From: "Sricharan" <sricharan-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org> To: 'Robin Murphy' <robin.murphy-5wv7dgnIgG8@public.gmane.org>, will.deacon-5wv7dgnIgG8@public.gmane.org, joro-zLv9SwRftAIdnm+yROfE0A@public.gmane.org, lorenzo.pieralisi-5wv7dgnIgG8@public.gmane.org, iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, linux-arm-msm-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, m.szyprowski-Sze3O3UU22JBDgjK7y7TUQ@public.gmane.org Subject: RE: [PATCH V5 05/12] drivers: platform: Configure dma operations at probe time Date: Fri, 20 Jan 2017 11:48:18 +0530 [thread overview] Message-ID: <000c01d272e5$020ba300$0622e900$@codeaurora.org> (raw) In-Reply-To: <327b5a9b-35f6-f6e0-4ef9-d2ed15bb4e49-5wv7dgnIgG8@public.gmane.org> Hi Robin, >On 19/01/17 15:05, Sricharan R wrote: >> Configuring DMA ops at probe time will allow deferring device probe when >> the IOMMU isn't available yet. The dma_configure for the device is >> now called from the generic device_attach callback just before the >> bus/driver probe is called. This way, configuring the DMA ops for the >> device would be called at the same place for all bus_types, hence the >> deferred probing mechanism should work for all buses as well. >> >> pci_bus_add_devices (platform/amba)(_device_create/driver_register) >> | | >> pci_bus_add_device (device_add/driver_register) >> | | >> device_attach device_initial_probe >> | | >> __device_attach_driver __device_attach_driver >> | >> driver_probe_device >> | >> really_probe >> | >> dma_configure >> >> Similarly on the device/driver_unregister path __device_release_driver is >> called which inturn calls dma_deconfigure. >> >> Signed-off-by: Sricharan R <sricharan@codeaurora.org> >> --- >> * Removed the dma configuration for the pci devices in case of DT >> from pci_dma_configure which was hanging outside separately and >> doing it in dma_configure function itself. >> >> drivers/base/dd.c | 9 +++++++++ >> drivers/base/dma-mapping.c | 32 ++++++++++++++++++++++++++++++++ >> drivers/of/platform.c | 5 +---- >> drivers/pci/probe.c | 5 +---- >> include/linux/dma-mapping.h | 3 +++ >> 5 files changed, 46 insertions(+), 8 deletions(-) >> >> diff --git a/drivers/base/dd.c b/drivers/base/dd.c >> index a1fbf55..4882f06 100644 >> --- a/drivers/base/dd.c >> +++ b/drivers/base/dd.c >> @@ -19,6 +19,7 @@ >> >> #include <linux/device.h> >> #include <linux/delay.h> >> +#include <linux/dma-mapping.h> >> #include <linux/module.h> >> #include <linux/kthread.h> >> #include <linux/wait.h> >> @@ -356,6 +357,10 @@ static int really_probe(struct device *dev, struct device_driver *drv) >> if (ret) >> goto pinctrl_bind_failed; >> >> + ret = dma_configure(dev); >> + if (ret) >> + goto dma_failed; >> + >> if (driver_sysfs_add(dev)) { >> printk(KERN_ERR "%s: driver_sysfs_add(%s) failed\n", >> __func__, dev_name(dev)); >> @@ -417,6 +422,8 @@ static int really_probe(struct device *dev, struct device_driver *drv) >> goto done; >> >> probe_failed: >> + dma_deconfigure(dev); >> +dma_failed: >> if (dev->bus) >> blocking_notifier_call_chain(&dev->bus->p->bus_notifier, >> BUS_NOTIFY_DRIVER_NOT_BOUND, dev); >> @@ -826,6 +833,8 @@ static void __device_release_driver(struct device *dev, struct device *parent) >> drv->remove(dev); >> >> device_links_driver_cleanup(dev); >> + dma_deconfigure(dev); >> + >> devres_release_all(dev); >> dev->driver = NULL; >> dev_set_drvdata(dev, NULL); >> diff --git a/drivers/base/dma-mapping.c b/drivers/base/dma-mapping.c >> index efd71cf..dfe6fd7 100644 >> --- a/drivers/base/dma-mapping.c >> +++ b/drivers/base/dma-mapping.c >> @@ -10,6 +10,7 @@ >> #include <linux/dma-mapping.h> >> #include <linux/export.h> >> #include <linux/gfp.h> >> +#include <linux/of_device.h> >> #include <linux/slab.h> >> #include <linux/vmalloc.h> >> >> @@ -341,3 +342,34 @@ void dma_common_free_remap(void *cpu_addr, size_t size, unsigned long vm_flags) >> vunmap(cpu_addr); >> } >> #endif >> + >> +/* >> + * Common configuration to enable DMA API use for a device >> + */ >> +#include <linux/pci.h> >> + >> +int dma_configure(struct device *dev) >> +{ >> + struct device *_dev = dev; >> + int is_pci = dev_is_pci(dev); >> + >> + if (is_pci) { >> + _dev = pci_get_host_bridge_device(to_pci_dev(dev)); >> + if (IS_ENABLED(CONFIG_OF) && _dev->parent && >> + _dev->parent->of_node) >> + _dev = _dev->parent; >> + } >> + >> + if (_dev->of_node) >> + of_dma_configure(dev, _dev->of_node); >> + >> + if (is_pci) >> + pci_put_host_bridge_device(_dev); > >There's a fun bug here - at this point _dev is the *parent* of the >bridge device, so we put the refcount on the wrong device (the platform >device representing the host controller, rather than the PCI device >representing its insides), which frees the guy we're in the middle of >probing, and things rapidly go wrong afterwards: > ha my bad, right. I will change to fix this. I was testing this with a couple of pci devices, but it did not show up. Thanks for pointing this out. I will borrow the variable names from the hunk below. >[ 1.461026] bus: 'platform': driver_probe_device: matched device >40000000.pcie-controller with driver pci-host-generic >[ 1.471640] bus: 'platform': really_probe: probing driver >pci-host-generic with device 40000000.pcie-controller >[ 1.481678] OF: PCI: host bridge /pcie-controller@40000000 ranges: > >... > >[ 2.158259] bus: 'pci': driver_probe_device: matched device >0000:02:10.0 with driver pcieport >[ 2.166716] bus: 'pci': really_probe: probing driver pcieport with >device 0000:02:10.0 >[ 2.174590] pci 0000:02:10.0: Driver pcieport requests probe deferral >[ 2.180978] pci 0000:02:10.0: Added to deferred list >[ 2.185915] bus: 'pci': driver_probe_device: matched device >0000:02:1f.0 with driver pcieport >[ 2.194366] bus: 'pci': really_probe: probing driver pcieport with >device 0000:02:1f.0 >[ 2.202237] pci 0000:02:1f.0: Driver pcieport requests probe deferral >[ 2.208625] pci 0000:02:1f.0: Added to deferred list >[ 2.213582] driver: 'pci-host-generic': driver_bound: bound to device >'���v����.pcie-controller' >[ 2.222293] bus: 'platform': really_probe: bound device >���v����.pcie-controller to driver pci-host-generic >[ 2.232041] Unable to handle kernel NULL pointer dereference at >virtual address 00000000 > >I recall debugging this same issue before, and I seem to have a local >version of this commit dated about 6 weeks ago where dma_configure() >looks like this: > >--->8--- >int dma_configure(struct device *dev) >{ > struct device *bridge = NULL, *dma_dev = dev; > > if (dev_is_pci(dev)) { > bridge = pci_get_host_bridge_device(to_pci_dev(dev)); > dma_dev = bridge->parent; > } > > if (dma_dev && dma_dev->of_node) { > of_dma_configure(dev, dma_dev->of_node); > } else if (dma_dev && has_acpi_companion(dma_dev)) { > struct acpi_device *adev = to_acpi_device_node(dma_dev->fwnode); > enum dev_dma_attr attr = acpi_get_dma_attr(adev); > > if (attr == DEV_DMA_NOT_SUPPORTED) > dev_warn(dev, "DMA not supported.\n"); > else > arch_setup_dma_ops(dev, 0, 0, NULL, > attr == DEV_DMA_COHERENT); > } > > if (bridge) > pci_put_host_bridge_device(bridge); > > return 0; >} >---8<--- > >I have a feeling I was having a go at tidying up the "PCI hacks" from v4 >myself, but other things ended up taking precedence. > Sure, i will repost with borrowing the names. Regards, Sricharan _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu
WARNING: multiple messages have this Message-ID (diff)
From: sricharan@codeaurora.org (Sricharan) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH V5 05/12] drivers: platform: Configure dma operations at probe time Date: Fri, 20 Jan 2017 11:48:18 +0530 [thread overview] Message-ID: <000c01d272e5$020ba300$0622e900$@codeaurora.org> (raw) In-Reply-To: <327b5a9b-35f6-f6e0-4ef9-d2ed15bb4e49@arm.com> Hi Robin, >On 19/01/17 15:05, Sricharan R wrote: >> Configuring DMA ops at probe time will allow deferring device probe when >> the IOMMU isn't available yet. The dma_configure for the device is >> now called from the generic device_attach callback just before the >> bus/driver probe is called. This way, configuring the DMA ops for the >> device would be called at the same place for all bus_types, hence the >> deferred probing mechanism should work for all buses as well. >> >> pci_bus_add_devices (platform/amba)(_device_create/driver_register) >> | | >> pci_bus_add_device (device_add/driver_register) >> | | >> device_attach device_initial_probe >> | | >> __device_attach_driver __device_attach_driver >> | >> driver_probe_device >> | >> really_probe >> | >> dma_configure >> >> Similarly on the device/driver_unregister path __device_release_driver is >> called which inturn calls dma_deconfigure. >> >> Signed-off-by: Sricharan R <sricharan@codeaurora.org> >> --- >> * Removed the dma configuration for the pci devices in case of DT >> from pci_dma_configure which was hanging outside separately and >> doing it in dma_configure function itself. >> >> drivers/base/dd.c | 9 +++++++++ >> drivers/base/dma-mapping.c | 32 ++++++++++++++++++++++++++++++++ >> drivers/of/platform.c | 5 +---- >> drivers/pci/probe.c | 5 +---- >> include/linux/dma-mapping.h | 3 +++ >> 5 files changed, 46 insertions(+), 8 deletions(-) >> >> diff --git a/drivers/base/dd.c b/drivers/base/dd.c >> index a1fbf55..4882f06 100644 >> --- a/drivers/base/dd.c >> +++ b/drivers/base/dd.c >> @@ -19,6 +19,7 @@ >> >> #include <linux/device.h> >> #include <linux/delay.h> >> +#include <linux/dma-mapping.h> >> #include <linux/module.h> >> #include <linux/kthread.h> >> #include <linux/wait.h> >> @@ -356,6 +357,10 @@ static int really_probe(struct device *dev, struct device_driver *drv) >> if (ret) >> goto pinctrl_bind_failed; >> >> + ret = dma_configure(dev); >> + if (ret) >> + goto dma_failed; >> + >> if (driver_sysfs_add(dev)) { >> printk(KERN_ERR "%s: driver_sysfs_add(%s) failed\n", >> __func__, dev_name(dev)); >> @@ -417,6 +422,8 @@ static int really_probe(struct device *dev, struct device_driver *drv) >> goto done; >> >> probe_failed: >> + dma_deconfigure(dev); >> +dma_failed: >> if (dev->bus) >> blocking_notifier_call_chain(&dev->bus->p->bus_notifier, >> BUS_NOTIFY_DRIVER_NOT_BOUND, dev); >> @@ -826,6 +833,8 @@ static void __device_release_driver(struct device *dev, struct device *parent) >> drv->remove(dev); >> >> device_links_driver_cleanup(dev); >> + dma_deconfigure(dev); >> + >> devres_release_all(dev); >> dev->driver = NULL; >> dev_set_drvdata(dev, NULL); >> diff --git a/drivers/base/dma-mapping.c b/drivers/base/dma-mapping.c >> index efd71cf..dfe6fd7 100644 >> --- a/drivers/base/dma-mapping.c >> +++ b/drivers/base/dma-mapping.c >> @@ -10,6 +10,7 @@ >> #include <linux/dma-mapping.h> >> #include <linux/export.h> >> #include <linux/gfp.h> >> +#include <linux/of_device.h> >> #include <linux/slab.h> >> #include <linux/vmalloc.h> >> >> @@ -341,3 +342,34 @@ void dma_common_free_remap(void *cpu_addr, size_t size, unsigned long vm_flags) >> vunmap(cpu_addr); >> } >> #endif >> + >> +/* >> + * Common configuration to enable DMA API use for a device >> + */ >> +#include <linux/pci.h> >> + >> +int dma_configure(struct device *dev) >> +{ >> + struct device *_dev = dev; >> + int is_pci = dev_is_pci(dev); >> + >> + if (is_pci) { >> + _dev = pci_get_host_bridge_device(to_pci_dev(dev)); >> + if (IS_ENABLED(CONFIG_OF) && _dev->parent && >> + _dev->parent->of_node) >> + _dev = _dev->parent; >> + } >> + >> + if (_dev->of_node) >> + of_dma_configure(dev, _dev->of_node); >> + >> + if (is_pci) >> + pci_put_host_bridge_device(_dev); > >There's a fun bug here - at this point _dev is the *parent* of the >bridge device, so we put the refcount on the wrong device (the platform >device representing the host controller, rather than the PCI device >representing its insides), which frees the guy we're in the middle of >probing, and things rapidly go wrong afterwards: > ha my bad, right. I will change to fix this. I was testing this with a couple of pci devices, but it did not show up. Thanks for pointing this out. I will borrow the variable names from the hunk below. >[ 1.461026] bus: 'platform': driver_probe_device: matched device >40000000.pcie-controller with driver pci-host-generic >[ 1.471640] bus: 'platform': really_probe: probing driver >pci-host-generic with device 40000000.pcie-controller >[ 1.481678] OF: PCI: host bridge /pcie-controller at 40000000 ranges: > >... > >[ 2.158259] bus: 'pci': driver_probe_device: matched device >0000:02:10.0 with driver pcieport >[ 2.166716] bus: 'pci': really_probe: probing driver pcieport with >device 0000:02:10.0 >[ 2.174590] pci 0000:02:10.0: Driver pcieport requests probe deferral >[ 2.180978] pci 0000:02:10.0: Added to deferred list >[ 2.185915] bus: 'pci': driver_probe_device: matched device >0000:02:1f.0 with driver pcieport >[ 2.194366] bus: 'pci': really_probe: probing driver pcieport with >device 0000:02:1f.0 >[ 2.202237] pci 0000:02:1f.0: Driver pcieport requests probe deferral >[ 2.208625] pci 0000:02:1f.0: Added to deferred list >[ 2.213582] driver: 'pci-host-generic': driver_bound: bound to device >'???v????.pcie-controller' >[ 2.222293] bus: 'platform': really_probe: bound device >???v????.pcie-controller to driver pci-host-generic >[ 2.232041] Unable to handle kernel NULL pointer dereference at >virtual address 00000000 > >I recall debugging this same issue before, and I seem to have a local >version of this commit dated about 6 weeks ago where dma_configure() >looks like this: > >--->8--- >int dma_configure(struct device *dev) >{ > struct device *bridge = NULL, *dma_dev = dev; > > if (dev_is_pci(dev)) { > bridge = pci_get_host_bridge_device(to_pci_dev(dev)); > dma_dev = bridge->parent; > } > > if (dma_dev && dma_dev->of_node) { > of_dma_configure(dev, dma_dev->of_node); > } else if (dma_dev && has_acpi_companion(dma_dev)) { > struct acpi_device *adev = to_acpi_device_node(dma_dev->fwnode); > enum dev_dma_attr attr = acpi_get_dma_attr(adev); > > if (attr == DEV_DMA_NOT_SUPPORTED) > dev_warn(dev, "DMA not supported.\n"); > else > arch_setup_dma_ops(dev, 0, 0, NULL, > attr == DEV_DMA_COHERENT); > } > > if (bridge) > pci_put_host_bridge_device(bridge); > > return 0; >} >---8<--- > >I have a feeling I was having a go at tidying up the "PCI hacks" from v4 >myself, but other things ended up taking precedence. > Sure, i will repost with borrowing the names. Regards, Sricharan
next prev parent reply other threads:[~2017-01-20 6:18 UTC|newest] Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top 2017-01-19 15:05 [PATCH V5 00/12] IOMMU probe deferral support Sricharan R 2017-01-19 15:05 ` Sricharan R [not found] ` <1484838356-24962-1-git-send-email-sricharan-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org> 2017-01-19 15:05 ` [PATCH V5 01/12] iommu/of: Refactor of_iommu_configure() for error handling Sricharan R 2017-01-19 15:05 ` Sricharan R 2017-01-19 15:05 ` [PATCH V5 02/12] iommu/of: Prepare for deferred IOMMU configuration Sricharan R 2017-01-19 15:05 ` Sricharan R 2017-01-19 15:05 ` [PATCH V5 03/12] of: dma: Move range size workaround to of_dma_get_range() Sricharan R 2017-01-19 15:05 ` Sricharan R 2017-01-19 15:05 ` [PATCH V5 04/12] of: dma: Make of_dma_deconfigure() public Sricharan R 2017-01-19 15:05 ` Sricharan R 2017-01-19 15:05 ` [PATCH V5 05/12] drivers: platform: Configure dma operations at probe time Sricharan R 2017-01-19 15:05 ` Sricharan R [not found] ` <1484838356-24962-6-git-send-email-sricharan-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org> 2017-01-19 16:48 ` Lorenzo Pieralisi 2017-01-19 16:48 ` Lorenzo Pieralisi 2017-01-20 5:34 ` Sricharan 2017-01-20 5:34 ` Sricharan 2017-01-19 17:49 ` Robin Murphy 2017-01-19 17:49 ` Robin Murphy [not found] ` <327b5a9b-35f6-f6e0-4ef9-d2ed15bb4e49-5wv7dgnIgG8@public.gmane.org> 2017-01-19 19:02 ` Lorenzo Pieralisi 2017-01-19 19:02 ` Lorenzo Pieralisi 2017-01-20 6:25 ` Sricharan 2017-01-20 6:25 ` Sricharan 2017-01-20 6:18 ` Sricharan [this message] 2017-01-20 6:18 ` Sricharan 2017-01-19 15:05 ` [PATCH V5 06/12] iommu: of: Handle IOMMU lookup failure with deferred probing or error Sricharan R 2017-01-19 15:05 ` Sricharan R 2017-01-19 15:05 ` [PATCH V5 07/12] arm64: dma-mapping: Remove the notifier trick to handle early setting of dma_ops Sricharan R 2017-01-19 15:05 ` Sricharan R [not found] ` <1484838356-24962-8-git-send-email-sricharan-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org> 2017-01-19 16:20 ` Will Deacon 2017-01-19 16:20 ` Will Deacon 2017-01-19 15:05 ` [PATCH V5 08/12] iommu/arm-smmu: Clean up early-probing workarounds Sricharan R 2017-01-19 15:05 ` Sricharan R [not found] ` <1484838356-24962-9-git-send-email-sricharan-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org> 2017-01-19 16:18 ` Will Deacon 2017-01-19 16:18 ` Will Deacon 2017-01-20 19:20 ` Sricharan 2017-01-20 19:20 ` Sricharan 2017-01-19 16:50 ` Lorenzo Pieralisi 2017-01-19 16:50 ` Lorenzo Pieralisi 2017-01-19 17:58 ` Robin Murphy 2017-01-19 17:58 ` Robin Murphy [not found] ` <b3f0f112-e4a0-fa4d-4f18-44484392a5e9-5wv7dgnIgG8@public.gmane.org> 2017-01-20 6:32 ` Sricharan 2017-01-20 6:32 ` Sricharan 2017-01-19 15:05 ` [PATCH V5 09/12] ACPI/IORT: Add function to check SMMUs drivers presence Sricharan R 2017-01-19 15:05 ` Sricharan R 2017-01-19 15:05 ` [PATCH V5 10/12] drivers: acpi: Configure acpi devices dma operation at probe time Sricharan R 2017-01-19 15:05 ` Sricharan R [not found] ` <1484838356-24962-11-git-send-email-sricharan-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org> 2017-01-19 16:42 ` Lorenzo Pieralisi 2017-01-19 16:42 ` Lorenzo Pieralisi 2017-01-20 5:31 ` Sricharan 2017-01-20 5:31 ` Sricharan 2017-01-19 15:05 ` [PATCH V5 11/12] drivers: acpi: Handle IOMMU lookup failure with deferred probing or error Sricharan R 2017-01-19 15:05 ` Sricharan R 2017-01-19 15:05 ` [PATCH V5 12/12] ACPI/IORT: Remove linker section for IORT entries probing Sricharan R 2017-01-19 15:05 ` Sricharan R
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to='000c01d272e5$020ba300$0622e900$@codeaurora.org' \ --to=sricharan-sgv2jx0feol9jmxxk+q4oq@public.gmane.org \ --cc=iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org \ --cc=joro-zLv9SwRftAIdnm+yROfE0A@public.gmane.org \ --cc=linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org \ --cc=linux-arm-msm-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \ --cc=lorenzo.pieralisi-5wv7dgnIgG8@public.gmane.org \ --cc=m.szyprowski-Sze3O3UU22JBDgjK7y7TUQ@public.gmane.org \ --cc=robin.murphy-5wv7dgnIgG8@public.gmane.org \ --cc=will.deacon-5wv7dgnIgG8@public.gmane.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.