Skip to contentSkip to conversation

tpm: Add support for Atmel I2C TPMs

PATCH In review Jason Gunthorpe 3 participants 7 messages Sep 11, 2013 – Oct 1, 2013 3 messages not in the archive

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 tpm: Add support for Atmel I2C TPMs

This is based on the work of Teddy Reed <[email protected]> published on GitHub:
https://github.com/theopolis/tpm-i2c-atmel.git
34894b988b67e0ae55088d6388e77b0dbf10c07d

That driver was never merged, I have taken it as a starting point, forward ported, tested and revised the driver:
- Make it broadly textually similar to the Infineon and Nuvoton I2C driver
- Place everything in a format suitable for mainline inclusion
- Use high level I2C functions i2c_smbus_write_i2c_block_data and i2c_master_recv for data xfer
- Use the timeout system from the core code, by faking out a status register
- checkpatch cleanups
- Testing on ARM Kirkwood, with this device tree, using a AT97SC3204T-X1A180

        tpm@29 {
                compatible = "atmel,at97sc3204t";
                reg = <0x29>;
        };

PENDING: Teddy will need to provide a Signed-Off line for this driver, although very little of his original text remained unchanged, the work to discover the protocol is entirely his and he will need to certify the Developer Certification of Origin that it is suitable for mainline inclusion.

[jgg: revised and tested]
---
 .../devicetree/bindings/i2c/trivial-devices.txt    |   1 +
 drivers/char/tpm/Kconfig                           |   9 +
 drivers/char/tpm/Makefile                          |   1 +
 drivers/char/tpm/tpm_i2c_atmel.c                   | 276 +++++++++++++++++++++
 4 files changed, 287 insertions(+)
 create mode 100644 drivers/char/tpm/tpm_i2c_atmel.c

This has been compile tested on 3.11 x86-64 and runtime tested on 3.7 ARM

I'm not yet 100% certain this driver works perfectly, for some reason my TPM will not create an EK, which is either due to a larger sent TPM messages becoming corrupt (combined with a TPM bug to not report an error code) or something wrong with the TPM itself.

There is already evidence that this TPM has non-compliant firmware, since several other commands do not work as the spec defines them to (eg something as simple as reading TPM_PERMANENT_FLAGS has problems).

I have chips with newer firmware, and will test with those soon...

Documentation/devicetree/bindings/i2c/trivial-devices.txt +1 -0
index c389465..de6ceb5 100644
@@ -15,6 +15,7 @@ adi,adt7461		+/-1C TDM Extended Temp Range I.C
 adt7461			+/-1C TDM Extended Temp Range I.C
 at,24c08		i2c serial eeprom  (24cxx)
 atmel,24c02		i2c serial eeprom  (24cxx)
+atmel,at97sc3204t       i2c trusted platform module (TPM)
 catalyst,24c32		i2c serial eeprom
 dallas,ds1307		64 x 8, Serial, I2C Real-Time Clock
 dallas,ds1338		I2C RTC with 56-Byte NV RAM
drivers/char/tpm/Kconfig +9 -0
index 5853712..171fa89 100644
@@ -33,6 +33,15 @@ config TCG_TIS
 	  from within Linux.  To compile this driver as a module, choose
 	  M here; the module will be called tpm_tis.
 
+config TCG_TIS_I2C_ATMEL
+	tristate "TPM Interface Specification 1.2 Interface (I2C - Atmel)"
+	depends on I2C
+	---help---
+	  If you have an Atmel I2C TPM security chip say Yes and it will be
+          accessible from within Linux.
+          To compile this driver as a module, choose M here; the module will
+	  be called tpm_tis_i2c_atmel.
+
 config TCG_TIS_I2C_INFINEON
 	tristate "TPM Interface Specification 1.2 Interface (I2C - Infineon)"
 	depends on I2C
drivers/char/tpm/Makefile +1 -0
index 468a6aa..c328aa9 100644
@@ -12,6 +12,7 @@ ifdef CONFIG_TCG_IBMVTPM
 endif
 endif
 obj-$(CONFIG_TCG_TIS) += tpm_tis.o
+obj-$(CONFIG_TCG_TIS_I2C_ATMEL) += tpm_i2c_atmel.o
 obj-$(CONFIG_TCG_TIS_I2C_INFINEON) += tpm_i2c_infineon.o
 obj-$(CONFIG_TCG_NUVOTON_I2C) += tpm_nuvoton_i2c.o
 obj-$(CONFIG_TCG_NSC) += tpm_nsc.o
drivers/char/tpm/tpm_i2c_atmel.c new file +276 -0

276 lines not on this page · raw message

Signature · 1 line
1.8.1.2
#1 lore raw

drivers/char/tpm/tpm_i2c_atmel.c line 61

Quoted 8 lines · Jason Gunthorpe

I'm not yet 100% certain this driver works perfectly, for some reason my TPM will not create an EK, which is either due to a larger sent TPM messages becoming corrupt (combined with a TPM bug to not report an error code) or something wrong with the TPM itself.

drivers/char/tpm/tpm_i2c_atmel.c new file +3 -0
+
+	status = i2c_smbus_write_i2c_block_data(client, buf[0], len - 1,
+						buf + 1);

Thankfully Peter pointed out that these calls only transfer 32 bytes, which causes the problem I observed. The TPM returns success with a corrupted response when you truncate the request, apparently.

Fixing it with:

+ status = i2c_master_send(client, buf, len);

Has the driver pass all my tests. I will post a v1 after a few days and collect any other comments.

Jason

maintainer
#2 lore raw

Hi Jason,

Am Donnerstag, 12. September 2013, 01:08:41 schrieb Jason Gunthorpe:

Quoted 18 lines · Jason Gunthorpe

drivers/char/tpm/tpm_i2c_atmel.c line 61

Quoted 8 lines

I'm not yet 100% certain this driver works perfectly, for some reason my TPM will not create an EK, which is either due to a larger sent TPM messages becoming corrupt (combined with a TPM bug to not report an error code) or something wrong with the TPM itself.

drivers/char/tpm/tpm_i2c_atmel.c new file +3 -0
+
+	status = i2c_smbus_write_i2c_block_data(client, buf[0], len - 1,
+						buf + 1);

Thankfully Peter pointed out that these calls only transfer 32 bytes, which causes the problem I observed. The TPM returns success with a corrupted response when you truncate the request, apparently.

Fixing it with:

(quoted diff) +1 -0
+	status = i2c_master_send(client, buf, len);
Quoted 2 lines

Has the driver pass all my tests. I will post a v1 after a few days and collect any other comments.

Yes this should probably do the thing ;)

Thanks for porting yet another driver.

I think I'll stage them (nuvoton, st and atmel) all in a temporary branch "testing-and-review" so it's easier for people to try them out. This branch might rebase and change all the time, but maybe it makes things easier (and patches can be squashed afterwards easily if needed).

But since you'll be sending a v2, I'll wait until the v2.

Thanks,
Peter

maintainer
#6 lore raw
In reply to Jason Gunthorpe

Am Mittwoch, 11. September 2013, 22:58:07 schrieb Jason Gunthorpe:

Quoted 22 lines

This is based on the work of Teddy Reed <[email protected]> published on GitHub:
https://github.com/theopolis/tpm-i2c-atmel.git
34894b988b67e0ae55088d6388e77b0dbf10c07d

That driver was never merged, I have taken it as a starting point, forward ported, tested and revised the driver:
- Make it broadly textually similar to the Infineon and Nuvoton I2C driver
- Place everything in a format suitable for mainline inclusion
- Use high level I2C functions i2c_smbus_write_i2c_block_data and i2c_master_recv for data xfer
- Use the timeout system from the core code, by faking out a status register
- checkpatch cleanups
- Testing on ARM Kirkwood, with this device tree, using a AT97SC3204T-X1A180

        tpm@29 {
                compatible = "atmel,at97sc3204t";
                reg = <0x29>;
        };
Quoted 8 lines

PENDING: Teddy will need to provide a Signed-Off line for this driver, although very little of his original text remained unchanged, the work to discover the protocol is entirely his and he will need to certify the Developer Certification of Origin that it is suitable for mainline inclusion.

[jgg: revised and tested]

Hi Jason,

thanks for porting yet another driver.
I think I'll stage them (nuvoton, st and atmel) all in a temporary branch "testing-and-review" so it's easier for people to try them out.

Thanks,
Peter

maintainer
#3 lore raw
In reply to Peter Hüwe

Hi,

Quoted 7 lines

I think I'll stage them (nuvoton, st and atmel) all in a temporary branch "testing-and-review" so it's easier for people to try them out. This branch might rebase and change all the time, but maybe it makes things easier (and patches can be squashed afterwards easily if needed).

But since you'll be sending a v2, I'll wait until the v2.

I staged everything here for the moment
https://github.com/PeterHuewe/linux-tpmdd/tree/testing-and-review

All three are not ready yet though.
Will rebase/update when the new versions come up.

Thanks,
Peter

#5 lore raw
In reply to a message not in the archive
Quoted 9 lines · Peter H?we

Hi,

Quoted 6 lines

I think I'll stage them (nuvoton, st and atmel) all in a temporary branch "testing-and-review" so it's easier for people to try them out. This branch might rebase and change all the time, but maybe it makes things easier (and patches can be squashed afterwards easily if needed).

Quoted 3 lines
Quoted 1 line

But since you'll be sending a v2, I'll wait until the v2.

I staged everything here for the moment
https://github.com/PeterHuewe/linux-tpmdd/tree/testing-and-review

I put up a branch on my github that has the drivers as well as parts of a major cleanup as we both vaugely discussed.

Not finished yet, but maybe by the time you get back. It is big, I'm just going to keep it on github for the moment.

commit dfd68b37beb8806b0a061af71436017d0b6c577e

Author: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
Date:   Sat Sep 14 19:21:12 2013 -0600

tpm: Pull everything related to /dev/tpmX into tpm-dev.c

This cleans up all the identical file_operations sprinkled about in the drivers.

To support this the way the chip structure is recovered in tpm_open is switched to use container_of rather than searching, otherwise this is just a copy of function content from tpm.c to tpm-dev.c and the resulting cleanup from making the functions static.

commit 9d83865080be4fce23c3247754c640aacd30c25e

Author: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
Date:   Sat Sep 14 18:51:54 2013 -0600

tpm: Store devname in the tpm_chip

Just put the memory directly in the chip structure, rather than a 2nd dedicated kalloc.

commit 4b05342eb38db16afb5698dd7b79ce21b56e669b

Author: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
Date:   Sat Sep 14 18:47:43 2013 -0600

tpm: Create a tpm_class_ops structure and use it

This replaces the static initialization of a tpm_vendor_specific structure in the drivers with the standard Linux idiom of providing a const structure of function pointers.

As the first clean up step this just changes the names and copies ops into chip->vendor exactly as before.

commit 4e1a392500e4f588ec685703d0c1066c93135eaa

Author: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
Date:   Sat Sep 14 17:55:30 2013 -0600

tpm: xen-tpmfront: Fix default durations

All the default durations were being set to 10 minutes which is way too long for the timeouts. Normal values for the longest duration are around 5 mins, and short duration ar around .5s.

Further, these are just the default, tpm_get_timeouts will set them to values from the TPM (or throw an error).

Just remove them.

commit 181416f763b9adf2cf8f9cf0ca4d377a16b255d7

Author: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
Date:   Sat Sep 14 17:36:29 2013 -0600

tpm atmel: Call request_region with the correct base

Commit e0dd03caf20d040a0a86b6bd74028ec9bda545f5 changed the code path here so that ateml_get_base_addr no longer directly altered the tpm_vendor_specific structure, and instead placed the base address on the stack.

The commit missed updating the request_region call, which would have resulted in request_region being called with 0 as the base address.

I don't know if request_region(0, ..) will fail, if so the driver has been broken since 2006 and we should remove it from the tree as it has no users.

commit c41042057dc3fd156c8cb264a18dac1ce2bdf90a

Author: Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
Date:   Sat Sep 14 16:57:58 2013 -0600

tpm: ibmvtpm: Use %zd formatting for size_t format arguments

This suppresses compile warnings on 32 bit builds.

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#9 lore raw
In reply to a message not in the archive
Quoted 6 lines · Teddy Reed

Looks great, thanks again Jason!
Also, if anyone would like some TSSOP AT97SC3204T's, please email me and I will send you a few along with breakouts and programmable crystal so you can test/play/break/etc, no charge, but I'd rather only ship within the USA.

Thanks Teddy,

I've updated the series with the revised commit message:
https://github.com/jgunthorpe/linux/commit/4378757fbf7771ffecc3b8779856dcc938f45234

tpm: Add support for Atmel I2C TPMs
This is based on the work of Teddy Reed <[email protected]> published on GitHub:
https://github.com/theopolis/tpm-i2c-atmel.git
34894b9

That driver was never merged, I have taken it as a starting port, forward ported, tested and revised the driver:
- Make it broadly textually similar to the Infineon and Nuvoton I2C driver
- Place everything in a format suitable for mainline inclusion
- Use high level I2C functions i2c_master_send and i2c_master_recv for data xfer
- Use the timeout system from the core code, by faking out a status register
- Only I2C transfer the number of bytes in the reply, not a fixed message size.
- checkpatch cleanups
- Testing on ARM Kirkwood, with this device tree, using a AT97SC3204T-X1A180

        tpm@29 {
                compatible = "atmel,at97sc3204t";
                reg = <0x29>;
        };

Jason
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html