From 6f6f42466d902c92f21b46a45e6af22d1d663607 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Mon, 25 Jun 2018 08:17:33 -0500 Subject: [PATCH 01/25] PCI: shpchp: Manage SHPC unconditionally on non-ACPI systems An SHPC can be operated either by platform firmware or by the OS. The OS uses a host bridge ACPI _OSC method to negotiate for control of SHPC. If firmware wants to prevent an OS from operating an SHPC, it must supply an _OSC method that declines to grant SHPC ownership to the OS. If acpi_pci_find_root() returns NULL, it means there's no ACPI host bridge device (PNP0A03 or PNP0A08) and hence no _OSC method, so the OS is always allowed to manage the SHPC. Fix a NULL pointer dereference when CONFIG_ACPI=y but the current hardware/firmware platform doesn't support ACPI. In that case, acpi_get_hp_hw_control_from_firmware() is implemented but acpi_pci_find_root() returns NULL. Fixes: 90cc0c3cc709 ("PCI: shpchp: Add shpchp_is_native()") Link: https://lkml.kernel.org/r/20180621164715.28160-1-marc.zyngier@arm.com Reported-by: Marc Zyngier Tested-by: Marc Zyngier Signed-off-by: Bjorn Helgaas Reviewed-by: Mika Westerberg --- drivers/pci/hotplug/acpi_pcihp.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/pci/hotplug/acpi_pcihp.c b/drivers/pci/hotplug/acpi_pcihp.c index 3979f89b250a..5bd6c1573295 100644 --- a/drivers/pci/hotplug/acpi_pcihp.c +++ b/drivers/pci/hotplug/acpi_pcihp.c @@ -7,7 +7,6 @@ * All rights reserved. * * Send feedback to - * */ #include @@ -87,8 +86,17 @@ int acpi_get_hp_hw_control_from_firmware(struct pci_dev *pdev) return 0; /* If _OSC exists, we should not evaluate OSHP */ + + /* + * If there's no ACPI host bridge (i.e., ACPI support is compiled + * into the kernel but the hardware platform doesn't support ACPI), + * there's nothing to do here. + */ host = pci_find_host_bridge(pdev->bus); root = acpi_pci_find_root(ACPI_HANDLE(&host->dev)); + if (!root) + return 0; + if (root->osc_support_set) goto no_control; From 68bb22e9d561ad7c20274eee92058048ac689b98 Mon Sep 17 00:00:00 2001 From: Alan Douglas Date: Fri, 22 Jun 2018 17:17:17 +0100 Subject: [PATCH 02/25] PCI: Initialize endpoint library before controllers The endpoint library must be initialized before its users, which are in drivers/pci/controllers. The endpoint initialization currently depends on link order. This corrects a kernel crash when loading the Cadence EP driver, since it calls devm_pci_epc_create() and this is only valid once the endpoint library has been initialized. Fixes: 6e0832fa432e ("PCI: Collect all native drivers under drivers/pci/controller/") Signed-off-by: Alan Douglas Signed-off-by: Bjorn Helgaas --- drivers/pci/Makefile | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile index 535201984b8b..1b2cfe51e8d7 100644 --- a/drivers/pci/Makefile +++ b/drivers/pci/Makefile @@ -28,10 +28,10 @@ obj-$(CONFIG_PCI_PF_STUB) += pci-pf-stub.o obj-$(CONFIG_PCI_ECAM) += ecam.o obj-$(CONFIG_XEN_PCIDEV_FRONTEND) += xen-pcifront.o -obj-y += controller/ -obj-y += switch/ - # Endpoint library must be initialized before its users obj-$(CONFIG_PCI_ENDPOINT) += endpoint/ +obj-y += controller/ +obj-y += switch/ + ccflags-$(CONFIG_PCI_DEBUG) := -DDEBUG From 925d31668d5b6c9849c50167f31f9b5cf9eec892 Mon Sep 17 00:00:00 2001 From: Lorenzo Pieralisi Date: Tue, 19 Jun 2018 12:21:05 +0100 Subject: [PATCH 03/25] PCI: controller: Move PCI_DOMAINS selection to arch Kconfig Commit 51bc085d6454 ("PCI: Improve host drivers compile test coverage") added configuration options to allow PCI host controller drivers to be compile tested on all architectures. Some host controller drivers (eg PCIE_ALTERA) config entries select the PCI_DOMAINS config option to enable PCI domains management in the kernel. Now that host controller drivers can be compiled on all architectures, this triggers build regressions on arches that do not implement the PCI_DOMAINS required API (ie pci_domain_nr()): drivers/ata/pata_ali.c: In function 'ali_init_chipset': drivers/ata/pata_ali.c:469:38: error: implicit declaration of function 'pci_domain_nr'; did you mean 'pci_iomap_wc'? Furthemore, some software configurations (ie Jailhouse) require a PCI_DOMAINS enabled kernel to configure multiple host controllers without having an explicit dependency on the ARM platform on which they run. Make PCI_DOMAINS a visible configuration option on ARM so that software configurations that need it can manually select it and move the PCI_DOMAINS selection from PCI controllers configuration file to ARM sub-arch config entries that currently require it, fixing the issue. Fixes: 51bc085d6454 ("PCI: Improve host drivers compile test coverage") Link: https://lkml.kernel.org/r/20180612170229.GA10141@roeck-us.net Reported-by: Guenter Roeck Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Acked-by: Jan Kiszka Acked-by: Ley Foon Tan Acked-by: Rob Herring Cc: Scott Branden Cc: Will Deacon Cc: Russell King Cc: Guenter Roeck --- arch/arm/Kconfig | 8 +++++++- arch/arm/mach-bcm/Kconfig | 1 + arch/arm/mach-socfpga/Kconfig | 1 + drivers/pci/controller/Kconfig | 3 --- 4 files changed, 9 insertions(+), 4 deletions(-) diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 54eeb8d00bc6..843edfd000be 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -1245,8 +1245,14 @@ config PCI VESA. If you have PCI, say Y, otherwise N. config PCI_DOMAINS - bool + bool "Support for multiple PCI domains" depends on PCI + help + Enable PCI domains kernel management. Say Y if your machine + has a PCI bus hierarchy that requires more than one PCI + domain (aka segment) to be correctly managed. Say N otherwise. + + If you don't know what to do here, say N. config PCI_DOMAINS_GENERIC def_bool PCI_DOMAINS diff --git a/arch/arm/mach-bcm/Kconfig b/arch/arm/mach-bcm/Kconfig index c46a728df44e..25aac6ee2ab1 100644 --- a/arch/arm/mach-bcm/Kconfig +++ b/arch/arm/mach-bcm/Kconfig @@ -20,6 +20,7 @@ config ARCH_BCM_IPROC select GPIOLIB select ARM_AMBA select PINCTRL + select PCI_DOMAINS if PCI help This enables support for systems based on Broadcom IPROC architected SoCs. The IPROC complex contains one or more ARM CPUs along with common diff --git a/arch/arm/mach-socfpga/Kconfig b/arch/arm/mach-socfpga/Kconfig index d0f62eacf59d..4adb901dd5eb 100644 --- a/arch/arm/mach-socfpga/Kconfig +++ b/arch/arm/mach-socfpga/Kconfig @@ -10,6 +10,7 @@ menuconfig ARCH_SOCFPGA select HAVE_ARM_SCU select HAVE_ARM_TWD if SMP select MFD_SYSCON + select PCI_DOMAINS if PCI if ARCH_SOCFPGA config SOCFPGA_SUSPEND diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig index 18fa09b3ac8f..cc9fa02d32a0 100644 --- a/drivers/pci/controller/Kconfig +++ b/drivers/pci/controller/Kconfig @@ -96,7 +96,6 @@ config PCI_HOST_GENERIC depends on OF select PCI_HOST_COMMON select IRQ_DOMAIN - select PCI_DOMAINS help Say Y here if you want to support a simple generic PCI host controller, such as the one emulated by kvmtool. @@ -138,7 +137,6 @@ config PCI_VERSATILE config PCIE_IPROC tristate - select PCI_DOMAINS help This enables the iProc PCIe core controller support for Broadcom's iProc family of SoCs. An appropriate bus interface driver needs @@ -176,7 +174,6 @@ config PCIE_IPROC_MSI config PCIE_ALTERA bool "Altera PCIe controller" depends on ARM || NIOS2 || COMPILE_TEST - select PCI_DOMAINS help Say Y here if you want to enable PCIe controller support on Altera FPGA. From 4050360f964694a3ac0c83badd1a441207c86889 Mon Sep 17 00:00:00 2001 From: Marek Vasut Date: Fri, 29 Jun 2018 13:47:38 -0500 Subject: [PATCH 04/25] PCI: rcar: Shut the PHY down in failpath If anything fails past phy_init_fn() and the system is a Gen3 with a PHY, the PHY will be left on and inited. This is caused by the phy_init_fn, which is in fact a pointer to rcar_pcie_phy_init_gen3() function, which starts the PHY, yet has no counterpart in the failpath. Add that counterpart. Fixes: 517ca93a7159 ("PCI: rcar: Add R-Car gen3 PHY support") Signed-off-by: Marek Vasut Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Geert Uytterhoeven Acked-by: Simon Horman Cc: Geert Uytterhoeven Cc: Phil Edworthy Cc: Wolfram Sang --- drivers/pci/controller/pcie-rcar.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c index 874d75c9ee4a..8b0d3206cbc4 100644 --- a/drivers/pci/controller/pcie-rcar.c +++ b/drivers/pci/controller/pcie-rcar.c @@ -1165,7 +1165,7 @@ static int rcar_pcie_probe(struct platform_device *pdev) if (rcar_pcie_hw_init(pcie)) { dev_info(dev, "PCIe link down\n"); err = -ENODEV; - goto err_clk_disable; + goto err_phy_shutdown; } data = rcar_pci_read_reg(pcie, MACSR); @@ -1177,7 +1177,7 @@ static int rcar_pcie_probe(struct platform_device *pdev) dev_err(dev, "failed to enable MSI support: %d\n", err); - goto err_clk_disable; + goto err_phy_shutdown; } } @@ -1191,6 +1191,12 @@ err_msi_teardown: if (IS_ENABLED(CONFIG_PCI_MSI)) rcar_pcie_teardown_msi(pcie); +err_phy_shutdown: + if (pcie->phy) { + phy_power_off(pcie->phy); + phy_exit(pcie->phy); + } + err_clk_disable: clk_disable_unprepare(pcie->bus_clk); From 3c5777c372b6eb2e17802b3dc4bd5ebea45d9bcc Mon Sep 17 00:00:00 2001 From: Marek Vasut Date: Fri, 29 Jun 2018 13:48:15 -0500 Subject: [PATCH 05/25] PCI: rcar: Clean up PHY init on failure If the Gen3 PHY fails to power up, the code does not undo the initialization caused by phy_init(). Add the missing failure handling to the rcar_pcie_phy_init_gen3() function. Fixes: 517ca93a7159 ("PCI: rcar: Add R-Car gen3 PHY support") Reported-by: Geert Uytterhoeven Signed-off-by: Marek Vasut Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Geert Uytterhoeven Acked-by: Simon Horman Cc: Geert Uytterhoeven Cc: Phil Edworthy Cc: Wolfram Sang --- drivers/pci/controller/pcie-rcar.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c index 8b0d3206cbc4..c8febb009454 100644 --- a/drivers/pci/controller/pcie-rcar.c +++ b/drivers/pci/controller/pcie-rcar.c @@ -680,7 +680,11 @@ static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie) if (err) return err; - return phy_power_on(pcie->phy); + err = phy_power_on(pcie->phy); + if (err) + phy_exit(pcie->phy); + + return err; } static int rcar_msi_alloc(struct rcar_msi *chip) From 9bc5f0833a55c7bf768d517b5fdcf89ead43b794 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Fri, 29 Jun 2018 13:49:04 -0500 Subject: [PATCH 06/25] PCI: controller: dwc: Do not let PCIE_DW_PLAT_HOST default to yes PCIE_DW_PLAT_HOST does not have any platform dependency, so it should not default to yes. Fixes: 1d906b22076e12cf ("PCI: dwc: Add support for EP mode") Signed-off-by: Geert Uytterhoeven Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Acked-by: Gustavo Pimentel --- drivers/pci/controller/dwc/Kconfig | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/pci/controller/dwc/Kconfig b/drivers/pci/controller/dwc/Kconfig index 16f52c626b4b..91b0194240a5 100644 --- a/drivers/pci/controller/dwc/Kconfig +++ b/drivers/pci/controller/dwc/Kconfig @@ -58,7 +58,6 @@ config PCIE_DW_PLAT_HOST depends on PCI && PCI_MSI_IRQ_DOMAIN select PCIE_DW_HOST select PCIE_DW_PLAT - default y help Enables support for the PCIe controller in the Designware IP to work in host mode. There are two instances of PCIe controller in From fb0de5b8dcc68da4a0d39b0f5fcd5368085ad7c1 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 29 Jun 2018 13:49:24 -0500 Subject: [PATCH 07/25] PCI: endpoint: Use after free in pci_epf_unregister_driver() We need to use list_for_each_entry_safe() because the pci_ep_cfs_remove_epf_group() function frees "group". Fixes: ef1433f717a2 ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry") Signed-off-by: Dan Carpenter [lorenzo.pieralisi@arm.com: updated commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Acked-by: Kishon Vijay Abraham I --- drivers/pci/endpoint/pci-epf-core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/pci/endpoint/pci-epf-core.c b/drivers/pci/endpoint/pci-epf-core.c index 523a8cab3bfb..bf53fad636a5 100644 --- a/drivers/pci/endpoint/pci-epf-core.c +++ b/drivers/pci/endpoint/pci-epf-core.c @@ -145,10 +145,10 @@ EXPORT_SYMBOL_GPL(pci_epf_alloc_space); */ void pci_epf_unregister_driver(struct pci_epf_driver *driver) { - struct config_group *group; + struct config_group *group, *tmp; mutex_lock(&pci_epf_mutex); - list_for_each_entry(group, &driver->epf_group, group_entry) + list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) pci_ep_cfs_remove_epf_group(group); list_del(&driver->epf_group); mutex_unlock(&pci_epf_mutex); From 8c3f9bd851a4d3acf0a0f222d4e9e41c0cd1ea8e Mon Sep 17 00:00:00 2001 From: Nicholas Mc Guire Date: Fri, 29 Jun 2018 13:49:54 -0500 Subject: [PATCH 08/25] PCI: xilinx: Add missing of_node_put() The call to of_get_next_child() returns a node pointer with refcount incremented thus it must be explicitly decremented here after the last usage. Fixes: 8961def56845 ("PCI: xilinx: Add Xilinx AXI PCIe Host Bridge IP driver") Signed-off-by: Nicholas Mc Guire [lorenzo.pieralisi@arm.com: reworked commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas --- drivers/pci/controller/pcie-xilinx.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/pci/controller/pcie-xilinx.c b/drivers/pci/controller/pcie-xilinx.c index b110a3a814e3..7b1389d8e2a5 100644 --- a/drivers/pci/controller/pcie-xilinx.c +++ b/drivers/pci/controller/pcie-xilinx.c @@ -509,6 +509,7 @@ static int xilinx_pcie_init_irq_domain(struct xilinx_pcie_port *port) port->leg_domain = irq_domain_add_linear(pcie_intc_node, PCI_NUM_INTX, &intx_domain_ops, port); + of_node_put(pcie_intc_node); if (!port->leg_domain) { dev_err(dev, "Failed to get a INTx IRQ domain\n"); return -ENODEV; From 342639d996f18bc0a4db2f42a84230c0a966dc94 Mon Sep 17 00:00:00 2001 From: Nicholas Mc Guire Date: Fri, 29 Jun 2018 13:50:10 -0500 Subject: [PATCH 09/25] PCI: xilinx-nwl: Add missing of_node_put() The call to of_get_next_child() returns a node pointer with refcount incremented thus it must be explicitly decremented here after the last usage. Fixes: ab597d35ef11 ("PCI: xilinx-nwl: Add support for Xilinx NWL PCIe Host Controller") Signed-off-by: Nicholas Mc Guire [lorenzo.pieralisi@arm.com: updated commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas --- drivers/pci/controller/pcie-xilinx-nwl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pcie-xilinx-nwl.c b/drivers/pci/controller/pcie-xilinx-nwl.c index 6a4bbb5b3de0..fb32840ce8e6 100644 --- a/drivers/pci/controller/pcie-xilinx-nwl.c +++ b/drivers/pci/controller/pcie-xilinx-nwl.c @@ -559,7 +559,7 @@ static int nwl_pcie_init_irq_domain(struct nwl_pcie *pcie) PCI_NUM_INTX, &legacy_domain_ops, pcie); - + of_node_put(legacy_intc_node); if (!pcie->legacy_irq_domain) { dev_err(dev, "failed to create IRQ domain\n"); return -ENOMEM; From 3dc6ddfedc2818eaaa36842fbb049191e0c5e50f Mon Sep 17 00:00:00 2001 From: Nicholas Mc Guire Date: Fri, 29 Jun 2018 13:50:27 -0500 Subject: [PATCH 10/25] PCI: faraday: Add missing of_node_put() The call to of_get_next_child() returns a node pointer with refcount incremented thus it must be explicitly decremented here in the error path and after the last usage. Fixes: d3c68e0a7e34 ("PCI: faraday: Add Faraday Technology FTPCI100 PCI Host Bridge driver") Signed-off-by: Nicholas Mc Guire [lorenzo.pieralisi@arm.com: updated commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij --- drivers/pci/controller/pci-ftpci100.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/pci/controller/pci-ftpci100.c b/drivers/pci/controller/pci-ftpci100.c index a1ebe9ed441f..20bb2564a6b3 100644 --- a/drivers/pci/controller/pci-ftpci100.c +++ b/drivers/pci/controller/pci-ftpci100.c @@ -355,11 +355,13 @@ static int faraday_pci_setup_cascaded_irq(struct faraday_pci *p) irq = of_irq_get(intc, 0); if (irq <= 0) { dev_err(p->dev, "failed to get parent IRQ\n"); + of_node_put(intc); return irq ?: -EINVAL; } p->irqdomain = irq_domain_add_linear(intc, PCI_NUM_INTX, &faraday_pci_irqdomain_ops, p); + of_node_put(intc); if (!p->irqdomain) { dev_err(p->dev, "failed to create Gemini PCI IRQ domain\n"); return -EINVAL; From 38972375ef7bdc7dd989bcb48d5448662a95bca2 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Fri, 29 Jun 2018 15:08:52 -0500 Subject: [PATCH 11/25] PCI/IOV: Reset total_VFs limit after detaching PF driver The TotalVFs register in the SR-IOV capability is the hardware limit on the number of VFs. A PF driver can limit the number of VFs further with pci_sriov_set_totalvfs(). When the PF driver is removed, reset any VF limit that was imposed by the driver because that limit may not apply to other drivers. Before 8d85a7a4f2c9 ("PCI/IOV: Allow PF drivers to limit total_VFs to 0"), pci_sriov_set_totalvfs(pdev, 0) meant "we can enable TotalVFs virtual functions", and the nfp driver used that to remove the VF limit when the driver unloads. 8d85a7a4f2c9 broke that because instead of removing the VF limit, pci_sriov_set_totalvfs(pdev, 0) actually sets the limit to zero, and that limit persists even if another driver is loaded. We could fix that by making the nfp driver reset the limit when it unloads, but it seems more robust to do it in the PCI core instead of relying on the driver. The regression scenario is: nfp_pci_probe (driver 1) ... nfp_pci_remove pci_sriov_set_totalvfs(pf->pdev, 0) # limits VFs to 0 ... nfp_pci_probe (driver 2) nfp_rtsym_read_le("nfd_vf_cfg_max_vfs") # no VF limit from firmware Now driver 2 is broken because the VF limit is still 0 from driver 1. Fixes: 8d85a7a4f2c9 ("PCI/IOV: Allow PF drivers to limit total_VFs to 0") Signed-off-by: Jakub Kicinski [bhelgaas: changelog, rename functions] Signed-off-by: Bjorn Helgaas --- drivers/pci/iov.c | 16 ++++++++++++++++ drivers/pci/pci-driver.c | 1 + drivers/pci/pci.h | 4 ++++ 3 files changed, 21 insertions(+) diff --git a/drivers/pci/iov.c b/drivers/pci/iov.c index d0d73dbbd5ca..0f04ae648cf1 100644 --- a/drivers/pci/iov.c +++ b/drivers/pci/iov.c @@ -574,6 +574,22 @@ void pci_iov_release(struct pci_dev *dev) sriov_release(dev); } +/** + * pci_iov_remove - clean up SR-IOV state after PF driver is detached + * @dev: the PCI device + */ +void pci_iov_remove(struct pci_dev *dev) +{ + struct pci_sriov *iov = dev->sriov; + + if (!dev->is_physfn) + return; + + iov->driver_max_VFs = iov->total_VFs; + if (iov->num_VFs) + pci_warn(dev, "driver left SR-IOV enabled after remove\n"); +} + /** * pci_iov_update_resource - update a VF BAR * @dev: the PCI device diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c index c125d53033c6..6792292b5fc7 100644 --- a/drivers/pci/pci-driver.c +++ b/drivers/pci/pci-driver.c @@ -445,6 +445,7 @@ static int pci_device_remove(struct device *dev) } pcibios_free_irq(pci_dev); pci_dev->driver = NULL; + pci_iov_remove(pci_dev); } /* Undo the runtime PM settings in local_pci_probe() */ diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h index c358e7a07f3f..882f1f9596df 100644 --- a/drivers/pci/pci.h +++ b/drivers/pci/pci.h @@ -311,6 +311,7 @@ static inline void pci_restore_ats_state(struct pci_dev *dev) #ifdef CONFIG_PCI_IOV int pci_iov_init(struct pci_dev *dev); void pci_iov_release(struct pci_dev *dev); +void pci_iov_remove(struct pci_dev *dev); void pci_iov_update_resource(struct pci_dev *dev, int resno); resource_size_t pci_sriov_resource_alignment(struct pci_dev *dev, int resno); void pci_restore_iov_state(struct pci_dev *dev); @@ -323,6 +324,9 @@ static inline int pci_iov_init(struct pci_dev *dev) } static inline void pci_iov_release(struct pci_dev *dev) +{ +} +static inline void pci_iov_remove(struct pci_dev *dev) { } static inline void pci_restore_iov_state(struct pci_dev *dev) From 83235822b8b4fe47ecbd6b6bcbcc902860ac00fc Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Fri, 29 Jun 2018 15:09:00 -0500 Subject: [PATCH 12/25] nfp: stop limiting VFs to 0 Before 8d85a7a4f2c9 ("PCI/IOV: Allow PF drivers to limit total_VFs to 0"), pci_sriov_set_totalvfs(pdev, 0) meant "we can enable TotalVFs virtual functions". After 8d85a7a4f2c9, it means "we can't enable *any* VFs". That broke this scenario where nfp intends to remove any limit on the number of VFs that can be enabled: nfp_pci_probe nfp_pcie_sriov_read_nfd_limit nfp_rtsym_read_le("nfd_vf_cfg_max_vfs", &err) pci_sriov_set_totalvfs(pf->pdev, 0) # if FW didn't expose a limit ... # userspace writes N to sysfs "sriov_numvfs": sriov_numvfs_store pci_sriov_get_totalvfs # now returns 0 return -ERANGE Prior to 8d85a7a4f2c9, pci_sriov_get_totalvfs() returned TotalVFs, but it now returns 0. Remove the pci_sriov_set_totalvfs(pdev, 0) calls so we don't limit the number of VFs that can be enabled. Fixes: 8d85a7a4f2c9 ("PCI/IOV: Allow PF drivers to limit total_VFs to 0") Signed-off-by: Jakub Kicinski [bhelgaas: changelog] Signed-off-by: Bjorn Helgaas --- drivers/net/ethernet/netronome/nfp/nfp_main.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.c b/drivers/net/ethernet/netronome/nfp/nfp_main.c index 46b76d5a726c..152283d7e59c 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_main.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_main.c @@ -240,7 +240,6 @@ static int nfp_pcie_sriov_read_nfd_limit(struct nfp_pf *pf) return pci_sriov_set_totalvfs(pf->pdev, pf->limit_vfs); pf->limit_vfs = ~0; - pci_sriov_set_totalvfs(pf->pdev, 0); /* 0 is unset */ /* Allow any setting for backwards compatibility if symbol not found */ if (err == -ENOENT) return 0; @@ -668,7 +667,7 @@ static int nfp_pci_probe(struct pci_dev *pdev, err = nfp_net_pci_probe(pf); if (err) - goto err_sriov_unlimit; + goto err_fw_unload; err = nfp_hwmon_register(pf); if (err) { @@ -680,8 +679,6 @@ static int nfp_pci_probe(struct pci_dev *pdev, err_net_remove: nfp_net_pci_remove(pf); -err_sriov_unlimit: - pci_sriov_set_totalvfs(pf->pdev, 0); err_fw_unload: kfree(pf->rtbl); nfp_mip_close(pf->mip); @@ -715,7 +712,6 @@ static void nfp_pci_remove(struct pci_dev *pdev) nfp_hwmon_unregister(pf); nfp_pcie_sriov_disable(pdev); - pci_sriov_set_totalvfs(pf->pdev, 0); nfp_net_pci_remove(pf); From 35a88a18d7ea58600e11590405bc93b08e16e7f5 Mon Sep 17 00:00:00 2001 From: Dexuan Cui Date: Mon, 9 Jul 2018 13:16:07 -0500 Subject: [PATCH 13/25] PCI: hv: Disable/enable IRQs rather than BH in hv_compose_msi_msg() Commit de0aa7b2f97d ("PCI: hv: Fix 2 hang issues in hv_compose_msi_msg()") uses local_bh_disable()/enable(), because hv_pci_onchannelcallback() can also run in tasklet context as the channel event callback, so bottom halves should be disabled to prevent a race condition. With CONFIG_PROVE_LOCKING=y in the recent mainline, or old kernels that don't have commit f71b74bca637 ("irq/softirqs: Use lockdep to assert IRQs are disabled/enabled"), when the upper layer IRQ code calls hv_compose_msi_msg() with local IRQs disabled, we'll see a warning at the beginning of __local_bh_enable_ip(): IRQs not enabled as expected WARNING: CPU: 0 PID: 408 at kernel/softirq.c:162 __local_bh_enable_ip The warning exposes an issue in de0aa7b2f97d: local_bh_enable() can potentially call do_softirq(), which is not supposed to run when local IRQs are disabled. Let's fix this by using local_irq_save()/restore() instead. Note: hv_pci_onchannelcallback() is not a hot path because it's only called when the PCI device is hot added and removed, which is infrequent. Fixes: de0aa7b2f97d ("PCI: hv: Fix 2 hang issues in hv_compose_msi_msg()") Signed-off-by: Dexuan Cui Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Haiyang Zhang Cc: stable@vger.kernel.org Cc: Stephen Hemminger Cc: K. Y. Srinivasan --- drivers/pci/controller/pci-hyperv.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/pci/controller/pci-hyperv.c b/drivers/pci/controller/pci-hyperv.c index 6cc5036ac83c..f6325f1a89e8 100644 --- a/drivers/pci/controller/pci-hyperv.c +++ b/drivers/pci/controller/pci-hyperv.c @@ -1073,6 +1073,7 @@ static void hv_compose_msi_msg(struct irq_data *data, struct msi_msg *msg) struct pci_bus *pbus; struct pci_dev *pdev; struct cpumask *dest; + unsigned long flags; struct compose_comp_ctxt comp; struct tran_int_desc *int_desc; struct { @@ -1164,14 +1165,15 @@ static void hv_compose_msi_msg(struct irq_data *data, struct msi_msg *msg) * the channel callback directly when channel->target_cpu is * the current CPU. When the higher level interrupt code * calls us with interrupt enabled, let's add the - * local_bh_disable()/enable() to avoid race. + * local_irq_save()/restore() to avoid race: + * hv_pci_onchannelcallback() can also run in tasklet. */ - local_bh_disable(); + local_irq_save(flags); if (hbus->hdev->channel->target_cpu == smp_processor_id()) hv_pci_onchannelcallback(hbus); - local_bh_enable(); + local_irq_restore(flags); if (hpdev->state == hv_pcichild_ejecting) { dev_err_once(&hbus->hdev->device, From a83a2173441698f7ac9867b93bcca21cf18a032d Mon Sep 17 00:00:00 2001 From: Kishon Vijay Abraham I Date: Mon, 9 Jul 2018 15:18:56 -0500 Subject: [PATCH 14/25] PCI: endpoint: Fix NULL pointer dereference error when CONFIGFS is disabled commit ef1433f717a2 ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry") while adding configfs entry for each pci_epf_device_id table entry introduced a NULL pointer dereference error when CONFIG_PCI_ENDPOINT_CONFIGFS is not enabled. Fix it here. Fixes: ef1433f717a2 ("PCI: endpoint: Create configfs entry for each pci_epf_device_id table entry") Signed-off-by: Kishon Vijay Abraham I [lorenzo.pieralisi: updated commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas --- drivers/pci/endpoint/pci-epf-core.c | 62 ++++++++++++++++++++--------- 1 file changed, 43 insertions(+), 19 deletions(-) diff --git a/drivers/pci/endpoint/pci-epf-core.c b/drivers/pci/endpoint/pci-epf-core.c index bf53fad636a5..825fa24427a3 100644 --- a/drivers/pci/endpoint/pci-epf-core.c +++ b/drivers/pci/endpoint/pci-epf-core.c @@ -137,6 +137,20 @@ void *pci_epf_alloc_space(struct pci_epf *epf, size_t size, enum pci_barno bar) } EXPORT_SYMBOL_GPL(pci_epf_alloc_space); +static void pci_epf_remove_cfs(struct pci_epf_driver *driver) +{ + struct config_group *group, *tmp; + + if (!IS_ENABLED(CONFIG_PCI_ENDPOINT_CONFIGFS)) + return; + + mutex_lock(&pci_epf_mutex); + list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) + pci_ep_cfs_remove_epf_group(group); + list_del(&driver->epf_group); + mutex_unlock(&pci_epf_mutex); +} + /** * pci_epf_unregister_driver() - unregister the PCI EPF driver * @driver: the PCI EPF driver that has to be unregistered @@ -145,17 +159,38 @@ EXPORT_SYMBOL_GPL(pci_epf_alloc_space); */ void pci_epf_unregister_driver(struct pci_epf_driver *driver) { - struct config_group *group, *tmp; - - mutex_lock(&pci_epf_mutex); - list_for_each_entry_safe(group, tmp, &driver->epf_group, group_entry) - pci_ep_cfs_remove_epf_group(group); - list_del(&driver->epf_group); - mutex_unlock(&pci_epf_mutex); + pci_epf_remove_cfs(driver); driver_unregister(&driver->driver); } EXPORT_SYMBOL_GPL(pci_epf_unregister_driver); +static int pci_epf_add_cfs(struct pci_epf_driver *driver) +{ + struct config_group *group; + const struct pci_epf_device_id *id; + + if (!IS_ENABLED(CONFIG_PCI_ENDPOINT_CONFIGFS)) + return 0; + + INIT_LIST_HEAD(&driver->epf_group); + + id = driver->id_table; + while (id->name[0]) { + group = pci_ep_cfs_add_epf_group(id->name); + if (IS_ERR(group)) { + pci_epf_remove_cfs(driver); + return PTR_ERR(group); + } + + mutex_lock(&pci_epf_mutex); + list_add_tail(&group->group_entry, &driver->epf_group); + mutex_unlock(&pci_epf_mutex); + id++; + } + + return 0; +} + /** * __pci_epf_register_driver() - register a new PCI EPF driver * @driver: structure representing PCI EPF driver @@ -167,8 +202,6 @@ int __pci_epf_register_driver(struct pci_epf_driver *driver, struct module *owner) { int ret; - struct config_group *group; - const struct pci_epf_device_id *id; if (!driver->ops) return -EINVAL; @@ -183,16 +216,7 @@ int __pci_epf_register_driver(struct pci_epf_driver *driver, if (ret) return ret; - INIT_LIST_HEAD(&driver->epf_group); - - id = driver->id_table; - while (id->name[0]) { - group = pci_ep_cfs_add_epf_group(id->name); - mutex_lock(&pci_epf_mutex); - list_add_tail(&group->group_entry, &driver->epf_group); - mutex_unlock(&pci_epf_mutex); - id++; - } + pci_epf_add_cfs(driver); return 0; } From a5fb9fb023a1435f2b42bccd7f547560f3a21dc3 Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:40:26 -0500 Subject: [PATCH 15/25] PCI: OF: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. Introduce the devm_pci_remap_iospace() managed API and replace the pci_remap_iospace() call with it to fix the bug. Fixes: dbf9826d5797 ("PCI: generic: Convert to DT resource parsing API") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: split commit/updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij --- drivers/pci/of.c | 2 +- drivers/pci/pci.c | 38 ++++++++++++++++++++++++++++++++++++++ include/linux/pci.h | 2 ++ 3 files changed, 41 insertions(+), 1 deletion(-) diff --git a/drivers/pci/of.c b/drivers/pci/of.c index d088c9147f10..69a60d6ebd73 100644 --- a/drivers/pci/of.c +++ b/drivers/pci/of.c @@ -612,7 +612,7 @@ int pci_parse_request_of_pci_ranges(struct device *dev, switch (resource_type(res)) { case IORESOURCE_IO: - err = pci_remap_iospace(res, iobase); + err = devm_pci_remap_iospace(dev, res, iobase); if (err) { dev_warn(dev, "error %d: failed to map resource %pR\n", err, res); diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 97acba712e4e..316496e99da9 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -3579,6 +3579,44 @@ void pci_unmap_iospace(struct resource *res) } EXPORT_SYMBOL(pci_unmap_iospace); +static void devm_pci_unmap_iospace(struct device *dev, void *ptr) +{ + struct resource **res = ptr; + + pci_unmap_iospace(*res); +} + +/** + * devm_pci_remap_iospace - Managed pci_remap_iospace() + * @dev: Generic device to remap IO address for + * @res: Resource describing the I/O space + * @phys_addr: physical address of range to be mapped + * + * Managed pci_remap_iospace(). Map is automatically unmapped on driver + * detach. + */ +int devm_pci_remap_iospace(struct device *dev, const struct resource *res, + phys_addr_t phys_addr) +{ + const struct resource **ptr; + int error; + + ptr = devres_alloc(devm_pci_unmap_iospace, sizeof(*ptr), GFP_KERNEL); + if (!ptr) + return -ENOMEM; + + error = pci_remap_iospace(res, phys_addr); + if (error) { + devres_free(ptr); + } else { + *ptr = res; + devres_add(dev, ptr); + } + + return error; +} +EXPORT_SYMBOL(devm_pci_remap_iospace); + /** * devm_pci_remap_cfgspace - Managed pci_remap_cfgspace() * @dev: Generic device to remap IO address for diff --git a/include/linux/pci.h b/include/linux/pci.h index 340029b2fb38..abd5d5e17aee 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -1240,6 +1240,8 @@ int pci_register_io_range(struct fwnode_handle *fwnode, phys_addr_t addr, unsigned long pci_address_to_pio(phys_addr_t addr); phys_addr_t pci_pio_to_address(unsigned long pio); int pci_remap_iospace(const struct resource *res, phys_addr_t phys_addr); +int devm_pci_remap_iospace(struct device *dev, const struct resource *res, + phys_addr_t phys_addr); void pci_unmap_iospace(struct resource *res); void __iomem *devm_pci_remap_cfgspace(struct device *dev, resource_size_t offset, From 925652d03589084850023d8924b376b7ee3c1b1b Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:40:34 -0500 Subject: [PATCH 16/25] PCI: xgene: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. The X-Gene PCI controller driver has the same issue. Replace pci_remap_iospace() with the devm_ managed version so that the pages get unmapped automagically on any probe failure. Fixes: 5f6b6ccdbe1c ("PCI: xgene: Add APM X-Gene PCIe driver") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij --- drivers/pci/controller/pci-xgene.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pci-xgene.c b/drivers/pci/controller/pci-xgene.c index d854d67e873c..ffda3e8b4742 100644 --- a/drivers/pci/controller/pci-xgene.c +++ b/drivers/pci/controller/pci-xgene.c @@ -423,7 +423,7 @@ static int xgene_pcie_map_ranges(struct xgene_pcie_port *port, case IORESOURCE_IO: xgene_pcie_setup_ob_reg(port, res, OMR3BARL, io_base, res->start - window->offset); - ret = pci_remap_iospace(res, io_base); + ret = devm_pci_remap_iospace(dev, res, io_base); if (ret < 0) return ret; break; From 0018b265adf7e251f90d3ca1c7c0e32e2a0ad262 Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:40:40 -0500 Subject: [PATCH 17/25] PCI: versatile: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. The Versatile PCI controller driver has the same issue. Replace pci_remap_iospace() with the devm_ managed version to fix the bug. Fixes: b7e78170efd4 ("PCI: versatile: Add DT-based ARM Versatile PB PCIe host driver") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij --- drivers/pci/controller/pci-versatile.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pci-versatile.c b/drivers/pci/controller/pci-versatile.c index 994f32061b32..f59ad2728c0b 100644 --- a/drivers/pci/controller/pci-versatile.c +++ b/drivers/pci/controller/pci-versatile.c @@ -82,7 +82,7 @@ static int versatile_pci_parse_request_of_pci_ranges(struct device *dev, switch (resource_type(res)) { case IORESOURCE_IO: - err = pci_remap_iospace(res, iobase); + err = devm_pci_remap_iospace(dev, res, iobase); if (err) { dev_warn(dev, "error %d: failed to map resource %pR\n", err, res); From fd07f5e19c6fcdfa318944764248cf44eb06e532 Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:40:46 -0500 Subject: [PATCH 18/25] PCI: designware: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver is left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. The DesignWare PCIe controller driver has the same issue. Replace devm_pci_remap_iospace() with a devm_ managed version to fix the bug. Fixes: cbce7900598c ("PCI: designware: Make driver arch-agnostic") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij Acked-by: Jingoo Han --- drivers/pci/controller/dwc/pcie-designware-host.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/pci/controller/dwc/pcie-designware-host.c b/drivers/pci/controller/dwc/pcie-designware-host.c index 781aa03aeede..29a05759a294 100644 --- a/drivers/pci/controller/dwc/pcie-designware-host.c +++ b/drivers/pci/controller/dwc/pcie-designware-host.c @@ -363,7 +363,8 @@ int dw_pcie_host_init(struct pcie_port *pp) resource_list_for_each_entry_safe(win, tmp, &bridge->windows) { switch (resource_type(win->res)) { case IORESOURCE_IO: - ret = pci_remap_iospace(win->res, pp->io_base); + ret = devm_pci_remap_iospace(dev, win->res, + pp->io_base); if (ret) { dev_warn(dev, "Error %d: failed to map resource %pR\n", ret, win->res); From 1df3e5b3feebf29a3ecfa0c0f06f79544ca573e4 Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:40:53 -0500 Subject: [PATCH 19/25] PCI: aardvark: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. The Aardvark PCI controller driver has the same issue. Replace pci_remap_iospace() with its devm_ managed version to fix the bug. Fixes: 8c39d710363c ("PCI: aardvark: Add Aardvark PCI host controller driver") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Thomas Petazzoni Reviewed-by: Linus Walleij --- drivers/pci/controller/pci-aardvark.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pci-aardvark.c b/drivers/pci/controller/pci-aardvark.c index d3172d5d3d35..0fae816fba39 100644 --- a/drivers/pci/controller/pci-aardvark.c +++ b/drivers/pci/controller/pci-aardvark.c @@ -849,7 +849,7 @@ static int advk_pcie_parse_request_of_pci_ranges(struct advk_pcie *pcie) 0, 0xF8000000, 0, lower_32_bits(res->start), OB_PCIE_IO); - err = pci_remap_iospace(res, iobase); + err = devm_pci_remap_iospace(dev, res, iobase); if (err) { dev_warn(dev, "error %d: failed to map resource %pR\n", err, res); From e30609454b39139a91faf631685f503b7ea3f27d Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:40:59 -0500 Subject: [PATCH 20/25] PCI: faraday: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. The Faraday PCI driver has the same issue. Replace pci_remap_iospace() with its devm_ managed version to fix the bug. Fixes: d3c68e0a7e34 ("PCI: faraday: Add Faraday Technology FTPCI100 PCI Host Bridge driver") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij --- drivers/pci/controller/pci-ftpci100.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pci-ftpci100.c b/drivers/pci/controller/pci-ftpci100.c index 20bb2564a6b3..bf5ece5d9291 100644 --- a/drivers/pci/controller/pci-ftpci100.c +++ b/drivers/pci/controller/pci-ftpci100.c @@ -503,7 +503,7 @@ static int faraday_pci_probe(struct platform_device *pdev) dev_err(dev, "illegal IO mem size\n"); return -EINVAL; } - ret = pci_remap_iospace(io, io_base); + ret = devm_pci_remap_iospace(dev, io, io_base); if (ret) { dev_warn(dev, "error %d: failed to map resource %pR\n", ret, io); From 438477b9a089e45ea7c8fb549553e52833117ed8 Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:41:05 -0500 Subject: [PATCH 21/25] PCI: mediatek: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. The MediaTek PCIe driver has the same issue. Replace devm_pci_remap_iospace() with its devm_ managed counterpart to fix the bug. Fixes: 637cfacae96f ("PCI: mediatek: Add MediaTek PCIe host controller support") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij --- drivers/pci/controller/pcie-mediatek.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pcie-mediatek.c b/drivers/pci/controller/pcie-mediatek.c index 0baabe30858f..861dda69f366 100644 --- a/drivers/pci/controller/pcie-mediatek.c +++ b/drivers/pci/controller/pcie-mediatek.c @@ -1109,7 +1109,7 @@ static int mtk_pcie_request_resources(struct mtk_pcie *pcie) if (err < 0) return err; - pci_remap_iospace(&pcie->pio, pcie->io.start); + devm_pci_remap_iospace(dev, &pcie->pio, pcie->io.start); return 0; } From 270ed733e68955049b693bea8f4a1efb293a96ae Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Wed, 18 Jul 2018 15:41:11 -0500 Subject: [PATCH 22/25] PCI: v3-semi: Fix I/O space page leak When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. The V3 Semiconductor PCI driver has the same issue. Replace devm_pci_remap_iospace() with its devm_ managed version to fix the bug. Fixes: 68a15eb7bd0c ("PCI: v3-semi: Add V3 Semiconductor PCI host driver") Signed-off-by: Sergei Shtylyov [lorenzo.pieralisi@arm.com: updated the commit log] Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas Reviewed-by: Linus Walleij --- drivers/pci/controller/pci-v3-semi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pci-v3-semi.c b/drivers/pci/controller/pci-v3-semi.c index 68b8bfbdb867..d219404bad92 100644 --- a/drivers/pci/controller/pci-v3-semi.c +++ b/drivers/pci/controller/pci-v3-semi.c @@ -537,7 +537,7 @@ static int v3_pci_setup_resource(struct v3_pci *v3, v3->io_bus_addr = io->start - win->offset; dev_dbg(dev, "I/O window %pR, bus addr %pap\n", io, &v3->io_bus_addr); - ret = pci_remap_iospace(io, io_base); + ret = devm_pci_remap_iospace(dev, io, io_base); if (ret) { dev_warn(dev, "error %d: failed to map resource %pR\n", From bd91b56cb3b27492963caeb5fccefe20a986ca8d Mon Sep 17 00:00:00 2001 From: Thomas Tai Date: Thu, 26 Jul 2018 12:13:04 -0500 Subject: [PATCH 23/25] PCI/AER: Work around use-after-free in pcie_do_fatal_recovery() When an fatal error is received by a non-bridge device, the device is removed, and pci_stop_and_remove_bus_device() deallocates the device structure. The freed device structure is used by subsequent code to send uevents and print messages. Hold a reference on the device until we're finished using it. This is not an ideal fix because pcie_do_fatal_recovery() should not use the device at all after removing it, but that's too big a project for right now. Fixes: 7e9084b36740 ("PCI/AER: Handle ERR_FATAL with removal and re-enumeration of devices") Signed-off-by: Thomas Tai [bhelgaas: changelog, reduce get/put coverage] Signed-off-by: Bjorn Helgaas --- drivers/pci/pcie/err.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/pci/pcie/err.c b/drivers/pci/pcie/err.c index f7ce0cb0b0b7..f02e334beb45 100644 --- a/drivers/pci/pcie/err.c +++ b/drivers/pci/pcie/err.c @@ -295,6 +295,7 @@ void pcie_do_fatal_recovery(struct pci_dev *dev, u32 service) parent = udev->subordinate; pci_lock_rescan_remove(); + pci_dev_get(dev); list_for_each_entry_safe_reverse(pdev, temp, &parent->devices, bus_list) { pci_dev_get(pdev); @@ -328,6 +329,7 @@ void pcie_do_fatal_recovery(struct pci_dev *dev, u32 service) pci_info(dev, "Device recovery from fatal error failed\n"); } + pci_dev_put(dev); pci_unlock_rescan_remove(); } From a54e43f993f8ec2f063b616a0e4d2b09e08d78a5 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 27 Jul 2018 17:10:39 -0500 Subject: [PATCH 24/25] PCI: mobiveil: Avoid integer overflow in IB_WIN_SIZE IB_WIN_SIZE is larger than INT_MAX so we need to cast it to u64. Fixes: 9af6bcb11e12 ("PCI: mobiveil: Add Mobiveil PCIe Host Bridge IP driver") Signed-off-by: Dan Carpenter Signed-off-by: Lorenzo Pieralisi Signed-off-by: Bjorn Helgaas --- drivers/pci/controller/pcie-mobiveil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/pcie-mobiveil.c b/drivers/pci/controller/pcie-mobiveil.c index 4d6c20e47bed..cf0aa7cee5b0 100644 --- a/drivers/pci/controller/pcie-mobiveil.c +++ b/drivers/pci/controller/pcie-mobiveil.c @@ -107,7 +107,7 @@ #define CFG_WINDOW_TYPE 0 #define IO_WINDOW_TYPE 1 #define MEM_WINDOW_TYPE 2 -#define IB_WIN_SIZE (256 * 1024 * 1024 * 1024) +#define IB_WIN_SIZE ((u64)256 * 1024 * 1024 * 1024) #define MAX_PIO_WINDOWS 8 /* Parameters for the waiting for link up routine */ From 44bda4b7d26e9fffed6d7152d98a2e9edaeb2a76 Mon Sep 17 00:00:00 2001 From: Hari Vyas Date: Tue, 3 Jul 2018 14:35:41 +0530 Subject: [PATCH 25/25] PCI: Fix is_added/is_busmaster race condition When a PCI device is detected, pdev->is_added is set to 1 and proc and sysfs entries are created. When the device is removed, pdev->is_added is checked for one and then device is detached with clearing of proc and sys entries and at end, pdev->is_added is set to 0. is_added and is_busmaster are bit fields in pci_dev structure sharing same memory location. A strange issue was observed with multiple removal and rescan of a PCIe NVMe device using sysfs commands where is_added flag was observed as zero instead of one while removing device and proc,sys entries are not cleared. This causes issue in later device addition with warning message "proc_dir_entry" already registered. Debugging revealed a race condition between the PCI core setting the is_added bit in pci_bus_add_device() and the NVMe driver reset work-queue setting the is_busmaster bit in pci_set_master(). As these fields are not handled atomically, that clears the is_added bit. Move the is_added bit to a separate private flag variable and use atomic functions to set and retrieve the device addition state. This avoids the race because is_added no longer shares a memory location with is_busmaster. Link: https://bugzilla.kernel.org/show_bug.cgi?id=200283 Signed-off-by: Hari Vyas Signed-off-by: Bjorn Helgaas Reviewed-by: Lukas Wunner Acked-by: Michael Ellerman --- arch/powerpc/kernel/pci-common.c | 4 +++- arch/powerpc/platforms/powernv/pci-ioda.c | 3 ++- arch/powerpc/platforms/pseries/setup.c | 3 ++- drivers/pci/bus.c | 6 +++--- drivers/pci/hotplug/acpiphp_glue.c | 2 +- drivers/pci/pci.h | 11 +++++++++++ drivers/pci/probe.c | 4 ++-- drivers/pci/remove.c | 5 +++-- include/linux/pci.h | 1 - 9 files changed, 27 insertions(+), 12 deletions(-) diff --git a/arch/powerpc/kernel/pci-common.c b/arch/powerpc/kernel/pci-common.c index fe9733ffffaa..471aac313b89 100644 --- a/arch/powerpc/kernel/pci-common.c +++ b/arch/powerpc/kernel/pci-common.c @@ -42,6 +42,8 @@ #include #include +#include "../../../drivers/pci/pci.h" + /* hose_spinlock protects accesses to the the phb_bitmap. */ static DEFINE_SPINLOCK(hose_spinlock); LIST_HEAD(hose_list); @@ -1014,7 +1016,7 @@ void pcibios_setup_bus_devices(struct pci_bus *bus) /* Cardbus can call us to add new devices to a bus, so ignore * those who are already fully discovered */ - if (dev->is_added) + if (pci_dev_is_added(dev)) continue; pcibios_setup_device(dev); diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c index 5bd0eb6681bc..70b2e1e0f23c 100644 --- a/arch/powerpc/platforms/powernv/pci-ioda.c +++ b/arch/powerpc/platforms/powernv/pci-ioda.c @@ -46,6 +46,7 @@ #include "powernv.h" #include "pci.h" +#include "../../../../drivers/pci/pci.h" #define PNV_IODA1_M64_NUM 16 /* Number of M64 BARs */ #define PNV_IODA1_M64_SEGS 8 /* Segments per M64 BAR */ @@ -3138,7 +3139,7 @@ static void pnv_pci_ioda_fixup_iov_resources(struct pci_dev *pdev) struct pci_dn *pdn; int mul, total_vfs; - if (!pdev->is_physfn || pdev->is_added) + if (!pdev->is_physfn || pci_dev_is_added(pdev)) return; pdn = pci_get_pdn(pdev); diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c index 139f0af6c3d9..8a4868a3964b 100644 --- a/arch/powerpc/platforms/pseries/setup.c +++ b/arch/powerpc/platforms/pseries/setup.c @@ -71,6 +71,7 @@ #include #include "pseries.h" +#include "../../../../drivers/pci/pci.h" int CMO_PrPSP = -1; int CMO_SecPSP = -1; @@ -664,7 +665,7 @@ static void pseries_pci_fixup_iov_resources(struct pci_dev *pdev) const int *indexes; struct device_node *dn = pci_device_to_OF_node(pdev); - if (!pdev->is_physfn || pdev->is_added) + if (!pdev->is_physfn || pci_dev_is_added(pdev)) return; /*Firmware must support open sriov otherwise dont configure*/ indexes = of_get_property(dn, "ibm,open-sriov-vf-bar-info", NULL); diff --git a/drivers/pci/bus.c b/drivers/pci/bus.c index 35b7fc87eac5..5cb40b2518f9 100644 --- a/drivers/pci/bus.c +++ b/drivers/pci/bus.c @@ -330,7 +330,7 @@ void pci_bus_add_device(struct pci_dev *dev) return; } - dev->is_added = 1; + pci_dev_assign_added(dev, true); } EXPORT_SYMBOL_GPL(pci_bus_add_device); @@ -347,14 +347,14 @@ void pci_bus_add_devices(const struct pci_bus *bus) list_for_each_entry(dev, &bus->devices, bus_list) { /* Skip already-added devices */ - if (dev->is_added) + if (pci_dev_is_added(dev)) continue; pci_bus_add_device(dev); } list_for_each_entry(dev, &bus->devices, bus_list) { /* Skip if device attach failed */ - if (!dev->is_added) + if (!pci_dev_is_added(dev)) continue; child = dev->subordinate; if (child) diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c index 3a17b290df5d..ef0b1b6ba86f 100644 --- a/drivers/pci/hotplug/acpiphp_glue.c +++ b/drivers/pci/hotplug/acpiphp_glue.c @@ -509,7 +509,7 @@ static void enable_slot(struct acpiphp_slot *slot) list_for_each_entry(dev, &bus->devices, bus_list) { /* Assume that newly added devices are powered on already. */ - if (!dev->is_added) + if (!pci_dev_is_added(dev)) dev->current_state = PCI_D0; } diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h index 882f1f9596df..08817253c8a2 100644 --- a/drivers/pci/pci.h +++ b/drivers/pci/pci.h @@ -288,6 +288,7 @@ struct pci_sriov { /* pci_dev priv_flags */ #define PCI_DEV_DISCONNECTED 0 +#define PCI_DEV_ADDED 1 static inline int pci_dev_set_disconnected(struct pci_dev *dev, void *unused) { @@ -300,6 +301,16 @@ static inline bool pci_dev_is_disconnected(const struct pci_dev *dev) return test_bit(PCI_DEV_DISCONNECTED, &dev->priv_flags); } +static inline void pci_dev_assign_added(struct pci_dev *dev, bool added) +{ + assign_bit(PCI_DEV_ADDED, &dev->priv_flags, added); +} + +static inline bool pci_dev_is_added(const struct pci_dev *dev) +{ + return test_bit(PCI_DEV_ADDED, &dev->priv_flags); +} + #ifdef CONFIG_PCI_ATS void pci_restore_ats_state(struct pci_dev *dev); #else diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index ac876e32de4b..611adcd9c169 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -2433,13 +2433,13 @@ int pci_scan_slot(struct pci_bus *bus, int devfn) dev = pci_scan_single_device(bus, devfn); if (!dev) return 0; - if (!dev->is_added) + if (!pci_dev_is_added(dev)) nr++; for (fn = next_fn(bus, dev, 0); fn > 0; fn = next_fn(bus, dev, fn)) { dev = pci_scan_single_device(bus, devfn + fn); if (dev) { - if (!dev->is_added) + if (!pci_dev_is_added(dev)) nr++; dev->multifunction = 1; } diff --git a/drivers/pci/remove.c b/drivers/pci/remove.c index 6f072eae4f7a..5e3d0dced2b8 100644 --- a/drivers/pci/remove.c +++ b/drivers/pci/remove.c @@ -19,11 +19,12 @@ static void pci_stop_dev(struct pci_dev *dev) { pci_pme_active(dev, false); - if (dev->is_added) { + if (pci_dev_is_added(dev)) { device_release_driver(&dev->dev); pci_proc_detach_device(dev); pci_remove_sysfs_dev_files(dev); - dev->is_added = 0; + + pci_dev_assign_added(dev, false); } if (dev->bus->self) diff --git a/include/linux/pci.h b/include/linux/pci.h index abd5d5e17aee..c133ccfa002e 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -368,7 +368,6 @@ struct pci_dev { unsigned int transparent:1; /* Subtractive decode bridge */ unsigned int multifunction:1; /* Multi-function device */ - unsigned int is_added:1; unsigned int is_busmaster:1; /* Is busmaster */ unsigned int no_msi:1; /* May not use MSI */ unsigned int no_64bit_msi:1; /* May only use 32-bit MSIs */