public inbox for linux-pm@vger.kernel.org
 help / color / mirror / Atom feed
From: Bjorn Helgaas <helgaas@kernel.org>
To: Kai-Heng Feng <kai.heng.feng@canonical.com>,
	Rajvi Jingar <rajvi.jingar@linux.intel.com>,
	"Rafael J . Wysocki" <rafael@kernel.org>
Cc: Koba Ko <koba.ko@canonical.com>,
	Mika Westerberg <mika.westerberg@linux.intel.com>,
	"David E . Box" <david.e.box@linux.intel.com>,
	Sathyanarayanan Kuppuswamy 
	<sathyanarayanan.kuppuswamy@linux.intel.com>,
	linux-pci@vger.kernel.org, linux-pm@vger.kernel.org,
	linux-kernel@vger.kernel.org, Bjorn Helgaas <bhelgaas@google.com>
Subject: [PATCH v4 3/9] PCI/PTM: Separate configuration and enable
Date: Fri,  9 Sep 2022 15:24:59 -0500	[thread overview]
Message-ID: <20220909202505.314195-4-helgaas@kernel.org> (raw)
In-Reply-To: <20220909202505.314195-1-helgaas@kernel.org>

From: Bjorn Helgaas <bhelgaas@google.com>

PTM configuration and enabling were previously mixed together:
pci_ptm_init() collected granularity info and enabled PTM for Root Ports
and Switch Upstream Ports; pci_enable_ptm() did the same for Endpoints.

Move everything related to the PTM Capability register to pci_ptm_init()
for all devices, and everything related to the PTM Control register to
pci_enable_ptm().

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 drivers/pci/pcie/ptm.c | 104 +++++++++++++++++++----------------------
 1 file changed, 49 insertions(+), 55 deletions(-)

diff --git a/drivers/pci/pcie/ptm.c b/drivers/pci/pcie/ptm.c
index 0df6cdfe38b4..ba1d50c965fa 100644
--- a/drivers/pci/pcie/ptm.c
+++ b/drivers/pci/pcie/ptm.c
@@ -99,25 +99,19 @@ static struct pci_dev *pci_upstream_ptm(struct pci_dev *dev)
 	return NULL;
 }
 
+/*
+ * Find the PTM Capability (if present) and extract the information we need
+ * to use it.
+ */
 void pci_ptm_init(struct pci_dev *dev)
 {
 	u16 ptm;
-	u32 cap, ctrl;
-	u8 local_clock;
+	u32 cap;
 	struct pci_dev *ups;
 
 	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;
-
 	ptm = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_PTM);
 	if (!ptm)
 		return;
@@ -126,76 +120,76 @@ void pci_ptm_init(struct pci_dev *dev)
 	pci_add_ext_cap_save_buffer(dev, PCI_EXT_CAP_ID_PTM, sizeof(u16));
 
 	pci_read_config_dword(dev, ptm + PCI_PTM_CAP, &cap);
-	local_clock = (cap & PCI_PTM_GRANULARITY_MASK) >> 8;
+	dev->ptm_granularity = (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.
+	 * 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 && ups->ptm_enabled) {
-		ctrl = PCI_PTM_CTRL_ENABLE;
+	if (ups) {
 		if (ups->ptm_granularity == 0)
 			dev->ptm_granularity = 0;
-		else if (ups->ptm_granularity > local_clock)
+		else if (ups->ptm_granularity > dev->ptm_granularity)
 			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;
+	} 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;
 	}
 
-	ctrl |= dev->ptm_granularity << 8;
-	pci_write_config_dword(dev, ptm + PCI_PTM_CTRL, ctrl);
-	dev->ptm_enabled = 1;
-
-	pci_ptm_info(dev);
+	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT ||
+	    pci_pcie_type(dev) == PCI_EXP_TYPE_UPSTREAM)
+		pci_enable_ptm(dev, NULL);
 }
 
+/**
+ * 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)
 {
-	u16 ptm;
-	u32 cap, ctrl;
+	u16 ptm = dev->ptm_cap;
 	struct pci_dev *ups;
+	u32 ctrl;
 
-	if (!pci_is_pcie(dev))
-		return -EINVAL;
-
-	ptm = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_PTM);
 	if (!ptm)
 		return -EINVAL;
 
-	dev->ptm_cap = ptm;
-	pci_read_config_dword(dev, ptm + 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.
+	 * 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.
 	 */
-	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ENDPOINT) {
+	if (!dev->ptm_root) {
 		ups = pci_upstream_ptm(dev);
 		if (!ups || !ups->ptm_enabled)
 			return -EINVAL;
-
-		dev->ptm_granularity = ups->ptm_granularity;
-	} else if (pci_pcie_type(dev) == PCI_EXP_TYPE_RC_END) {
-		dev->ptm_granularity = 0;
-	} else
-		return -EINVAL;
+	}
 
 	ctrl = PCI_PTM_CTRL_ENABLE;
 	ctrl |= dev->ptm_granularity << 8;
+	if (dev->ptm_root)
+		ctrl |= PCI_PTM_CTRL_ROOT;
+
 	pci_write_config_dword(dev, ptm + PCI_PTM_CTRL, ctrl);
 	dev->ptm_enabled = 1;
 
-- 
2.25.1


  parent reply	other threads:[~2022-09-09 20:25 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-09 20:24 [PATCH v4 0/9] PCI/PM: Always disable PTM for all devices during suspend Bjorn Helgaas
2022-09-09 20:24 ` [PATCH v4 1/9] PCI/PTM: Cache PTM Capability offset Bjorn Helgaas
2022-09-09 23:38   ` Sathyanarayanan Kuppuswamy
2022-09-09 20:24 ` [PATCH v4 2/9] PCI/PTM: Add pci_upstream_ptm() helper Bjorn Helgaas
2022-09-09 23:38   ` Sathyanarayanan Kuppuswamy
2022-09-09 20:24 ` Bjorn Helgaas [this message]
2022-09-09 20:25 ` [PATCH v4 4/9] PCI/PTM: Add pci_suspend_ptm() and pci_resume_ptm() Bjorn Helgaas
2022-09-09 23:44   ` Sathyanarayanan Kuppuswamy
2022-09-09 20:25 ` [PATCH v4 5/9] PCI/PTM: Move pci_ptm_info() body into its only caller Bjorn Helgaas
2022-09-09 23:45   ` Sathyanarayanan Kuppuswamy
2022-09-09 20:25 ` [PATCH v4 6/9] PCI/PTM: Preserve RsvdP bits in PTM Control register Bjorn Helgaas
2022-09-09 23:46   ` Sathyanarayanan Kuppuswamy
2022-09-09 20:25 ` [PATCH v4 7/9] PCI/PTM: Reorder functions in logical order Bjorn Helgaas
2022-09-09 20:25 ` [PATCH v4 8/9] PCI/PTM: Consolidate PTM interface declarations Bjorn Helgaas
2022-09-09 20:25 ` [PATCH v4 9/9] PCI/PM: Always disable PTM for all devices during suspend Bjorn Helgaas
2022-09-12  4:16 ` [PATCH v4 0/9] " Mika Westerberg
2022-09-12 20:36   ` Bjorn Helgaas
2022-09-13  8:13     ` Rafael J. Wysocki

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220909202505.314195-4-helgaas@kernel.org \
    --to=helgaas@kernel.org \
    --cc=bhelgaas@google.com \
    --cc=david.e.box@linux.intel.com \
    --cc=kai.heng.feng@canonical.com \
    --cc=koba.ko@canonical.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=mika.westerberg@linux.intel.com \
    --cc=rafael@kernel.org \
    --cc=rajvi.jingar@linux.intel.com \
    --cc=sathyanarayanan.kuppuswamy@linux.intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox