__compat_only_sysfs_link_entry_to_kobj() was unconditionally called for
TPM1 chips, which caused crash on Acer C720 laptop where DSM for the
ACPI object did not exist.
There are two reasons for unwanted behavior:
* The code did not check whether
__compat_only_sysfs_link_entry_to_kobj() returned -ENOENT. This is
OK. It just meanst that ppi is not available.
* The code did not clean up properly. Compat link should added only
after all other init is done.
This patch sorts out these issues.
Fixes: 9b774d5cf2db
Reported-by: Jeremiah Mahler <[email protected]>
Signed-off-by: Jarkko Sakkinen <[email protected]>
Tested-by: Jeremiah Mahler <[email protected]>
---
drivers/char/tpm/tpm-chip.c | 18 ++++++++++--------
1 file changed, 10 insertions(+), 8 deletions(-)
diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
index a5cdce7..45cc39a 100644
--- a/drivers/char/tpm/tpm-chip.c
+++ b/drivers/char/tpm/tpm-chip.c
@@ -226,14 +226,6 @@ int tpm_chip_register(struct tpm_chip *chip)
if (rc)
goto out_err;
- if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
- rc = __compat_only_sysfs_link_entry_to_kobj(&chip->pdev->kobj,
- &chip->dev.kobj,
- "ppi");
- if (rc)
- goto out_err;
- }
-
/* Make the chip available. */
spin_lock(&driver_lock);
list_add_tail_rcu(&chip->list, &tpm_chip_list);
@@ -241,6 +233,16 @@ int tpm_chip_register(struct tpm_chip *chip)
chip->flags |= TPM_CHIP_FLAG_REGISTERED;
+ if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
+ rc = __compat_only_sysfs_link_entry_to_kobj(&chip->pdev->kobj,
+ &chip->dev.kobj,
+ "ppi");
+ if (rc && rc != -ENOENT) {
+ tpm_chip_unregister(chip);
+ return rc;
+ }
+ }
+
return 0;
out_err:
tpm1_chip_unregister(chip);
--
2.5.0
On Sun, Nov 08, 2015 at 09:51:07AM +0200, Jarkko Sakkinen wrote:
> + if (rc && rc != -ENOENT) {
> + tpm_chip_unregister(chip);
> + return rc;
> + }
> + }
This is still goofy looking, the list_add_tail_rcu should be the last
thing done and cannot fail. Just code this with the usual goto based
unwind and goto to do tpm_dev_del_device for the above failure.
> +
> return 0;
> out_err:
> tpm1_chip_unregister(chip);
Which avoids calling tpm1_chip_unregister and tpm_chip_unregister in
the same function, which looks so wrong to a casual read..
Jason
On Mon, Nov 09, 2015 at 03:47:20PM -0700, Jason Gunthorpe wrote:
> On Sun, Nov 08, 2015 at 09:51:07AM +0200, Jarkko Sakkinen wrote:
> > + if (rc && rc != -ENOENT) {
> > + tpm_chip_unregister(chip);
> > + return rc;
> > + }
> > + }
>
> This is still goofy looking, the list_add_tail_rcu should be the last
> thing done and cannot fail. Just code this with the usual goto based
> unwind and goto to do tpm_dev_del_device for the above failure.
Agreed that this would be cleaner.
> > +
> > return 0;
> > out_err:
> > tpm1_chip_unregister(chip);
>
> Which avoids calling tpm1_chip_unregister and tpm_chip_unregister in
> the same function, which looks so wrong to a casual read..
I'm not going to change the current patch at this point of release
cycle but this clean up definitely makes sense later on.
> Jason
/Jarkko
Jarkko,
On Sun, Nov 08, 2015 at 09:51:07AM +0200, Jarkko Sakkinen wrote:
> __compat_only_sysfs_link_entry_to_kobj() was unconditionally called for
> TPM1 chips, which caused crash on Acer C720 laptop where DSM for the
> ACPI object did not exist.
>
> There are two reasons for unwanted behavior:
>
> * The code did not check whether
> __compat_only_sysfs_link_entry_to_kobj() returned -ENOENT. This is
> OK. It just meanst that ppi is not available.
> * The code did not clean up properly. Compat link should added only
> after all other init is done.
>
> This patch sorts out these issues.
>
> Fixes: 9b774d5cf2db
> Reported-by: Jeremiah Mahler <[email protected]>
> Signed-off-by: Jarkko Sakkinen <[email protected]>
> Tested-by: Jeremiah Mahler <[email protected]>
> ---
> drivers/char/tpm/tpm-chip.c | 18 ++++++++++--------
> 1 file changed, 10 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
> index a5cdce7..45cc39a 100644
> --- a/drivers/char/tpm/tpm-chip.c
> +++ b/drivers/char/tpm/tpm-chip.c
> @@ -226,14 +226,6 @@ int tpm_chip_register(struct tpm_chip *chip)
> if (rc)
> goto out_err;
>
> - if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
> - rc = __compat_only_sysfs_link_entry_to_kobj(&chip->pdev->kobj,
> - &chip->dev.kobj,
> - "ppi");
> - if (rc)
> - goto out_err;
> - }
> -
> /* Make the chip available. */
> spin_lock(&driver_lock);
> list_add_tail_rcu(&chip->list, &tpm_chip_list);
> @@ -241,6 +233,16 @@ int tpm_chip_register(struct tpm_chip *chip)
>
> chip->flags |= TPM_CHIP_FLAG_REGISTERED;
>
> + if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
> + rc = __compat_only_sysfs_link_entry_to_kobj(&chip->pdev->kobj,
> + &chip->dev.kobj,
> + "ppi");
> + if (rc && rc != -ENOENT) {
> + tpm_chip_unregister(chip);
> + return rc;
> + }
> + }
> +
> return 0;
> out_err:
> tpm1_chip_unregister(chip);
> --
> 2.5.0
>
This patch doesn't apply to the latest linux-next (20151109). I think I
may be missing a dependent patch. Do you know which one?
--
- Jeremiah Mahler
On Mon, Nov 09, 2015 at 10:11:41PM -0800, Jeremiah Mahler wrote:
> Jarkko,
>
> On Sun, Nov 08, 2015 at 09:51:07AM +0200, Jarkko Sakkinen wrote:
> > __compat_only_sysfs_link_entry_to_kobj() was unconditionally called for
> > TPM1 chips, which caused crash on Acer C720 laptop where DSM for the
> > ACPI object did not exist.
> >
> > There are two reasons for unwanted behavior:
> >
> > * The code did not check whether
> > __compat_only_sysfs_link_entry_to_kobj() returned -ENOENT. This is
> > OK. It just meanst that ppi is not available.
> > * The code did not clean up properly. Compat link should added only
> > after all other init is done.
> >
> > This patch sorts out these issues.
> >
> > Fixes: 9b774d5cf2db
> > Reported-by: Jeremiah Mahler <[email protected]>
> > Signed-off-by: Jarkko Sakkinen <[email protected]>
> > Tested-by: Jeremiah Mahler <[email protected]>
> > ---
> > drivers/char/tpm/tpm-chip.c | 18 ++++++++++--------
> > 1 file changed, 10 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c
> > index a5cdce7..45cc39a 100644
> > --- a/drivers/char/tpm/tpm-chip.c
> > +++ b/drivers/char/tpm/tpm-chip.c
> > @@ -226,14 +226,6 @@ int tpm_chip_register(struct tpm_chip *chip)
> > if (rc)
> > goto out_err;
> >
> > - if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
> > - rc = __compat_only_sysfs_link_entry_to_kobj(&chip->pdev->kobj,
> > - &chip->dev.kobj,
> > - "ppi");
> > - if (rc)
> > - goto out_err;
> > - }
> > -
> > /* Make the chip available. */
> > spin_lock(&driver_lock);
> > list_add_tail_rcu(&chip->list, &tpm_chip_list);
> > @@ -241,6 +233,16 @@ int tpm_chip_register(struct tpm_chip *chip)
> >
> > chip->flags |= TPM_CHIP_FLAG_REGISTERED;
> >
> > + if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
> > + rc = __compat_only_sysfs_link_entry_to_kobj(&chip->pdev->kobj,
> > + &chip->dev.kobj,
> > + "ppi");
> > + if (rc && rc != -ENOENT) {
> > + tpm_chip_unregister(chip);
> > + return rc;
> > + }
> > + }
> > +
> > return 0;
> > out_err:
> > tpm1_chip_unregister(chip);
> > --
> > 2.5.0
> >
>
> This patch doesn't apply to the latest linux-next (20151109). I think I
> may be missing a dependent patch. Do you know which one?
It is expected as there are 7 other bug fixes that I'm including to the
next pull rquest and this is on top.
> - Jeremiah Mahler
/Jarkko