x86: Avoid lazy mm refcounting

XXX:
  arch/x86/power/cpu.c:   load_mm_ldt(current->active_mm);        /* This does lldt */

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>

arch/Kconfig

diff --git a/arch/Kconfig b/arch/Kconfig
index f4b210ab..fafa4e1 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -475,6 +475,26 @@
 	  irqs disabled over activate_mm. Architectures that do IPI based TLB
 	  shootdowns should enable this.
 
+# This option allows MMU_LAZY_TLB_REFCOUNT=n. It ensures no CPUs are using an
+# mm as a lazy tlb beyond its last reference count, by shooting down these
+# users before the mm is deallocated. __mmdrop() first IPIs all CPUs that may
+# be using the mm as a lazy tlb, so that they may switch themselves to using
+# init_mm for their active mm. mm_cpumask(mm) is used to determine which CPUs
+# may be using mm as a lazy tlb mm.
+#
+# To implement this, an arch *must*:
+# - At the time of the final mmdrop of the mm, ensure mm_cpumask(mm) contains
+#   at least all possible CPUs in which the mm is lazy.
+# - It must meet the requirements for MMU_LAZY_TLB_REFCOUNT=n (see above).
+config MMU_LAZY_TLB_SHOOTDOWN
+	bool
+
+# This option allows MMU_LAZY_TLB_REFCOUNT=n. It ensures no CPUs are using an
+# mm as a lazy tlb beyond its last reference count by relying on
+# tlb_gather_mmu_fullmm().
+config MMU_LAZY_TLB_IMPLICIT_SHOOTDOWN
+	bool
+
 # Use normal mm refcounting for MMU_LAZY_TLB kernel thread references.
 # MMU_LAZY_TLB_REFCOUNT=n can improve the scalability of context switching
 # to/from kernel threads when the same mm is running on a lot of CPUs (a large
@@ -491,21 +511,7 @@
 # converted already).
 config MMU_LAZY_TLB_REFCOUNT
 	def_bool y
-	depends on !MMU_LAZY_TLB_SHOOTDOWN
-
-# This option allows MMU_LAZY_TLB_REFCOUNT=n. It ensures no CPUs are using an
-# mm as a lazy tlb beyond its last reference count, by shooting down these
-# users before the mm is deallocated. __mmdrop() first IPIs all CPUs that may
-# be using the mm as a lazy tlb, so that they may switch themselves to using
-# init_mm for their active mm. mm_cpumask(mm) is used to determine which CPUs
-# may be using mm as a lazy tlb mm.
-#
-# To implement this, an arch *must*:
-# - At the time of the final mmdrop of the mm, ensure mm_cpumask(mm) contains
-#   at least all possible CPUs in which the mm is lazy.
-# - It must meet the requirements for MMU_LAZY_TLB_REFCOUNT=n (see above).
-config MMU_LAZY_TLB_SHOOTDOWN
-	bool
+	depends on !MMU_LAZY_TLB_SHOOTDOWN && !MMU_LAZY_TLB_IMPLICIT_SHOOTDOWN
 
 config ARCH_HAVE_NMI_SAFE_CMPXCHG
 	bool

arch/x86/Kconfig

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 76a2389..afc409f 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -305,6 +305,7 @@
 	select FUNCTION_ALIGNMENT_4B
 	imply IMA_SECURE_AND_OR_TRUSTED_BOOT    if EFI
 	select HAVE_DYNAMIC_FTRACE_NO_PATCHABLE
+	select MMU_LAZY_TLB_IMPLICIT_SHOOTDOWN
 
 config INSTRUCTION_DECODER
 	def_bool y

kernel/fork.c

diff --git a/kernel/fork.c b/kernel/fork.c
index 10917c3..5830e7a 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -861,7 +861,8 @@ static void do_shoot_lazy_tlb(void *arg)
 
 static void cleanup_lazy_tlbs(struct mm_struct *mm)
 {
-	if (!IS_ENABLED(CONFIG_MMU_LAZY_TLB_SHOOTDOWN)) {
+	if (!IS_ENABLED(CONFIG_MMU_LAZY_TLB_SHOOTDOWN) ||
+	    IS_ENABLED(CONFIG_MMU_LAZY_TLB_IMPLICIT_SHOOTDOWN)) {
 		/*
 		 * In this case, lazy tlb mms are refounted and would not reach
 		 * __mmdrop until all CPUs have switched away and mmdrop()ed.