The CCP driver now has symbols that can be used by i2c-designware-amdpsp
to handle the communication regarding i2c arbitration with the PSP for
both Cezanne and Mendocino based designs.
Utilize those symbols.
v7->v8:
* Pick up tags
* Drop hunk for error message change
* Fix kernel robot build failures
Mario Limonciello (6):
crypto: ccp: Drop extra doorbell checks
crypto: ccp: Bump up doorbell debug message to error
crypto: ccp: Return doorbell status code as an argument
crypto: ccp: Use lower 8 bytes to communicate with doorbell command
register
i2c: designware: Use PCI PSP driver for communication
i2c: designware: Add doorbell support for Mendocino
drivers/crypto/ccp/platform-access.c | 31 +--
drivers/i2c/busses/Kconfig | 5 +-
drivers/i2c/busses/i2c-designware-amdpsp.c | 197 +++++---------------
drivers/i2c/busses/i2c-designware-core.h | 1 -
drivers/i2c/busses/i2c-designware-platdrv.c | 1 -
include/linux/psp-platform-access.h | 5 +-
6 files changed, 62 insertions(+), 178 deletions(-)
base-commit: d6cb9ab4bfeaaa69147948e42d7cff080db82d07
--
2.34.1
Currently the PSP semaphore communication base address is discovered
by using an MSR that is not architecturally guaranteed for future
platforms. Also the mailbox that is utilized for communication with
the PSP may have other consumers in the kernel, so it's better to
make all communication go through a single driver.
Signed-off-by: Mario Limonciello <[email protected]>
Reviewed-by: Mark Hasemeyer <[email protected]>
Acked-by: Jarkko Nikula <[email protected]>
Tested-by: Mark Hasemeyer <[email protected]>
---
v7->v8:
* Drop hunk changing error message
* Stop using imply. As shown by failed configs for kernel robot we need
stronger dependencies. Model them off of I2C_DESIGNWARE_BAYTRAIL which
has a similar relationship
---
drivers/i2c/busses/Kconfig | 4 +-
drivers/i2c/busses/i2c-designware-amdpsp.c | 175 +++-----------------
drivers/i2c/busses/i2c-designware-core.h | 1 -
drivers/i2c/busses/i2c-designware-platdrv.c | 1 -
include/linux/psp-platform-access.h | 1 +
5 files changed, 29 insertions(+), 153 deletions(-)
diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig
index 25eb4e8fd22f..89f8b75043d0 100644
--- a/drivers/i2c/busses/Kconfig
+++ b/drivers/i2c/busses/Kconfig
@@ -566,9 +566,11 @@ config I2C_DESIGNWARE_PLATFORM
config I2C_DESIGNWARE_AMDPSP
bool "AMD PSP I2C semaphore support"
- depends on X86_MSR
depends on ACPI
+ depends on CRYPTO_DEV_SP_PSP
depends on I2C_DESIGNWARE_PLATFORM
+ depends on (I2C_DESIGNWARE_PLATFORM=y && CRYPTO_DEV_CCP_DD=y) || \
+ (I2C_DESIGNWARE_PLATFORM=m && CRYPTO_DEV_CCP_DD)
help
This driver enables managed host access to the selected I2C bus shared
between AMD CPU and AMD PSP.
diff --git a/drivers/i2c/busses/i2c-designware-amdpsp.c b/drivers/i2c/busses/i2c-designware-amdpsp.c
index 652e6b64bd5f..12870dc44bdb 100644
--- a/drivers/i2c/busses/i2c-designware-amdpsp.c
+++ b/drivers/i2c/busses/i2c-designware-amdpsp.c
@@ -1,35 +1,20 @@
// SPDX-License-Identifier: GPL-2.0
-#include <linux/bitfield.h>
-#include <linux/bits.h>
#include <linux/i2c.h>
-#include <linux/io-64-nonatomic-lo-hi.h>
+#include <linux/psp-platform-access.h>
#include <linux/psp.h>
-#include <linux/types.h>
#include <linux/workqueue.h>
-#include <asm/msr.h>
-
#include "i2c-designware-core.h"
-#define MSR_AMD_PSP_ADDR 0xc00110a2
-#define PSP_MBOX_OFFSET 0x10570
-#define PSP_CMD_TIMEOUT_US (500 * USEC_PER_MSEC)
-
#define PSP_I2C_RESERVATION_TIME_MS 100
-#define PSP_I2C_REQ_BUS_CMD 0x64
#define PSP_I2C_REQ_RETRY_CNT 400
#define PSP_I2C_REQ_RETRY_DELAY_US (25 * USEC_PER_MSEC)
#define PSP_I2C_REQ_STS_OK 0x0
#define PSP_I2C_REQ_STS_BUS_BUSY 0x1
#define PSP_I2C_REQ_STS_INV_PARAM 0x3
-struct psp_req_buffer_hdr {
- u32 total_size;
- u32 status;
-};
-
enum psp_i2c_req_type {
PSP_I2C_REQ_ACQUIRE,
PSP_I2C_REQ_RELEASE,
@@ -41,119 +26,12 @@ struct psp_i2c_req {
enum psp_i2c_req_type type;
};
-struct psp_mbox {
- u32 cmd_fields;
- u64 i2c_req_addr;
-} __packed;
-
static DEFINE_MUTEX(psp_i2c_access_mutex);
static unsigned long psp_i2c_sem_acquired;
-static void __iomem *mbox_iomem;
static u32 psp_i2c_access_count;
static bool psp_i2c_mbox_fail;
static struct device *psp_i2c_dev;
-/*
- * Implementation of PSP-x86 i2c-arbitration mailbox introduced for AMD Cezanne
- * family of SoCs.
- */
-
-static int psp_get_mbox_addr(unsigned long *mbox_addr)
-{
- unsigned long long psp_mmio;
-
- if (rdmsrl_safe(MSR_AMD_PSP_ADDR, &psp_mmio))
- return -EIO;
-
- *mbox_addr = (unsigned long)(psp_mmio + PSP_MBOX_OFFSET);
-
- return 0;
-}
-
-static int psp_mbox_probe(void)
-{
- unsigned long mbox_addr;
- int ret;
-
- ret = psp_get_mbox_addr(&mbox_addr);
- if (ret)
- return ret;
-
- mbox_iomem = ioremap(mbox_addr, sizeof(struct psp_mbox));
- if (!mbox_iomem)
- return -ENOMEM;
-
- return 0;
-}
-
-/* Recovery field should be equal 0 to start sending commands */
-static int psp_check_mbox_recovery(struct psp_mbox __iomem *mbox)
-{
- u32 tmp;
-
- tmp = readl(&mbox->cmd_fields);
-
- return FIELD_GET(PSP_CMDRESP_RECOVERY, tmp);
-}
-
-static int psp_wait_cmd(struct psp_mbox __iomem *mbox)
-{
- u32 tmp, expected;
-
- /* Expect mbox_cmd to be cleared and the response bit to be set by PSP */
- expected = FIELD_PREP(PSP_CMDRESP_RESP, 1);
-
- /*
- * Check for readiness of PSP mailbox in a tight loop in order to
- * process further as soon as command was consumed.
- */
- return readl_poll_timeout(&mbox->cmd_fields, tmp, (tmp == expected),
- 0, PSP_CMD_TIMEOUT_US);
-}
-
-/* Status equal to 0 means that PSP succeed processing command */
-static u32 psp_check_mbox_sts(struct psp_mbox __iomem *mbox)
-{
- u32 cmd_reg;
-
- cmd_reg = readl(&mbox->cmd_fields);
-
- return FIELD_GET(PSP_CMDRESP_STS, cmd_reg);
-}
-
-static int psp_send_cmd(struct psp_i2c_req *req)
-{
- struct psp_mbox __iomem *mbox = mbox_iomem;
- phys_addr_t req_addr;
- u32 cmd_reg;
-
- if (psp_check_mbox_recovery(mbox))
- return -EIO;
-
- if (psp_wait_cmd(mbox))
- return -EBUSY;
-
- /*
- * Fill mailbox with address of command-response buffer, which will be
- * used for sending i2c requests as well as reading status returned by
- * PSP. Use physical address of buffer, since PSP will map this region.
- */
- req_addr = __psp_pa((void *)req);
- writeq(req_addr, &mbox->i2c_req_addr);
-
- /* Write command register to trigger processing */
- cmd_reg = FIELD_PREP(PSP_CMDRESP_CMD, PSP_I2C_REQ_BUS_CMD);
- writel(cmd_reg, &mbox->cmd_fields);
-
- if (psp_wait_cmd(mbox))
- return -ETIMEDOUT;
-
- if (psp_check_mbox_sts(mbox))
- return -EIO;
-
- return 0;
-}
-
/* Helper to verify status returned by PSP */
static int check_i2c_req_sts(struct psp_i2c_req *req)
{
@@ -173,22 +51,25 @@ static int check_i2c_req_sts(struct psp_i2c_req *req)
}
}
-static int psp_send_check_i2c_req(struct psp_i2c_req *req)
+/*
+ * Errors in x86-PSP i2c-arbitration protocol may occur at two levels:
+ * 1. mailbox communication - PSP is not operational or some IO errors with
+ * basic communication had happened.
+ * 2. i2c-requests - PSP refuses to grant i2c arbitration to x86 for too long.
+ *
+ * In order to distinguish between these in error handling code all mailbox
+ * communication errors on the first level (from CCP symbols) will be passed
+ * up and if -EIO is returned the second level will be checked.
+ */
+static int psp_send_i2c_req_cezanne(struct psp_i2c_req *req)
{
- /*
- * Errors in x86-PSP i2c-arbitration protocol may occur at two levels:
- * 1. mailbox communication - PSP is not operational or some IO errors
- * with basic communication had happened;
- * 2. i2c-requests - PSP refuses to grant i2c arbitration to x86 for too
- * long.
- * In order to distinguish between these two in error handling code, all
- * errors on the first level (returned by psp_send_cmd) are shadowed by
- * -EIO.
- */
- if (psp_send_cmd(req))
- return -EIO;
+ int ret;
- return check_i2c_req_sts(req);
+ ret = psp_send_platform_access_msg(PSP_I2C_REQ_BUS_CMD, (struct psp_request *)req);
+ if (ret == -EIO)
+ return check_i2c_req_sts(req);
+
+ return ret;
}
static int psp_send_i2c_req(enum psp_i2c_req_type i2c_req_type)
@@ -202,11 +83,11 @@ static int psp_send_i2c_req(enum psp_i2c_req_type i2c_req_type)
if (!req)
return -ENOMEM;
- req->hdr.total_size = sizeof(*req);
+ req->hdr.payload_size = sizeof(*req);
req->type = i2c_req_type;
start = jiffies;
- ret = read_poll_timeout(psp_send_check_i2c_req, status,
+ ret = read_poll_timeout(psp_send_i2c_req_cezanne, status,
(status != -EBUSY),
PSP_I2C_REQ_RETRY_DELAY_US,
PSP_I2C_REQ_RETRY_CNT * PSP_I2C_REQ_RETRY_DELAY_US,
@@ -381,7 +262,8 @@ static const struct i2c_lock_operations i2c_dw_psp_lock_ops = {
int i2c_dw_amdpsp_probe_lock_support(struct dw_i2c_dev *dev)
{
- int ret;
+ if (!IS_REACHABLE(CONFIG_CRYPTO_DEV_CCP_DD))
+ return -ENODEV;
if (!dev)
return -ENODEV;
@@ -393,11 +275,10 @@ int i2c_dw_amdpsp_probe_lock_support(struct dw_i2c_dev *dev)
if (psp_i2c_dev)
return -EEXIST;
- psp_i2c_dev = dev->dev;
+ if (psp_check_platform_access_status())
+ return -EPROBE_DEFER;
- ret = psp_mbox_probe();
- if (ret)
- return ret;
+ psp_i2c_dev = dev->dev;
dev_info(psp_i2c_dev, "I2C bus managed by AMD PSP\n");
@@ -411,9 +292,3 @@ int i2c_dw_amdpsp_probe_lock_support(struct dw_i2c_dev *dev)
return 0;
}
-
-/* Unmap area used as a mailbox with PSP */
-void i2c_dw_amdpsp_remove_lock_support(struct dw_i2c_dev *dev)
-{
- iounmap(mbox_iomem);
-}
diff --git a/drivers/i2c/busses/i2c-designware-core.h b/drivers/i2c/busses/i2c-designware-core.h
index 050d8c63ad3c..c5d87aae39c6 100644
--- a/drivers/i2c/busses/i2c-designware-core.h
+++ b/drivers/i2c/busses/i2c-designware-core.h
@@ -383,7 +383,6 @@ int i2c_dw_baytrail_probe_lock_support(struct dw_i2c_dev *dev);
#if IS_ENABLED(CONFIG_I2C_DESIGNWARE_AMDPSP)
int i2c_dw_amdpsp_probe_lock_support(struct dw_i2c_dev *dev);
-void i2c_dw_amdpsp_remove_lock_support(struct dw_i2c_dev *dev);
#endif
int i2c_dw_validate_speed(struct dw_i2c_dev *dev);
diff --git a/drivers/i2c/busses/i2c-designware-platdrv.c b/drivers/i2c/busses/i2c-designware-platdrv.c
index 74182db03a88..89ad88c54754 100644
--- a/drivers/i2c/busses/i2c-designware-platdrv.c
+++ b/drivers/i2c/busses/i2c-designware-platdrv.c
@@ -214,7 +214,6 @@ static const struct i2c_dw_semaphore_callbacks i2c_dw_semaphore_cb_table[] = {
#ifdef CONFIG_I2C_DESIGNWARE_AMDPSP
{
.probe = i2c_dw_amdpsp_probe_lock_support,
- .remove = i2c_dw_amdpsp_remove_lock_support,
},
#endif
{}
diff --git a/include/linux/psp-platform-access.h b/include/linux/psp-platform-access.h
index 1b661341d8f3..75da8f5f7ad8 100644
--- a/include/linux/psp-platform-access.h
+++ b/include/linux/psp-platform-access.h
@@ -7,6 +7,7 @@
enum psp_platform_access_msg {
PSP_CMD_NONE = 0x0,
+ PSP_I2C_REQ_BUS_CMD = 0x64,
};
struct psp_req_buffer_hdr {
--
2.34.1
Mendocino and later platform don't use the platform feature mailbox for
communication for I2C arbitration, they rely upon ringing a doorbell.
Detect the platform by the device ID of the root port and choose the
appropriate method.
Link: https://lore.kernel.org/linux-i2c/[email protected]/
Signed-off-by: Mario Limonciello <[email protected]>
Acked-by: Jarkko Nikula <[email protected]>
Reviewed-by: Mark Hasemeyer <[email protected]>
Tested-by: Mark Hasemeyer <[email protected]>
---
drivers/i2c/busses/Kconfig | 1 +
drivers/i2c/busses/i2c-designware-amdpsp.c | 26 +++++++++++++++++++++-
2 files changed, 26 insertions(+), 1 deletion(-)
diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig
index 89f8b75043d0..4b4323bbf268 100644
--- a/drivers/i2c/busses/Kconfig
+++ b/drivers/i2c/busses/Kconfig
@@ -568,6 +568,7 @@ config I2C_DESIGNWARE_AMDPSP
bool "AMD PSP I2C semaphore support"
depends on ACPI
depends on CRYPTO_DEV_SP_PSP
+ depends on PCI
depends on I2C_DESIGNWARE_PLATFORM
depends on (I2C_DESIGNWARE_PLATFORM=y && CRYPTO_DEV_CCP_DD=y) || \
(I2C_DESIGNWARE_PLATFORM=m && CRYPTO_DEV_CCP_DD)
diff --git a/drivers/i2c/busses/i2c-designware-amdpsp.c b/drivers/i2c/busses/i2c-designware-amdpsp.c
index 12870dc44bdb..63454b06e5da 100644
--- a/drivers/i2c/busses/i2c-designware-amdpsp.c
+++ b/drivers/i2c/busses/i2c-designware-amdpsp.c
@@ -1,6 +1,7 @@
// SPDX-License-Identifier: GPL-2.0
#include <linux/i2c.h>
+#include <linux/pci.h>
#include <linux/psp-platform-access.h>
#include <linux/psp.h>
#include <linux/workqueue.h>
@@ -32,6 +33,8 @@ static u32 psp_i2c_access_count;
static bool psp_i2c_mbox_fail;
static struct device *psp_i2c_dev;
+static int (*_psp_send_i2c_req)(struct psp_i2c_req *req);
+
/* Helper to verify status returned by PSP */
static int check_i2c_req_sts(struct psp_i2c_req *req)
{
@@ -72,6 +75,17 @@ static int psp_send_i2c_req_cezanne(struct psp_i2c_req *req)
return ret;
}
+static int psp_send_i2c_req_doorbell(struct psp_i2c_req *req)
+{
+ int ret;
+
+ ret = psp_ring_platform_doorbell(req->type, &req->hdr.status);
+ if (ret == -EIO)
+ return check_i2c_req_sts(req);
+
+ return ret;
+}
+
static int psp_send_i2c_req(enum psp_i2c_req_type i2c_req_type)
{
struct psp_i2c_req *req;
@@ -87,7 +101,7 @@ static int psp_send_i2c_req(enum psp_i2c_req_type i2c_req_type)
req->type = i2c_req_type;
start = jiffies;
- ret = read_poll_timeout(psp_send_i2c_req_cezanne, status,
+ ret = read_poll_timeout(_psp_send_i2c_req, status,
(status != -EBUSY),
PSP_I2C_REQ_RETRY_DELAY_US,
PSP_I2C_REQ_RETRY_CNT * PSP_I2C_REQ_RETRY_DELAY_US,
@@ -262,6 +276,8 @@ static const struct i2c_lock_operations i2c_dw_psp_lock_ops = {
int i2c_dw_amdpsp_probe_lock_support(struct dw_i2c_dev *dev)
{
+ struct pci_dev *rdev;
+
if (!IS_REACHABLE(CONFIG_CRYPTO_DEV_CCP_DD))
return -ENODEV;
@@ -275,6 +291,14 @@ int i2c_dw_amdpsp_probe_lock_support(struct dw_i2c_dev *dev)
if (psp_i2c_dev)
return -EEXIST;
+ /* Cezanne uses platform mailbox, Mendocino and later use doorbell */
+ rdev = pci_get_domain_bus_and_slot(0, 0, PCI_DEVFN(0, 0));
+ if (rdev->device == 0x1630)
+ _psp_send_i2c_req = psp_send_i2c_req_cezanne;
+ else
+ _psp_send_i2c_req = psp_send_i2c_req_doorbell;
+ pci_dev_put(rdev);
+
if (psp_check_platform_access_status())
return -EPROBE_DEFER;
--
2.34.1
On Mon, Apr 03, 2023 at 01:32:09PM -0500, Mario Limonciello wrote:
> The CCP driver now has symbols that can be used by i2c-designware-amdpsp
> to handle the communication regarding i2c arbitration with the PSP for
> both Cezanne and Mendocino based designs.
How should this go upstream, i.e. are there dependencies? Shall I pick
the I2C patches or is it better if all goes via the crypto tree?
On 4/6/23 01:42, Wolfram Sang wrote:
> On Mon, Apr 03, 2023 at 01:32:09PM -0500, Mario Limonciello wrote:
>> The CCP driver now has symbols that can be used by i2c-designware-amdpsp
>> to handle the communication regarding i2c arbitration with the PSP for
>> both Cezanne and Mendocino based designs.
> How should this go upstream, i.e. are there dependencies? Shall I pick
> the I2C patches or is it better if all goes via the crypto tree?
>
IMO it's better to go through the crypto tree. There are dependencies
in the crypto part from the earlier series that was merged.
On Mon, Apr 03, 2023 at 01:32:14PM -0500, Mario Limonciello wrote:
> Currently the PSP semaphore communication base address is discovered
> by using an MSR that is not architecturally guaranteed for future
> platforms. Also the mailbox that is utilized for communication with
> the PSP may have other consumers in the kernel, so it's better to
> make all communication go through a single driver.
>
> Signed-off-by: Mario Limonciello <[email protected]>
> Reviewed-by: Mark Hasemeyer <[email protected]>
> Acked-by: Jarkko Nikula <[email protected]>
> Tested-by: Mark Hasemeyer <[email protected]>
Acked-by: Wolfram Sang <[email protected]>
On Mon, Apr 03, 2023 at 01:32:15PM -0500, Mario Limonciello wrote:
> Mendocino and later platform don't use the platform feature mailbox for
> communication for I2C arbitration, they rely upon ringing a doorbell.
>
> Detect the platform by the device ID of the root port and choose the
> appropriate method.
>
> Link: https://lore.kernel.org/linux-i2c/[email protected]/
> Signed-off-by: Mario Limonciello <[email protected]>
> Acked-by: Jarkko Nikula <[email protected]>
> Reviewed-by: Mark Hasemeyer <[email protected]>
> Tested-by: Mark Hasemeyer <[email protected]>
Acked-by: Wolfram Sang <[email protected]>
On Fri, Apr 07, 2023 at 08:17:36AM +0200, Wolfram Sang wrote:
> On Mon, Apr 03, 2023 at 01:32:14PM -0500, Mario Limonciello wrote:
> > Currently the PSP semaphore communication base address is discovered
> > by using an MSR that is not architecturally guaranteed for future
> > platforms. Also the mailbox that is utilized for communication with
> > the PSP may have other consumers in the kernel, so it's better to
> > make all communication go through a single driver.
> >
> > Signed-off-by: Mario Limonciello <[email protected]>
> > Reviewed-by: Mark Hasemeyer <[email protected]>
> > Acked-by: Jarkko Nikula <[email protected]>
> > Tested-by: Mark Hasemeyer <[email protected]>
>
> Acked-by: Wolfram Sang <[email protected]>
Thanks, but as the original patch wasn't posted to linux-crypto
so I can't pick this up.
Mario, could you repost just these two patches with the Acks added?
Cheers,
--
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt