[PATCH v7 07/10] kernel/jump_label: abstract jump_entry member accessors

Ard Biesheuvel ard.biesheuvel at linaro.org
Fri Jan 5 10:29:09 PST 2018


On 5 January 2018 at 18:22, Catalin Marinas <catalin.marinas at arm.com> wrote:
> On Fri, Jan 05, 2018 at 06:01:33PM +0000, Ard Biesheuvel wrote:
>> On 5 January 2018 at 17:58, Catalin Marinas <catalin.marinas at arm.com> wrote:
>> > On Tue, Jan 02, 2018 at 08:05:46PM +0000, Ard Biesheuvel wrote:
>> >> diff --git a/arch/arm/include/asm/jump_label.h b/arch/arm/include/asm/jump_label.h
>> >> index e12d7d096fc0..7b05b404063a 100644
>> >> --- a/arch/arm/include/asm/jump_label.h
>> >> +++ b/arch/arm/include/asm/jump_label.h
>> >> @@ -45,5 +45,32 @@ struct jump_entry {
>> >>       jump_label_t key;
>> >>  };
>> >>
>> >> +static inline jump_label_t jump_entry_code(const struct jump_entry *entry)
>> >> +{
>> >> +     return entry->code;
>> >> +}
>> >> +
>> >> +static inline struct static_key *jump_entry_key(const struct jump_entry *entry)
>> >> +{
>> >> +     return (struct static_key *)((unsigned long)entry->key & ~1UL);
>> >> +}
>> >> +
>> >> +static inline bool jump_entry_is_branch(const struct jump_entry *entry)
>> >> +{
>> >> +     return (unsigned long)entry->key & 1UL;
>> >> +}
>> >> +
>> >> +static inline bool jump_entry_is_module_init(const struct jump_entry *entry)
>> >> +{
>> >> +     return entry->code == 0;
>> >> +}
>> >> +
>> >> +static inline void jump_entry_set_module_init(struct jump_entry *entry)
>> >> +{
>> >> +     entry->code = 0;
>> >> +}
>> >> +
>> >> +#define jump_label_swap              NULL
>> >
>> > Is there any difference between these functions on any of the
>> > architectures touched? Even with the relative offset, arm64 and x86
>> > looked the same to me (well, I may have missed some detail).
>>
>> No, the latter two are identical everywhere, and the others are the
>> same modulo absolute vs relative.
>>
>> The issue is that the struct definition is per-arch so the accessors
>> should be as well.
>
> Up to this patch, even the jump_entry structure is the same on all
> architectures (the jump_label_t type differs).
>
> With relative offset, can you not just define jump_label_t to s32? At a
> quick grep in mainline, it doesn't seem to be used outside the structure
> definition.
>

I think we can just remove jump_label_t entirely, and replace it with
unsigned long for absolute, and s32 for relative. Maybe I am missing
something, but things like

#ifdef CONFIG_X86_64
typedef u64 jump_label_t;
#else
typedef u32 jump_label_t;
#endif

seem a bit pointless to me anyway.


>> Perhaps I should introduce two variants two asm-generic, similar to
>> how we have different flavors of unaligned accessors.
>
> You could as well define them directly in kernel/jump_label.h or, if
> used outside this file, include/linux/jump_label.h.
>

Perhaps I should define a Kconfig symbol after all for relative jump
labels, and just keep everything in the same file. The question is
whether I should use CONFIG_HAVE_ARCH_PREL32_RELOCATIONS for this as
well.



More information about the linux-arm-kernel mailing list