[PATCH] tpm: fix suspend/resume paths for TPM 2.0

Fixed suspend/resume paths for TPM 2.0 and consolidated all the
associated code to the tpm_pm_suspend() and tpm_pm_resume()
functions. Resume path should be handled by the firmware, i.e.
Startup(CLEAR) for hibernate and Startup(STATE) for suspend.

There might be some non-PC embedded devices in the future where
Startup() is not the handled by the FW but fixing the code for
those IMHO should be postponed until there is hardware available
to test the fixes although extra Startup in the driver code is
essentially a NOP.

Reported-by: Peter Hüwe <redacted>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>
---
 drivers/char/tpm/tpm-interface.c | 10 ++++++++--
 drivers/char/tpm/tpm_crb.c       | 16 +---------------
 drivers/char/tpm/tpm_tis.c       | 22 ++++++++++------------
 3 files changed, 19 insertions(+), 29 deletions(-)

diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c
index bf53a37..93c8b90fd 100644
--- a/drivers/char/tpm/tpm-interface.c
+++ b/drivers/char/tpm/tpm-interface.c

@@ -901,8 +901,13 @@ int tpm_pm_suspend(struct device *dev)
 	if (chip == NULL)
 		return -ENODEV;
 
-	if (chip->flags & TPM_CHIP_FLAG_TPM2)
-		return tpm2_shutdown(chip, TPM2_SU_CLEAR);
+	if (chip->flags & TPM_CHIP_FLAG_TPM2) {
+		rc = tpm2_shutdown(chip, TPM2_SU_STATE);
+		if (rc < 0)
+			return rc;
+
+		return 0;
+	}
 
 	/* for buggy tpm, flush pcrs with extend to selected dummy */
 	if (tpm_suspend_pcr) {

@@ -952,6 +957,7 @@ EXPORT_SYMBOL_GPL(tpm_pm_suspend);
 int tpm_pm_resume(struct device *dev)
 {
 	struct tpm_chip *chip = dev_get_drvdata(dev);
+	int ret;
 
 	if (chip == NULL)
 		return -ENODEV;

diff --git a/drivers/char/tpm/tpm_crb.c b/drivers/char/tpm/tpm_crb.c
index 3dd23cf..3e080f5 100644
--- a/drivers/char/tpm/tpm_crb.c
+++ b/drivers/char/tpm/tpm_crb.c

@@ -95,21 +95,7 @@ struct crb_priv {
 	u8 __iomem *rsp;
 };
 
-#ifdef CONFIG_PM_SLEEP
-static int crb_resume(struct device *dev)
-{
-	int rc;
-	struct tpm_chip *chip = dev_get_drvdata(dev);
-
-	rc = tpm2_shutdown(chip, TPM2_SU_STATE);
-	if (!rc)
-		rc = tpm2_do_selftest(chip);
-
-	return rc;
-}
-#endif
-
-static SIMPLE_DEV_PM_OPS(crb_pm, tpm_pm_suspend, crb_resume);
+static SIMPLE_DEV_PM_OPS(crb_pm, tpm_pm_suspend, tpm_pm_resume);
 
 static u8 crb_status(struct tpm_chip *chip)
 {

diff --git a/drivers/char/tpm/tpm_tis.c b/drivers/char/tpm/tpm_tis.c
index 6725bef..c105eb7 100644
--- a/drivers/char/tpm/tpm_tis.c
+++ b/drivers/char/tpm/tpm_tis.c

@@ -865,25 +865,23 @@ static void tpm_tis_reenable_interrupts(struct tpm_chip *chip)
 static int tpm_tis_resume(struct device *dev)
 {
 	struct tpm_chip *chip = dev_get_drvdata(dev);
-	int ret = 0;
+	int ret;
 
 	if (chip->vendor.irq)
 		tpm_tis_reenable_interrupts(chip);
 
-	if (chip->flags & TPM_CHIP_FLAG_TPM2) {
-		/* NOP if firmware properly does this. */
-		tpm2_startup(chip, TPM2_SU_STATE);
+	ret = tpm_pm_resume(dev);
+	if (ret)
+		return ret;
 
-		ret = tpm2_shutdown(chip, TPM2_SU_STATE);
-		if (!ret)
-			ret = tpm2_do_selftest(chip);
-	} else {
-		ret = tpm_pm_resume(dev);
-		if (!ret)
-			tpm_do_selftest(chip);
+	/* TPM 1.2 requires self-test on resume. */
+	if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
+		ret = tpm_do_selftest(chip);
+		if (ret < 0)
+			return ret;
 	}
 
-	return ret;
+	return 0;
 }
 #endif
 

-- 
2.1.4

On Tue, 27 Jan 2015, Jarkko Sakkinen wrote:

Fixed suspend/resume paths for TPM 2.0 and consolidated all the
associated code to the tpm_pm_suspend() and tpm_pm_resume()
functions. Resume path should be handled by the firmware, i.e.
Startup(CLEAR) for hibernate and Startup(STATE) for suspend.

There might be some non-PC embedded devices in the future where
Startup() is not the handled by the FW but fixing the code for
those IMHO should be postponed until there is hardware available
to test the fixes although extra Startup in the driver code is
essentially a NOP.

Reported-by: Peter Hüwe <PeterHuewe@gmx.de>
Signed-off-by: Jarkko Sakkinen <redacted>
---

...

...

+	/* TPM 1.2 requires self-test on resume. */
+	if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
+		ret = tpm_do_selftest(chip);
+		if (ret < 0)
+			return ret;

Just to note, the return value from tpm_do_selftest() on TPM 1.2 chips was 
previously ignored. Mine does return 0.

On Tue, 2015-01-27 at 16:52 +0000, Scot Doyle wrote:

On Tue, 27 Jan 2015, Jarkko Sakkinen wrote:

quoted

Fixed suspend/resume paths for TPM 2.0 and consolidated all the
associated code to the tpm_pm_suspend() and tpm_pm_resume()
functions. Resume path should be handled by the firmware, i.e.
Startup(CLEAR) for hibernate and Startup(STATE) for suspend.

There might be some non-PC embedded devices in the future where
Startup() is not the handled by the FW but fixing the code for
those IMHO should be postponed until there is hardware available
to test the fixes although extra Startup in the driver code is
essentially a NOP.

Reported-by: Peter Hüwe <redacted>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>
---

...

quoted

--- a/drivers/char/tpm/tpm_tis.c
+++ b/drivers/char/tpm/tpm_tis.c

@@ -865,25 +865,23 @@ static void tpm_tis_reenable_interrupts(struct tpm_chip *chip)
 static int tpm_tis_resume(struct device *dev)
 {

...

quoted

+	/* TPM 1.2 requires self-test on resume. */
+	if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
+		ret = tpm_do_selftest(chip);
+		if (ret < 0)
+			return ret;

Just to note, the return value from tpm_do_selftest() on TPM 1.2 chips was 
previously ignored. Mine does return 0.

Right. I can update the patch to ignore return value if the majority
wants that.

/Jarkko

On Tue, Jan 27, 2015 at 06:57:22PM +0200, Jarkko Sakkinen wrote:

quoted

quoted

+	/* TPM 1.2 requires self-test on resume. */
+	if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
+		ret = tpm_do_selftest(chip);
+		if (ret < 0)
+			return ret;

Just to note, the return value from tpm_do_selftest() on TPM 1.2 chips was 
previously ignored. Mine does return 0.

Right. I can update the patch to ignore return value if the majority
wants that.

What happens to the system when pnp_driver.resume() returns failure?

Should tpm ever report failure on resume to the rest of the kernel?

Shouldn't this stuff be in tpm_pm_resume common code anyhow?

Jason

On Tue, 2015-01-27 at 10:03 -0700, Jason Gunthorpe wrote:

On Tue, Jan 27, 2015 at 06:57:22PM +0200, Jarkko Sakkinen wrote:

quoted

quoted

quoted

+	/* TPM 1.2 requires self-test on resume. */
+	if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
+		ret = tpm_do_selftest(chip);
+		if (ret < 0)
+			return ret;

Just to note, the return value from tpm_do_selftest() on TPM 1.2 chips was 
previously ignored. Mine does return 0.

Right. I can update the patch to ignore return value if the majority
wants that.

What happens to the system when pnp_driver.resume() returns failure?

Should tpm ever report failure on resume to the rest of the kernel?

Shouldn't this stuff be in tpm_pm_resume common code anyhow?

I think it should but not in the scope of this bug fix IMHO.

Jason

/Jarkko

On Tue, 2015-01-27 at 19:23 +0200, Jarkko Sakkinen wrote:

On Tue, 2015-01-27 at 10:03 -0700, Jason Gunthorpe wrote:

quoted

On Tue, Jan 27, 2015 at 06:57:22PM +0200, Jarkko Sakkinen wrote:

quoted

quoted

quoted

+	/* TPM 1.2 requires self-test on resume. */
+	if (!(chip->flags & TPM_CHIP_FLAG_TPM2)) {
+		ret = tpm_do_selftest(chip);
+		if (ret < 0)
+			return ret;

Just to note, the return value from tpm_do_selftest() on TPM 1.2 chips was 
previously ignored. Mine does return 0.

Right. I can update the patch to ignore return value if the majority
wants that.

What happens to the system when pnp_driver.resume() returns failure?

Should tpm ever report failure on resume to the rest of the kernel?

Shouldn't this stuff be in tpm_pm_resume common code anyhow?

I think it should but not in the scope of this bug fix IMHO.

This may sound stupid but maybe I should not handle the return value of
tpm_do_selftest() with the same reasoning (not in the scope of this fix)
because it modifies semantics and my fix only fixes TPM 2.0 stuff.

I could leave a comment there that this return value is not handle as a
remainder.

quoted

Jason

/Jarkko