diff mbox series

[v7,4/6] tpm: tpm_tis_spi: Support cr50 devices

Message ID 20190920183240.181420-5-swboyd@chromium.org (mailing list archive)
State New, archived
Headers show
Series tpm: Add driver for cr50 | expand

Commit Message

Stephen Boyd Sept. 20, 2019, 6:32 p.m. UTC
From: Andrey Pronin <apronin@chromium.org>

Add TPM2.0 PTP FIFO compatible SPI interface for chips with Cr50
firmware. The firmware running on the currently supported H1 Secure
Microcontroller requires a special driver to handle its specifics:

 - need to ensure a certain delay between SPI transactions, or else
   the chip may miss some part of the next transaction
 - if there is no SPI activity for some time, it may go to sleep,
   and needs to be waken up before sending further commands
 - access to vendor-specific registers

Cr50 firmware has a requirement to wait for the TPM to wakeup before
sending commands over the SPI bus. Otherwise, the firmware could be in
deep sleep and not respond. The method to wait for the device to wakeup
is slightly different than the usual flow control mechanism described in
the TCG SPI spec. Add a completion to tpm_tis_spi_transfer() before we
start a SPI transfer so we can keep track of the last time the TPM
driver accessed the SPI bus to support the flow control mechanism.

Split the cr50 logic off into a different file to keep it out of the
normal code flow of the existing SPI driver while making it all part of
the same module when the code is optionally compiled into the same
module. Export a new function, tpm_tis_spi_init(), and the associated
read/write/transfer APIs so that we can do this. Make the cr50 code wrap
the tpm_tis_spi_phy struct with its own struct to override the behavior
of tpm_tis_spi_transfer() by supplying a custom flow control hook. This
shares the most code between the core driver and the cr50 support
without combining everything into the core driver or exporting module
symbols.

Signed-off-by: Andrey Pronin <apronin@chromium.org>
Cc: Andrey Pronin <apronin@chromium.org>
Cc: Duncan Laurie <dlaurie@chromium.org>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Guenter Roeck <groeck@chromium.org>
Cc: Alexander Steffen <Alexander.Steffen@infineon.com>
Cc: Heiko Stuebner <heiko@sntech.de>
[swboyd@chromium.org: Replace boilerplate with SPDX tag, drop
suspended bit and remove ifdef checks in cr50.h, migrate to functions
exported in tpm_tis_spi.h, combine into one module instead of two]
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
---
 drivers/char/tpm/Kconfig            |   7 +
 drivers/char/tpm/Makefile           |   4 +-
 drivers/char/tpm/tpm_tis_spi.c      |  78 ++++---
 drivers/char/tpm/tpm_tis_spi.h      |  53 +++++
 drivers/char/tpm/tpm_tis_spi_cr50.c | 321 ++++++++++++++++++++++++++++
 5 files changed, 431 insertions(+), 32 deletions(-)
 create mode 100644 drivers/char/tpm/tpm_tis_spi.h
 create mode 100644 drivers/char/tpm/tpm_tis_spi_cr50.c

Comments

Jarkko Sakkinen Oct. 6, 2019, 10:32 p.m. UTC | #1
On Fri, Sep 20, 2019 at 11:32:38AM -0700, Stephen Boyd wrote:
> From: Andrey Pronin <apronin@chromium.org>
> 
> Add TPM2.0 PTP FIFO compatible SPI interface for chips with Cr50
> firmware. The firmware running on the currently supported H1 Secure
> Microcontroller requires a special driver to handle its specifics:
> 
>  - need to ensure a certain delay between SPI transactions, or else
>    the chip may miss some part of the next transaction
>  - if there is no SPI activity for some time, it may go to sleep,
>    and needs to be waken up before sending further commands
>  - access to vendor-specific registers
> 
> Cr50 firmware has a requirement to wait for the TPM to wakeup before
> sending commands over the SPI bus. Otherwise, the firmware could be in
> deep sleep and not respond. The method to wait for the device to wakeup
> is slightly different than the usual flow control mechanism described in
> the TCG SPI spec. Add a completion to tpm_tis_spi_transfer() before we
> start a SPI transfer so we can keep track of the last time the TPM
> driver accessed the SPI bus to support the flow control mechanism.
> 
> Split the cr50 logic off into a different file to keep it out of the
> normal code flow of the existing SPI driver while making it all part of
> the same module when the code is optionally compiled into the same
> module. Export a new function, tpm_tis_spi_init(), and the associated
> read/write/transfer APIs so that we can do this. Make the cr50 code wrap
> the tpm_tis_spi_phy struct with its own struct to override the behavior
> of tpm_tis_spi_transfer() by supplying a custom flow control hook. This
> shares the most code between the core driver and the cr50 support
> without combining everything into the core driver or exporting module
> symbols.
> 
> Signed-off-by: Andrey Pronin <apronin@chromium.org>
> Cc: Andrey Pronin <apronin@chromium.org>
> Cc: Duncan Laurie <dlaurie@chromium.org>
> Cc: Jason Gunthorpe <jgg@ziepe.ca>
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: Guenter Roeck <groeck@chromium.org>
> Cc: Alexander Steffen <Alexander.Steffen@infineon.com>
> Cc: Heiko Stuebner <heiko@sntech.de>

First, I apologize for such a long latency (two weeks).

I think this looks legit now.

Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>

/Jarkko
Alexander Steffen Feb. 3, 2020, 9:13 a.m. UTC | #2
On 20.09.2019 20:32, Stephen Boyd wrote:
> diff --git a/drivers/char/tpm/Makefile b/drivers/char/tpm/Makefile
> index a01c4cab902a..c96439f11c85 100644
> --- a/drivers/char/tpm/Makefile
> +++ b/drivers/char/tpm/Makefile
> @@ -21,7 +21,9 @@ tpm-$(CONFIG_EFI) += eventlog/efi.o
>   tpm-$(CONFIG_OF) += eventlog/of.o
>   obj-$(CONFIG_TCG_TIS_CORE) += tpm_tis_core.o
>   obj-$(CONFIG_TCG_TIS) += tpm_tis.o
> -obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi.o
> +obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi_mod.o
> +tpm_tis_spi_mod-y := tpm_tis_spi.o
> +tpm_tis_spi_mod-$(CONFIG_TCG_TIS_SPI_CR50) += tpm_tis_spi_cr50.o
>   obj-$(CONFIG_TCG_TIS_I2C_ATMEL) += tpm_i2c_atmel.o
>   obj-$(CONFIG_TCG_TIS_I2C_INFINEON) += tpm_i2c_infineon.o
>   obj-$(CONFIG_TCG_TIS_I2C_NUVOTON) += tpm_i2c_nuvoton.o

This renames the driver module from tpm_tis_spi to tpm_tis_spi_mod, was 
this done intentionally? When trying to upgrade the kernel, this just 
broke my test system, since all scripts expect to be able to load 
tpm_tis_spi, which does not exist anymore with that change.

Alexander
Stephen Boyd Feb. 4, 2020, 12:37 a.m. UTC | #3
Quoting Alexander Steffen (2020-02-03 01:13:29)
> On 20.09.2019 20:32, Stephen Boyd wrote:
> > diff --git a/drivers/char/tpm/Makefile b/drivers/char/tpm/Makefile
> > index a01c4cab902a..c96439f11c85 100644
> > --- a/drivers/char/tpm/Makefile
> > +++ b/drivers/char/tpm/Makefile
> > @@ -21,7 +21,9 @@ tpm-$(CONFIG_EFI) += eventlog/efi.o
> >   tpm-$(CONFIG_OF) += eventlog/of.o
> >   obj-$(CONFIG_TCG_TIS_CORE) += tpm_tis_core.o
> >   obj-$(CONFIG_TCG_TIS) += tpm_tis.o
> > -obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi.o
> > +obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi_mod.o
> > +tpm_tis_spi_mod-y := tpm_tis_spi.o
> > +tpm_tis_spi_mod-$(CONFIG_TCG_TIS_SPI_CR50) += tpm_tis_spi_cr50.o
> >   obj-$(CONFIG_TCG_TIS_I2C_ATMEL) += tpm_i2c_atmel.o
> >   obj-$(CONFIG_TCG_TIS_I2C_INFINEON) += tpm_i2c_infineon.o
> >   obj-$(CONFIG_TCG_TIS_I2C_NUVOTON) += tpm_i2c_nuvoton.o
> 
> This renames the driver module from tpm_tis_spi to tpm_tis_spi_mod, was 
> this done intentionally? When trying to upgrade the kernel, this just 
> broke my test system, since all scripts expect to be able to load 
> tpm_tis_spi, which does not exist anymore with that change.
> 

I mentioned this during the review of this patch set. I thought nobody
would care, given that it's just a module name.

Can your scripts load the module based on something besides the module
name? Perhaps by using device attributes instead?
Alexander Steffen Feb. 4, 2020, 7:15 a.m. UTC | #4
On 04.02.2020 01:37, Stephen Boyd wrote:
> Quoting Alexander Steffen (2020-02-03 01:13:29)
>> On 20.09.2019 20:32, Stephen Boyd wrote:
>>> diff --git a/drivers/char/tpm/Makefile b/drivers/char/tpm/Makefile
>>> index a01c4cab902a..c96439f11c85 100644
>>> --- a/drivers/char/tpm/Makefile
>>> +++ b/drivers/char/tpm/Makefile
>>> @@ -21,7 +21,9 @@ tpm-$(CONFIG_EFI) += eventlog/efi.o
>>>    tpm-$(CONFIG_OF) += eventlog/of.o
>>>    obj-$(CONFIG_TCG_TIS_CORE) += tpm_tis_core.o
>>>    obj-$(CONFIG_TCG_TIS) += tpm_tis.o
>>> -obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi.o
>>> +obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi_mod.o
>>> +tpm_tis_spi_mod-y := tpm_tis_spi.o
>>> +tpm_tis_spi_mod-$(CONFIG_TCG_TIS_SPI_CR50) += tpm_tis_spi_cr50.o
>>>    obj-$(CONFIG_TCG_TIS_I2C_ATMEL) += tpm_i2c_atmel.o
>>>    obj-$(CONFIG_TCG_TIS_I2C_INFINEON) += tpm_i2c_infineon.o
>>>    obj-$(CONFIG_TCG_TIS_I2C_NUVOTON) += tpm_i2c_nuvoton.o
>>
>> This renames the driver module from tpm_tis_spi to tpm_tis_spi_mod, was
>> this done intentionally? When trying to upgrade the kernel, this just
>> broke my test system, since all scripts expect to be able to load
>> tpm_tis_spi, which does not exist anymore with that change.
>>
> 
> I mentioned this during the review of this patch set. I thought nobody
> would care, given that it's just a module name.
> 
> Can your scripts load the module based on something besides the module
> name? Perhaps by using device attributes instead?

The scripts are effectively using modprobe/rmmod/etc. and those need the 
name. modprobe can be fixed by defining an alias, but this does not work 
for rmmod. Many other things also depend on the name, e.g. module 
blacklisting or the output of lsmod, where people might now get confused 
by the difference between "tpm_tis_spi_mod" and "tpm_tis_i2c". Also, 
there are many tutorials out there, that explicitly tell users to run 
something like "modprobe tpm_tis_spi", which won't work anymore now.

So, if there is a good reason to break compatibility, I'm fine with 
that. But in this case, isn't there some way to achieve the desired 
functionality without changing the name? Even if it is a little more 
complex than the three-line change above, that would probably be worth it.

Alexander
Jarkko Sakkinen Feb. 4, 2020, noon UTC | #5
On Tue, Feb 04, 2020 at 08:15:49AM +0100, Alexander Steffen wrote:
> The scripts are effectively using modprobe/rmmod/etc. and those need the
> name. modprobe can be fixed by defining an alias, but this does not work for
> rmmod. Many other things also depend on the name, e.g. module blacklisting
> or the output of lsmod, where people might now get confused by the
> difference between "tpm_tis_spi_mod" and "tpm_tis_i2c". Also, there are many
> tutorials out there, that explicitly tell users to run something like
> "modprobe tpm_tis_spi", which won't work anymore now.
> 
> So, if there is a good reason to break compatibility, I'm fine with that.
> But in this case, isn't there some way to achieve the desired functionality
> without changing the name? Even if it is a little more complex than the
> three-line change above, that would probably be worth it.

Nope, you're right. I'll work out a fix for this and
add you as reported-by.

/Jarkko
diff mbox series

Patch

diff --git a/drivers/char/tpm/Kconfig b/drivers/char/tpm/Kconfig
index 88a3c06fc153..ecbffae14941 100644
--- a/drivers/char/tpm/Kconfig
+++ b/drivers/char/tpm/Kconfig
@@ -67,6 +67,13 @@  config TCG_TIS_SPI
 	  within Linux. To compile this driver as a module, choose  M here;
 	  the module will be called tpm_tis_spi.
 
+config TCG_TIS_SPI_CR50
+	bool "Cr50 SPI Interface"
+	depends on TCG_TIS_SPI
+	---help---
+	  If you have a H1 secure module running Cr50 firmware on SPI bus,
+	  say Yes and it will be accessible from within Linux.
+
 config TCG_TIS_I2C_ATMEL
 	tristate "TPM Interface Specification 1.2 Interface (I2C - Atmel)"
 	depends on I2C
diff --git a/drivers/char/tpm/Makefile b/drivers/char/tpm/Makefile
index a01c4cab902a..c96439f11c85 100644
--- a/drivers/char/tpm/Makefile
+++ b/drivers/char/tpm/Makefile
@@ -21,7 +21,9 @@  tpm-$(CONFIG_EFI) += eventlog/efi.o
 tpm-$(CONFIG_OF) += eventlog/of.o
 obj-$(CONFIG_TCG_TIS_CORE) += tpm_tis_core.o
 obj-$(CONFIG_TCG_TIS) += tpm_tis.o
-obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi.o
+obj-$(CONFIG_TCG_TIS_SPI) += tpm_tis_spi_mod.o
+tpm_tis_spi_mod-y := tpm_tis_spi.o
+tpm_tis_spi_mod-$(CONFIG_TCG_TIS_SPI_CR50) += tpm_tis_spi_cr50.o
 obj-$(CONFIG_TCG_TIS_I2C_ATMEL) += tpm_i2c_atmel.o
 obj-$(CONFIG_TCG_TIS_I2C_INFINEON) += tpm_i2c_infineon.o
 obj-$(CONFIG_TCG_TIS_I2C_NUVOTON) += tpm_i2c_nuvoton.o
diff --git a/drivers/char/tpm/tpm_tis_spi.c b/drivers/char/tpm/tpm_tis_spi.c
index b3ed85671dd8..5e4253e7c080 100644
--- a/drivers/char/tpm/tpm_tis_spi.c
+++ b/drivers/char/tpm/tpm_tis_spi.c
@@ -20,6 +20,7 @@ 
  * Dorn and Kyleen Hall and Jarko Sakkinnen.
  */
 
+#include <linux/completion.h>
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/moduleparam.h>
@@ -31,27 +32,16 @@ 
 
 #include <linux/spi/spi.h>
 #include <linux/gpio.h>
+#include <linux/of_device.h>
 #include <linux/of_irq.h>
 #include <linux/of_gpio.h>
 #include <linux/tpm.h>
 #include "tpm.h"
 #include "tpm_tis_core.h"
+#include "tpm_tis_spi.h"
 
 #define MAX_SPI_FRAMESIZE 64
 
-struct tpm_tis_spi_phy {
-	struct tpm_tis_data priv;
-	struct spi_device *spi_device;
-	int (*flow_control)(struct tpm_tis_spi_phy *phy,
-			    struct spi_transfer *xfer);
-	u8 *iobuf;
-};
-
-static inline struct tpm_tis_spi_phy *to_tpm_tis_spi_phy(struct tpm_tis_data *data)
-{
-	return container_of(data, struct tpm_tis_spi_phy, priv);
-}
-
 /*
  * TCG SPI flow control is documented in section 6.4 of the spec[1]. In short,
  * keep trying to read from the device until MISO goes high indicating the
@@ -87,8 +77,8 @@  static int tpm_tis_spi_flow_control(struct tpm_tis_spi_phy *phy,
 	return 0;
 }
 
-static int tpm_tis_spi_transfer(struct tpm_tis_data *data, u32 addr, u16 len,
-				u8 *in, const u8 *out)
+int tpm_tis_spi_transfer(struct tpm_tis_data *data, u32 addr, u16 len,
+			 u8 *in, const u8 *out)
 {
 	struct tpm_tis_spi_phy *phy = to_tpm_tis_spi_phy(data);
 	int ret = 0;
@@ -136,6 +126,7 @@  static int tpm_tis_spi_transfer(struct tpm_tis_data *data, u32 addr, u16 len,
 
 		spi_message_init(&m);
 		spi_message_add_tail(&spi_xfer, &m);
+		reinit_completion(&phy->ready);
 		ret = spi_sync_locked(phy->spi_device, &m);
 		if (ret < 0)
 			goto exit;
@@ -165,7 +156,7 @@  static int tpm_tis_spi_write_bytes(struct tpm_tis_data *data, u32 addr,
 	return tpm_tis_spi_transfer(data, addr, len, NULL, value);
 }
 
-static int tpm_tis_spi_read16(struct tpm_tis_data *data, u32 addr, u16 *result)
+int tpm_tis_spi_read16(struct tpm_tis_data *data, u32 addr, u16 *result)
 {
 	__le16 result_le;
 	int rc;
@@ -178,7 +169,7 @@  static int tpm_tis_spi_read16(struct tpm_tis_data *data, u32 addr, u16 *result)
 	return rc;
 }
 
-static int tpm_tis_spi_read32(struct tpm_tis_data *data, u32 addr, u32 *result)
+int tpm_tis_spi_read32(struct tpm_tis_data *data, u32 addr, u32 *result)
 {
 	__le32 result_le;
 	int rc;
@@ -191,7 +182,7 @@  static int tpm_tis_spi_read32(struct tpm_tis_data *data, u32 addr, u32 *result)
 	return rc;
 }
 
-static int tpm_tis_spi_write32(struct tpm_tis_data *data, u32 addr, u32 value)
+int tpm_tis_spi_write32(struct tpm_tis_data *data, u32 addr, u32 value)
 {
 	__le32 value_le;
 	int rc;
@@ -203,6 +194,18 @@  static int tpm_tis_spi_write32(struct tpm_tis_data *data, u32 addr, u32 value)
 	return rc;
 }
 
+int tpm_tis_spi_init(struct spi_device *spi, struct tpm_tis_spi_phy *phy,
+		     int irq, const struct tpm_tis_phy_ops *phy_ops)
+{
+	phy->iobuf = devm_kmalloc(&spi->dev, MAX_SPI_FRAMESIZE, GFP_KERNEL);
+	if (!phy->iobuf)
+		return -ENOMEM;
+
+	phy->spi_device = spi;
+
+	return tpm_tis_core_init(&spi->dev, &phy->priv, irq, phy_ops, NULL);
+}
+
 static const struct tpm_tis_phy_ops tpm_spi_phy_ops = {
 	.read_bytes = tpm_tis_spi_read_bytes,
 	.write_bytes = tpm_tis_spi_write_bytes,
@@ -221,11 +224,6 @@  static int tpm_tis_spi_probe(struct spi_device *dev)
 	if (!phy)
 		return -ENOMEM;
 
-	phy->spi_device = dev;
-
-	phy->iobuf = devm_kmalloc(&dev->dev, MAX_SPI_FRAMESIZE, GFP_KERNEL);
-	if (!phy->iobuf)
-		return -ENOMEM;
 	phy->flow_control = tpm_tis_spi_flow_control;
 
 	/* If the SPI device has an IRQ then use that */
@@ -234,11 +232,27 @@  static int tpm_tis_spi_probe(struct spi_device *dev)
 	else
 		irq = -1;
 
-	return tpm_tis_core_init(&dev->dev, &phy->priv, irq, &tpm_spi_phy_ops,
-				 NULL);
+	init_completion(&phy->ready);
+	return tpm_tis_spi_init(dev, phy, irq, &tpm_spi_phy_ops);
+}
+
+typedef int (*tpm_tis_spi_probe_func)(struct spi_device *);
+
+static int tpm_tis_spi_driver_probe(struct spi_device *spi)
+{
+	const struct spi_device_id *spi_dev_id = spi_get_device_id(spi);
+	tpm_tis_spi_probe_func probe_func;
+
+	probe_func = of_device_get_match_data(&spi->dev);
+	if (!probe_func && spi_dev_id)
+		probe_func = (tpm_tis_spi_probe_func)spi_dev_id->driver_data;
+	if (!probe_func)
+		return -ENODEV;
+
+	return probe_func(spi);
 }
 
-static SIMPLE_DEV_PM_OPS(tpm_tis_pm, tpm_pm_suspend, tpm_tis_resume);
+static SIMPLE_DEV_PM_OPS(tpm_tis_pm, tpm_pm_suspend, tpm_tis_spi_resume);
 
 static int tpm_tis_spi_remove(struct spi_device *dev)
 {
@@ -250,15 +264,17 @@  static int tpm_tis_spi_remove(struct spi_device *dev)
 }
 
 static const struct spi_device_id tpm_tis_spi_id[] = {
-	{"tpm_tis_spi", 0},
+	{ "tpm_tis_spi", (unsigned long)tpm_tis_spi_probe },
+	{ "cr50", (unsigned long)cr50_spi_probe },
 	{}
 };
 MODULE_DEVICE_TABLE(spi, tpm_tis_spi_id);
 
 static const struct of_device_id of_tis_spi_match[] = {
-	{ .compatible = "st,st33htpm-spi", },
-	{ .compatible = "infineon,slb9670", },
-	{ .compatible = "tcg,tpm_tis-spi", },
+	{ .compatible = "st,st33htpm-spi", .data = tpm_tis_spi_probe },
+	{ .compatible = "infineon,slb9670", .data = tpm_tis_spi_probe },
+	{ .compatible = "tcg,tpm_tis-spi", .data = tpm_tis_spi_probe },
+	{ .compatible = "google,cr50", .data = cr50_spi_probe },
 	{}
 };
 MODULE_DEVICE_TABLE(of, of_tis_spi_match);
@@ -277,7 +293,7 @@  static struct spi_driver tpm_tis_spi_driver = {
 		.of_match_table = of_match_ptr(of_tis_spi_match),
 		.acpi_match_table = ACPI_PTR(acpi_tis_spi_match),
 	},
-	.probe = tpm_tis_spi_probe,
+	.probe = tpm_tis_spi_driver_probe,
 	.remove = tpm_tis_spi_remove,
 	.id_table = tpm_tis_spi_id,
 };
diff --git a/drivers/char/tpm/tpm_tis_spi.h b/drivers/char/tpm/tpm_tis_spi.h
new file mode 100644
index 000000000000..bba73979c368
--- /dev/null
+++ b/drivers/char/tpm/tpm_tis_spi.h
@@ -0,0 +1,53 @@ 
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2015 Infineon Technologies AG
+ * Copyright (C) 2016 STMicroelectronics SAS
+ */
+
+#ifndef TPM_TIS_SPI_H
+#define TPM_TIS_SPI_H
+
+#include "tpm_tis_core.h"
+
+struct tpm_tis_spi_phy {
+	struct tpm_tis_data priv;
+	struct spi_device *spi_device;
+	int (*flow_control)(struct tpm_tis_spi_phy *phy,
+			     struct spi_transfer *xfer);
+	struct completion ready;
+	unsigned long wake_after;
+
+	u8 *iobuf;
+};
+
+static inline struct tpm_tis_spi_phy *to_tpm_tis_spi_phy(struct tpm_tis_data *data)
+{
+	return container_of(data, struct tpm_tis_spi_phy, priv);
+}
+
+extern int tpm_tis_spi_init(struct spi_device *spi, struct tpm_tis_spi_phy *phy,
+			    int irq, const struct tpm_tis_phy_ops *phy_ops);
+
+extern int tpm_tis_spi_transfer(struct tpm_tis_data *data, u32 addr, u16 len,
+				u8 *in, const u8 *out);
+
+extern int tpm_tis_spi_read16(struct tpm_tis_data *data, u32 addr, u16 *result);
+extern int tpm_tis_spi_read32(struct tpm_tis_data *data, u32 addr, u32 *result);
+extern int tpm_tis_spi_write32(struct tpm_tis_data *data, u32 addr, u32 value);
+
+#ifdef CONFIG_TCG_TIS_SPI_CR50
+extern int cr50_spi_probe(struct spi_device *spi);
+#else
+static inline int cr50_spi_probe(struct spi_device *spi)
+{
+	return -ENODEV;
+}
+#endif
+
+#if defined(CONFIG_PM_SLEEP) && defined(CONFIG_TCG_TIS_SPI_CR50)
+extern int tpm_tis_spi_resume(struct device *dev);
+#else
+#define tpm_tis_spi_resume	NULL
+#endif
+
+#endif
diff --git a/drivers/char/tpm/tpm_tis_spi_cr50.c b/drivers/char/tpm/tpm_tis_spi_cr50.c
new file mode 100644
index 000000000000..187a023c2556
--- /dev/null
+++ b/drivers/char/tpm/tpm_tis_spi_cr50.c
@@ -0,0 +1,321 @@ 
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2016 Google, Inc
+ *
+ * This device driver implements a TCG PTP FIFO interface over SPI for chips
+ * with Cr50 firmware.
+ * It is based on tpm_tis_spi driver by Peter Huewe and Christophe Ricard.
+ */
+
+#include <linux/completion.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/pm.h>
+#include <linux/spi/spi.h>
+#include <linux/wait.h>
+
+#include "tpm_tis_core.h"
+#include "tpm_tis_spi.h"
+
+/*
+ * Cr50 timing constants:
+ * - can go to sleep not earlier than after CR50_SLEEP_DELAY_MSEC.
+ * - needs up to CR50_WAKE_START_DELAY_USEC to wake after sleep.
+ * - requires waiting for "ready" IRQ, if supported; or waiting for at least
+ *   CR50_NOIRQ_ACCESS_DELAY_MSEC between transactions, if IRQ is not supported.
+ * - waits for up to CR50_FLOW_CONTROL for flow control 'ready' indication.
+ */
+#define CR50_SLEEP_DELAY_MSEC			1000
+#define CR50_WAKE_START_DELAY_USEC		1000
+#define CR50_NOIRQ_ACCESS_DELAY			msecs_to_jiffies(2)
+#define CR50_READY_IRQ_TIMEOUT			msecs_to_jiffies(TPM2_TIMEOUT_A)
+#define CR50_FLOW_CONTROL			msecs_to_jiffies(TPM2_TIMEOUT_A)
+#define MAX_IRQ_CONFIRMATION_ATTEMPTS		3
+
+#define TPM_CR50_FW_VER(l)			(0x0f90 | ((l) << 12))
+#define TPM_CR50_MAX_FW_VER_LEN			64
+
+struct cr50_spi_phy {
+	struct tpm_tis_spi_phy spi_phy;
+
+	struct mutex time_track_mutex;
+	unsigned long last_access;
+
+	unsigned long access_delay;
+
+	unsigned int irq_confirmation_attempt;
+	bool irq_needs_confirmation;
+	bool irq_confirmed;
+};
+
+static inline struct cr50_spi_phy *to_cr50_spi_phy(struct tpm_tis_spi_phy *phy)
+{
+	return container_of(phy, struct cr50_spi_phy, spi_phy);
+}
+
+/*
+ * The cr50 interrupt handler just signals waiting threads that the
+ * interrupt was asserted.  It does not do any processing triggered
+ * by interrupts but is instead used to avoid fixed delays.
+ */
+static irqreturn_t cr50_spi_irq_handler(int dummy, void *dev_id)
+{
+	struct cr50_spi_phy *cr50_phy = dev_id;
+
+	cr50_phy->irq_confirmed = true;
+	complete(&cr50_phy->spi_phy.ready);
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * Cr50 needs to have at least some delay between consecutive
+ * transactions. Make sure we wait.
+ */
+static void cr50_ensure_access_delay(struct cr50_spi_phy *phy)
+{
+	unsigned long allowed_access = phy->last_access + phy->access_delay;
+	unsigned long time_now = jiffies;
+	struct device *dev = &phy->spi_phy.spi_device->dev;
+
+	/*
+	 * Note: There is a small chance, if Cr50 is not accessed in a few days,
+	 * that time_in_range will not provide the correct result after the wrap
+	 * around for jiffies. In this case, we'll have an unneeded short delay,
+	 * which is fine.
+	 */
+	if (time_in_range_open(time_now, phy->last_access, allowed_access)) {
+		unsigned long remaining, timeout = allowed_access - time_now;
+
+		remaining = wait_for_completion_timeout(&phy->spi_phy.ready,
+							timeout);
+		if (!remaining && phy->irq_confirmed)
+			dev_warn(dev, "Timeout waiting for TPM ready IRQ\n");
+	}
+
+	if (phy->irq_needs_confirmation) {
+		unsigned int attempt = ++phy->irq_confirmation_attempt;
+
+		if (phy->irq_confirmed) {
+			phy->irq_needs_confirmation = false;
+			phy->access_delay = CR50_READY_IRQ_TIMEOUT;
+			dev_info(dev, "TPM ready IRQ confirmed on attempt %u\n",
+				 attempt);
+		} else if (attempt > MAX_IRQ_CONFIRMATION_ATTEMPTS) {
+			phy->irq_needs_confirmation = false;
+			dev_warn(dev, "IRQ not confirmed - will use delays\n");
+		}
+	}
+}
+
+/*
+ * Cr50 might go to sleep if there is no SPI activity for some time and
+ * miss the first few bits/bytes on the bus. In such case, wake it up
+ * by asserting CS and give it time to start up.
+ */
+static bool cr50_needs_waking(struct cr50_spi_phy *phy)
+{
+	/*
+	 * Note: There is a small chance, if Cr50 is not accessed in a few days,
+	 * that time_in_range will not provide the correct result after the wrap
+	 * around for jiffies. In this case, we'll probably timeout or read
+	 * incorrect value from TPM_STS and just retry the operation.
+	 */
+	return !time_in_range_open(jiffies, phy->last_access,
+				   phy->spi_phy.wake_after);
+}
+
+static void cr50_wake_if_needed(struct cr50_spi_phy *cr50_phy)
+{
+	struct tpm_tis_spi_phy *phy = &cr50_phy->spi_phy;
+
+	if (cr50_needs_waking(cr50_phy)) {
+		/* Assert CS, wait 1 msec, deassert CS */
+		struct spi_transfer spi_cs_wake = { .delay_usecs = 1000 };
+
+		spi_sync_transfer(phy->spi_device, &spi_cs_wake, 1);
+		/* Wait for it to fully wake */
+		usleep_range(CR50_WAKE_START_DELAY_USEC,
+			     CR50_WAKE_START_DELAY_USEC * 2);
+	}
+
+	/* Reset the time when we need to wake Cr50 again */
+	phy->wake_after = jiffies + msecs_to_jiffies(CR50_SLEEP_DELAY_MSEC);
+}
+
+/*
+ * Flow control: clock the bus and wait for cr50 to set LSB before
+ * sending/receiving data. TCG PTP spec allows it to happen during
+ * the last byte of header, but cr50 never does that in practice,
+ * and earlier versions had a bug when it was set too early, so don't
+ * check for it during header transfer.
+ */
+static int cr50_spi_flow_control(struct tpm_tis_spi_phy *phy,
+				 struct spi_transfer *spi_xfer)
+{
+	struct device *dev = &phy->spi_device->dev;
+	unsigned long timeout = jiffies + CR50_FLOW_CONTROL;
+	struct spi_message m;
+	int ret;
+
+	spi_xfer->len = 1;
+
+	do {
+		spi_message_init(&m);
+		spi_message_add_tail(spi_xfer, &m);
+		ret = spi_sync_locked(phy->spi_device, &m);
+		if (ret < 0)
+			return ret;
+
+		if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "Timeout during flow control\n");
+			return -EBUSY;
+		}
+	} while (!(phy->iobuf[0] & 0x01));
+
+	return 0;
+}
+
+static int tpm_tis_spi_cr50_transfer(struct tpm_tis_data *data, u32 addr, u16 len,
+				     u8 *in, const u8 *out)
+{
+	struct tpm_tis_spi_phy *phy = to_tpm_tis_spi_phy(data);
+	struct cr50_spi_phy *cr50_phy = to_cr50_spi_phy(phy);
+	int ret;
+
+	mutex_lock(&cr50_phy->time_track_mutex);
+	/*
+	 * Do this outside of spi_bus_lock in case cr50 is not the
+	 * only device on that spi bus.
+	 */
+	cr50_ensure_access_delay(cr50_phy);
+	cr50_wake_if_needed(cr50_phy);
+
+	ret = tpm_tis_spi_transfer(data, addr, len, in, out);
+
+	cr50_phy->last_access = jiffies;
+	mutex_unlock(&cr50_phy->time_track_mutex);
+
+	return ret;
+}
+
+static int tpm_tis_spi_cr50_read_bytes(struct tpm_tis_data *data, u32 addr,
+				       u16 len, u8 *result)
+{
+	return tpm_tis_spi_cr50_transfer(data, addr, len, result, NULL);
+}
+
+static int tpm_tis_spi_cr50_write_bytes(struct tpm_tis_data *data, u32 addr,
+					u16 len, const u8 *value)
+{
+	return tpm_tis_spi_cr50_transfer(data, addr, len, NULL, value);
+}
+
+static const struct tpm_tis_phy_ops tpm_spi_cr50_phy_ops = {
+	.read_bytes = tpm_tis_spi_cr50_read_bytes,
+	.write_bytes = tpm_tis_spi_cr50_write_bytes,
+	.read16 = tpm_tis_spi_read16,
+	.read32 = tpm_tis_spi_read32,
+	.write32 = tpm_tis_spi_write32,
+};
+
+static void cr50_print_fw_version(struct tpm_tis_data *data)
+{
+	struct tpm_tis_spi_phy *phy = to_tpm_tis_spi_phy(data);
+	int i, len = 0;
+	char fw_ver[TPM_CR50_MAX_FW_VER_LEN + 1];
+	char fw_ver_block[4];
+
+	/*
+	 * Write anything to TPM_CR50_FW_VER to start from the beginning
+	 * of the version string
+	 */
+	tpm_tis_write8(data, TPM_CR50_FW_VER(data->locality), 0);
+
+	/* Read the string, 4 bytes at a time, until we get '\0' */
+	do {
+		tpm_tis_read_bytes(data, TPM_CR50_FW_VER(data->locality), 4,
+				   fw_ver_block);
+		for (i = 0; i < 4 && fw_ver_block[i]; ++len, ++i)
+			fw_ver[len] = fw_ver_block[i];
+	} while (i == 4 && len < TPM_CR50_MAX_FW_VER_LEN);
+	fw_ver[len] = '\0';
+
+	dev_info(&phy->spi_device->dev, "Cr50 firmware version: %s\n", fw_ver);
+}
+
+int cr50_spi_probe(struct spi_device *spi)
+{
+	struct tpm_tis_spi_phy *phy;
+	struct cr50_spi_phy *cr50_phy;
+	int ret;
+	struct tpm_chip *chip;
+
+	cr50_phy = devm_kzalloc(&spi->dev, sizeof(*cr50_phy), GFP_KERNEL);
+	if (!cr50_phy)
+		return -ENOMEM;
+
+	phy = &cr50_phy->spi_phy;
+	phy->flow_control = cr50_spi_flow_control;
+	phy->wake_after = jiffies;
+	init_completion(&phy->ready);
+
+	cr50_phy->access_delay = CR50_NOIRQ_ACCESS_DELAY;
+	cr50_phy->last_access = jiffies;
+	mutex_init(&cr50_phy->time_track_mutex);
+
+	if (spi->irq > 0) {
+		ret = devm_request_irq(&spi->dev, spi->irq, cr50_spi_irq_handler,
+				       IRQF_TRIGGER_RISING | IRQF_ONESHOT,
+				       "cr50_spi", cr50_phy);
+		if (ret < 0) {
+			if (ret == -EPROBE_DEFER)
+				return ret;
+			dev_warn(&spi->dev, "Requesting IRQ %d failed: %d\n",
+				 spi->irq, ret);
+			/*
+			 * This is not fatal, the driver will fall back to
+			 * delays automatically, since ready will never
+			 * be completed without a registered irq handler.
+			 * So, just fall through.
+			 */
+		} else {
+			/*
+			 * IRQ requested, let's verify that it is actually
+			 * triggered, before relying on it.
+			 */
+			cr50_phy->irq_needs_confirmation = true;
+		}
+	} else {
+		dev_warn(&spi->dev,
+			 "No IRQ - will use delays between transactions.\n");
+	}
+
+	ret = tpm_tis_spi_init(spi, phy, -1, &tpm_spi_cr50_phy_ops);
+	if (ret)
+		return ret;
+
+	cr50_print_fw_version(&phy->priv);
+
+	chip = dev_get_drvdata(&spi->dev);
+	chip->flags |= TPM_CHIP_FLAG_FIRMWARE_POWER_MANAGED;
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+int tpm_tis_spi_resume(struct device *dev)
+{
+	struct tpm_chip *chip = dev_get_drvdata(dev);
+	struct tpm_tis_data *data = dev_get_drvdata(&chip->dev);
+	struct tpm_tis_spi_phy *phy = to_tpm_tis_spi_phy(data);
+	/*
+	 * Jiffies not increased during suspend, so we need to reset
+	 * the time to wake Cr50 after resume.
+	 */
+	phy->wake_after = jiffies;
+
+	return tpm_tis_resume(dev);
+}
+#endif