On Fri, Nov 01, 2024 at 11:07:15PM +0200, Jarkko Sakkinen wrote:
> On Fri Nov 1, 2024 at 10:23 PM EET, Jerry Snitselaar wrote:
> > On Fri, Nov 01, 2024 at 02:21:56AM +0200, Jarkko Sakkinen wrote:
> > > Setting TPM_CHIP_FLAG_SUSPENDED in the end of tpm_pm_suspend() can be racy
> > > according, as this leaves window for tpm_hwrng_read() to be called while
> > > the operation is in progress. The recent bug report gives also evidence of
> > > this behaviour.
> > >
> > > Aadress this by locking the TPM chip before checking any chip->flags both
> > > in tpm_pm_suspend() and tpm_hwrng_read(). Move TPM_CHIP_FLAG_SUSPENDED
> > > check inside tpm_get_random() so that it will be always checked only when
> > > the lock is reserved.
> > >
> > > Cc: stable@xxxxxxxxxxxxxxx # v6.4+
> > > Fixes: 99d464506255 ("tpm: Prevent hwrng from activating during resume")
> > > Reported-by: Mike Seo <mikeseohyungjin@xxxxxxxxx>
> > > Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219383
> > > Signed-off-by: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
> > > ---
> > > v3:
> > > - Check TPM_CHIP_FLAG_SUSPENDED inside tpm_get_random() so that it is
> > > also done under the lock (suggested by Jerry Snitselaar).
> > > v2:
> > > - Addressed my own remark:
> > > https://lore.kernel.org/linux-integrity/D59JAI6RR2CD.G5E5T4ZCZ49W@xxxxxxxxxx/
> > > ---
> > > drivers/char/tpm/tpm-chip.c | 4 ----
> > > drivers/char/tpm/tpm-interface.c | 32 ++++++++++++++++++++++----------
> > > 2 files changed, 22 insertions(+), 14 deletions(-)
> > >
> > > diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
> > > index 1ff99a7091bb..7df7abaf3e52 100644
> > > --- a/drivers/char/tpm/tpm-chip.c
> > > +++ b/drivers/char/tpm/tpm-chip.c
> > > @@ -525,10 +525,6 @@ static int tpm_hwrng_read(struct hwrng *rng, void *data, size_t max, bool wait)
> > > {
> > > struct tpm_chip *chip = container_of(rng, struct tpm_chip, hwrng);
> > >
> > > - /* Give back zero bytes, as TPM chip has not yet fully resumed: */
> > > - if (chip->flags & TPM_CHIP_FLAG_SUSPENDED)
> > > - return 0;
> > > -
> > > return tpm_get_random(chip, data, max);
> > > }
> > >
> > > diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c
> > > index 8134f002b121..b1daa0d7b341 100644
> > > --- a/drivers/char/tpm/tpm-interface.c
> > > +++ b/drivers/char/tpm/tpm-interface.c
> > > @@ -370,6 +370,13 @@ int tpm_pm_suspend(struct device *dev)
> > > if (!chip)
> > > return -ENODEV;
> > >
> > > + rc = tpm_try_get_ops(chip);
> > > + if (rc) {
> > > + /* Can be safely set out of locks, as no action cannot race: */
> > > + chip->flags |= TPM_CHIP_FLAG_SUSPENDED;
> > > + goto out;
> > > + }
> > > +
> > > if (chip->flags & TPM_CHIP_FLAG_ALWAYS_POWERED)
> > > goto suspended;
> > >
> > > @@ -377,21 +384,19 @@ int tpm_pm_suspend(struct device *dev)
> > > !pm_suspend_via_firmware())
> > > goto suspended;
> > >
> > > - rc = tpm_try_get_ops(chip);
> > > - if (!rc) {
> > > - if (chip->flags & TPM_CHIP_FLAG_TPM2) {
> > > - tpm2_end_auth_session(chip);
> > > - tpm2_shutdown(chip, TPM2_SU_STATE);
> > > - } else {
> > > - rc = tpm1_pm_suspend(chip, tpm_suspend_pcr);
> > > - }
> > > -
> > > - tpm_put_ops(chip);
> > > + if (chip->flags & TPM_CHIP_FLAG_TPM2) {
> > > + tpm2_end_auth_session(chip);
> > > + tpm2_shutdown(chip, TPM2_SU_STATE);
> > > + goto suspended;
> > > }
> > >
> > > + rc = tpm1_pm_suspend(chip, tpm_suspend_pcr);
> > > +
> >
> >
> > I imagine the above still be wrapped in an else with the if (chip->flags & TPM_CHIP_FLAG_TPM2)
> > otherwise it will call tpm1_pm_suspend for both tpm1 and tpm2 devices, yes?
> >
> > So:
> >
> > if (chip->flags & TPM_CHIP_FLAG_TPM2) {
> > tpm2_end_auth_session(chip);
> > tpm2_shutdown(chip, TPM2_SU_STATE);
> > goto suspended;
> > } else {
> > rc = tpm1_pm_suspend(chip, tpm_suspend_pcr);
> > }
> >
> >
> > Other than that I think it looks good.
>
> It should be fine because after tpm2_shutdown() is called there is "goto
> suspended;". This is IMHO more readable as it matches the structure of
> previous exits before it. In future if this needs to be improved it will
> easier to move the logic to a helper function (e.g. __tpm_pm_suspend())
> where gotos are substituted with return-statements.
>
> BR, Jarkko
>
Heh, yep.
Reviewed-by: Jerry Snitselaar <jsnitsel@xxxxxxxxxx>
