| Message ID | 2001450.KHHLaZsp3O@vostro.rjw.lan (mailing list archive) |
|---|---|
| State | New, archived |
| Delegated to: | Bjorn Helgaas |
| Headers | show |
On Mon, Apr 13, 2015, at 11:23, Rafael J. Wysocki wrote: > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > > Refine the mechanism introduced by commit f244d8b623da (ACPIPHP / radeon > / nouveau: Fix VGA switcheroo problem related to hotplug) to propagate > the ignore_hotplug setting of the device to its parent bridge in case > hotplug notifications related to the graphics adapter switching are > given for the bridge rather than for the device itself (the need to > be ignored in both cases). I do apologise if this is a stupid question, but is there any chance the bridge will be connected to other devices that do require hotplug handling, and not just to the GPU?
On Tuesday, April 14, 2015 12:28:12 PM Henrique de Moraes Holschuh wrote: > On Mon, Apr 13, 2015, at 11:23, Rafael J. Wysocki wrote: > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > > > > Refine the mechanism introduced by commit f244d8b623da (ACPIPHP / radeon > > / nouveau: Fix VGA switcheroo problem related to hotplug) to propagate > > the ignore_hotplug setting of the device to its parent bridge in case > > hotplug notifications related to the graphics adapter switching are > > given for the bridge rather than for the device itself (the need to > > be ignored in both cases). > > I do apologise if this is a stupid question, but is there any chance the > bridge will be connected to other devices that do require hotplug handling, > and not just to the GPU? The bridge is actually a downstream PCIe port holding the GPU, so no. :-) Rafael -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Apr 14, 2015 at 8:03 PM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote: > On Tuesday, April 14, 2015 12:28:12 PM Henrique de Moraes Holschuh wrote: >> On Mon, Apr 13, 2015, at 11:23, Rafael J. Wysocki wrote: >> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com> >> > >> > Refine the mechanism introduced by commit f244d8b623da (ACPIPHP / radeon >> > / nouveau: Fix VGA switcheroo problem related to hotplug) to propagate >> > the ignore_hotplug setting of the device to its parent bridge in case >> > hotplug notifications related to the graphics adapter switching are >> > given for the bridge rather than for the device itself (the need to >> > be ignored in both cases). >> >> I do apologise if this is a stupid question, but is there any chance the >> bridge will be connected to other devices that do require hotplug handling, >> and not just to the GPU? > > The bridge is actually a downstream PCIe port holding the GPU, so no. :-) When radeon/nouveau call pci_ignore_hotplug(), that's the case, but in general all we know is that pci_ignore_hotplug() receives a PCI device. We don't know whether it's PCI or PCIe. In the hotplug topologies I'm familiar with, a bridge only leads to one hot-pluggable slot, but I don't remember anything that would guarantee that. For PCIe, I think there can only be one slot, but for PCI I would think it possible to have one bridge leading to several hotpluggable slots, with the hotplug controller(s) being separate from the bridge. Bjorn -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Apr 15, 2015 at 2:55 PM, Bjorn Helgaas <bhelgaas@google.com> wrote: > On Tue, Apr 14, 2015 at 8:03 PM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote: >> On Tuesday, April 14, 2015 12:28:12 PM Henrique de Moraes Holschuh wrote: >>> On Mon, Apr 13, 2015, at 11:23, Rafael J. Wysocki wrote: >>> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com> >>> > >>> > Refine the mechanism introduced by commit f244d8b623da (ACPIPHP / radeon >>> > / nouveau: Fix VGA switcheroo problem related to hotplug) to propagate >>> > the ignore_hotplug setting of the device to its parent bridge in case >>> > hotplug notifications related to the graphics adapter switching are >>> > given for the bridge rather than for the device itself (the need to >>> > be ignored in both cases). >>> >>> I do apologise if this is a stupid question, but is there any chance the >>> bridge will be connected to other devices that do require hotplug handling, >>> and not just to the GPU? >> >> The bridge is actually a downstream PCIe port holding the GPU, so no. :-) > > When radeon/nouveau call pci_ignore_hotplug(), that's the case, but in > general all we know is that pci_ignore_hotplug() receives a PCI > device. We don't know whether it's PCI or PCIe. In the hotplug > topologies I'm familiar with, a bridge only leads to one hot-pluggable > slot, but I don't remember anything that would guarantee that. For > PCIe, I think there can only be one slot, but for PCI I would think it > possible to have one bridge leading to several hotpluggable slots, > with the hotplug controller(s) being separate from the bridge. Realistically, the switcheroo people are the only users of pci_ignore_hotplug() today and if somebody wants the hotplug events to be ignored for him and perhaps not for someone else on the same bridge, then something is seriously broken about that system anyway. Still, it is up to you whether or not to apply the patch. :-) Rafael -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Apr 13, 2015 at 04:23:36PM +0200, Rafael J. Wysocki wrote: > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > > Refine the mechanism introduced by commit f244d8b623da (ACPIPHP / radeon > / nouveau: Fix VGA switcheroo problem related to hotplug) to propagate > the ignore_hotplug setting of the device to its parent bridge in case > hotplug notifications related to the graphics adapter switching are > given for the bridge rather than for the device itself (the need to > be ignored in both cases). > > Link: https://bugzilla.kernel.org/show_bug.cgi?id=61891 > Link: https://bugs.freedesktop.org/show_bug.cgi?id=88927 > Reported-and-tested-by: tiagdtd-lava <tiagdtd-lava@yahoo.de> > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Applied to pci/hotplug for v4.2, thanks! I marked it for stable and added a note to connect it to b440bde74f04 ("PCI: Add pci_ignore_hotplug() to ignore hotplug events for a device") > --- > > "Stable" material AFAICS. > > --- > drivers/pci/pci.c | 11 +++++++++++ > include/linux/pci.h | 6 +----- > 2 files changed, 12 insertions(+), 5 deletions(-) > > Index: linux-pm/drivers/pci/pci.c > =================================================================== > --- linux-pm.orig/drivers/pci/pci.c > +++ linux-pm/drivers/pci/pci.c > @@ -4319,6 +4319,17 @@ bool pci_device_is_present(struct pci_de > } > EXPORT_SYMBOL_GPL(pci_device_is_present); > > +void pci_ignore_hotplug(struct pci_dev *dev) > +{ > + struct pci_dev *bridge = dev->bus->self; > + > + dev->ignore_hotplug = 1; > + /* Propagate the "ignore hotplug" setting to the parent bridge. */ > + if (bridge) > + bridge->ignore_hotplug = 1; > +} > +EXPORT_SYMBOL_GPL(pci_ignore_hotplug); > + > #define RESOURCE_ALIGNMENT_PARAM_SIZE COMMAND_LINE_SIZE > static char resource_alignment_param[RESOURCE_ALIGNMENT_PARAM_SIZE] = {0}; > static DEFINE_SPINLOCK(resource_alignment_lock); > Index: linux-pm/include/linux/pci.h > =================================================================== > --- linux-pm.orig/include/linux/pci.h > +++ linux-pm/include/linux/pci.h > @@ -1002,6 +1002,7 @@ int __must_check pci_assign_resource(str > int __must_check pci_reassign_resource(struct pci_dev *dev, int i, resource_size_t add_size, resource_size_t align); > int pci_select_bars(struct pci_dev *dev, unsigned long flags); > bool pci_device_is_present(struct pci_dev *pdev); > +void pci_ignore_hotplug(struct pci_dev *dev); > > /* ROM control related routines */ > int pci_enable_rom(struct pci_dev *pdev); > @@ -1039,11 +1040,6 @@ bool pci_dev_run_wake(struct pci_dev *de > bool pci_check_pme_status(struct pci_dev *dev); > void pci_pme_wakeup_bus(struct pci_bus *bus); > > -static inline void pci_ignore_hotplug(struct pci_dev *dev) > -{ > - dev->ignore_hotplug = 1; > -} > - > static inline int pci_enable_wake(struct pci_dev *dev, pci_power_t state, > bool enable) > { > -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Friday, May 22, 2015 08:40:37 PM Bjorn Helgaas wrote: > On Mon, Apr 13, 2015 at 04:23:36PM +0200, Rafael J. Wysocki wrote: > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > > > > Refine the mechanism introduced by commit f244d8b623da (ACPIPHP / radeon > > / nouveau: Fix VGA switcheroo problem related to hotplug) to propagate > > the ignore_hotplug setting of the device to its parent bridge in case > > hotplug notifications related to the graphics adapter switching are > > given for the bridge rather than for the device itself (the need to > > be ignored in both cases). > > > > Link: https://bugzilla.kernel.org/show_bug.cgi?id=61891 > > Link: https://bugs.freedesktop.org/show_bug.cgi?id=88927 > > Reported-and-tested-by: tiagdtd-lava <tiagdtd-lava@yahoo.de> > > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> > > Applied to pci/hotplug for v4.2, thanks! > > I marked it for stable and added a note to connect it to b440bde74f04 > ("PCI: Add pci_ignore_hotplug() to ignore hotplug events for a device") Thanks a lot! -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Index: linux-pm/drivers/pci/pci.c =================================================================== --- linux-pm.orig/drivers/pci/pci.c +++ linux-pm/drivers/pci/pci.c @@ -4319,6 +4319,17 @@ bool pci_device_is_present(struct pci_de } EXPORT_SYMBOL_GPL(pci_device_is_present); +void pci_ignore_hotplug(struct pci_dev *dev) +{ + struct pci_dev *bridge = dev->bus->self; + + dev->ignore_hotplug = 1; + /* Propagate the "ignore hotplug" setting to the parent bridge. */ + if (bridge) + bridge->ignore_hotplug = 1; +} +EXPORT_SYMBOL_GPL(pci_ignore_hotplug); + #define RESOURCE_ALIGNMENT_PARAM_SIZE COMMAND_LINE_SIZE static char resource_alignment_param[RESOURCE_ALIGNMENT_PARAM_SIZE] = {0}; static DEFINE_SPINLOCK(resource_alignment_lock); Index: linux-pm/include/linux/pci.h =================================================================== --- linux-pm.orig/include/linux/pci.h +++ linux-pm/include/linux/pci.h @@ -1002,6 +1002,7 @@ int __must_check pci_assign_resource(str int __must_check pci_reassign_resource(struct pci_dev *dev, int i, resource_size_t add_size, resource_size_t align); int pci_select_bars(struct pci_dev *dev, unsigned long flags); bool pci_device_is_present(struct pci_dev *pdev); +void pci_ignore_hotplug(struct pci_dev *dev); /* ROM control related routines */ int pci_enable_rom(struct pci_dev *pdev); @@ -1039,11 +1040,6 @@ bool pci_dev_run_wake(struct pci_dev *de bool pci_check_pme_status(struct pci_dev *dev); void pci_pme_wakeup_bus(struct pci_bus *bus); -static inline void pci_ignore_hotplug(struct pci_dev *dev) -{ - dev->ignore_hotplug = 1; -} - static inline int pci_enable_wake(struct pci_dev *dev, pci_power_t state, bool enable) {