Merge branch 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 page table isolation updates from Thomas Gleixner:
 "This is the final set of enabling page table isolation on x86:

   - Infrastructure patches for handling the extra page tables.

   - Patches which map the various bits and pieces which are required to
     get in and out of user space into the user space visible page
     tables.

   - The required changes to have CR3 switching in the entry/exit code.

   - Optimizations for the CR3 switching along with documentation how
     the ASID/PCID mechanism works.

   - Updates to dump pagetables to cover the user space page tables for
     W+X scans and extra debugfs files to analyze both the kernel and
     the user space visible page tables

  The whole functionality is compile time controlled via a config switch
  and can be turned on/off on the command line as well"

* 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (32 commits)
  x86/ldt: Make the LDT mapping RO
  x86/mm/dump_pagetables: Allow dumping current pagetables
  x86/mm/dump_pagetables: Check user space page table for WX pages
  x86/mm/dump_pagetables: Add page table directory to the debugfs VFS hierarchy
  x86/mm/pti: Add Kconfig
  x86/dumpstack: Indicate in Oops whether PTI is configured and enabled
  x86/mm: Clarify the whole ASID/kernel PCID/user PCID naming
  x86/mm: Use INVPCID for __native_flush_tlb_single()
  x86/mm: Optimize RESTORE_CR3
  x86/mm: Use/Fix PCID to optimize user/kernel switches
  x86/mm: Abstract switching CR3
  x86/mm: Allow flushing for future ASID switches
  x86/pti: Map the vsyscall page if needed
  x86/pti: Put the LDT in its own PGD if PTI is on
  x86/mm/64: Make a full PGD-entry size hole in the memory map
  x86/events/intel/ds: Map debug buffers in cpu_entry_area
  x86/cpu_entry_area: Add debugstore entries to cpu_entry_area
  x86/mm/pti: Map ESPFIX into user space
  x86/mm/pti: Share entry text PMD
  x86/entry: Align entry text section to PMD boundary
  ...

16 files changed, 462 insertions, 53 deletions

diff --git a/arch/x86/include/asm/cpu_entry_area.h b/arch/x86/include/asm/cpu_entry_area.h
index 2fbc69a0916ed..4a7884b8dca55 100644
--- a/arch/x86/include/asm/cpu_entry_area.h
+++ b/arch/x86/include/asm/cpu_entry_area.h
@@ -5,6 +5,7 @@
 
 #include <linux/percpu-defs.h>
 #include <asm/processor.h>
+#include <asm/intel_ds.h>
 
 /*
  * cpu_entry_area is a percpu region that contains things needed by the CPU
@@ -40,6 +41,18 @@ struct cpu_entry_area {
 	 */
 	char exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ + DEBUG_STKSZ];
 #endif
+#ifdef CONFIG_CPU_SUP_INTEL
+	/*
+	 * Per CPU debug store for Intel performance monitoring. Wastes a
+	 * full page at the moment.
+	 */
+	struct debug_store cpu_debug_store;
+	/*
+	 * The actual PEBS/BTS buffers must be mapped to user space
+	 * Reserve enough fixmap PTEs.
+	 */
+	struct debug_store_buffers cpu_debug_buffers;
+#endif
 };
 
 #define CPU_ENTRY_AREA_SIZE	(sizeof(struct cpu_entry_area))
diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h
index 800104c8a3edf..07cdd17157050 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -197,11 +197,12 @@
 #define X86_FEATURE_CAT_L3		( 7*32+ 4) /* Cache Allocation Technology L3 */
 #define X86_FEATURE_CAT_L2		( 7*32+ 5) /* Cache Allocation Technology L2 */
 #define X86_FEATURE_CDP_L3		( 7*32+ 6) /* Code and Data Prioritization L3 */
+#define X86_FEATURE_INVPCID_SINGLE	( 7*32+ 7) /* Effectively INVPCID && CR4.PCIDE=1 */
 
 #define X86_FEATURE_HW_PSTATE		( 7*32+ 8) /* AMD HW-PState */
 #define X86_FEATURE_PROC_FEEDBACK	( 7*32+ 9) /* AMD ProcFeedbackInterface */
 #define X86_FEATURE_SME			( 7*32+10) /* AMD Secure Memory Encryption */
-
+#define X86_FEATURE_PTI			( 7*32+11) /* Kernel Page Table Isolation enabled */
 #define X86_FEATURE_INTEL_PPIN		( 7*32+14) /* Intel Processor Inventory Number */
 #define X86_FEATURE_INTEL_PT		( 7*32+15) /* Intel Processor Trace */
 #define X86_FEATURE_AVX512_4VNNIW	( 7*32+16) /* AVX-512 Neural Network Instructions */
@@ -340,5 +341,6 @@
 #define X86_BUG_SWAPGS_FENCE		X86_BUG(11) /* SWAPGS without input dep on GS */
 #define X86_BUG_MONITOR			X86_BUG(12) /* IPI required to wake up remote CPU */
 #define X86_BUG_AMD_E400		X86_BUG(13) /* CPU is among the affected by Erratum 400 */
+#define X86_BUG_CPU_INSECURE		X86_BUG(14) /* CPU is insecure and needs kernel page table isolation */
 
 #endif /* _ASM_X86_CPUFEATURES_H */
diff --git a/arch/x86/include/asm/desc.h b/arch/x86/include/asm/desc.h
index ec8be07c0cda5..13c5ee878a477 100644
--- a/arch/x86/include/asm/desc.h
+++ b/arch/x86/include/asm/desc.h
@@ -21,6 +21,8 @@ static inline void fill_ldt(struct desc_struct *desc, const struct user_desc *in
 
 	desc->type		= (info->read_exec_only ^ 1) << 1;
 	desc->type	       |= info->contents << 2;
+	/* Set the ACCESS bit so it can be mapped RO */
+	desc->type	       |= 1;
 
 	desc->s			= 1;
 	desc->dpl		= 0x3;
diff --git a/arch/x86/include/asm/disabled-features.h b/arch/x86/include/asm/disabled-features.h
index 14d6d50073142..b027633e73003 100644
--- a/arch/x86/include/asm/disabled-features.h
+++ b/arch/x86/include/asm/disabled-features.h
@@ -50,6 +50,12 @@
 # define DISABLE_LA57	(1<<(X86_FEATURE_LA57 & 31))
 #endif
 
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+# define DISABLE_PTI		0
+#else
+# define DISABLE_PTI		(1 << (X86_FEATURE_PTI & 31))
+#endif
+
 /*
  * Make sure to add features to the correct mask
  */
@@ -60,7 +66,7 @@
 #define DISABLED_MASK4	(DISABLE_PCID)
 #define DISABLED_MASK5	0
 #define DISABLED_MASK6	0
-#define DISABLED_MASK7	0
+#define DISABLED_MASK7	(DISABLE_PTI)
 #define DISABLED_MASK8	0
 #define DISABLED_MASK9	(DISABLE_MPX)
 #define DISABLED_MASK10	0
diff --git a/arch/x86/include/asm/intel_ds.h b/arch/x86/include/asm/intel_ds.h
new file mode 100644
index 0000000000000..62a9f4966b429
--- /dev/null
+++ b/arch/x86/include/asm/intel_ds.h
@@ -0,0 +1,36 @@
+#ifndef _ASM_INTEL_DS_H
+#define _ASM_INTEL_DS_H
+
+#include <linux/percpu-defs.h>
+
+#define BTS_BUFFER_SIZE		(PAGE_SIZE << 4)
+#define PEBS_BUFFER_SIZE	(PAGE_SIZE << 4)
+
+/* The maximal number of PEBS events: */
+#define MAX_PEBS_EVENTS		8
+
+/*
+ * A debug store configuration.
+ *
+ * We only support architectures that use 64bit fields.
+ */
+struct debug_store {
+	u64	bts_buffer_base;
+	u64	bts_index;
+	u64	bts_absolute_maximum;
+	u64	bts_interrupt_threshold;
+	u64	pebs_buffer_base;
+	u64	pebs_index;
+	u64	pebs_absolute_maximum;
+	u64	pebs_interrupt_threshold;
+	u64	pebs_event_reset[MAX_PEBS_EVENTS];
+} __aligned(PAGE_SIZE);
+
+DECLARE_PER_CPU_PAGE_ALIGNED(struct debug_store, cpu_debug_store);
+
+struct debug_store_buffers {
+	char	bts_buffer[BTS_BUFFER_SIZE];
+	char	pebs_buffer[PEBS_BUFFER_SIZE];
+};
+
+#endif
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 5ede7cae1d673..c931b88982a0f 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -50,10 +50,33 @@ struct ldt_struct {
 	 * call gates.  On native, we could merge the ldt_struct and LDT
 	 * allocations, but it's not worth trying to optimize.
 	 */
-	struct desc_struct *entries;
-	unsigned int nr_entries;
+	struct desc_struct	*entries;
+	unsigned int		nr_entries;
+
+	/*
+	 * If PTI is in use, then the entries array is not mapped while we're
+	 * in user mode.  The whole array will be aliased at the addressed
+	 * given by ldt_slot_va(slot).  We use two slots so that we can allocate
+	 * and map, and enable a new LDT without invalidating the mapping
+	 * of an older, still-in-use LDT.
+	 *
+	 * slot will be -1 if this LDT doesn't have an alias mapping.
+	 */
+	int			slot;
 };
 
+/* This is a multiple of PAGE_SIZE. */
+#define LDT_SLOT_STRIDE (LDT_ENTRIES * LDT_ENTRY_SIZE)
+
+static inline void *ldt_slot_va(int slot)
+{
+#ifdef CONFIG_X86_64
+	return (void *)(LDT_BASE_ADDR + LDT_SLOT_STRIDE * slot);
+#else
+	BUG();
+#endif
+}
+
 /*
  * Used for LDT copy/destruction.
  */
@@ -64,6 +87,7 @@ static inline void init_new_context_ldt(struct mm_struct *mm)
 }
 int ldt_dup_context(struct mm_struct *oldmm, struct mm_struct *mm);
 void destroy_context_ldt(struct mm_struct *mm);
+void ldt_arch_exit_mmap(struct mm_struct *mm);
 #else	/* CONFIG_MODIFY_LDT_SYSCALL */
 static inline void init_new_context_ldt(struct mm_struct *mm) { }
 static inline int ldt_dup_context(struct mm_struct *oldmm,
@@ -71,7 +95,8 @@ static inline int ldt_dup_context(struct mm_struct *oldmm,
 {
 	return 0;
 }
-static inline void destroy_context_ldt(struct mm_struct *mm) {}
+static inline void destroy_context_ldt(struct mm_struct *mm) { }
+static inline void ldt_arch_exit_mmap(struct mm_struct *mm) { }
 #endif
 
 static inline void load_mm_ldt(struct mm_struct *mm)
@@ -96,10 +121,31 @@ static inline void load_mm_ldt(struct mm_struct *mm)
 	 * that we can see.
 	 */
 
-	if (unlikely(ldt))
-		set_ldt(ldt->entries, ldt->nr_entries);
-	else
+	if (unlikely(ldt)) {
+		if (static_cpu_has(X86_FEATURE_PTI)) {
+			if (WARN_ON_ONCE((unsigned long)ldt->slot > 1)) {
+				/*
+				 * Whoops -- either the new LDT isn't mapped
+				 * (if slot == -1) or is mapped into a bogus
+				 * slot (if slot > 1).
+				 */
+				clear_LDT();
+				return;
+			}
+
+			/*
+			 * If page table isolation is enabled, ldt->entries
+			 * will not be mapped in the userspace pagetables.
+			 * Tell the CPU to access the LDT through the alias
+			 * at ldt_slot_va(ldt->slot).
+			 */
+			set_ldt(ldt_slot_va(ldt->slot), ldt->nr_entries);
+		} else {
+			set_ldt(ldt->entries, ldt->nr_entries);
+		}
+	} else {
 		clear_LDT();
+	}
 #else
 	clear_LDT();
 #endif
@@ -194,6 +240,7 @@ static inline int arch_dup_mmap(struct mm_struct *oldmm, struct mm_struct *mm)
 static inline void arch_exit_mmap(struct mm_struct *mm)
 {
 	paravirt_arch_exit_mmap(mm);
+	ldt_arch_exit_mmap(mm);
 }
 
 #ifdef CONFIG_X86_64
diff --git a/arch/x86/include/asm/pgalloc.h b/arch/x86/include/asm/pgalloc.h
index 4b5e1eafada73..aff42e1da6ee1 100644
--- a/arch/x86/include/asm/pgalloc.h
+++ b/arch/x86/include/asm/pgalloc.h
@@ -30,6 +30,17 @@ static inline void paravirt_release_p4d(unsigned long pfn) {}
  */
 extern gfp_t __userpte_alloc_gfp;
 
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+/*
+ * Instead of one PGD, we acquire two PGDs.  Being order-1, it is
+ * both 8k in size and 8k-aligned.  That lets us just flip bit 12
+ * in a pointer to swap between the two 4k halves.
+ */
+#define PGD_ALLOCATION_ORDER 1
+#else
+#define PGD_ALLOCATION_ORDER 0
+#endif
+
 /*
  * Allocate and free page tables.
  */
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 95e2dfd755218..e42b8943cb1a3 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -28,6 +28,7 @@ extern pgd_t early_top_pgt[PTRS_PER_PGD];
 int __init __early_make_pgtable(unsigned long address, pmdval_t pmd);
 
 void ptdump_walk_pgd_level(struct seq_file *m, pgd_t *pgd);
+void ptdump_walk_pgd_level_debugfs(struct seq_file *m, pgd_t *pgd, bool user);
 void ptdump_walk_pgd_level_checkwx(void);
 
 #ifdef CONFIG_DEBUG_WX
@@ -841,7 +842,12 @@ static inline pud_t *pud_offset(p4d_t *p4d, unsigned long address)
 
 static inline int p4d_bad(p4d_t p4d)
 {
-	return (p4d_flags(p4d) & ~(_KERNPG_TABLE | _PAGE_USER)) != 0;
+	unsigned long ignore_flags = _KERNPG_TABLE | _PAGE_USER;
+
+	if (IS_ENABLED(CONFIG_PAGE_TABLE_ISOLATION))
+		ignore_flags |= _PAGE_NX;
+
+	return (p4d_flags(p4d) & ~ignore_flags) != 0;
 }
 #endif  /* CONFIG_PGTABLE_LEVELS > 3 */
 
@@ -875,7 +881,12 @@ static inline p4d_t *p4d_offset(pgd_t *pgd, unsigned long address)
 
 static inline int pgd_bad(pgd_t pgd)
 {
-	return (pgd_flags(pgd) & ~_PAGE_USER) != _KERNPG_TABLE;
+	unsigned long ignore_flags = _PAGE_USER;
+
+	if (IS_ENABLED(CONFIG_PAGE_TABLE_ISOLATION))
+		ignore_flags |= _PAGE_NX;
+
+	return (pgd_flags(pgd) & ~ignore_flags) != _KERNPG_TABLE;
 }
 
 static inline int pgd_none(pgd_t pgd)
@@ -904,7 +915,11 @@ static inline int pgd_none(pgd_t pgd)
  * pgd_offset() returns a (pgd_t *)
  * pgd_index() is used get the offset into the pgd page's array of pgd_t's;
  */
-#define pgd_offset(mm, address) ((mm)->pgd + pgd_index((address)))
+#define pgd_offset_pgd(pgd, address) (pgd + pgd_index((address)))
+/*
+ * a shortcut to get a pgd_t in a given mm
+ */
+#define pgd_offset(mm, address) pgd_offset_pgd((mm)->pgd, (address))
 /*
  * a shortcut which implies the use of the kernel's pgd, instead
  * of a process's
@@ -1106,7 +1121,14 @@ static inline int pud_write(pud_t pud)
  */
 static inline void clone_pgd_range(pgd_t *dst, pgd_t *src, int count)
 {
-       memcpy(dst, src, count * sizeof(pgd_t));
+	memcpy(dst, src, count * sizeof(pgd_t));
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+	if (!static_cpu_has(X86_FEATURE_PTI))
+		return;
+	/* Clone the user space pgd as well */
+	memcpy(kernel_to_user_pgdp(dst), kernel_to_user_pgdp(src),
+	       count * sizeof(pgd_t));
+#endif
 }
 
 #define PTE_SHIFT ilog2(PTRS_PER_PTE)
diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h
index e9f05331e732a..81462e9a34f6a 100644
--- a/arch/x86/include/asm/pgtable_64.h
+++ b/arch/x86/include/asm/pgtable_64.h
@@ -131,9 +131,97 @@ static inline pud_t native_pudp_get_and_clear(pud_t *xp)
 #endif
 }
 
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+/*
+ * All top-level PAGE_TABLE_ISOLATION page tables are order-1 pages
+ * (8k-aligned and 8k in size).  The kernel one is at the beginning 4k and
+ * the user one is in the last 4k.  To switch between them, you
+ * just need to flip the 12th bit in their addresses.
+ */
+#define PTI_PGTABLE_SWITCH_BIT	PAGE_SHIFT
+
+/*
+ * This generates better code than the inline assembly in
+ * __set_bit().
+ */
+static inline void *ptr_set_bit(void *ptr, int bit)
+{
+	unsigned long __ptr = (unsigned long)ptr;
+
+	__ptr |= BIT(bit);
+	return (void *)__ptr;
+}
+static inline void *ptr_clear_bit(void *ptr, int bit)
+{
+	unsigned long __ptr = (unsigned long)ptr;
+
+	__ptr &= ~BIT(bit);
+	return (void *)__ptr;
+}
+
+static inline pgd_t *kernel_to_user_pgdp(pgd_t *pgdp)
+{
+	return ptr_set_bit(pgdp, PTI_PGTABLE_SWITCH_BIT);
+}
+
+static inline pgd_t *user_to_kernel_pgdp(pgd_t *pgdp)
+{
+	return ptr_clear_bit(pgdp, PTI_PGTABLE_SWITCH_BIT);
+}
+
+static inline p4d_t *kernel_to_user_p4dp(p4d_t *p4dp)
+{
+	return ptr_set_bit(p4dp, PTI_PGTABLE_SWITCH_BIT);
+}
+
+static inline p4d_t *user_to_kernel_p4dp(p4d_t *p4dp)
+{
+	return ptr_clear_bit(p4dp, PTI_PGTABLE_SWITCH_BIT);
+}
+#endif /* CONFIG_PAGE_TABLE_ISOLATION */
+
+/*
+ * Page table pages are page-aligned.  The lower half of the top
+ * level is used for userspace and the top half for the kernel.
+ *
+ * Returns true for parts of the PGD that map userspace and
+ * false for the parts that map the kernel.
+ */
+static inline bool pgdp_maps_userspace(void *__ptr)
+{
+	unsigned long ptr = (unsigned long)__ptr;
+
+	return (ptr & ~PAGE_MASK) < (PAGE_SIZE / 2);
+}
+
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+pgd_t __pti_set_user_pgd(pgd_t *pgdp, pgd_t pgd);
+
+/*
+ * Take a PGD location (pgdp) and a pgd value that needs to be set there.
+ * Populates the user and returns the resulting PGD that must be set in
+ * the kernel copy of the page tables.
+ */
+static inline pgd_t pti_set_user_pgd(pgd_t *pgdp, pgd_t pgd)
+{
+	if (!static_cpu_has(X86_FEATURE_PTI))
+		return pgd;
+	return __pti_set_user_pgd(pgdp, pgd);
+}
+#else
+static inline pgd_t pti_set_user_pgd(pgd_t *pgdp, pgd_t pgd)
+{
+	return pgd;
+}
+#endif
+
 static inline void native_set_p4d(p4d_t *p4dp, p4d_t p4d)
 {
+#if defined(CONFIG_PAGE_TABLE_ISOLATION) && !defined(CONFIG_X86_5LEVEL)
+	p4dp->pgd = pti_set_user_pgd(&p4dp->pgd, p4d.pgd);
+#else
 	*p4dp = p4d;
+#endif
 }
 
 static inline void native_p4d_clear(p4d_t *p4d)
@@ -147,7 +235,11 @@ static inline void native_p4d_clear(p4d_t *p4d)
 
 static inline void native_set_pgd(pgd_t *pgdp, pgd_t pgd)
 {
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+	*pgdp = pti_set_user_pgd(pgdp, pgd);
+#else
 	*pgdp = pgd;
+#endif
 }
 
 static inline void native_pgd_clear(pgd_t *pgd)
diff --git a/arch/x86/include/asm/pgtable_64_types.h b/arch/x86/include/asm/pgtable_64_types.h
index 3d27831bc58df..b97a539bcdeee 100644
--- a/arch/x86/include/asm/pgtable_64_types.h
+++ b/arch/x86/include/asm/pgtable_64_types.h
@@ -79,13 +79,17 @@ typedef struct { pteval_t pte; } pte_t;
 #define MAXMEM			_AC(__AC(1, UL) << MAX_PHYSMEM_BITS, UL)
 
 #ifdef CONFIG_X86_5LEVEL
-# define VMALLOC_SIZE_TB	_AC(16384, UL)
-# define __VMALLOC_BASE		_AC(0xff92000000000000, UL)
+# define VMALLOC_SIZE_TB	_AC(12800, UL)
+# define __VMALLOC_BASE		_AC(0xffa0000000000000, UL)
 # define __VMEMMAP_BASE		_AC(0xffd4000000000000, UL)
+# define LDT_PGD_ENTRY		_AC(-112, UL)
+# define LDT_BASE_ADDR		(LDT_PGD_ENTRY << PGDIR_SHIFT)
 #else
 # define VMALLOC_SIZE_TB	_AC(32, UL)
 # define __VMALLOC_BASE		_AC(0xffffc90000000000, UL)
 # define __VMEMMAP_BASE		_AC(0xffffea0000000000, UL)
+# define LDT_PGD_ENTRY		_AC(-4, UL)
+# define LDT_BASE_ADDR		(LDT_PGD_ENTRY << PGDIR_SHIFT)
 #endif
 
 #ifdef CONFIG_RANDOMIZE_MEMORY
diff --git a/arch/x86/include/asm/processor-flags.h b/arch/x86/include/asm/processor-flags.h
index 43212a43ee69f..6a60fea90b9d9 100644
--- a/arch/x86/include/asm/processor-flags.h
+++ b/arch/x86/include/asm/processor-flags.h
@@ -38,6 +38,11 @@
 #define CR3_ADDR_MASK	__sme_clr(0x7FFFFFFFFFFFF000ull)
 #define CR3_PCID_MASK	0xFFFull
 #define CR3_NOFLUSH	BIT_ULL(63)
+
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+# define X86_CR3_PTI_SWITCH_BIT	11
+#endif
+
 #else
 /*
  * CR3_ADDR_MASK needs at least bits 31:5 set on PAE systems, and we save
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index cad8dab266bce..d3a67fba200ae 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -852,13 +852,22 @@ static inline void spin_lock_prefetch(const void *x)
 
 #else
 /*
- * User space process size. 47bits minus one guard page.  The guard
- * page is necessary on Intel CPUs: if a SYSCALL instruction is at
- * the highest possible canonical userspace address, then that
- * syscall will enter the kernel with a non-canonical return
- * address, and SYSRET will explode dangerously.  We avoid this
- * particular problem by preventing anything from being mapped
- * at the maximum canonical address.
+ * User space process size.  This is the first address outside the user range.
+ * There are a few constraints that determine this:
+ *
+ * On Intel CPUs, if a SYSCALL instruction is at the highest canonical
+ * address, then that syscall will enter the kernel with a
+ * non-canonical return address, and SYSRET will explode dangerously.
+ * We avoid this particular problem by preventing anything executable
+ * from being mapped at the maximum canonical address.
+ *
+ * On AMD CPUs in the Ryzen family, there's a nasty bug in which the
+ * CPUs malfunction if they execute code from the highest canonical page.
+ * They'll speculate right off the end of the canonical space, and
+ * bad things happen.  This is worked around in the same way as the
+ * Intel problem.
+ *
+ * With page table isolation enabled, we map the LDT in ... [stay tuned]
  */
 #define TASK_SIZE_MAX	((1UL << __VIRTUAL_MASK_SHIFT) - PAGE_SIZE)
 
diff --git a/arch/x86/include/asm/pti.h b/arch/x86/include/asm/pti.h
new file mode 100644
index 0000000000000..0b5ef05b2d2d9
--- /dev/null
+++ b/arch/x86/include/asm/pti.h
@@ -0,0 +1,14 @@
+// SPDX-License-Identifier: GPL-2.0
+#ifndef _ASM_X86_PTI_H
+#define _ASM_X86_PTI_H
+#ifndef __ASSEMBLY__
+
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+extern void pti_init(void);
+extern void pti_check_boottime_disable(void);
+#else
+static inline void pti_check_boottime_disable(void) { }
+#endif
+
+#endif /* __ASSEMBLY__ */
+#endif /* _ASM_X86_PTI_H */
diff --git a/arch/x86/include/asm/tlbflush.h b/arch/x86/include/asm/tlbflush.h
index e1884cf35257b..f68f9c836cca0 100644
--- a/arch/x86/include/asm/tlbflush.h
+++ b/arch/x86/include/asm/tlbflush.h
@@ -10,38 +10,90 @@
 #include <asm/special_insns.h>
 #include <asm/smp.h>
 #include <asm/invpcid.h>
+#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
+
 /*
  * When enabled, PAGE_TABLE_ISOLATION consumes a single bit for
  * user/kernel switches
  */
-#define PTI_CONSUMED_ASID_BITS		0
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+# define PTI_CONSUMED_PCID_BITS	1
+#else
+# define PTI_CONSUMED_PCID_BITS	0
+#endif
+
+#define CR3_AVAIL_PCID_BITS (X86_CR3_PCID_BITS - PTI_CONSUMED_PCID_BITS)
 
-#define CR3_AVAIL_ASID_BITS (CR3_HW_ASID_BITS - PTI_CONSUMED_ASID_BITS)
 /*
  * 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_ASID_BITS) - 2)
+#define MAX_ASID_AVAILABLE ((1 << CR3_AVAIL_PCID_BITS) - 2)
 
+/*
+ * 6 because 6 should be plenty and struct tlb_state will fit in two cache
+ * lines.
+ */
+#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);
+
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+	/*
+	 * Make sure that the dynamic ASID space does not confict with the
+	 * bit we are using to switch between user and kernel ASIDs.
+	 */
+	BUILD_BUG_ON(TLB_NR_DYN_ASIDS >= (1 << X86_CR3_PTI_SWITCH_BIT));
+
+	/*
+	 * The ASID being passed in here should have respected the
+	 * MAX_ASID_AVAILABLE and thus never have the switch bit set.
+	 */
+	VM_WARN_ON_ONCE(asid & (1 << X86_CR3_PTI_SWITCH_BIT));
+#endif
 	/*
+	 * The dynamically-assigned ASIDs that get passed in are small
+	 * (<TLB_NR_DYN_ASIDS).  They never have the high switch bit set,
+	 * so do not bother to clear it.
+	 *
 	 * If PCID is on, ASID-aware code paths put the ASID+1 into the
 	 * PCID bits.  This serves two purposes.  It prevents a nasty
 	 * situation in which PCID-unaware code saves CR3, loads some other
@@ -53,6 +105,18 @@ static inline u16 kern_pcid(u16 asid)
 	return asid + 1;
 }
 
+/*
+ * Given @asid, compute uPCID
+ */
+static inline u16 user_pcid(u16 asid)
+{
+	u16 ret = kern_pcid(asid);
+#ifdef CONFIG_PAGE_TABLE_ISOLATION
+	ret |= 1 << X86_CR3_PTI_SWITCH_BIT;
+#endif
+	return ret;
+}
+
 struct pgd_t;
 static inline unsigned long build_cr3(pgd_t *pgd, u16 asid)
 {
@@ -95,12 +159,6 @@ static inline bool tlb_defer_switch_to_init_mm(void)
 	return !static_cpu_has(X86_FEATURE_PCID);
 }
 
-/*
- * 6 because 6 should be plenty and struct tlb_state will fit in
- * two cache lines.
- */
-#define TLB_NR_DYN_ASIDS 6
-
 struct tlb_context {
 	u64 ctx_id;
 	u64 tlb_gen;
@@ -135,6 +193,24 @@ struct tlb_state {
 	bool is_lazy;
 
 	/*
+	 * If set we changed the page tables in such a way that we
+	 * needed an invalidation of all contexts (aka. PCIDs / ASIDs).
+	 * This tells us to go invalidate all the non-loaded ctxs[]
+	 * on the next context switch.
+	 *
+	 * The current ctx was kept up-to-date as it ran and does not
+	 * need to be invalidated.
+	 */
+	bool invalidate_other;
+
+	/*
+	 * Mask that contains TLB_NR_DYN_ASIDS+1 bits to indicate
+	 * the corresponding user PCID needs a flush next time we
+	 * switch to it; see SWITCH_TO_USER_CR3.
+	 */
+	unsigned short user_pcid_flush_mask;
+
+	/*
 	 * Access to this CR4 shadow and to H/W CR4 is protected by
 	 * disabling interrupts when modifying either one.
 	 */
@@ -215,6 +291,14 @@ static inline unsigned long cr4_read_shadow(void)
 }
 
 /*
+ * Mark all other ASIDs as invalid, preserves the current.
+ */
+static inline void invalidate_other_asid(void)
+{
+	this_cpu_write(cpu_tlbstate.invalidate_other, true);
+}
+
+/*
  * Save some of cr4 feature set we're using (e.g.  Pentium 4MB
  * enable and PPro Global page enable), so that any CPU's that boot
  * up after us can get the correct flags.  This should only be used
@@ -234,14 +318,41 @@ static inline void cr4_set_bits_and_update_boot(unsigned long mask)
 extern void initialize_tlbstate_and_flush(void);
 
 /*
+ * Given an ASID, flush the corresponding user ASID.  We can delay this
+ * until the next time we switch to it.
+ *
+ * See SWITCH_TO_USER_CR3.
+ */
+static inline void invalidate_user_asid(u16 asid)
+{
+	/* There is no user ASID if address space separation is off */
+	if (!IS_ENABLED(CONFIG_PAGE_TABLE_ISOLATION))
+		return;
+
+	/*
+	 * We only have a single ASID if PCID is off and the CR3
+	 * write will have flushed it.
+	 */
+	if (!cpu_feature_enabled(X86_FEATURE_PCID))
+		return;
+
+	if (!static_cpu_has(X86_FEATURE_PTI))
+		return;
+
+	__set_bit(kern_pcid(asid),
+		  (unsigned long *)this_cpu_ptr(&cpu_tlbstate.user_pcid_flush_mask));
+}
+
+/*
  * flush the entire current user mapping
  */
 static inline void __native_flush_tlb(void)
 {
+	invalidate_user_asid(this_cpu_read(cpu_tlbstate.loaded_mm_asid));
 	/*
-	 * If current->mm == NULL then we borrow a mm which may change during a
-	 * task switch and therefore we must not be preempted while we write CR3
-	 * back:
+	 * If current->mm == NULL then we borrow a mm which may change
+	 * during a task switch and therefore we must not be preempted
+	 * while we write CR3 back:
 	 */
 	preempt_disable();
 	native_write_cr3(__native_read_cr3());
@@ -259,6 +370,8 @@ static inline void __native_flush_tlb_global(void)
 		/*
 		 * Using INVPCID is considerably faster than a pair of writes
 		 * to CR4 sandwiched inside an IRQ flag save/restore.
+		 *
+		 * Note, this works with CR4.PCIDE=0 or 1.
 		 */
 		invpcid_flush_all();
 		return;
@@ -285,7 +398,21 @@ static inline void __native_flush_tlb_global(void)
  */
 static inline void __native_flush_tlb_single(unsigned long addr)
 {
+	u32 loaded_mm_asid = this_cpu_read(cpu_tlbstate.loaded_mm_asid);
+
 	asm volatile("invlpg (%0)" ::"r" (addr) : "memory");
+
+	if (!static_cpu_has(X86_FEATURE_PTI))
+		return;
+
+	/*
+	 * Some platforms #GP if we call invpcid(type=1/2) before CR4.PCIDE=1.
+	 * Just use invalidate_user_asid() in case we are called early.
+	 */
+	if (!this_cpu_has(X86_FEATURE_INVPCID_SINGLE))
+		invalidate_user_asid(loaded_mm_asid);
+	else
+		invpcid_flush_one(user_pcid(loaded_mm_asid), addr);
 }
 
 /*
@@ -301,14 +428,6 @@ static inline void __flush_tlb_all(void)
 		 */
 		__flush_tlb();
 	}
-
-	/*
-	 * Note: if we somehow had PCID but not PGE, then this wouldn't work --
-	 * we'd end up flushing kernel translations for the current ASID but
-	 * we might fail to flush kernel translations for other cached ASIDs.
-	 *
-	 * To avoid this issue, we force PCID off if PGE is off.
-	 */
 }
 
 /*
@@ -318,6 +437,16 @@ static inline void __flush_tlb_one(unsigned long addr)
 {
 	count_vm_tlb_event(NR_TLB_LOCAL_FLUSH_ONE);
 	__flush_tlb_single(addr);
+
+	if (!static_cpu_has(X86_FEATURE_PTI))
+		return;
+
+	/*
+	 * __flush_tlb_single() will have cleared the TLB entry for this ASID,
+	 * but since kernel space is replicated across all, we must also
+	 * invalidate all others.
+	 */
+	invalidate_other_asid();
 }
 
 #define TLB_FLUSH_ALL	-1UL
@@ -378,6 +507,17 @@ static inline void flush_tlb_page(struct vm_area_struct *vma, unsigned long a)
 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)
 {
diff --git a/arch/x86/include/asm/vsyscall.h b/arch/x86/include/asm/vsyscall.h
index d9a7c659009c9..b986b2ca688a0 100644
--- a/arch/x86/include/asm/vsyscall.h
+++ b/arch/x86/include/asm/vsyscall.h
@@ -7,6 +7,7 @@
 
 #ifdef CONFIG_X86_VSYSCALL_EMULATION
 extern void map_vsyscall(void);
+extern void set_vsyscall_pgtable_user_bits(pgd_t *root);
 
 /*
  * Called on instruction fetch fault in vsyscall page.
diff --git a/arch/x86/include/uapi/asm/processor-flags.h b/arch/x86/include/uapi/asm/processor-flags.h
index 7e1e730396ae0..bcba3c643e63d 100644
--- a/arch/x86/include/uapi/asm/processor-flags.h
+++ b/arch/x86/include/uapi/asm/processor-flags.h
@@ -78,7 +78,12 @@
 #define X86_CR3_PWT		_BITUL(X86_CR3_PWT_BIT)
 #define X86_CR3_PCD_BIT		4 /* Page Cache Disable */
 #define X86_CR3_PCD		_BITUL(X86_CR3_PCD_BIT)
-#define X86_CR3_PCID_MASK	_AC(0x00000fff,UL) /* PCID Mask */
+
+#define X86_CR3_PCID_BITS	12
+#define X86_CR3_PCID_MASK	(_AC((1UL << X86_CR3_PCID_BITS) - 1, UL))
+
+#define X86_CR3_PCID_NOFLUSH_BIT 63 /* Preserve old PCID */
+#define X86_CR3_PCID_NOFLUSH    _BITULL(X86_CR3_PCID_NOFLUSH_BIT)
 
 /*
  * Intel CPU features in CR4