| Message ID | 746a5cd6-1446-eda4-8b23-03c1cac30b8d@suse.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | eef6a53d9931b080c26c272a0d1207e60b4c1bb5 |
| Headers | show |
| Series | Xen: drop exports of {set,clear}_foreign_p2m_mapping() | expand |
On Tue, 9 Mar 2021, Jan Beulich wrote: > They're only used internally, and the layering violation they contain > (x86) or imply (Arm) of calling HYPERVISOR_grant_table_op() strongly > advise against any (uncontrolled) use from a module. The functions also > never had users except the ones from drivers/xen/grant-table.c forever > since their introduction in 3.15. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Stefano Stabellini <sstabellini@kernel.org> > --- > It's not clear to me why Arm doesn't have merely stubs just like x86 has > when PV support is disabled. ARM implements the two functions to keep track of foreign pages mapped into the system. It uses swiotlb-xen to translate gfn to mfn for foreign pages. Local pages are 1:1. This way, an IOMMU is not required. > --- a/arch/arm/xen/p2m.c > +++ b/arch/arm/xen/p2m.c > @@ -130,7 +130,6 @@ int set_foreign_p2m_mapping(struct gntta > > return 0; > } > -EXPORT_SYMBOL_GPL(set_foreign_p2m_mapping); > > int clear_foreign_p2m_mapping(struct gnttab_unmap_grant_ref *unmap_ops, > struct gnttab_unmap_grant_ref *kunmap_ops, > @@ -145,7 +144,6 @@ int clear_foreign_p2m_mapping(struct gnt > > return 0; > } > -EXPORT_SYMBOL_GPL(clear_foreign_p2m_mapping); > > bool __set_phys_to_machine_multi(unsigned long pfn, > unsigned long mfn, unsigned long nr_pages) > --- a/arch/x86/xen/p2m.c > +++ b/arch/x86/xen/p2m.c > @@ -776,7 +776,6 @@ int set_foreign_p2m_mapping(struct gntta > out: > return ret; > } > -EXPORT_SYMBOL_GPL(set_foreign_p2m_mapping); > > int clear_foreign_p2m_mapping(struct gnttab_unmap_grant_ref *unmap_ops, > struct gnttab_unmap_grant_ref *kunmap_ops, > @@ -802,7 +801,6 @@ int clear_foreign_p2m_mapping(struct gnt > > return ret; > } > -EXPORT_SYMBOL_GPL(clear_foreign_p2m_mapping); > > #ifdef CONFIG_XEN_DEBUG_FS > #include <linux/debugfs.h> >
On 3/9/21 12:00 PM, Jan Beulich wrote: > They're only used internally, and the layering violation they contain > (x86) or imply (Arm) of calling HYPERVISOR_grant_table_op() strongly > advise against any (uncontrolled) use from a module. The functions also > never had users except the ones from drivers/xen/grant-table.c forever > since their introduction in 3.15. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> Applied to for-linus-5.12b
--- a/arch/arm/xen/p2m.c +++ b/arch/arm/xen/p2m.c @@ -130,7 +130,6 @@ int set_foreign_p2m_mapping(struct gntta return 0; } -EXPORT_SYMBOL_GPL(set_foreign_p2m_mapping); int clear_foreign_p2m_mapping(struct gnttab_unmap_grant_ref *unmap_ops, struct gnttab_unmap_grant_ref *kunmap_ops, @@ -145,7 +144,6 @@ int clear_foreign_p2m_mapping(struct gnt return 0; } -EXPORT_SYMBOL_GPL(clear_foreign_p2m_mapping); bool __set_phys_to_machine_multi(unsigned long pfn, unsigned long mfn, unsigned long nr_pages) --- a/arch/x86/xen/p2m.c +++ b/arch/x86/xen/p2m.c @@ -776,7 +776,6 @@ int set_foreign_p2m_mapping(struct gntta out: return ret; } -EXPORT_SYMBOL_GPL(set_foreign_p2m_mapping); int clear_foreign_p2m_mapping(struct gnttab_unmap_grant_ref *unmap_ops, struct gnttab_unmap_grant_ref *kunmap_ops, @@ -802,7 +801,6 @@ int clear_foreign_p2m_mapping(struct gnt return ret; } -EXPORT_SYMBOL_GPL(clear_foreign_p2m_mapping); #ifdef CONFIG_XEN_DEBUG_FS #include <linux/debugfs.h>
They're only used internally, and the layering violation they contain (x86) or imply (Arm) of calling HYPERVISOR_grant_table_op() strongly advise against any (uncontrolled) use from a module. The functions also never had users except the ones from drivers/xen/grant-table.c forever since their introduction in 3.15. Signed-off-by: Jan Beulich <jbeulich@suse.com> --- It's not clear to me why Arm doesn't have merely stubs just like x86 has when PV support is disabled.