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

Pull x86 mm updates from Ingo Molnar:
 "The main changes in this cycle were:

   - Update and clean up x86 fault handling, by Andy Lutomirski.

   - Drop usage of __flush_tlb_all() in kernel_physical_mapping_init()
     and related fallout, by Dan Williams.

   - CPA cleanups and reorganization by Peter Zijlstra: simplify the
     flow and remove a few warts.

   - Other misc cleanups"

* 'x86-mm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (29 commits)
  x86/mm/dump_pagetables: Use DEFINE_SHOW_ATTRIBUTE()
  x86/mm/cpa: Rename @addrinarray to @numpages
  x86/mm/cpa: Better use CLFLUSHOPT
  x86/mm/cpa: Fold cpa_flush_range() and cpa_flush_array() into a single cpa_flush() function
  x86/mm/cpa: Make cpa_data::numpages invariant
  x86/mm/cpa: Optimize cpa_flush_array() TLB invalidation
  x86/mm/cpa: Simplify the code after making cpa->vaddr invariant
  x86/mm/cpa: Make cpa_data::vaddr invariant
  x86/mm/cpa: Add __cpa_addr() helper
  x86/mm/cpa: Add ARRAY and PAGES_ARRAY selftests
  x86/mm: Drop usage of __flush_tlb_all() in kernel_physical_mapping_init()
  x86/mm: Validate kernel_physical_mapping_init() PTE population
  generic/pgtable: Introduce set_pte_safe()
  generic/pgtable: Introduce {p4d,pgd}_same()
  generic/pgtable: Make {pmd, pud}_same() unconditionally available
  x86/fault: Clean up the page fault oops decoder a bit
  x86/fault: Decode page fault OOPSes better
  x86/vsyscall/64: Use X86_PF constants in the simulated #PF error code
  x86/oops: Show the correct CS value in show_regs()
  x86/fault: Don't try to recover from an implicit supervisor access
  ...

arch/x86/entry/vsyscall/vsyscall_64.c

@@ -102,7 +102,7 @@ static bool write_ok_or_segv(unsigned long ptr, size_t size)
 	if (!access_ok(VERIFY_WRITE, (void __user *)ptr, size)) {
 		struct thread_struct *thread = &current->thread;
 
-		thread->error_code	= 6;  /* user fault, no page, write */
+		thread->error_code	= X86_PF_USER | X86_PF_WRITE;
 		thread->cr2		= ptr;
 		thread->trap_nr		= X86_TRAP_PF;
 

arch/x86/include/asm/disabled-features.h

@@ -16,6 +16,12 @@
 # define DISABLE_MPX	(1<<(X86_FEATURE_MPX & 31))
 #endif
 
+#ifdef CONFIG_X86_SMAP
+# define DISABLE_SMAP	0
+#else
+# define DISABLE_SMAP	(1<<(X86_FEATURE_SMAP & 31))
+#endif
+
 #ifdef CONFIG_X86_INTEL_UMIP
 # define DISABLE_UMIP	0
 #else
@@ -68,7 +74,7 @@
 #define DISABLED_MASK6	0
 #define DISABLED_MASK7	(DISABLE_PTI)
 #define DISABLED_MASK8	0
-#define DISABLED_MASK9	(DISABLE_MPX)
+#define DISABLED_MASK9	(DISABLE_MPX|DISABLE_SMAP)
 #define DISABLED_MASK10	0
 #define DISABLED_MASK11	0
 #define DISABLED_MASK12	0

arch/x86/include/asm/pgalloc.h

@@ -80,6 +80,13 @@ static inline void pmd_populate_kernel(struct mm_struct *mm,
 	set_pmd(pmd, __pmd(__pa(pte) | _PAGE_TABLE));
 }
 
+static inline void pmd_populate_kernel_safe(struct mm_struct *mm,
+				       pmd_t *pmd, pte_t *pte)
+{
+	paravirt_alloc_pte(mm, __pa(pte) >> PAGE_SHIFT);
+	set_pmd_safe(pmd, __pmd(__pa(pte) | _PAGE_TABLE));
+}
+
 static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd,
 				struct page *pte)
 {
@@ -132,6 +139,12 @@ static inline void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd)
 	paravirt_alloc_pmd(mm, __pa(pmd) >> PAGE_SHIFT);
 	set_pud(pud, __pud(_PAGE_TABLE | __pa(pmd)));
 }
+
+static inline void pud_populate_safe(struct mm_struct *mm, pud_t *pud, pmd_t *pmd)
+{
+	paravirt_alloc_pmd(mm, __pa(pmd) >> PAGE_SHIFT);
+	set_pud_safe(pud, __pud(_PAGE_TABLE | __pa(pmd)));
+}
 #endif	/* CONFIG_X86_PAE */
 
 #if CONFIG_PGTABLE_LEVELS > 3
@@ -141,6 +154,12 @@ static inline void p4d_populate(struct mm_struct *mm, p4d_t *p4d, pud_t *pud)
 	set_p4d(p4d, __p4d(_PAGE_TABLE | __pa(pud)));
 }
 
+static inline void p4d_populate_safe(struct mm_struct *mm, p4d_t *p4d, pud_t *pud)
+{
+	paravirt_alloc_pud(mm, __pa(pud) >> PAGE_SHIFT);
+	set_p4d_safe(p4d, __p4d(_PAGE_TABLE | __pa(pud)));
+}
+
 static inline pud_t *pud_alloc_one(struct mm_struct *mm, unsigned long addr)
 {
 	gfp_t gfp = GFP_KERNEL_ACCOUNT;
@@ -173,6 +192,14 @@ static inline void pgd_populate(struct mm_struct *mm, pgd_t *pgd, p4d_t *p4d)
 	set_pgd(pgd, __pgd(_PAGE_TABLE | __pa(p4d)));
 }
 
+static inline void pgd_populate_safe(struct mm_struct *mm, pgd_t *pgd, p4d_t *p4d)
+{
+	if (!pgtable_l5_enabled())
+		return;
+	paravirt_alloc_p4d(mm, __pa(p4d) >> PAGE_SHIFT);
+	set_pgd_safe(pgd, __pgd(_PAGE_TABLE | __pa(p4d)));
+}
+
 static inline p4d_t *p4d_alloc_one(struct mm_struct *mm, unsigned long addr)
 {
 	gfp_t gfp = GFP_KERNEL_ACCOUNT;

arch/x86/kernel/process_64.c

@@ -68,7 +68,7 @@ void __show_regs(struct pt_regs *regs, enum show_regs_mode mode)
 	unsigned long cr0 = 0L, cr2 = 0L, cr3 = 0L, cr4 = 0L, fs, gs, shadowgs;
 	unsigned long d0, d1, d2, d3, d6, d7;
 	unsigned int fsindex, gsindex;
-	unsigned int ds, cs, es;
+	unsigned int ds, es;
 
 	show_iret_regs(regs);
 
@@ -100,7 +100,6 @@ void __show_regs(struct pt_regs *regs, enum show_regs_mode mode)
 	}
 
 	asm("movl %%ds,%0" : "=r" (ds));
-	asm("movl %%cs,%0" : "=r" (cs));
 	asm("movl %%es,%0" : "=r" (es));
 	asm("movl %%fs,%0" : "=r" (fsindex));
 	asm("movl %%gs,%0" : "=r" (gsindex));
@@ -116,7 +115,7 @@ void __show_regs(struct pt_regs *regs, enum show_regs_mode mode)
 
 	printk(KERN_DEFAULT "FS:  %016lx(%04x) GS:%016lx(%04x) knlGS:%016lx\n",
 	       fs, fsindex, gs, gsindex, shadowgs);
-	printk(KERN_DEFAULT "CS:  %04x DS: %04x ES: %04x CR0: %016lx\n", cs, ds,
+	printk(KERN_DEFAULT "CS:  %04lx DS: %04x ES: %04x CR0: %016lx\n", regs->cs, ds,
 			es, cr0);
 	printk(KERN_DEFAULT "CR2: %016lx CR3: %016lx CR4: %016lx\n", cr2, cr3,
 			cr4);

arch/x86/mm/debug_pagetables.c

@@ -10,20 +10,9 @@ static int ptdump_show(struct seq_file *m, void *v)
 	return 0;
 }
 
-static int ptdump_open(struct inode *inode, struct file *filp)
-{
-	return single_open(filp, ptdump_show, NULL);
-}
-
-static const struct file_operations ptdump_fops = {
-	.owner		= THIS_MODULE,
-	.open		= ptdump_open,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
-};
+DEFINE_SHOW_ATTRIBUTE(ptdump);
 
-static int ptdump_show_curknl(struct seq_file *m, void *v)
+static int ptdump_curknl_show(struct seq_file *m, void *v)
 {
 	if (current->mm->pgd) {
 		down_read(&current->mm->mmap_sem);
@@ -33,23 +22,12 @@ static int ptdump_show_curknl(struct seq_file *m, void *v)
 	return 0;
 }
 
-static int ptdump_open_curknl(struct inode *inode, struct file *filp)
-{
-	return single_open(filp, ptdump_show_curknl, NULL);
-}
-
-static const struct file_operations ptdump_curknl_fops = {
-	.owner		= THIS_MODULE,
-	.open		= ptdump_open_curknl,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
-};
+DEFINE_SHOW_ATTRIBUTE(ptdump_curknl);
 
 #ifdef CONFIG_PAGE_TABLE_ISOLATION
 static struct dentry *pe_curusr;
 
-static int ptdump_show_curusr(struct seq_file *m, void *v)
+static int ptdump_curusr_show(struct seq_file *m, void *v)
 {
 	if (current->mm->pgd) {
 		down_read(&current->mm->mmap_sem);
@@ -59,42 +37,20 @@ static int ptdump_show_curusr(struct seq_file *m, void *v)
 	return 0;
 }
 
-static int ptdump_open_curusr(struct inode *inode, struct file *filp)
-{
-	return single_open(filp, ptdump_show_curusr, NULL);
-}
-
-static const struct file_operations ptdump_curusr_fops = {
-	.owner		= THIS_MODULE,
-	.open		= ptdump_open_curusr,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
-};
+DEFINE_SHOW_ATTRIBUTE(ptdump_curusr);
 #endif
 
 #if defined(CONFIG_EFI) && defined(CONFIG_X86_64)
 static struct dentry *pe_efi;
 
-static int ptdump_show_efi(struct seq_file *m, void *v)
+static int ptdump_efi_show(struct seq_file *m, void *v)
 {
 	if (efi_mm.pgd)
 		ptdump_walk_pgd_level_debugfs(m, efi_mm.pgd, false);
 	return 0;
 }
 
-static int ptdump_open_efi(struct inode *inode, struct file *filp)
-{
-	return single_open(filp, ptdump_show_efi, NULL);
-}
-
-static const struct file_operations ptdump_efi_fops = {
-	.owner		= THIS_MODULE,
-	.open		= ptdump_open_efi,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
-};
+DEFINE_SHOW_ATTRIBUTE(ptdump_efi);
 #endif
 
 static struct dentry *dir, *pe_knl, *pe_curknl;

arch/x86/mm/fault.c

@@ -27,6 +27,7 @@
 #include <asm/vm86.h>			/* struct vm86			*/
 #include <asm/mmu_context.h>		/* vma_pkey()			*/
 #include <asm/efi.h>			/* efi_recover_from_page_fault()*/
+#include <asm/desc.h>			/* store_idt(), ...		*/
 
 #define CREATE_TRACE_POINTS
 #include <asm/trace/exceptions.h>
@@ -571,10 +572,55 @@ static int is_f00f_bug(struct pt_regs *regs, unsigned long address)
 	return 0;
 }
 
+static void show_ldttss(const struct desc_ptr *gdt, const char *name, u16 index)
+{
+	u32 offset = (index >> 3) * sizeof(struct desc_struct);
+	unsigned long addr;
+	struct ldttss_desc desc;
+
+	if (index == 0) {
+		pr_alert("%s: NULL\n", name);
+		return;
+	}
+
+	if (offset + sizeof(struct ldttss_desc) >= gdt->size) {
+		pr_alert("%s: 0x%hx -- out of bounds\n", name, index);
+		return;
+	}
+
+	if (probe_kernel_read(&desc, (void *)(gdt->address + offset),
+			      sizeof(struct ldttss_desc))) {
+		pr_alert("%s: 0x%hx -- GDT entry is not readable\n",
+			 name, index);
+		return;
+	}
+
+	addr = desc.base0 | (desc.base1 << 16) | (desc.base2 << 24);
+#ifdef CONFIG_X86_64
+	addr |= ((u64)desc.base3 << 32);
+#endif
+	pr_alert("%s: 0x%hx -- base=0x%lx limit=0x%x\n",
+		 name, index, addr, (desc.limit0 | (desc.limit1 << 16)));
+}
+
+/*
+ * This helper function transforms the #PF error_code bits into
+ * "[PROT] [USER]" type of descriptive, almost human-readable error strings:
+ */
+static void err_str_append(unsigned long error_code, char *buf, unsigned long mask, const char *txt)
+{
+	if (error_code & mask) {
+		if (buf[0])
+			strcat(buf, " ");
+		strcat(buf, txt);
+	}
+}
+
 static void
-show_fault_oops(struct pt_regs *regs, unsigned long error_code,
-		unsigned long address)
+show_fault_oops(struct pt_regs *regs, unsigned long error_code, unsigned long address)
 {
+	char err_txt[64];
+
 	if (!oops_may_print())
 		return;
 
@@ -602,6 +648,52 @@ show_fault_oops(struct pt_regs *regs, unsigned long error_code,
 		 address < PAGE_SIZE ? "NULL pointer dereference" : "paging request",
 		 (void *)address);
 
+	err_txt[0] = 0;
+
+	/*
+	 * Note: length of these appended strings including the separation space and the
+	 * zero delimiter must fit into err_txt[].
+	 */
+	err_str_append(error_code, err_txt, X86_PF_PROT,  "[PROT]" );
+	err_str_append(error_code, err_txt, X86_PF_WRITE, "[WRITE]");
+	err_str_append(error_code, err_txt, X86_PF_USER,  "[USER]" );
+	err_str_append(error_code, err_txt, X86_PF_RSVD,  "[RSVD]" );
+	err_str_append(error_code, err_txt, X86_PF_INSTR, "[INSTR]");
+	err_str_append(error_code, err_txt, X86_PF_PK,    "[PK]"   );
+
+	pr_alert("#PF error: %s\n", error_code ? err_txt : "[normal kernel read fault]");
+
+	if (!(error_code & X86_PF_USER) && user_mode(regs)) {
+		struct desc_ptr idt, gdt;
+		u16 ldtr, tr;
+
+		pr_alert("This was a system access from user code\n");
+
+		/*
+		 * This can happen for quite a few reasons.  The more obvious
+		 * ones are faults accessing the GDT, or LDT.  Perhaps
+		 * surprisingly, if the CPU tries to deliver a benign or
+		 * contributory exception from user code and gets a page fault
+		 * during delivery, the page fault can be delivered as though
+		 * it originated directly from user code.  This could happen
+		 * due to wrong permissions on the IDT, GDT, LDT, TSS, or
+		 * kernel or IST stack.
+		 */
+		store_idt(&idt);
+
+		/* Usable even on Xen PV -- it's just slow. */
+		native_store_gdt(&gdt);
+
+		pr_alert("IDT: 0x%lx (limit=0x%hx) GDT: 0x%lx (limit=0x%hx)\n",
+			 idt.address, idt.size, gdt.address, gdt.size);
+
+		store_ldt(ldtr);
+		show_ldttss(&gdt, "LDTR", ldtr);
+
+		store_tr(tr);
+		show_ldttss(&gdt, "TR", tr);
+	}
+
 	dump_pagetable(address);
 }
 
@@ -621,16 +713,30 @@ pgtable_bad(struct pt_regs *regs, unsigned long error_code,
 	       tsk->comm, address);
 	dump_pagetable(address);
 
-	tsk->thread.cr2		= address;
-	tsk->thread.trap_nr	= X86_TRAP_PF;
-	tsk->thread.error_code	= error_code;
-
 	if (__die("Bad pagetable", regs, error_code))
 		sig = 0;
 
 	oops_end(flags, regs, sig);
 }
 
+static void set_signal_archinfo(unsigned long address,
+				unsigned long error_code)
+{
+	struct task_struct *tsk = current;
+
+	/*
+	 * To avoid leaking information about the kernel page
+	 * table layout, pretend that user-mode accesses to
+	 * kernel addresses are always protection faults.
+	 */
+	if (address >= TASK_SIZE_MAX)
+		error_code |= X86_PF_PROT;
+
+	tsk->thread.trap_nr = X86_TRAP_PF;
+	tsk->thread.error_code = error_code | X86_PF_USER;
+	tsk->thread.cr2 = address;
+}
+
 static noinline void
 no_context(struct pt_regs *regs, unsigned long error_code,
 	   unsigned long address, int signal, int si_code)
@@ -639,6 +745,15 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 	unsigned long flags;
 	int sig;
 
+	if (user_mode(regs)) {
+		/*
+		 * This is an implicit supervisor-mode access from user
+		 * mode.  Bypass all the kernel-mode recovery code and just
+		 * OOPS.
+		 */
+		goto oops;
+	}
+
 	/* Are we prepared to handle this kernel fault? */
 	if (fixup_exception(regs, X86_TRAP_PF, error_code, address)) {
 		/*
@@ -656,9 +771,7 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 		 * faulting through the emulate_vsyscall() logic.
 		 */
 		if (current->thread.sig_on_uaccess_err && signal) {
-			tsk->thread.trap_nr = X86_TRAP_PF;
-			tsk->thread.error_code = error_code | X86_PF_USER;
-			tsk->thread.cr2 = address;
+			set_signal_archinfo(address, error_code);
 
 			/* XXX: hwpoison faults will set the wrong code. */
 			force_sig_fault(signal, si_code, (void __user *)address,
@@ -726,6 +839,7 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 	if (IS_ENABLED(CONFIG_EFI))
 		efi_recover_from_page_fault(address);
 
+oops:
 	/*
 	 * Oops. The kernel tried to access some bad page. We'll have to
 	 * terminate things with extreme prejudice:
@@ -737,10 +851,6 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 	if (task_stack_end_corrupted(tsk))
 		printk(KERN_EMERG "Thread overran stack, or stack corrupted\n");
 
-	tsk->thread.cr2		= address;
-	tsk->thread.trap_nr	= X86_TRAP_PF;
-	tsk->thread.error_code	= error_code;
-
 	sig = SIGKILL;
 	if (__die("Oops", regs, error_code))
 		sig = 0;
@@ -794,7 +904,7 @@ __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
 	struct task_struct *tsk = current;
 
 	/* User mode accesses just cause a SIGSEGV */
-	if (error_code & X86_PF_USER) {
+	if (user_mode(regs) && (error_code & X86_PF_USER)) {
 		/*
 		 * It's possible to have interrupts off here:
 		 */
@@ -821,9 +931,7 @@ __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
 		if (likely(show_unhandled_signals))
 			show_signal_msg(regs, error_code, address, tsk);
 
-		tsk->thread.cr2		= address;
-		tsk->thread.error_code	= error_code;
-		tsk->thread.trap_nr	= X86_TRAP_PF;
+		set_signal_archinfo(address, error_code);
 
 		if (si_code == SEGV_PKUERR)
 			force_sig_pkuerr((void __user *)address, pkey);
@@ -937,9 +1045,7 @@ do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address,
 	if (is_prefetch(regs, error_code, address))
 		return;
 
-	tsk->thread.cr2		= address;
-	tsk->thread.error_code	= error_code;
-	tsk->thread.trap_nr	= X86_TRAP_PF;
+	set_signal_archinfo(address, error_code);
 
 #ifdef CONFIG_MEMORY_FAILURE
 	if (fault & (VM_FAULT_HWPOISON|VM_FAULT_HWPOISON_LARGE)) {
@@ -1148,23 +1254,6 @@ static int fault_in_kernel_space(unsigned long address)
 	return address >= TASK_SIZE_MAX;
 }
 
-static inline bool smap_violation(int error_code, struct pt_regs *regs)
-{
-	if (!IS_ENABLED(CONFIG_X86_SMAP))
-		return false;
-
-	if (!static_cpu_has(X86_FEATURE_SMAP))
-		return false;
-
-	if (error_code & X86_PF_USER)
-		return false;
-
-	if (!user_mode(regs) && (regs->flags & X86_EFLAGS_AC))
-		return false;
-
-	return true;
-}
-
 /*
  * Called for all faults where 'address' is part of the kernel address
  * space.  Might get called for faults that originate from *code* that
@@ -1230,7 +1319,6 @@ void do_user_addr_fault(struct pt_regs *regs,
 			unsigned long hw_error_code,
 			unsigned long address)
 {
-	unsigned long sw_error_code;
 	struct vm_area_struct *vma;
 	struct task_struct *tsk;
 	struct mm_struct *mm;
@@ -1252,10 +1340,16 @@ void do_user_addr_fault(struct pt_regs *regs,
 		pgtable_bad(regs, hw_error_code, address);
 
 	/*
-	 * Check for invalid kernel (supervisor) access to user
-	 * pages in the user address space.
+	 * If SMAP is on, check for invalid kernel (supervisor) access to user
+	 * pages in the user address space.  The odd case here is WRUSS,
+	 * which, according to the preliminary documentation, does not respect
+	 * SMAP and will have the USER bit set so, in all cases, SMAP
+	 * enforcement appears to be consistent with the USER bit.
 	 */
-	if (unlikely(smap_violation(hw_error_code, regs))) {
+	if (unlikely(cpu_feature_enabled(X86_FEATURE_SMAP) &&
+		     !(hw_error_code & X86_PF_USER) &&
+		     !(regs->flags & X86_EFLAGS_AC)))
+	{
 		bad_area_nosemaphore(regs, hw_error_code, address);
 		return;
 	}
@@ -1269,13 +1363,6 @@ void do_user_addr_fault(struct pt_regs *regs,
 		return;
 	}
 
-	/*
-	 * hw_error_code is literally the "page fault error code" passed to
-	 * the kernel directly from the hardware.  But, we will shortly be
-	 * modifying it in software, so give it a new name.
-	 */
-	sw_error_code = hw_error_code;
-
 	/*
 	 * It's safe to allow irq's after cr2 has been saved and the
 	 * vmalloc fault has been handled.
@@ -1285,26 +1372,6 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 */
 	if (user_mode(regs)) {
 		local_irq_enable();
-		/*
-		 * Up to this point, X86_PF_USER set in hw_error_code
-		 * indicated a user-mode access.  But, after this,
-		 * X86_PF_USER in sw_error_code will indicate either
-		 * that, *or* an implicit kernel(supervisor)-mode access
-		 * which originated from user mode.
-		 */
-		if (!(hw_error_code & X86_PF_USER)) {
-			/*
-			 * The CPU was in user mode, but the CPU says
-			 * the fault was not a user-mode access.
-			 * Must be an implicit kernel-mode access,
-			 * which we do not expect to happen in the
-			 * user address space.
-			 */
-			pr_warn_once("kernel-mode error from user-mode: %lx\n",
-					hw_error_code);
-
-			sw_error_code |= X86_PF_USER;
-		}
 		flags |= FAULT_FLAG_USER;
 	} else {
 		if (regs->flags & X86_EFLAGS_IF)
@@ -1313,9 +1380,9 @@ void do_user_addr_fault(struct pt_regs *regs,
 
 	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
-	if (sw_error_code & X86_PF_WRITE)
+	if (hw_error_code & X86_PF_WRITE)
 		flags |= FAULT_FLAG_WRITE;
-	if (sw_error_code & X86_PF_INSTR)
+	if (hw_error_code & X86_PF_INSTR)
 		flags |= FAULT_FLAG_INSTRUCTION;
 
 #ifdef CONFIG_X86_64
@@ -1328,7 +1395,7 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 * The vsyscall page does not have a "real" VMA, so do this
 	 * emulation before we go searching for VMAs.
 	 */
-	if ((sw_error_code & X86_PF_INSTR) && is_vsyscall_vaddr(address)) {
+	if ((hw_error_code & X86_PF_INSTR) && is_vsyscall_vaddr(address)) {
 		if (emulate_vsyscall(regs, address))
 			return;
 	}
@@ -1344,18 +1411,15 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 * Only do the expensive exception table search when we might be at
 	 * risk of a deadlock.  This happens if we
 	 * 1. Failed to acquire mmap_sem, and
-	 * 2. The access did not originate in userspace.  Note: either the
-	 *    hardware or earlier page fault code may set X86_PF_USER
-	 *    in sw_error_code.
+	 * 2. The access did not originate in userspace.
 	 */
 	if (unlikely(!down_read_trylock(&mm->mmap_sem))) {
-		if (!(sw_error_code & X86_PF_USER) &&
-		    !search_exception_tables(regs->ip)) {
+		if (!user_mode(regs) && !search_exception_tables(regs->ip)) {
 			/*
 			 * Fault from code in kernel from
 			 * which we do not expect faults.
 			 */
-			bad_area_nosemaphore(regs, sw_error_code, address);
+			bad_area_nosemaphore(regs, hw_error_code, address);
 			return;
 		}
 retry:
@@ -1371,29 +1435,17 @@ void do_user_addr_fault(struct pt_regs *regs,
 
 	vma = find_vma(mm, address);
 	if (unlikely(!vma)) {
-		bad_area(regs, sw_error_code, address);
+		bad_area(regs, hw_error_code, address);
 		return;
 	}
 	if (likely(vma->vm_start <= address))
 		goto good_area;
 	if (unlikely(!(vma->vm_flags & VM_GROWSDOWN))) {
-		bad_area(regs, sw_error_code, address);
+		bad_area(regs, hw_error_code, address);
 		return;
 	}
-	if (sw_error_code & X86_PF_USER) {
-		/*
-		 * Accessing the stack below %sp is always a bug.
-		 * The large cushion allows instructions like enter
-		 * and pusha to work. ("enter $65535, $31" pushes
-		 * 32 pointers and then decrements %sp by 65535.)
-		 */
-		if (unlikely(address + 65536 + 32 * sizeof(unsigned long) < regs->sp)) {
-			bad_area(regs, sw_error_code, address);
-			return;
-		}
-	}
 	if (unlikely(expand_stack(vma, address))) {
-		bad_area(regs, sw_error_code, address);
+		bad_area(regs, hw_error_code, address);
 		return;
 	}
 
@@ -1402,8 +1454,8 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 * we can handle it..
 	 */
 good_area:
-	if (unlikely(access_error(sw_error_code, vma))) {
-		bad_area_access_error(regs, sw_error_code, address, vma);
+	if (unlikely(access_error(hw_error_code, vma))) {
+		bad_area_access_error(regs, hw_error_code, address, vma);
 		return;
 	}
 
@@ -1442,13 +1494,13 @@ void do_user_addr_fault(struct pt_regs *regs,
 			return;
 
 		/* Not returning to user mode? Handle exceptions or die: */
-		no_context(regs, sw_error_code, address, SIGBUS, BUS_ADRERR);
+		no_context(regs, hw_error_code, address, SIGBUS, BUS_ADRERR);
 		return;
 	}
 
 	up_read(&mm->mmap_sem);
 	if (unlikely(fault & VM_FAULT_ERROR)) {
-		mm_fault_error(regs, sw_error_code, address, fault);
+		mm_fault_error(regs, hw_error_code, address, fault);
 		return;
 	}
 

arch/x86/mm/init_64.c

@@ -432,7 +432,7 @@ phys_pte_init(pte_t *pte_page, unsigned long paddr, unsigned long paddr_end,
 					     E820_TYPE_RAM) &&
 			    !e820__mapped_any(paddr & PAGE_MASK, paddr_next,
 					     E820_TYPE_RESERVED_KERN))
-				set_pte(pte, __pte(0));
+				set_pte_safe(pte, __pte(0));
 			continue;
 		}
 
@@ -452,7 +452,7 @@ phys_pte_init(pte_t *pte_page, unsigned long paddr, unsigned long paddr_end,
 			pr_info("   pte=%p addr=%lx pte=%016lx\n", pte, paddr,
 				pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL).pte);
 		pages++;
-		set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, prot));
+		set_pte_safe(pte, pfn_pte(paddr >> PAGE_SHIFT, prot));
 		paddr_last = (paddr & PAGE_MASK) + PAGE_SIZE;
 	}
 
@@ -487,7 +487,7 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long paddr, unsigned long paddr_end,
 					     E820_TYPE_RAM) &&
 			    !e820__mapped_any(paddr & PMD_MASK, paddr_next,
 					     E820_TYPE_RESERVED_KERN))
-				set_pmd(pmd, __pmd(0));
+				set_pmd_safe(pmd, __pmd(0));
 			continue;
 		}
 
@@ -524,7 +524,7 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long paddr, unsigned long paddr_end,
 		if (page_size_mask & (1<<PG_LEVEL_2M)) {
 			pages++;
 			spin_lock(&init_mm.page_table_lock);
-			set_pte((pte_t *)pmd,
+			set_pte_safe((pte_t *)pmd,
 				pfn_pte((paddr & PMD_MASK) >> PAGE_SHIFT,
 					__pgprot(pgprot_val(prot) | _PAGE_PSE)));
 			spin_unlock(&init_mm.page_table_lock);
@@ -536,7 +536,7 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long paddr, unsigned long paddr_end,
 		paddr_last = phys_pte_init(pte, paddr, paddr_end, new_prot);
 
 		spin_lock(&init_mm.page_table_lock);
-		pmd_populate_kernel(&init_mm, pmd, pte);
+		pmd_populate_kernel_safe(&init_mm, pmd, pte);
 		spin_unlock(&init_mm.page_table_lock);
 	}
 	update_page_count(PG_LEVEL_2M, pages);
@@ -573,7 +573,7 @@ phys_pud_init(pud_t *pud_page, unsigned long paddr, unsigned long paddr_end,
 					     E820_TYPE_RAM) &&
 			    !e820__mapped_any(paddr & PUD_MASK, paddr_next,
 					     E820_TYPE_RESERVED_KERN))
-				set_pud(pud, __pud(0));
+				set_pud_safe(pud, __pud(0));
 			continue;
 		}
 
@@ -584,7 +584,6 @@ phys_pud_init(pud_t *pud_page, unsigned long paddr, unsigned long paddr_end,
 							   paddr_end,
 							   page_size_mask,
 							   prot);
-				__flush_tlb_all();
 				continue;
 			}
 			/*
@@ -611,7 +610,7 @@ phys_pud_init(pud_t *pud_page, unsigned long paddr, unsigned long paddr_end,
 		if (page_size_mask & (1<<PG_LEVEL_1G)) {
 			pages++;
 			spin_lock(&init_mm.page_table_lock);
-			set_pte((pte_t *)pud,
+			set_pte_safe((pte_t *)pud,
 				pfn_pte((paddr & PUD_MASK) >> PAGE_SHIFT,
 					PAGE_KERNEL_LARGE));
 			spin_unlock(&init_mm.page_table_lock);
@@ -624,10 +623,9 @@ phys_pud_init(pud_t *pud_page, unsigned long paddr, unsigned long paddr_end,
 					   page_size_mask, prot);
 
 		spin_lock(&init_mm.page_table_lock);
-		pud_populate(&init_mm, pud, pmd);
+		pud_populate_safe(&init_mm, pud, pmd);
 		spin_unlock(&init_mm.page_table_lock);
 	}
-	__flush_tlb_all();
 
 	update_page_count(PG_LEVEL_1G, pages);
 
@@ -659,7 +657,7 @@ phys_p4d_init(p4d_t *p4d_page, unsigned long paddr, unsigned long paddr_end,
 					     E820_TYPE_RAM) &&
 			    !e820__mapped_any(paddr & P4D_MASK, paddr_next,
 					     E820_TYPE_RESERVED_KERN))
-				set_p4d(p4d, __p4d(0));
+				set_p4d_safe(p4d, __p4d(0));
 			continue;
 		}
 
@@ -668,7 +666,6 @@ phys_p4d_init(p4d_t *p4d_page, unsigned long paddr, unsigned long paddr_end,
 			paddr_last = phys_pud_init(pud, paddr,
 					paddr_end,
 					page_size_mask);
-			__flush_tlb_all();
 			continue;
 		}
 
@@ -677,10 +674,9 @@ phys_p4d_init(p4d_t *p4d_page, unsigned long paddr, unsigned long paddr_end,
 					   page_size_mask);
 
 		spin_lock(&init_mm.page_table_lock);
-		p4d_populate(&init_mm, p4d, pud);
+		p4d_populate_safe(&init_mm, p4d, pud);
 		spin_unlock(&init_mm.page_table_lock);
 	}
-	__flush_tlb_all();
 
 	return paddr_last;
 }
@@ -723,9 +719,9 @@ kernel_physical_mapping_init(unsigned long paddr_start,
 
 		spin_lock(&init_mm.page_table_lock);
 		if (pgtable_l5_enabled())
-			pgd_populate(&init_mm, pgd, p4d);
+			pgd_populate_safe(&init_mm, pgd, p4d);
 		else
-			p4d_populate(&init_mm, p4d_offset(pgd, vaddr), (pud_t *) p4d);
+			p4d_populate_safe(&init_mm, p4d_offset(pgd, vaddr), (pud_t *) p4d);
 		spin_unlock(&init_mm.page_table_lock);
 		pgd_changed = true;
 	}
@@ -733,8 +729,6 @@ kernel_physical_mapping_init(unsigned long paddr_start,
 	if (pgd_changed)
 		sync_global_pgds(vaddr_start, vaddr_end - 1);
 
-	__flush_tlb_all();
-
 	return paddr_last;
 }
 

arch/x86/mm/mm_internal.h

@@ -19,4 +19,6 @@ extern int after_bootmem;
 
 void update_cache_mode_entry(unsigned entry, enum page_cache_mode cache);
 
+extern unsigned long tlb_single_page_flush_ceiling;
+
 #endif	/* __X86_MM_INTERNAL_H */

arch/x86/mm/pageattr-test.c

@@ -23,7 +23,8 @@
 static __read_mostly int print = 1;
 
 enum {
-	NTEST			= 400,
+	NTEST			= 3 * 100,
+	NPAGES			= 100,
 #ifdef CONFIG_X86_64
 	LPS			= (1 << PMD_SHIFT),
 #elif defined(CONFIG_X86_PAE)
@@ -110,6 +111,9 @@ static int print_split(struct split_state *s)
 static unsigned long addr[NTEST];
 static unsigned int len[NTEST];
 
+static struct page *pages[NPAGES];
+static unsigned long addrs[NPAGES];
+
 /* Change the global bit on random pages in the direct mapping */
 static int pageattr_test(void)
 {
@@ -120,7 +124,6 @@ static int pageattr_test(void)
 	unsigned int level;
 	int i, k;
 	int err;
-	unsigned long test_addr;
 
 	if (print)
 		printk(KERN_INFO "CPA self-test:\n");
@@ -137,7 +140,7 @@ static int pageattr_test(void)
 		unsigned long pfn = prandom_u32() % max_pfn_mapped;
 
 		addr[i] = (unsigned long)__va(pfn << PAGE_SHIFT);
-		len[i] = prandom_u32() % 100;
+		len[i] = prandom_u32() % NPAGES;
 		len[i] = min_t(unsigned long, len[i], max_pfn_mapped - pfn - 1);
 
 		if (len[i] == 0)
@@ -167,14 +170,29 @@ static int pageattr_test(void)
 				break;
 			}
 			__set_bit(pfn + k, bm);
+			addrs[k] = addr[i] + k*PAGE_SIZE;
+			pages[k] = pfn_to_page(pfn + k);
 		}
 		if (!addr[i] || !pte || !k) {
 			addr[i] = 0;
 			continue;
 		}
 
-		test_addr = addr[i];
-		err = change_page_attr_set(&test_addr, len[i], PAGE_CPA_TEST, 0);
+		switch (i % 3) {
+		case 0:
+			err = change_page_attr_set(&addr[i], len[i], PAGE_CPA_TEST, 0);
+			break;
+
+		case 1:
+			err = change_page_attr_set(addrs, len[1], PAGE_CPA_TEST, 1);
+			break;
+
+		case 2:
+			err = cpa_set_pages_array(pages, len[i], PAGE_CPA_TEST);
+			break;
+		}
+
+
 		if (err < 0) {
 			printk(KERN_ERR "CPA %d failed %d\n", i, err);
 			failed++;
@@ -206,8 +224,7 @@ static int pageattr_test(void)
 			failed++;
 			continue;
 		}
-		test_addr = addr[i];
-		err = change_page_attr_clear(&test_addr, len[i], PAGE_CPA_TEST, 0);
+		err = change_page_attr_clear(&addr[i], len[i], PAGE_CPA_TEST, 0);
 		if (err < 0) {
 			printk(KERN_ERR "CPA reverting failed: %d\n", err);
 			failed++;

arch/x86/mm/pageattr.c

@@ -26,6 +26,8 @@
 #include <asm/pat.h>
 #include <asm/set_memory.h>
 
+#include "mm_internal.h"
+
 /*
  * The current flushing context - we pass it instead of 5 arguments:
  */
@@ -35,11 +37,11 @@ struct cpa_data {
 	pgprot_t	mask_set;
 	pgprot_t	mask_clr;
 	unsigned long	numpages;
-	int		flags;
+	unsigned long	curpage;
 	unsigned long	pfn;
-	unsigned	force_split		: 1,
+	unsigned int	flags;
+	unsigned int	force_split		: 1,
 			force_static_prot	: 1;
-	int		curpage;
 	struct page	**pages;
 };
 
@@ -228,19 +230,28 @@ static bool __cpa_pfn_in_highmap(unsigned long pfn)
 
 #endif
 
+static unsigned long __cpa_addr(struct cpa_data *cpa, unsigned long idx)
+{
+	if (cpa->flags & CPA_PAGES_ARRAY) {
+		struct page *page = cpa->pages[idx];
+
+		if (unlikely(PageHighMem(page)))
+			return 0;
+
+		return (unsigned long)page_address(page);
+	}
+
+	if (cpa->flags & CPA_ARRAY)
+		return cpa->vaddr[idx];
+
+	return *cpa->vaddr + idx * PAGE_SIZE;
+}
+
 /*
  * Flushing functions
  */
 
-/**
- * clflush_cache_range - flush a cache range with clflush
- * @vaddr:	virtual start address
- * @size:	number of bytes to flush
- *
- * clflushopt is an unordered instruction which needs fencing with mfence or
- * sfence to avoid ordering issues.
- */
-void clflush_cache_range(void *vaddr, unsigned int size)
+static void clflush_cache_range_opt(void *vaddr, unsigned int size)
 {
 	const unsigned long clflush_size = boot_cpu_data.x86_clflush_size;
 	void *p = (void *)((unsigned long)vaddr & ~(clflush_size - 1));
@@ -249,11 +260,22 @@ void clflush_cache_range(void *vaddr, unsigned int size)
 	if (p >= vend)
 		return;
 
-	mb();
-
 	for (; p < vend; p += clflush_size)
 		clflushopt(p);
+}
 
+/**
+ * clflush_cache_range - flush a cache range with clflush
+ * @vaddr:	virtual start address
+ * @size:	number of bytes to flush
+ *
+ * CLFLUSHOPT is an unordered instruction which needs fencing with MFENCE or
+ * SFENCE to avoid ordering issues.
+ */
+void clflush_cache_range(void *vaddr, unsigned int size)
+{
+	mb();
+	clflush_cache_range_opt(vaddr, size);
 	mb();
 }
 EXPORT_SYMBOL_GPL(clflush_cache_range);
@@ -285,87 +307,49 @@ static void cpa_flush_all(unsigned long cache)
 	on_each_cpu(__cpa_flush_all, (void *) cache, 1);
 }
 
-static bool __inv_flush_all(int cache)
+void __cpa_flush_tlb(void *data)
 {
-	BUG_ON(irqs_disabled() && !early_boot_irqs_disabled);
+	struct cpa_data *cpa = data;
+	unsigned int i;
 
-	if (cache && !static_cpu_has(X86_FEATURE_CLFLUSH)) {
-		cpa_flush_all(cache);
-		return true;
-	}
-
-	return false;
+	for (i = 0; i < cpa->numpages; i++)
+		__flush_tlb_one_kernel(__cpa_addr(cpa, i));
 }
 
-static void cpa_flush_range(unsigned long start, int numpages, int cache)
+static void cpa_flush(struct cpa_data *data, int cache)
 {
-	unsigned int i, level;
-	unsigned long addr;
+	struct cpa_data *cpa = data;
+	unsigned int i;
 
-	WARN_ON(PAGE_ALIGN(start) != start);
-
-	if (__inv_flush_all(cache))
-		return;
-
-	flush_tlb_kernel_range(start, start + PAGE_SIZE * numpages);
+	BUG_ON(irqs_disabled() && !early_boot_irqs_disabled);
 
-	if (!cache)
+	if (cache && !static_cpu_has(X86_FEATURE_CLFLUSH)) {
+		cpa_flush_all(cache);
 		return;
-
-	/*
-	 * We only need to flush on one CPU,
-	 * clflush is a MESI-coherent instruction that
-	 * will cause all other CPUs to flush the same
-	 * cachelines:
-	 */
-	for (i = 0, addr = start; i < numpages; i++, addr += PAGE_SIZE) {
-		pte_t *pte = lookup_address(addr, &level);
-
-		/*
-		 * Only flush present addresses:
-		 */
-		if (pte && (pte_val(*pte) & _PAGE_PRESENT))
-			clflush_cache_range((void *) addr, PAGE_SIZE);
 	}
-}
 
-static void cpa_flush_array(unsigned long baddr, unsigned long *start,
-			    int numpages, int cache,
-			    int in_flags, struct page **pages)
-{
-	unsigned int i, level;
-
-	if (__inv_flush_all(cache))
-		return;
-
-	flush_tlb_all();
+	if (cpa->numpages <= tlb_single_page_flush_ceiling)
+		on_each_cpu(__cpa_flush_tlb, cpa, 1);
+	else
+		flush_tlb_all();
 
 	if (!cache)
 		return;
 
-	/*
-	 * We only need to flush on one CPU,
-	 * clflush is a MESI-coherent instruction that
-	 * will cause all other CPUs to flush the same
-	 * cachelines:
-	 */
-	for (i = 0; i < numpages; i++) {
-		unsigned long addr;
-		pte_t *pte;
-
-		if (in_flags & CPA_PAGES_ARRAY)
-			addr = (unsigned long)page_address(pages[i]);
-		else
-			addr = start[i];
+	mb();
+	for (i = 0; i < cpa->numpages; i++) {
+		unsigned long addr = __cpa_addr(cpa, i);
+		unsigned int level;
 
-		pte = lookup_address(addr, &level);
+		pte_t *pte = lookup_address(addr, &level);
 
 		/*
 		 * Only flush present addresses:
 		 */
 		if (pte && (pte_val(*pte) & _PAGE_PRESENT))
-			clflush_cache_range((void *)addr, PAGE_SIZE);
+			clflush_cache_range_opt((void *)addr, PAGE_SIZE);
 	}
+	mb();
 }
 
 static bool overlaps(unsigned long r1_start, unsigned long r1_end,
@@ -1476,15 +1460,7 @@ static int __change_page_attr(struct cpa_data *cpa, int primary)
 	unsigned int level;
 	pte_t *kpte, old_pte;
 
-	if (cpa->flags & CPA_PAGES_ARRAY) {
-		struct page *page = cpa->pages[cpa->curpage];
-		if (unlikely(PageHighMem(page)))
-			return 0;
-		address = (unsigned long)page_address(page);
-	} else if (cpa->flags & CPA_ARRAY)
-		address = cpa->vaddr[cpa->curpage];
-	else
-		address = *cpa->vaddr;
+	address = __cpa_addr(cpa, cpa->curpage);
 repeat:
 	kpte = _lookup_address_cpa(cpa, address, &level);
 	if (!kpte)
@@ -1565,22 +1541,14 @@ static int cpa_process_alias(struct cpa_data *cpa)
 	 * No need to redo, when the primary call touched the direct
 	 * mapping already:
 	 */
-	if (cpa->flags & CPA_PAGES_ARRAY) {
-		struct page *page = cpa->pages[cpa->curpage];
-		if (unlikely(PageHighMem(page)))
-			return 0;
-		vaddr = (unsigned long)page_address(page);
-	} else if (cpa->flags & CPA_ARRAY)
-		vaddr = cpa->vaddr[cpa->curpage];
-	else
-		vaddr = *cpa->vaddr;
-
+	vaddr = __cpa_addr(cpa, cpa->curpage);
 	if (!(within(vaddr, PAGE_OFFSET,
 		    PAGE_OFFSET + (max_pfn_mapped << PAGE_SHIFT)))) {
 
 		alias_cpa = *cpa;
 		alias_cpa.vaddr = &laddr;
 		alias_cpa.flags &= ~(CPA_PAGES_ARRAY | CPA_ARRAY);
+		alias_cpa.curpage = 0;
 
 		ret = __change_page_attr_set_clr(&alias_cpa, 0);
 		if (ret)
@@ -1600,6 +1568,7 @@ static int cpa_process_alias(struct cpa_data *cpa)
 		alias_cpa = *cpa;
 		alias_cpa.vaddr = &temp_cpa_vaddr;
 		alias_cpa.flags &= ~(CPA_PAGES_ARRAY | CPA_ARRAY);
+		alias_cpa.curpage = 0;
 
 		/*
 		 * The high mapping range is imprecise, so ignore the
@@ -1615,14 +1584,15 @@ static int cpa_process_alias(struct cpa_data *cpa)
 static int __change_page_attr_set_clr(struct cpa_data *cpa, int checkalias)
 {
 	unsigned long numpages = cpa->numpages;
-	int ret;
+	unsigned long rempages = numpages;
+	int ret = 0;
 
-	while (numpages) {
+	while (rempages) {
 		/*
 		 * Store the remaining nr of pages for the large page
 		 * preservation check.
 		 */
-		cpa->numpages = numpages;
+		cpa->numpages = rempages;
 		/* for array changes, we can't use large page */
 		if (cpa->flags & (CPA_ARRAY | CPA_PAGES_ARRAY))
 			cpa->numpages = 1;
@@ -1633,12 +1603,12 @@ static int __change_page_attr_set_clr(struct cpa_data *cpa, int checkalias)
 		if (!debug_pagealloc_enabled())
 			spin_unlock(&cpa_lock);
 		if (ret)
-			return ret;
+			goto out;
 
 		if (checkalias) {
 			ret = cpa_process_alias(cpa);
 			if (ret)
-				return ret;
+				goto out;
 		}
 
 		/*
@@ -1646,15 +1616,15 @@ static int __change_page_attr_set_clr(struct cpa_data *cpa, int checkalias)
 		 * CPA operation. Either a large page has been
 		 * preserved or a single page update happened.
 		 */
-		BUG_ON(cpa->numpages > numpages || !cpa->numpages);
-		numpages -= cpa->numpages;
-		if (cpa->flags & (CPA_PAGES_ARRAY | CPA_ARRAY))
-			cpa->curpage++;
-		else
-			*cpa->vaddr += cpa->numpages * PAGE_SIZE;
-
+		BUG_ON(cpa->numpages > rempages || !cpa->numpages);
+		rempages -= cpa->numpages;
+		cpa->curpage += cpa->numpages;
 	}
-	return 0;
+
+out:
+	/* Restore the original numpages */
+	cpa->numpages = numpages;
+	return ret;
 }
 
 /*
@@ -1687,7 +1657,6 @@ static int change_page_attr_set_clr(unsigned long *addr, int numpages,
 {
 	struct cpa_data cpa;
 	int ret, cache, checkalias;
-	unsigned long baddr = 0;
 
 	memset(&cpa, 0, sizeof(cpa));
 
@@ -1721,11 +1690,6 @@ static int change_page_attr_set_clr(unsigned long *addr, int numpages,
 			 */
 			WARN_ON_ONCE(1);
 		}
-		/*
-		 * Save address for cache flush. *addr is modified in the call
-		 * to __change_page_attr_set_clr() below.
-		 */
-		baddr = make_addr_canonical_again(*addr);
 	}
 
 	/* Must avoid aliasing mappings in the highmem code */
@@ -1773,13 +1737,7 @@ static int change_page_attr_set_clr(unsigned long *addr, int numpages,
 		goto out;
 	}
 
-	if (cpa.flags & (CPA_PAGES_ARRAY | CPA_ARRAY)) {
-		cpa_flush_array(baddr, addr, numpages, cache,
-				cpa.flags, pages);
-	} else {
-		cpa_flush_range(baddr, numpages, cache);
-	}
-
+	cpa_flush(&cpa, cache);
 out:
 	return ret;
 }
@@ -1850,14 +1808,14 @@ int set_memory_uc(unsigned long addr, int numpages)
 }
 EXPORT_SYMBOL(set_memory_uc);
 
-static int _set_memory_array(unsigned long *addr, int addrinarray,
+static int _set_memory_array(unsigned long *addr, int numpages,
 		enum page_cache_mode new_type)
 {
 	enum page_cache_mode set_type;
 	int i, j;
 	int ret;
 
-	for (i = 0; i < addrinarray; i++) {
+	for (i = 0; i < numpages; i++) {
 		ret = reserve_memtype(__pa(addr[i]), __pa(addr[i]) + PAGE_SIZE,
 					new_type, NULL);
 		if (ret)
@@ -1868,11 +1826,11 @@ static int _set_memory_array(unsigned long *addr, int addrinarray,
 	set_type = (new_type == _PAGE_CACHE_MODE_WC) ?
 				_PAGE_CACHE_MODE_UC_MINUS : new_type;
 
-	ret = change_page_attr_set(addr, addrinarray,
+	ret = change_page_attr_set(addr, numpages,
 				   cachemode2pgprot(set_type), 1);
 
 	if (!ret && new_type == _PAGE_CACHE_MODE_WC)
-		ret = change_page_attr_set_clr(addr, addrinarray,
+		ret = change_page_attr_set_clr(addr, numpages,
 					       cachemode2pgprot(
 						_PAGE_CACHE_MODE_WC),
 					       __pgprot(_PAGE_CACHE_MASK),
@@ -1889,36 +1847,34 @@ static int _set_memory_array(unsigned long *addr, int addrinarray,
 	return ret;
 }
 
-int set_memory_array_uc(unsigned long *addr, int addrinarray)
+int set_memory_array_uc(unsigned long *addr, int numpages)
 {
-	return _set_memory_array(addr, addrinarray, _PAGE_CACHE_MODE_UC_MINUS);
+	return _set_memory_array(addr, numpages, _PAGE_CACHE_MODE_UC_MINUS);
 }
 EXPORT_SYMBOL(set_memory_array_uc);
 
-int set_memory_array_wc(unsigned long *addr, int addrinarray)
+int set_memory_array_wc(unsigned long *addr, int numpages)
 {
-	return _set_memory_array(addr, addrinarray, _PAGE_CACHE_MODE_WC);
+	return _set_memory_array(addr, numpages, _PAGE_CACHE_MODE_WC);
 }
 EXPORT_SYMBOL(set_memory_array_wc);
 
-int set_memory_array_wt(unsigned long *addr, int addrinarray)
+int set_memory_array_wt(unsigned long *addr, int numpages)
 {
-	return _set_memory_array(addr, addrinarray, _PAGE_CACHE_MODE_WT);
+	return _set_memory_array(addr, numpages, _PAGE_CACHE_MODE_WT);
 }
 EXPORT_SYMBOL_GPL(set_memory_array_wt);
 
 int _set_memory_wc(unsigned long addr, int numpages)
 {
 	int ret;
-	unsigned long addr_copy = addr;
 
 	ret = change_page_attr_set(&addr, numpages,
 				   cachemode2pgprot(_PAGE_CACHE_MODE_UC_MINUS),
 				   0);
 	if (!ret) {
-		ret = change_page_attr_set_clr(&addr_copy, numpages,
-					       cachemode2pgprot(
-						_PAGE_CACHE_MODE_WC),
+		ret = change_page_attr_set_clr(&addr, numpages,
+					       cachemode2pgprot(_PAGE_CACHE_MODE_WC),
 					       __pgprot(_PAGE_CACHE_MASK),
 					       0, 0, NULL);
 	}
@@ -1985,18 +1941,18 @@ int set_memory_wb(unsigned long addr, int numpages)
 }
 EXPORT_SYMBOL(set_memory_wb);
 
-int set_memory_array_wb(unsigned long *addr, int addrinarray)
+int set_memory_array_wb(unsigned long *addr, int numpages)
 {
 	int i;
 	int ret;
 
 	/* WB cache mode is hard wired to all cache attribute bits being 0 */
-	ret = change_page_attr_clear(addr, addrinarray,
+	ret = change_page_attr_clear(addr, numpages,
 				      __pgprot(_PAGE_CACHE_MASK), 1);
 	if (ret)
 		return ret;
 
-	for (i = 0; i < addrinarray; i++)
+	for (i = 0; i < numpages; i++)
 		free_memtype(__pa(addr[i]), __pa(addr[i]) + PAGE_SIZE);
 
 	return 0;
@@ -2066,7 +2022,6 @@ int set_memory_global(unsigned long addr, int numpages)
 static int __set_memory_enc_dec(unsigned long addr, int numpages, bool enc)
 {
 	struct cpa_data cpa;
-	unsigned long start;
 	int ret;
 
 	/* Nothing to do if memory encryption is not active */
@@ -2077,8 +2032,6 @@ static int __set_memory_enc_dec(unsigned long addr, int numpages, bool enc)
 	if (WARN_ONCE(addr & ~PAGE_MASK, "misaligned address: %#lx\n", addr))
 		addr &= PAGE_MASK;
 
-	start = addr;
-
 	memset(&cpa, 0, sizeof(cpa));
 	cpa.vaddr = &addr;
 	cpa.numpages = numpages;
@@ -2093,18 +2046,18 @@ static int __set_memory_enc_dec(unsigned long addr, int numpages, bool enc)
 	/*
 	 * Before changing the encryption attribute, we need to flush caches.
 	 */
-	cpa_flush_range(start, numpages, 1);
+	cpa_flush(&cpa, 1);
 
 	ret = __change_page_attr_set_clr(&cpa, 1);
 
 	/*
-	 * After changing the encryption attribute, we need to flush TLBs
-	 * again in case any speculative TLB caching occurred (but no need
-	 * to flush caches again).  We could just use cpa_flush_all(), but
-	 * in case TLB flushing gets optimized in the cpa_flush_range()
-	 * path use the same logic as above.
+	 * After changing the encryption attribute, we need to flush TLBs again
+	 * in case any speculative TLB caching occurred (but no need to flush
+	 * caches again).  We could just use cpa_flush_all(), but in case TLB
+	 * flushing gets optimized in the cpa_flush() path use the same logic
+	 * as above.
 	 */
-	cpa_flush_range(start, numpages, 0);
+	cpa_flush(&cpa, 0);
 
 	return ret;
 }
@@ -2129,7 +2082,7 @@ int set_pages_uc(struct page *page, int numpages)
 }
 EXPORT_SYMBOL(set_pages_uc);
 
-static int _set_pages_array(struct page **pages, int addrinarray,
+static int _set_pages_array(struct page **pages, int numpages,
 		enum page_cache_mode new_type)
 {
 	unsigned long start;
@@ -2139,7 +2092,7 @@ static int _set_pages_array(struct page **pages, int addrinarray,
 	int free_idx;
 	int ret;
 
-	for (i = 0; i < addrinarray; i++) {
+	for (i = 0; i < numpages; i++) {
 		if (PageHighMem(pages[i]))
 			continue;
 		start = page_to_pfn(pages[i]) << PAGE_SHIFT;
@@ -2152,10 +2105,10 @@ static int _set_pages_array(struct page **pages, int addrinarray,
 	set_type = (new_type == _PAGE_CACHE_MODE_WC) ?
 				_PAGE_CACHE_MODE_UC_MINUS : new_type;
 
-	ret = cpa_set_pages_array(pages, addrinarray,
+	ret = cpa_set_pages_array(pages, numpages,
 				  cachemode2pgprot(set_type));
 	if (!ret && new_type == _PAGE_CACHE_MODE_WC)
-		ret = change_page_attr_set_clr(NULL, addrinarray,
+		ret = change_page_attr_set_clr(NULL, numpages,
 					       cachemode2pgprot(
 						_PAGE_CACHE_MODE_WC),
 					       __pgprot(_PAGE_CACHE_MASK),
@@ -2175,21 +2128,21 @@ static int _set_pages_array(struct page **pages, int addrinarray,
 	return -EINVAL;
 }
 
-int set_pages_array_uc(struct page **pages, int addrinarray)
+int set_pages_array_uc(struct page **pages, int numpages)
 {
-	return _set_pages_array(pages, addrinarray, _PAGE_CACHE_MODE_UC_MINUS);
+	return _set_pages_array(pages, numpages, _PAGE_CACHE_MODE_UC_MINUS);
 }
 EXPORT_SYMBOL(set_pages_array_uc);
 
-int set_pages_array_wc(struct page **pages, int addrinarray)
+int set_pages_array_wc(struct page **pages, int numpages)
 {
-	return _set_pages_array(pages, addrinarray, _PAGE_CACHE_MODE_WC);
+	return _set_pages_array(pages, numpages, _PAGE_CACHE_MODE_WC);
 }
 EXPORT_SYMBOL(set_pages_array_wc);
 
-int set_pages_array_wt(struct page **pages, int addrinarray)
+int set_pages_array_wt(struct page **pages, int numpages)
 {
-	return _set_pages_array(pages, addrinarray, _PAGE_CACHE_MODE_WT);
+	return _set_pages_array(pages, numpages, _PAGE_CACHE_MODE_WT);
 }
 EXPORT_SYMBOL_GPL(set_pages_array_wt);
 
@@ -2201,7 +2154,7 @@ int set_pages_wb(struct page *page, int numpages)
 }
 EXPORT_SYMBOL(set_pages_wb);
 
-int set_pages_array_wb(struct page **pages, int addrinarray)
+int set_pages_array_wb(struct page **pages, int numpages)
 {
 	int retval;
 	unsigned long start;
@@ -2209,12 +2162,12 @@ int set_pages_array_wb(struct page **pages, int addrinarray)
 	int i;
 
 	/* WB cache mode is hard wired to all cache attribute bits being 0 */
-	retval = cpa_clear_pages_array(pages, addrinarray,
+	retval = cpa_clear_pages_array(pages, numpages,
 			__pgprot(_PAGE_CACHE_MASK));
 	if (retval)
 		return retval;
 
-	for (i = 0; i < addrinarray; i++) {
+	for (i = 0; i < numpages; i++) {
 		if (PageHighMem(pages[i]))
 			continue;
 		start = page_to_pfn(pages[i]) << PAGE_SHIFT;

arch/x86/mm/tlb.c

@@ -15,6 +15,8 @@
 #include <asm/apic.h>
 #include <asm/uv/uv.h>
 
+#include "mm_internal.h"
+
 /*
  *	TLB flushing, formerly SMP-only
  *		c/o Linus Torvalds.
@@ -721,7 +723,7 @@ void native_flush_tlb_others(const struct cpumask *cpumask,
  *
  * This is in units of pages.
  */
-static unsigned long tlb_single_page_flush_ceiling __read_mostly = 33;
+unsigned long tlb_single_page_flush_ceiling __read_mostly = 33;
 
 void flush_tlb_mm_range(struct mm_struct *mm, unsigned long start,
 				unsigned long end, unsigned int stride_shift,

include/asm-generic/5level-fixup.h

@@ -26,6 +26,7 @@
 #define p4d_clear(p4d)			pgd_clear(p4d)
 #define p4d_val(p4d)			pgd_val(p4d)
 #define p4d_populate(mm, p4d, pud)	pgd_populate(mm, p4d, pud)
+#define p4d_populate_safe(mm, p4d, pud)	pgd_populate(mm, p4d, pud)
 #define p4d_page(p4d)			pgd_page(p4d)
 #define p4d_page_vaddr(p4d)		pgd_page_vaddr(p4d)
 

include/asm-generic/pgtable-nop4d-hack.h

@@ -31,6 +31,7 @@ static inline void pgd_clear(pgd_t *pgd)	{ }
 #define pud_ERROR(pud)				(pgd_ERROR((pud).pgd))
 
 #define pgd_populate(mm, pgd, pud)		do { } while (0)
+#define pgd_populate_safe(mm, pgd, pud)		do { } while (0)
 /*
  * (puds are folded into pgds so this doesn't get actually called,
  * but the define is needed for a generic inline function.)

include/asm-generic/pgtable-nop4d.h

@@ -26,6 +26,7 @@ static inline void pgd_clear(pgd_t *pgd)	{ }
 #define p4d_ERROR(p4d)				(pgd_ERROR((p4d).pgd))
 
 #define pgd_populate(mm, pgd, p4d)		do { } while (0)
+#define pgd_populate_safe(mm, pgd, p4d)		do { } while (0)
 /*
  * (p4ds are folded into pgds so this doesn't get actually called,
  * but the define is needed for a generic inline function.)

include/asm-generic/pgtable-nopud.h

@@ -35,6 +35,7 @@ static inline void p4d_clear(p4d_t *p4d)	{ }
 #define pud_ERROR(pud)				(p4d_ERROR((pud).p4d))
 
 #define p4d_populate(mm, p4d, pud)		do { } while (0)
+#define p4d_populate_safe(mm, p4d, pud)		do { } while (0)
 /*
  * (puds are folded into p4ds so this doesn't get actually called,
  * but the define is needed for a generic inline function.)

include/asm-generic/pgtable.h

@@ -375,7 +375,6 @@ static inline int pte_unused(pte_t pte)
 #endif
 
 #ifndef __HAVE_ARCH_PMD_SAME
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
 static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b)
 {
 	return pmd_val(pmd_a) == pmd_val(pmd_b);
@@ -385,21 +384,60 @@ static inline int pud_same(pud_t pud_a, pud_t pud_b)
 {
 	return pud_val(pud_a) == pud_val(pud_b);
 }
-#else /* CONFIG_TRANSPARENT_HUGEPAGE */
-static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b)
+#endif
+
+#ifndef __HAVE_ARCH_P4D_SAME
+static inline int p4d_same(p4d_t p4d_a, p4d_t p4d_b)
 {
-	BUILD_BUG();
-	return 0;
+	return p4d_val(p4d_a) == p4d_val(p4d_b);
 }
+#endif
 
-static inline int pud_same(pud_t pud_a, pud_t pud_b)
+#ifndef __HAVE_ARCH_PGD_SAME
+static inline int pgd_same(pgd_t pgd_a, pgd_t pgd_b)
 {
-	BUILD_BUG();
-	return 0;
+	return pgd_val(pgd_a) == pgd_val(pgd_b);
 }
-#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 #endif
 
+/*
+ * Use set_p*_safe(), and elide TLB flushing, when confident that *no*
+ * TLB flush will be required as a result of the "set". For example, use
+ * in scenarios where it is known ahead of time that the routine is
+ * setting non-present entries, or re-setting an existing entry to the
+ * same value. Otherwise, use the typical "set" helpers and flush the
+ * TLB.
+ */
+#define set_pte_safe(ptep, pte) \
+({ \
+	WARN_ON_ONCE(pte_present(*ptep) && !pte_same(*ptep, pte)); \
+	set_pte(ptep, pte); \
+})
+
+#define set_pmd_safe(pmdp, pmd) \
+({ \
+	WARN_ON_ONCE(pmd_present(*pmdp) && !pmd_same(*pmdp, pmd)); \
+	set_pmd(pmdp, pmd); \
+})
+
+#define set_pud_safe(pudp, pud) \
+({ \
+	WARN_ON_ONCE(pud_present(*pudp) && !pud_same(*pudp, pud)); \
+	set_pud(pudp, pud); \
+})
+
+#define set_p4d_safe(p4dp, p4d) \
+({ \
+	WARN_ON_ONCE(p4d_present(*p4dp) && !p4d_same(*p4dp, p4d)); \
+	set_p4d(p4dp, p4d); \
+})
+
+#define set_pgd_safe(pgdp, pgd) \
+({ \
+	WARN_ON_ONCE(pgd_present(*pgdp) && !pgd_same(*pgdp, pgd)); \
+	set_pgd(pgdp, pgd); \
+})
+
 #ifndef __HAVE_ARCH_DO_SWAP_PAGE
 /*
  * Some architectures support metadata associated with a page. When a