From: Alexandru Gagniuc <mr.nuke.me@gmail.com>
To: helgaas@kernel.org
Cc: alex_gagniuc@dellteam.com, Austin.Bolen@dell.com,
Shyam.Iyer@dell.com, Alexandru Gagniuc <mr.nuke.me@gmail.com>,
Bjorn Helgaas <bhelgaas@google.com>,
"Rafael J. Wysocki" <rafael.j.wysocki@intel.com>,
Keith Busch <keith.busch@intel.com>,
Oza Pawandeep <poza@codeaurora.org>,
Mika Westerberg <mika.westerberg@linux.intel.com>,
Frederick Lawler <fred@fredlawl.com>,
Lukas Wunner <lukas@wunner.de>,
linux-kernel@vger.kernel.org, linux-pci@vger.kernel.org
Subject: [PATCH v2] PCI: pciehp: Report degraded links via link bandwidth notification
Date: Fri, 7 Dec 2018 12:20:00 -0600 [thread overview]
Message-ID: <20181207182021.16344-1-mr.nuke.me@gmail.com> (raw)
In-Reply-To: <20181129230454.GF178809@google.com>
A warning is generated when a PCIe device is probed with a degraded
link, but there was no similar mechanism to warn when the link becomes
degraded after probing. The Link Bandwidth Notification provides this
mechanism.
Use the link bandwidth notification interrupt to detect bandwidth
changes, and rescan the bandwidth, looking for the weakest point. This
is the same logic used in probe().
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
---
Changes since v1:
* Layer on top of the pcie port service drivers, instead of hotplug service.
This patch needs to be applied on top of:
PCI: Add missing include to drivers/pci.h
Anticipated FAQ:
Q: Why is this unconditionally compiled in?
A: The symmetrical check in pci probe() is also always compiled in.
Q: Why call module_init() instead of adding a call in pcie_init_services() ?
A: A call in pcie_init_services() also requires a prototype in portdrv.h, a
non-static implementation in bw_notification.c. Using module_init() is
functionally equivalent, and takes less code.
Q: Why print only on degraded links and not when a link is back to full speed?
For symmetry with PCI probe(). Although I see a benefit in conveying that a
link is back to full speed, I expect this to be extremely rare. Secondary bus
reset is usually needed to retrain at full bandwidth.
drivers/pci/pcie/Makefile | 1 +
drivers/pci/pcie/bw_notification.c | 107 +++++++++++++++++++++++++++++
drivers/pci/pcie/portdrv.h | 4 +-
drivers/pci/pcie/portdrv_core.c | 14 ++--
4 files changed, 121 insertions(+), 5 deletions(-)
create mode 100644 drivers/pci/pcie/bw_notification.c
diff --git a/drivers/pci/pcie/Makefile b/drivers/pci/pcie/Makefile
index ab514083d5d4..f1d7bc1e5efa 100644
--- a/drivers/pci/pcie/Makefile
+++ b/drivers/pci/pcie/Makefile
@@ -3,6 +3,7 @@
# Makefile for PCI Express features and port driver
pcieportdrv-y := portdrv_core.o portdrv_pci.o err.o
+pcieportdrv-y += bw_notification.o
obj-$(CONFIG_PCIEPORTBUS) += pcieportdrv.o
diff --git a/drivers/pci/pcie/bw_notification.c b/drivers/pci/pcie/bw_notification.c
new file mode 100644
index 000000000000..64391ea9a172
--- /dev/null
+++ b/drivers/pci/pcie/bw_notification.c
@@ -0,0 +1,107 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * PCI Express Bandwidth notification services driver
+ * Author: Alexandru Gagniuc <mr.nuke.me@gmail.com>
+ *
+ * Copyright (C) 2018, Dell Inc
+ *
+ * The PCIe bandwidth notification provides a way to notify the operating system
+ * when the link width or data rate changes. This capability is required for all
+ * root ports and downstream ports supporting links wider than x1 and/or
+ * multiple link speeds.
+ *
+ * This service port driver hooks into the bandwidth notification interrupt and
+ * warns when links become degraded in operation.
+ */
+
+#include <linux/module.h>
+
+#include "../pci.h"
+#include "portdrv.h"
+
+static bool pcie_link_bandwidth_notification_supported(struct pci_dev *dev)
+{
+ int ret;
+ u32 lnk_cap;
+
+ ret = pcie_capability_read_dword(dev, PCI_EXP_LNKCAP, &lnk_cap);
+ return (ret == PCIBIOS_SUCCESSFUL) && (lnk_cap & PCI_EXP_LNKCAP_LBNC);
+}
+
+static void pcie_enable_link_bandwidth_notification(struct pci_dev *dev)
+{
+ u16 lnk_ctl;
+
+ pcie_capability_read_word(dev, PCI_EXP_LNKCTL, &lnk_ctl);
+ lnk_ctl |= PCI_EXP_LNKCTL_LBMIE;
+ pcie_capability_write_word(dev, PCI_EXP_LNKCTL, lnk_ctl);
+}
+
+static void pcie_disable_link_bandwidth_notification(struct pci_dev *dev)
+{
+ u16 lnk_ctl;
+
+ pcie_capability_read_word(dev, PCI_EXP_LNKCTL, &lnk_ctl);
+ lnk_ctl &= ~PCI_EXP_LNKCTL_LBMIE;
+ pcie_capability_write_word(dev, PCI_EXP_LNKCTL, lnk_ctl);
+}
+
+static irqreturn_t pcie_bw_notification_irq(int irq, void *context)
+{
+ struct pcie_device *srv = context;
+ struct pci_dev *port = srv->port;
+ struct pci_dev *dev;
+ u16 link_status, events;
+ int ret;
+
+ ret = pcie_capability_read_word(port, PCI_EXP_LNKSTA, &link_status);
+ events = link_status & PCI_EXP_LNKSTA_LBMS;
+
+ if (!events || ret != PCIBIOS_SUCCESSFUL)
+ return IRQ_NONE;
+
+ /* Print status from upstream link partner, not this downstream port. */
+ list_for_each_entry(dev, &port->subordinate->devices, bus_list)
+ __pcie_print_link_status(dev, false);
+
+ pcie_capability_write_word(port, PCI_EXP_LNKSTA, events);
+ return IRQ_HANDLED;
+}
+
+static int pcie_bandwidth_notification_probe(struct pcie_device *srv)
+{
+ int ret;
+
+ /* Single-width or single-speed ports do not have to support this. */
+ if (!pcie_link_bandwidth_notification_supported(srv->port))
+ return -ENODEV;
+
+ ret = devm_request_irq(&srv->device, srv->irq, pcie_bw_notification_irq,
+ IRQF_SHARED, "PCIe BW notif", srv);
+ if (ret)
+ return ret;
+
+ pcie_enable_link_bandwidth_notification(srv->port);
+
+ return 0;
+}
+
+static void pcie_bandwidth_notification_remove(struct pcie_device *srv)
+{
+ pcie_disable_link_bandwidth_notification(srv->port);
+}
+
+static struct pcie_port_service_driver pcie_bandwidth_notification_driver = {
+ .name = "pcie_bw_notification",
+ .port_type = PCI_EXP_TYPE_DOWNSTREAM,
+ .service = PCIE_PORT_SERVICE_BwNOTIF,
+ .probe = pcie_bandwidth_notification_probe,
+ .remove = pcie_bandwidth_notification_remove,
+};
+
+static int __init pcie_bandwidth_notification_service_init(void)
+{
+ return pcie_port_service_register(&pcie_bandwidth_notification_driver);
+}
+
+module_init(pcie_bandwidth_notification_service_init);
diff --git a/drivers/pci/pcie/portdrv.h b/drivers/pci/pcie/portdrv.h
index e495f04394d0..46652469ffaa 100644
--- a/drivers/pci/pcie/portdrv.h
+++ b/drivers/pci/pcie/portdrv.h
@@ -20,8 +20,10 @@
#define PCIE_PORT_SERVICE_HP (1 << PCIE_PORT_SERVICE_HP_SHIFT)
#define PCIE_PORT_SERVICE_DPC_SHIFT 3 /* Downstream Port Containment */
#define PCIE_PORT_SERVICE_DPC (1 << PCIE_PORT_SERVICE_DPC_SHIFT)
+#define PCIE_PORT_SERVICE_BWNOTIF_SHIFT 4 /* Bandwidth notification */
+#define PCIE_PORT_SERVICE_BwNOTIF (1 << PCIE_PORT_SERVICE_BWNOTIF_SHIFT)
-#define PCIE_PORT_DEVICE_MAXSERVICES 4
+#define PCIE_PORT_DEVICE_MAXSERVICES 5
#ifdef CONFIG_PCIEAER
int pcie_aer_init(void);
diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c
index f458ac9cb70c..86455ff7ced9 100644
--- a/drivers/pci/pcie/portdrv_core.c
+++ b/drivers/pci/pcie/portdrv_core.c
@@ -99,7 +99,7 @@ static int pcie_message_numbers(struct pci_dev *dev, int mask,
*/
static int pcie_port_enable_irq_vec(struct pci_dev *dev, int *irqs, int mask)
{
- int nr_entries, nvec;
+ int nr_entries, nvec, pcie_irq;
u32 pme = 0, aer = 0, dpc = 0;
/* Allocate the maximum possible number of MSI/MSI-X vectors */
@@ -136,9 +136,12 @@ static int pcie_port_enable_irq_vec(struct pci_dev *dev, int *irqs, int mask)
}
/* PME and hotplug share an MSI/MSI-X vector */
- if (mask & (PCIE_PORT_SERVICE_PME | PCIE_PORT_SERVICE_HP)) {
- irqs[PCIE_PORT_SERVICE_PME_SHIFT] = pci_irq_vector(dev, pme);
- irqs[PCIE_PORT_SERVICE_HP_SHIFT] = pci_irq_vector(dev, pme);
+ if (mask & (PCIE_PORT_SERVICE_PME | PCIE_PORT_SERVICE_HP |
+ PCIE_PORT_SERVICE_BwNOTIF)) {
+ pcie_irq = pci_irq_vector(dev, pme);
+ irqs[PCIE_PORT_SERVICE_PME_SHIFT] = pcie_irq;
+ irqs[PCIE_PORT_SERVICE_HP_SHIFT] = pcie_irq;
+ irqs[PCIE_PORT_SERVICE_BWNOTIF_SHIFT] = pcie_irq;
}
if (mask & PCIE_PORT_SERVICE_AER)
@@ -250,6 +253,9 @@ static int get_port_device_capability(struct pci_dev *dev)
pci_aer_available() && services & PCIE_PORT_SERVICE_AER)
services |= PCIE_PORT_SERVICE_DPC;
+ if (pci_pcie_type(dev) == PCI_EXP_TYPE_DOWNSTREAM)
+ services |= PCIE_PORT_SERVICE_BwNOTIF;
+
return services;
}
--
2.17.1
next prev parent reply other threads:[~2018-12-07 18:21 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-29 0:08 [PATCH] PCI: pciehp: Report degraded links via link bandwidth notification Alexandru Gagniuc
2018-11-29 16:06 ` Mika Westerberg
2018-11-29 19:00 ` Alex_Gagniuc
2018-11-29 19:30 ` Mika Westerberg
2018-11-29 17:35 ` Bjorn Helgaas
2018-11-29 18:57 ` Alex_Gagniuc
2018-11-29 19:13 ` Lukas Wunner
2018-11-29 23:04 ` Bjorn Helgaas
2018-11-29 23:24 ` Alex_Gagniuc
2018-12-07 18:20 ` Alexandru Gagniuc [this message]
2018-12-27 19:33 ` [PATCH v2] " Alex G.
2019-02-25 2:28 ` Lukas Wunner
2019-02-27 20:21 ` Alex_Gagniuc
2019-02-28 6:43 ` Lukas Wunner
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=20181207182021.16344-1-mr.nuke.me@gmail.com \
--to=mr.nuke.me@gmail.com \
--cc=Austin.Bolen@dell.com \
--cc=Shyam.Iyer@dell.com \
--cc=alex_gagniuc@dellteam.com \
--cc=bhelgaas@google.com \
--cc=fred@fredlawl.com \
--cc=helgaas@kernel.org \
--cc=keith.busch@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lukas@wunner.de \
--cc=mika.westerberg@linux.intel.com \
--cc=poza@codeaurora.org \
--cc=rafael.j.wysocki@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;
as well as URLs for NNTP newsgroup(s).