This set of patches expands the VMBus driver to include device tree
support. This feature allows for a kernel boot without the use of ACPI
tables, resulting in a smaller memory footprint and potentially faster
boot times. This is tested by enabling CONFIG_FLAT and OF_EARLY_FLATTREE
for x86.
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.
The fourth patch introduces the device tree documentation for VMBus.
The fifth patch adds device tree support to the VMBus driver. Currently
this is tested only for x86 and it may not work for other archs.
[V8]
- Remove the auto select of CONFIG_OF
- Remove the dependency on !ACPI for OF_EARLY_FLATTREE
- Used acpi_disabled instead of #ifdef and hence added a dummy function
for vmbus_acpi_add
- GFP_ATOMIC -> GFP_KERNEL
- used range.flags instead of hard coding flags
- used __maybe_unused for acpi device id, removed #ifdef
[V7]
- Use cpu_addr instead of bus_addr.
- Updated Documentation accordingly.
[V6]
- define acpi_sleep_state_supported in acpi header for !ACPI,
dropped the changes done in hv_common.c under #ifdef CONFIG_ACPI
- "Devicetree" instead of "device tree"
- Remove initialize of ret
- set "np = pdev->dev.of_node;" on declarartion
- remove one error print
- use bus_addr instead of pci_addr
[V5]
- Removed #else for device tree parsing code. This should help better
test coverage.
- Fix macro '__maybe_unused' warning
- Added below options in Kconfig to enable device tree options for HYPERV
select OF if !ACPI
select OF_EARLY_FLATTREE if !ACPI
- moved dt documantation to bus folder
- update the dt node to have 'bus' as parent node instead of 'soc'. Also
added compatible and ranges property for parent node.
- Made sure dt_binding_check have no error/varnings for microsoft,vmbus.yaml file
- Fix commit messages and add Reviwed-by from Michael for first 3 patches
[V4]
- rebased which fixed return type of 'vmbus_mmio_remove' from int to void
- used __maybe_unused for 'vmbus_of_match' and safeguard vmbus_acpi_device_ids
under #ifdef
[V3]
- Changed the logic to use generic api (for_each_of_range) for parsing "ranges".
- Remove dependency of ACPI for HYPERV in case of x86.
- Removed "device tree bindings" from title and patch subject.
- Removed duplicate vendor prefix, used microsoft instead of msft.
- Use 'soc' in example of device tree documantation for parent node.
- Fixed compatible schemas error generated in other modules referring to
virtio.
- Drop hex notation and leading zeros from device tree cell properties.
- Added missing full stop at the end of commit message.
- Typos fix: s/Initaly/Initially/ and s/hibernate/hibernation/.
- Replace to_acpi_device with ACPI_COMPANION which simplify the logic.
- Added more info in cover letter aboutsystem under test.
[v2]
- Convert VMBus acpi device to platform device, and added device tree support
in separate patch. This enables using same driver structure for both the flows.
- In Device tree documentation, changed virtio folder to hypervisor and moved
VMBus documentation there.
- Moved bindings before Device tree patch.
- Removed stale ".data" and ".name" field from of_device match table.
- Removed debug print.
Saurabh Sengar (5):
drivers/clocksource/hyper-v: non ACPI support in hyperv clock
ACPI: bus: Add stub acpi_sleep_state_supported() in non-ACPI cases
Drivers: hv: vmbus: Convert acpi_device to more generic
platform_device
dt-bindings: bus: VMBus
Driver: VMBus: Add Devicetree support
.../bindings/bus/microsoft,vmbus.yaml | 54 ++++++++
MAINTAINERS | 1 +
drivers/clocksource/hyperv_timer.c | 15 ++-
drivers/hv/Kconfig | 5 +-
drivers/hv/vmbus_drv.c | 120 ++++++++++++++----
include/linux/acpi.h | 5 +
6 files changed, 170 insertions(+), 30 deletions(-)
create mode 100644 Documentation/devicetree/bindings/bus/microsoft,vmbus.yaml
--
2.34.1
acpi_sleep_state_supported() is defined only when CONFIG_ACPI=y. The
function is in acpi_bus.h, and acpi_bus.h can only be used in
CONFIG_ACPI=y cases. Add the stub function to linux/acpi.h to make
compilation successful for !CONFIG_ACPI cases.
Signed-off-by: Saurabh Sengar <[email protected]>
Acked-by: Rafael J. Wysocki <[email protected]>
---
include/linux/acpi.h | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index efff750f326d..d331f76b0c19 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -1075,6 +1075,11 @@ static inline u32 acpi_osc_ctx_get_cxl_control(struct acpi_osc_context *context)
return 0;
}
+static inline bool acpi_sleep_state_supported(u8 sleep_state)
+{
+ return false;
+}
+
#endif /* !CONFIG_ACPI */
#ifdef CONFIG_ACPI_HOTPLUG_IOAPIC
--
2.34.1
Add a placeholder function for the hv_setup_stimer0_irq API to accommodate
systems without ACPI support. Since this function is not utilized on
x86/x64 systems and non-ACPI support is only intended for x86/x64 systems,
a placeholder function is sufficient for now and can be improved upon if
necessary in the future.
Signed-off-by: Saurabh Sengar <[email protected]>
Reviewed-by: Michael Kelley <[email protected]>
---
drivers/clocksource/hyperv_timer.c | 15 +++++++++++++--
1 file changed, 13 insertions(+), 2 deletions(-)
diff --git a/drivers/clocksource/hyperv_timer.c b/drivers/clocksource/hyperv_timer.c
index c0cef92b12b8..f32948c8a96f 100644
--- a/drivers/clocksource/hyperv_timer.c
+++ b/drivers/clocksource/hyperv_timer.c
@@ -49,7 +49,7 @@ static bool direct_mode_enabled;
static int stimer0_irq = -1;
static int stimer0_message_sint;
-static DEFINE_PER_CPU(long, stimer0_evt);
+static __maybe_unused DEFINE_PER_CPU(long, stimer0_evt);
/*
* Common code for stimer0 interrupts coming via Direct Mode or
@@ -68,7 +68,7 @@ EXPORT_SYMBOL_GPL(hv_stimer0_isr);
* stimer0 interrupt handler for architectures that support
* per-cpu interrupts, which also implies Direct Mode.
*/
-static irqreturn_t hv_stimer0_percpu_isr(int irq, void *dev_id)
+static irqreturn_t __maybe_unused hv_stimer0_percpu_isr(int irq, void *dev_id)
{
hv_stimer0_isr();
return IRQ_HANDLED;
@@ -196,6 +196,7 @@ void __weak hv_remove_stimer0_handler(void)
{
};
+#ifdef CONFIG_ACPI
/* Called only on architectures with per-cpu IRQs (i.e., not x86/x64) */
static int hv_setup_stimer0_irq(void)
{
@@ -230,6 +231,16 @@ static void hv_remove_stimer0_irq(void)
stimer0_irq = -1;
}
}
+#else
+static int hv_setup_stimer0_irq(void)
+{
+ return 0;
+}
+
+static void hv_remove_stimer0_irq(void)
+{
+}
+#endif
/* hv_stimer_alloc - Global initialization of the clockevent and stimer0 */
int hv_stimer_alloc(bool have_percpu_irqs)
--
2.34.1
Add dt-bindings for Hyper-V VMBus.
Signed-off-by: Saurabh Sengar <[email protected]>
Reviewed-by: Rob Herring <[email protected]>
---
.../bindings/bus/microsoft,vmbus.yaml | 54 +++++++++++++++++++
MAINTAINERS | 1 +
2 files changed, 55 insertions(+)
create mode 100644 Documentation/devicetree/bindings/bus/microsoft,vmbus.yaml
diff --git a/Documentation/devicetree/bindings/bus/microsoft,vmbus.yaml b/Documentation/devicetree/bindings/bus/microsoft,vmbus.yaml
new file mode 100644
index 000000000000..a8d40c766dcd
--- /dev/null
+++ b/Documentation/devicetree/bindings/bus/microsoft,vmbus.yaml
@@ -0,0 +1,54 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/bus/microsoft,vmbus.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Microsoft Hyper-V VMBus
+
+maintainers:
+ - Saurabh Sengar <[email protected]>
+
+description:
+ VMBus is a software bus that implement the protocols for communication
+ between the root or host OS and guest OSs (virtual machines).
+
+properties:
+ compatible:
+ const: microsoft,vmbus
+
+ ranges: true
+
+ '#address-cells':
+ const: 2
+
+ '#size-cells':
+ const: 1
+
+required:
+ - compatible
+ - ranges
+ - '#address-cells'
+ - '#size-cells'
+
+additionalProperties: false
+
+examples:
+ - |
+ soc {
+ #address-cells = <2>;
+ #size-cells = <1>;
+ bus {
+ compatible = "simple-bus";
+ #address-cells = <2>;
+ #size-cells = <1>;
+ ranges;
+
+ vmbus@ff0000000 {
+ compatible = "microsoft,vmbus";
+ #address-cells = <2>;
+ #size-cells = <1>;
+ ranges = <0x0f 0xf0000000 0x0f 0xf0000000 0x10000000>;
+ };
+ };
+ };
diff --git a/MAINTAINERS b/MAINTAINERS
index 25a0981c74b6..c4f4c8cd9276 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -9544,6 +9544,7 @@ S: Supported
T: git git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux.git
F: Documentation/ABI/stable/sysfs-bus-vmbus
F: Documentation/ABI/testing/debugfs-hyperv
+F: Documentation/devicetree/bindings/bus/microsoft,vmbus.yaml
F: Documentation/virt/hyperv
F: Documentation/networking/device_drivers/ethernet/microsoft/netvsc.rst
F: arch/arm64/hyperv
--
2.34.1
VMBus driver code currently has direct dependency on ACPI and struct
acpi_device. As a staging step toward optionally configuring based on
Devicetree instead of ACPI, use a more generic platform device to reduce
the dependency on ACPI where possible, though the dependency on ACPI
is not completely removed. Also rename the function vmbus_acpi_remove()
to the more generic vmbus_mmio_remove().
Signed-off-by: Saurabh Sengar <[email protected]>
Reviewed-by: Michael Kelley <[email protected]>
---
drivers/hv/vmbus_drv.c | 58 +++++++++++++++++++++++++-----------------
1 file changed, 35 insertions(+), 23 deletions(-)
diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
index 229353f1e9c2..3ad2fa2b92e7 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 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_DMA_COHERENT);
return 0;
}
@@ -2081,7 +2082,7 @@ 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;
child_device_obj->device.release = vmbus_device_release;
child_device_obj->device.dma_parms = &child_device_obj->dma_parms;
@@ -2253,7 +2254,7 @@ 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;
@@ -2432,13 +2433,14 @@ 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 acpi_device *device = ACPI_COMPANION(&pdev->dev);
- hv_acpi_dev = device;
+ hv_dev = &device->dev;
/*
* Older versions of Hyper-V for ARM64 fail to include the _CCA
@@ -2480,10 +2482,21 @@ 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)
{
@@ -2649,15 +2662,15 @@ 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)
@@ -2741,12 +2754,11 @@ 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;
}
@@ -2776,8 +2788,8 @@ 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;
}
@@ -2830,7 +2842,7 @@ 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);
}
--
2.34.1
Update the driver to support Devicetree boot as well along with ACPI.
At present the Devicetree parsing only provides the mmio region info
and is not the exact copy of ACPI parsing. This is sufficient to cater
all the current Devicetree usecases for VMBus.
Currently Devicetree is supported only for x86 systems.
Signed-off-by: Saurabh Sengar <[email protected]>
---
[V8]
- Remove the auto select of CONFIG_OF
- Remove the dependency on !ACPI for OF_EARLY_FLATTREE
- Used acpi_disabled instead of #ifdef and hence added a dummy function
for vmbus_acpi_add
- GFP_ATOMIC -> GFP_KERNEL
- used range.flags instead of hard coding flags
- used __maybe_unused for acpi device id, removed #ifdef
drivers/hv/Kconfig | 5 ++--
drivers/hv/vmbus_drv.c | 64 +++++++++++++++++++++++++++++++++++++++---
2 files changed, 63 insertions(+), 6 deletions(-)
diff --git a/drivers/hv/Kconfig b/drivers/hv/Kconfig
index 0747a8f1fcee..47132b30b7ee 100644
--- a/drivers/hv/Kconfig
+++ b/drivers/hv/Kconfig
@@ -4,11 +4,12 @@ menu "Microsoft Hyper-V guest support"
config HYPERV
tristate "Microsoft Hyper-V client drivers"
- depends on ACPI && ((X86 && X86_LOCAL_APIC && HYPERVISOR_GUEST) \
- || (ARM64 && !CPU_BIG_ENDIAN))
+ depends on (X86 && X86_LOCAL_APIC && HYPERVISOR_GUEST) \
+ || (ACPI && ARM64 && !CPU_BIG_ENDIAN)
select PARAVIRT
select X86_HV_CALLBACK_VECTOR if X86
select VMAP_PFN
+ select OF_EARLY_FLATTREE if OF
help
Select this option to run Linux as a Hyper-V client operating
system.
diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
index 3ad2fa2b92e7..15097e1f3f2b 100644
--- a/drivers/hv/vmbus_drv.c
+++ b/drivers/hv/vmbus_drv.c
@@ -20,6 +20,7 @@
#include <linux/completion.h>
#include <linux/hyperv.h>
#include <linux/kernel_stat.h>
+#include <linux/of_address.h>
#include <linux/clockchips.h>
#include <linux/cpu.h>
#include <linux/sched/isolation.h>
@@ -2143,7 +2144,7 @@ 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.
@@ -2253,6 +2254,7 @@ static acpi_status vmbus_walk_resources(struct acpi_resource *res, void *ctx)
return AE_OK;
}
+#endif
static void vmbus_mmio_remove(void)
{
@@ -2273,7 +2275,7 @@ 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;
@@ -2433,6 +2435,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size)
}
EXPORT_SYMBOL_GPL(vmbus_free_mmio);
+#ifdef CONFIG_ACPI
static int vmbus_acpi_add(struct platform_device *pdev)
{
acpi_status result;
@@ -2485,10 +2488,52 @@ static int vmbus_acpi_add(struct platform_device *pdev)
vmbus_mmio_remove();
return ret_val;
}
+#else
+static int vmbus_acpi_add(struct platform_device *pdev)
+{
+ return 0;
+}
+#endif
+
+static int vmbus_device_add(struct platform_device *pdev)
+{
+ struct resource **cur_res = &hyperv_mmio;
+ struct of_range range;
+ struct of_range_parser parser;
+ struct device_node *np = pdev->dev.of_node;
+ int ret;
+
+ hv_dev = &pdev->dev;
+
+ ret = of_range_parser_init(&parser, np);
+ if (ret)
+ return ret;
+
+ for_each_of_range(&parser, &range) {
+ struct resource *res;
+
+ res = kzalloc(sizeof(*res), GFP_KERNEL);
+ if (!res)
+ return -ENOMEM;
+
+ res->name = "hyperv mmio";
+ res->flags = range.flags;
+ res->start = range.cpu_addr;
+ res->end = range.cpu_addr + range.size;
+
+ *cur_res = res;
+ cur_res = &res->sibling;
+ }
+
+ return ret;
+}
static int vmbus_platform_driver_probe(struct platform_device *pdev)
{
- return vmbus_acpi_add(pdev);
+ if (!acpi_disabled)
+ return vmbus_acpi_add(pdev);
+ else
+ return vmbus_device_add(pdev);
}
static int vmbus_platform_driver_remove(struct platform_device *pdev)
@@ -2634,7 +2679,17 @@ static int vmbus_bus_resume(struct device *dev)
#define vmbus_bus_resume NULL
#endif /* CONFIG_PM_SLEEP */
-static const struct acpi_device_id vmbus_acpi_device_ids[] = {
+static const __maybe_unused struct of_device_id vmbus_of_match[] = {
+ {
+ .compatible = "microsoft,vmbus",
+ },
+ {
+ /* sentinel */
+ },
+};
+MODULE_DEVICE_TABLE(of, vmbus_of_match);
+
+static const __maybe_unused struct acpi_device_id vmbus_acpi_device_ids[] = {
{"VMBUS", 0},
{"VMBus", 0},
{"", 0},
@@ -2668,6 +2723,7 @@ 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,
}
--
2.34.1
From: Saurabh Sengar <[email protected]> Sent: Tuesday, March 14, 2023 2:16 AM
>
> Update the driver to support Devicetree boot as well along with ACPI.
> At present the Devicetree parsing only provides the mmio region info
> and is not the exact copy of ACPI parsing. This is sufficient to cater
> all the current Devicetree usecases for VMBus.
>
> Currently Devicetree is supported only for x86 systems.
>
> Signed-off-by: Saurabh Sengar <[email protected]>
> ---
> [V8]
> - Remove the auto select of CONFIG_OF
> - Remove the dependency on !ACPI for OF_EARLY_FLATTREE
> - Used acpi_disabled instead of #ifdef and hence added a dummy function
> for vmbus_acpi_add
> - GFP_ATOMIC -> GFP_KERNEL
> - used range.flags instead of hard coding flags
> - used __maybe_unused for acpi device id, removed #ifdef
>
> drivers/hv/Kconfig | 5 ++--
> drivers/hv/vmbus_drv.c | 64 +++++++++++++++++++++++++++++++++++++++---
> 2 files changed, 63 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/hv/Kconfig b/drivers/hv/Kconfig
> index 0747a8f1fcee..47132b30b7ee 100644
> --- a/drivers/hv/Kconfig
> +++ b/drivers/hv/Kconfig
> @@ -4,11 +4,12 @@ menu "Microsoft Hyper-V guest support"
>
> config HYPERV
> tristate "Microsoft Hyper-V client drivers"
> - depends on ACPI && ((X86 && X86_LOCAL_APIC && HYPERVISOR_GUEST) \
> - || (ARM64 && !CPU_BIG_ENDIAN))
> + depends on (X86 && X86_LOCAL_APIC && HYPERVISOR_GUEST) \
> + || (ACPI && ARM64 && !CPU_BIG_ENDIAN)
> select PARAVIRT
> select X86_HV_CALLBACK_VECTOR if X86
> select VMAP_PFN
> + select OF_EARLY_FLATTREE if OF
> help
> Select this option to run Linux as a Hyper-V client operating
> system.
> diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
> index 3ad2fa2b92e7..15097e1f3f2b 100644
> --- a/drivers/hv/vmbus_drv.c
> +++ b/drivers/hv/vmbus_drv.c
> @@ -20,6 +20,7 @@
> #include <linux/completion.h>
> #include <linux/hyperv.h>
> #include <linux/kernel_stat.h>
> +#include <linux/of_address.h>
> #include <linux/clockchips.h>
> #include <linux/cpu.h>
> #include <linux/sched/isolation.h>
> @@ -2143,7 +2144,7 @@ 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.
> @@ -2253,6 +2254,7 @@ static acpi_status vmbus_walk_resources(struct acpi_resource
> *res, void *ctx)
>
> return AE_OK;
> }
> +#endif
>
> static void vmbus_mmio_remove(void)
> {
> @@ -2273,7 +2275,7 @@ 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;
> @@ -2433,6 +2435,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t
> size)
> }
> EXPORT_SYMBOL_GPL(vmbus_free_mmio);
>
> +#ifdef CONFIG_ACPI
> static int vmbus_acpi_add(struct platform_device *pdev)
> {
> acpi_status result;
> @@ -2485,10 +2488,52 @@ static int vmbus_acpi_add(struct platform_device *pdev)
> vmbus_mmio_remove();
> return ret_val;
> }
> +#else
> +static int vmbus_acpi_add(struct platform_device *pdev)
> +{
> + return 0;
> +}
> +#endif
> +
> +static int vmbus_device_add(struct platform_device *pdev)
> +{
> + struct resource **cur_res = &hyperv_mmio;
> + struct of_range range;
> + struct of_range_parser parser;
> + struct device_node *np = pdev->dev.of_node;
> + int ret;
> +
> + hv_dev = &pdev->dev;
> +
> + ret = of_range_parser_init(&parser, np);
> + if (ret)
> + return ret;
> +
> + for_each_of_range(&parser, &range) {
> + struct resource *res;
> +
> + res = kzalloc(sizeof(*res), GFP_KERNEL);
> + if (!res)
> + return -ENOMEM;
I should have looked at this more closely in the previous
revision. But if this error path is taken, does any cleanup
need to be done of 'res' entries that were allocated in
previous iterations of the for_each_of_range() loop? It
seems like the cleanup and releasing of previously allocated
memory should be done.
> +
> + res->name = "hyperv mmio";
> + res->flags = range.flags;
> + res->start = range.cpu_addr;
> + res->end = range.cpu_addr + range.size;
> +
> + *cur_res = res;
> + cur_res = &res->sibling;
> + }
> +
> + return ret;
> +}
>
> static int vmbus_platform_driver_probe(struct platform_device *pdev)
> {
> - return vmbus_acpi_add(pdev);
> + if (!acpi_disabled)
> + return vmbus_acpi_add(pdev);
> + else
> + return vmbus_device_add(pdev);
Nit: Usually when there's a negated test with if/else, it's best to flip the
if and else clauses so as to eliminate the negation. It's just slightly less
semantic load for the human reader to parse through. So:
if (acpi_disabled)
return vmbus_device_add(pdev);
else
return vmbus_acpi_add(pdev);
Everything else looks good to me.
Michael
> }
>
> static int vmbus_platform_driver_remove(struct platform_device *pdev)
> @@ -2634,7 +2679,17 @@ static int vmbus_bus_resume(struct device *dev)
> #define vmbus_bus_resume NULL
> #endif /* CONFIG_PM_SLEEP */
>
> -static const struct acpi_device_id vmbus_acpi_device_ids[] = {
> +static const __maybe_unused struct of_device_id vmbus_of_match[] = {
> + {
> + .compatible = "microsoft,vmbus",
> + },
> + {
> + /* sentinel */
> + },
> +};
> +MODULE_DEVICE_TABLE(of, vmbus_of_match);
> +
> +static const __maybe_unused struct acpi_device_id vmbus_acpi_device_ids[] = {
> {"VMBUS", 0},
> {"VMBus", 0},
> {"", 0},
> @@ -2668,6 +2723,7 @@ 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,
> }
> --
> 2.34.1
On Tue, Mar 14, 2023 at 09:21:32PM +0000, Michael Kelley (LINUX) wrote:
> From: Saurabh Sengar <[email protected]> Sent: Tuesday, March 14, 2023 2:16 AM
> >
> > Update the driver to support Devicetree boot as well along with ACPI.
> > At present the Devicetree parsing only provides the mmio region info
> > and is not the exact copy of ACPI parsing. This is sufficient to cater
> > all the current Devicetree usecases for VMBus.
> >
> > Currently Devicetree is supported only for x86 systems.
> >
> > Signed-off-by: Saurabh Sengar <[email protected]>
> > ---
> > [V8]
> > - Remove the auto select of CONFIG_OF
> > - Remove the dependency on !ACPI for OF_EARLY_FLATTREE
> > - Used acpi_disabled instead of #ifdef and hence added a dummy function
> > for vmbus_acpi_add
> > - GFP_ATOMIC -> GFP_KERNEL
> > - used range.flags instead of hard coding flags
> > - used __maybe_unused for acpi device id, removed #ifdef
> >
> > drivers/hv/Kconfig | 5 ++--
> > drivers/hv/vmbus_drv.c | 64 +++++++++++++++++++++++++++++++++++++++---
> > 2 files changed, 63 insertions(+), 6 deletions(-)
> >
> > diff --git a/drivers/hv/Kconfig b/drivers/hv/Kconfig
> > index 0747a8f1fcee..47132b30b7ee 100644
> > --- a/drivers/hv/Kconfig
> > +++ b/drivers/hv/Kconfig
> > @@ -4,11 +4,12 @@ menu "Microsoft Hyper-V guest support"
> >
> > config HYPERV
> > tristate "Microsoft Hyper-V client drivers"
> > - depends on ACPI && ((X86 && X86_LOCAL_APIC && HYPERVISOR_GUEST) \
> > - || (ARM64 && !CPU_BIG_ENDIAN))
> > + depends on (X86 && X86_LOCAL_APIC && HYPERVISOR_GUEST) \
> > + || (ACPI && ARM64 && !CPU_BIG_ENDIAN)
> > select PARAVIRT
> > select X86_HV_CALLBACK_VECTOR if X86
> > select VMAP_PFN
> > + select OF_EARLY_FLATTREE if OF
> > help
> > Select this option to run Linux as a Hyper-V client operating
> > system.
> > diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
> > index 3ad2fa2b92e7..15097e1f3f2b 100644
> > --- a/drivers/hv/vmbus_drv.c
> > +++ b/drivers/hv/vmbus_drv.c
> > @@ -20,6 +20,7 @@
> > #include <linux/completion.h>
> > #include <linux/hyperv.h>
> > #include <linux/kernel_stat.h>
> > +#include <linux/of_address.h>
> > #include <linux/clockchips.h>
> > #include <linux/cpu.h>
> > #include <linux/sched/isolation.h>
> > @@ -2143,7 +2144,7 @@ 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.
> > @@ -2253,6 +2254,7 @@ static acpi_status vmbus_walk_resources(struct acpi_resource
> > *res, void *ctx)
> >
> > return AE_OK;
> > }
> > +#endif
> >
> > static void vmbus_mmio_remove(void)
> > {
> > @@ -2273,7 +2275,7 @@ 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;
> > @@ -2433,6 +2435,7 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t
> > size)
> > }
> > EXPORT_SYMBOL_GPL(vmbus_free_mmio);
> >
> > +#ifdef CONFIG_ACPI
> > static int vmbus_acpi_add(struct platform_device *pdev)
> > {
> > acpi_status result;
> > @@ -2485,10 +2488,52 @@ static int vmbus_acpi_add(struct platform_device *pdev)
> > vmbus_mmio_remove();
> > return ret_val;
> > }
> > +#else
> > +static int vmbus_acpi_add(struct platform_device *pdev)
> > +{
> > + return 0;
> > +}
> > +#endif
> > +
> > +static int vmbus_device_add(struct platform_device *pdev)
> > +{
> > + struct resource **cur_res = &hyperv_mmio;
> > + struct of_range range;
> > + struct of_range_parser parser;
> > + struct device_node *np = pdev->dev.of_node;
> > + int ret;
> > +
> > + hv_dev = &pdev->dev;
> > +
> > + ret = of_range_parser_init(&parser, np);
> > + if (ret)
> > + return ret;
> > +
> > + for_each_of_range(&parser, &range) {
> > + struct resource *res;
> > +
> > + res = kzalloc(sizeof(*res), GFP_KERNEL);
> > + if (!res)
> > + return -ENOMEM;
>
> I should have looked at this more closely in the previous
> revision. But if this error path is taken, does any cleanup
> need to be done of 'res' entries that were allocated in
> previous iterations of the for_each_of_range() loop? It
> seems like the cleanup and releasing of previously allocated
> memory should be done.
No problem, thanks for pointing this.
For now we have only one range in device tree so this case may not occur,
but for completeness I agree we should add that, in future we might
have multiple ranges supported.
I will add vmbus_mmio_remove() in case of error here, I guess that
should be sufficient.
>
> > +
> > + res->name = "hyperv mmio";
> > + res->flags = range.flags;
> > + res->start = range.cpu_addr;
> > + res->end = range.cpu_addr + range.size;
> > +
> > + *cur_res = res;
> > + cur_res = &res->sibling;
> > + }
> > +
> > + return ret;
> > +}
> >
> > static int vmbus_platform_driver_probe(struct platform_device *pdev)
> > {
> > - return vmbus_acpi_add(pdev);
> > + if (!acpi_disabled)
> > + return vmbus_acpi_add(pdev);
> > + else
> > + return vmbus_device_add(pdev);
>
> Nit: Usually when there's a negated test with if/else, it's best to flip the
> if and else clauses so as to eliminate the negation. It's just slightly less
> semantic load for the human reader to parse through. So:
>
> if (acpi_disabled)
> return vmbus_device_add(pdev);
> else
> return vmbus_acpi_add(pdev);
>
> Everything else looks good to me.
Sure will fix.
Regards,
Saurabh
>
> Michael
>
> > }
> >
> > static int vmbus_platform_driver_remove(struct platform_device *pdev)
> > @@ -2634,7 +2679,17 @@ static int vmbus_bus_resume(struct device *dev)
> > #define vmbus_bus_resume NULL
> > #endif /* CONFIG_PM_SLEEP */
> >
> > -static const struct acpi_device_id vmbus_acpi_device_ids[] = {
> > +static const __maybe_unused struct of_device_id vmbus_of_match[] = {
> > + {
> > + .compatible = "microsoft,vmbus",
> > + },
> > + {
> > + /* sentinel */
> > + },
> > +};
> > +MODULE_DEVICE_TABLE(of, vmbus_of_match);
> > +
> > +static const __maybe_unused struct acpi_device_id vmbus_acpi_device_ids[] = {
> > {"VMBUS", 0},
> > {"VMBus", 0},
> > {"", 0},
> > @@ -2668,6 +2723,7 @@ 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,
> > }
> > --
> > 2.34.1