| Message ID | 128d17f3625807a26875b7a419fb56610424b18c.1726048521.git.oleksii.kurochko@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Move {acpi_}device_init() and device_get_class() to common code | expand |
On 11.09.2024 12:04, Oleksii Kurochko wrote: > --- a/xen/common/Makefile > +++ b/xen/common/Makefile > @@ -6,6 +6,7 @@ obj-$(CONFIG_HYPFS_CONFIG) += config_data.o > obj-$(CONFIG_CORE_PARKING) += core_parking.o > obj-y += cpu.o > obj-$(CONFIG_DEBUG_TRACE) += debugtrace.o > +obj-$(call or,$(CONFIG_HAS_DEVICE_TREE),$(CONFIG_HAS_ACPI)) += device.o I can't spot any HAS_ACPI in the tree. And if this was switched to CONFIG_ACPI I'd further ask why the file needs building on x86. Also I think I'd prefer to avoid the of the "or" macro here: obj-$(CONFIG_ACPI) += device.o obj-$(CONFIG_HAS_DEVICE_TREE) += device.o ought to be quite fine. There's de-duplication somewhere for what $(obj-y) lists. > --- /dev/null > +++ b/xen/common/device.c > @@ -0,0 +1,82 @@ > +/* SPDX-License-Identifier: GPL-2.0-or-later */ > +/* > + * Based on the code from: > + * xen/arch/arm/device.c > + */ > + > +#include <xen/bug.h> > +#include <xen/device_tree.h> > +#include <xen/errno.h> > +#include <xen/init.h> > + > +#include <asm-generic/device.h> > + > +#ifdef CONFIG_ACPI > +extern const struct acpi_device_desc _asdevice[], _aedevice[]; Why does this live separately here, rather than ... > +#endif > + > +#ifdef CONFIG_HAS_DEVICE_TREE > + > +extern const struct device_desc _sdevice[], _edevice[]; ... like this ... > +int __init device_init(struct dt_device_node *dev, enum device_class class, > + const void *data) > +{ > + const struct device_desc *desc; > + > + ASSERT(dev != NULL); > + > + if ( !dt_device_is_available(dev) || dt_device_for_passthrough(dev) ) > + return -ENODEV; > + > + for ( desc = _sdevice; desc != _edevice; desc++ ) > + { > + if ( desc->class != class ) > + continue; > + > + if ( dt_match_node(desc->dt_match, dev) ) > + { > + ASSERT(desc->init != NULL); > + > + return desc->init(dev, data); > + } > + } > + > + return -EBADF; > +} > + > +enum device_class device_get_class(const struct dt_device_node *dev) > +{ > + const struct device_desc *desc; > + > + ASSERT(dev != NULL); > + > + for ( desc = _sdevice; desc != _edevice; desc++ ) > + { > + if ( dt_match_node(desc->dt_match, dev) ) > + return desc->class; > + } > + > + return DEVICE_UNKNOWN; > +} > + > +#endif > + > +#ifdef CONFIG_ACPI ... in the section where it's needed? Leaving just one #ifdef for ACPI. Jan
On Thu, 2024-09-12 at 17:28 +0200, Jan Beulich wrote: > On 11.09.2024 12:04, Oleksii Kurochko wrote: > > --- a/xen/common/Makefile > > +++ b/xen/common/Makefile > > @@ -6,6 +6,7 @@ obj-$(CONFIG_HYPFS_CONFIG) += config_data.o > > obj-$(CONFIG_CORE_PARKING) += core_parking.o > > obj-y += cpu.o > > obj-$(CONFIG_DEBUG_TRACE) += debugtrace.o > > +obj-$(call or,$(CONFIG_HAS_DEVICE_TREE),$(CONFIG_HAS_ACPI)) += > > device.o > > I can't spot any HAS_ACPI in the tree. And if this was switched to > CONFIG_ACPI > I'd further ask why the file needs building on x86. Oh, there is no need for building this on x86. With what you suggested here ... > > Also I think I'd prefer to avoid the of the "or" macro here: > > obj-$(CONFIG_ACPI) += device.o > obj-$(CONFIG_HAS_DEVICE_TREE) += device.o ... IIUC it will fix the issue with building this file for x86 as CONFIG_ACPI depends on (ARM_64 && ARM_EFI). > > ought to be quite fine. There's de-duplication somewhere for what > $(obj-y) lists. > > > --- /dev/null > > +++ b/xen/common/device.c > > @@ -0,0 +1,82 @@ > > +/* SPDX-License-Identifier: GPL-2.0-or-later */ > > +/* > > + * Based on the code from: > > + * xen/arch/arm/device.c > > + */ > > + > > +#include <xen/bug.h> > > +#include <xen/device_tree.h> > > +#include <xen/errno.h> > > +#include <xen/init.h> > > + > > +#include <asm-generic/device.h> > > + > > +#ifdef CONFIG_ACPI > > +extern const struct acpi_device_desc _asdevice[], _aedevice[]; > > Why does this live separately here, rather than ... > > > +#endif > > + > > +#ifdef CONFIG_HAS_DEVICE_TREE > > + > > +extern const struct device_desc _sdevice[], _edevice[]; > > ... like this ... > > > +int __init device_init(struct dt_device_node *dev, enum > > device_class class, > > + const void *data) > > +{ > > + const struct device_desc *desc; > > + > > + ASSERT(dev != NULL); > > + > > + if ( !dt_device_is_available(dev) || > > dt_device_for_passthrough(dev) ) > > + return -ENODEV; > > + > > + for ( desc = _sdevice; desc != _edevice; desc++ ) > > + { > > + if ( desc->class != class ) > > + continue; > > + > > + if ( dt_match_node(desc->dt_match, dev) ) > > + { > > + ASSERT(desc->init != NULL); > > + > > + return desc->init(dev, data); > > + } > > + } > > + > > + return -EBADF; > > +} > > + > > +enum device_class device_get_class(const struct dt_device_node > > *dev) > > +{ > > + const struct device_desc *desc; > > + > > + ASSERT(dev != NULL); > > + > > + for ( desc = _sdevice; desc != _edevice; desc++ ) > > + { > > + if ( dt_match_node(desc->dt_match, dev) ) > > + return desc->class; > > + } > > + > > + return DEVICE_UNKNOWN; > > +} > > + > > +#endif > > + > > +#ifdef CONFIG_ACPI > > ... in the section where it's needed? Leaving just one #ifdef for > ACPI. Just tried to follow the approach that first is going variables declaration/definitions and then functions. But I am okay to move it to CONFIG_ACPI #ifdef CONFIG_ACPI ... #endif. Thanks. ~ Oleksii
On 13.09.2024 16:35, oleksii.kurochko@gmail.com wrote: > On Thu, 2024-09-12 at 17:28 +0200, Jan Beulich wrote: >> On 11.09.2024 12:04, Oleksii Kurochko wrote: >>> --- a/xen/common/Makefile >>> +++ b/xen/common/Makefile >>> @@ -6,6 +6,7 @@ obj-$(CONFIG_HYPFS_CONFIG) += config_data.o >>> obj-$(CONFIG_CORE_PARKING) += core_parking.o >>> obj-y += cpu.o >>> obj-$(CONFIG_DEBUG_TRACE) += debugtrace.o >>> +obj-$(call or,$(CONFIG_HAS_DEVICE_TREE),$(CONFIG_HAS_ACPI)) += >>> device.o >> >> I can't spot any HAS_ACPI in the tree. And if this was switched to >> CONFIG_ACPI >> I'd further ask why the file needs building on x86. > Oh, there is no need for building this on x86. With what you suggested > here ... > >> >> Also I think I'd prefer to avoid the of the "or" macro here: >> >> obj-$(CONFIG_ACPI) += device.o >> obj-$(CONFIG_HAS_DEVICE_TREE) += device.o > ... IIUC it will fix the issue with building this file for x86 as > CONFIG_ACPI depends on (ARM_64 && ARM_EFI). Except that "depends on" is itself Arm-only, so won't affect x86. Or else x86 would end up without ACPI support, which would mean full breakage on about every system. Jan
On Fri, 2024-09-13 at 19:45 +0200, Jan Beulich wrote: > On 13.09.2024 16:35, oleksii.kurochko@gmail.com wrote: > > On Thu, 2024-09-12 at 17:28 +0200, Jan Beulich wrote: > > > On 11.09.2024 12:04, Oleksii Kurochko wrote: > > > > --- a/xen/common/Makefile > > > > +++ b/xen/common/Makefile > > > > @@ -6,6 +6,7 @@ obj-$(CONFIG_HYPFS_CONFIG) += config_data.o > > > > obj-$(CONFIG_CORE_PARKING) += core_parking.o > > > > obj-y += cpu.o > > > > obj-$(CONFIG_DEBUG_TRACE) += debugtrace.o > > > > +obj-$(call or,$(CONFIG_HAS_DEVICE_TREE),$(CONFIG_HAS_ACPI)) += > > > > device.o > > > > > > I can't spot any HAS_ACPI in the tree. And if this was switched > > > to > > > CONFIG_ACPI > > > I'd further ask why the file needs building on x86. > > Oh, there is no need for building this on x86. With what you > > suggested > > here ... > > > > > > > > Also I think I'd prefer to avoid the of the "or" macro here: > > > > > > obj-$(CONFIG_ACPI) += device.o > > > obj-$(CONFIG_HAS_DEVICE_TREE) += device.o > > ... IIUC it will fix the issue with building this file for x86 as > > CONFIG_ACPI depends on (ARM_64 && ARM_EFI). > > Except that "depends on" is itself Arm-only, so won't affect x86. > Or else x86 would end up without ACPI support, which would mean > full breakage on about every system. There is another CONFIG_ACPI in xen/drivers/acpi which is equal to 'y' for x86 so it seems to me that it is needed another config ( GENERIC_DEVICE_INIT ? ) which will be disabled for x86 by default so device.o won't be compiled for x86. Have I overlooked something or better option exist? Probably it would be better to use "and" macro? ~ Oleksii
diff --git a/xen/arch/arm/device.c b/xen/arch/arm/device.c index 3e02cff008..5610cddcba 100644 --- a/xen/arch/arm/device.c +++ b/xen/arch/arm/device.c @@ -8,79 +8,12 @@ * Copyright (C) 2013 Linaro Limited. */ -#include <asm/device.h> -#include <asm/setup.h> +#include <xen/device_tree.h> #include <xen/errno.h> -#include <xen/init.h> #include <xen/iocap.h> #include <xen/lib.h> -extern const struct device_desc _sdevice[], _edevice[]; - -#ifdef CONFIG_ACPI -extern const struct acpi_device_desc _asdevice[], _aedevice[]; -#endif - -int __init device_init(struct dt_device_node *dev, enum device_class class, - const void *data) -{ - const struct device_desc *desc; - - ASSERT(dev != NULL); - - if ( !dt_device_is_available(dev) || dt_device_for_passthrough(dev) ) - return -ENODEV; - - for ( desc = _sdevice; desc != _edevice; desc++ ) - { - if ( desc->class != class ) - continue; - - if ( dt_match_node(desc->dt_match, dev) ) - { - ASSERT(desc->init != NULL); - - return desc->init(dev, data); - } - - } - - return -EBADF; -} - -#ifdef CONFIG_ACPI -int __init acpi_device_init(enum device_class class, const void *data, int class_type) -{ - const struct acpi_device_desc *desc; - - for ( desc = _asdevice; desc != _aedevice; desc++ ) - { - if ( ( desc->class != class ) || ( desc->class_type != class_type ) ) - continue; - - ASSERT(desc->init != NULL); - - return desc->init(data); - } - - return -EBADF; -} -#endif - -enum device_class device_get_class(const struct dt_device_node *dev) -{ - const struct device_desc *desc; - - ASSERT(dev != NULL); - - for ( desc = _sdevice; desc != _edevice; desc++ ) - { - if ( dt_match_node(desc->dt_match, dev) ) - return desc->class; - } - - return DEVICE_UNKNOWN; -} +#include <asm/setup.h> int map_irq_to_domain(struct domain *d, unsigned int irq, bool need_mapping, const char *devname) diff --git a/xen/common/Makefile b/xen/common/Makefile index fc52e0857d..703e26de1d 100644 --- a/xen/common/Makefile +++ b/xen/common/Makefile @@ -6,6 +6,7 @@ obj-$(CONFIG_HYPFS_CONFIG) += config_data.o obj-$(CONFIG_CORE_PARKING) += core_parking.o obj-y += cpu.o obj-$(CONFIG_DEBUG_TRACE) += debugtrace.o +obj-$(call or,$(CONFIG_HAS_DEVICE_TREE),$(CONFIG_HAS_ACPI)) += device.o obj-$(CONFIG_HAS_DEVICE_TREE) += device-tree/ obj-$(CONFIG_IOREQ_SERVER) += dm.o obj-y += domain.o diff --git a/xen/common/device.c b/xen/common/device.c new file mode 100644 index 0000000000..e34a7dc88e --- /dev/null +++ b/xen/common/device.c @@ -0,0 +1,82 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Based on the code from: + * xen/arch/arm/device.c + */ + +#include <xen/bug.h> +#include <xen/device_tree.h> +#include <xen/errno.h> +#include <xen/init.h> + +#include <asm-generic/device.h> + +#ifdef CONFIG_ACPI +extern const struct acpi_device_desc _asdevice[], _aedevice[]; +#endif + +#ifdef CONFIG_HAS_DEVICE_TREE + +extern const struct device_desc _sdevice[], _edevice[]; + +int __init device_init(struct dt_device_node *dev, enum device_class class, + const void *data) +{ + const struct device_desc *desc; + + ASSERT(dev != NULL); + + if ( !dt_device_is_available(dev) || dt_device_for_passthrough(dev) ) + return -ENODEV; + + for ( desc = _sdevice; desc != _edevice; desc++ ) + { + if ( desc->class != class ) + continue; + + if ( dt_match_node(desc->dt_match, dev) ) + { + ASSERT(desc->init != NULL); + + return desc->init(dev, data); + } + } + + return -EBADF; +} + +enum device_class device_get_class(const struct dt_device_node *dev) +{ + const struct device_desc *desc; + + ASSERT(dev != NULL); + + for ( desc = _sdevice; desc != _edevice; desc++ ) + { + if ( dt_match_node(desc->dt_match, dev) ) + return desc->class; + } + + return DEVICE_UNKNOWN; +} + +#endif + +#ifdef CONFIG_ACPI +int __init acpi_device_init(enum device_class class, const void *data, int class_type) +{ + const struct acpi_device_desc *desc; + + for ( desc = _asdevice; desc != _aedevice; desc++ ) + { + if ( ( desc->class != class ) || ( desc->class_type != class_type ) ) + continue; + + ASSERT(desc->init != NULL); + + return desc->init(data); + } + + return -EBADF; +} +#endif
Remove the device initialization code from `xen/arch/arm/device.c` and move it to the common code to avoid duplication and make it accessible for both ARM and other architectures. device_get_class(), device_init(), _sdevice[] and _edevice[] are wrapped by "#ifdef CONFIG_HAS_DEVICE_TREE" for the case if an arch doesn't support device tree. Remove unnecessary inclusions of <asm/device.h> and <xen/init.h> from `xen/arch/arm/device.c` as no code in the file relies on these headers. Fix the inclusion order by moving <asm/setup.h> after <xen/*> headers to resolve a compilation error: ./include/public/xen.h:968:35: error: unknown type name 'uint64_t' 968 | __DEFINE_XEN_GUEST_HANDLE(uint64, uint64_t); | ^~~~~~~~ ./include/public/arch-arm.h:191:21: note: in definition of macro '___DEFINE_XEN_GUEST_HANDLE' 191 | typedef union { type *p; uint64_aligned_t q; } \ | ^~~~ ./include/public/xen.h:968:1: note: in expansion of macro '__DEFINE_XEN_GUEST_HANDLE' 968 | __DEFINE_XEN_GUEST_HANDLE(uint64, uint64_t); because <asm/setup.h> includes <public/version.h>, which in turn includes "xen.h", which requires <xen/types.h> to be processed correctly. Additionally, add <xen/device_tree.h> to `device.c` as functions from this header are used within the file. Additionally, the Makefile is updated to compile `device.o` conditionally based on whether `CONFIG_HAS_DEVICE_TREE` or `CONFIG_HAS_ACPI` is enabled, ensuring the correct device initialization code is included as needed. Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com> --- xen/arch/arm/device.c | 71 ++----------------------------------- xen/common/Makefile | 1 + xen/common/device.c | 82 +++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 85 insertions(+), 69 deletions(-) create mode 100644 xen/common/device.c