[v6,13/13] riscv: mm: Always use an ASID to flush mm contexts

Commit Message

Samuel Holland March 27, 2024, 4:49 a.m. UTC

Even if multiple ASIDs are not supported, using the single-ASID variant
of the sfence.vma instruction preserves TLB entries for global (kernel)
pages. So it is always more efficient to use the single-ASID code path.

Reviewed-by: Alexandre Ghiti <[e-mail]>
Signed-off-by: Samuel Holland <[e-mail]>
---

(no changes since v5)

Changes in v5:
 - Leave use_asid_allocator declared in asm/mmu_context.h

Changes in v4:
 - There is now only one copy of __flush_tlb_range()

Changes in v2:
 - Update both copies of __flush_tlb_range()

 arch/riscv/mm/tlbflush.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

Patch

diff --git a/arch/riscv/mm/tlbflush.c b/arch/riscv/mm/tlbflush.c
index 35266dd9a9a2..44e7ed4e194f 100644
--- a/arch/riscv/mm/tlbflush.c
+++ b/arch/riscv/mm/tlbflush.c
@@ -109,8 +109,7 @@  static void __flush_tlb_range(struct cpumask *cmask, unsigned long asid,
 
 static inline unsigned long get_mm_asid(struct mm_struct *mm)
 {
-	return static_branch_unlikely(&use_asid_allocator) ?
-			cntx2asid(atomic_long_read(&mm->context.id)) : FLUSH_TLB_NO_ASID;
+	return cntx2asid(atomic_long_read(&mm->context.id));
 }
 
 void flush_tlb_mm(struct mm_struct *mm)