Re: [PATCH 10/12] tpm: fix up the tpm_class shutdown_pre pointer when created

From: Jarkko Sakkinen
Date: Tue Mar 14 2023 - 09:03:31 EST


On Tue, Mar 14, 2023 at 01:53:59PM +0100, Greg Kroah-Hartman wrote:
> On Tue, Mar 14, 2023 at 01:09:04PM +0200, Jarkko Sakkinen wrote:
> > On Mon, Mar 13, 2023 at 07:18:41PM +0100, Greg Kroah-Hartman wrote:
> > > Do not wait until long after the struct class has been created to set
> > > the shutdown_pre pointer for the tpm_class, assign it right away.
> > >
> > > This is the only in-kernel offender that tries to modify the
> > > device->class pointer contents after it has been assigned to a device,
> > > so fix that up by doing the function pointer assignment before it is
> > > matched with the device. Because of this, the patch should go through
> > > the driver core tree to allow later changes to struct device to be
> > > possible.
> > >
> > > Cc: Peter Huewe <peterhuewe@xxxxxx>
> > > Cc: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
> > > Cc: Jason Gunthorpe <jgg@xxxxxxxx>
> > > Cc: linux-integrity@xxxxxxxxxxxxxxx
> > > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > > ---
> > > drivers/char/tpm/tpm-chip.c | 3 +--
> > > drivers/char/tpm/tpm-interface.c | 1 +
> > > drivers/char/tpm/tpm.h | 1 +
> > > 3 files changed, 3 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
> > > index b99f55f2d4fd..7c444209a256 100644
> > > --- a/drivers/char/tpm/tpm-chip.c
> > > +++ b/drivers/char/tpm/tpm-chip.c
> > > @@ -282,7 +282,7 @@ static void tpm_dev_release(struct device *dev)
> > > *
> > > * Return: always 0 (i.e. success)
> > > */
> > > -static int tpm_class_shutdown(struct device *dev)
> > > +int tpm_class_shutdown(struct device *dev)
> > > {
> > > struct tpm_chip *chip = container_of(dev, struct tpm_chip, dev);
> > >
> > > @@ -337,7 +337,6 @@ struct tpm_chip *tpm_chip_alloc(struct device *pdev,
> > > device_initialize(&chip->dev);
> > >
> > > chip->dev.class = tpm_class;
> > > - chip->dev.class->shutdown_pre = tpm_class_shutdown;
> > > chip->dev.release = tpm_dev_release;
> > > chip->dev.parent = pdev;
> > > chip->dev.groups = chip->groups;
> > > diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c
> > > index 8763c820d1f8..43e23a04433a 100644
> > > --- a/drivers/char/tpm/tpm-interface.c
> > > +++ b/drivers/char/tpm/tpm-interface.c
> > > @@ -467,6 +467,7 @@ static int __init tpm_init(void)
> > > int rc;
> > >
> > > tpm_class = class_create("tpm");
> > > + tpm_class->shutdown_pre = tpm_class_shutdown;
> > > if (IS_ERR(tpm_class)) {
> > > pr_err("couldn't create tpm class\n");
> > > return PTR_ERR(tpm_class);
> > > diff --git a/drivers/char/tpm/tpm.h b/drivers/char/tpm/tpm.h
> > > index 24ee4e1cc452..a45eb39db0c4 100644
> > > --- a/drivers/char/tpm/tpm.h
> > > +++ b/drivers/char/tpm/tpm.h
> > > @@ -183,6 +183,7 @@ int tpm1_get_pcr_allocation(struct tpm_chip *chip);
> > > unsigned long tpm_calc_ordinal_duration(struct tpm_chip *chip, u32 ordinal);
> > > int tpm_pm_suspend(struct device *dev);
> > > int tpm_pm_resume(struct device *dev);
> > > +int tpm_class_shutdown(struct device *dev);
> > >
> > > static inline void tpm_msleep(unsigned int delay_msec)
> > > {
> > > --
> > > 2.39.2
> > >
> >
> > Reviewed-by: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
> >
> > Are you going to pick this?
>
> Yes, as the changelog text said, I'd like to take it through my tree,
> thanks!

OK, right. Just wanted to prevent any races, thanks.

BR, Jarkko