| Message ID | 20210324232825.1157363-1-npiggin@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | sparc32: add stub pud_page define for walking huge vmalloc page tables | expand |
From: Nicholas Piggin <npiggin@gmail.com> Date: Thu, 25 Mar 2021 09:28:25 +1000 > Similarly to the stub p4d_page in sparc64, add a stub pud_page, this > is needed for hugepages in the vmap page tables to be walked without > ifdefs, which should be no functional change for sparc32. > > Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Acked-by: David S. Miller <davem@davemloft.net>
Hi all, On Thu, 25 Mar 2021 09:28:25 +1000 Nicholas Piggin <npiggin@gmail.com> wrote: > > Similarly to the stub p4d_page in sparc64, add a stub pud_page, this > is needed for hugepages in the vmap page tables to be walked without > ifdefs, which should be no functional change for sparc32. > > Signed-off-by: Nicholas Piggin <npiggin@gmail.com> > --- > > This should go in Andrew's mm tree which has a compile error on sparc32 > introduced by the huge vmalloc pages, noticed by Stephen. > > Thanks, > Nick > > arch/sparc/include/asm/pgtable_32.h | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/arch/sparc/include/asm/pgtable_32.h b/arch/sparc/include/asm/pgtable_32.h > index 632cdb959542..a5cf79c149fe 100644 > --- a/arch/sparc/include/asm/pgtable_32.h > +++ b/arch/sparc/include/asm/pgtable_32.h > @@ -321,6 +321,9 @@ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot) > pgprot_val(newprot)); > } > > +/* only used by the huge vmap code, should never be called */ > +#define pud_page(pud) NULL > + > struct seq_file; > void mmu_info(struct seq_file *m); > > -- > 2.23.0 > I have used this instead of my hack fix in linux-next today.
diff --git a/arch/sparc/include/asm/pgtable_32.h b/arch/sparc/include/asm/pgtable_32.h index 632cdb959542..a5cf79c149fe 100644 --- a/arch/sparc/include/asm/pgtable_32.h +++ b/arch/sparc/include/asm/pgtable_32.h @@ -321,6 +321,9 @@ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot) pgprot_val(newprot)); } +/* only used by the huge vmap code, should never be called */ +#define pud_page(pud) NULL + struct seq_file; void mmu_info(struct seq_file *m);
Similarly to the stub p4d_page in sparc64, add a stub pud_page, this is needed for hugepages in the vmap page tables to be walked without ifdefs, which should be no functional change for sparc32. Signed-off-by: Nicholas Piggin <npiggin@gmail.com> --- This should go in Andrew's mm tree which has a compile error on sparc32 introduced by the huge vmalloc pages, noticed by Stephen. Thanks, Nick arch/sparc/include/asm/pgtable_32.h | 3 +++ 1 file changed, 3 insertions(+)