Message ID | 1675188609-20913-1-git-send-email-ssengar@linux.microsoft.com |
---|---|
Headers | show |
Series | Device tree support for Hyper-V VMBus driver | expand |
On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar <ssengar@linux.microsoft.com> wrote: > > Update the driver to support device tree boot as well along with ACPI. > At present the device tree parsing only provides the mmio region info > and is not the exact copy of ACPI parsing. This is sufficient to cater > all the current device tree usecases for VMBus. > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > --- > drivers/hv/vmbus_drv.c | 75 ++++++++++++++++++++++++++++++++++++++++-- > 1 file changed, 73 insertions(+), 2 deletions(-) > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > index 49030e756b9f..1741f1348f9f 100644 > --- a/drivers/hv/vmbus_drv.c > +++ b/drivers/hv/vmbus_drv.c > @@ -2152,7 +2152,7 @@ void vmbus_device_unregister(struct hv_device *device_obj) > device_unregister(&device_obj->device); > } > > - > +#ifdef CONFIG_ACPI > /* > * VMBUS is an acpi enumerated device. Get the information we > * need from DSDT. > @@ -2262,6 +2262,7 @@ static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx) > > return AE_OK; > } > +#endif > > static void vmbus_mmio_remove(void) > { > @@ -2282,7 +2283,7 @@ static void vmbus_mmio_remove(void) > } > } > > -static void vmbus_reserve_fb(void) > +static void __maybe_unused vmbus_reserve_fb(void) > { > resource_size_t start = 0, size; > struct pci_dev *pdev; > @@ -2442,6 +2443,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size) > } > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > +#ifdef CONFIG_ACPI It's better to put C 'if (!IS_ENABLED(CONFIG_ACPI)' code in the > static int vmbus_acpi_add(struct platform_device *pdev) > { > acpi_status result; > @@ -2496,10 +2498,68 @@ static int vmbus_acpi_add(struct platform_device *pdev) > vmbus_mmio_remove(); > return ret_val; > } > +#else > + > +static int vmbus_device_add(struct platform_device *pdev) > +{ > + struct resource **cur_res = &hyperv_mmio; > + struct device_node *np; > + u32 *ranges, len; > + u64 start; > + int nr_ranges, child_cells = 2, cur_cell = 0, ret = 0; > + > + hv_dev = pdev; > + np = pdev->dev.of_node; > + > + nr_ranges = device_property_count_u32(&pdev->dev, "ranges"); Parsing ranges yourself is a bad sign. It's a standard property and we have functions which handle it. If those don't work, then something is wrong with your DT or they need to be fixed/expanded. > + if (nr_ranges < 0) > + return nr_ranges; > + ranges = kcalloc(nr_ranges, sizeof(u32), GFP_KERNEL); > + if (!ranges) > + return -ENOMEM; > + > + if (device_property_read_u32_array(&pdev->dev, "ranges", ranges, nr_ranges)) { > + ret = -EINVAL; > + goto free_ranges; > + } > + > + while (cur_cell < nr_ranges) { > + struct resource *res; > + > + /* The first u64 in the ranges description isn't used currently. */ > + cur_cell = cur_cell + child_cells; > + start = ranges[cur_cell++]; > + start = (start << 32) | ranges[cur_cell++]; > + len = ranges[cur_cell++]; To expand my last point, the format of ranges is <child_addr parent_addr length>. That's not what your 'ranges' has. You've also just ignored '#address-cells' and '#size-cells'. > + > + res = kzalloc(sizeof(*res), GFP_ATOMIC); > + if (!res) { > + ret = -ENOMEM; > + goto free_ranges; > + } > + > + res->name = "hyperv mmio"; > + res->flags = IORESOURCE_MEM | IORESOURCE_MEM_64; > + res->start = start; > + res->end = start + len; > + > + *cur_res = res; > + cur_res = &res->sibling; > + } > + > +free_ranges: > + kfree(ranges); > + return ret; > +} > +#endif > > static int vmbus_platform_driver_probe(struct platform_device *pdev) > { > +#ifdef CONFIG_ACPI > return vmbus_acpi_add(pdev); > +#else > + return vmbus_device_add(pdev); > +#endif > } > > static int vmbus_platform_driver_remove(struct platform_device *pdev) > @@ -2645,6 +2705,16 @@ static int vmbus_bus_resume(struct device *dev) > #define vmbus_bus_resume NULL > #endif /* CONFIG_PM_SLEEP */ > > +static const struct of_device_id vmbus_of_match[] = { > + { > + .compatible = "msft,vmbus", > + }, > + { > + /* sentinel */ > + }, > +}; > +MODULE_DEVICE_TABLE(of, vmbus_of_match); > + > static const struct acpi_device_id vmbus_acpi_device_ids[] = { > {"VMBUS", 0}, > {"VMBus", 0}, > @@ -2679,6 +2749,7 @@ static struct platform_driver vmbus_platform_driver = { > .driver = { > .name = "vmbus", > .acpi_match_table = ACPI_PTR(vmbus_acpi_device_ids), > + .of_match_table = of_match_ptr(vmbus_of_match), > .pm = &vmbus_bus_pm, > .probe_type = PROBE_FORCE_SYNCHRONOUS, > } > -- > 2.25.1 >
On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar <ssengar@linux.microsoft.com> wrote: > > This set of patches expands the VMBus driver to include device tree > support. > > The first two patches enable compilation of Hyper-V APIs in a non-ACPI > build. > > The third patch converts the VMBus driver from acpi to more generic > platform driver. > > Further to add device tree documentation for VMBus, it needs to club with > other virtualization driver's documentation. For this rename the virtio > folder to more generic hypervisor, so that all the hypervisor based > devices can co-exist in a single place in device tree documentation. The > fourth patch does this renaming. > > The fifth patch introduces the device tree documentation for VMBus. > > The sixth patch adds device tree support to the VMBus driver. Currently > this is tested only for x86 and it may not work for other archs. I can read all the patches and see *what* they do. You don't really need to list that here. I'm still wondering *why*. That is what the cover letter and commit messages should answer. Why do you need DT support? How does this even work on x86? FDT is only enabled for CE4100 platform. Rob
On Tue, Jan 31, 2023 at 02:27:51PM -0600, Rob Herring wrote: > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > <ssengar@linux.microsoft.com> wrote: > > > > This set of patches expands the VMBus driver to include device tree > > support. > > > > The first two patches enable compilation of Hyper-V APIs in a non-ACPI > > build. > > > > The third patch converts the VMBus driver from acpi to more generic > > platform driver. > > > > Further to add device tree documentation for VMBus, it needs to club with > > other virtualization driver's documentation. For this rename the virtio > > folder to more generic hypervisor, so that all the hypervisor based > > devices can co-exist in a single place in device tree documentation. The > > fourth patch does this renaming. > > > > The fifth patch introduces the device tree documentation for VMBus. > > > > The sixth patch adds device tree support to the VMBus driver. Currently > > this is tested only for x86 and it may not work for other archs. > > I can read all the patches and see *what* they do. You don't really > need to list that here. I'm still wondering *why*. That is what the > cover letter and commit messages should answer. Why do you need DT > support? How does this even work on x86? FDT is only enabled for > CE4100 platform. HI Rob, Thanks for your comments. We are working on a solution where kernel is booted without ACPI tables to keep the overall system's memory footprints slim and possibly faster boot time. We have tested this by enabling CONFIG_OF for x86. I can add this info in cover letter in next version. Regards, Saurabh > > Rob
On Tue, Jan 31, 2023 at 06:04:49PM -0800, Saurabh Singh Sengar wrote: > On Tue, Jan 31, 2023 at 02:27:51PM -0600, Rob Herring wrote: > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > <ssengar@linux.microsoft.com> wrote: > > > > > > This set of patches expands the VMBus driver to include device tree > > > support. > > > > > > The first two patches enable compilation of Hyper-V APIs in a non-ACPI > > > build. > > > > > > The third patch converts the VMBus driver from acpi to more generic > > > platform driver. > > > > > > Further to add device tree documentation for VMBus, it needs to club with > > > other virtualization driver's documentation. For this rename the virtio > > > folder to more generic hypervisor, so that all the hypervisor based > > > devices can co-exist in a single place in device tree documentation. The > > > fourth patch does this renaming. > > > > > > The fifth patch introduces the device tree documentation for VMBus. > > > > > > The sixth patch adds device tree support to the VMBus driver. Currently > > > this is tested only for x86 and it may not work for other archs. > > > > I can read all the patches and see *what* they do. You don't really > > need to list that here. I'm still wondering *why*. That is what the > > cover letter and commit messages should answer. Why do you need DT > > support? How does this even work on x86? FDT is only enabled for > > CE4100 platform. > > HI Rob, > > Thanks for your comments. > We are working on a solution where kernel is booted without ACPI tables to keep > the overall system's memory footprints slim and possibly faster boot time. > We have tested this by enabling CONFIG_OF for x86. It's CONFIG_OF_EARLY_FLATTREE which you would need and that's not user selectable. At a minimum, you need some kconfig changes. Where are those? Also see my comment on v1 about running DT validation on your dtb. I'm sure running it would point out other issues. Such as the root level comaptible string(s) need to be documented. You need cpu nodes, interrupt controller, timers, etc. Those all have to be documented. Rob
On Wed, Feb 01, 2023 at 08:51:46AM -0600, Rob Herring wrote: > On Tue, Jan 31, 2023 at 06:04:49PM -0800, Saurabh Singh Sengar wrote: > > On Tue, Jan 31, 2023 at 02:27:51PM -0600, Rob Herring wrote: > > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > > <ssengar@linux.microsoft.com> wrote: > > > > > > > > This set of patches expands the VMBus driver to include device tree > > > > support. > > > > > > > > The first two patches enable compilation of Hyper-V APIs in a non-ACPI > > > > build. > > > > > > > > The third patch converts the VMBus driver from acpi to more generic > > > > platform driver. > > > > > > > > Further to add device tree documentation for VMBus, it needs to club with > > > > other virtualization driver's documentation. For this rename the virtio > > > > folder to more generic hypervisor, so that all the hypervisor based > > > > devices can co-exist in a single place in device tree documentation. The > > > > fourth patch does this renaming. > > > > > > > > The fifth patch introduces the device tree documentation for VMBus. > > > > > > > > The sixth patch adds device tree support to the VMBus driver. Currently > > > > this is tested only for x86 and it may not work for other archs. > > > > > > I can read all the patches and see *what* they do. You don't really > > > need to list that here. I'm still wondering *why*. That is what the > > > cover letter and commit messages should answer. Why do you need DT > > > support? How does this even work on x86? FDT is only enabled for > > > CE4100 platform. > > > > HI Rob, > > > > Thanks for your comments. > > We are working on a solution where kernel is booted without ACPI tables to keep > > the overall system's memory footprints slim and possibly faster boot time. > > We have tested this by enabling CONFIG_OF for x86. > > It's CONFIG_OF_EARLY_FLATTREE which you would need and that's not user > selectable. At a minimum, you need some kconfig changes. Where are > those? You are right we have define a new config flag in Kconfig, and selected CONFIG_OF and CONFIG_OF_EARLY_FLATTREE. We are working on upstreaming that patch as well however that will be a separate patch series. > > Also see my comment on v1 about running DT validation on your dtb. I'm > sure running it would point out other issues. Such as the root level > comaptible string(s) need to be documented. You need cpu nodes, > interrupt controller, timers, etc. Those all have to be documented. I will be changing the parent node to soc node as suggested by Krzysztof in other thread. soc { #address-cells = <2>; #size-cells = <2>; vmbus@ff0000000 { #address-cells = <2>; #size-cells = <1>; compatible = "Microsoft,vmbus"; ranges = <0x00 0x00 0x0f 0xf0000000 0x10000000>; }; }; This will be sufficient. Regards, Saurabh > > Rob
On Tue, Jan 31, 2023 at 02:12:53PM -0600, Rob Herring wrote: > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > <ssengar@linux.microsoft.com> wrote: > > > > Update the driver to support device tree boot as well along with ACPI. > > At present the device tree parsing only provides the mmio region info > > and is not the exact copy of ACPI parsing. This is sufficient to cater > > all the current device tree usecases for VMBus. > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > --- > > drivers/hv/vmbus_drv.c | 75 ++++++++++++++++++++++++++++++++++++++++-- > > 1 file changed, 73 insertions(+), 2 deletions(-) > > > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > > index 49030e756b9f..1741f1348f9f 100644 > > --- a/drivers/hv/vmbus_drv.c > > +++ b/drivers/hv/vmbus_drv.c > > @@ -2152,7 +2152,7 @@ void vmbus_device_unregister(struct hv_device *device_obj) > > device_unregister(&device_obj->device); > > } > > > > - > > +#ifdef CONFIG_ACPI > > /* > > * VMBUS is an acpi enumerated device. Get the information we > > * need from DSDT. > > @@ -2262,6 +2262,7 @@ static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx) > > > > return AE_OK; > > } > > +#endif > > > > static void vmbus_mmio_remove(void) > > { > > @@ -2282,7 +2283,7 @@ static void vmbus_mmio_remove(void) > > } > > } > > > > -static void vmbus_reserve_fb(void) > > +static void __maybe_unused vmbus_reserve_fb(void) > > { > > resource_size_t start = 0, size; > > struct pci_dev *pdev; > > @@ -2442,6 +2443,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size) > > } > > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > > > +#ifdef CONFIG_ACPI > > It's better to put C 'if (!IS_ENABLED(CONFIG_ACPI)' code in the I wanted to have separate function for ACPI and device tree flow, which can be easily maintained with #ifdef. Please let me know if its fine. > > > static int vmbus_acpi_add(struct platform_device *pdev) > > { > > acpi_status result; > > @@ -2496,10 +2498,68 @@ static int vmbus_acpi_add(struct platform_device *pdev) > > vmbus_mmio_remove(); > > return ret_val; > > } > > +#else > > + > > +static int vmbus_device_add(struct platform_device *pdev) > > +{ > > + struct resource **cur_res = &hyperv_mmio; > > + struct device_node *np; > > + u32 *ranges, len; > > + u64 start; > > + int nr_ranges, child_cells = 2, cur_cell = 0, ret = 0; > > + > > + hv_dev = pdev; > > + np = pdev->dev.of_node; > > + > > + nr_ranges = device_property_count_u32(&pdev->dev, "ranges"); > > Parsing ranges yourself is a bad sign. It's a standard property and we > have functions which handle it. If those don't work, then something is > wrong with your DT or they need to be fixed/expanded. I find all the standard functions which parse "ranges" property are doing much more then I need. Our requirement is to only pass the mmio memory range and size, I couldn't find any standard API doing this. I see some of the drivers are using these APIs to parse ranges property hence I follwed those examples. I will be happy to improve it if I get any better alternative. > > > + if (nr_ranges < 0) > > + return nr_ranges; > > + ranges = kcalloc(nr_ranges, sizeof(u32), GFP_KERNEL); > > + if (!ranges) > > + return -ENOMEM; > > + > > + if (device_property_read_u32_array(&pdev->dev, "ranges", ranges, nr_ranges)) { > > + ret = -EINVAL; > > + goto free_ranges; > > + } > > + > > + while (cur_cell < nr_ranges) { > > + struct resource *res; > > + > > + /* The first u64 in the ranges description isn't used currently. */ > > + cur_cell = cur_cell + child_cells; > > + start = ranges[cur_cell++]; > > + start = (start << 32) | ranges[cur_cell++]; > > + len = ranges[cur_cell++]; > > To expand my last point, the format of ranges is <child_addr > parent_addr length>. That's not what your 'ranges' has. You've also > just ignored '#address-cells' and '#size-cells'. Got it. However I need to check if there is any standard API which can give me these values, otherwise I may have to parse these as well :( Regards, Saurabh > > > + > > + res = kzalloc(sizeof(*res), GFP_ATOMIC); > > + if (!res) { > > + ret = -ENOMEM; > > + goto free_ranges; > > + } > > + > > + res->name = "hyperv mmio"; > > + res->flags = IORESOURCE_MEM | IORESOURCE_MEM_64; > > + res->start = start; > > + res->end = start + len; > > + > > + *cur_res = res; > > + cur_res = &res->sibling; > > + } > > + > > +free_ranges: > > + kfree(ranges); > > + return ret; > > +} > > +#endif > > > > static int vmbus_platform_driver_probe(struct platform_device *pdev) > > { > > +#ifdef CONFIG_ACPI > > return vmbus_acpi_add(pdev); > > +#else > > + return vmbus_device_add(pdev); > > +#endif > > } > > > > static int vmbus_platform_driver_remove(struct platform_device *pdev) > > @@ -2645,6 +2705,16 @@ static int vmbus_bus_resume(struct device *dev) > > #define vmbus_bus_resume NULL > > #endif /* CONFIG_PM_SLEEP */ > > > > +static const struct of_device_id vmbus_of_match[] = { > > + { > > + .compatible = "msft,vmbus", > > + }, > > + { > > + /* sentinel */ > > + }, > > +}; > > +MODULE_DEVICE_TABLE(of, vmbus_of_match); > > + > > static const struct acpi_device_id vmbus_acpi_device_ids[] = { > > {"VMBUS", 0}, > > {"VMBus", 0}, > > @@ -2679,6 +2749,7 @@ static struct platform_driver vmbus_platform_driver = { > > .driver = { > > .name = "vmbus", > > .acpi_match_table = ACPI_PTR(vmbus_acpi_device_ids), > > + .of_match_table = of_match_ptr(vmbus_of_match), > > .pm = &vmbus_bus_pm, > > .probe_type = PROBE_FORCE_SYNCHRONOUS, > > } > > -- > > 2.25.1 > >
On 01/02/2023 17:34, Saurabh Singh Sengar wrote: >> Also see my comment on v1 about running DT validation on your dtb. I'm >> sure running it would point out other issues. Such as the root level >> comaptible string(s) need to be documented. You need cpu nodes, >> interrupt controller, timers, etc. Those all have to be documented. > > I will be changing the parent node to soc node as suggested by Krzysztof > in other thread. > > soc { > #address-cells = <2>; > #size-cells = <2>; > > vmbus@ff0000000 { > #address-cells = <2>; > #size-cells = <1>; > compatible = "Microsoft,vmbus"; > ranges = <0x00 0x00 0x0f 0xf0000000 0x10000000>; > }; > }; > > This will be sufficient. It will be ok for the example, but will not be ok for supporting your use case. Please solve all the points from Rob's comment above. Where is their documentation? Best regards, Krzysztof
On Wed, Feb 01, 2023 at 08:51:33AM -0800, Saurabh Singh Sengar wrote: > On Tue, Jan 31, 2023 at 02:12:53PM -0600, Rob Herring wrote: > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > <ssengar@linux.microsoft.com> wrote: > > > > > > Update the driver to support device tree boot as well along with ACPI. > > > At present the device tree parsing only provides the mmio region info > > > and is not the exact copy of ACPI parsing. This is sufficient to cater > > > all the current device tree usecases for VMBus. > > > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > > --- > > > drivers/hv/vmbus_drv.c | 75 ++++++++++++++++++++++++++++++++++++++++-- > > > 1 file changed, 73 insertions(+), 2 deletions(-) > > > > > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > > > index 49030e756b9f..1741f1348f9f 100644 > > > --- a/drivers/hv/vmbus_drv.c > > > +++ b/drivers/hv/vmbus_drv.c > > > @@ -2152,7 +2152,7 @@ void vmbus_device_unregister(struct hv_device *device_obj) > > > device_unregister(&device_obj->device); > > > } > > > > > > - > > > +#ifdef CONFIG_ACPI > > > /* > > > * VMBUS is an acpi enumerated device. Get the information we > > > * need from DSDT. > > > @@ -2262,6 +2262,7 @@ static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx) > > > > > > return AE_OK; > > > } > > > +#endif > > > > > > static void vmbus_mmio_remove(void) > > > { > > > @@ -2282,7 +2283,7 @@ static void vmbus_mmio_remove(void) > > > } > > > } > > > > > > -static void vmbus_reserve_fb(void) > > > +static void __maybe_unused vmbus_reserve_fb(void) > > > { > > > resource_size_t start = 0, size; > > > struct pci_dev *pdev; > > > @@ -2442,6 +2443,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size) > > > } > > > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > > > > > +#ifdef CONFIG_ACPI > > > > It's better to put C 'if (!IS_ENABLED(CONFIG_ACPI)' code in the > > I wanted to have separate function for ACPI and device tree flow, which > can be easily maintained with #ifdef. Please let me know if its fine. Yes, you can have separate functions: static int vmbus_acpi_add(struct platform_device *pdev) { if (!IS_ENABLED(CONFIG_ACPI)) return -ENODEV; ... } The compiler will throw away the function in the end if CONFIG_ACPI is not enabled. That is easier for us to maintain because it reduces the combinations to build. > > > > > > static int vmbus_acpi_add(struct platform_device *pdev) > > > { > > > acpi_status result; > > > @@ -2496,10 +2498,68 @@ static int vmbus_acpi_add(struct platform_device *pdev) > > > vmbus_mmio_remove(); > > > return ret_val; > > > } > > > +#else > > > + > > > +static int vmbus_device_add(struct platform_device *pdev) > > > +{ > > > + struct resource **cur_res = &hyperv_mmio; > > > + struct device_node *np; > > > + u32 *ranges, len; > > > + u64 start; > > > + int nr_ranges, child_cells = 2, cur_cell = 0, ret = 0; > > > + > > > + hv_dev = pdev; > > > + np = pdev->dev.of_node; > > > + > > > + nr_ranges = device_property_count_u32(&pdev->dev, "ranges"); > > > > Parsing ranges yourself is a bad sign. It's a standard property and we > > have functions which handle it. If those don't work, then something is > > wrong with your DT or they need to be fixed/expanded. > > I find all the standard functions which parse "ranges" property are doing > much more then I need. Our requirement is to only pass the mmio memory range > and size, I couldn't find any standard API doing this. You can't just change how standard properties work to suit your needs. We shouldn't even be having this discussion because we have tools to check all this now. dtc does some and dtschema does a lot more. > I see some of the drivers are using these APIs to parse ranges property hence > I follwed those examples. I will be happy to improve it if I get any better > alternative. You can always find bad examples to follow... > > > + if (nr_ranges < 0) > > > + return nr_ranges; > > > + ranges = kcalloc(nr_ranges, sizeof(u32), GFP_KERNEL); > > > + if (!ranges) > > > + return -ENOMEM; > > > + > > > + if (device_property_read_u32_array(&pdev->dev, "ranges", ranges, nr_ranges)) { > > > + ret = -EINVAL; > > > + goto free_ranges; > > > + } > > > + > > > + while (cur_cell < nr_ranges) { > > > + struct resource *res; > > > + > > > + /* The first u64 in the ranges description isn't used currently. */ > > > + cur_cell = cur_cell + child_cells; > > > + start = ranges[cur_cell++]; > > > + start = (start << 32) | ranges[cur_cell++]; > > > + len = ranges[cur_cell++]; > > > > To expand my last point, the format of ranges is <child_addr > > parent_addr length>. That's not what your 'ranges' has. You've also > > just ignored '#address-cells' and '#size-cells'. > > Got it. However I need to check if there is any standard API which can > give me these values, otherwise I may have to parse these as well :( for_each_of_range() That is not how linux works. When the core code doesn't do what you want, you adapt it to your needs. You don't work around it. Read this[1]. Rob [1] https://lwn.net/Articles/443531/
From: Saurabh Sengar <ssengar@linux.microsoft.com> Sent: Tuesday, January 31, 2023 10:10 AM > > acpi_sleep_state_supported API is only define for CONFIG_ACPI flag and > thus it can't be used for non-ACPI builds. Initaly there won't be s/Initaly/Initially/ > hibernate support for non ACPI builds. s/hibernate/hibernation/ > > This change will help adding device tree support in subsequent commits. > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > --- > drivers/hv/hv_common.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/hv/hv_common.c b/drivers/hv/hv_common.c > index 52a6f89ccdbd..370ec20d1993 100644 > --- a/drivers/hv/hv_common.c > +++ b/drivers/hv/hv_common.c > @@ -234,7 +234,11 @@ EXPORT_SYMBOL_GPL(hv_setup_dma_ops); > > bool hv_is_hibernation_supported(void) > { > +#ifdef CONFIG_ACPI > return !hv_root_partition && acpi_sleep_state_supported(ACPI_STATE_S4); > +#else > + return false; > +#endif Is this patch needed? If CONFIG_ACPI is not set, then per arch/x86/Kconfig, CONFIG_ACPI_SYSTEM_POWER_STATES_SUPPORT is not selected. In that case, the #ifdef in include/acpi/acpi_bus.h provides a stub for acpi_sleep_state_supported() that returns "false". So it seems like the existing code should compile and correctly return "false" when CONFIG_ACPI is not set. Michael > } > EXPORT_SYMBOL_GPL(hv_is_hibernation_supported); > > -- > 2.25.1
On Wed, Feb 01, 2023 at 11:46:38AM -0600, Rob Herring wrote: > On Wed, Feb 01, 2023 at 08:51:33AM -0800, Saurabh Singh Sengar wrote: > > On Tue, Jan 31, 2023 at 02:12:53PM -0600, Rob Herring wrote: > > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > > <ssengar@linux.microsoft.com> wrote: > > > > (...) > > > > @@ -2442,6 +2443,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size) > > > > } > > > > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > > > > > > > +#ifdef CONFIG_ACPI > > > > > > It's better to put C 'if (!IS_ENABLED(CONFIG_ACPI)' code in the > > > > I wanted to have separate function for ACPI and device tree flow, which > > can be easily maintained with #ifdef. Please let me know if its fine. > > Yes, you can have separate functions: > > static int vmbus_acpi_add(struct platform_device *pdev) > { > if (!IS_ENABLED(CONFIG_ACPI)) > return -ENODEV; > > ... > } > > The compiler will throw away the function in the end if CONFIG_ACPI is > not enabled. > > That is easier for us to maintain because it reduces the combinations to > build. Thanks, Will fix this in v3. > > > > > > > > > > static int vmbus_acpi_add(struct platform_device *pdev) > > > > { > > > > acpi_status result; > > > > @@ -2496,10 +2498,68 @@ static int vmbus_acpi_add(struct platform_device *pdev) > > > > vmbus_mmio_remove(); > > > > return ret_val; > > > > } > > > > +#else > > > > + > > > > +static int vmbus_device_add(struct platform_device *pdev) > > > > +{ > > > > + struct resource **cur_res = &hyperv_mmio; > > > > + struct device_node *np; > > > > + u32 *ranges, len; > > > > + u64 start; > > > > + int nr_ranges, child_cells = 2, cur_cell = 0, ret = 0; > > > > + > > > > + hv_dev = pdev; > > > > + np = pdev->dev.of_node; > > > > + > > > > + nr_ranges = device_property_count_u32(&pdev->dev, "ranges"); > > > > > > Parsing ranges yourself is a bad sign. It's a standard property and we > > > have functions which handle it. If those don't work, then something is > > > wrong with your DT or they need to be fixed/expanded. > > > > I find all the standard functions which parse "ranges" property are doing > > much more then I need. Our requirement is to only pass the mmio memory range > > and size, I couldn't find any standard API doing this. > > You can't just change how standard properties work to suit your needs. > > We shouldn't even be having this discussion because we have tools to > check all this now. dtc does some and dtschema does a lot more. > > > I see some of the drivers are using these APIs to parse ranges property hence > > I follwed those examples. I will be happy to improve it if I get any better > > alternative. > > You can always find bad examples to follow... > > > > > + if (nr_ranges < 0) > > > > + return nr_ranges; > > > > + ranges = kcalloc(nr_ranges, sizeof(u32), GFP_KERNEL); > > > > + if (!ranges) > > > > + return -ENOMEM; > > > > + > > > > + if (device_property_read_u32_array(&pdev->dev, "ranges", ranges, nr_ranges)) { > > > > + ret = -EINVAL; > > > > + goto free_ranges; > > > > + } > > > > + > > > > + while (cur_cell < nr_ranges) { > > > > + struct resource *res; > > > > + > > > > + /* The first u64 in the ranges description isn't used currently. */ > > > > + cur_cell = cur_cell + child_cells; > > > > + start = ranges[cur_cell++]; > > > > + start = (start << 32) | ranges[cur_cell++]; > > > > + len = ranges[cur_cell++]; > > > > > > To expand my last point, the format of ranges is <child_addr > > > parent_addr length>. That's not what your 'ranges' has. You've also > > > just ignored '#address-cells' and '#size-cells'. > > > > Got it. However I need to check if there is any standard API which can > > give me these values, otherwise I may have to parse these as well :( > > for_each_of_range() > > That is not how linux works. When the core code doesn't do what you > want, you adapt it to your needs. You don't work around it. Read > this[1]. > > Rob > > [1] https://lwn.net/Articles/443531/ Thanks I will work on this suggestion and fix this up in next version.
From: Saurabh Sengar <ssengar@linux.microsoft.com> Sent: Tuesday, January 31, 2023 10:10 AM > > Use more generic platform device instead of acpi device. Also rename the > function vmbus_acpi_remove to more generic name vmbus_mmio_remove. > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > --- > drivers/hv/vmbus_drv.c | 78 +++++++++++++++++++++++++----------------- > 1 file changed, 46 insertions(+), 32 deletions(-) > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > index d24dd65b33d4..49030e756b9f 100644 > --- a/drivers/hv/vmbus_drv.c > +++ b/drivers/hv/vmbus_drv.c > @@ -12,6 +12,7 @@ > #include <linux/init.h> > #include <linux/module.h> > #include <linux/device.h> > +#include <linux/platform_device.h> > #include <linux/interrupt.h> > #include <linux/sysctl.h> > #include <linux/slab.h> > @@ -44,7 +45,7 @@ struct vmbus_dynid { > struct hv_vmbus_device_id id; > }; > > -static struct acpi_device *hv_acpi_dev; > +static struct platform_device *hv_dev; > > static int hyperv_cpuhp_online; > > @@ -143,7 +144,7 @@ static DEFINE_MUTEX(hyperv_mmio_lock); > > static int vmbus_exists(void) > { > - if (hv_acpi_dev == NULL) > + if (hv_dev == NULL) > return -ENODEV; > > return 0; > @@ -932,7 +933,7 @@ static int vmbus_dma_configure(struct device *child_device) > * On x86/x64 coherence is assumed and these calls have no effect. > */ > hv_setup_dma_ops(child_device, > - device_get_dma_attr(&hv_acpi_dev->dev) == DEV_DMA_COHERENT); > + device_get_dma_attr(&hv_dev->dev) == DEV_DMA_COHERENT); > return 0; > } > > @@ -2090,7 +2091,7 @@ int vmbus_device_register(struct hv_device > *child_device_obj) > &child_device_obj->channel->offermsg.offer.if_instance); > > child_device_obj->device.bus = &hv_bus; > - child_device_obj->device.parent = &hv_acpi_dev->dev; > + child_device_obj->device.parent = &hv_dev->dev; > child_device_obj->device.release = vmbus_device_release; > > child_device_obj->device.dma_parms = &child_device_obj->dma_parms; > @@ -2262,7 +2263,7 @@ static acpi_status vmbus_walk_resources(struct > acpi_resource *res, void *ctx) > return AE_OK; > } > > -static void vmbus_acpi_remove(struct acpi_device *device) > +static void vmbus_mmio_remove(void) > { > struct resource *cur_res; > struct resource *next_res; > @@ -2441,13 +2442,15 @@ void vmbus_free_mmio(resource_size_t start, > resource_size_t size) > } > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > -static int vmbus_acpi_add(struct acpi_device *device) > +static int vmbus_acpi_add(struct platform_device *pdev) > { > acpi_status result; > int ret_val = -ENODEV; > - struct acpi_device *ancestor; > + struct platform_device *ancestor; > + struct acpi_device *adev = to_acpi_device(&pdev->dev); This doesn't work. The argument to vmbus_acpi_add() is a struct platform_device, which has a struct device embedded in it (not a pointer). to_acpi_device() takes a struct device as an argument, assuming that the struct device is embedded in a struct acpi_device, which is not the case here. The resulting local variable adev is actually pointing to some (perhaps negative) offset within the struct platform_device, and uses of adev are getting unknown random data from within (or before) the struct platform_device. > > - hv_acpi_dev = device; > + hv_dev = pdev; > + adev->fwnode.dev = &pdev->dev; > > /* > * Older versions of Hyper-V for ARM64 fail to include the _CCA > @@ -2456,15 +2459,16 @@ static int vmbus_acpi_add(struct acpi_device *device) > * up the ACPI device to behave as if _CCA is present and indicates > * hardware coherence. > */ > - ACPI_COMPANION_SET(&device->dev, device); > + ACPI_COMPANION_SET(&pdev->dev, ACPI_COMPANION(&pdev->dev)); This statement seems tautological. If ACPI_COMPANION(&pdev->dev) returns a valid result, why would the ACPI companion for &pdev->dev need to be set? The original code was setting the ACPI companion for the embedded struct device to be the struct acpi_device. I forget why this wasn't already done for the VMBus device when it was originally parsed from the ACPI DSDT ... > if (IS_ENABLED(CONFIG_ACPI_CCA_REQUIRED) && > - device_get_dma_attr(&device->dev) == DEV_DMA_NOT_SUPPORTED) { > + device_get_dma_attr(&pdev->dev) == DEV_DMA_NOT_SUPPORTED) { > + struct acpi_device *adev_node = ACPI_COMPANION(&pdev->dev); If earlier code in this function can get a correct pointer to the struct acpi_device, then this statement shouldn't be necessary. You already have it. > pr_info("No ACPI _CCA found; assuming coherent device I/O\n"); > - device->flags.cca_seen = true; > - device->flags.coherent_dma = true; > + adev_node->flags.cca_seen = true; > + adev_node->flags.coherent_dma = true; > } > > - result = acpi_walk_resources(device->handle, METHOD_NAME__CRS, > + result = acpi_walk_resources(ACPI_HANDLE(&pdev->dev), METHOD_NAME__CRS, Again, if you have a correct pointer to the struct acpi_device, then adev->handle (like the original code) should be simpler than looking it up again with ACPI_HANDLE(). > vmbus_walk_resources, NULL); > > if (ACPI_FAILURE(result)) > @@ -2473,9 +2477,9 @@ static int vmbus_acpi_add(struct acpi_device *device) > * Some ancestor of the vmbus acpi device (Gen1 or Gen2 > * firmware) is the VMOD that has the mmio ranges. Get that. > */ > - for (ancestor = acpi_dev_parent(device); ancestor; > - ancestor = acpi_dev_parent(ancestor)) { > - result = acpi_walk_resources(ancestor->handle, METHOD_NAME__CRS, > + for (ancestor = to_platform_device(pdev->dev.parent); ancestor; > + ancestor = to_platform_device(ancestor->dev.parent)) { > + result = acpi_walk_resources(ACPI_HANDLE(&ancestor->dev), METHOD_NAME__CRS, Similarly, if you get a correct pointer to the struct acpi_device, does the above code need any changes? I'm hoping not. > vmbus_walk_resources, NULL); > > if (ACPI_FAILURE(result)) > @@ -2489,10 +2493,21 @@ static int vmbus_acpi_add(struct acpi_device *device) > > acpi_walk_err: > if (ret_val) > - vmbus_acpi_remove(device); > + vmbus_mmio_remove(); > return ret_val; > } > > +static int vmbus_platform_driver_probe(struct platform_device *pdev) > +{ > + return vmbus_acpi_add(pdev); > +} > + > +static int vmbus_platform_driver_remove(struct platform_device *pdev) > +{ > + vmbus_mmio_remove(); > + return 0; > +} > + > #ifdef CONFIG_PM_SLEEP > static int vmbus_bus_suspend(struct device *dev) > { > @@ -2658,15 +2673,15 @@ static const struct dev_pm_ops vmbus_bus_pm = { > .restore_noirq = vmbus_bus_resume > }; > > -static struct acpi_driver vmbus_acpi_driver = { > - .name = "vmbus", > - .ids = vmbus_acpi_device_ids, > - .ops = { > - .add = vmbus_acpi_add, > - .remove = vmbus_acpi_remove, > - }, > - .drv.pm = &vmbus_bus_pm, > - .drv.probe_type = PROBE_FORCE_SYNCHRONOUS, > +static struct platform_driver vmbus_platform_driver = { > + .probe = vmbus_platform_driver_probe, > + .remove = vmbus_platform_driver_remove, > + .driver = { > + .name = "vmbus", > + .acpi_match_table = ACPI_PTR(vmbus_acpi_device_ids), > + .pm = &vmbus_bus_pm, > + .probe_type = PROBE_FORCE_SYNCHRONOUS, > + } > }; > > static void hv_kexec_handler(void) > @@ -2750,12 +2765,11 @@ static int __init hv_acpi_init(void) > /* > * Get ACPI resources first. > */ > - ret = acpi_bus_register_driver(&vmbus_acpi_driver); > - > + ret = platform_driver_register(&vmbus_platform_driver); > if (ret) > return ret; > > - if (!hv_acpi_dev) { > + if (!hv_dev) { > ret = -ENODEV; > goto cleanup; > } > @@ -2785,8 +2799,8 @@ static int __init hv_acpi_init(void) > return 0; > > cleanup: > - acpi_bus_unregister_driver(&vmbus_acpi_driver); > - hv_acpi_dev = NULL; > + platform_driver_unregister(&vmbus_platform_driver); > + hv_dev = NULL; > return ret; > } > > @@ -2839,7 +2853,7 @@ static void __exit vmbus_exit(void) > > cpuhp_remove_state(hyperv_cpuhp_online); > hv_synic_free(); > - acpi_bus_unregister_driver(&vmbus_acpi_driver); > + platform_driver_unregister(&vmbus_platform_driver); > } > > > -- > 2.25.1
On Wed, Feb 01, 2023 at 05:47:44PM +0000, Michael Kelley (LINUX) wrote: > From: Saurabh Sengar <ssengar@linux.microsoft.com> Sent: Tuesday, January 31, 2023 10:10 AM > > > > acpi_sleep_state_supported API is only define for CONFIG_ACPI flag and > > thus it can't be used for non-ACPI builds. Initaly there won't be > > s/Initaly/Initially/ OK > > > hibernate support for non ACPI builds. > > s/hibernate/hibernation/ OK > > > > > This change will help adding device tree support in subsequent commits. > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > --- > > drivers/hv/hv_common.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > > diff --git a/drivers/hv/hv_common.c b/drivers/hv/hv_common.c > > index 52a6f89ccdbd..370ec20d1993 100644 > > --- a/drivers/hv/hv_common.c > > +++ b/drivers/hv/hv_common.c > > @@ -234,7 +234,11 @@ EXPORT_SYMBOL_GPL(hv_setup_dma_ops); > > > > bool hv_is_hibernation_supported(void) > > { > > +#ifdef CONFIG_ACPI > > return !hv_root_partition && acpi_sleep_state_supported(ACPI_STATE_S4); > > +#else > > + return false; > > +#endif > > Is this patch needed? If CONFIG_ACPI is not set, then per > arch/x86/Kconfig, CONFIG_ACPI_SYSTEM_POWER_STATES_SUPPORT > is not selected. In that case, the #ifdef in include/acpi/acpi_bus.h > provides a stub for acpi_sleep_state_supported() that returns "false". > So it seems like the existing code should compile and correctly return > "false" when CONFIG_ACPI is not set. You are right, if CONFIG_ACPI_SYSTEM_POWER_STATES_SUPPORT is not set acpi_sleep_state_supported will return false, but this is applicable only when CONFIG_ACPI is enable. If CONFIG_ACPI is not enable both these functions are not defined. Regards, Saurabh > > Michael > > > } > > EXPORT_SYMBOL_GPL(hv_is_hibernation_supported); > > > > -- > > 2.25.1
On Wed, Feb 01, 2023 at 06:32:29PM +0000, Michael Kelley (LINUX) wrote: > From: Saurabh Sengar <ssengar@linux.microsoft.com> Sent: Tuesday, January 31, 2023 10:10 AM > > > > Use more generic platform device instead of acpi device. Also rename the > > function vmbus_acpi_remove to more generic name vmbus_mmio_remove. > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > --- > > drivers/hv/vmbus_drv.c | 78 +++++++++++++++++++++++++----------------- > > 1 file changed, 46 insertions(+), 32 deletions(-) > > > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > > index d24dd65b33d4..49030e756b9f 100644 > > --- a/drivers/hv/vmbus_drv.c > > +++ b/drivers/hv/vmbus_drv.c > > @@ -12,6 +12,7 @@ > > #include <linux/init.h> > > #include <linux/module.h> > > #include <linux/device.h> > > +#include <linux/platform_device.h> > > #include <linux/interrupt.h> > > #include <linux/sysctl.h> > > #include <linux/slab.h> > > @@ -44,7 +45,7 @@ struct vmbus_dynid { > > struct hv_vmbus_device_id id; > > }; > > > > -static struct acpi_device *hv_acpi_dev; > > +static struct platform_device *hv_dev; > > > > static int hyperv_cpuhp_online; > > > > @@ -143,7 +144,7 @@ static DEFINE_MUTEX(hyperv_mmio_lock); > > > > static int vmbus_exists(void) > > { > > - if (hv_acpi_dev == NULL) > > + if (hv_dev == NULL) > > return -ENODEV; > > > > return 0; > > @@ -932,7 +933,7 @@ static int vmbus_dma_configure(struct device *child_device) > > * On x86/x64 coherence is assumed and these calls have no effect. > > */ > > hv_setup_dma_ops(child_device, > > - device_get_dma_attr(&hv_acpi_dev->dev) == DEV_DMA_COHERENT); > > + device_get_dma_attr(&hv_dev->dev) == DEV_DMA_COHERENT); > > return 0; > > } > > > > @@ -2090,7 +2091,7 @@ int vmbus_device_register(struct hv_device > > *child_device_obj) > > &child_device_obj->channel->offermsg.offer.if_instance); > > > > child_device_obj->device.bus = &hv_bus; > > - child_device_obj->device.parent = &hv_acpi_dev->dev; > > + child_device_obj->device.parent = &hv_dev->dev; > > child_device_obj->device.release = vmbus_device_release; > > > > child_device_obj->device.dma_parms = &child_device_obj->dma_parms; > > @@ -2262,7 +2263,7 @@ static acpi_status vmbus_walk_resources(struct > > acpi_resource *res, void *ctx) > > return AE_OK; > > } > > > > -static void vmbus_acpi_remove(struct acpi_device *device) > > +static void vmbus_mmio_remove(void) > > { > > struct resource *cur_res; > > struct resource *next_res; > > @@ -2441,13 +2442,15 @@ void vmbus_free_mmio(resource_size_t start, > > resource_size_t size) > > } > > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > > > -static int vmbus_acpi_add(struct acpi_device *device) > > +static int vmbus_acpi_add(struct platform_device *pdev) > > { > > acpi_status result; > > int ret_val = -ENODEV; > > - struct acpi_device *ancestor; > > + struct platform_device *ancestor; > > + struct acpi_device *adev = to_acpi_device(&pdev->dev); > > This doesn't work. The argument to vmbus_acpi_add() is a struct > platform_device, which has a struct device embedded in it (not a > pointer). to_acpi_device() takes a struct device as an argument, > assuming that the struct device is embedded in a struct > acpi_device, which is not the case here. The resulting local > variable adev is actually pointing to some (perhaps negative) > offset within the struct platform_device, and uses of adev are > getting unknown random data from within (or before) the > struct platform_device. Thanks for review. If I understand correctly you are saying to change the argument of to_acpi_device from pointer to 'struct device'. I tried changing this statement to: "struct acpi_device *adev = to_acpi_device(pdev->dev);" However on compiling this code I see this error: ./include/linux/container_of.h:19:28: error: invalid type argument of unary ‘*’ (have ‘struct device’) Please let me know if I missunderstood your suggestion. I would like to mention that this code has been tested and verified. The purpose of this change was to assign the device node to the acpi firmware node, as seen in the line "adev->fwnode.dev = &pdev->dev;". This was necessary because without it, the mlx driver would crash as it searches for the device in the acpi firmware node, which would otherwise be NULL. I am confident that the addresses are correctly assigned, as any issues with the assignments would have caused a panic. Regards, Saurabh > > > > > - hv_acpi_dev = device; > > + hv_dev = pdev; > > + adev->fwnode.dev = &pdev->dev; > > > > /* > > * Older versions of Hyper-V for ARM64 fail to include the _CCA > > @@ -2456,15 +2459,16 @@ static int vmbus_acpi_add(struct acpi_device *device) > > * up the ACPI device to behave as if _CCA is present and indicates > > * hardware coherence. > > */ > > - ACPI_COMPANION_SET(&device->dev, device); > > + ACPI_COMPANION_SET(&pdev->dev, ACPI_COMPANION(&pdev->dev)); > > This statement seems tautological. If ACPI_COMPANION(&pdev->dev) > returns a valid result, why would the ACPI companion for &pdev->dev > need to be set? The original code was setting the ACPI companion for the > embedded struct device to be the struct acpi_device. I forget why this > wasn't already done for the VMBus device when it was originally parsed > from the ACPI DSDT ... > > > if (IS_ENABLED(CONFIG_ACPI_CCA_REQUIRED) && > > - device_get_dma_attr(&device->dev) == DEV_DMA_NOT_SUPPORTED) { > > + device_get_dma_attr(&pdev->dev) == DEV_DMA_NOT_SUPPORTED) { > > + struct acpi_device *adev_node = ACPI_COMPANION(&pdev->dev); > > If earlier code in this function can get a correct pointer to the struct acpi_device, > then this statement shouldn't be necessary. You already have it. > > > pr_info("No ACPI _CCA found; assuming coherent device I/O\n"); > > - device->flags.cca_seen = true; > > - device->flags.coherent_dma = true; > > + adev_node->flags.cca_seen = true; > > + adev_node->flags.coherent_dma = true; > > } > > > > - result = acpi_walk_resources(device->handle, METHOD_NAME__CRS, > > + result = acpi_walk_resources(ACPI_HANDLE(&pdev->dev), METHOD_NAME__CRS, > > Again, if you have a correct pointer to the struct acpi_device, then adev->handle > (like the original code) should be simpler than looking it up again with ACPI_HANDLE(). > > > vmbus_walk_resources, NULL); > > > > if (ACPI_FAILURE(result)) > > @@ -2473,9 +2477,9 @@ static int vmbus_acpi_add(struct acpi_device *device) > > * Some ancestor of the vmbus acpi device (Gen1 or Gen2 > > * firmware) is the VMOD that has the mmio ranges. Get that. > > */ > > - for (ancestor = acpi_dev_parent(device); ancestor; > > - ancestor = acpi_dev_parent(ancestor)) { > > - result = acpi_walk_resources(ancestor->handle, METHOD_NAME__CRS, > > + for (ancestor = to_platform_device(pdev->dev.parent); ancestor; > > + ancestor = to_platform_device(ancestor->dev.parent)) { > > + result = acpi_walk_resources(ACPI_HANDLE(&ancestor->dev), METHOD_NAME__CRS, > > Similarly, if you get a correct pointer to the struct acpi_device, does the above > code need any changes? I'm hoping not. > > > vmbus_walk_resources, NULL); > > > > if (ACPI_FAILURE(result)) > > @@ -2489,10 +2493,21 @@ static int vmbus_acpi_add(struct acpi_device *device) > > > > acpi_walk_err: > > if (ret_val) > > - vmbus_acpi_remove(device); > > + vmbus_mmio_remove(); > > return ret_val; > > } > > > > +static int vmbus_platform_driver_probe(struct platform_device *pdev) > > +{ > > + return vmbus_acpi_add(pdev); > > +} > > + > > +static int vmbus_platform_driver_remove(struct platform_device *pdev) > > +{ > > + vmbus_mmio_remove(); > > + return 0; > > +} > > + > > #ifdef CONFIG_PM_SLEEP > > static int vmbus_bus_suspend(struct device *dev) > > { > > @@ -2658,15 +2673,15 @@ static const struct dev_pm_ops vmbus_bus_pm = { > > .restore_noirq = vmbus_bus_resume > > }; > > > > -static struct acpi_driver vmbus_acpi_driver = { > > - .name = "vmbus", > > - .ids = vmbus_acpi_device_ids, > > - .ops = { > > - .add = vmbus_acpi_add, > > - .remove = vmbus_acpi_remove, > > - }, > > - .drv.pm = &vmbus_bus_pm, > > - .drv.probe_type = PROBE_FORCE_SYNCHRONOUS, > > +static struct platform_driver vmbus_platform_driver = { > > + .probe = vmbus_platform_driver_probe, > > + .remove = vmbus_platform_driver_remove, > > + .driver = { > > + .name = "vmbus", > > + .acpi_match_table = ACPI_PTR(vmbus_acpi_device_ids), > > + .pm = &vmbus_bus_pm, > > + .probe_type = PROBE_FORCE_SYNCHRONOUS, > > + } > > }; > > > > static void hv_kexec_handler(void) > > @@ -2750,12 +2765,11 @@ static int __init hv_acpi_init(void) > > /* > > * Get ACPI resources first. > > */ > > - ret = acpi_bus_register_driver(&vmbus_acpi_driver); > > - > > + ret = platform_driver_register(&vmbus_platform_driver); > > if (ret) > > return ret; > > > > - if (!hv_acpi_dev) { > > + if (!hv_dev) { > > ret = -ENODEV; > > goto cleanup; > > } > > @@ -2785,8 +2799,8 @@ static int __init hv_acpi_init(void) > > return 0; > > > > cleanup: > > - acpi_bus_unregister_driver(&vmbus_acpi_driver); > > - hv_acpi_dev = NULL; > > + platform_driver_unregister(&vmbus_platform_driver); > > + hv_dev = NULL; > > return ret; > > } > > > > @@ -2839,7 +2853,7 @@ static void __exit vmbus_exit(void) > > > > cpuhp_remove_state(hyperv_cpuhp_online); > > hv_synic_free(); > > - acpi_bus_unregister_driver(&vmbus_acpi_driver); > > + platform_driver_unregister(&vmbus_platform_driver); > > } > > > > > > -- > > 2.25.1
From: Saurabh Singh Sengar <ssengar@linux.microsoft.com> Sent: Thursday, February 2, 2023 6:49 AM > > On Wed, Feb 01, 2023 at 05:47:44PM +0000, Michael Kelley (LINUX) wrote: > > From: Saurabh Sengar <ssengar@linux.microsoft.com> Sent: Tuesday, January 31, > 2023 10:10 AM > > > > > > acpi_sleep_state_supported API is only define for CONFIG_ACPI flag and > > > thus it can't be used for non-ACPI builds. Initaly there won't be > > > > s/Initaly/Initially/ > > OK > > > > > > hibernate support for non ACPI builds. > > > > s/hibernate/hibernation/ > > OK > > > > > > > > > This change will help adding device tree support in subsequent commits. > > > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > > --- > > > drivers/hv/hv_common.c | 4 ++++ > > > 1 file changed, 4 insertions(+) > > > > > > diff --git a/drivers/hv/hv_common.c b/drivers/hv/hv_common.c > > > index 52a6f89ccdbd..370ec20d1993 100644 > > > --- a/drivers/hv/hv_common.c > > > +++ b/drivers/hv/hv_common.c > > > @@ -234,7 +234,11 @@ EXPORT_SYMBOL_GPL(hv_setup_dma_ops); > > > > > > bool hv_is_hibernation_supported(void) > > > { > > > +#ifdef CONFIG_ACPI > > > return !hv_root_partition && acpi_sleep_state_supported(ACPI_STATE_S4); > > > +#else > > > + return false; > > > +#endif > > > > Is this patch needed? If CONFIG_ACPI is not set, then per > > arch/x86/Kconfig, CONFIG_ACPI_SYSTEM_POWER_STATES_SUPPORT > > is not selected. In that case, the #ifdef in include/acpi/acpi_bus.h > > provides a stub for acpi_sleep_state_supported() that returns "false". > > So it seems like the existing code should compile and correctly return > > "false" when CONFIG_ACPI is not set. > > You are right, if CONFIG_ACPI_SYSTEM_POWER_STATES_SUPPORT is not set > acpi_sleep_state_supported will return false, but this is applicable only > when CONFIG_ACPI is enable. If CONFIG_ACPI is not enable both these > functions are not defined. > Indeed, you are right. Most of include/acpi/acpi_bus.h is bracketed with #ifdef CONFIG_ACPI, including the stub acpi_sleep_state_supported(). Oh well. :-( Michael
On Wed, Feb 01, 2023 at 06:32:29PM +0000, Michael Kelley (LINUX) wrote: > From: Saurabh Sengar <ssengar@linux.microsoft.com> Sent: Tuesday, January 31, 2023 10:10 AM > > > > Use more generic platform device instead of acpi device. Also rename the > > function vmbus_acpi_remove to more generic name vmbus_mmio_remove. > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > --- > > drivers/hv/vmbus_drv.c | 78 +++++++++++++++++++++++++----------------- > > 1 file changed, 46 insertions(+), 32 deletions(-) > > > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > > index d24dd65b33d4..49030e756b9f 100644 > > --- a/drivers/hv/vmbus_drv.c > > +++ b/drivers/hv/vmbus_drv.c > > @@ -12,6 +12,7 @@ > > #include <linux/init.h> > > #include <linux/module.h> > > #include <linux/device.h> > > +#include <linux/platform_device.h> > > #include <linux/interrupt.h> > > #include <linux/sysctl.h> > > #include <linux/slab.h> > > @@ -44,7 +45,7 @@ struct vmbus_dynid { > > struct hv_vmbus_device_id id; > > }; > > > > -static struct acpi_device *hv_acpi_dev; > > +static struct platform_device *hv_dev; > > > > static int hyperv_cpuhp_online; > > > > @@ -143,7 +144,7 @@ static DEFINE_MUTEX(hyperv_mmio_lock); > > > > static int vmbus_exists(void) > > { > > - if (hv_acpi_dev == NULL) > > + if (hv_dev == NULL) > > return -ENODEV; > > > > return 0; > > @@ -932,7 +933,7 @@ static int vmbus_dma_configure(struct device *child_device) > > * On x86/x64 coherence is assumed and these calls have no effect. > > */ > > hv_setup_dma_ops(child_device, > > - device_get_dma_attr(&hv_acpi_dev->dev) == DEV_DMA_COHERENT); > > + device_get_dma_attr(&hv_dev->dev) == DEV_DMA_COHERENT); > > return 0; > > } > > > > @@ -2090,7 +2091,7 @@ int vmbus_device_register(struct hv_device > > *child_device_obj) > > &child_device_obj->channel->offermsg.offer.if_instance); > > > > child_device_obj->device.bus = &hv_bus; > > - child_device_obj->device.parent = &hv_acpi_dev->dev; > > + child_device_obj->device.parent = &hv_dev->dev; > > child_device_obj->device.release = vmbus_device_release; > > > > child_device_obj->device.dma_parms = &child_device_obj->dma_parms; > > @@ -2262,7 +2263,7 @@ static acpi_status vmbus_walk_resources(struct > > acpi_resource *res, void *ctx) > > return AE_OK; > > } > > > > -static void vmbus_acpi_remove(struct acpi_device *device) > > +static void vmbus_mmio_remove(void) > > { > > struct resource *cur_res; > > struct resource *next_res; > > @@ -2441,13 +2442,15 @@ void vmbus_free_mmio(resource_size_t start, > > resource_size_t size) > > } > > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > > > -static int vmbus_acpi_add(struct acpi_device *device) > > +static int vmbus_acpi_add(struct platform_device *pdev) > > { > > acpi_status result; > > int ret_val = -ENODEV; > > - struct acpi_device *ancestor; > > + struct platform_device *ancestor; > > + struct acpi_device *adev = to_acpi_device(&pdev->dev); > > This doesn't work. The argument to vmbus_acpi_add() is a struct > platform_device, which has a struct device embedded in it (not a > pointer). to_acpi_device() takes a struct device as an argument, > assuming that the struct device is embedded in a struct > acpi_device, which is not the case here. The resulting local > variable adev is actually pointing to some (perhaps negative) > offset within the struct platform_device, and uses of adev are > getting unknown random data from within (or before) the > struct platform_device. Please discard my earlier reply on this. I will fix it in V3. Thanks for pointing this. > > > > > - hv_acpi_dev = device; > > + hv_dev = pdev; > > + adev->fwnode.dev = &pdev->dev; > > > > /* > > * Older versions of Hyper-V for ARM64 fail to include the _CCA > > @@ -2456,15 +2459,16 @@ static int vmbus_acpi_add(struct acpi_device *device) > > * up the ACPI device to behave as if _CCA is present and indicates > > * hardware coherence. > > */ > > - ACPI_COMPANION_SET(&device->dev, device); > > + ACPI_COMPANION_SET(&pdev->dev, ACPI_COMPANION(&pdev->dev)); > > This statement seems tautological. If ACPI_COMPANION(&pdev->dev) > returns a valid result, why would the ACPI companion for &pdev->dev > need to be set? The original code was setting the ACPI companion for the > embedded struct device to be the struct acpi_device. I forget why this > wasn't already done for the VMBus device when it was originally parsed > from the ACPI DSDT ... This need to be changed to: ACPI_COMPANION_SET(&adev_node->dev, adev_node) will fix this as well. > > > if (IS_ENABLED(CONFIG_ACPI_CCA_REQUIRED) && > > - device_get_dma_attr(&device->dev) == DEV_DMA_NOT_SUPPORTED) { > > + device_get_dma_attr(&pdev->dev) == DEV_DMA_NOT_SUPPORTED) { > > + struct acpi_device *adev_node = ACPI_COMPANION(&pdev->dev); > > If earlier code in this function can get a correct pointer to the struct acpi_device, > then this statement shouldn't be necessary. You already have it. agree, will fix > > > pr_info("No ACPI _CCA found; assuming coherent device I/O\n"); > > - device->flags.cca_seen = true; > > - device->flags.coherent_dma = true; > > + adev_node->flags.cca_seen = true; > > + adev_node->flags.coherent_dma = true; > > } > > > > - result = acpi_walk_resources(device->handle, METHOD_NAME__CRS, > > + result = acpi_walk_resources(ACPI_HANDLE(&pdev->dev), METHOD_NAME__CRS, > > Again, if you have a correct pointer to the struct acpi_device, then adev->handle > (like the original code) should be simpler than looking it up again with ACPI_HANDLE(). OK > > > vmbus_walk_resources, NULL); > > > > if (ACPI_FAILURE(result)) > > @@ -2473,9 +2477,9 @@ static int vmbus_acpi_add(struct acpi_device *device) > > * Some ancestor of the vmbus acpi device (Gen1 or Gen2 > > * firmware) is the VMOD that has the mmio ranges. Get that. > > */ > > - for (ancestor = acpi_dev_parent(device); ancestor; > > - ancestor = acpi_dev_parent(ancestor)) { > > - result = acpi_walk_resources(ancestor->handle, METHOD_NAME__CRS, > > + for (ancestor = to_platform_device(pdev->dev.parent); ancestor; > > + ancestor = to_platform_device(ancestor->dev.parent)) { > > + result = acpi_walk_resources(ACPI_HANDLE(&ancestor->dev), METHOD_NAME__CRS, > > Similarly, if you get a correct pointer to the struct acpi_device, does the above > code need any changes? I'm hoping not. Will try to clean this up as well. Regards, Saurabh > > > vmbus_walk_resources, NULL); > > > > if (ACPI_FAILURE(result)) > > @@ -2489,10 +2493,21 @@ static int vmbus_acpi_add(struct acpi_device *device) > > > > acpi_walk_err: > > if (ret_val) > > - vmbus_acpi_remove(device); > > + vmbus_mmio_remove(); > > return ret_val; > > } > > > > +static int vmbus_platform_driver_probe(struct platform_device *pdev) > > +{ > > + return vmbus_acpi_add(pdev); > > +} > > + > > +static int vmbus_platform_driver_remove(struct platform_device *pdev) > > +{ > > + vmbus_mmio_remove(); > > + return 0; > > +} > > + > > #ifdef CONFIG_PM_SLEEP > > static int vmbus_bus_suspend(struct device *dev) > > { > > @@ -2658,15 +2673,15 @@ static const struct dev_pm_ops vmbus_bus_pm = { > > .restore_noirq = vmbus_bus_resume > > }; > > > > -static struct acpi_driver vmbus_acpi_driver = { > > - .name = "vmbus", > > - .ids = vmbus_acpi_device_ids, > > - .ops = { > > - .add = vmbus_acpi_add, > > - .remove = vmbus_acpi_remove, > > - }, > > - .drv.pm = &vmbus_bus_pm, > > - .drv.probe_type = PROBE_FORCE_SYNCHRONOUS, > > +static struct platform_driver vmbus_platform_driver = { > > + .probe = vmbus_platform_driver_probe, > > + .remove = vmbus_platform_driver_remove, > > + .driver = { > > + .name = "vmbus", > > + .acpi_match_table = ACPI_PTR(vmbus_acpi_device_ids), > > + .pm = &vmbus_bus_pm, > > + .probe_type = PROBE_FORCE_SYNCHRONOUS, > > + } > > }; > > > > static void hv_kexec_handler(void) > > @@ -2750,12 +2765,11 @@ static int __init hv_acpi_init(void) > > /* > > * Get ACPI resources first. > > */ > > - ret = acpi_bus_register_driver(&vmbus_acpi_driver); > > - > > + ret = platform_driver_register(&vmbus_platform_driver); > > if (ret) > > return ret; > > > > - if (!hv_acpi_dev) { > > + if (!hv_dev) { > > ret = -ENODEV; > > goto cleanup; > > } > > @@ -2785,8 +2799,8 @@ static int __init hv_acpi_init(void) > > return 0; > > > > cleanup: > > - acpi_bus_unregister_driver(&vmbus_acpi_driver); > > - hv_acpi_dev = NULL; > > + platform_driver_unregister(&vmbus_platform_driver); > > + hv_dev = NULL; > > return ret; > > } > > > > @@ -2839,7 +2853,7 @@ static void __exit vmbus_exit(void) > > > > cpuhp_remove_state(hyperv_cpuhp_online); > > hv_synic_free(); > > - acpi_bus_unregister_driver(&vmbus_acpi_driver); > > + platform_driver_unregister(&vmbus_platform_driver); > > } > > > > > > -- > > 2.25.1
On Wed, Feb 01, 2023 at 11:46:38AM -0600, Rob Herring wrote: > On Wed, Feb 01, 2023 at 08:51:33AM -0800, Saurabh Singh Sengar wrote: > > On Tue, Jan 31, 2023 at 02:12:53PM -0600, Rob Herring wrote: > > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > > <ssengar@linux.microsoft.com> wrote: > > > > > > > > Update the driver to support device tree boot as well along with ACPI. > > > > At present the device tree parsing only provides the mmio region info > > > > and is not the exact copy of ACPI parsing. This is sufficient to cater > > > > all the current device tree usecases for VMBus. > > > > > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > > > --- > > > > drivers/hv/vmbus_drv.c | 75 ++++++++++++++++++++++++++++++++++++++++-- > > > > 1 file changed, 73 insertions(+), 2 deletions(-) > > > > > > > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > > > > index 49030e756b9f..1741f1348f9f 100644 > > > > --- a/drivers/hv/vmbus_drv.c > > > > +++ b/drivers/hv/vmbus_drv.c > > > > @@ -2152,7 +2152,7 @@ void vmbus_device_unregister(struct hv_device *device_obj) > > > > device_unregister(&device_obj->device); > > > > } (...) > > > > struct pci_dev *pdev; > > > > @@ -2442,6 +2443,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size) > > > > } > > > > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > > > > > > > +#ifdef CONFIG_ACPI > > > > > > It's better to put C 'if (!IS_ENABLED(CONFIG_ACPI)' code in the > > > > I wanted to have separate function for ACPI and device tree flow, which > > can be easily maintained with #ifdef. Please let me know if its fine. > > Yes, you can have separate functions: > > static int vmbus_acpi_add(struct platform_device *pdev) > { > if (!IS_ENABLED(CONFIG_ACPI)) > return -ENODEV; > > ... > } > > The compiler will throw away the function in the end if CONFIG_ACPI is > not enabled. > > That is easier for us to maintain because it reduces the combinations to > build. > I tried removing #ifdef CONFIG_ACPI and use C's if(!IS_ENABLED(CONFIG_ACPI)) but looks compiler is not optimizing out the rest of function, it still throwing errors for acpi functions. This doesn't look 1:1 replacement to me. Please let me know if I have missunderstood any of your suggestion. drivers/hv/vmbus_drv.c:2175:8: error: implicit declaration of function ‘acpi_dev_resource_interrupt’ [-Werror=implicit-function- > > > > > > > > > static int vmbus_acpi_add(struct platform_device *pdev) > > > > { > > > > acpi_status result; > > > > @@ -2496,10 +2498,68 @@ static int vmbus_acpi_add(struct platform_device *pdev) > > [1] https://lwn.net/Articles/443531/
On Wed, Feb 01, 2023 at 06:15:23PM +0100, Krzysztof Kozlowski wrote: > On 01/02/2023 17:34, Saurabh Singh Sengar wrote: > >> Also see my comment on v1 about running DT validation on your dtb. I'm > >> sure running it would point out other issues. Such as the root level > >> comaptible string(s) need to be documented. You need cpu nodes, > >> interrupt controller, timers, etc. Those all have to be documented. > > > > I will be changing the parent node to soc node as suggested by Krzysztof > > in other thread. > > > > soc { > > #address-cells = <2>; > > #size-cells = <2>; > > > > vmbus@ff0000000 { > > #address-cells = <2>; > > #size-cells = <1>; > > compatible = "Microsoft,vmbus"; > > ranges = <0x00 0x00 0x0f 0xf0000000 0x10000000>; > > }; > > }; > > > > This will be sufficient. > > It will be ok for the example, but will not be ok for supporting your > use case. Please solve all the points from Rob's comment above. Where is > their documentation? > > Best regards, > Krzysztof Hi Rob/ Krzysztof, I am happy to update the documentation as requested. Please note that, apart from CPUs, there is no other device node in the tree. Here are some of the info related to our system: Timers: VMBus code uses a Hyper-V Synthetic timer and there is no device tree node or ACPI method required for this. This is implemented as drivers/clocksource/hyperv_timer.c Interrupt controller: The hypervisor virtualizes interrupt delivery to virtual processors. This is done through the use of a synthetic interrupt controller (SynIC) which is an extension of a virtualized local APIC. In the cpu DT nodes we have APIC ids. Below are the cpu nodes we use, please suggest if I need to update any document for it. cpus { #address-cells = <1>; #size-cells = <0>; cpu@0 { device_type = "cpu"; reg = <0>; status = "okay"; }; cpu@1 { device_type = "cpu"; reg = <1>; status = "okay"; }; }; Regards, Saurabh
On Fri, Feb 3, 2023 at 11:36 AM Saurabh Singh Sengar <ssengar@linux.microsoft.com> wrote: > > On Wed, Feb 01, 2023 at 11:46:38AM -0600, Rob Herring wrote: > > On Wed, Feb 01, 2023 at 08:51:33AM -0800, Saurabh Singh Sengar wrote: > > > On Tue, Jan 31, 2023 at 02:12:53PM -0600, Rob Herring wrote: > > > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > > > <ssengar@linux.microsoft.com> wrote: > > > > > > > > > > Update the driver to support device tree boot as well along with ACPI. > > > > > At present the device tree parsing only provides the mmio region info > > > > > and is not the exact copy of ACPI parsing. This is sufficient to cater > > > > > all the current device tree usecases for VMBus. > > > > > > > > > > Signed-off-by: Saurabh Sengar <ssengar@linux.microsoft.com> > > > > > --- > > > > > drivers/hv/vmbus_drv.c | 75 ++++++++++++++++++++++++++++++++++++++++-- > > > > > 1 file changed, 73 insertions(+), 2 deletions(-) > > > > > > > > > > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c > > > > > index 49030e756b9f..1741f1348f9f 100644 > > > > > --- a/drivers/hv/vmbus_drv.c > > > > > +++ b/drivers/hv/vmbus_drv.c > > > > > @@ -2152,7 +2152,7 @@ void vmbus_device_unregister(struct hv_device *device_obj) > > > > > device_unregister(&device_obj->device); > > > > > } > (...) > > > > > struct pci_dev *pdev; > > > > > @@ -2442,6 +2443,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size) > > > > > } > > > > > EXPORT_SYMBOL_GPL(vmbus_free_mmio); > > > > > > > > > > +#ifdef CONFIG_ACPI > > > > > > > > It's better to put C 'if (!IS_ENABLED(CONFIG_ACPI)' code in the > > > > > > I wanted to have separate function for ACPI and device tree flow, which > > > can be easily maintained with #ifdef. Please let me know if its fine. > > > > Yes, you can have separate functions: > > > > static int vmbus_acpi_add(struct platform_device *pdev) > > { > > if (!IS_ENABLED(CONFIG_ACPI)) > > return -ENODEV; > > > > ... > > } > > > > The compiler will throw away the function in the end if CONFIG_ACPI is > > not enabled. > > > > That is easier for us to maintain because it reduces the combinations to > > build. > > > > I tried removing #ifdef CONFIG_ACPI and use C's if(!IS_ENABLED(CONFIG_ACPI)) but looks > compiler is not optimizing out the rest of function, it still throwing errors > for acpi functions. This doesn't look 1:1 replacement to me. > Please let me know if I have missunderstood any of your suggestion. > > drivers/hv/vmbus_drv.c:2175:8: error: implicit declaration of function ‘acpi_dev_resource_interrupt’ [-Werror=implicit-function- That's a failure of the ACPI headers not having empty function declarations. The DT functions do... Also, this is just a broken assumption: #ifdef CONFIG_ACPI #else // Assume DT #endif Both ACPI and DT can be enabled at the same time. They may be mutually exclusive for a platform, but not the kernel. For distro kernels, both will be enabled typically if the arch supports both. On arm64, DT is never disabled because the boot interface is always DT. Furthermore, this makes compile testing your code difficult. The arm64 defconfig, allmodconfig and allyesconfig all will not build the DT code. The same for x86. This means all the CI builds that happen can't build test this. Rob
On Wed, Feb 1, 2023 at 10:34 AM Saurabh Singh Sengar <ssengar@linux.microsoft.com> wrote: > > On Wed, Feb 01, 2023 at 08:51:46AM -0600, Rob Herring wrote: > > On Tue, Jan 31, 2023 at 06:04:49PM -0800, Saurabh Singh Sengar wrote: > > > On Tue, Jan 31, 2023 at 02:27:51PM -0600, Rob Herring wrote: > > > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > > > <ssengar@linux.microsoft.com> wrote: > > > > > > > > > > This set of patches expands the VMBus driver to include device tree > > > > > support. > > > > > > > > > > The first two patches enable compilation of Hyper-V APIs in a non-ACPI > > > > > build. > > > > > > > > > > The third patch converts the VMBus driver from acpi to more generic > > > > > platform driver. > > > > > > > > > > Further to add device tree documentation for VMBus, it needs to club with > > > > > other virtualization driver's documentation. For this rename the virtio > > > > > folder to more generic hypervisor, so that all the hypervisor based > > > > > devices can co-exist in a single place in device tree documentation. The > > > > > fourth patch does this renaming. > > > > > > > > > > The fifth patch introduces the device tree documentation for VMBus. > > > > > > > > > > The sixth patch adds device tree support to the VMBus driver. Currently > > > > > this is tested only for x86 and it may not work for other archs. > > > > > > > > I can read all the patches and see *what* they do. You don't really > > > > need to list that here. I'm still wondering *why*. That is what the > > > > cover letter and commit messages should answer. Why do you need DT > > > > support? How does this even work on x86? FDT is only enabled for > > > > CE4100 platform. > > > > > > HI Rob, > > > > > > Thanks for your comments. > > > We are working on a solution where kernel is booted without ACPI tables to keep > > > the overall system's memory footprints slim and possibly faster boot time. > > > We have tested this by enabling CONFIG_OF for x86. > > > > It's CONFIG_OF_EARLY_FLATTREE which you would need and that's not user > > selectable. At a minimum, you need some kconfig changes. Where are > > those? > > You are right we have define a new config flag in Kconfig, and selected CONFIG_OF > and CONFIG_OF_EARLY_FLATTREE. We are working on upstreaming that patch as well > however that will be a separate patch series. Fair enough, but that should come first IMO. Really I just want to see a complete picture. That can be a reference to a git branch(es) or other patch series. But again, what I want to see in particular is the actual DT and validation run on it. > > Also see my comment on v1 about running DT validation on your dtb. I'm > > sure running it would point out other issues. Such as the root level > > comaptible string(s) need to be documented. You need cpu nodes, > > interrupt controller, timers, etc. Those all have to be documented. > > I will be changing the parent node to soc node as suggested by Krzysztof > in other thread. Another issue yes, but orthogonal to my comments. > > soc { > #address-cells = <2>; > #size-cells = <2>; You are missing 'ranges' here. Without it, addresses aren't translatable. You are also missing 'compatible = "simple-bus";'. This happens to work on x86 because of legacy reasons, but we don't want new cases added. > > vmbus@ff0000000 { > #address-cells = <2>; > #size-cells = <1>; > compatible = "Microsoft,vmbus"; 'Microsoft' is not a vendor prefix. > ranges = <0x00 0x00 0x0f 0xf0000000 0x10000000>; > }; > }; > > This will be sufficient. All these comments are unnecessary because the tools will now check these things and we shouldn't have to. Rob
On Tue, Jan 31, 2023 at 06:04:49PM -0800, Saurabh Singh Sengar wrote: > > We are working on a solution where kernel is booted without ACPI tables to keep > the overall system's memory footprints slim and possibly faster boot time. > We have tested this by enabling CONFIG_OF for x86. > Very interesting. Do you comparison on how slow/fast DT version w.r.t ACPI and similarly for the memory footprint ? It would be good to add those information as well. I know some systems just use ACPI static tables as they don't run ACPICA runtime interpretter, was that experimented as well or it wasn't used due to some specific(details please) limitations without it. > I can add this info in cover letter in next version. Yes please.
On Tue, Feb 07, 2023 at 11:38:55AM -0600, Rob Herring wrote: > On Fri, Feb 3, 2023 at 11:36 AM Saurabh Singh Sengar > <ssengar@linux.microsoft.com> wrote: > > > > On Wed, Feb 01, 2023 at 11:46:38AM -0600, Rob Herring wrote: > > > On Wed, Feb 01, 2023 at 08:51:33AM -0800, Saurabh Singh Sengar wrote: > > > > On Tue, Jan 31, 2023 at 02:12:53PM -0600, Rob Herring wrote: > > > > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > > > > <ssengar@linux.microsoft.com> wrote: > > > > I wanted to have separate function for ACPI and device tree flow, which (...) > > > > can be easily maintained with #ifdef. Please let me know if its fine. > > > > > > Yes, you can have separate functions: > > > > > > static int vmbus_acpi_add(struct platform_device *pdev) > > > { > > > if (!IS_ENABLED(CONFIG_ACPI)) > > > return -ENODEV; > > > > > > ... > > > } > > > > > > The compiler will throw away the function in the end if CONFIG_ACPI is > > > not enabled. > > > > > > That is easier for us to maintain because it reduces the combinations to > > > build. > > > > > > > I tried removing #ifdef CONFIG_ACPI and use C's if(!IS_ENABLED(CONFIG_ACPI)) but looks > > compiler is not optimizing out the rest of function, it still throwing errors > > for acpi functions. This doesn't look 1:1 replacement to me. > > Please let me know if I have missunderstood any of your suggestion. > > > > drivers/hv/vmbus_drv.c:2175:8: error: implicit declaration of function ‘acpi_dev_resource_interrupt’ [-Werror=implicit-function- > > That's a failure of the ACPI headers not having empty function > declarations. The DT functions do... > > Also, this is just a broken assumption: > > #ifdef CONFIG_ACPI > > #else > // Assume DT > #endif > > Both ACPI and DT can be enabled at the same time. They may be mutually > exclusive for a platform, but not the kernel. For distro kernels, both > will be enabled typically if the arch supports both. On arm64, DT is > never disabled because the boot interface is always DT. > > Furthermore, this makes compile testing your code difficult. The arm64 > defconfig, allmodconfig and allyesconfig all will not build the DT > code. The same for x86. This means all the CI builds that happen can't > build test this. Thanks for letting me know the challanges with testing. My intention was to give ACPI higher priority, in case ACPI is enabled system should go ACPI flow, OF flow should be used only when ACPI is disabled. I can replace #else part with #ifdef CONFIG_OF if that helps. Regards, Saurabh > > Rob
On Tue, Feb 07, 2023 at 11:53:54AM -0600, Rob Herring wrote: > On Wed, Feb 1, 2023 at 10:34 AM Saurabh Singh Sengar > <ssengar@linux.microsoft.com> wrote: > > > > On Wed, Feb 01, 2023 at 08:51:46AM -0600, Rob Herring wrote: > > > On Tue, Jan 31, 2023 at 06:04:49PM -0800, Saurabh Singh Sengar wrote: > > > > On Tue, Jan 31, 2023 at 02:27:51PM -0600, Rob Herring wrote: > > > > > On Tue, Jan 31, 2023 at 12:10 PM Saurabh Sengar > > > > > <ssengar@linux.microsoft.com> wrote: > > > > > > > > > > > > This set of patches expands the VMBus driver to include device tree > > > > > > support. (...) > > You are right we have define a new config flag in Kconfig, and selected CONFIG_OF > > and CONFIG_OF_EARLY_FLATTREE. We are working on upstreaming that patch as well > > however that will be a separate patch series. > > Fair enough, but that should come first IMO. Really I just want to see > a complete picture. That can be a reference to a git branch(es) or > other patch series. But again, what I want to see in particular is the > actual DT and validation run on it. Thank you for explaining the concern. I now understand it fully. I have come to the realization that enabling the vmbus device tree should not be impacted by any changes. To address this, I will add the following lines to the HYPERV Kconfig definition I used for testing: select OF if !ACPI select OF_EARLY_FLATTREE if !ACPI" > > > > Also see my comment on v1 about running DT validation on your dtb. I'm > > > sure running it would point out other issues. Such as the root level > > > comaptible string(s) need to be documented. You need cpu nodes, > > > interrupt controller, timers, etc. Those all have to be documented. > > > > I will be changing the parent node to soc node as suggested by Krzysztof > > in other thread. > > Another issue yes, but orthogonal to my comments. > > > > > soc { > > #address-cells = <2>; > > #size-cells = <2>; > > You are missing 'ranges' here. Without it, addresses aren't translatable. > > You are also missing 'compatible = "simple-bus";'. This happens to > work on x86 because of legacy reasons, but we don't want new cases > added. I am a bit unclear on the reason for adding the ranges property in the root node. To provide more context, I have included my full device tree below. I believe that having the ranges property in the VMBus device node is sufficient. Please let me know if this can be improved. /dts-v1/; / { #address-cells = <0>; #size-cells = <0>; model = "microsoft,test"; cpus { #address-cells = <0x01>; #size-cells = <0x00>; cpu@0 { device_type = "cpu"; reg = <0x00>; status = "okay"; }; cpu@1 { device_type = "cpu"; reg = <0x01>; status = "okay"; }; }; vmbus@ff0000000 { #address-cells = <2>; #size-cells = <2>; compatible = "microsoft,vmbus"; ranges = <0x0f 0xf0000000 0x0f 0xf0000000 0x0 0x10000000>; }; }; > > > > > vmbus@ff0000000 { > > #address-cells = <2>; > > #size-cells = <1>; > > compatible = "Microsoft,vmbus"; > > 'Microsoft' is not a vendor prefix. > > > ranges = <0x00 0x00 0x0f 0xf0000000 0x10000000>; > > }; > > }; > > > > This will be sufficient. > > All these comments are unnecessary because the tools will now check > these things and we shouldn't have to. Agree, and its fixed in latest version. Regards, Saurabh > > Rob