From: Peter Zijlstra <pet...@infradead.org>
Ideally we'd also use sparse to enforce this separation so it becomes much
more difficult to mess up.
Signed-off-by: Peter Zijlstra (Intel) <pet...@infradead.org>
Signed-off-by: Ingo Molnar <mi...@kernel.org>
Signed-off-by: Thomas Gleixner <t...@linutronix.de>
Cc: Andy Lutomirski <l...@kernel.org>
Cc: Boris Ostrovsky <boris.ostrov...@oracle.com>
Cc: Borislav Petkov <b...@alien8.de>
Cc: Brian Gerst <brge...@gmail.com>
Cc: Dave Hansen <dave.han...@linux.intel.com>
Cc: David Laight <david.lai...@aculab.com>
Cc: Denys Vlasenko <dvlas...@redhat.com>
Cc: Eduardo Valentin <edu...@amazon.com>
Cc: Greg KH <gre...@linuxfoundation.org>
Cc: H. Peter Anvin <h...@zytor.com>
Cc: Josh Poimboeuf <jpoim...@redhat.com>
Cc: Juergen Gross <jgr...@suse.com>
Cc: Linus Torvalds <torva...@linux-foundation.org>
Cc: Will Deacon <will.dea...@arm.com>
Cc: aligu...@amazon.com
Cc: daniel.gr...@iaik.tugraz.at
Cc: hu...@google.com
Cc: keesc...@google.com
Cc: linux...@kvack.org
---
arch/x86/include/asm/tlbflush.h | 55 +++++++++++++++++++++++++++++++---------
1 file changed, 43 insertions(+), 12 deletions(-)
--- a/arch/x86/include/asm/tlbflush.h
+++ b/arch/x86/include/asm/tlbflush.h
@@ -13,16 +13,33 @@
#include <asm/pti.h>
#include <asm/processor-flags.h>
-static inline u64 inc_mm_tlb_gen(struct mm_struct *mm)
-{
- /*
- * Bump the generation count. This also serves as a full barrier
- * that synchronizes with switch_mm(): callers are required to order
- * their read of mm_cpumask after their writes to the paging
- * structures.
- */
- return atomic64_inc_return(&mm->context.tlb_gen);
-}
+/*
+ * The x86 feature is called PCID (Process Context IDentifier). It is similar
+ * to what is traditionally called ASID on the RISC processors.
+ *
+ * We don't use the traditional ASID implementation, where each process/mm gets
+ * its own ASID and flush/restart when we run out of ASID space.
+ *
+ * Instead we have a small per-cpu array of ASIDs and cache the last few mm's
+ * that came by on this CPU, allowing cheaper switch_mm between processes on
+ * this CPU.
+ *
+ * We end up with different spaces for different things. To avoid confusion we
+ * use different names for each of them:
+ *
+ * ASID - [0, TLB_NR_DYN_ASIDS-1]
+ * the canonical identifier for an mm
+ *
+ * kPCID - [1, TLB_NR_DYN_ASIDS]
+ * the value we write into the PCID part of CR3; corresponds to the
+ * ASID+1, because PCID 0 is special.
+ *
+ * uPCID - [2048 + 1, 2048 + TLB_NR_DYN_ASIDS]
+ * for KPTI each mm has two address spaces and thus needs two
+ * PCID values, but we can still do with a single ASID denomination
+ * for each mm. Corresponds to kPCID + 2048.
+ *
+ */
/* There are 12 bits of space for ASIDS in CR3 */
#define CR3_HW_ASID_BITS 12
@@ -41,7 +58,7 @@ static inline u64 inc_mm_tlb_gen(struct
/*
* ASIDs are zero-based: 0->MAX_AVAIL_ASID are valid. -1 below to account
- * for them being zero-based. Another -1 is because ASID 0 is reserved for
+ * for them being zero-based. Another -1 is because PCID 0 is reserved for
* use by non-PCID-aware users.
*/
#define MAX_ASID_AVAILABLE ((1 << CR3_AVAIL_PCID_BITS) - 2)
@@ -52,6 +69,9 @@ static inline u64 inc_mm_tlb_gen(struct
*/
#define TLB_NR_DYN_ASIDS 6
+/*
+ * Given @asid, compute kPCID
+ */
static inline u16 kern_pcid(u16 asid)
{
VM_WARN_ON_ONCE(asid > MAX_ASID_AVAILABLE);
@@ -86,7 +106,7 @@ static inline u16 kern_pcid(u16 asid)
}
/*
- * The user PCID is just the kernel one, plus the "switch bit".
+ * Given @asid, compute uPCID
*/
static inline u16 user_pcid(u16 asid)
{
@@ -487,6 +507,17 @@ static inline void flush_tlb_page(struct
void native_flush_tlb_others(const struct cpumask *cpumask,
const struct flush_tlb_info *info);
+static inline u64 inc_mm_tlb_gen(struct mm_struct *mm)
+{
+ /*
+ * Bump the generation count. This also serves as a full barrier
+ * that synchronizes with switch_mm(): callers are required to order
+ * their read of mm_cpumask after their writes to the paging
+ * structures.
+ */
+ return atomic64_inc_return(&mm->context.tlb_gen);
+}
+
static inline void arch_tlbbatch_add_mm(struct arch_tlbflush_unmap_batch
*batch,
struct mm_struct *mm)
{
