Merge branch 'pci/pm'

- Cache the PTM capability offset instead of searching for it every time
  (Bjorn Helgaas)

- Separate PTM configuration from PTM enable (Bjorn Helgaas)

- Add pci_suspend_ptm() and pci_resume_ptm() to disable and re-enable PTM
  on suspend/resume so some Root Ports can safely enter a lower-power PM
  state (Bjorn Helgaas)

- Disable PTM for all devices during suspend; previously we only did this
  for Root Ports and even then only in certain cases (Bjorn Helgaas)

- Simplify pci_pm_suspend_noirq() (Rajvi Jingar)

- Reduce the delay after transitions to/from D3hot by using usleep_range()
  instead of msleep(), which reduces the typical delay from 19ms to 10ms
  (Sajid Dalvi, Will McVicker)

* pci/pm:
  PCI/PM: Reduce D3hot delay with usleep_range()
  PCI/PM: Simplify pci_pm_suspend_noirq()
  PCI/PM: Always disable PTM for all devices during suspend
  PCI/PTM: Consolidate PTM interface declarations
  PCI/PTM: Reorder functions in logical order
  PCI/PTM: Preserve RsvdP bits in PTM Control register
  PCI/PTM: Move pci_ptm_info() body into its only caller
  PCI/PTM: Add pci_suspend_ptm() and pci_resume_ptm()
  PCI/PTM: Separate configuration and enable
  PCI/PTM: Add pci_upstream_ptm() helper
  PCI/PTM: Cache PTM Capability offset

4 files changed, 205 insertions, 183 deletions

diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
index 49238ddd39ee..107d77f3c846 100644
--- a/drivers/pci/pci-driver.c
+++ b/drivers/pci/pci-driver.c
@@ -774,6 +774,12 @@ static int pci_pm_suspend(struct device *dev)
 
 	pci_dev->skip_bus_pm = false;
 
+	/*
+	 * Disabling PTM allows some systems, e.g., Intel mobile chips
+	 * since Coffee Lake, to enter a lower-power PM state.
+	 */
+	pci_suspend_ptm(pci_dev);
+
 	if (pci_has_legacy_pm_support(pci_dev))
 		return pci_legacy_suspend(dev, PMSG_SUSPEND);
 
@@ -867,20 +873,15 @@ static int pci_pm_suspend_noirq(struct device *dev)
 		}
 	}
 
-	if (pci_dev->skip_bus_pm) {
+	if (!pci_dev->state_saved) {
+		pci_save_state(pci_dev);
+
 		/*
-		 * Either the device is a bridge with a child in D0 below it, or
-		 * the function is running for the second time in a row without
-		 * going through full resume, which is possible only during
-		 * suspend-to-idle in a spurious wakeup case.  The device should
-		 * be in D0 at this point, but if it is a bridge, it may be
-		 * necessary to save its state.
+		 * If the device is a bridge with a child in D0 below it,
+		 * it needs to stay in D0, so check skip_bus_pm to avoid
+		 * putting it into a low-power state in that case.
 		 */
-		if (!pci_dev->state_saved)
-			pci_save_state(pci_dev);
-	} else if (!pci_dev->state_saved) {
-		pci_save_state(pci_dev);
-		if (pci_power_manageable(pci_dev))
+		if (!pci_dev->skip_bus_pm && pci_power_manageable(pci_dev))
 			pci_prepare_to_sleep(pci_dev);
 	}
 
@@ -987,6 +988,8 @@ static int pci_pm_resume(struct device *dev)
 	if (pci_dev->state_saved)
 		pci_restore_standard_config(pci_dev);
 
+	pci_resume_ptm(pci_dev);
+
 	if (pci_has_legacy_pm_support(pci_dev))
 		return pci_legacy_resume(dev);
 
@@ -1274,6 +1277,8 @@ static int pci_pm_runtime_suspend(struct device *dev)
 	pci_power_t prev = pci_dev->current_state;
 	int error;
 
+	pci_suspend_ptm(pci_dev);
+
 	/*
 	 * If pci_dev->driver is not set (unbound), we leave the device in D0,
 	 * but it may go to D3cold when the bridge above it runtime suspends.
@@ -1335,6 +1340,7 @@ static int pci_pm_runtime_resume(struct device *dev)
 	 * D3cold when the bridge above it runtime suspended.
 	 */
 	pci_pm_default_resume_early(pci_dev);
+	pci_resume_ptm(pci_dev);
 
 	if (!pci_dev->driver)
 		return 0;
diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index 68a49fbaabde..2127aba3550b 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -66,13 +66,15 @@ struct pci_pme_device {
 
 static void pci_dev_d3_sleep(struct pci_dev *dev)
 {
-	unsigned int delay = dev->d3hot_delay;
-
-	if (delay < pci_pm_d3hot_delay)
-		delay = pci_pm_d3hot_delay;
-
-	if (delay)
-		msleep(delay);
+	unsigned int delay_ms = max(dev->d3hot_delay, pci_pm_d3hot_delay);
+	unsigned int upper;
+
+	if (delay_ms) {
+		/* Use a 20% upper bound, 1ms minimum */
+		upper = max(DIV_ROUND_CLOSEST(delay_ms, 5), 1U);
+		usleep_range(delay_ms * USEC_PER_MSEC,
+			     (delay_ms + upper) * USEC_PER_MSEC);
+	}
 }
 
 bool pci_reset_supported(struct pci_dev *dev)
@@ -2708,24 +2710,12 @@ int pci_prepare_to_sleep(struct pci_dev *dev)
 	if (target_state == PCI_POWER_ERROR)
 		return -EIO;
 
-	/*
-	 * There are systems (for example, Intel mobile chips since Coffee
-	 * Lake) where the power drawn while suspended can be significantly
-	 * reduced by disabling PTM on PCIe root ports as this allows the
-	 * port to enter a lower-power PM state and the SoC to reach a
-	 * lower-power idle state as a whole.
-	 */
-	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT)
-		pci_disable_ptm(dev);
-
 	pci_enable_wake(dev, target_state, wakeup);
 
 	error = pci_set_power_state(dev, target_state);
 
-	if (error) {
+	if (error)
 		pci_enable_wake(dev, target_state, false);
-		pci_restore_ptm_state(dev);
-	}
 
 	return error;
 }
@@ -2766,24 +2756,12 @@ int pci_finish_runtime_suspend(struct pci_dev *dev)
 	if (target_state == PCI_POWER_ERROR)
 		return -EIO;
 
-	/*
-	 * There are systems (for example, Intel mobile chips since Coffee
-	 * Lake) where the power drawn while suspended can be significantly
-	 * reduced by disabling PTM on PCIe root ports as this allows the
-	 * port to enter a lower-power PM state and the SoC to reach a
-	 * lower-power idle state as a whole.
-	 */
-	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT)
-		pci_disable_ptm(dev);
-
 	__pci_enable_wake(dev, target_state, pci_dev_run_wake(dev));
 
 	error = pci_set_power_state(dev, target_state);
 
-	if (error) {
+	if (error)
 		pci_enable_wake(dev, target_state, false);
-		pci_restore_ptm_state(dev);
-	}
 
 	return error;
 }
diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h
index 365a844ec430..46af1195c047 100644
--- a/drivers/pci/pci.h
+++ b/drivers/pci/pci.h
@@ -505,13 +505,17 @@ static inline int pci_iov_bus_range(struct pci_bus *bus)
 #endif /* CONFIG_PCI_IOV */
 
 #ifdef CONFIG_PCIE_PTM
+void pci_ptm_init(struct pci_dev *dev);
 void pci_save_ptm_state(struct pci_dev *dev);
 void pci_restore_ptm_state(struct pci_dev *dev);
-void pci_disable_ptm(struct pci_dev *dev);
+void pci_suspend_ptm(struct pci_dev *dev);
+void pci_resume_ptm(struct pci_dev *dev);
 #else
+static inline void pci_ptm_init(struct pci_dev *dev) { }
 static inline void pci_save_ptm_state(struct pci_dev *dev) { }
 static inline void pci_restore_ptm_state(struct pci_dev *dev) { }
-static inline void pci_disable_ptm(struct pci_dev *dev) { }
+static inline void pci_suspend_ptm(struct pci_dev *dev) { }
+static inline void pci_resume_ptm(struct pci_dev *dev) { }
 #endif
 
 unsigned long pci_cardbus_resource_alignment(struct resource *);
@@ -579,12 +583,6 @@ static inline void pcie_set_ecrc_checking(struct pci_dev *dev) { }
 static inline void pcie_ecrc_get_policy(char *str) { }
 #endif
 
-#ifdef CONFIG_PCIE_PTM
-void pci_ptm_init(struct pci_dev *dev);
-#else
-static inline void pci_ptm_init(struct pci_dev *dev) { }
-#endif
-
 struct pci_dev_reset_methods {
 	u16 vendor;
 	u16 device;
diff --git a/drivers/pci/pcie/ptm.c b/drivers/pci/pcie/ptm.c
index 368a254e3124..b4e5f553467c 100644
--- a/drivers/pci/pcie/ptm.c
+++ b/drivers/pci/pcie/ptm.c
@@ -9,30 +9,38 @@
 #include <linux/pci.h>
 #include "../pci.h"
 
-static void pci_ptm_info(struct pci_dev *dev)
+/*
+ * If the next upstream device supports PTM, return it; otherwise return
+ * NULL.  PTM Messages are local, so both link partners must support it.
+ */
+static struct pci_dev *pci_upstream_ptm(struct pci_dev *dev)
 {
-	char clock_desc[8];
+	struct pci_dev *ups = pci_upstream_bridge(dev);
 
-	switch (dev->ptm_granularity) {
-	case 0:
-		snprintf(clock_desc, sizeof(clock_desc), "unknown");
-		break;
-	case 255:
-		snprintf(clock_desc, sizeof(clock_desc), ">254ns");
-		break;
-	default:
-		snprintf(clock_desc, sizeof(clock_desc), "%uns",
-			 dev->ptm_granularity);
-		break;
-	}
-	pci_info(dev, "PTM enabled%s, %s granularity\n",
-		 dev->ptm_root ? " (root)" : "", clock_desc);
+	/*
+	 * Switch Downstream Ports are not permitted to have a PTM
+	 * capability; their PTM behavior is controlled by the Upstream
+	 * Port (PCIe r5.0, sec 7.9.16), so if the upstream bridge is a
+	 * Switch Downstream Port, look up one more level.
+	 */
+	if (ups && pci_pcie_type(ups) == PCI_EXP_TYPE_DOWNSTREAM)
+		ups = pci_upstream_bridge(ups);
+
+	if (ups && ups->ptm_cap)
+		return ups;
+
+	return NULL;
 }
 
-void pci_disable_ptm(struct pci_dev *dev)
+/*
+ * Find the PTM Capability (if present) and extract the information we need
+ * to use it.
+ */
+void pci_ptm_init(struct pci_dev *dev)
 {
-	int ptm;
-	u16 ctrl;
+	u16 ptm;
+	u32 cap;
+	struct pci_dev *ups;
 
 	if (!pci_is_pcie(dev))
 		return;
@@ -41,21 +49,47 @@ void pci_disable_ptm(struct pci_dev *dev)
 	if (!ptm)
 		return;
 
-	pci_read_config_word(dev, ptm + PCI_PTM_CTRL, &ctrl);
-	ctrl &= ~(PCI_PTM_CTRL_ENABLE | PCI_PTM_CTRL_ROOT);
-	pci_write_config_word(dev, ptm + PCI_PTM_CTRL, ctrl);
+	dev->ptm_cap = ptm;
+	pci_add_ext_cap_save_buffer(dev, PCI_EXT_CAP_ID_PTM, sizeof(u32));
+
+	pci_read_config_dword(dev, ptm + PCI_PTM_CAP, &cap);
+	dev->ptm_granularity = (cap & PCI_PTM_GRANULARITY_MASK) >> 8;
+
+	/*
+	 * Per the spec recommendation (PCIe r6.0, sec 7.9.15.3), select the
+	 * furthest upstream Time Source as the PTM Root.  For Endpoints,
+	 * "the Effective Granularity is the maximum Local Clock Granularity
+	 * reported by the PTM Root and all intervening PTM Time Sources."
+	 */
+	ups = pci_upstream_ptm(dev);
+	if (ups) {
+		if (ups->ptm_granularity == 0)
+			dev->ptm_granularity = 0;
+		else if (ups->ptm_granularity > dev->ptm_granularity)
+			dev->ptm_granularity = ups->ptm_granularity;
+	} else if (cap & PCI_PTM_CAP_ROOT) {
+		dev->ptm_root = 1;
+	} else if (pci_pcie_type(dev) == PCI_EXP_TYPE_RC_END) {
+
+		/*
+		 * Per sec 7.9.15.3, this should be the Local Clock
+		 * Granularity of the associated Time Source.  But it
+		 * doesn't say how to find that Time Source.
+		 */
+		dev->ptm_granularity = 0;
+	}
+
+	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT ||
+	    pci_pcie_type(dev) == PCI_EXP_TYPE_UPSTREAM)
+		pci_enable_ptm(dev, NULL);
 }
 
 void pci_save_ptm_state(struct pci_dev *dev)
 {
-	int ptm;
+	u16 ptm = dev->ptm_cap;
 	struct pci_cap_saved_state *save_state;
-	u16 *cap;
+	u32 *cap;
 
-	if (!pci_is_pcie(dev))
-		return;
-
-	ptm = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_PTM);
 	if (!ptm)
 		return;
 
@@ -63,146 +97,152 @@ void pci_save_ptm_state(struct pci_dev *dev)
 	if (!save_state)
 		return;
 
-	cap = (u16 *)&save_state->cap.data[0];
-	pci_read_config_word(dev, ptm + PCI_PTM_CTRL, cap);
+	cap = (u32 *)&save_state->cap.data[0];
+	pci_read_config_dword(dev, ptm + PCI_PTM_CTRL, cap);
 }
 
 void pci_restore_ptm_state(struct pci_dev *dev)
 {
+	u16 ptm = dev->ptm_cap;
 	struct pci_cap_saved_state *save_state;
-	int ptm;
-	u16 *cap;
+	u32 *cap;
 
-	if (!pci_is_pcie(dev))
+	if (!ptm)
 		return;
 
 	save_state = pci_find_saved_ext_cap(dev, PCI_EXT_CAP_ID_PTM);
-	ptm = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_PTM);
-	if (!save_state || !ptm)
+	if (!save_state)
 		return;
 
-	cap = (u16 *)&save_state->cap.data[0];
-	pci_write_config_word(dev, ptm + PCI_PTM_CTRL, *cap);
+	cap = (u32 *)&save_state->cap.data[0];
+	pci_write_config_dword(dev, ptm + PCI_PTM_CTRL, *cap);
 }
 
-void pci_ptm_init(struct pci_dev *dev)
+/* Enable PTM in the Control register if possible */
+static int __pci_enable_ptm(struct pci_dev *dev)
 {
-	int pos;
-	u32 cap, ctrl;
-	u8 local_clock;
+	u16 ptm = dev->ptm_cap;
 	struct pci_dev *ups;
+	u32 ctrl;
 
-	if (!pci_is_pcie(dev))
-		return;
-
-	/*
-	 * Enable PTM only on interior devices (root ports, switch ports,
-	 * etc.) on the assumption that it causes no link traffic until an
-	 * endpoint enables it.
-	 */
-	if ((pci_pcie_type(dev) == PCI_EXP_TYPE_ENDPOINT ||
-	     pci_pcie_type(dev) == PCI_EXP_TYPE_RC_END))
-		return;
+	if (!ptm)
+		return -EINVAL;
 
 	/*
-	 * Switch Downstream Ports are not permitted to have a PTM
-	 * capability; their PTM behavior is controlled by the Upstream
-	 * Port (PCIe r5.0, sec 7.9.16).
+	 * A device uses local PTM Messages to request time information
+	 * from a PTM Root that's farther upstream.  Every device along the
+	 * path must support PTM and have it enabled so it can handle the
+	 * messages.  Therefore, if this device is not a PTM Root, the
+	 * upstream link partner must have PTM enabled before we can enable
+	 * PTM.
 	 */
-	ups = pci_upstream_bridge(dev);
-	if (pci_pcie_type(dev) == PCI_EXP_TYPE_DOWNSTREAM &&
-	    ups && ups->ptm_enabled) {
-		dev->ptm_granularity = ups->ptm_granularity;
-		dev->ptm_enabled = 1;
-		return;
+	if (!dev->ptm_root) {
+		ups = pci_upstream_ptm(dev);
+		if (!ups || !ups->ptm_enabled)
+			return -EINVAL;
 	}
 
-	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_PTM);
-	if (!pos)
-		return;
-
-	pci_add_ext_cap_save_buffer(dev, PCI_EXT_CAP_ID_PTM, sizeof(u16));
-
-	pci_read_config_dword(dev, pos + PCI_PTM_CAP, &cap);
-	local_clock = (cap & PCI_PTM_GRANULARITY_MASK) >> 8;
-
-	/*
-	 * There's no point in enabling PTM unless it's enabled in the
-	 * upstream device or this device can be a PTM Root itself.  Per
-	 * the spec recommendation (PCIe r3.1, sec 7.32.3), select the
-	 * furthest upstream Time Source as the PTM Root.
-	 */
-	if (ups && ups->ptm_enabled) {
-		ctrl = PCI_PTM_CTRL_ENABLE;
-		if (ups->ptm_granularity == 0)
-			dev->ptm_granularity = 0;
-		else if (ups->ptm_granularity > local_clock)
-			dev->ptm_granularity = ups->ptm_granularity;
-	} else {
-		if (cap & PCI_PTM_CAP_ROOT) {
-			ctrl = PCI_PTM_CTRL_ENABLE | PCI_PTM_CTRL_ROOT;
-			dev->ptm_root = 1;
-			dev->ptm_granularity = local_clock;
-		} else
-			return;
-	}
+	pci_read_config_dword(dev, ptm + PCI_PTM_CTRL, &ctrl);
 
+	ctrl |= PCI_PTM_CTRL_ENABLE;
+	ctrl &= ~PCI_PTM_GRANULARITY_MASK;
 	ctrl |= dev->ptm_granularity << 8;
-	pci_write_config_dword(dev, pos + PCI_PTM_CTRL, ctrl);
-	dev->ptm_enabled = 1;
+	if (dev->ptm_root)
+		ctrl |= PCI_PTM_CTRL_ROOT;
 
-	pci_ptm_info(dev);
+	pci_write_config_dword(dev, ptm + PCI_PTM_CTRL, ctrl);
+	return 0;
 }
 
+/**
+ * pci_enable_ptm() - Enable Precision Time Measurement
+ * @dev: PCI device
+ * @granularity: pointer to return granularity
+ *
+ * Enable Precision Time Measurement for @dev.  If successful and
+ * @granularity is non-NULL, return the Effective Granularity.
+ *
+ * Return: zero if successful, or -EINVAL if @dev lacks a PTM Capability or
+ * is not a PTM Root and lacks an upstream path of PTM-enabled devices.
+ */
 int pci_enable_ptm(struct pci_dev *dev, u8 *granularity)
 {
-	int pos;
-	u32 cap, ctrl;
-	struct pci_dev *ups;
-
-	if (!pci_is_pcie(dev))
-		return -EINVAL;
-
-	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_PTM);
-	if (!pos)
-		return -EINVAL;
-
-	pci_read_config_dword(dev, pos + PCI_PTM_CAP, &cap);
-	if (!(cap & PCI_PTM_CAP_REQ))
-		return -EINVAL;
-
-	/*
-	 * For a PCIe Endpoint, PTM is only useful if the endpoint can
-	 * issue PTM requests to upstream devices that have PTM enabled.
-	 *
-	 * For Root Complex Integrated Endpoints, there is no upstream
-	 * device, so there must be some implementation-specific way to
-	 * associate the endpoint with a time source.
-	 */
-	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ENDPOINT) {
-		ups = pci_upstream_bridge(dev);
-		if (!ups || !ups->ptm_enabled)
-			return -EINVAL;
+	int rc;
+	char clock_desc[8];
 
-		dev->ptm_granularity = ups->ptm_granularity;
-	} else if (pci_pcie_type(dev) == PCI_EXP_TYPE_RC_END) {
-		dev->ptm_granularity = 0;
-	} else
-		return -EINVAL;
+	rc = __pci_enable_ptm(dev);
+	if (rc)
+		return rc;
 
-	ctrl = PCI_PTM_CTRL_ENABLE;
-	ctrl |= dev->ptm_granularity << 8;
-	pci_write_config_dword(dev, pos + PCI_PTM_CTRL, ctrl);
 	dev->ptm_enabled = 1;
 
-	pci_ptm_info(dev);
-
 	if (granularity)
 		*granularity = dev->ptm_granularity;
+
+	switch (dev->ptm_granularity) {
+	case 0:
+		snprintf(clock_desc, sizeof(clock_desc), "unknown");
+		break;
+	case 255:
+		snprintf(clock_desc, sizeof(clock_desc), ">254ns");
+		break;
+	default:
+		snprintf(clock_desc, sizeof(clock_desc), "%uns",
+			 dev->ptm_granularity);
+		break;
+	}
+	pci_info(dev, "PTM enabled%s, %s granularity\n",
+		 dev->ptm_root ? " (root)" : "", clock_desc);
+
 	return 0;
 }
 EXPORT_SYMBOL(pci_enable_ptm);
 
+static void __pci_disable_ptm(struct pci_dev *dev)
+{
+	u16 ptm = dev->ptm_cap;
+	u32 ctrl;
+
+	if (!ptm)
+		return;
+
+	pci_read_config_dword(dev, ptm + PCI_PTM_CTRL, &ctrl);
+	ctrl &= ~(PCI_PTM_CTRL_ENABLE | PCI_PTM_CTRL_ROOT);
+	pci_write_config_dword(dev, ptm + PCI_PTM_CTRL, ctrl);
+}
+
+/**
+ * pci_disable_ptm() - Disable Precision Time Measurement
+ * @dev: PCI device
+ *
+ * Disable Precision Time Measurement for @dev.
+ */
+void pci_disable_ptm(struct pci_dev *dev)
+{
+	if (dev->ptm_enabled) {
+		__pci_disable_ptm(dev);
+		dev->ptm_enabled = 0;
+	}
+}
+EXPORT_SYMBOL(pci_disable_ptm);
+
+/*
+ * Disable PTM, but preserve dev->ptm_enabled so we silently re-enable it on
+ * resume if necessary.
+ */
+void pci_suspend_ptm(struct pci_dev *dev)
+{
+	if (dev->ptm_enabled)
+		__pci_disable_ptm(dev);
+}
+
+/* If PTM was enabled before suspend, re-enable it when resuming */
+void pci_resume_ptm(struct pci_dev *dev)
+{
+	if (dev->ptm_enabled)
+		__pci_enable_ptm(dev);
+}
+
 bool pcie_ptm_enabled(struct pci_dev *dev)
 {
 	if (!dev)