bcXj7zL@kernel.org \
--to=jarkko@kernel.org \
--cc=LinoSanfilippo@gmx.de \
--cc=jgg@ziepe.ca \
--cc=l.sanfilippo@kunbus.com \
--cc=linux-integrity@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@mniewoehner.de \
--cc=lukas@wunner.de \
--cc=p.rosenberger@kunbus.com \
--cc=peterhuewe@gmx.de \
--cc=stefanb@linux.vnet.ibm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome