| Message ID | 20171226023646.17722-5-jeffy.chen@rock-chips.com (mailing list archive) |
|---|---|
| State | New, archived |
| Delegated to: | Bjorn Helgaas |
| Headers | show |
On Tue, Dec 26, 2017 at 10:36:45AM +0800, Jeffy Chen wrote: > Add of_pci_setup_wake_irq() and of_pci_teardown_wake_irq() to handle > the PCIe WAKE# interrupt. > > Also use the dedicated wakeirq infrastructure to simplify it. > > And add pci-of.c to enable/disable the wakeup irq in noirq stage to > avoid possible irq storm. > > Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com> > --- > > Changes in v13: > Fix compiler error reported by kbuild test robot <fengguang.wu@intel.com> > > Changes in v12: > Enable the wake irq in noirq stage to avoid possible irq storm. > > Changes in v11: > Only support 1-per-device PCIe WAKE# pin as suggested. > > Changes in v10: > Use device_set_wakeup_capable() instead of device_set_wakeup_enable(), > since dedicated wakeirq will be lost in device_set_wakeup_enable(false). > > Changes in v9: > Fix check error in .cleanup(). > Move dedicated wakeirq setup to setup() callback and use > device_set_wakeup_enable() to enable/disable. > > Changes in v8: > Add pci-of.c and use platform_pm_ops to handle the PCIe WAKE# signal. > > Changes in v7: > Move PCIE_WAKE handling into pci core. > > Changes in v6: > Fix device_init_wake error handling, and add some comments. > > Changes in v5: > Rebase. > > Changes in v3: > Fix error handling. > > Changes in v2: > Use dev_pm_set_dedicated_wake_irq. > > drivers/of/of_pci_irq.c | 49 +++++++++++++++++++++++++++++++ Please move this to drivers/pci/of.c (or perhaps create pci/of_irq.c). > drivers/pci/Makefile | 1 + > drivers/pci/pci-driver.c | 10 +++++++ > drivers/pci/pci-of.c | 75 ++++++++++++++++++++++++++++++++++++++++++++++++ We already have drivers/pci/of.c. It's not clear what the difference is from the filenames. Either merge with of.c or perhaps of-pm.c. > include/linux/of_pci.h | 9 ++++++ > 5 files changed, 144 insertions(+) > create mode 100644 drivers/pci/pci-of.c > > diff --git a/drivers/of/of_pci_irq.c b/drivers/of/of_pci_irq.c > index d39565d5477b..abec3a44853b 100644 > --- a/drivers/of/of_pci_irq.c > +++ b/drivers/of/of_pci_irq.c > @@ -1,8 +1,57 @@ > #include <linux/kernel.h> > #include <linux/of_pci.h> > #include <linux/of_irq.h> > +#include <linux/pm_wakeirq.h> > #include <linux/export.h> > > +int of_pci_setup_wake_irq(struct pci_dev *pdev) > +{ > + struct pci_dev *ppdev; > + struct device_node *dn; > + int ret, irq; > + > + /* Get the pci_dev of our parent. Hopefully it's a port. */ > + ppdev = pdev->bus->self; > + /* Nope, it's a host bridge. */ > + if (!ppdev) > + return 0; > + > + dn = pci_device_to_OF_node(ppdev); > + if (!dn) > + return 0; > + > + irq = of_irq_get_byname(dn, "wakeup"); > + if (irq == -EPROBE_DEFER) { > + return irq; > + } else if (irq < 0) { > + /* Ignore other errors, since a missing wakeup is non-fatal. */ > + dev_info(&pdev->dev, "cannot get wakeup interrupt: %d\n", irq); > + return 0; > + } > + > + device_init_wakeup(&pdev->dev, true); > + > + ret = dev_pm_set_dedicated_wake_irq(&pdev->dev, irq); > + if (ret < 0) { > + dev_err(&pdev->dev, "failed to set wake IRQ: %d\n", ret); > + device_init_wakeup(&pdev->dev, false); > + return ret; > + } > + > + /* Start out disabled to avoid irq storm */ > + dev_pm_disable_wake_irq(&pdev->dev); > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(of_pci_setup_wake_irq); > + > +void of_pci_teardown_wake_irq(struct pci_dev *pdev) > +{ > + dev_pm_clear_wake_irq(&pdev->dev); > + device_init_wakeup(&pdev->dev, false); > +} > +EXPORT_SYMBOL_GPL(of_pci_teardown_wake_irq); > + > /** > * of_irq_parse_pci - Resolve the interrupt for a PCI device > * @pdev: the device whose interrupt is to be resolved > diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile > index c7819b973df7..d0182c82162a 100644 > --- a/drivers/pci/Makefile > +++ b/drivers/pci/Makefile > @@ -29,6 +29,7 @@ obj-$(CONFIG_PCI_IOV) += iov.o > # ACPI _DSM provided firmware instance and string name > # > obj-$(CONFIG_ACPI) += pci-acpi.o > +obj-$(CONFIG_OF) += pci-of.o > > # SMBIOS provided firmware instance and labels > obj-$(CONFIG_PCI_LABEL) += pci-label.o > diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c > index d79dbc377b9c..b4475ff35d97 100644 > --- a/drivers/pci/pci-driver.c > +++ b/drivers/pci/pci-driver.c > @@ -17,6 +17,7 @@ > #include <linux/slab.h> > #include <linux/sched.h> > #include <linux/cpu.h> > +#include <linux/of_pci.h> > #include <linux/pm_runtime.h> > #include <linux/suspend.h> > #include <linux/kexec.h> > @@ -421,10 +422,17 @@ static int pci_device_probe(struct device *dev) > if (error < 0) > return error; > > + error = of_pci_setup_wake_irq(pci_dev); > + if (error < 0) { > + pcibios_free_irq(pci_dev); > + return error; > + } > + > pci_dev_get(pci_dev); > if (pci_device_can_probe(pci_dev)) { > error = __pci_device_probe(drv, pci_dev); > if (error) { > + of_pci_teardown_wake_irq(pci_dev); > pcibios_free_irq(pci_dev); > pci_dev_put(pci_dev); > } > @@ -438,6 +446,8 @@ static int pci_device_remove(struct device *dev) > struct pci_dev *pci_dev = to_pci_dev(dev); > struct pci_driver *drv = pci_dev->driver; > > + of_pci_teardown_wake_irq(pci_dev); > + > if (drv) { > if (drv->remove) { > pm_runtime_get_sync(dev); > diff --git a/drivers/pci/pci-of.c b/drivers/pci/pci-of.c > new file mode 100644 > index 000000000000..ad413b2de508 > --- /dev/null > +++ b/drivers/pci/pci-of.c > @@ -0,0 +1,75 @@ > +/* > + * File: pci-of.c > + * Purpose: Provide PCI PM/wakeup support in OF systems > + * > + * Copyright (C) 2017 Google, Inc. > + * Brian Norris <briannorris@chromium.org> > + * > + * Copyright (C) 2017 Rockchip Electronics Co., Ltd. > + * Jeffy Chen <jeffy.chen@rock-chips.com> > + * > + * This program is free software; you can redistribute it and/or modify it > + * under the terms and conditions of the GNU General Public License, > + * version 2, as published by the Free Software Foundation. > + */ > + > +#include <linux/acpi.h> > +#include <linux/init.h> > +#include <linux/module.h> > +#include <linux/pci.h> > +#include <linux/pm_wakeirq.h> > +#include "pci.h" > + > +static bool of_pci_power_manageable(struct pci_dev *dev) > +{ > + return false; > +} > + > +static int of_pci_set_power_state(struct pci_dev *dev, pci_power_t state) > +{ > + return -ENOSYS; > +} > + > +static pci_power_t of_pci_get_power_state(struct pci_dev *dev) > +{ > + return PCI_UNKNOWN; > +} > + > +static pci_power_t of_pci_choose_state(struct pci_dev *pdev) > +{ > + return PCI_POWER_ERROR; > +} > + > +static int of_pci_wakeup(struct pci_dev *dev, bool enable) > +{ > + if (enable) > + dev_pm_enable_wake_irq(&dev->dev); > + else > + dev_pm_disable_wake_irq(&dev->dev); > + return 0; > +} > + > +static bool of_pci_need_resume(struct pci_dev *dev) > +{ > + return false; > +} > + > +static const struct pci_platform_pm_ops of_pci_platform_pm = { > + .is_manageable = of_pci_power_manageable, > + .set_state = of_pci_set_power_state, > + .get_state = of_pci_get_power_state, > + .choose_state = of_pci_choose_state, > + .set_wakeup = of_pci_wakeup, > + .need_resume = of_pci_need_resume, > +}; > + > +static int __init of_pci_init(void) > +{ > + if (!acpi_disabled) > + return 0; > + > + pci_set_platform_pm(&of_pci_platform_pm); I guess no DT based system will override this? > + > + return 0; > +} > +arch_initcall(of_pci_init);
Hi Rob, Thanks for your reply. On 12/27/2017 07:56 AM, Rob Herring wrote: >> > >> > drivers/of/of_pci_irq.c | 49 +++++++++++++++++++++++++++++++ > Please move this to drivers/pci/of.c (or perhaps create pci/of_irq.c). > >> > drivers/pci/Makefile | 1 + >> > drivers/pci/pci-driver.c | 10 +++++++ >> > drivers/pci/pci-of.c | 75 ++++++++++++++++++++++++++++++++++++++++++++++++ > We already have drivers/pci/of.c. It's not clear what the difference is > from the filenames. Either merge with of.c or perhaps of-pm.c. this new file does something similar to the pci-acpi.c and pci-mid.c.. and i am agree the naming is not clear, maybe we can rename both of those files to something like pci-pm-***.c? Hi Rafael, do you think this would make sense? >> > include/linux/of_pci.h | 9 ++++++ >> > 5 files changed, 144 insertions(+) >> > create mode 100644 drivers/pci/pci-of.c >> > >> >diff --git a/drivers/of/of_pci_irq.c b/drivers/of/of_pci_irq.c >> >index d39565d5477b..abec3a44853b 100644 >> >--- a/drivers/of/of_pci_irq.c >> >+++ b/drivers/of/of_pci_irq.c >> >@@ -1,8 +1,57 @@ >> > #include <linux/kernel.h> >> > #include <linux/of_pci.h> >> > #include <linux/of_irq.h> >> >+#include <linux/pm_wakeirq.h> >> > #include <linux/export.h> >> > >> >+int of_pci_setup_wake_irq(struct pci_dev *pdev) >> >+{ >> >+ struct pci_dev *ppdev; >> >+ struct device_node *dn; >> >+ int ret, irq; >> >+ >> >+ /* Get the pci_dev of our parent. Hopefully it's a port. */ >> >+ ppdev = pdev->bus->self; >> >+ /* Nope, it's a host bridge. */ >> >+ if (!ppdev) >> >+ return 0; >> >+ >> >+ dn = pci_device_to_OF_node(ppdev); >> >+ if (!dn) >> >+ return 0; >> >+ >> >+ irq = of_irq_get_byname(dn, "wakeup"); >> >+ if (irq == -EPROBE_DEFER) { >> >+ return irq; >> >+ } else if (irq < 0) { >> >+ /* Ignore other errors, since a missing wakeup is non-fatal. */ >> >+ dev_info(&pdev->dev, "cannot get wakeup interrupt: %d\n", irq); >> >+ return 0; >> >+ } >> >+ >> >+ device_init_wakeup(&pdev->dev, true); >> >+ >> >+ ret = dev_pm_set_dedicated_wake_irq(&pdev->dev, irq); >> >+ if (ret < 0) { >> >+ dev_err(&pdev->dev, "failed to set wake IRQ: %d\n", ret); >> >+ device_init_wakeup(&pdev->dev, false); >> >+ return ret; >> >+ } >> >+ >> >+ /* Start out disabled to avoid irq storm */ >> >+ dev_pm_disable_wake_irq(&pdev->dev); >> >+ >> >+ return 0; >> >+} >> >+EXPORT_SYMBOL_GPL(of_pci_setup_wake_irq); >> >+ >> >+void of_pci_teardown_wake_irq(struct pci_dev *pdev) >> >+{ >> >+ dev_pm_clear_wake_irq(&pdev->dev); >> >+ device_init_wakeup(&pdev->dev, false); >> >+} >> >+EXPORT_SYMBOL_GPL(of_pci_teardown_wake_irq); >> >+ >> > /** >> > * of_irq_parse_pci - Resolve the interrupt for a PCI device >> > * @pdev: the device whose interrupt is to be resolved >> >diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile >> >index c7819b973df7..d0182c82162a 100644 >> >--- a/drivers/pci/Makefile >> >+++ b/drivers/pci/Makefile >> >@@ -29,6 +29,7 @@ obj-$(CONFIG_PCI_IOV) += iov.o >> > # ACPI _DSM provided firmware instance and string name >> > # >> > obj-$(CONFIG_ACPI) += pci-acpi.o >> >+obj-$(CONFIG_OF) += pci-of.o >> > >> > # SMBIOS provided firmware instance and labels >> > obj-$(CONFIG_PCI_LABEL) += pci-label.o >> >diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c >> >index d79dbc377b9c..b4475ff35d97 100644 >> >--- a/drivers/pci/pci-driver.c >> >+++ b/drivers/pci/pci-driver.c >> >@@ -17,6 +17,7 @@ >> > #include <linux/slab.h> >> > #include <linux/sched.h> >> > #include <linux/cpu.h> >> >+#include <linux/of_pci.h> >> > #include <linux/pm_runtime.h> >> > #include <linux/suspend.h> >> > #include <linux/kexec.h> >> >@@ -421,10 +422,17 @@ static int pci_device_probe(struct device *dev) >> > if (error < 0) >> > return error; >> > >> >+ error = of_pci_setup_wake_irq(pci_dev); >> >+ if (error < 0) { >> >+ pcibios_free_irq(pci_dev); >> >+ return error; >> >+ } >> >+ >> > pci_dev_get(pci_dev); >> > if (pci_device_can_probe(pci_dev)) { >> > error = __pci_device_probe(drv, pci_dev); >> > if (error) { >> >+ of_pci_teardown_wake_irq(pci_dev); >> > pcibios_free_irq(pci_dev); >> > pci_dev_put(pci_dev); >> > } >> >@@ -438,6 +446,8 @@ static int pci_device_remove(struct device *dev) >> > struct pci_dev *pci_dev = to_pci_dev(dev); >> > struct pci_driver *drv = pci_dev->driver; >> > >> >+ of_pci_teardown_wake_irq(pci_dev); >> >+ >> > if (drv) { >> > if (drv->remove) { >> > pm_runtime_get_sync(dev); >> >diff --git a/drivers/pci/pci-of.c b/drivers/pci/pci-of.c >> >new file mode 100644 >> >index 000000000000..ad413b2de508 >> >--- /dev/null >> >+++ b/drivers/pci/pci-of.c >> >@@ -0,0 +1,75 @@ >> >+/* >> >+ * File: pci-of.c >> >+ * Purpose: Provide PCI PM/wakeup support in OF systems >> >+ * >> >+ * Copyright (C) 2017 Google, Inc. >> >+ * Brian Norris<briannorris@chromium.org> >> >+ * >> >+ * Copyright (C) 2017 Rockchip Electronics Co., Ltd. >> >+ * Jeffy Chen<jeffy.chen@rock-chips.com> >> >+ * >> >+ * This program is free software; you can redistribute it and/or modify it >> >+ * under the terms and conditions of the GNU General Public License, >> >+ * version 2, as published by the Free Software Foundation. >> >+ */ >> >+ >> >+#include <linux/acpi.h> >> >+#include <linux/init.h> >> >+#include <linux/module.h> >> >+#include <linux/pci.h> >> >+#include <linux/pm_wakeirq.h> >> >+#include "pci.h" >> >+ >> >+static bool of_pci_power_manageable(struct pci_dev *dev) >> >+{ >> >+ return false; >> >+} >> >+ >> >+static int of_pci_set_power_state(struct pci_dev *dev, pci_power_t state) >> >+{ >> >+ return -ENOSYS; >> >+} >> >+ >> >+static pci_power_t of_pci_get_power_state(struct pci_dev *dev) >> >+{ >> >+ return PCI_UNKNOWN; >> >+} >> >+ >> >+static pci_power_t of_pci_choose_state(struct pci_dev *pdev) >> >+{ >> >+ return PCI_POWER_ERROR; >> >+} >> >+ >> >+static int of_pci_wakeup(struct pci_dev *dev, bool enable) >> >+{ >> >+ if (enable) >> >+ dev_pm_enable_wake_irq(&dev->dev); >> >+ else >> >+ dev_pm_disable_wake_irq(&dev->dev); >> >+ return 0; >> >+} >> >+ >> >+static bool of_pci_need_resume(struct pci_dev *dev) >> >+{ >> >+ return false; >> >+} >> >+ >> >+static const struct pci_platform_pm_ops of_pci_platform_pm = { >> >+ .is_manageable = of_pci_power_manageable, >> >+ .set_state = of_pci_set_power_state, >> >+ .get_state = of_pci_get_power_state, >> >+ .choose_state = of_pci_choose_state, >> >+ .set_wakeup = of_pci_wakeup, >> >+ .need_resume = of_pci_need_resume, >> >+}; >> >+ >> >+static int __init of_pci_init(void) >> >+{ >> >+ if (!acpi_disabled) >> >+ return 0; >> >+ >> >+ pci_set_platform_pm(&of_pci_platform_pm); > I guess no DT based system will override this? i think the !acpi_disabled means acpi been disabled or CONFIG_ACPI is undefined. and pci-acpi.c would only work when we have CONFIG_ACPI. but i have no idea about pci-mid.c or would it possible to have more platform pm ops in the future...maybe we should add some dependency in the Kconfig? >
On Tue, Dec 26, 2017 at 7:32 PM, JeffyChen <jeffy.chen@rock-chips.com> wrote: > Hi Rob, > > Thanks for your reply. > > On 12/27/2017 07:56 AM, Rob Herring wrote: >>> >>> > >>> > drivers/of/of_pci_irq.c | 49 +++++++++++++++++++++++++++++++ >> >> Please move this to drivers/pci/of.c (or perhaps create pci/of_irq.c). >> >>> > drivers/pci/Makefile | 1 + >>> > drivers/pci/pci-driver.c | 10 +++++++ >>> > drivers/pci/pci-of.c | 75 >>> > ++++++++++++++++++++++++++++++++++++++++++++++++ >> >> We already have drivers/pci/of.c. It's not clear what the difference is >> from the filenames. Either merge with of.c or perhaps of-pm.c. > > > this new file does something similar to the pci-acpi.c and pci-mid.c.. pci-acpi.c has similar things to pci/of.c. The naming is just not consistent. Also, I plan to move the rest of drivers/of/of_pci* to drivers/pci. > and i am agree the naming is not clear, maybe we can rename both of those > files to something like pci-pm-***.c? At least pci-acpi.c is more than just PM functions, so that doesn't make sense. Given that all the ACPI related functions are in 1 file, we should do the same for DT. > > Hi Rafael, do you think this would make sense? > > [...] >>> >+static int __init of_pci_init(void) >>> >+{ >>> >+ if (!acpi_disabled) >>> >+ return 0; >>> >+ >>> >+ pci_set_platform_pm(&of_pci_platform_pm); >> >> I guess no DT based system will override this? > > > i think the !acpi_disabled means acpi been disabled or CONFIG_ACPI is > undefined. > > and pci-acpi.c would only work when we have CONFIG_ACPI. > > but i have no idea about pci-mid.c or would it possible to have more > platform pm ops in the future...maybe we should add some dependency in the > Kconfig? It's probably fine given there are only 2 other implementations so far. Rob
On Wed, Dec 27, 2017 at 4:30 PM, Rob Herring <robh@kernel.org> wrote: > On Tue, Dec 26, 2017 at 7:32 PM, JeffyChen <jeffy.chen@rock-chips.com> wrote: >> Hi Rob, >> >> Thanks for your reply. >> >> On 12/27/2017 07:56 AM, Rob Herring wrote: >>>> >>>> > >>>> > drivers/of/of_pci_irq.c | 49 +++++++++++++++++++++++++++++++ >>> >>> Please move this to drivers/pci/of.c (or perhaps create pci/of_irq.c). >>> >>>> > drivers/pci/Makefile | 1 + >>>> > drivers/pci/pci-driver.c | 10 +++++++ >>>> > drivers/pci/pci-of.c | 75 >>>> > ++++++++++++++++++++++++++++++++++++++++++++++++ >>> >>> We already have drivers/pci/of.c. It's not clear what the difference is >>> from the filenames. Either merge with of.c or perhaps of-pm.c. >> >> >> this new file does something similar to the pci-acpi.c and pci-mid.c.. > > pci-acpi.c has similar things to pci/of.c. The naming is just not consistent. > > Also, I plan to move the rest of drivers/of/of_pci* to drivers/pci. > >> and i am agree the naming is not clear, maybe we can rename both of those >> files to something like pci-pm-***.c? > > At least pci-acpi.c is more than just PM functions, so that doesn't > make sense. Given that all the ACPI related functions are in 1 file, > we should do the same for DT. Agreed. Thanks, Rafael
On Wed, 2017-12-27 at 09:30 -0600, Rob Herring wrote: > On Tue, Dec 26, 2017 at 7:32 PM, JeffyChen <jeffy.chen@rock-chips.com> > wrote: > this new file does something similar to the pci-acpi.c and pci-mid.c.. > pci-acpi.c has similar things to pci/of.c. The naming is just not > consistent. > > and i am agree the naming is not clear, maybe we can rename both of > > those > > files to something like pci-pm-***.c? > > At least pci-acpi.c is more than just PM functions, so that doesn't > make sense. Given that all the ACPI related functions are in 1 file, > we should do the same for DT. > > but i have no idea about pci-mid.c or would it possible to have more > > platform pm ops in the future...maybe we should add some dependency > > in the > > Kconfig? You may consider pci-mid.c as pci-sfi.c to some extend. So, in that sense it more looks like pci-of.c would be the name for DT case.
diff --git a/drivers/of/of_pci_irq.c b/drivers/of/of_pci_irq.c index d39565d5477b..abec3a44853b 100644 --- a/drivers/of/of_pci_irq.c +++ b/drivers/of/of_pci_irq.c @@ -1,8 +1,57 @@ #include <linux/kernel.h> #include <linux/of_pci.h> #include <linux/of_irq.h> +#include <linux/pm_wakeirq.h> #include <linux/export.h> +int of_pci_setup_wake_irq(struct pci_dev *pdev) +{ + struct pci_dev *ppdev; + struct device_node *dn; + int ret, irq; + + /* Get the pci_dev of our parent. Hopefully it's a port. */ + ppdev = pdev->bus->self; + /* Nope, it's a host bridge. */ + if (!ppdev) + return 0; + + dn = pci_device_to_OF_node(ppdev); + if (!dn) + return 0; + + irq = of_irq_get_byname(dn, "wakeup"); + if (irq == -EPROBE_DEFER) { + return irq; + } else if (irq < 0) { + /* Ignore other errors, since a missing wakeup is non-fatal. */ + dev_info(&pdev->dev, "cannot get wakeup interrupt: %d\n", irq); + return 0; + } + + device_init_wakeup(&pdev->dev, true); + + ret = dev_pm_set_dedicated_wake_irq(&pdev->dev, irq); + if (ret < 0) { + dev_err(&pdev->dev, "failed to set wake IRQ: %d\n", ret); + device_init_wakeup(&pdev->dev, false); + return ret; + } + + /* Start out disabled to avoid irq storm */ + dev_pm_disable_wake_irq(&pdev->dev); + + return 0; +} +EXPORT_SYMBOL_GPL(of_pci_setup_wake_irq); + +void of_pci_teardown_wake_irq(struct pci_dev *pdev) +{ + dev_pm_clear_wake_irq(&pdev->dev); + device_init_wakeup(&pdev->dev, false); +} +EXPORT_SYMBOL_GPL(of_pci_teardown_wake_irq); + /** * of_irq_parse_pci - Resolve the interrupt for a PCI device * @pdev: the device whose interrupt is to be resolved diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile index c7819b973df7..d0182c82162a 100644 --- a/drivers/pci/Makefile +++ b/drivers/pci/Makefile @@ -29,6 +29,7 @@ obj-$(CONFIG_PCI_IOV) += iov.o # ACPI _DSM provided firmware instance and string name # obj-$(CONFIG_ACPI) += pci-acpi.o +obj-$(CONFIG_OF) += pci-of.o # SMBIOS provided firmware instance and labels obj-$(CONFIG_PCI_LABEL) += pci-label.o diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c index d79dbc377b9c..b4475ff35d97 100644 --- a/drivers/pci/pci-driver.c +++ b/drivers/pci/pci-driver.c @@ -17,6 +17,7 @@ #include <linux/slab.h> #include <linux/sched.h> #include <linux/cpu.h> +#include <linux/of_pci.h> #include <linux/pm_runtime.h> #include <linux/suspend.h> #include <linux/kexec.h> @@ -421,10 +422,17 @@ static int pci_device_probe(struct device *dev) if (error < 0) return error; + error = of_pci_setup_wake_irq(pci_dev); + if (error < 0) { + pcibios_free_irq(pci_dev); + return error; + } + pci_dev_get(pci_dev); if (pci_device_can_probe(pci_dev)) { error = __pci_device_probe(drv, pci_dev); if (error) { + of_pci_teardown_wake_irq(pci_dev); pcibios_free_irq(pci_dev); pci_dev_put(pci_dev); } @@ -438,6 +446,8 @@ static int pci_device_remove(struct device *dev) struct pci_dev *pci_dev = to_pci_dev(dev); struct pci_driver *drv = pci_dev->driver; + of_pci_teardown_wake_irq(pci_dev); + if (drv) { if (drv->remove) { pm_runtime_get_sync(dev); diff --git a/drivers/pci/pci-of.c b/drivers/pci/pci-of.c new file mode 100644 index 000000000000..ad413b2de508 --- /dev/null +++ b/drivers/pci/pci-of.c @@ -0,0 +1,75 @@ +/* + * File: pci-of.c + * Purpose: Provide PCI PM/wakeup support in OF systems + * + * Copyright (C) 2017 Google, Inc. + * Brian Norris <briannorris@chromium.org> + * + * Copyright (C) 2017 Rockchip Electronics Co., Ltd. + * Jeffy Chen <jeffy.chen@rock-chips.com> + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + */ + +#include <linux/acpi.h> +#include <linux/init.h> +#include <linux/module.h> +#include <linux/pci.h> +#include <linux/pm_wakeirq.h> +#include "pci.h" + +static bool of_pci_power_manageable(struct pci_dev *dev) +{ + return false; +} + +static int of_pci_set_power_state(struct pci_dev *dev, pci_power_t state) +{ + return -ENOSYS; +} + +static pci_power_t of_pci_get_power_state(struct pci_dev *dev) +{ + return PCI_UNKNOWN; +} + +static pci_power_t of_pci_choose_state(struct pci_dev *pdev) +{ + return PCI_POWER_ERROR; +} + +static int of_pci_wakeup(struct pci_dev *dev, bool enable) +{ + if (enable) + dev_pm_enable_wake_irq(&dev->dev); + else + dev_pm_disable_wake_irq(&dev->dev); + return 0; +} + +static bool of_pci_need_resume(struct pci_dev *dev) +{ + return false; +} + +static const struct pci_platform_pm_ops of_pci_platform_pm = { + .is_manageable = of_pci_power_manageable, + .set_state = of_pci_set_power_state, + .get_state = of_pci_get_power_state, + .choose_state = of_pci_choose_state, + .set_wakeup = of_pci_wakeup, + .need_resume = of_pci_need_resume, +}; + +static int __init of_pci_init(void) +{ + if (!acpi_disabled) + return 0; + + pci_set_platform_pm(&of_pci_platform_pm); + + return 0; +} +arch_initcall(of_pci_init); diff --git a/include/linux/of_pci.h b/include/linux/of_pci.h index bf588a05d0d0..80fe8ae3393e 100644 --- a/include/linux/of_pci.h +++ b/include/linux/of_pci.h @@ -10,6 +10,8 @@ struct of_phandle_args; struct device_node; #ifdef CONFIG_OF_PCI +int of_pci_setup_wake_irq(struct pci_dev *pdev); +void of_pci_teardown_wake_irq(struct pci_dev *pdev); int of_irq_parse_pci(const struct pci_dev *pdev, struct of_phandle_args *out_irq); struct device_node *of_pci_find_child_device(struct device_node *parent, unsigned int devfn); @@ -23,6 +25,13 @@ int of_pci_map_rid(struct device_node *np, u32 rid, const char *map_name, const char *map_mask_name, struct device_node **target, u32 *id_out); #else +static inline int of_pci_setup_wake_irq(struct pci_dev *pdev) +{ + return 0; +} + +static void of_pci_teardown_wake_irq(struct pci_dev *pdev) { }; + static inline int of_irq_parse_pci(const struct pci_dev *pdev, struct of_phandle_args *out_irq) { return 0;
Add of_pci_setup_wake_irq() and of_pci_teardown_wake_irq() to handle the PCIe WAKE# interrupt. Also use the dedicated wakeirq infrastructure to simplify it. And add pci-of.c to enable/disable the wakeup irq in noirq stage to avoid possible irq storm. Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com> --- Changes in v13: Fix compiler error reported by kbuild test robot <fengguang.wu@intel.com> Changes in v12: Enable the wake irq in noirq stage to avoid possible irq storm. Changes in v11: Only support 1-per-device PCIe WAKE# pin as suggested. Changes in v10: Use device_set_wakeup_capable() instead of device_set_wakeup_enable(), since dedicated wakeirq will be lost in device_set_wakeup_enable(false). Changes in v9: Fix check error in .cleanup(). Move dedicated wakeirq setup to setup() callback and use device_set_wakeup_enable() to enable/disable. Changes in v8: Add pci-of.c and use platform_pm_ops to handle the PCIe WAKE# signal. Changes in v7: Move PCIE_WAKE handling into pci core. Changes in v6: Fix device_init_wake error handling, and add some comments. Changes in v5: Rebase. Changes in v3: Fix error handling. Changes in v2: Use dev_pm_set_dedicated_wake_irq. drivers/of/of_pci_irq.c | 49 +++++++++++++++++++++++++++++++ drivers/pci/Makefile | 1 + drivers/pci/pci-driver.c | 10 +++++++ drivers/pci/pci-of.c | 75 ++++++++++++++++++++++++++++++++++++++++++++++++ include/linux/of_pci.h | 9 ++++++ 5 files changed, 144 insertions(+) create mode 100644 drivers/pci/pci-of.c