[E][unstable][PATCH 0/8] thunderbolt: ICL support

classic Classic list List threaded Threaded
10 messages Options
Reply | Threaded
Open this post in threaded view
|

[E][unstable][PATCH 0/8] thunderbolt: ICL support

Timo Aaltonen-6
BugLink: http://bugs.launchpad.net/bugs/1844680

These patches from 5.4 add Thunderbolt support for Ice Lake.

Mika Westerberg (8):
  thunderbolt: Correct path indices for PCIe tunnel
  thunderbolt: Move NVM upgrade support flag to struct icm
  thunderbolt: Use 32-bit writes when writing ring producer/consumer
  thunderbolt: Do not fail adding switch if some port is not implemented
  thunderbolt: Hide switch attributes that are not set
  thunderbolt: Expose active parts of NVM even if upgrade is not
    supported
  thunderbolt: Add support for Intel Ice Lake
  ACPI / property: Add two new Thunderbolt property GUIDs to the list

 drivers/acpi/property.c        |   6 +
 drivers/thunderbolt/Makefile   |   2 +-
 drivers/thunderbolt/ctl.c      |  23 +++-
 drivers/thunderbolt/eeprom.c   |   4 -
 drivers/thunderbolt/icm.c      | 194 ++++++++++++++++++++++++++++++---
 drivers/thunderbolt/nhi.c      | 134 +++++++++++++++++++++--
 drivers/thunderbolt/nhi.h      |  22 ++++
 drivers/thunderbolt/nhi_ops.c  | 179 ++++++++++++++++++++++++++++++
 drivers/thunderbolt/nhi_regs.h |  37 +++++++
 drivers/thunderbolt/switch.c   |  52 +++++++--
 drivers/thunderbolt/tb_msgs.h  |  16 ++-
 drivers/thunderbolt/tunnel.c   |   4 +-
 include/linux/thunderbolt.h    |   2 +
 13 files changed, 622 insertions(+), 53 deletions(-)
 create mode 100644 drivers/thunderbolt/nhi_ops.c

--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 1/8] thunderbolt: Correct path indices for PCIe tunnel

Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

PCIe tunnel path indices got mixed up when we added support for tunnels
between switches that are not adjacent. This did not affect the
functionality as it is just an index but fix it now nevertheless to make
the code easier to understand.

Reported-by: Rajmohan Mani <[hidden email]>
Fixes: 8c7acaaf020f ("thunderbolt: Extend tunnel creation to more than 2 adjacent switches")
Signed-off-by: Mika Westerberg <[hidden email]>
Reviewed-by: Yehezkel Bernat <[hidden email]>
(cherry picked from commit ce19f91eae43e39d5a1da55344756ab5a3c7e8d1)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/thunderbolt/tunnel.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/thunderbolt/tunnel.c b/drivers/thunderbolt/tunnel.c
index 31d0234837e4..5a99234826e7 100644
--- a/drivers/thunderbolt/tunnel.c
+++ b/drivers/thunderbolt/tunnel.c
@@ -211,7 +211,7 @@ struct tb_tunnel *tb_tunnel_alloc_pci(struct tb *tb, struct tb_port *up,
  return NULL;
  }
  tb_pci_init_path(path);
- tunnel->paths[TB_PCI_PATH_UP] = path;
+ tunnel->paths[TB_PCI_PATH_DOWN] = path;
 
  path = tb_path_alloc(tb, up, TB_PCI_HOPID, down, TB_PCI_HOPID, 0,
      "PCIe Up");
@@ -220,7 +220,7 @@ struct tb_tunnel *tb_tunnel_alloc_pci(struct tb *tb, struct tb_port *up,
  return NULL;
  }
  tb_pci_init_path(path);
- tunnel->paths[TB_PCI_PATH_DOWN] = path;
+ tunnel->paths[TB_PCI_PATH_UP] = path;
 
  return tunnel;
 }
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 2/8] thunderbolt: Move NVM upgrade support flag to struct icm

Timo Aaltonen-6
In reply to this post by Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

This is depends on the controller and on the platform/CPU we are
running. Move it to struct icm so we can set it per controller.

Signed-off-by: Mika Westerberg <[hidden email]>
Reviewed-by: Yehezkel Bernat <[hidden email]>
Tested-by: Mario Limonciello <[hidden email]>
(cherry picked from commit f437c24bf694b0293f835dea8c25e3a5c1433d07)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/thunderbolt/icm.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

diff --git a/drivers/thunderbolt/icm.c b/drivers/thunderbolt/icm.c
index fbdcef56a676..e9835ab35465 100644
--- a/drivers/thunderbolt/icm.c
+++ b/drivers/thunderbolt/icm.c
@@ -55,6 +55,7 @@
  * @safe_mode: ICM is in safe mode
  * @max_boot_acl: Maximum number of preboot ACL entries (%0 if not supported)
  * @rpm: Does the controller support runtime PM (RTD3)
+ * @can_upgrade_nvm: Can the NVM firmware be upgrade on this controller
  * @is_supported: Checks if we can support ICM on this controller
  * @cio_reset: Trigger CIO reset
  * @get_mode: Read and return the ICM firmware mode (optional)
@@ -74,6 +75,7 @@ struct icm {
  int vnd_cap;
  bool safe_mode;
  bool rpm;
+ bool can_upgrade_nvm;
  bool (*is_supported)(struct tb *tb);
  int (*cio_reset)(struct tb *tb);
  int (*get_mode)(struct tb *tb);
@@ -1913,12 +1915,7 @@ static int icm_start(struct tb *tb)
  if (IS_ERR(tb->root_switch))
  return PTR_ERR(tb->root_switch);
 
- /*
- * NVM upgrade has not been tested on Apple systems and they
- * don't provide images publicly either. To be on the safe side
- * prevent root switch NVM upgrade on Macs for now.
- */
- tb->root_switch->no_nvm_upgrade = x86_apple_machine;
+ tb->root_switch->no_nvm_upgrade = !icm->can_upgrade_nvm;
  tb->root_switch->rpm = icm->rpm;
 
  ret = tb_switch_add(tb->root_switch);
@@ -2021,6 +2018,7 @@ struct tb *icm_probe(struct tb_nhi *nhi)
  switch (nhi->pdev->device) {
  case PCI_DEVICE_ID_INTEL_FALCON_RIDGE_2C_NHI:
  case PCI_DEVICE_ID_INTEL_FALCON_RIDGE_4C_NHI:
+ icm->can_upgrade_nvm = true;
  icm->is_supported = icm_fr_is_supported;
  icm->get_route = icm_fr_get_route;
  icm->save_devices = icm_fr_save_devices;
@@ -2038,6 +2036,13 @@ struct tb *icm_probe(struct tb_nhi *nhi)
  case PCI_DEVICE_ID_INTEL_ALPINE_RIDGE_C_4C_NHI:
  case PCI_DEVICE_ID_INTEL_ALPINE_RIDGE_C_2C_NHI:
  icm->max_boot_acl = ICM_AR_PREBOOT_ACL_ENTRIES;
+ /*
+ * NVM upgrade has not been tested on Apple systems and
+ * they don't provide images publicly either. To be on
+ * the safe side prevent root switch NVM upgrade on Macs
+ * for now.
+ */
+ icm->can_upgrade_nvm = !x86_apple_machine;
  icm->is_supported = icm_ar_is_supported;
  icm->cio_reset = icm_ar_cio_reset;
  icm->get_mode = icm_ar_get_mode;
@@ -2054,6 +2059,7 @@ struct tb *icm_probe(struct tb_nhi *nhi)
  case PCI_DEVICE_ID_INTEL_TITAN_RIDGE_2C_NHI:
  case PCI_DEVICE_ID_INTEL_TITAN_RIDGE_4C_NHI:
  icm->max_boot_acl = ICM_AR_PREBOOT_ACL_ENTRIES;
+ icm->can_upgrade_nvm = !x86_apple_machine;
  icm->is_supported = icm_ar_is_supported;
  icm->cio_reset = icm_tr_cio_reset;
  icm->get_mode = icm_ar_get_mode;
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 3/8] thunderbolt: Use 32-bit writes when writing ring producer/consumer

Timo Aaltonen-6
In reply to this post by Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

The register access should be using 32-bit reads/writes according to the
datasheet. With the previous generation hardware 16-bit writes have been
working but starting with ICL this is not the case anymore so fix
producer/consumer register update to use correct width register address.

Signed-off-by: Mika Westerberg <[hidden email]>
Reviewed-by: Yehezkel Bernat <[hidden email]>
Tested-by: Mario Limonciello <[hidden email]>
(cherry picked from commit 943795219d3cb9f8ce6ce51cad3ffe1f61e95c6b)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/thunderbolt/nhi.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)

diff --git a/drivers/thunderbolt/nhi.c b/drivers/thunderbolt/nhi.c
index 27fbe62c7ddd..9c782706e652 100644
--- a/drivers/thunderbolt/nhi.c
+++ b/drivers/thunderbolt/nhi.c
@@ -143,9 +143,20 @@ static void __iomem *ring_options_base(struct tb_ring *ring)
  return io;
 }
 
-static void ring_iowrite16desc(struct tb_ring *ring, u32 value, u32 offset)
+static void ring_iowrite_cons(struct tb_ring *ring, u16 cons)
 {
- iowrite16(value, ring_desc_base(ring) + offset);
+ /*
+ * The other 16-bits in the register is read-only and writes to it
+ * are ignored by the hardware so we can save one ioread32() by
+ * filling the read-only bits with zeroes.
+ */
+ iowrite32(cons, ring_desc_base(ring) + 8);
+}
+
+static void ring_iowrite_prod(struct tb_ring *ring, u16 prod)
+{
+ /* See ring_iowrite_cons() above for explanation */
+ iowrite32(prod << 16, ring_desc_base(ring) + 8);
 }
 
 static void ring_iowrite32desc(struct tb_ring *ring, u32 value, u32 offset)
@@ -197,7 +208,10 @@ static void ring_write_descriptors(struct tb_ring *ring)
  descriptor->sof = frame->sof;
  }
  ring->head = (ring->head + 1) % ring->size;
- ring_iowrite16desc(ring, ring->head, ring->is_tx ? 10 : 8);
+ if (ring->is_tx)
+ ring_iowrite_prod(ring, ring->head);
+ else
+ ring_iowrite_cons(ring, ring->head);
  }
 }
 
@@ -662,7 +676,7 @@ void tb_ring_stop(struct tb_ring *ring)
 
  ring_iowrite32options(ring, 0, 0);
  ring_iowrite64desc(ring, 0, 0);
- ring_iowrite16desc(ring, 0, ring->is_tx ? 10 : 8);
+ ring_iowrite32desc(ring, 0, 8);
  ring_iowrite32desc(ring, 0, 12);
  ring->head = 0;
  ring->tail = 0;
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 4/8] thunderbolt: Do not fail adding switch if some port is not implemented

Timo Aaltonen-6
In reply to this post by Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

There are two ways to mark a port as unimplemented. Typical way is to
return port type as TB_TYPE_INACTIVE when its config space is read.
Alternatively if the port is not physically present (such as ports 10
and 11 in ICL) reading from port config space returns
TB_CFG_ERROR_INVALID_CONFIG_SPACE instead. Currently the driver bails
out from adding the switch if it receives any error during port
inititialization which is wrong.

Handle this properly and just leave the port as TB_TYPE_INACTIVE before
continuing to the next port.

This also allows us to get rid of special casing for Light Ridge port 5
in eeprom.c.

Signed-off-by: Mika Westerberg <[hidden email]>
Reviewed-by: Yehezkel Bernat <[hidden email]>
Tested-by: Mario Limonciello <[hidden email]>
(cherry picked from commit d94dcbb10183f3b384c84e65724d2b753aa53c4d)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/thunderbolt/ctl.c    | 23 +++++++++++++++++++----
 drivers/thunderbolt/eeprom.c |  4 ----
 drivers/thunderbolt/switch.c |  8 +++++++-
 3 files changed, 26 insertions(+), 9 deletions(-)

diff --git a/drivers/thunderbolt/ctl.c b/drivers/thunderbolt/ctl.c
index 2427d73be731..2ec1af8f7968 100644
--- a/drivers/thunderbolt/ctl.c
+++ b/drivers/thunderbolt/ctl.c
@@ -930,6 +930,23 @@ struct tb_cfg_result tb_cfg_write_raw(struct tb_ctl *ctl, const void *buffer,
  return res;
 }
 
+static int tb_cfg_get_error(struct tb_ctl *ctl, enum tb_cfg_space space,
+    const struct tb_cfg_result *res)
+{
+ /*
+ * For unimplemented ports access to port config space may return
+ * TB_CFG_ERROR_INVALID_CONFIG_SPACE (alternatively their type is
+ * set to TB_TYPE_INACTIVE). In the former case return -ENODEV so
+ * that the caller can mark the port as disabled.
+ */
+ if (space == TB_CFG_PORT &&
+    res->tb_error == TB_CFG_ERROR_INVALID_CONFIG_SPACE)
+ return -ENODEV;
+
+ tb_cfg_print_error(ctl, res);
+ return -EIO;
+}
+
 int tb_cfg_read(struct tb_ctl *ctl, void *buffer, u64 route, u32 port,
  enum tb_cfg_space space, u32 offset, u32 length)
 {
@@ -942,8 +959,7 @@ int tb_cfg_read(struct tb_ctl *ctl, void *buffer, u64 route, u32 port,
 
  case 1:
  /* Thunderbolt error, tb_error holds the actual number */
- tb_cfg_print_error(ctl, &res);
- return -EIO;
+ return tb_cfg_get_error(ctl, space, &res);
 
  case -ETIMEDOUT:
  tb_ctl_warn(ctl, "timeout reading config space %u from %#x\n",
@@ -969,8 +985,7 @@ int tb_cfg_write(struct tb_ctl *ctl, const void *buffer, u64 route, u32 port,
 
  case 1:
  /* Thunderbolt error, tb_error holds the actual number */
- tb_cfg_print_error(ctl, &res);
- return -EIO;
+ return tb_cfg_get_error(ctl, space, &res);
 
  case -ETIMEDOUT:
  tb_ctl_warn(ctl, "timeout writing config space %u to %#x\n",
diff --git a/drivers/thunderbolt/eeprom.c b/drivers/thunderbolt/eeprom.c
index 81e8ac4c5805..3b2fb4aafd92 100644
--- a/drivers/thunderbolt/eeprom.c
+++ b/drivers/thunderbolt/eeprom.c
@@ -525,10 +525,6 @@ int tb_drom_read(struct tb_switch *sw)
  sw->ports[3].dual_link_port = &sw->ports[4];
  sw->ports[4].dual_link_port = &sw->ports[3];
 
- /* Port 5 is inaccessible on this gen 1 controller */
- if (sw->config.device_id == PCI_DEVICE_ID_INTEL_LIGHT_RIDGE)
- sw->ports[5].disabled = true;
-
  return 0;
  }
 
diff --git a/drivers/thunderbolt/switch.c b/drivers/thunderbolt/switch.c
index 5668a44e0653..64f845fe4690 100644
--- a/drivers/thunderbolt/switch.c
+++ b/drivers/thunderbolt/switch.c
@@ -611,8 +611,14 @@ static int tb_init_port(struct tb_port *port)
  int cap;
 
  res = tb_port_read(port, &port->config, TB_CFG_PORT, 0, 8);
- if (res)
+ if (res) {
+ if (res == -ENODEV) {
+ tb_dbg(port->sw->tb, " Port %d: not implemented\n",
+       port->port);
+ return 0;
+ }
  return res;
+ }
 
  /* Port 0 is the switch itself and has no PHY. */
  if (port->config.type == TB_TYPE_PORT && port->port != 0) {
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 5/8] thunderbolt: Hide switch attributes that are not set

Timo Aaltonen-6
In reply to this post by Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

Thunderbolt host routers may not always contain DROM that includes
device identification information. This is mostly needed for Ice Lake
systems but some Falcon Ridge controllers on PCs also do not have DROM.

In that case hide the identification attributes.

Signed-off-by: Mika Westerberg <[hidden email]>
Reviewed-by: Yehezkel Bernat <[hidden email]>
Tested-by: Mario Limonciello <[hidden email]>
(cherry picked from commit 58f414fa435cf728a82f435bac4781da86afb623)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/thunderbolt/switch.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/thunderbolt/switch.c b/drivers/thunderbolt/switch.c
index 64f845fe4690..bd3eaaf34ea1 100644
--- a/drivers/thunderbolt/switch.c
+++ b/drivers/thunderbolt/switch.c
@@ -1337,7 +1337,19 @@ static umode_t switch_attr_is_visible(struct kobject *kobj,
  struct device *dev = container_of(kobj, struct device, kobj);
  struct tb_switch *sw = tb_to_switch(dev);
 
- if (attr == &dev_attr_key.attr) {
+ if (attr == &dev_attr_device.attr) {
+ if (!sw->device)
+ return 0;
+ } else if (attr == &dev_attr_device_name.attr) {
+ if (!sw->device_name)
+ return 0;
+ } else if (attr == &dev_attr_vendor.attr)  {
+ if (!sw->vendor)
+ return 0;
+ } else if (attr == &dev_attr_vendor_name.attr)  {
+ if (!sw->vendor_name)
+ return 0;
+ } else if (attr == &dev_attr_key.attr) {
  if (tb_route(sw) &&
     sw->tb->security_level == TB_SECURITY_SECURE &&
     sw->security_level == TB_SECURITY_SECURE)
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 6/8] thunderbolt: Expose active parts of NVM even if upgrade is not supported

Timo Aaltonen-6
In reply to this post by Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

Ice Lake Thunderbolt controller NVM firmware is part of the BIOS image
which means it is not writable through the DMA port anymore. However, we
can still read it so we can keep nvm_version and active parts of NVM.
This way users still can find out the active NVM version and other
potentially useful information directly from Linux.

Signed-off-by: Mika Westerberg <[hidden email]>
Reviewed-by: Yehezkel Bernat <[hidden email]>
Tested-by: Mario Limonciello <[hidden email]>
(cherry picked from commit 3f415e5ee18b0097755afc3ac3a5640b196a239e)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/thunderbolt/switch.c | 28 +++++++++++++++++++---------
 1 file changed, 19 insertions(+), 9 deletions(-)

diff --git a/drivers/thunderbolt/switch.c b/drivers/thunderbolt/switch.c
index bd3eaaf34ea1..531f11fecf75 100644
--- a/drivers/thunderbolt/switch.c
+++ b/drivers/thunderbolt/switch.c
@@ -364,12 +364,14 @@ static int tb_switch_nvm_add(struct tb_switch *sw)
  nvm->active = nvm_dev;
  }
 
- nvm_dev = register_nvmem(sw, nvm->id, NVM_MAX_SIZE, false);
- if (IS_ERR(nvm_dev)) {
- ret = PTR_ERR(nvm_dev);
- goto err_nvm_active;
+ if (!sw->no_nvm_upgrade) {
+ nvm_dev = register_nvmem(sw, nvm->id, NVM_MAX_SIZE, false);
+ if (IS_ERR(nvm_dev)) {
+ ret = PTR_ERR(nvm_dev);
+ goto err_nvm_active;
+ }
+ nvm->non_active = nvm_dev;
  }
- nvm->non_active = nvm_dev;
 
  sw->nvm = nvm;
  return 0;
@@ -398,7 +400,8 @@ static void tb_switch_nvm_remove(struct tb_switch *sw)
  if (!nvm->authenticating)
  nvm_clear_auth_status(sw);
 
- nvmem_unregister(nvm->non_active);
+ if (nvm->non_active)
+ nvmem_unregister(nvm->non_active);
  if (nvm->active)
  nvmem_unregister(nvm->active);
  ida_simple_remove(&nvm_ida, nvm->id);
@@ -1355,8 +1358,11 @@ static umode_t switch_attr_is_visible(struct kobject *kobj,
     sw->security_level == TB_SECURITY_SECURE)
  return attr->mode;
  return 0;
- } else if (attr == &dev_attr_nvm_authenticate.attr ||
-   attr == &dev_attr_nvm_version.attr) {
+ } else if (attr == &dev_attr_nvm_authenticate.attr) {
+ if (sw->dma_port && !sw->no_nvm_upgrade)
+ return attr->mode;
+ return 0;
+ } else if (attr == &dev_attr_nvm_version.attr) {
  if (sw->dma_port)
  return attr->mode;
  return 0;
@@ -1707,13 +1713,17 @@ static int tb_switch_add_dma_port(struct tb_switch *sw)
  break;
  }
 
- if (sw->no_nvm_upgrade)
+ /* Root switch DMA port requires running firmware */
+ if (!tb_route(sw) && sw->config.enabled)
  return 0;
 
  sw->dma_port = dma_port_alloc(sw);
  if (!sw->dma_port)
  return 0;
 
+ if (sw->no_nvm_upgrade)
+ return 0;
+
  /*
  * Check status of the previous flash authentication. If there
  * is one we need to power cycle the switch in any case to make
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 7/8] thunderbolt: Add support for Intel Ice Lake

Timo Aaltonen-6
In reply to this post by Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

The Thunderbolt controller is integrated into the Ice Lake CPU itself
and requires special flows to power it on and off using force power bit
in NHI VSEC registers. Runtime PM (RTD3) and Sx flows also differ from
the discrete solutions. Now the firmware notifies the driver whether
RTD3 entry or exit are possible. The driver is responsible of sending
Go2Sx command through link controller mailbox when system enters Sx
states (suspend-to-mem/disk). Rest of the ICM firwmare flows follow
Titan Ridge.

Signed-off-by: Raanan Avargil <[hidden email]>
Signed-off-by: Mika Westerberg <[hidden email]>
Reviewed-by: Yehezkel Bernat <[hidden email]>
Tested-by: Mario Limonciello <[hidden email]>
(cherry picked from commit 3cdb9446a117d5d63af823bde6fe6babc312e77b)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/thunderbolt/Makefile   |   2 +-
 drivers/thunderbolt/icm.c      | 176 ++++++++++++++++++++++++++++++--
 drivers/thunderbolt/nhi.c      | 112 +++++++++++++++++++--
 drivers/thunderbolt/nhi.h      |  22 ++++
 drivers/thunderbolt/nhi_ops.c  | 179 +++++++++++++++++++++++++++++++++
 drivers/thunderbolt/nhi_regs.h |  37 +++++++
 drivers/thunderbolt/switch.c   |   2 +
 drivers/thunderbolt/tb_msgs.h  |  16 ++-
 include/linux/thunderbolt.h    |   2 +
 9 files changed, 526 insertions(+), 22 deletions(-)
 create mode 100644 drivers/thunderbolt/nhi_ops.c

diff --git a/drivers/thunderbolt/Makefile b/drivers/thunderbolt/Makefile
index 3f55cb3c81b2..001187c577bf 100644
--- a/drivers/thunderbolt/Makefile
+++ b/drivers/thunderbolt/Makefile
@@ -1,4 +1,4 @@
 # SPDX-License-Identifier: GPL-2.0-only
 obj-${CONFIG_THUNDERBOLT} := thunderbolt.o
-thunderbolt-objs := nhi.o ctl.o tb.o switch.o cap.o path.o tunnel.o eeprom.o
+thunderbolt-objs := nhi.o nhi_ops.o ctl.o tb.o switch.o cap.o path.o tunnel.o eeprom.o
 thunderbolt-objs += domain.o dma_port.o icm.o property.o xdomain.o lc.o
diff --git a/drivers/thunderbolt/icm.c b/drivers/thunderbolt/icm.c
index e9835ab35465..245588f691e7 100644
--- a/drivers/thunderbolt/icm.c
+++ b/drivers/thunderbolt/icm.c
@@ -56,16 +56,19 @@
  * @max_boot_acl: Maximum number of preboot ACL entries (%0 if not supported)
  * @rpm: Does the controller support runtime PM (RTD3)
  * @can_upgrade_nvm: Can the NVM firmware be upgrade on this controller
+ * @veto: Is RTD3 veto in effect
  * @is_supported: Checks if we can support ICM on this controller
  * @cio_reset: Trigger CIO reset
  * @get_mode: Read and return the ICM firmware mode (optional)
  * @get_route: Find a route string for given switch
  * @save_devices: Ask ICM to save devices to ACL when suspending (optional)
  * @driver_ready: Send driver ready message to ICM
+ * @set_uuid: Set UUID for the root switch (optional)
  * @device_connected: Handle device connected ICM message
  * @device_disconnected: Handle device disconnected ICM message
  * @xdomain_connected - Handle XDomain connected ICM message
  * @xdomain_disconnected - Handle XDomain disconnected ICM message
+ * @rtd3_veto: Handle RTD3 veto notification ICM message
  */
 struct icm {
  struct mutex request_lock;
@@ -76,6 +79,7 @@ struct icm {
  bool safe_mode;
  bool rpm;
  bool can_upgrade_nvm;
+ bool veto;
  bool (*is_supported)(struct tb *tb);
  int (*cio_reset)(struct tb *tb);
  int (*get_mode)(struct tb *tb);
@@ -84,6 +88,7 @@ struct icm {
  int (*driver_ready)(struct tb *tb,
     enum tb_security_level *security_level,
     size_t *nboot_acl, bool *rpm);
+ void (*set_uuid)(struct tb *tb);
  void (*device_connected)(struct tb *tb,
  const struct icm_pkg_header *hdr);
  void (*device_disconnected)(struct tb *tb,
@@ -92,6 +97,7 @@ struct icm {
   const struct icm_pkg_header *hdr);
  void (*xdomain_disconnected)(struct tb *tb,
      const struct icm_pkg_header *hdr);
+ void (*rtd3_veto)(struct tb *tb, const struct icm_pkg_header *hdr);
 };
 
 struct icm_notification {
@@ -296,6 +302,43 @@ static int icm_request(struct tb *tb, const void *request, size_t request_size,
  return -ETIMEDOUT;
 }
 
+/*
+ * If rescan is queued to run (we are resuming), postpone it to give the
+ * firmware some more time to send device connected notifications for next
+ * devices in the chain.
+ */
+static void icm_postpone_rescan(struct tb *tb)
+{
+ struct icm *icm = tb_priv(tb);
+
+ if (delayed_work_pending(&icm->rescan_work))
+ mod_delayed_work(tb->wq, &icm->rescan_work,
+ msecs_to_jiffies(500));
+}
+
+static void icm_veto_begin(struct tb *tb)
+{
+ struct icm *icm = tb_priv(tb);
+
+ if (!icm->veto) {
+ icm->veto = true;
+ /* Keep the domain powered while veto is in effect */
+ pm_runtime_get(&tb->dev);
+ }
+}
+
+static void icm_veto_end(struct tb *tb)
+{
+ struct icm *icm = tb_priv(tb);
+
+ if (icm->veto) {
+ icm->veto = false;
+ /* Allow the domain suspend now */
+ pm_runtime_mark_last_busy(&tb->dev);
+ pm_runtime_put_autosuspend(&tb->dev);
+ }
+}
+
 static bool icm_fr_is_supported(struct tb *tb)
 {
  return !x86_apple_machine;
@@ -519,14 +562,16 @@ static int icm_fr_disconnect_xdomain_paths(struct tb *tb, struct tb_xdomain *xd)
  return 0;
 }
 
-static void add_switch(struct tb_switch *parent_sw, u64 route,
-       const uuid_t *uuid, const u8 *ep_name,
-       size_t ep_name_size, u8 connection_id, u8 connection_key,
-       u8 link, u8 depth, enum tb_security_level security_level,
-       bool authorized, bool boot)
+static struct tb_switch *add_switch(struct tb_switch *parent_sw, u64 route,
+    const uuid_t *uuid, const u8 *ep_name,
+    size_t ep_name_size, u8 connection_id,
+    u8 connection_key, u8 link, u8 depth,
+    enum tb_security_level security_level,
+    bool authorized, bool boot)
 {
  const struct intel_vss *vss;
  struct tb_switch *sw;
+ int ret;
 
  pm_runtime_get_sync(&parent_sw->dev);
 
@@ -557,14 +602,18 @@ static void add_switch(struct tb_switch *parent_sw, u64 route,
  tb_port_at(route, parent_sw)->remote = tb_upstream_port(sw);
  tb_upstream_port(sw)->remote = tb_port_at(route, parent_sw);
 
- if (tb_switch_add(sw)) {
+ ret = tb_switch_add(sw);
+ if (ret) {
  tb_port_at(tb_route(sw), parent_sw)->remote = NULL;
  tb_switch_put(sw);
+ sw = ERR_PTR(ret);
  }
 
 out:
  pm_runtime_mark_last_busy(&parent_sw->dev);
  pm_runtime_put_autosuspend(&parent_sw->dev);
+
+ return sw;
 }
 
 static void update_switch(struct tb_switch *parent_sw, struct tb_switch *sw,
@@ -656,6 +705,8 @@ icm_fr_device_connected(struct tb *tb, const struct icm_pkg_header *hdr)
  u64 route;
  int ret;
 
+ icm_postpone_rescan(tb);
+
  link = pkg->link_info & ICM_LINK_INFO_LINK_MASK;
  depth = (pkg->link_info & ICM_LINK_INFO_DEPTH_MASK) >>
  ICM_LINK_INFO_DEPTH_SHIFT;
@@ -1086,7 +1137,8 @@ static int icm_tr_disconnect_xdomain_paths(struct tb *tb, struct tb_xdomain *xd)
 }
 
 static void
-icm_tr_device_connected(struct tb *tb, const struct icm_pkg_header *hdr)
+__icm_tr_device_connected(struct tb *tb, const struct icm_pkg_header *hdr,
+  bool force_rtd3)
 {
  const struct icm_tr_event_device_connected *pkg =
  (const struct icm_tr_event_device_connected *)hdr;
@@ -1096,6 +1148,8 @@ icm_tr_device_connected(struct tb *tb, const struct icm_pkg_header *hdr)
  bool authorized, boot;
  u64 route;
 
+ icm_postpone_rescan(tb);
+
  /*
  * Currently we don't use the QoS information coming with the
  * device connected message so simply just ignore that extra
@@ -1151,13 +1205,21 @@ icm_tr_device_connected(struct tb *tb, const struct icm_pkg_header *hdr)
  return;
  }
 
- add_switch(parent_sw, route, &pkg->ep_uuid, (const u8 *)pkg->ep_name,
-   sizeof(pkg->ep_name), pkg->connection_id,
-   0, 0, 0, security_level, authorized, boot);
+ sw = add_switch(parent_sw, route, &pkg->ep_uuid, (const u8 *)pkg->ep_name,
+ sizeof(pkg->ep_name), pkg->connection_id, 0, 0, 0,
+ security_level, authorized, boot);
+ if (!IS_ERR(sw) && force_rtd3)
+ sw->rpm = true;
 
  tb_switch_put(parent_sw);
 }
 
+static void
+icm_tr_device_connected(struct tb *tb, const struct icm_pkg_header *hdr)
+{
+ __icm_tr_device_connected(tb, hdr, false);
+}
+
 static void
 icm_tr_device_disconnected(struct tb *tb, const struct icm_pkg_header *hdr)
 {
@@ -1468,6 +1530,61 @@ static int icm_ar_set_boot_acl(struct tb *tb, const uuid_t *uuids,
  return 0;
 }
 
+static int
+icm_icl_driver_ready(struct tb *tb, enum tb_security_level *security_level,
+    size_t *nboot_acl, bool *rpm)
+{
+ struct icm_tr_pkg_driver_ready_response reply;
+ struct icm_pkg_driver_ready request = {
+ .hdr.code = ICM_DRIVER_READY,
+ };
+ int ret;
+
+ memset(&reply, 0, sizeof(reply));
+ ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
+  1, 20000);
+ if (ret)
+ return ret;
+
+ /* Ice Lake always supports RTD3 */
+ if (rpm)
+ *rpm = true;
+
+ return 0;
+}
+
+static void icm_icl_set_uuid(struct tb *tb)
+{
+ struct tb_nhi *nhi = tb->nhi;
+ u32 uuid[4];
+
+ pci_read_config_dword(nhi->pdev, VS_CAP_10, &uuid[0]);
+ pci_read_config_dword(nhi->pdev, VS_CAP_11, &uuid[1]);
+ uuid[2] = 0xffffffff;
+ uuid[3] = 0xffffffff;
+
+ tb->root_switch->uuid = kmemdup(uuid, sizeof(uuid), GFP_KERNEL);
+}
+
+static void
+icm_icl_device_connected(struct tb *tb, const struct icm_pkg_header *hdr)
+{
+ __icm_tr_device_connected(tb, hdr, true);
+}
+
+static void icm_icl_rtd3_veto(struct tb *tb, const struct icm_pkg_header *hdr)
+{
+ const struct icm_icl_event_rtd3_veto *pkg =
+ (const struct icm_icl_event_rtd3_veto *)hdr;
+
+ tb_dbg(tb, "ICM rtd3 veto=0x%08x\n", pkg->veto_reason);
+
+ if (pkg->veto_reason)
+ icm_veto_begin(tb);
+ else
+ icm_veto_end(tb);
+}
+
 static void icm_handle_notification(struct work_struct *work)
 {
  struct icm_notification *n = container_of(work, typeof(*n), work);
@@ -1495,6 +1612,9 @@ static void icm_handle_notification(struct work_struct *work)
  case ICM_EVENT_XDOMAIN_DISCONNECTED:
  icm->xdomain_disconnected(tb, n->pkg);
  break;
+ case ICM_EVENT_RTD3_VETO:
+ icm->rtd3_veto(tb, n->pkg);
+ break;
  }
  }
 
@@ -1853,6 +1973,13 @@ static void icm_complete(struct tb *tb)
  if (tb->nhi->going_away)
  return;
 
+ /*
+ * If RTD3 was vetoed before we entered system suspend allow it
+ * again now before driver ready is sent. Firmware sends a new RTD3
+ * veto if it is still the case after we have sent it driver ready
+ * command.
+ */
+ icm_veto_end(tb);
  icm_unplug_children(tb->root_switch);
 
  /*
@@ -1918,6 +2045,9 @@ static int icm_start(struct tb *tb)
  tb->root_switch->no_nvm_upgrade = !icm->can_upgrade_nvm;
  tb->root_switch->rpm = icm->rpm;
 
+ if (icm->set_uuid)
+ icm->set_uuid(tb);
+
  ret = tb_switch_add(tb->root_switch);
  if (ret) {
  tb_switch_put(tb->root_switch);
@@ -2002,6 +2132,19 @@ static const struct tb_cm_ops icm_tr_ops = {
  .disconnect_xdomain_paths = icm_tr_disconnect_xdomain_paths,
 };
 
+/* Ice Lake */
+static const struct tb_cm_ops icm_icl_ops = {
+ .driver_ready = icm_driver_ready,
+ .start = icm_start,
+ .stop = icm_stop,
+ .complete = icm_complete,
+ .runtime_suspend = icm_runtime_suspend,
+ .runtime_resume = icm_runtime_resume,
+ .handle_event = icm_handle_event,
+ .approve_xdomain_paths = icm_tr_approve_xdomain_paths,
+ .disconnect_xdomain_paths = icm_tr_disconnect_xdomain_paths,
+};
+
 struct tb *icm_probe(struct tb_nhi *nhi)
 {
  struct icm *icm;
@@ -2070,6 +2213,19 @@ struct tb *icm_probe(struct tb_nhi *nhi)
  icm->xdomain_disconnected = icm_tr_xdomain_disconnected;
  tb->cm_ops = &icm_tr_ops;
  break;
+
+ case PCI_DEVICE_ID_INTEL_ICL_NHI0:
+ case PCI_DEVICE_ID_INTEL_ICL_NHI1:
+ icm->is_supported = icm_ar_is_supported;
+ icm->driver_ready = icm_icl_driver_ready;
+ icm->set_uuid = icm_icl_set_uuid;
+ icm->device_connected = icm_icl_device_connected;
+ icm->device_disconnected = icm_tr_device_disconnected;
+ icm->xdomain_connected = icm_tr_xdomain_connected;
+ icm->xdomain_disconnected = icm_tr_xdomain_disconnected;
+ icm->rtd3_veto = icm_icl_rtd3_veto;
+ tb->cm_ops = &icm_icl_ops;
+ break;
  }
 
  if (!icm->is_supported || !icm->is_supported(tb)) {
diff --git a/drivers/thunderbolt/nhi.c b/drivers/thunderbolt/nhi.c
index 9c782706e652..641b21b54460 100644
--- a/drivers/thunderbolt/nhi.c
+++ b/drivers/thunderbolt/nhi.c
@@ -16,6 +16,7 @@
 #include <linux/interrupt.h>
 #include <linux/module.h>
 #include <linux/delay.h>
+#include <linux/property.h>
 
 #include "nhi.h"
 #include "nhi_regs.h"
@@ -859,12 +860,52 @@ static irqreturn_t nhi_msi(int irq, void *data)
  return IRQ_HANDLED;
 }
 
-static int nhi_suspend_noirq(struct device *dev)
+static int __nhi_suspend_noirq(struct device *dev, bool wakeup)
 {
  struct pci_dev *pdev = to_pci_dev(dev);
  struct tb *tb = pci_get_drvdata(pdev);
+ struct tb_nhi *nhi = tb->nhi;
+ int ret;
+
+ ret = tb_domain_suspend_noirq(tb);
+ if (ret)
+ return ret;
+
+ if (nhi->ops && nhi->ops->suspend_noirq) {
+ ret = nhi->ops->suspend_noirq(tb->nhi, wakeup);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+static int nhi_suspend_noirq(struct device *dev)
+{
+ return __nhi_suspend_noirq(dev, device_may_wakeup(dev));
+}
+
+static bool nhi_wake_supported(struct pci_dev *pdev)
+{
+ u8 val;
+
+ /*
+ * If power rails are sustainable for wakeup from S4 this
+ * property is set by the BIOS.
+ */
+ if (device_property_read_u8(&pdev->dev, "WAKE_SUPPORTED", &val))
+ return !!val;
+
+ return true;
+}
+
+static int nhi_poweroff_noirq(struct device *dev)
+{
+ struct pci_dev *pdev = to_pci_dev(dev);
+ bool wakeup;
 
- return tb_domain_suspend_noirq(tb);
+ wakeup = device_may_wakeup(dev) && nhi_wake_supported(pdev);
+ return __nhi_suspend_noirq(dev, wakeup);
 }
 
 static void nhi_enable_int_throttling(struct tb_nhi *nhi)
@@ -887,16 +928,24 @@ static int nhi_resume_noirq(struct device *dev)
 {
  struct pci_dev *pdev = to_pci_dev(dev);
  struct tb *tb = pci_get_drvdata(pdev);
+ struct tb_nhi *nhi = tb->nhi;
+ int ret;
 
  /*
  * Check that the device is still there. It may be that the user
  * unplugged last device which causes the host controller to go
  * away on PCs.
  */
- if (!pci_device_is_present(pdev))
- tb->nhi->going_away = true;
- else
+ if (!pci_device_is_present(pdev)) {
+ nhi->going_away = true;
+ } else {
+ if (nhi->ops && nhi->ops->resume_noirq) {
+ ret = nhi->ops->resume_noirq(nhi);
+ if (ret)
+ return ret;
+ }
  nhi_enable_int_throttling(tb->nhi);
+ }
 
  return tb_domain_resume_noirq(tb);
 }
@@ -929,16 +978,35 @@ static int nhi_runtime_suspend(struct device *dev)
 {
  struct pci_dev *pdev = to_pci_dev(dev);
  struct tb *tb = pci_get_drvdata(pdev);
+ struct tb_nhi *nhi = tb->nhi;
+ int ret;
+
+ ret = tb_domain_runtime_suspend(tb);
+ if (ret)
+ return ret;
 
- return tb_domain_runtime_suspend(tb);
+ if (nhi->ops && nhi->ops->runtime_suspend) {
+ ret = nhi->ops->runtime_suspend(tb->nhi);
+ if (ret)
+ return ret;
+ }
+ return 0;
 }
 
 static int nhi_runtime_resume(struct device *dev)
 {
  struct pci_dev *pdev = to_pci_dev(dev);
  struct tb *tb = pci_get_drvdata(pdev);
+ struct tb_nhi *nhi = tb->nhi;
+ int ret;
+
+ if (nhi->ops && nhi->ops->runtime_resume) {
+ ret = nhi->ops->runtime_resume(nhi);
+ if (ret)
+ return ret;
+ }
 
- nhi_enable_int_throttling(tb->nhi);
+ nhi_enable_int_throttling(nhi);
  return tb_domain_runtime_resume(tb);
 }
 
@@ -966,6 +1034,9 @@ static void nhi_shutdown(struct tb_nhi *nhi)
  flush_work(&nhi->interrupt_work);
  }
  ida_destroy(&nhi->msix_ida);
+
+ if (nhi->ops && nhi->ops->shutdown)
+ nhi->ops->shutdown(nhi);
 }
 
 static int nhi_init_msi(struct tb_nhi *nhi)
@@ -1010,12 +1081,27 @@ static int nhi_init_msi(struct tb_nhi *nhi)
  return 0;
 }
 
+static bool nhi_imr_valid(struct pci_dev *pdev)
+{
+ u8 val;
+
+ if (!device_property_read_u8(&pdev->dev, "IMR_VALID", &val))
+ return !!val;
+
+ return true;
+}
+
 static int nhi_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 {
  struct tb_nhi *nhi;
  struct tb *tb;
  int res;
 
+ if (!nhi_imr_valid(pdev)) {
+ dev_warn(&pdev->dev, "firmware image not valid, aborting\n");
+ return -ENODEV;
+ }
+
  res = pcim_enable_device(pdev);
  if (res) {
  dev_err(&pdev->dev, "cannot enable PCI device, aborting\n");
@@ -1033,6 +1119,7 @@ static int nhi_probe(struct pci_dev *pdev, const struct pci_device_id *id)
  return -ENOMEM;
 
  nhi->pdev = pdev;
+ nhi->ops = (const struct tb_nhi_ops *)id->driver_data;
  /* cannot fail - table is allocated bin pcim_iomap_regions */
  nhi->iobase = pcim_iomap_table(pdev)[0];
  nhi->hop_count = ioread32(nhi->iobase + REG_HOP_COUNT) & 0x3ff;
@@ -1065,6 +1152,12 @@ static int nhi_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
  pci_set_master(pdev);
 
+ if (nhi->ops && nhi->ops->init) {
+ res = nhi->ops->init(nhi);
+ if (res)
+ return res;
+ }
+
  tb = icm_probe(nhi);
  if (!tb)
  tb = tb_probe(nhi);
@@ -1125,6 +1218,7 @@ static const struct dev_pm_ops nhi_pm_ops = {
  .restore_noirq = nhi_resume_noirq,
  .suspend = nhi_suspend,
  .freeze = nhi_suspend,
+ .poweroff_noirq = nhi_poweroff_noirq,
  .poweroff = nhi_suspend,
  .complete = nhi_complete,
  .runtime_suspend = nhi_runtime_suspend,
@@ -1172,6 +1266,10 @@ static struct pci_device_id nhi_ids[] = {
  { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_ALPINE_RIDGE_C_USBONLY_NHI) },
  { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_TITAN_RIDGE_2C_NHI) },
  { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_TITAN_RIDGE_4C_NHI) },
+ { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_ICL_NHI0),
+  .driver_data = (kernel_ulong_t)&icl_nhi_ops },
+ { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_ICL_NHI1),
+  .driver_data = (kernel_ulong_t)&icl_nhi_ops },
 
  { 0,}
 };
diff --git a/drivers/thunderbolt/nhi.h b/drivers/thunderbolt/nhi.h
index 1b5d47ecd3ed..b7b973949f8e 100644
--- a/drivers/thunderbolt/nhi.h
+++ b/drivers/thunderbolt/nhi.h
@@ -30,6 +30,26 @@ enum nhi_mailbox_cmd {
 int nhi_mailbox_cmd(struct tb_nhi *nhi, enum nhi_mailbox_cmd cmd, u32 data);
 enum nhi_fw_mode nhi_mailbox_mode(struct tb_nhi *nhi);
 
+/**
+ * struct tb_nhi_ops - NHI specific optional operations
+ * @init: NHI specific initialization
+ * @suspend_noirq: NHI specific suspend_noirq hook
+ * @resume_noirq: NHI specific resume_noirq hook
+ * @runtime_suspend: NHI specific runtime_suspend hook
+ * @runtime_resume: NHI specific runtime_resume hook
+ * @shutdown: NHI specific shutdown
+ */
+struct tb_nhi_ops {
+ int (*init)(struct tb_nhi *nhi);
+ int (*suspend_noirq)(struct tb_nhi *nhi, bool wakeup);
+ int (*resume_noirq)(struct tb_nhi *nhi);
+ int (*runtime_suspend)(struct tb_nhi *nhi);
+ int (*runtime_resume)(struct tb_nhi *nhi);
+ void (*shutdown)(struct tb_nhi *nhi);
+};
+
+extern const struct tb_nhi_ops icl_nhi_ops;
+
 /*
  * PCI IDs used in this driver from Win Ridge forward. There is no
  * need for the PCI quirk anymore as we will use ICM also on Apple
@@ -51,5 +71,7 @@ enum nhi_fw_mode nhi_mailbox_mode(struct tb_nhi *nhi);
 #define PCI_DEVICE_ID_INTEL_TITAN_RIDGE_4C_BRIDGE 0x15ea
 #define PCI_DEVICE_ID_INTEL_TITAN_RIDGE_4C_NHI 0x15eb
 #define PCI_DEVICE_ID_INTEL_TITAN_RIDGE_DD_BRIDGE 0x15ef
+#define PCI_DEVICE_ID_INTEL_ICL_NHI1 0x8a0d
+#define PCI_DEVICE_ID_INTEL_ICL_NHI0 0x8a17
 
 #endif
diff --git a/drivers/thunderbolt/nhi_ops.c b/drivers/thunderbolt/nhi_ops.c
new file mode 100644
index 000000000000..61cd09cef943
--- /dev/null
+++ b/drivers/thunderbolt/nhi_ops.c
@@ -0,0 +1,179 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * NHI specific operations
+ *
+ * Copyright (C) 2019, Intel Corporation
+ * Author: Mika Westerberg <[hidden email]>
+ */
+
+#include <linux/delay.h>
+#include <linux/suspend.h>
+
+#include "nhi.h"
+#include "nhi_regs.h"
+#include "tb.h"
+
+/* Ice Lake specific NHI operations */
+
+#define ICL_LC_MAILBOX_TIMEOUT 500 /* ms */
+
+static int check_for_device(struct device *dev, void *data)
+{
+ return tb_is_switch(dev);
+}
+
+static bool icl_nhi_is_device_connected(struct tb_nhi *nhi)
+{
+ struct tb *tb = pci_get_drvdata(nhi->pdev);
+ int ret;
+
+ ret = device_for_each_child(&tb->root_switch->dev, NULL,
+    check_for_device);
+ return ret > 0;
+}
+
+static int icl_nhi_force_power(struct tb_nhi *nhi, bool power)
+{
+ u32 vs_cap;
+
+ /*
+ * The Thunderbolt host controller is present always in Ice Lake
+ * but the firmware may not be loaded and running (depending
+ * whether there is device connected and so on). Each time the
+ * controller is used we need to "Force Power" it first and wait
+ * for the firmware to indicate it is up and running. This "Force
+ * Power" is really not about actually powering on/off the
+ * controller so it is accessible even if "Force Power" is off.
+ *
+ * The actual power management happens inside shared ACPI power
+ * resources using standard ACPI methods.
+ */
+ pci_read_config_dword(nhi->pdev, VS_CAP_22, &vs_cap);
+ if (power) {
+ vs_cap &= ~VS_CAP_22_DMA_DELAY_MASK;
+ vs_cap |= 0x22 << VS_CAP_22_DMA_DELAY_SHIFT;
+ vs_cap |= VS_CAP_22_FORCE_POWER;
+ } else {
+ vs_cap &= ~VS_CAP_22_FORCE_POWER;
+ }
+ pci_write_config_dword(nhi->pdev, VS_CAP_22, vs_cap);
+
+ if (power) {
+ unsigned int retries = 10;
+ u32 val;
+
+ /* Wait until the firmware tells it is up and running */
+ do {
+ pci_read_config_dword(nhi->pdev, VS_CAP_9, &val);
+ if (val & VS_CAP_9_FW_READY)
+ return 0;
+ msleep(250);
+ } while (--retries);
+
+ return -ETIMEDOUT;
+ }
+
+ return 0;
+}
+
+static void icl_nhi_lc_mailbox_cmd(struct tb_nhi *nhi, enum icl_lc_mailbox_cmd cmd)
+{
+ u32 data;
+
+ pci_read_config_dword(nhi->pdev, VS_CAP_19, &data);
+ data = (cmd << VS_CAP_19_CMD_SHIFT) & VS_CAP_19_CMD_MASK;
+ pci_write_config_dword(nhi->pdev, VS_CAP_19, data | VS_CAP_19_VALID);
+}
+
+static int icl_nhi_lc_mailbox_cmd_complete(struct tb_nhi *nhi, int timeout)
+{
+ unsigned long end;
+ u32 data;
+
+ if (!timeout)
+ goto clear;
+
+ end = jiffies + msecs_to_jiffies(timeout);
+ do {
+ pci_read_config_dword(nhi->pdev, VS_CAP_18, &data);
+ if (data & VS_CAP_18_DONE)
+ goto clear;
+ msleep(100);
+ } while (time_before(jiffies, end));
+
+ return -ETIMEDOUT;
+
+clear:
+ /* Clear the valid bit */
+ pci_write_config_dword(nhi->pdev, VS_CAP_19, 0);
+ return 0;
+}
+
+static void icl_nhi_set_ltr(struct tb_nhi *nhi)
+{
+ u32 max_ltr, ltr;
+
+ pci_read_config_dword(nhi->pdev, VS_CAP_16, &max_ltr);
+ max_ltr &= 0xffff;
+ /* Program the same value for both snoop and no-snoop */
+ ltr = max_ltr << 16 | max_ltr;
+ pci_write_config_dword(nhi->pdev, VS_CAP_15, ltr);
+}
+
+static int icl_nhi_suspend(struct tb_nhi *nhi)
+{
+ int ret;
+
+ if (icl_nhi_is_device_connected(nhi))
+ return 0;
+
+ /*
+ * If there is no device connected we need to perform both: a
+ * handshake through LC mailbox and force power down before
+ * entering D3.
+ */
+ icl_nhi_lc_mailbox_cmd(nhi, ICL_LC_PREPARE_FOR_RESET);
+ ret = icl_nhi_lc_mailbox_cmd_complete(nhi, ICL_LC_MAILBOX_TIMEOUT);
+ if (ret)
+ return ret;
+
+ return icl_nhi_force_power(nhi, false);
+}
+
+static int icl_nhi_suspend_noirq(struct tb_nhi *nhi, bool wakeup)
+{
+ enum icl_lc_mailbox_cmd cmd;
+
+ if (!pm_suspend_via_firmware())
+ return icl_nhi_suspend(nhi);
+
+ cmd = wakeup ? ICL_LC_GO2SX : ICL_LC_GO2SX_NO_WAKE;
+ icl_nhi_lc_mailbox_cmd(nhi, cmd);
+ return icl_nhi_lc_mailbox_cmd_complete(nhi, ICL_LC_MAILBOX_TIMEOUT);
+}
+
+static int icl_nhi_resume(struct tb_nhi *nhi)
+{
+ int ret;
+
+ ret = icl_nhi_force_power(nhi, true);
+ if (ret)
+ return ret;
+
+ icl_nhi_set_ltr(nhi);
+ return 0;
+}
+
+static void icl_nhi_shutdown(struct tb_nhi *nhi)
+{
+ icl_nhi_force_power(nhi, false);
+}
+
+const struct tb_nhi_ops icl_nhi_ops = {
+ .init = icl_nhi_resume,
+ .suspend_noirq = icl_nhi_suspend_noirq,
+ .resume_noirq = icl_nhi_resume,
+ .runtime_suspend = icl_nhi_suspend,
+ .runtime_resume = icl_nhi_resume,
+ .shutdown = icl_nhi_shutdown,
+};
diff --git a/drivers/thunderbolt/nhi_regs.h b/drivers/thunderbolt/nhi_regs.h
index a60bd98c1d04..0d4970dcef84 100644
--- a/drivers/thunderbolt/nhi_regs.h
+++ b/drivers/thunderbolt/nhi_regs.h
@@ -124,4 +124,41 @@ struct ring_desc {
 #define REG_FW_STS_ICM_EN_INVERT BIT(1)
 #define REG_FW_STS_ICM_EN BIT(0)
 
+/* ICL NHI VSEC registers */
+
+/* FW ready */
+#define VS_CAP_9 0xc8
+#define VS_CAP_9_FW_READY BIT(31)
+/* UUID */
+#define VS_CAP_10 0xcc
+#define VS_CAP_11 0xd0
+/* LTR */
+#define VS_CAP_15 0xe0
+#define VS_CAP_16 0xe4
+/* TBT2PCIe */
+#define VS_CAP_18 0xec
+#define VS_CAP_18_DONE BIT(0)
+/* PCIe2TBT */
+#define VS_CAP_19 0xf0
+#define VS_CAP_19_VALID BIT(0)
+#define VS_CAP_19_CMD_SHIFT 1
+#define VS_CAP_19_CMD_MASK GENMASK(7, 1)
+/* Force power */
+#define VS_CAP_22 0xfc
+#define VS_CAP_22_FORCE_POWER BIT(1)
+#define VS_CAP_22_DMA_DELAY_MASK GENMASK(31, 24)
+#define VS_CAP_22_DMA_DELAY_SHIFT 24
+
+/**
+ * enum icl_lc_mailbox_cmd - ICL specific LC mailbox commands
+ * @ICL_LC_GO2SX: Ask LC to enter Sx without wake
+ * @ICL_LC_GO2SX_NO_WAKE: Ask LC to enter Sx with wake
+ * @ICL_LC_PREPARE_FOR_RESET: Prepare LC for reset
+ */
+enum icl_lc_mailbox_cmd {
+ ICL_LC_GO2SX = 0x02,
+ ICL_LC_GO2SX_NO_WAKE = 0x03,
+ ICL_LC_PREPARE_FOR_RESET = 0x21,
+};
+
 #endif
diff --git a/drivers/thunderbolt/switch.c b/drivers/thunderbolt/switch.c
index 531f11fecf75..410bf1bceeee 100644
--- a/drivers/thunderbolt/switch.c
+++ b/drivers/thunderbolt/switch.c
@@ -1470,6 +1470,8 @@ static int tb_switch_get_generation(struct tb_switch *sw)
  case PCI_DEVICE_ID_INTEL_TITAN_RIDGE_2C_BRIDGE:
  case PCI_DEVICE_ID_INTEL_TITAN_RIDGE_4C_BRIDGE:
  case PCI_DEVICE_ID_INTEL_TITAN_RIDGE_DD_BRIDGE:
+ case PCI_DEVICE_ID_INTEL_ICL_NHI0:
+ case PCI_DEVICE_ID_INTEL_ICL_NHI1:
  return 3;
 
  default:
diff --git a/drivers/thunderbolt/tb_msgs.h b/drivers/thunderbolt/tb_msgs.h
index afbe1d29bb03..4b641e4ee0c5 100644
--- a/drivers/thunderbolt/tb_msgs.h
+++ b/drivers/thunderbolt/tb_msgs.h
@@ -104,10 +104,11 @@ enum icm_pkg_code {
 };
 
 enum icm_event_code {
- ICM_EVENT_DEVICE_CONNECTED = 3,
- ICM_EVENT_DEVICE_DISCONNECTED = 4,
- ICM_EVENT_XDOMAIN_CONNECTED = 6,
- ICM_EVENT_XDOMAIN_DISCONNECTED = 7,
+ ICM_EVENT_DEVICE_CONNECTED = 0x3,
+ ICM_EVENT_DEVICE_DISCONNECTED = 0x4,
+ ICM_EVENT_XDOMAIN_CONNECTED = 0x6,
+ ICM_EVENT_XDOMAIN_DISCONNECTED = 0x7,
+ ICM_EVENT_RTD3_VETO = 0xa,
 };
 
 struct icm_pkg_header {
@@ -463,6 +464,13 @@ struct icm_tr_pkg_disconnect_xdomain_response {
  uuid_t remote_uuid;
 };
 
+/* Ice Lake messages */
+
+struct icm_icl_event_rtd3_veto {
+ struct icm_pkg_header hdr;
+ u32 veto_reason;
+};
+
 /* XDomain messages */
 
 struct tb_xdomain_header {
diff --git a/include/linux/thunderbolt.h b/include/linux/thunderbolt.h
index 2d7e012db03f..ece782ef5466 100644
--- a/include/linux/thunderbolt.h
+++ b/include/linux/thunderbolt.h
@@ -429,6 +429,7 @@ static inline struct tb_xdomain *tb_service_parent(struct tb_service *svc)
  * @lock: Must be held during ring creation/destruction. Is acquired by
  *  interrupt_work when dispatching interrupts to individual rings.
  * @pdev: Pointer to the PCI device
+ * @ops: NHI specific optional ops
  * @iobase: MMIO space of the NHI
  * @tx_rings: All Tx rings available on this host controller
  * @rx_rings: All Rx rings available on this host controller
@@ -442,6 +443,7 @@ static inline struct tb_xdomain *tb_service_parent(struct tb_service *svc)
 struct tb_nhi {
  spinlock_t lock;
  struct pci_dev *pdev;
+ const struct tb_nhi_ops *ops;
  void __iomem *iobase;
  struct tb_ring **tx_rings;
  struct tb_ring **rx_rings;
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[PATCH 8/8] ACPI / property: Add two new Thunderbolt property GUIDs to the list

Timo Aaltonen-6
In reply to this post by Timo Aaltonen-6
From: Mika Westerberg <[hidden email]>

BugLink: http://bugs.launchpad.net/bugs/1844680

Ice Lake Thunderbolt controller includes two new device property
compatible properties that we need to be able to extract in the driver
so add them to the growing array of GUIDs.

Signed-off-by: Mika Westerberg <[hidden email]>
Acked-by: Rafael J. Wysocki <[hidden email]>
(cherry picked from commit dfda204198848b47bdb98ab83b94dbb7c7692b55)
Signed-off-by: Timo Aaltonen <[hidden email]>
---
 drivers/acpi/property.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c
index ea3d700da3ca..e095334eaec9 100644
--- a/drivers/acpi/property.c
+++ b/drivers/acpi/property.c
@@ -39,6 +39,12 @@ static const guid_t prp_guids[] = {
  /* External facing port GUID: efcc06cc-73ac-4bc3-bff0-76143807c389 */
  GUID_INIT(0xefcc06cc, 0x73ac, 0x4bc3,
   0xbf, 0xf0, 0x76, 0x14, 0x38, 0x07, 0xc3, 0x89),
+ /* Thunderbolt GUID for IMR_VALID: c44d002f-69f9-4e7d-a904-a7baabdf43f7 */
+ GUID_INIT(0xc44d002f, 0x69f9, 0x4e7d,
+  0xa9, 0x04, 0xa7, 0xba, 0xab, 0xdf, 0x43, 0xf7),
+ /* Thunderbolt GUID for WAKE_SUPPORTED: 6c501103-c189-4296-ba72-9bf5a26ebe5d */
+ GUID_INIT(0x6c501103, 0xc189, 0x4296,
+  0xba, 0x72, 0x9b, 0xf5, 0xa2, 0x6e, 0xbe, 0x5d),
 };
 
 /* ACPI _DSD data subnodes GUID: dbb8e3e6-5886-4ba6-8795-1319f52a966b */
--
2.20.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

APPLIED: [E][unstable][PATCH 0/8] thunderbolt: ICL support

Seth Forshee
In reply to this post by Timo Aaltonen-6
On Wed, Sep 25, 2019 at 01:05:53PM +0300, Timo Aaltonen wrote:
> BugLink: http://bugs.launchpad.net/bugs/1844680
>
> These patches from 5.4 add Thunderbolt support for Ice Lake.

Applied to eoan/master-next. Unstable is currently inactive until
5.4-rc1 comes out. Thanks!

--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team