m68k: mm: Improve kernel_page_table()
With the PTE-tables now only being 256 bytes, allocating a full page for them is a giant waste. Start by improving the boot time allocator such that init_mm initialization will at least have optimal memory density. Much thanks to Will Deacon in help with debugging and ferreting out lost information on these dusty MMUs. Notes: - _TABLE_MASK is reduced to account for the shorter (256 byte) alignment of pte-tables, per the manual, table entries should only ever have state in the low 4 bits (Used,WrProt,Desc1,Desc0) so it is still longer than strictly required. (Thanks Will!!!) - Also use kernel_page_table() for the 020/030 zero_pgtable case and consequently remove the zero_pgtable init hack (will fix up later). Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Will Deacon <will@kernel.org> Acked-by: Greg Ungerer <gerg@linux-m68k.org> Tested-by: Michael Schmitz <schmitzmic@gmail.com> Tested-by: Greg Ungerer <gerg@linux-m68k.org> Link: https://lore.kernel.org/r/20200131125403.768263973@infradead.org Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
This commit is contained in:

committed by
Geert Uytterhoeven

parent
ef22d8abd8
commit
ef9285f69f
@@ -42,7 +42,6 @@ EXPORT_SYMBOL(empty_zero_page);
|
||||
|
||||
#if !defined(CONFIG_SUN3) && !defined(CONFIG_COLDFIRE)
|
||||
extern void init_pointer_table(unsigned long ptable);
|
||||
extern pmd_t *zero_pgtable;
|
||||
#endif
|
||||
|
||||
#ifdef CONFIG_MMU
|
||||
@@ -135,10 +134,6 @@ static inline void init_pointer_tables(void)
|
||||
if (pud_present(*pud))
|
||||
init_pointer_table(pgd_page_vaddr(kernel_pg_dir[i]));
|
||||
}
|
||||
|
||||
/* insert also pointer table that we used to unmap the zero page */
|
||||
if (zero_pgtable)
|
||||
init_pointer_table((unsigned long)zero_pgtable);
|
||||
#endif
|
||||
}
|
||||
|
||||
|
Reference in New Issue
Block a user