Hi Al,
Please take these as bug reports to Redhat maintained patches. I am unsure of the correct fixes in these cases.
I have pushed these to the mustang topic branch and will push the SPCR one to the other topics as well.
Thanks
Graeme
SPCR is broken in CONFIG_ACPI=n the function headers are defined in asm/acpi.h which is not included when CONFIG_ACPI=n
Signed-off-by: Graeme Gregory graeme.gregory@linaro.org --- include/linux/acpi.h | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/include/linux/acpi.h b/include/linux/acpi.h index ae29a09..a75275e 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -813,4 +813,12 @@ static inline struct acpi_device *acpi_get_next_child(struct device *dev,
#endif
+#if !defined(CONFIG_ACPI) +static inline bool acpi_spcr_console_check(struct acpi_device *adev, + char *name, int index) +{ + return false; +} +#endif + #endif /*_LINUX_ACPI_H*/
I have just realised the SPRC patch series in acpi-topic-mustang is different to the one in acpi-topic-seattle which does not suffer from this problem as the definitions are in include/linux/acpi.h instead and this function is already stubbed.
Graeme
On 14 July 2015 at 11:17, Graeme Gregory graeme.gregory@linaro.org wrote:
SPCR is broken in CONFIG_ACPI=n the function headers are defined in asm/acpi.h which is not included when CONFIG_ACPI=n
Signed-off-by: Graeme Gregory graeme.gregory@linaro.org
include/linux/acpi.h | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/include/linux/acpi.h b/include/linux/acpi.h index ae29a09..a75275e 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -813,4 +813,12 @@ static inline struct acpi_device *acpi_get_next_child(struct device *dev,
#endif
+#if !defined(CONFIG_ACPI) +static inline bool acpi_spcr_console_check(struct acpi_device *adev,
char *name, int index)+{
return false;+} +#endif
#endif /*_LINUX_ACPI_H*/
2.1.4
This just leads to compile errors so stub the _probe function in the CONFIG_ACPI=n case.
Signed-off-by: Graeme Gregory graeme.gregory@linaro.org --- virt/kvm/arm/vgic-v2.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/virt/kvm/arm/vgic-v2.c b/virt/kvm/arm/vgic-v2.c index 5ba2ea6..5d75118 100644 --- a/virt/kvm/arm/vgic-v2.c +++ b/virt/kvm/arm/vgic-v2.c @@ -27,7 +27,7 @@
#include <linux/irqchip/arm-gic.h>
-#include <asm/acpi.h> +//#include <asm/acpi.h> #include <asm/kvm_emulate.h> #include <asm/kvm_arm.h> #include <asm/kvm_mmu.h> @@ -260,6 +260,7 @@ out: return ret; }
+#ifdef CONFIG_ACPI struct acpi_madt_generic_interrupt *vgic_acpi; static void gic_get_acpi_header(struct acpi_subtable_header *header) { @@ -329,3 +330,11 @@ int vgic_v2_acpi_probe(const struct vgic_ops **ops, out: return ret; } +#else +int vgic_v2_acpi_probe(const struct vgic_ops **ops, + const struct vgic_params **params) +{ + return -ENOSYS; +} +#endif +
On 14 July 2015 at 11:17, Graeme Gregory graeme.gregory@linaro.org wrote:
This just leads to compile errors so stub the _probe function in the CONFIG_ACPI=n case.
Signed-off-by: Graeme Gregory graeme.gregory@linaro.org
virt/kvm/arm/vgic-v2.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/virt/kvm/arm/vgic-v2.c b/virt/kvm/arm/vgic-v2.c index 5ba2ea6..5d75118 100644 --- a/virt/kvm/arm/vgic-v2.c +++ b/virt/kvm/arm/vgic-v2.c @@ -27,7 +27,7 @@
#include <linux/irqchip/arm-gic.h>
-#include <asm/acpi.h> +//#include <asm/acpi.h>
Ignore this, I forgot to remove it before committing!
Graeme
#include <asm/kvm_emulate.h> #include <asm/kvm_arm.h> #include <asm/kvm_mmu.h> @@ -260,6 +260,7 @@ out: return ret; }
+#ifdef CONFIG_ACPI struct acpi_madt_generic_interrupt *vgic_acpi; static void gic_get_acpi_header(struct acpi_subtable_header *header) { @@ -329,3 +330,11 @@ int vgic_v2_acpi_probe(const struct vgic_ops **ops, out: return ret; } +#else +int vgic_v2_acpi_probe(const struct vgic_ops **ops,
const struct vgic_params **params)+{
return -ENOSYS;+} +#endif
-- 2.1.4
-
G Gregory -
Graeme Gregory