lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <154992573941.7864.17514554950814966890.stgit@bhelgaas-glaptop.roam.corp.google.com>
Date:   Mon, 11 Feb 2019 16:55:39 -0600
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     linux-pci@...r.kernel.org
Cc:     RussianNeuroMancer <russianneuromancer@...ru>,
        David Ward <david.ward@...mit.edu>,
        Frederick Lawler <fred@...dlawl.com>,
        Patrick Talbert <ptalbert@...hat.com>,
        Lukas Wunner <lukas@...ner.de>,
        Srinath Mannam <srinath.mannam@...adcom.com>,
        "Rafael J. Wysocki" <rafael.j.wysocki@...el.com>,
        Vidya Sagar <vidyas@...dia.com>,
        Emmanuel Grumbach <emmanuel.grumbach@...el.com>,
        linux-kernel@...r.kernel.org
Subject: [PATCH v1 1/2] PCI/ASPM: Use LTR if already enabled by platform

From: Bjorn Helgaas <bhelgaas@...gle.com>

RussianNeuroMancer reported that the Intel 7265 wifi on a Dell Venue 11 Pro
7140 table stopped working after wakeup from suspend and bisected the
problem to 9ab105deb60f ("PCI/ASPM: Disable ASPM L1.2 Substate if we don't
have LTR").  David Ward reported the same problem on a Dell Latitude 7350.

After af8bb9f89838 ("PCI/ACPI: Request LTR control from platform before
using it"), we don't enable LTR unless the platform has granted LTR control
to us.  In addition, we don't notice if the platform had already enabled
LTR itself.

After 9ab105deb60f ("PCI/ASPM: Disable ASPM L1.2 Substate if we don't have
LTR"), we avoid using LTR if we don't think the path to the device has LTR
enabled.

The combination means that if the platform itself enables LTR but declines
to give the OS control over LTR, we unnecessarily avoided using ASPM L1.2.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=201469
Fixes: 9ab105deb60f ("PCI/ASPM: Disable ASPM L1.2 Substate if we don't have LTR")
Fixes: af8bb9f89838 ("PCI/ACPI: Request LTR control from platform before using it")
Reported-by: RussianNeuroMancer <russianneuromancer@...ru>
Reported-by: David Ward <david.ward@...mit.edu>
CC: stable@...r.kernel.org	# v4.18+
---
 drivers/pci/probe.c |   36 +++++++++++++++++++++++-------------
 1 file changed, 23 insertions(+), 13 deletions(-)

diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c
index 257b9f6f2ebb..c46a3fcb341e 100644
--- a/drivers/pci/probe.c
+++ b/drivers/pci/probe.c
@@ -2071,11 +2071,8 @@ static void pci_configure_ltr(struct pci_dev *dev)
 {
 #ifdef CONFIG_PCIEASPM
 	struct pci_host_bridge *host = pci_find_host_bridge(dev->bus);
-	u32 cap;
 	struct pci_dev *bridge;
-
-	if (!host->native_ltr)
-		return;
+	u32 cap, ctl;
 
 	if (!pci_is_pcie(dev))
 		return;
@@ -2084,22 +2081,35 @@ static void pci_configure_ltr(struct pci_dev *dev)
 	if (!(cap & PCI_EXP_DEVCAP2_LTR))
 		return;
 
-	/*
-	 * Software must not enable LTR in an Endpoint unless the Root
-	 * Complex and all intermediate Switches indicate support for LTR.
-	 * PCIe r3.1, sec 6.18.
-	 */
-	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT)
-		dev->ltr_path = 1;
-	else {
+	pcie_capability_read_dword(dev, PCI_EXP_DEVCTL2, &ctl);
+	if (ctl & PCI_EXP_DEVCTL2_LTR_EN) {
+		if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT) {
+			dev->ltr_path = 1;
+			return;
+		}
+
 		bridge = pci_upstream_bridge(dev);
 		if (bridge && bridge->ltr_path)
 			dev->ltr_path = 1;
+
+		return;
 	}
 
-	if (dev->ltr_path)
+	if (!host->native_ltr)
+		return;
+
+	/*
+	 * Software must not enable LTR in an Endpoint unless the Root
+	 * Complex and all intermediate Switches indicate support for LTR.
+	 * PCIe r4.0, sec 6.18.
+	 */
+	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT ||
+	    ((bridge = pci_upstream_bridge(dev)) &&
+	      bridge->ltr_path)) {
 		pcie_capability_set_word(dev, PCI_EXP_DEVCTL2,
 					 PCI_EXP_DEVCTL2_LTR_EN);
+		dev->ltr_path = 1;
+	}
 #endif
 }
 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ