| Message ID | 20181003175052.31917-1-james.morse@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | asm-generic/pgtable-nop?d.h: define folded with a value for use in C | expand |
On Wed, Oct 03, 2018 at 06:50:52PM +0100, James Morse wrote: > It turns out "if (__is_defined(__PAGETABLE_PMD_FOLDED))" isn't equivalent > to "#ifdef __PAGETABLE_PMD_FOLDED". (who knew!) > > kconfig.h's __is_defined() expects a define of the form > "#define CONFIG_BOOGER 1". But these nop?d headers just have > "#define __PAGETABLE_PMD_FOLDED". I wonder why we're using __is_defined() at all on arm64. It looks like some internal kconfig.h helper for IS_ENABLED() etc. Even if it looks nicer, I'd rather go for an #ifdef in the arm64 set_pmd() code.
On Thu, Oct 04, 2018 at 10:45:00AM +0100, Catalin Marinas wrote: > On Wed, Oct 03, 2018 at 06:50:52PM +0100, James Morse wrote: > > It turns out "if (__is_defined(__PAGETABLE_PMD_FOLDED))" isn't equivalent > > to "#ifdef __PAGETABLE_PMD_FOLDED". (who knew!) > > > > kconfig.h's __is_defined() expects a define of the form > > "#define CONFIG_BOOGER 1". But these nop?d headers just have > > "#define __PAGETABLE_PMD_FOLDED". > > I wonder why we're using __is_defined() at all on arm64. It looks like > some internal kconfig.h helper for IS_ENABLED() etc. Even if it looks > nicer, I'd rather go for an #ifdef in the arm64 set_pmd() code. This was my fault entirely; sorry about that. I'd wanted to ensure that we got build coverage regardless of the kernel configuration, which was why i'd used __is_defined(). Thanks, Mark.
diff --git a/include/asm-generic/pgtable-nopmd.h b/include/asm-generic/pgtable-nopmd.h index f35f6e8149e4..fcb4769a075a 100644 --- a/include/asm-generic/pgtable-nopmd.h +++ b/include/asm-generic/pgtable-nopmd.h @@ -8,7 +8,7 @@ struct mm_struct; -#define __PAGETABLE_PMD_FOLDED +#define __PAGETABLE_PMD_FOLDED 1 /* * Having the pmd type consist of a pud gets the size right, and allows diff --git a/include/asm-generic/pgtable-nopud.h b/include/asm-generic/pgtable-nopud.h index e950b9c50f34..d300dbcddaf3 100644 --- a/include/asm-generic/pgtable-nopud.h +++ b/include/asm-generic/pgtable-nopud.h @@ -9,7 +9,7 @@ #else #include <asm-generic/pgtable-nop4d.h> -#define __PAGETABLE_PUD_FOLDED +#define __PAGETABLE_PUD_FOLDED 1 /* * Having the pud type consist of a p4d gets the size right, and allows
It turns out "if (__is_defined(__PAGETABLE_PMD_FOLDED))" isn't equivalent to "#ifdef __PAGETABLE_PMD_FOLDED". (who knew!) kconfig.h's __is_defined() expects a define of the form "#define CONFIG_BOOGER 1". But these nop?d headers just have "#define __PAGETABLE_PMD_FOLDED". This means ____is_defined()'s triplet passed to __take_second_arg() is 'empty-string, 1 0' in both cases, meaning these symbols can't be used from C. (they are always false). asm-generic gets away with this as its using the pre-processor's defined() macro on this, not the C __is_defined(). Add the expected '1'. Signed-off-by: James Morse <james.morse@arm.com> CC: Mark Rutland <mark.rutland@arm.com> --- This fixes an issue with patches queued for arm64. Can it go via the arm64 tree? Link: lore.kernel.org/r/0f9a3abc-4890-faf5-ee7e-18434641b858@arm.com include/asm-generic/pgtable-nopmd.h | 2 +- include/asm-generic/pgtable-nopud.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)