Skip to contentSkip to conversation

tpm: Autodetect itpm devices

PATCH In review Matthew Garrett 7 participants 11 messages Oct 21, 2010 – Nov 29, 2010 TPM device driver

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 tpm: Autodetect itpm devices

Some Lenovos have TPMs that require a quirk to function correctly. This can be autodetected by checking whether the device has a _HID of INTC0102. This is an invalid PNPid, and as such is discarded by the pnp layer - however it's still present in the ACPI code, so we can pull it out that way. This means that the quirk won't be automatically applied on non-ACPI systems, but without ACPI we don't have any way to identify the chip anyway so I don't think that's a great concern.

---
 drivers/char/tpm/tpm_tis.c |   24 ++++++++++++++++++++++++
 1 files changed, 24 insertions(+), 0 deletions(-)
drivers/char/tpm/tpm_tis.c +24 -0
index 1030f84..c17a305 100644
@@ -25,6 +25,7 @@
 #include <linux/slab.h>
 #include <linux/interrupt.h>
 #include <linux/wait.h>
+#include <linux/acpi.h>
 #include "tpm.h"
 
 #define TPM_HEADER_SIZE 10
@@ -78,6 +79,26 @@ enum tis_defaults {
 static LIST_HEAD(tis_chips);
 static DEFINE_SPINLOCK(tis_lock);
 
+#ifdef CONFIG_ACPI
+static int is_itpm(struct pnp_dev *dev)
+{
+	struct acpi_device *acpi = pnp_acpi_device(dev);
+	struct acpi_hardware_id *id;
+
+	list_for_each_entry(id, &acpi->pnp.ids, list) {
+		if (!strcmp("INTC0102", id->id))
+			return 1;
+	}
+
+	return 0;
+}
+#else
+static int is_itpm(struct pnp_dev *dev)
+{
+	return 0;
+}
+#endif
+
 static int check_locality(struct tpm_chip *chip, int l)
 {
 	if ((ioread8(chip->vendor.iobase + TPM_ACCESS(l)) &
@@ -472,6 +493,9 @@ static int tpm_tis_init(struct device *dev, resource_size_t start,
 		 "1.2 TPM (device-id 0x%X, rev-id %d)\n",
 		 vendor >> 16, ioread8(chip->vendor.iobase + TPM_RID(0)));
 
+	if (is_itpm(to_pnp_dev(dev)))
+		itpm = 1;
+
 	if (itpm)
 		dev_info(dev, "Intel iTPM workaround enabled\n");
 
Signature · 1 line
1.7.3.1
Acked-by on v1 1/1
#1 lore raw
Quoted 9 lines · Matthew Garrett

Some Lenovos have TPMs that require a quirk to function correctly. This can be autodetected by checking whether the device has a _HID of INTC0102. This is an invalid PNPid, and as such is discarded by the pnp layer - however it's still present in the ACPI code, so we can pull it out that way. This means that the quirk won't be automatically applied on non-ACPI systems, but without ACPI we don't have any way to identify the chip anyway so I don't think that's a great concern.

Thanks Matthew,

I'd be glad to remove the module option after this patch, but the non-ACPI case wouldn't remain covered indeed.

#2 lore raw

[2.6.37-rc1 REGRESSION] suspend fails on x200 (was Re: [PATCH] tpm: Autodetect itpm devices)

In reply to Matthew Garrett
Quoted 7 lines · Matthew Garrett

Some Lenovos have TPMs that require a quirk to function correctly. This can be autodetected by checking whether the device has a _HID of INTC0102. This is an invalid PNPid, and as such is discarded by the pnp layer - however it's still present in the ACPI code, so we can pull it out that way. This means that the quirk won't be automatically applied on non-ACPI systems, but without ACPI we don't have any way to identify the chip anyway so I don't think that's a great concern.

Matthew,

Thanks for coming up with a clean way to integrate this iTPM support!

Unfortunately, automagically loading tpm_tis results in a suspend regression on my X200:

(hand-transcribed, sorry for any typos)

[   16.759044] tpm_tis 00:0a: 1.2 TPM (device-id 0x1020, rev-id 6)
[ 7537.636134] PM: Entering mem sleep
[ 7537.636221] Suspending console(s) (use no_console_suspend to debug)
[ 7537.636668] sd 0:0:0:0: [sda] Synchronizing SCSI cache
[ 7537.658111] tpm_tis 00:0a: tpm_transmit: tpm_send: error -5
[ 7537.658119] legacy_suspend(): pnp_bus_suspend+0x0/0x6a returns -5
[ 7537.658122] PM: Device 00:0a failed to suspend: error -5
[ 7537.681940] sd 0:0:0:0: [sda] Stopping disk
[ 7538.564024] PM: Some devices failed to suspend

I suspect that it would have failed previously if I'd forced the module to load, but since tpm_tis now loads automatically as of 2.6.37-rc1, I have to "rmmod tpm_tis" for suspend to work.

-andy

#3 lore raw

Re: [2.6.37-rc1 REGRESSION] suspend fails on x200 (was Re: [PATCH] tpm: Autodetect itpm devices)

Quoted 3 lines · Andy Isaacson

I suspect that it would have failed previously if I'd forced the module to load, but since tpm_tis now loads automatically as of 2.6.37-rc1, I have to "rmmod tpm_tis" for suspend to work.

Hm. That's what this patch was intended to fix. Can you check that the tree you're using has it merged? The autoloading comes from the fix to the pnp code to use _CID even if _HID is malformed, so now we have a pnp device when we previously didn't.

Signature · 1 line
Matthew Garrett | [email protected]
#4 lore raw

Re: suspend fails on x200 (was Re: [PATCH] tpm: Autodetect itpm devices)

Quoted 9 lines · Matthew Garrett
Quoted 3 lines · Andy Isaacson

I suspect that it would have failed previously if I'd forced the module to load, but since tpm_tis now loads automatically as of 2.6.37-rc1, I have to "rmmod tpm_tis" for suspend to work.

Hm. That's what this patch was intended to fix. Can you check that the tree you're using has it merged? The autoloading comes from the fix to the pnp code to use _CID even if _HID is malformed, so now we have a pnp device when we previously didn't.

Hmmm, I am testing with straight-up .37-rc1, c8ddb2713c62. It does not include your patch.

This time around, still running -rc1, I ran pm-suspend and it suspended
+ resumed fine. The second time I ran pm-suspend, it switched to text console and started flashing the moon LED, but hung there with the LCD backlight still on and the LED moon flashing.

So, definitely something exciting going on... and not a simple suspend regression, and definitely not caused by this patch, my apologies. Subject line updated accordingly.

I'll try applying "Autodetect itpm devices" and report back.

-andy

#5 lore raw

Re: suspend fails on x200 (was Re: [PATCH] tpm: Autodetect itpm devices)

Quoted 19 lines · Andy Isaacson
Quoted 9 lines · Matthew Garrett
Quoted 3 lines · Andy Isaacson

I suspect that it would have failed previously if I'd forced the module to load, but since tpm_tis now loads automatically as of 2.6.37-rc1, I have to "rmmod tpm_tis" for suspend to work.

Hm. That's what this patch was intended to fix. Can you check that the tree you're using has it merged? The autoloading comes from the fix to the pnp code to use _CID even if _HID is malformed, so now we have a pnp device when we previously didn't.

Hmmm, I am testing with straight-up .37-rc1, c8ddb2713c62. It does not include your patch.

This time around, still running -rc1, I ran pm-suspend and it suspended
+ resumed fine. The second time I ran pm-suspend, it switched to text console and started flashing the moon LED, but hung there with the LCD backlight still on and the LED moon flashing.

Quoted 5 lines

So, definitely something exciting going on... and not a simple suspend regression, and definitely not caused by this patch, my apologies. Subject line updated accordingly.

I'll try applying "Autodetect itpm devices" and report back.

With your patch applied on top of -rc1, suspend + resume seems to work fine, and I don't have any error messages from tpm in dmesg.

Thanks, problem solved. (Still no idea what caused that one hang-with-blinking-moon, but it doesn't appear to be connected.)

-andy

#6 lore raw

Re: suspend fails on x200 (was Re: [PATCH] tpm: Autodetect itpm devices)

Quoted 24 lines · Andy Isaacson
Quoted 23 lines · Andy Isaacson
Quoted 9 lines · Matthew Garrett
Quoted 3 lines · Andy Isaacson

I suspect that it would have failed previously if I'd forced the module to load, but since tpm_tis now loads automatically as of 2.6.37-rc1, I have to "rmmod tpm_tis" for suspend to work.

Hm. That's what this patch was intended to fix. Can you check that the tree you're using has it merged? The autoloading comes from the fix to the pnp code to use _CID even if _HID is malformed, so now we have a pnp device when we previously didn't.

Hmmm, I am testing with straight-up .37-rc1, c8ddb2713c62. It does not include your patch.

This time around, still running -rc1, I ran pm-suspend and it suspended
+ resumed fine. The second time I ran pm-suspend, it switched to text console and started flashing the moon LED, but hung there with the LCD backlight still on and the LED moon flashing.

So, definitely something exciting going on... and not a simple suspend regression, and definitely not caused by this patch, my apologies. Subject line updated accordingly.

Quoted 7 lines
Quoted 1 line

I'll try applying "Autodetect itpm devices" and report back.

With your patch applied on top of -rc1, suspend + resume seems to work fine, and I don't have any error messages from tpm in dmesg.

Thanks, problem solved. (Still no idea what caused that one hang-with-blinking-moon, but it doesn't appear to be connected.)

Please give us a pointer to the patch, there are other users experiencing this issue.

Thanks,
Rafael

#7 lore raw

Re: suspend fails on x200 (was Re: [PATCH] tpm: Autodetect itpm devices)

Quoted 6 lines · Rafael J. Wysocki
Quoted 6 lines
Quoted 5 lines

So, definitely something exciting going on... and not a simple suspend regression, and definitely not caused by this patch, my apologies. Subject line updated accordingly.

I'll try applying "Autodetect itpm devices" and report back.

Quoted 8 lines
Quoted 5 lines

With your patch applied on top of -rc1, suspend + resume seems to work fine, and I don't have any error messages from tpm in dmesg.

Thanks, problem solved. (Still no idea what caused that one hang-with-blinking-moon, but it doesn't appear to be connected.)

Please give us a pointer to the patch, there are other users experiencing this issue.

My iTPM-causes-suspend-to-fail is fixed by the head message in this thread, <[email protected]>,
http://lkml.org/lkml/2010/10/21/456

-andy

Tested-by on v1 1/1
#8 lore raw
In reply to Matthew Garrett

[ added James Morris and Rafael to CC ]

Quoted 9 lines · Matthew Garrett

Some Lenovos have TPMs that require a quirk to function correctly. This can be autodetected by checking whether the device has a _HID of INTC0102. This is an invalid PNPid, and as such is discarded by the pnp layer - however it's still present in the ACPI code, so we can pull it out that way. This means that the quirk won't be automatically applied on non-ACPI systems, but without ACPI we don't have any way to identify the chip anyway so I don't think that's a great concern.

This definitely is .37 material still, as it fixes suspend regression on various machines, as reported here:

	http://lkml.org/lkml/2010/11/1/493
	http://lkml.org/lkml/2010/11/29/219

James, any word on this? It doesn't seem to be present either in Linus' tree, nor in linux-next.

Thanks.

Quoted 54 lines
---
 drivers/char/tpm/tpm_tis.c |   24 ++++++++++++++++++++++++
 1 files changed, 24 insertions(+), 0 deletions(-)
drivers/char/tpm/tpm_tis.c +24 -0
index 1030f84..c17a305 100644
@@ -25,6 +25,7 @@
 #include <linux/slab.h>
 #include <linux/interrupt.h>
 #include <linux/wait.h>
+#include <linux/acpi.h>
 #include "tpm.h"
 
 #define TPM_HEADER_SIZE 10
@@ -78,6 +79,26 @@ enum tis_defaults {
 static LIST_HEAD(tis_chips);
 static DEFINE_SPINLOCK(tis_lock);
 
+#ifdef CONFIG_ACPI
+static int is_itpm(struct pnp_dev *dev)
+{
+	struct acpi_device *acpi = pnp_acpi_device(dev);
+	struct acpi_hardware_id *id;
+
+	list_for_each_entry(id, &acpi->pnp.ids, list) {
+		if (!strcmp("INTC0102", id->id))
+			return 1;
+	}
+
+	return 0;
+}
+#else
+static int is_itpm(struct pnp_dev *dev)
+{
+	return 0;
+}
+#endif
+
 static int check_locality(struct tpm_chip *chip, int l)
 {
 	if ((ioread8(chip->vendor.iobase + TPM_ACCESS(l)) &
@@ -472,6 +493,9 @@ static int tpm_tis_init(struct device *dev, resource_size_t start,
 		 "1.2 TPM (device-id 0x%X, rev-id %d)\n",
 		 vendor >> 16, ioread8(chip->vendor.iobase + TPM_RID(0)));
 
+	if (is_itpm(to_pnp_dev(dev)))
+		itpm = 1;
+
 	if (itpm)
 		dev_info(dev, "Intel iTPM workaround enabled\n");
 
Signature · 2 lines
Jiri Kosina
SUSE Labs, Novell Inc.
#9 lore raw
Quoted 4 lines · Jiri Kosina
Quoted 1 line
Quoted 8 lines

This definitely is .37 material still, as it fixes suspend regression on various machines, as reported here:

	http://lkml.org/lkml/2010/11/1/493
	http://lkml.org/lkml/2010/11/29/219

James, any word on this? It doesn't seem to be present either in Linus' tree, nor in linux-next.

I'm pretty sure this is the first I've seen of this, which would explain why it's not in my tree (Rajiv: as mainatiner please collect and forward TPM patches to me, or at least cc me when acking them).

I can push this to Linus.

Quoted 62 lines

Thanks.

Quoted 54 lines
---
 drivers/char/tpm/tpm_tis.c |   24 ++++++++++++++++++++++++
 1 files changed, 24 insertions(+), 0 deletions(-)
drivers/char/tpm/tpm_tis.c +24 -0
index 1030f84..c17a305 100644
@@ -25,6 +25,7 @@
 #include <linux/slab.h>
 #include <linux/interrupt.h>
 #include <linux/wait.h>
+#include <linux/acpi.h>
 #include "tpm.h"
 
 #define TPM_HEADER_SIZE 10
@@ -78,6 +79,26 @@ enum tis_defaults {
 static LIST_HEAD(tis_chips);
 static DEFINE_SPINLOCK(tis_lock);
 
+#ifdef CONFIG_ACPI
+static int is_itpm(struct pnp_dev *dev)
+{
+	struct acpi_device *acpi = pnp_acpi_device(dev);
+	struct acpi_hardware_id *id;
+
+	list_for_each_entry(id, &acpi->pnp.ids, list) {
+		if (!strcmp("INTC0102", id->id))
+			return 1;
+	}
+
+	return 0;
+}
+#else
+static int is_itpm(struct pnp_dev *dev)
+{
+	return 0;
+}
+#endif
+
 static int check_locality(struct tpm_chip *chip, int l)
 {
 	if ((ioread8(chip->vendor.iobase + TPM_ACCESS(l)) &
@@ -472,6 +493,9 @@ static int tpm_tis_init(struct device *dev, resource_size_t start,
 		 "1.2 TPM (device-id 0x%X, rev-id %d)\n",
 		 vendor >> 16, ioread8(chip->vendor.iobase + TPM_RID(0)));
 
+	if (is_itpm(to_pnp_dev(dev)))
+		itpm = 1;
+
 	if (itpm)
 		dev_info(dev, "Intel iTPM workaround enabled\n");
 
Signature · 2 lines
Jiri Kosina
SUSE Labs, Novell Inc.
Signature · 2 lines
James Morris
<[email protected]>
#10 lore raw

Em 29-11-2010 20:08, James Morris escreveu:

Quoted 7 lines · Jiri Kosina
Quoted 6 lines
Quoted 1 line

This definitely is .37 material still, as it fixes suspend regression on various machines, as reported here:

Quoted 9 lines
Quoted 5 lines
	http://lkml.org/lkml/2010/11/1/493
	http://lkml.org/lkml/2010/11/29/219

James, any word on this? It doesn't seem to be present either in Linus' tree, nor in linux-next.

I'm pretty sure this is the first I've seen of this, which would explain why it's not in my tree (Rajiv: as mainatiner please collect and forward TPM patches to me, or at least cc me when acking them).

I normally CC you when acking, don't know what happened this time, sorry.

Rajiv