From 12609013c43acaa5e547062ffc2fcfc4dcddde2e Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Mon, 18 Nov 2019 15:49:22 +0100
Subject: [PATCH 01/17] x86/setup: Clean up the header portion of setup.c

In 20 years we accumulated 89 #include lines in setup.c,
but we only need 30 of them (!) ...

Get rid of the excessive ones, and while at it, sort the
remaining ones alphabetically.

Also get rid of the incomplete changelogs at the header of the file,
and explain better what this file does.

Cc: linux-kernel@vger.kernel.org
Cc: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/kernel/setup.c | 126 ++++++++--------------------------------
 1 file changed, 24 insertions(+), 102 deletions(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index cedfe2077a69..7778ee065470 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -2,123 +2,45 @@
 /*
  *  Copyright (C) 1995  Linus Torvalds
  *
- *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
- *
- *  Memory region support
- *	David Parsons <orc@pell.chi.il.us>, July-August 1999
- *
- *  Added E820 sanitization routine (removes overlapping memory regions);
- *  Brian Moyle <bmoyle@mvista.com>, February 2001
- *
- * Moved CPU detection code to cpu/${cpu}.c
- *    Patrick Mochel <mochel@osdl.org>, March 2002
- *
- *  Provisions for empty E820 memory regions (reported by certain BIOSes).
- *  Alex Achenbach <xela@slit.de>, December 2002.
- *
+ * This file contains the setup_arch() code, which handles the architecture-dependent
+ * parts of early kernel initialization.
  */
-
-/*
- * This file handles the architecture-dependent parts of initialization
- */
-
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/mmzone.h>
-#include <linux/screen_info.h>
-#include <linux/ioport.h>
-#include <linux/acpi.h>
-#include <linux/sfi.h>
-#include <linux/apm_bios.h>
-#include <linux/initrd.h>
-#include <linux/memblock.h>
-#include <linux/seq_file.h>
 #include <linux/console.h>
-#include <linux/root_dev.h>
-#include <linux/highmem.h>
-#include <linux/export.h>
-#include <linux/efi.h>
-#include <linux/init.h>
-#include <linux/edd.h>
-#include <linux/iscsi_ibft.h>
-#include <linux/nodemask.h>
-#include <linux/kexec.h>
+#include <linux/crash_dump.h>
 #include <linux/dmi.h>
-#include <linux/pfn.h>
-#include <linux/pci.h>
-#include <asm/pci-direct.h>
+#include <linux/efi.h>
 #include <linux/init_ohci1394_dma.h>
-#include <linux/kvm_para.h>
-#include <linux/dma-contiguous.h>
-#include <xen/xen.h>
+#include <linux/initrd.h>
+#include <linux/iscsi_ibft.h>
+#include <linux/memblock.h>
+#include <linux/pci.h>
+#include <linux/root_dev.h>
+#include <linux/sfi.h>
+#include <linux/tboot.h>
+#include <linux/usb/xhci-dbgp.h>
+
 #include <uapi/linux/mount.h>
 
-#include <linux/errno.h>
-#include <linux/kernel.h>
-#include <linux/stddef.h>
-#include <linux/unistd.h>
-#include <linux/ptrace.h>
-#include <linux/user.h>
-#include <linux/delay.h>
+#include <xen/xen.h>
 
-#include <linux/kallsyms.h>
-#include <linux/cpufreq.h>
-#include <linux/dma-mapping.h>
-#include <linux/ctype.h>
-#include <linux/uaccess.h>
-
-#include <linux/percpu.h>
-#include <linux/crash_dump.h>
-#include <linux/tboot.h>
-#include <linux/jiffies.h>
-#include <linux/mem_encrypt.h>
-#include <linux/sizes.h>
-
-#include <linux/usb/xhci-dbgp.h>
-#include <video/edid.h>
-
-#include <asm/mtrr.h>
 #include <asm/apic.h>
-#include <asm/realmode.h>
-#include <asm/e820/api.h>
-#include <asm/mpspec.h>
-#include <asm/setup.h>
-#include <asm/efi.h>
-#include <asm/timer.h>
-#include <asm/i8259.h>
-#include <asm/sections.h>
-#include <asm/io_apic.h>
-#include <asm/ist.h>
-#include <asm/setup_arch.h>
 #include <asm/bios_ebda.h>
-#include <asm/cacheflush.h>
-#include <asm/processor.h>
 #include <asm/bugs.h>
-#include <asm/kasan.h>
-
-#include <asm/vsyscall.h>
 #include <asm/cpu.h>
-#include <asm/desc.h>
-#include <asm/dma.h>
-#include <asm/iommu.h>
+#include <asm/efi.h>
 #include <asm/gart.h>
-#include <asm/mmu_context.h>
-#include <asm/proto.h>
-
-#include <asm/paravirt.h>
 #include <asm/hypervisor.h>
-#include <asm/olpc_ofw.h>
-
-#include <asm/percpu.h>
-#include <asm/topology.h>
-#include <asm/apicdef.h>
-#include <asm/amd_nb.h>
-#include <asm/mce.h>
-#include <asm/alternative.h>
-#include <asm/prom.h>
-#include <asm/microcode.h>
+#include <asm/io_apic.h>
+#include <asm/kasan.h>
 #include <asm/kaslr.h>
+#include <asm/mce.h>
+#include <asm/mtrr.h>
+#include <asm/olpc_ofw.h>
+#include <asm/pci-direct.h>
+#include <asm/prom.h>
+#include <asm/proto.h>
 #include <asm/unwind.h>
+#include <asm/vsyscall.h>
 
 /*
  * max_low_pfn_mapped: highest direct mapped pfn under 4GB

From 360db4ace3117ac1d9936d529f59c653e337b0f5 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Mon, 18 Nov 2019 16:03:39 +0100
Subject: [PATCH 02/17] x86/setup: Enhance the comments

Update various comments, fix outright mistakes and meaningless descriptions.

Also harmonize the style across the file, both in form and in language.

Cc: linux-kernel@vger.kernel.org
Cc: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/kernel/setup.c | 41 +++++++++++++++++++++++++----------------
 1 file changed, 25 insertions(+), 16 deletions(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 7778ee065470..b5ac9932bcf6 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -43,11 +43,11 @@
 #include <asm/vsyscall.h>
 
 /*
- * max_low_pfn_mapped: highest direct mapped pfn under 4GB
- * max_pfn_mapped:     highest direct mapped pfn over 4GB
+ * max_low_pfn_mapped: highest directly mapped pfn < 4 GB
+ * max_pfn_mapped:     highest directly mapped pfn > 4 GB
  *
  * The direct mapping only covers E820_TYPE_RAM regions, so the ranges and gaps are
- * represented by pfn_mapped
+ * represented by pfn_mapped[].
  */
 unsigned long max_low_pfn_mapped;
 unsigned long max_pfn_mapped;
@@ -57,14 +57,23 @@ RESERVE_BRK(dmi_alloc, 65536);
 #endif
 
 
-static __initdata unsigned long _brk_start = (unsigned long)__brk_base;
-unsigned long _brk_end = (unsigned long)__brk_base;
+/*
+ * Range of the BSS area. The size of the BSS area is determined
+ * at link time, with RESERVE_BRK*() facility reserving additional
+ * chunks.
+ */
+static __initdata
+unsigned long _brk_start = (unsigned long)__brk_base;
+unsigned long _brk_end   = (unsigned long)__brk_base;
 
 struct boot_params boot_params;
 
 /*
- * Machine setup..
+ * These are the four main kernel memory regions, we put them into
+ * the resource tree so that kdump tools and other debugging tools
+ * recover it:
  */
+
 static struct resource rodata_resource = {
 	.name	= "Kernel rodata",
 	.start	= 0,
@@ -95,16 +104,16 @@ static struct resource bss_resource = {
 
 
 #ifdef CONFIG_X86_32
-/* cpu data as detected by the assembly code in head_32.S */
+/* CPU data as detected by the assembly code in head_32.S */
 struct cpuinfo_x86 new_cpu_data;
 
-/* common cpu data for all cpus */
+/* Common CPU data for all CPUs */
 struct cpuinfo_x86 boot_cpu_data __read_mostly;
 EXPORT_SYMBOL(boot_cpu_data);
 
 unsigned int def_to_bigsmp;
 
-/* for MCA, but anyone else can use it if they want */
+/* For MCA, but anyone else can use it if they want */
 unsigned int machine_id;
 unsigned int machine_submodel_id;
 unsigned int BIOS_revision;
@@ -390,15 +399,15 @@ static void __init memblock_x86_reserve_range_setup_data(void)
 /*
  * Keep the crash kernel below this limit.
  *
- * On 32 bits earlier kernels would limit the kernel to the low 512 MiB
+ * Earlier 32-bits kernels would limit the kernel to the low 512 MB range
  * due to mapping restrictions.
  *
- * On 64bit, kdump kernel need be restricted to be under 64TB, which is
+ * 64-bit kdump kernels need to be restricted to be under 64 TB, which is
  * the upper limit of system RAM in 4-level paging mode. Since the kdump
- * jumping could be from 5-level to 4-level, the jumping will fail if
- * kernel is put above 64TB, and there's no way to detect the paging mode
- * of the kernel which will be loaded for dumping during the 1st kernel
- * bootup.
+ * jump could be from 5-level paging to 4-level paging, the jump will fail if
+ * the kernel is put above 64 TB, and during the 1st kernel bootup there's
+ * no good way to detect the paging mode of the target kernel which will be
+ * loaded for dumping.
  */
 #ifdef CONFIG_X86_32
 # define CRASH_ADDR_LOW_MAX	SZ_512M
@@ -809,7 +818,7 @@ void __init setup_arch(char **cmdline_p)
 	/*
 	 * Note: Quark X1000 CPUs advertise PGE incorrectly and require
 	 * a cr3 based tlb flush, so the following __flush_tlb_all()
-	 * will not flush anything because the cpu quirk which clears
+	 * will not flush anything because the CPU quirk which clears
 	 * X86_FEATURE_PGE has not been invoked yet. Though due to the
 	 * load_cr3() above the TLB has been flushed already. The
 	 * quirk is invoked before subsequent calls to __flush_tlb_all()

From aee7f91369a80d2cb9bba198331479cc9bfc0ade Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 10 Dec 2019 10:05:45 +0100
Subject: [PATCH 03/17] x86/mm/pat: Update the comments in pat.c and
 pat_interval.c and refresh the code a bit

Tidy up the code:

 - add comments explaining the PAT code, the role of the functions and the logic

 - fix various typos and grammar while at it

 - simplify the file-scope memtype_interval_*() namespace to interval_*()

 - simplify stylistic complications such as unnecessary linebreaks
   or convoluted control flow

 - use the simpler '#ifdef CONFIG_*' pattern instead of '#if defined(CONFIG_*)' pattern

 - remove the non-idiomatic newline between late_initcall() and its function definition

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/mm/pat.c          | 53 ++++++++++++++++++++++++++++---------
 arch/x86/mm/pat_interval.c | 54 ++++++++++++++++++++++----------------
 2 files changed, 72 insertions(+), 35 deletions(-)

diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index 2d758e19ef22..560ac5137a2f 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -1,11 +1,34 @@
 // SPDX-License-Identifier: GPL-2.0-only
 /*
- * Handle caching attributes in page tables (PAT)
+ * Page Attribute Table (PAT) support: handle memory caching attributes in page tables.
  *
  * Authors: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
  *          Suresh B Siddha <suresh.b.siddha@intel.com>
  *
  * Loosely based on earlier PAT patchset from Eric Biederman and Andi Kleen.
+ *
+ * Basic principles:
+ *
+ * PAT is a CPU feature supported by all modern x86 CPUs, to allow the firmware and
+ * the kernel to set one of a handful of 'caching type' attributes for physical
+ * memory ranges: uncached, write-combining, write-through, write-protected,
+ * and the most commonly used and default attribute: write-back caching.
+ *
+ * PAT support supercedes and augments MTRR support in a compatible fashion: MTRR is
+ * a hardware interface to enumerate a limited number of physical memory ranges
+ * and set their caching attributes explicitly, programmed into the CPU via MSRs.
+ * Even modern CPUs have MTRRs enabled - but these are typically not touched
+ * by the kernel or by user-space (such as the X server), we rely on PAT for any
+ * additional cache attribute logic.
+ *
+ * PAT doesn't work via explicit memory ranges, but uses page table entries to add
+ * cache attribute information to the mapped memory range: there's 3 bits used,
+ * (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT), with the 8 possible values mapped by the
+ * CPU to actual cache attributes via an MSR loaded into the CPU (MSR_IA32_CR_PAT).
+ *
+ * ( There's a metric ton of finer details, such as compatibility with CPU quirks
+ *   that only support 4 types of PAT entries, and interaction with MTRRs, see
+ *   below for details. )
  */
 
 #include <linux/seq_file.h>
@@ -839,7 +862,7 @@ int phys_mem_access_prot_allowed(struct file *file, unsigned long pfn,
 }
 
 /*
- * Change the memory type for the physial address range in kernel identity
+ * Change the memory type for the physical address range in kernel identity
  * mapping space if that range is a part of identity map.
  */
 int kernel_map_sync_memtype(u64 base, unsigned long size,
@@ -851,15 +874,14 @@ int kernel_map_sync_memtype(u64 base, unsigned long size,
 		return 0;
 
 	/*
-	 * some areas in the middle of the kernel identity range
-	 * are not mapped, like the PCI space.
+	 * Some areas in the middle of the kernel identity range
+	 * are not mapped, for example the PCI space.
 	 */
 	if (!page_is_ram(base >> PAGE_SHIFT))
 		return 0;
 
 	id_sz = (__pa(high_memory-1) <= base + size) ?
-				__pa(high_memory) - base :
-				size;
+				__pa(high_memory) - base : size;
 
 	if (ioremap_change_attr((unsigned long)__va(base), id_sz, pcm) < 0) {
 		pr_info("x86/PAT: %s:%d ioremap_change_attr failed %s for [mem %#010Lx-%#010Lx]\n",
@@ -1099,6 +1121,10 @@ EXPORT_SYMBOL_GPL(pgprot_writethrough);
 
 #if defined(CONFIG_DEBUG_FS) && defined(CONFIG_X86_PAT)
 
+/*
+ * We are allocating a temporary printout-entry to be passed
+ * between seq_start()/next() and seq_show():
+ */
 static struct memtype *memtype_get_idx(loff_t pos)
 {
 	struct memtype *print_entry;
@@ -1112,12 +1138,13 @@ static struct memtype *memtype_get_idx(loff_t pos)
 	ret = memtype_copy_nth_element(print_entry, pos);
 	spin_unlock(&memtype_lock);
 
-	if (!ret) {
-		return print_entry;
-	} else {
+	/* Free it on error: */
+	if (ret) {
 		kfree(print_entry);
 		return NULL;
 	}
+
+	return print_entry;
 }
 
 static void *memtype_seq_start(struct seq_file *seq, loff_t *pos)
@@ -1144,8 +1171,11 @@ static int memtype_seq_show(struct seq_file *seq, void *v)
 {
 	struct memtype *print_entry = (struct memtype *)v;
 
-	seq_printf(seq, "%s @ 0x%Lx-0x%Lx\n", cattr_name(print_entry->type),
-			print_entry->start, print_entry->end);
+	seq_printf(seq, "%s @ 0x%Lx-0x%Lx\n",
+			cattr_name(print_entry->type),
+			print_entry->start,
+			print_entry->end);
+
 	kfree(print_entry);
 
 	return 0;
@@ -1178,7 +1208,6 @@ static int __init pat_memtype_list_init(void)
 	}
 	return 0;
 }
-
 late_initcall(pat_memtype_list_init);
 
 #endif /* CONFIG_DEBUG_FS && CONFIG_X86_PAT */
diff --git a/arch/x86/mm/pat_interval.c b/arch/x86/mm/pat_interval.c
index 6855362eaf21..2abc47558818 100644
--- a/arch/x86/mm/pat_interval.c
+++ b/arch/x86/mm/pat_interval.c
@@ -25,25 +25,27 @@
  * physical memory areas. Without proper tracking, conflicting memory
  * types in different mappings can cause CPU cache corruption.
  *
- * The tree is an interval tree (augmented rbtree) with tree ordered
- * on starting address. Tree can contain multiple entries for
+ * The tree is an interval tree (augmented rbtree) which tree is ordered
+ * by the starting address. The tree can contain multiple entries for
  * different regions which overlap. All the aliases have the same
- * cache attributes of course.
+ * cache attributes of course, as enforced by the PAT logic.
  *
  * memtype_lock protects the rbtree.
  */
-static inline u64 memtype_interval_start(struct memtype *memtype)
+
+static inline u64 interval_start(struct memtype *memtype)
 {
 	return memtype->start;
 }
 
-static inline u64 memtype_interval_end(struct memtype *memtype)
+static inline u64 interval_end(struct memtype *memtype)
 {
 	return memtype->end - 1;
 }
+
 INTERVAL_TREE_DEFINE(struct memtype, rb, u64, subtree_max_end,
-		     memtype_interval_start, memtype_interval_end,
-		     static, memtype_interval)
+		     interval_start, interval_end,
+		     static, interval)
 
 static struct rb_root_cached memtype_rbroot = RB_ROOT_CACHED;
 
@@ -56,7 +58,7 @@ static struct memtype *memtype_match(u64 start, u64 end, int match_type)
 {
 	struct memtype *match;
 
-	match = memtype_interval_iter_first(&memtype_rbroot, start, end-1);
+	match = interval_iter_first(&memtype_rbroot, start, end-1);
 	while (match != NULL && match->start < end) {
 		if ((match_type == MEMTYPE_EXACT_MATCH) &&
 		    (match->start == start) && (match->end == end))
@@ -66,7 +68,7 @@ static struct memtype *memtype_match(u64 start, u64 end, int match_type)
 		    (match->start < start) && (match->end == end))
 			return match;
 
-		match = memtype_interval_iter_next(match, start, end-1);
+		match = interval_iter_next(match, start, end-1);
 	}
 
 	return NULL; /* Returns NULL if there is no match */
@@ -79,7 +81,7 @@ static int memtype_check_conflict(u64 start, u64 end,
 	struct memtype *match;
 	enum page_cache_mode found_type = reqtype;
 
-	match = memtype_interval_iter_first(&memtype_rbroot, start, end-1);
+	match = interval_iter_first(&memtype_rbroot, start, end-1);
 	if (match == NULL)
 		goto success;
 
@@ -89,12 +91,12 @@ static int memtype_check_conflict(u64 start, u64 end,
 	dprintk("Overlap at 0x%Lx-0x%Lx\n", match->start, match->end);
 	found_type = match->type;
 
-	match = memtype_interval_iter_next(match, start, end-1);
+	match = interval_iter_next(match, start, end-1);
 	while (match) {
 		if (match->type != found_type)
 			goto failure;
 
-		match = memtype_interval_iter_next(match, start, end-1);
+		match = interval_iter_next(match, start, end-1);
 	}
 success:
 	if (newtype)
@@ -106,11 +108,11 @@ failure:
 	pr_info("x86/PAT: %s:%d conflicting memory types %Lx-%Lx %s<->%s\n",
 		current->comm, current->pid, start, end,
 		cattr_name(found_type), cattr_name(match->type));
+
 	return -EBUSY;
 }
 
-int memtype_check_insert(struct memtype *new,
-			 enum page_cache_mode *ret_type)
+int memtype_check_insert(struct memtype *new, enum page_cache_mode *ret_type)
 {
 	int err = 0;
 
@@ -121,7 +123,7 @@ int memtype_check_insert(struct memtype *new,
 	if (ret_type)
 		new->type = *ret_type;
 
-	memtype_interval_insert(new, &memtype_rbroot);
+	interval_insert(new, &memtype_rbroot);
 	return 0;
 }
 
@@ -145,12 +147,13 @@ struct memtype *memtype_erase(u64 start, u64 end)
 
 	if (data->start == start) {
 		/* munmap: erase this node */
-		memtype_interval_remove(data, &memtype_rbroot);
+		interval_remove(data, &memtype_rbroot);
 	} else {
 		/* mremap: update the end value of this node */
-		memtype_interval_remove(data, &memtype_rbroot);
+		interval_remove(data, &memtype_rbroot);
 		data->end = start;
-		memtype_interval_insert(data, &memtype_rbroot);
+		interval_insert(data, &memtype_rbroot);
+
 		return NULL;
 	}
 
@@ -159,19 +162,24 @@ struct memtype *memtype_erase(u64 start, u64 end)
 
 struct memtype *memtype_lookup(u64 addr)
 {
-	return memtype_interval_iter_first(&memtype_rbroot, addr,
-					   addr + PAGE_SIZE-1);
+	return interval_iter_first(&memtype_rbroot, addr, addr + PAGE_SIZE-1);
 }
 
-#if defined(CONFIG_DEBUG_FS)
+/*
+ * Debugging helper, copy the Nth entry of the tree into a
+ * a copy for printout. This allows us to print out the tree
+ * via debugfs, without holding the memtype_lock too long:
+ */
+#ifdef CONFIG_DEBUG_FS
 int memtype_copy_nth_element(struct memtype *out, loff_t pos)
 {
 	struct memtype *match;
 	int i = 1;
 
-	match = memtype_interval_iter_first(&memtype_rbroot, 0, ULONG_MAX);
+	match = interval_iter_first(&memtype_rbroot, 0, ULONG_MAX);
+
 	while (match && pos != i) {
-		match = memtype_interval_iter_next(match, 0, ULONG_MAX);
+		match = interval_iter_next(match, 0, ULONG_MAX);
 		i++;
 	}
 

From 5557e831f68878ef202c0e5296235442cff9b41e Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 19 Nov 2019 09:50:19 +0100
Subject: [PATCH 04/17] x86/mm/pat: Disambiguate PAT-disabled boot messages

Right now we have these four types of PAT-disabled boot messages:

  x86/PAT: PAT support disabled.
  x86/PAT: PAT MSR is 0, disabled.
  x86/PAT: MTRRs disabled, skipping PAT initialization too.
  x86/PAT: PAT not supported by CPU.

The first message is ambiguous in that it doesn't signal that PAT is off
due to a boot option.

The second message doesn't really make it clear that this is the MSR value
during early bootup and it's the firmware environment that disabled PAT
support.

The fourth message doesn't really make it clear that we disable PAT support
because CONFIG_MTRR is off in the kernel.

Clarify, harmonize and fix the spelling in these user-visible messages:

  x86/PAT: PAT support disabled via boot option.
  x86/PAT: PAT support disabled by the firmware.
  x86/PAT: PAT support disabled because CONFIG_MTRR is disabled in the kernel.
  x86/PAT: PAT not supported by the CPU.

Also add a fifth message, in case PAT support is disabled at build time:

  x86/PAT: PAT support disabled because CONFIG_X86_PAT is disabled in the kernel.

Previously we'd just silently return from pat_init() without giving any indication
that PAT support is off.

Finally, clarify/extend some of the comments related to PAT initialization.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/mtrr.h |  2 +-
 arch/x86/mm/pat.c           | 20 ++++++++++++++------
 2 files changed, 15 insertions(+), 7 deletions(-)

diff --git a/arch/x86/include/asm/mtrr.h b/arch/x86/include/asm/mtrr.h
index dbff1456d215..3337d2233aef 100644
--- a/arch/x86/include/asm/mtrr.h
+++ b/arch/x86/include/asm/mtrr.h
@@ -86,7 +86,7 @@ static inline void mtrr_centaur_report_mcr(int mcr, u32 lo, u32 hi)
 }
 static inline void mtrr_bp_init(void)
 {
-	pat_disable("MTRRs disabled, skipping PAT initialization too.");
+	pat_disable("PAT support disabled because CONFIG_MTRR is disabled in the kernel.");
 }
 
 #define mtrr_ap_init() do {} while (0)
diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index 560ac5137a2f..e26b81ca5b95 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -66,7 +66,11 @@ static bool __read_mostly pat_disabled = !IS_ENABLED(CONFIG_X86_PAT);
 static bool __read_mostly pat_initialized;
 static bool __read_mostly init_cm_done;
 
-void pat_disable(const char *reason)
+/*
+ * PAT support is enabled by default, but can be disabled for
+ * various user-requested or hardware-forced reasons:
+ */
+void pat_disable(const char *msg_reason)
 {
 	if (pat_disabled)
 		return;
@@ -77,12 +81,12 @@ void pat_disable(const char *reason)
 	}
 
 	pat_disabled = true;
-	pr_info("x86/PAT: %s\n", reason);
+	pr_info("x86/PAT: %s\n", msg_reason);
 }
 
 static int __init nopat(char *str)
 {
-	pat_disable("PAT support disabled.");
+	pat_disable("PAT support disabled via boot option.");
 	return 0;
 }
 early_param("nopat", nopat);
@@ -238,13 +242,13 @@ static void pat_bsp_init(u64 pat)
 	u64 tmp_pat;
 
 	if (!boot_cpu_has(X86_FEATURE_PAT)) {
-		pat_disable("PAT not supported by CPU.");
+		pat_disable("PAT not supported by the CPU.");
 		return;
 	}
 
 	rdmsrl(MSR_IA32_CR_PAT, tmp_pat);
 	if (!tmp_pat) {
-		pat_disable("PAT MSR is 0, disabled.");
+		pat_disable("PAT support disabled by the firmware.");
 		return;
 	}
 
@@ -314,7 +318,7 @@ void init_cache_modes(void)
 }
 
 /**
- * pat_init - Initialize PAT MSR and PAT table
+ * pat_init - Initialize the PAT MSR and PAT table on the current CPU
  *
  * This function initializes PAT MSR and PAT table with an OS-defined value
  * to enable additional cache attributes, WC, WT and WP.
@@ -328,6 +332,10 @@ void pat_init(void)
 	u64 pat;
 	struct cpuinfo_x86 *c = &boot_cpu_data;
 
+#ifndef CONFIG_X86_PAT
+	pr_info_once("x86/PAT: PAT support disabled because CONFIG_X86_PAT is disabled in the kernel.\n");
+#endif
+
 	if (pat_disabled)
 		return;
 

From ef35b0fcee23d7636daa4bed62bceeaba4eed918 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 19 Nov 2019 09:38:26 +0100
Subject: [PATCH 05/17] x86/mm/pat: Create fixed width output in
 /sys/kernel/debug/x86/pat_memtype_list, similar to the E820 debug printouts

Before:

  write-back @ 0xbdfa9000-0xbdfaa000
  write-back @ 0xbdfaa000-0xbdfab000
  write-back @ 0xbdfab000-0xbdfac000
  uncached-minus @ 0xc0000000-0xd0000000
  uncached-minus @ 0xd0900000-0xd0920000
  uncached-minus @ 0xd0920000-0xd0940000
  uncached-minus @ 0xd0940000-0xd0960000
  uncached-minus @ 0xd0960000-0xd0980000
  uncached-minus @ 0xd0980000-0xd0981000
  uncached-minus @ 0xd0990000-0xd0991000
  uncached-minus @ 0xd09a0000-0xd09a1000
  uncached-minus @ 0xd09b0000-0xd09b1000
  uncached-minus @ 0xd0d00000-0xd0d01000
  uncached-minus @ 0xd0d10000-0xd0d11000
  uncached-minus @ 0xd0d20000-0xd0d21000
  uncached-minus @ 0xd0d40000-0xd0d41000
  uncached-minus @ 0xfed00000-0xfed01000
  uncached-minus @ 0xfed1f000-0xfed20000
  uncached-minus @ 0xfed40000-0xfed41000

After:

  PAT: [mem 0x00000000bdf8e000-0x00000000bdfa8000] write-back
  PAT: [mem 0x00000000bdfa9000-0x00000000bdfaa000] write-back
  PAT: [mem 0x00000000bdfaa000-0x00000000bdfab000] write-back
  PAT: [mem 0x00000000bdfab000-0x00000000bdfac000] write-back
  PAT: [mem 0x00000000c0000000-0x00000000d0000000] uncached-minus
  PAT: [mem 0x00000000d0900000-0x00000000d0920000] uncached-minus
  PAT: [mem 0x00000000d0920000-0x00000000d0940000] uncached-minus
  PAT: [mem 0x00000000d0940000-0x00000000d0960000] uncached-minus
  PAT: [mem 0x00000000d0960000-0x00000000d0980000] uncached-minus
  PAT: [mem 0x00000000d0980000-0x00000000d0981000] uncached-minus
  PAT: [mem 0x00000000d0990000-0x00000000d0991000] uncached-minus
  PAT: [mem 0x00000000d09a0000-0x00000000d09a1000] uncached-minus
  PAT: [mem 0x00000000d09b0000-0x00000000d09b1000] uncached-minus
  PAT: [mem 0x00000000fed1f000-0x00000000fed20000] uncached-minus
  PAT: [mem 0x00000000fed40000-0x00000000fed41000] uncached-minus

The advantage is that it's easier to parse at a glance - and the tree is ordered
by start address, which is now reflected in putting the start address in the first
column.

This is also now similar to how we print e820 entries:

  BIOS-e820: [mem 0x00000000bafda000-0x00000000bb3d3fff] usable
  BIOS-e820: [mem 0x00000000bb3d4000-0x00000000bdd2efff] reserved
  BIOS-e820: [mem 0x00000000bdd2f000-0x00000000bddccfff] ACPI NVS
  BIOS-e820: [mem 0x00000000bddcd000-0x00000000bdea0fff] ACPI data
  BIOS-e820: [mem 0x00000000bdea1000-0x00000000bdf2efff] ACPI NVS

Since this is a debugfs file not used by tools there's no known ABI dependencies.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/mm/pat.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index e26b81ca5b95..ea7da7e62e17 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -1179,10 +1179,10 @@ static int memtype_seq_show(struct seq_file *seq, void *v)
 {
 	struct memtype *print_entry = (struct memtype *)v;
 
-	seq_printf(seq, "%s @ 0x%Lx-0x%Lx\n",
-			cattr_name(print_entry->type),
+	seq_printf(seq, "PAT: [mem 0x%016Lx-0x%016Lx] %s\n",
 			print_entry->start,
-			print_entry->end);
+			print_entry->end,
+			cattr_name(print_entry->type));
 
 	kfree(print_entry);
 

From 47553d42c55f7b85e72ce6f3a18030102f8f93a3 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 19 Nov 2019 10:18:56 +0100
Subject: [PATCH 06/17] x86/mm/pat: Simplify the free_memtype() control flow

Simplify/streamline the quirky handling of the pat_pagerange_is_ram() logic,
and get rid of the 'err' local variable.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/mm/pat.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index ea7da7e62e17..af049920e59a 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -656,7 +656,6 @@ int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 
 int free_memtype(u64 start, u64 end)
 {
-	int err = -EINVAL;
 	int is_range_ram;
 	struct memtype *entry;
 
@@ -671,14 +670,10 @@ int free_memtype(u64 start, u64 end)
 		return 0;
 
 	is_range_ram = pat_pagerange_is_ram(start, end);
-	if (is_range_ram == 1) {
-
-		err = free_ram_pages_type(start, end);
-
-		return err;
-	} else if (is_range_ram < 0) {
+	if (is_range_ram == 1)
+		return free_ram_pages_type(start, end);
+	if (is_range_ram < 0)
 		return -EINVAL;
-	}
 
 	spin_lock(&memtype_lock);
 	entry = memtype_erase(start, end);

From baf65855baac302bbacf0ce17ed99b9c0940b930 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 10 Dec 2019 10:07:23 +0100
Subject: [PATCH 07/17] x86/mm/pat: Harmonize 'struct memtype *' local variable
 and function parameter use

We have quite a zoo of 'struct memtype' variable nomenclature:

  new
  entry
  print_entry
  data
  match
  out
  memtype

Beyond the randomness, some of these are outright confusing, especially
when used in larger functions.

Standardize them:

  entry
  entry_new
  entry_old
  entry_print
  entry_match
  entry_out

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/mm/pat.c          | 51 ++++++++++-----------
 arch/x86/mm/pat_internal.h |  6 +--
 arch/x86/mm/pat_interval.c | 91 +++++++++++++++++++-------------------
 3 files changed, 75 insertions(+), 73 deletions(-)

diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index af049920e59a..4a1804903d97 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -576,7 +576,7 @@ static u64 sanitize_phys(u64 address)
 int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 		    enum page_cache_mode *new_type)
 {
-	struct memtype *new;
+	struct memtype *entry_new;
 	enum page_cache_mode actual_type;
 	int is_range_ram;
 	int err = 0;
@@ -624,22 +624,22 @@ int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 		return -EINVAL;
 	}
 
-	new  = kzalloc(sizeof(struct memtype), GFP_KERNEL);
-	if (!new)
+	entry_new = kzalloc(sizeof(struct memtype), GFP_KERNEL);
+	if (!entry_new)
 		return -ENOMEM;
 
-	new->start	= start;
-	new->end	= end;
-	new->type	= actual_type;
+	entry_new->start = start;
+	entry_new->end	 = end;
+	entry_new->type	 = actual_type;
 
 	spin_lock(&memtype_lock);
 
-	err = memtype_check_insert(new, new_type);
+	err = memtype_check_insert(entry_new, new_type);
 	if (err) {
 		pr_info("x86/PAT: reserve_memtype failed [mem %#010Lx-%#010Lx], track %s, req %s\n",
 			start, end - 1,
-			cattr_name(new->type), cattr_name(req_type));
-		kfree(new);
+			cattr_name(entry_new->type), cattr_name(req_type));
+		kfree(entry_new);
 		spin_unlock(&memtype_lock);
 
 		return err;
@@ -648,7 +648,7 @@ int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 	spin_unlock(&memtype_lock);
 
 	dprintk("reserve_memtype added [mem %#010Lx-%#010Lx], track %s, req %s, ret %s\n",
-		start, end - 1, cattr_name(new->type), cattr_name(req_type),
+		start, end - 1, cattr_name(entry_new->type), cattr_name(req_type),
 		new_type ? cattr_name(*new_type) : "-");
 
 	return err;
@@ -657,7 +657,7 @@ int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 int free_memtype(u64 start, u64 end)
 {
 	int is_range_ram;
-	struct memtype *entry;
+	struct memtype *entry_old;
 
 	if (!pat_enabled())
 		return 0;
@@ -676,16 +676,16 @@ int free_memtype(u64 start, u64 end)
 		return -EINVAL;
 
 	spin_lock(&memtype_lock);
-	entry = memtype_erase(start, end);
+	entry_old = memtype_erase(start, end);
 	spin_unlock(&memtype_lock);
 
-	if (IS_ERR(entry)) {
+	if (IS_ERR(entry_old)) {
 		pr_info("x86/PAT: %s:%d freeing invalid memtype [mem %#010Lx-%#010Lx]\n",
 			current->comm, current->pid, start, end - 1);
 		return -EINVAL;
 	}
 
-	kfree(entry);
+	kfree(entry_old);
 
 	dprintk("free_memtype request [mem %#010Lx-%#010Lx]\n", start, end - 1);
 
@@ -726,6 +726,7 @@ static enum page_cache_mode lookup_memtype(u64 paddr)
 		rettype = _PAGE_CACHE_MODE_UC_MINUS;
 
 	spin_unlock(&memtype_lock);
+
 	return rettype;
 }
 
@@ -1130,24 +1131,24 @@ EXPORT_SYMBOL_GPL(pgprot_writethrough);
  */
 static struct memtype *memtype_get_idx(loff_t pos)
 {
-	struct memtype *print_entry;
+	struct memtype *entry_print;
 	int ret;
 
-	print_entry  = kzalloc(sizeof(struct memtype), GFP_KERNEL);
-	if (!print_entry)
+	entry_print  = kzalloc(sizeof(struct memtype), GFP_KERNEL);
+	if (!entry_print)
 		return NULL;
 
 	spin_lock(&memtype_lock);
-	ret = memtype_copy_nth_element(print_entry, pos);
+	ret = memtype_copy_nth_element(entry_print, pos);
 	spin_unlock(&memtype_lock);
 
 	/* Free it on error: */
 	if (ret) {
-		kfree(print_entry);
+		kfree(entry_print);
 		return NULL;
 	}
 
-	return print_entry;
+	return entry_print;
 }
 
 static void *memtype_seq_start(struct seq_file *seq, loff_t *pos)
@@ -1172,14 +1173,14 @@ static void memtype_seq_stop(struct seq_file *seq, void *v)
 
 static int memtype_seq_show(struct seq_file *seq, void *v)
 {
-	struct memtype *print_entry = (struct memtype *)v;
+	struct memtype *entry_print = (struct memtype *)v;
 
 	seq_printf(seq, "PAT: [mem 0x%016Lx-0x%016Lx] %s\n",
-			print_entry->start,
-			print_entry->end,
-			cattr_name(print_entry->type));
+			entry_print->start,
+			entry_print->end,
+			cattr_name(entry_print->type));
 
-	kfree(print_entry);
+	kfree(entry_print);
 
 	return 0;
 }
diff --git a/arch/x86/mm/pat_internal.h b/arch/x86/mm/pat_internal.h
index 79a06684349e..23ce8cdac159 100644
--- a/arch/x86/mm/pat_internal.h
+++ b/arch/x86/mm/pat_internal.h
@@ -29,13 +29,13 @@ static inline char *cattr_name(enum page_cache_mode pcm)
 }
 
 #ifdef CONFIG_X86_PAT
-extern int memtype_check_insert(struct memtype *new,
+extern int memtype_check_insert(struct memtype *entry_new,
 				enum page_cache_mode *new_type);
 extern struct memtype *memtype_erase(u64 start, u64 end);
 extern struct memtype *memtype_lookup(u64 addr);
-extern int memtype_copy_nth_element(struct memtype *out, loff_t pos);
+extern int memtype_copy_nth_element(struct memtype *entry_out, loff_t pos);
 #else
-static inline int memtype_check_insert(struct memtype *new,
+static inline int memtype_check_insert(struct memtype *entry_new,
 				       enum page_cache_mode *new_type)
 { return 0; }
 static inline struct memtype *memtype_erase(u64 start, u64 end)
diff --git a/arch/x86/mm/pat_interval.c b/arch/x86/mm/pat_interval.c
index 2abc47558818..3c983de5f492 100644
--- a/arch/x86/mm/pat_interval.c
+++ b/arch/x86/mm/pat_interval.c
@@ -33,14 +33,14 @@
  * memtype_lock protects the rbtree.
  */
 
-static inline u64 interval_start(struct memtype *memtype)
+static inline u64 interval_start(struct memtype *entry)
 {
-	return memtype->start;
+	return entry->start;
 }
 
-static inline u64 interval_end(struct memtype *memtype)
+static inline u64 interval_end(struct memtype *entry)
 {
-	return memtype->end - 1;
+	return entry->end - 1;
 }
 
 INTERVAL_TREE_DEFINE(struct memtype, rb, u64, subtree_max_end,
@@ -56,19 +56,20 @@ enum {
 
 static struct memtype *memtype_match(u64 start, u64 end, int match_type)
 {
-	struct memtype *match;
+	struct memtype *entry_match;
 
-	match = interval_iter_first(&memtype_rbroot, start, end-1);
-	while (match != NULL && match->start < end) {
+	entry_match = interval_iter_first(&memtype_rbroot, start, end-1);
+
+	while (entry_match != NULL && entry_match->start < end) {
 		if ((match_type == MEMTYPE_EXACT_MATCH) &&
-		    (match->start == start) && (match->end == end))
-			return match;
+		    (entry_match->start == start) && (entry_match->end == end))
+			return entry_match;
 
 		if ((match_type == MEMTYPE_END_MATCH) &&
-		    (match->start < start) && (match->end == end))
-			return match;
+		    (entry_match->start < start) && (entry_match->end == end))
+			return entry_match;
 
-		match = interval_iter_next(match, start, end-1);
+		entry_match = interval_iter_next(entry_match, start, end-1);
 	}
 
 	return NULL; /* Returns NULL if there is no match */
@@ -78,25 +79,25 @@ static int memtype_check_conflict(u64 start, u64 end,
 				  enum page_cache_mode reqtype,
 				  enum page_cache_mode *newtype)
 {
-	struct memtype *match;
+	struct memtype *entry_match;
 	enum page_cache_mode found_type = reqtype;
 
-	match = interval_iter_first(&memtype_rbroot, start, end-1);
-	if (match == NULL)
+	entry_match = interval_iter_first(&memtype_rbroot, start, end-1);
+	if (entry_match == NULL)
 		goto success;
 
-	if (match->type != found_type && newtype == NULL)
+	if (entry_match->type != found_type && newtype == NULL)
 		goto failure;
 
-	dprintk("Overlap at 0x%Lx-0x%Lx\n", match->start, match->end);
-	found_type = match->type;
+	dprintk("Overlap at 0x%Lx-0x%Lx\n", entry_match->start, entry_match->end);
+	found_type = entry_match->type;
 
-	match = interval_iter_next(match, start, end-1);
-	while (match) {
-		if (match->type != found_type)
+	entry_match = interval_iter_next(entry_match, start, end-1);
+	while (entry_match) {
+		if (entry_match->type != found_type)
 			goto failure;
 
-		match = interval_iter_next(match, start, end-1);
+		entry_match = interval_iter_next(entry_match, start, end-1);
 	}
 success:
 	if (newtype)
@@ -107,29 +108,29 @@ success:
 failure:
 	pr_info("x86/PAT: %s:%d conflicting memory types %Lx-%Lx %s<->%s\n",
 		current->comm, current->pid, start, end,
-		cattr_name(found_type), cattr_name(match->type));
+		cattr_name(found_type), cattr_name(entry_match->type));
 
 	return -EBUSY;
 }
 
-int memtype_check_insert(struct memtype *new, enum page_cache_mode *ret_type)
+int memtype_check_insert(struct memtype *entry_new, enum page_cache_mode *ret_type)
 {
 	int err = 0;
 
-	err = memtype_check_conflict(new->start, new->end, new->type, ret_type);
+	err = memtype_check_conflict(entry_new->start, entry_new->end, entry_new->type, ret_type);
 	if (err)
 		return err;
 
 	if (ret_type)
-		new->type = *ret_type;
+		entry_new->type = *ret_type;
 
-	interval_insert(new, &memtype_rbroot);
+	interval_insert(entry_new, &memtype_rbroot);
 	return 0;
 }
 
 struct memtype *memtype_erase(u64 start, u64 end)
 {
-	struct memtype *data;
+	struct memtype *entry_old;
 
 	/*
 	 * Since the memtype_rbroot tree allows overlapping ranges,
@@ -138,26 +139,26 @@ struct memtype *memtype_erase(u64 start, u64 end)
 	 * it then checks with END_MATCH, i.e. shrink the size of a node
 	 * from the end for the mremap case.
 	 */
-	data = memtype_match(start, end, MEMTYPE_EXACT_MATCH);
-	if (!data) {
-		data = memtype_match(start, end, MEMTYPE_END_MATCH);
-		if (!data)
+	entry_old = memtype_match(start, end, MEMTYPE_EXACT_MATCH);
+	if (!entry_old) {
+		entry_old = memtype_match(start, end, MEMTYPE_END_MATCH);
+		if (!entry_old)
 			return ERR_PTR(-EINVAL);
 	}
 
-	if (data->start == start) {
+	if (entry_old->start == start) {
 		/* munmap: erase this node */
-		interval_remove(data, &memtype_rbroot);
+		interval_remove(entry_old, &memtype_rbroot);
 	} else {
 		/* mremap: update the end value of this node */
-		interval_remove(data, &memtype_rbroot);
-		data->end = start;
-		interval_insert(data, &memtype_rbroot);
+		interval_remove(entry_old, &memtype_rbroot);
+		entry_old->end = start;
+		interval_insert(entry_old, &memtype_rbroot);
 
 		return NULL;
 	}
 
-	return data;
+	return entry_old;
 }
 
 struct memtype *memtype_lookup(u64 addr)
@@ -171,20 +172,20 @@ struct memtype *memtype_lookup(u64 addr)
  * via debugfs, without holding the memtype_lock too long:
  */
 #ifdef CONFIG_DEBUG_FS
-int memtype_copy_nth_element(struct memtype *out, loff_t pos)
+int memtype_copy_nth_element(struct memtype *entry_out, loff_t pos)
 {
-	struct memtype *match;
+	struct memtype *entry_match;
 	int i = 1;
 
-	match = interval_iter_first(&memtype_rbroot, 0, ULONG_MAX);
+	entry_match = interval_iter_first(&memtype_rbroot, 0, ULONG_MAX);
 
-	while (match && pos != i) {
-		match = interval_iter_next(match, 0, ULONG_MAX);
+	while (entry_match && pos != i) {
+		entry_match = interval_iter_next(entry_match, 0, ULONG_MAX);
 		i++;
 	}
 
-	if (match) { /* pos == i */
-		*out = *match;
+	if (entry_match) { /* pos == i */
+		*entry_out = *entry_match;
 		return 0;
 	} else {
 		return 1;

From d891b9219d2a73640d2f2a216ecb6fb29d832013 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 19 Nov 2019 11:46:36 +0100
Subject: [PATCH 08/17] x86/mm/pat: Clean up PAT initialization flags

Right now we have these variables that impact the PAT initialization sequence:

  pat_disabled
  boot_cpu_done
  pat_initialized
  init_cm_done

Some have a pat_ prefix, some not, and the naming is random,
which makes their purpose rather opaque.

Name them consistently and according to their role:

  pat_disabled
  pat_bp_initialized
  pat_bp_enabled
  pat_cm_initialized

Also rename pat_bsp_init() => pat_bp_init(), to use the canonical
abbreviation.

Also add a warning for double calls of init_cache_modes(), the call chains
leading to this are complex and I couldn't convince myself that we never
call this function twice - so utilize the flag for a debug check.

No change in functionality intended.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/mm/pat.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index 4a1804903d97..f1677fad1735 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -61,10 +61,10 @@
 #undef pr_fmt
 #define pr_fmt(fmt) "" fmt
 
-static bool __read_mostly boot_cpu_done;
+static bool __read_mostly pat_bp_initialized;
 static bool __read_mostly pat_disabled = !IS_ENABLED(CONFIG_X86_PAT);
-static bool __read_mostly pat_initialized;
-static bool __read_mostly init_cm_done;
+static bool __read_mostly pat_bp_enabled;
+static bool __read_mostly pat_cm_initialized;
 
 /*
  * PAT support is enabled by default, but can be disabled for
@@ -75,7 +75,7 @@ void pat_disable(const char *msg_reason)
 	if (pat_disabled)
 		return;
 
-	if (boot_cpu_done) {
+	if (pat_bp_initialized) {
 		WARN_ONCE(1, "x86/PAT: PAT cannot be disabled after initialization\n");
 		return;
 	}
@@ -93,7 +93,7 @@ early_param("nopat", nopat);
 
 bool pat_enabled(void)
 {
-	return pat_initialized;
+	return pat_bp_enabled;
 }
 EXPORT_SYMBOL_GPL(pat_enabled);
 
@@ -224,6 +224,8 @@ static void __init_cache_modes(u64 pat)
 	char pat_msg[33];
 	int i;
 
+	WARN_ON_ONCE(pat_cm_initialized);
+
 	pat_msg[32] = 0;
 	for (i = 7; i >= 0; i--) {
 		cache = pat_get_cache_mode((pat >> (i * 8)) & 7,
@@ -232,12 +234,12 @@ static void __init_cache_modes(u64 pat)
 	}
 	pr_info("x86/PAT: Configuration [0-7]: %s\n", pat_msg);
 
-	init_cm_done = true;
+	pat_cm_initialized = true;
 }
 
 #define PAT(x, y)	((u64)PAT_ ## y << ((x)*8))
 
-static void pat_bsp_init(u64 pat)
+static void pat_bp_init(u64 pat)
 {
 	u64 tmp_pat;
 
@@ -253,7 +255,7 @@ static void pat_bsp_init(u64 pat)
 	}
 
 	wrmsrl(MSR_IA32_CR_PAT, pat);
-	pat_initialized = true;
+	pat_bp_enabled = true;
 
 	__init_cache_modes(pat);
 }
@@ -275,7 +277,7 @@ void init_cache_modes(void)
 {
 	u64 pat = 0;
 
-	if (init_cm_done)
+	if (pat_cm_initialized)
 		return;
 
 	if (boot_cpu_has(X86_FEATURE_PAT)) {
@@ -395,9 +397,9 @@ void pat_init(void)
 		      PAT(4, WB) | PAT(5, WP) | PAT(6, UC_MINUS) | PAT(7, WT);
 	}
 
-	if (!boot_cpu_done) {
-		pat_bsp_init(pat);
-		boot_cpu_done = true;
+	if (!pat_bp_initialized) {
+		pat_bp_init(pat);
+		pat_bp_initialized = true;
 	} else {
 		pat_ap_init(pat);
 	}

From f9b57cf80c8b585614ba223732be0d8f19d558d8 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Tue, 10 Dec 2019 10:08:09 +0100
Subject: [PATCH 09/17] x86/mm/pat: Move the memtype related files to
 arch/x86/mm/pat/

- pat.c offers, dominantly, the memtype APIs - so rename it to memtype.c.

- pageattr.c is offering, primarily, the set_memory*() page attribute APIs,
  which is offered via the <asm/set_memory.h> header: name the .c file
  along the same pattern.

I.e. perform these renames, and move them all next to each other in arch/x86/mm/pat/:

    pat.c             => memtype.c
    pat_internal.h    => memtype.h
    pat_interval.c    => memtype_interval.c

    pageattr.c        => set_memory.c
    pageattr-test.c   => cpa-test.c

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/mm/Makefile                                   | 8 ++++----
 arch/x86/mm/pat/Makefile                               | 5 +++++
 arch/x86/mm/{pageattr-test.c => pat/cpa-test.c}        | 0
 arch/x86/mm/{pat.c => pat/memtype.c}                   | 4 ++--
 arch/x86/mm/{pat_internal.h => pat/memtype.h}          | 6 +++---
 arch/x86/mm/{pat_interval.c => pat/memtype_interval.c} | 2 +-
 arch/x86/mm/{pageattr.c => pat/set_memory.c}           | 4 ++--
 7 files changed, 17 insertions(+), 12 deletions(-)
 create mode 100644 arch/x86/mm/pat/Makefile
 rename arch/x86/mm/{pageattr-test.c => pat/cpa-test.c} (100%)
 rename arch/x86/mm/{pat.c => pat/memtype.c} (99%)
 rename arch/x86/mm/{pat_internal.h => pat/memtype.h} (94%)
 rename arch/x86/mm/{pat_interval.c => pat/memtype_interval.c} (99%)
 rename arch/x86/mm/{pageattr.c => pat/set_memory.c} (99%)

diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 3b89c201ac26..345848f270e3 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -12,8 +12,10 @@ CFLAGS_REMOVE_mem_encrypt.o		= -pg
 CFLAGS_REMOVE_mem_encrypt_identity.o	= -pg
 endif
 
-obj-y	:=  init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
-	    pat.o pgtable.o physaddr.o setup_nx.o tlb.o cpu_entry_area.o maccess.o
+obj-y				:=  init.o init_$(BITS).o fault.o ioremap.o extable.o mmap.o \
+				    pgtable.o physaddr.o setup_nx.o tlb.o cpu_entry_area.o maccess.o
+
+obj-y				+= pat/
 
 # Make sure __phys_addr has no stackprotector
 nostackp := $(call cc-option, -fno-stack-protector)
@@ -23,8 +25,6 @@ CFLAGS_mem_encrypt_identity.o	:= $(nostackp)
 
 CFLAGS_fault.o := -I $(srctree)/$(src)/../include/asm/trace
 
-obj-$(CONFIG_X86_PAT)		+= pat_interval.o
-
 obj-$(CONFIG_X86_32)		+= pgtable_32.o iomap_32.o
 
 obj-$(CONFIG_HUGETLB_PAGE)	+= hugetlbpage.o
diff --git a/arch/x86/mm/pat/Makefile b/arch/x86/mm/pat/Makefile
new file mode 100644
index 000000000000..ea464c995161
--- /dev/null
+++ b/arch/x86/mm/pat/Makefile
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0
+
+obj-y				:= set_memory.o memtype.o
+
+obj-$(CONFIG_X86_PAT)		+= memtype_interval.o
diff --git a/arch/x86/mm/pageattr-test.c b/arch/x86/mm/pat/cpa-test.c
similarity index 100%
rename from arch/x86/mm/pageattr-test.c
rename to arch/x86/mm/pat/cpa-test.c
diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat/memtype.c
similarity index 99%
rename from arch/x86/mm/pat.c
rename to arch/x86/mm/pat/memtype.c
index f1677fad1735..76532f080795 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat/memtype.c
@@ -55,8 +55,8 @@
 #include <asm/pat.h>
 #include <asm/io.h>
 
-#include "pat_internal.h"
-#include "mm_internal.h"
+#include "memtype.h"
+#include "../mm_internal.h"
 
 #undef pr_fmt
 #define pr_fmt(fmt) "" fmt
diff --git a/arch/x86/mm/pat_internal.h b/arch/x86/mm/pat/memtype.h
similarity index 94%
rename from arch/x86/mm/pat_internal.h
rename to arch/x86/mm/pat/memtype.h
index 23ce8cdac159..cacecdbceb55 100644
--- a/arch/x86/mm/pat_internal.h
+++ b/arch/x86/mm/pat/memtype.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0 */
-#ifndef __PAT_INTERNAL_H_
-#define __PAT_INTERNAL_H_
+#ifndef __MEMTYPE_H_
+#define __MEMTYPE_H_
 
 extern int pat_debug_enable;
 
@@ -46,4 +46,4 @@ static inline int memtype_copy_nth_element(struct memtype *out, loff_t pos)
 { return 0; }
 #endif
 
-#endif /* __PAT_INTERNAL_H_ */
+#endif /* __MEMTYPE_H_ */
diff --git a/arch/x86/mm/pat_interval.c b/arch/x86/mm/pat/memtype_interval.c
similarity index 99%
rename from arch/x86/mm/pat_interval.c
rename to arch/x86/mm/pat/memtype_interval.c
index 3c983de5f492..a7fbbdd88eeb 100644
--- a/arch/x86/mm/pat_interval.c
+++ b/arch/x86/mm/pat/memtype_interval.c
@@ -18,7 +18,7 @@
 #include <asm/pgtable.h>
 #include <asm/pat.h>
 
-#include "pat_internal.h"
+#include "memtype.h"
 
 /*
  * The memtype tree keeps track of memory type for specific
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pat/set_memory.c
similarity index 99%
rename from arch/x86/mm/pageattr.c
rename to arch/x86/mm/pat/set_memory.c
index 1b99ad05b117..8fbefee6fc6d 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -27,7 +27,7 @@
 #include <asm/pat.h>
 #include <asm/set_memory.h>
 
-#include "mm_internal.h"
+#include "../mm_internal.h"
 
 /*
  * The current flushing context - we pass it instead of 5 arguments:
@@ -2281,5 +2281,5 @@ int __init kernel_unmap_pages_in_pgd(pgd_t *pgd, unsigned long address,
  * be exposed to the rest of the kernel. Include these directly here.
  */
 #ifdef CONFIG_CPA_DEBUG
-#include "pageattr-test.c"
+#include "cpa-test.c"
 #endif

From ecdd6ee77b73d11fcf2ca6739e4d1fe590446599 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Wed, 20 Nov 2019 15:30:44 +0100
Subject: [PATCH 10/17] x86/mm/pat: Standardize on memtype_*() prefix for APIs

Half of our memtype APIs are memtype_ prefixed, the other half are _memtype suffixed:

	reserve_memtype()
	free_memtype()
	kernel_map_sync_memtype()
	io_reserve_memtype()
	io_free_memtype()

	memtype_check_insert()
	memtype_erase()
	memtype_lookup()
	memtype_copy_nth_element()

Use prefixes consistently, like most other modern kernel APIs:

	reserve_memtype()		=> memtype_reserve()
	free_memtype()			=> memtype_free()
	kernel_map_sync_memtype()	=> memtype_kernel_map_sync()
	io_reserve_memtype()		=> memtype_reserve_io()
	io_free_memtype()		=> memtype_free_io()

	memtype_check_insert()		=> memtype_check_insert()
	memtype_erase()			=> memtype_erase()
	memtype_lookup()		=> memtype_lookup()
	memtype_copy_nth_element()	=> memtype_copy_nth_element()

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/pat.h   | 10 ++++----
 arch/x86/mm/iomap_32.c       |  4 ++--
 arch/x86/mm/ioremap.c        | 10 ++++----
 arch/x86/mm/pat/memtype.c    | 44 ++++++++++++++++++------------------
 arch/x86/mm/pat/set_memory.c | 16 ++++++-------
 5 files changed, 42 insertions(+), 42 deletions(-)

diff --git a/arch/x86/include/asm/pat.h b/arch/x86/include/asm/pat.h
index 92015c65fa2a..4a9a97d930e7 100644
--- a/arch/x86/include/asm/pat.h
+++ b/arch/x86/include/asm/pat.h
@@ -10,17 +10,17 @@ void pat_disable(const char *reason);
 extern void pat_init(void);
 extern void init_cache_modes(void);
 
-extern int reserve_memtype(u64 start, u64 end,
+extern int memtype_reserve(u64 start, u64 end,
 		enum page_cache_mode req_pcm, enum page_cache_mode *ret_pcm);
-extern int free_memtype(u64 start, u64 end);
+extern int memtype_free(u64 start, u64 end);
 
-extern int kernel_map_sync_memtype(u64 base, unsigned long size,
+extern int memtype_kernel_map_sync(u64 base, unsigned long size,
 		enum page_cache_mode pcm);
 
-int io_reserve_memtype(resource_size_t start, resource_size_t end,
+int memtype_reserve_io(resource_size_t start, resource_size_t end,
 			enum page_cache_mode *pcm);
 
-void io_free_memtype(resource_size_t start, resource_size_t end);
+void memtype_free_io(resource_size_t start, resource_size_t end);
 
 bool pat_pfn_immune_to_uc_mtrr(unsigned long pfn);
 
diff --git a/arch/x86/mm/iomap_32.c b/arch/x86/mm/iomap_32.c
index 6748b4c2baff..4a0762ebe051 100644
--- a/arch/x86/mm/iomap_32.c
+++ b/arch/x86/mm/iomap_32.c
@@ -26,7 +26,7 @@ int iomap_create_wc(resource_size_t base, unsigned long size, pgprot_t *prot)
 	if (!is_io_mapping_possible(base, size))
 		return -EINVAL;
 
-	ret = io_reserve_memtype(base, base + size, &pcm);
+	ret = memtype_reserve_io(base, base + size, &pcm);
 	if (ret)
 		return ret;
 
@@ -40,7 +40,7 @@ EXPORT_SYMBOL_GPL(iomap_create_wc);
 
 void iomap_free(resource_size_t base, unsigned long size)
 {
-	io_free_memtype(base, base + size);
+	memtype_free_io(base, base + size);
 }
 EXPORT_SYMBOL_GPL(iomap_free);
 
diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index b3a2936377b5..e49de6cbc64e 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -196,10 +196,10 @@ __ioremap_caller(resource_size_t phys_addr, unsigned long size,
 	phys_addr &= PHYSICAL_PAGE_MASK;
 	size = PAGE_ALIGN(last_addr+1) - phys_addr;
 
-	retval = reserve_memtype(phys_addr, (u64)phys_addr + size,
+	retval = memtype_reserve(phys_addr, (u64)phys_addr + size,
 						pcm, &new_pcm);
 	if (retval) {
-		printk(KERN_ERR "ioremap reserve_memtype failed %d\n", retval);
+		printk(KERN_ERR "ioremap memtype_reserve failed %d\n", retval);
 		return NULL;
 	}
 
@@ -255,7 +255,7 @@ __ioremap_caller(resource_size_t phys_addr, unsigned long size,
 	area->phys_addr = phys_addr;
 	vaddr = (unsigned long) area->addr;
 
-	if (kernel_map_sync_memtype(phys_addr, size, pcm))
+	if (memtype_kernel_map_sync(phys_addr, size, pcm))
 		goto err_free_area;
 
 	if (ioremap_page_range(vaddr, vaddr + size, phys_addr, prot))
@@ -275,7 +275,7 @@ __ioremap_caller(resource_size_t phys_addr, unsigned long size,
 err_free_area:
 	free_vm_area(area);
 err_free_memtype:
-	free_memtype(phys_addr, phys_addr + size);
+	memtype_free(phys_addr, phys_addr + size);
 	return NULL;
 }
 
@@ -451,7 +451,7 @@ void iounmap(volatile void __iomem *addr)
 		return;
 	}
 
-	free_memtype(p->phys_addr, p->phys_addr + get_vm_area_size(p));
+	memtype_free(p->phys_addr, p->phys_addr + get_vm_area_size(p));
 
 	/* Finally remove it */
 	o = remove_vm_area((void __force *)addr);
diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
index 76532f080795..7ed37354e521 100644
--- a/arch/x86/mm/pat/memtype.c
+++ b/arch/x86/mm/pat/memtype.c
@@ -575,7 +575,7 @@ static u64 sanitize_phys(u64 address)
  * available type in new_type in case of no error. In case of any error
  * it will return a negative return value.
  */
-int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
+int memtype_reserve(u64 start, u64 end, enum page_cache_mode req_type,
 		    enum page_cache_mode *new_type)
 {
 	struct memtype *entry_new;
@@ -638,7 +638,7 @@ int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 
 	err = memtype_check_insert(entry_new, new_type);
 	if (err) {
-		pr_info("x86/PAT: reserve_memtype failed [mem %#010Lx-%#010Lx], track %s, req %s\n",
+		pr_info("x86/PAT: memtype_reserve failed [mem %#010Lx-%#010Lx], track %s, req %s\n",
 			start, end - 1,
 			cattr_name(entry_new->type), cattr_name(req_type));
 		kfree(entry_new);
@@ -649,14 +649,14 @@ int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 
 	spin_unlock(&memtype_lock);
 
-	dprintk("reserve_memtype added [mem %#010Lx-%#010Lx], track %s, req %s, ret %s\n",
+	dprintk("memtype_reserve added [mem %#010Lx-%#010Lx], track %s, req %s, ret %s\n",
 		start, end - 1, cattr_name(entry_new->type), cattr_name(req_type),
 		new_type ? cattr_name(*new_type) : "-");
 
 	return err;
 }
 
-int free_memtype(u64 start, u64 end)
+int memtype_free(u64 start, u64 end)
 {
 	int is_range_ram;
 	struct memtype *entry_old;
@@ -689,7 +689,7 @@ int free_memtype(u64 start, u64 end)
 
 	kfree(entry_old);
 
-	dprintk("free_memtype request [mem %#010Lx-%#010Lx]\n", start, end - 1);
+	dprintk("memtype_free request [mem %#010Lx-%#010Lx]\n", start, end - 1);
 
 	return 0;
 }
@@ -752,7 +752,7 @@ bool pat_pfn_immune_to_uc_mtrr(unsigned long pfn)
 EXPORT_SYMBOL_GPL(pat_pfn_immune_to_uc_mtrr);
 
 /**
- * io_reserve_memtype - Request a memory type mapping for a region of memory
+ * memtype_reserve_io - Request a memory type mapping for a region of memory
  * @start: start (physical address) of the region
  * @end: end (physical address) of the region
  * @type: A pointer to memtype, with requested type. On success, requested
@@ -761,7 +761,7 @@ EXPORT_SYMBOL_GPL(pat_pfn_immune_to_uc_mtrr);
  * On success, returns 0
  * On failure, returns non-zero
  */
-int io_reserve_memtype(resource_size_t start, resource_size_t end,
+int memtype_reserve_io(resource_size_t start, resource_size_t end,
 			enum page_cache_mode *type)
 {
 	resource_size_t size = end - start;
@@ -771,47 +771,47 @@ int io_reserve_memtype(resource_size_t start, resource_size_t end,
 
 	WARN_ON_ONCE(iomem_map_sanity_check(start, size));
 
-	ret = reserve_memtype(start, end, req_type, &new_type);
+	ret = memtype_reserve(start, end, req_type, &new_type);
 	if (ret)
 		goto out_err;
 
 	if (!is_new_memtype_allowed(start, size, req_type, new_type))
 		goto out_free;
 
-	if (kernel_map_sync_memtype(start, size, new_type) < 0)
+	if (memtype_kernel_map_sync(start, size, new_type) < 0)
 		goto out_free;
 
 	*type = new_type;
 	return 0;
 
 out_free:
-	free_memtype(start, end);
+	memtype_free(start, end);
 	ret = -EBUSY;
 out_err:
 	return ret;
 }
 
 /**
- * io_free_memtype - Release a memory type mapping for a region of memory
+ * memtype_free_io - Release a memory type mapping for a region of memory
  * @start: start (physical address) of the region
  * @end: end (physical address) of the region
  */
-void io_free_memtype(resource_size_t start, resource_size_t end)
+void memtype_free_io(resource_size_t start, resource_size_t end)
 {
-	free_memtype(start, end);
+	memtype_free(start, end);
 }
 
 int arch_io_reserve_memtype_wc(resource_size_t start, resource_size_t size)
 {
 	enum page_cache_mode type = _PAGE_CACHE_MODE_WC;
 
-	return io_reserve_memtype(start, start + size, &type);
+	return memtype_reserve_io(start, start + size, &type);
 }
 EXPORT_SYMBOL(arch_io_reserve_memtype_wc);
 
 void arch_io_free_memtype_wc(resource_size_t start, resource_size_t size)
 {
-	io_free_memtype(start, start + size);
+	memtype_free_io(start, start + size);
 }
 EXPORT_SYMBOL(arch_io_free_memtype_wc);
 
@@ -871,7 +871,7 @@ int phys_mem_access_prot_allowed(struct file *file, unsigned long pfn,
  * Change the memory type for the physical address range in kernel identity
  * mapping space if that range is a part of identity map.
  */
-int kernel_map_sync_memtype(u64 base, unsigned long size,
+int memtype_kernel_map_sync(u64 base, unsigned long size,
 			    enum page_cache_mode pcm)
 {
 	unsigned long id_sz;
@@ -901,7 +901,7 @@ int kernel_map_sync_memtype(u64 base, unsigned long size,
 
 /*
  * Internal interface to reserve a range of physical memory with prot.
- * Reserved non RAM regions only and after successful reserve_memtype,
+ * Reserved non RAM regions only and after successful memtype_reserve,
  * this func also keeps identity mapping (if any) in sync with this new prot.
  */
 static int reserve_pfn_range(u64 paddr, unsigned long size, pgprot_t *vma_prot,
@@ -938,14 +938,14 @@ static int reserve_pfn_range(u64 paddr, unsigned long size, pgprot_t *vma_prot,
 		return 0;
 	}
 
-	ret = reserve_memtype(paddr, paddr + size, want_pcm, &pcm);
+	ret = memtype_reserve(paddr, paddr + size, want_pcm, &pcm);
 	if (ret)
 		return ret;
 
 	if (pcm != want_pcm) {
 		if (strict_prot ||
 		    !is_new_memtype_allowed(paddr, size, want_pcm, pcm)) {
-			free_memtype(paddr, paddr + size);
+			memtype_free(paddr, paddr + size);
 			pr_err("x86/PAT: %s:%d map pfn expected mapping type %s for [mem %#010Lx-%#010Lx], got %s\n",
 			       current->comm, current->pid,
 			       cattr_name(want_pcm),
@@ -963,8 +963,8 @@ static int reserve_pfn_range(u64 paddr, unsigned long size, pgprot_t *vma_prot,
 				     cachemode2protval(pcm));
 	}
 
-	if (kernel_map_sync_memtype(paddr, size, pcm) < 0) {
-		free_memtype(paddr, paddr + size);
+	if (memtype_kernel_map_sync(paddr, size, pcm) < 0) {
+		memtype_free(paddr, paddr + size);
 		return -EINVAL;
 	}
 	return 0;
@@ -980,7 +980,7 @@ static void free_pfn_range(u64 paddr, unsigned long size)
 
 	is_ram = pat_pagerange_is_ram(paddr, paddr + size);
 	if (is_ram == 0)
-		free_memtype(paddr, paddr + size);
+		memtype_free(paddr, paddr + size);
 }
 
 /*
diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index 8fbefee6fc6d..3e5e98b78bf3 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -1801,7 +1801,7 @@ int set_memory_uc(unsigned long addr, int numpages)
 	/*
 	 * for now UC MINUS. see comments in ioremap()
 	 */
-	ret = reserve_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE,
+	ret = memtype_reserve(__pa(addr), __pa(addr) + numpages * PAGE_SIZE,
 			      _PAGE_CACHE_MODE_UC_MINUS, NULL);
 	if (ret)
 		goto out_err;
@@ -1813,7 +1813,7 @@ int set_memory_uc(unsigned long addr, int numpages)
 	return 0;
 
 out_free:
-	free_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE);
+	memtype_free(__pa(addr), __pa(addr) + numpages * PAGE_SIZE);
 out_err:
 	return ret;
 }
@@ -1839,14 +1839,14 @@ int set_memory_wc(unsigned long addr, int numpages)
 {
 	int ret;
 
-	ret = reserve_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE,
+	ret = memtype_reserve(__pa(addr), __pa(addr) + numpages * PAGE_SIZE,
 		_PAGE_CACHE_MODE_WC, NULL);
 	if (ret)
 		return ret;
 
 	ret = _set_memory_wc(addr, numpages);
 	if (ret)
-		free_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE);
+		memtype_free(__pa(addr), __pa(addr) + numpages * PAGE_SIZE);
 
 	return ret;
 }
@@ -1873,7 +1873,7 @@ int set_memory_wb(unsigned long addr, int numpages)
 	if (ret)
 		return ret;
 
-	free_memtype(__pa(addr), __pa(addr) + numpages * PAGE_SIZE);
+	memtype_free(__pa(addr), __pa(addr) + numpages * PAGE_SIZE);
 	return 0;
 }
 EXPORT_SYMBOL(set_memory_wb);
@@ -2014,7 +2014,7 @@ static int _set_pages_array(struct page **pages, int numpages,
 			continue;
 		start = page_to_pfn(pages[i]) << PAGE_SHIFT;
 		end = start + PAGE_SIZE;
-		if (reserve_memtype(start, end, new_type, NULL))
+		if (memtype_reserve(start, end, new_type, NULL))
 			goto err_out;
 	}
 
@@ -2040,7 +2040,7 @@ err_out:
 			continue;
 		start = page_to_pfn(pages[i]) << PAGE_SHIFT;
 		end = start + PAGE_SIZE;
-		free_memtype(start, end);
+		memtype_free(start, end);
 	}
 	return -EINVAL;
 }
@@ -2089,7 +2089,7 @@ int set_pages_array_wb(struct page **pages, int numpages)
 			continue;
 		start = page_to_pfn(pages[i]) << PAGE_SHIFT;
 		end = start + PAGE_SIZE;
-		free_memtype(start, end);
+		memtype_free(start, end);
 	}
 
 	return 0;

From eb243d1d28663c9b92010973a6a3ffa947f682ba Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Wed, 20 Nov 2019 15:33:57 +0100
Subject: [PATCH 11/17] x86/mm/pat: Rename <asm/pat.h> => <asm/memtype.h>

pat.h is a file whose main purpose is to provide the memtype_*() APIs.

PAT is the low level hardware mechanism - but the high level abstraction
is memtype.

So name the header <memtype.h> as well - this goes hand in hand with memtype.c
and memtype_interval.c.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/{pat.h => memtype.h} | 6 +++---
 arch/x86/include/asm/mtrr.h               | 2 +-
 arch/x86/include/asm/pci.h                | 2 +-
 arch/x86/kernel/cpu/common.c              | 2 +-
 arch/x86/kernel/cpu/mtrr/generic.c        | 2 +-
 arch/x86/kernel/cpu/mtrr/mtrr.c           | 2 +-
 arch/x86/kernel/cpu/scattered.c           | 2 +-
 arch/x86/kernel/cpu/topology.c            | 2 +-
 arch/x86/kernel/x86_init.c                | 2 +-
 arch/x86/kvm/mmu/mmu.c                    | 2 +-
 arch/x86/mm/iomap_32.c                    | 2 +-
 arch/x86/mm/ioremap.c                     | 2 +-
 arch/x86/mm/pat/memtype.c                 | 2 +-
 arch/x86/mm/pat/memtype_interval.c        | 2 +-
 arch/x86/mm/pat/set_memory.c              | 2 +-
 arch/x86/pci/i386.c                       | 2 +-
 arch/x86/xen/mmu_pv.c                     | 2 +-
 drivers/infiniband/hw/mlx5/main.c         | 2 +-
 drivers/media/pci/ivtv/ivtvfb.c           | 2 +-
 19 files changed, 21 insertions(+), 21 deletions(-)
 rename arch/x86/include/asm/{pat.h => memtype.h} (88%)

diff --git a/arch/x86/include/asm/pat.h b/arch/x86/include/asm/memtype.h
similarity index 88%
rename from arch/x86/include/asm/pat.h
rename to arch/x86/include/asm/memtype.h
index 4a9a97d930e7..ec18e38ae391 100644
--- a/arch/x86/include/asm/pat.h
+++ b/arch/x86/include/asm/memtype.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0 */
-#ifndef _ASM_X86_PAT_H
-#define _ASM_X86_PAT_H
+#ifndef _ASM_X86_MEMTYPE_H
+#define _ASM_X86_MEMTYPE_H
 
 #include <linux/types.h>
 #include <asm/pgtable_types.h>
@@ -24,4 +24,4 @@ void memtype_free_io(resource_size_t start, resource_size_t end);
 
 bool pat_pfn_immune_to_uc_mtrr(unsigned long pfn);
 
-#endif /* _ASM_X86_PAT_H */
+#endif /* _ASM_X86_MEMTYPE_H */
diff --git a/arch/x86/include/asm/mtrr.h b/arch/x86/include/asm/mtrr.h
index 3337d2233aef..829df26fd7a3 100644
--- a/arch/x86/include/asm/mtrr.h
+++ b/arch/x86/include/asm/mtrr.h
@@ -24,7 +24,7 @@
 #define _ASM_X86_MTRR_H
 
 #include <uapi/asm/mtrr.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 
 
 /*
diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h
index 90d0731fdcb6..c1fdd43fe187 100644
--- a/arch/x86/include/asm/pci.h
+++ b/arch/x86/include/asm/pci.h
@@ -9,7 +9,7 @@
 #include <linux/scatterlist.h>
 #include <linux/numa.h>
 #include <asm/io.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/x86_init.h>
 
 struct pci_sysdata {
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 2e4d90294fe6..9d6a35a4586e 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -49,7 +49,7 @@
 #include <asm/cpu.h>
 #include <asm/mce.h>
 #include <asm/msr.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/microcode.h>
 #include <asm/microcode_intel.h>
 #include <asm/intel-family.h>
diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c
index aa5c064a6a22..51b9190c628b 100644
--- a/arch/x86/kernel/cpu/mtrr/generic.c
+++ b/arch/x86/kernel/cpu/mtrr/generic.c
@@ -15,7 +15,7 @@
 #include <asm/tlbflush.h>
 #include <asm/mtrr.h>
 #include <asm/msr.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 
 #include "mtrr.h"
 
diff --git a/arch/x86/kernel/cpu/mtrr/mtrr.c b/arch/x86/kernel/cpu/mtrr/mtrr.c
index 507039c20128..6a80f36b5d59 100644
--- a/arch/x86/kernel/cpu/mtrr/mtrr.c
+++ b/arch/x86/kernel/cpu/mtrr/mtrr.c
@@ -52,7 +52,7 @@
 #include <asm/e820/api.h>
 #include <asm/mtrr.h>
 #include <asm/msr.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 
 #include "mtrr.h"
 
diff --git a/arch/x86/kernel/cpu/scattered.c b/arch/x86/kernel/cpu/scattered.c
index adf9b71386ef..62b137c3c97a 100644
--- a/arch/x86/kernel/cpu/scattered.c
+++ b/arch/x86/kernel/cpu/scattered.c
@@ -4,7 +4,7 @@
  */
 #include <linux/cpu.h>
 
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/apic.h>
 #include <asm/processor.h>
 
diff --git a/arch/x86/kernel/cpu/topology.c b/arch/x86/kernel/cpu/topology.c
index ee48c3fc8a65..d3a0791bc052 100644
--- a/arch/x86/kernel/cpu/topology.c
+++ b/arch/x86/kernel/cpu/topology.c
@@ -7,7 +7,7 @@
 
 #include <linux/cpu.h>
 #include <asm/apic.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/processor.h>
 
 #include "cpu.h"
diff --git a/arch/x86/kernel/x86_init.c b/arch/x86/kernel/x86_init.c
index ce89430a7f80..23e25f3034c2 100644
--- a/arch/x86/kernel/x86_init.c
+++ b/arch/x86/kernel/x86_init.c
@@ -20,7 +20,7 @@
 #include <asm/irq.h>
 #include <asm/io_apic.h>
 #include <asm/hpet.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/tsc.h>
 #include <asm/iommu.h>
 #include <asm/mach_traps.h>
diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 6f92b40d798c..a32b847a8089 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -40,7 +40,7 @@
 #include <linux/kthread.h>
 
 #include <asm/page.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/cmpxchg.h>
 #include <asm/e820/api.h>
 #include <asm/io.h>
diff --git a/arch/x86/mm/iomap_32.c b/arch/x86/mm/iomap_32.c
index 4a0762ebe051..f60398aeb644 100644
--- a/arch/x86/mm/iomap_32.c
+++ b/arch/x86/mm/iomap_32.c
@@ -4,7 +4,7 @@
  */
 
 #include <asm/iomap.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <linux/export.h>
 #include <linux/highmem.h>
 
diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index e49de6cbc64e..44e4beb4239f 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -24,7 +24,7 @@
 #include <asm/pgtable.h>
 #include <asm/tlbflush.h>
 #include <asm/pgalloc.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/setup.h>
 
 #include "physaddr.h"
diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
index 7ed37354e521..394be8611748 100644
--- a/arch/x86/mm/pat/memtype.c
+++ b/arch/x86/mm/pat/memtype.c
@@ -52,7 +52,7 @@
 #include <asm/mtrr.h>
 #include <asm/page.h>
 #include <asm/msr.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/io.h>
 
 #include "memtype.h"
diff --git a/arch/x86/mm/pat/memtype_interval.c b/arch/x86/mm/pat/memtype_interval.c
index a7fbbdd88eeb..a07e4882bf36 100644
--- a/arch/x86/mm/pat/memtype_interval.c
+++ b/arch/x86/mm/pat/memtype_interval.c
@@ -16,7 +16,7 @@
 #include <linux/gfp.h>
 
 #include <asm/pgtable.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 
 #include "memtype.h"
 
diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index 3e5e98b78bf3..d4ab493b1647 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -24,7 +24,7 @@
 #include <linux/uaccess.h>
 #include <asm/pgalloc.h>
 #include <asm/proto.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/set_memory.h>
 
 #include "../mm_internal.h"
diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c
index 9df652d3d927..fa855bbaebaf 100644
--- a/arch/x86/pci/i386.c
+++ b/arch/x86/pci/i386.c
@@ -34,7 +34,7 @@
 #include <linux/errno.h>
 #include <linux/memblock.h>
 
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/e820/api.h>
 #include <asm/pci_x86.h>
 #include <asm/io_apic.h>
diff --git a/arch/x86/xen/mmu_pv.c b/arch/x86/xen/mmu_pv.c
index c8dbee62ec2a..bbba8b17829a 100644
--- a/arch/x86/xen/mmu_pv.c
+++ b/arch/x86/xen/mmu_pv.c
@@ -67,7 +67,7 @@
 #include <asm/linkage.h>
 #include <asm/page.h>
 #include <asm/init.h>
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #include <asm/smp.h>
 #include <asm/tlb.h>
 
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 51100350b688..c0c2c56b7a80 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -40,7 +40,7 @@
 #include <linux/slab.h>
 #include <linux/bitmap.h>
 #if defined(CONFIG_X86)
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #endif
 #include <linux/sched.h>
 #include <linux/sched/mm.h>
diff --git a/drivers/media/pci/ivtv/ivtvfb.c b/drivers/media/pci/ivtv/ivtvfb.c
index 95a56cce9b65..1daf9e07cad7 100644
--- a/drivers/media/pci/ivtv/ivtvfb.c
+++ b/drivers/media/pci/ivtv/ivtvfb.c
@@ -37,7 +37,7 @@
 #include <linux/ivtvfb.h>
 
 #ifdef CONFIG_X86_64
-#include <asm/pat.h>
+#include <asm/memtype.h>
 #endif
 
 /* card parameters */

From 533d49b37a2b532354d3841a142173b8321818df Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Wed, 20 Nov 2019 15:45:32 +0100
Subject: [PATCH 12/17] x86/mm/pat: Clean up <asm/memtype.h> externs

Half of the declarations have an 'extern', half of them not,
use 'extern' consistently.

This makes grepping for APIs easier, such as:

  dagon:~/tip> git grep -E '\<memtype_.*\(' arch/x86/ | grep extern
  arch/x86/include/asm/memtype.h:extern int memtype_reserve(u64 start, u64 end,
  arch/x86/include/asm/memtype.h:extern int memtype_free(u64 start, u64 end);
  arch/x86/include/asm/memtype.h:extern int memtype_kernel_map_sync(u64 base, unsigned long size,
  arch/x86/include/asm/memtype.h:extern int memtype_reserve_io(resource_size_t start, resource_size_t end,
  arch/x86/include/asm/memtype.h:extern void memtype_free_io(resource_size_t start, resource_size_t end);
  arch/x86/mm/pat/memtype.h:extern int memtype_check_insert(struct memtype *entry_new,
  arch/x86/mm/pat/memtype.h:extern struct memtype *memtype_erase(u64 start, u64 end);
  arch/x86/mm/pat/memtype.h:extern struct memtype *memtype_lookup(u64 addr);
  arch/x86/mm/pat/memtype.h:extern int memtype_copy_nth_element(struct memtype *entry_out, loff_t pos);

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/memtype.h | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/arch/x86/include/asm/memtype.h b/arch/x86/include/asm/memtype.h
index ec18e38ae391..9c2447b3555d 100644
--- a/arch/x86/include/asm/memtype.h
+++ b/arch/x86/include/asm/memtype.h
@@ -5,8 +5,8 @@
 #include <linux/types.h>
 #include <asm/pgtable_types.h>
 
-bool pat_enabled(void);
-void pat_disable(const char *reason);
+extern bool pat_enabled(void);
+extern void pat_disable(const char *reason);
 extern void pat_init(void);
 extern void init_cache_modes(void);
 
@@ -17,11 +17,11 @@ extern int memtype_free(u64 start, u64 end);
 extern int memtype_kernel_map_sync(u64 base, unsigned long size,
 		enum page_cache_mode pcm);
 
-int memtype_reserve_io(resource_size_t start, resource_size_t end,
+extern int memtype_reserve_io(resource_size_t start, resource_size_t end,
 			enum page_cache_mode *pcm);
 
-void memtype_free_io(resource_size_t start, resource_size_t end);
+extern void memtype_free_io(resource_size_t start, resource_size_t end);
 
-bool pat_pfn_immune_to_uc_mtrr(unsigned long pfn);
+extern bool pat_pfn_immune_to_uc_mtrr(unsigned long pfn);
 
 #endif /* _ASM_X86_MEMTYPE_H */

From b75baaf3a81e71680f3d50965c9330b36993fbad Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Wed, 20 Nov 2019 15:57:04 +0100
Subject: [PATCH 13/17] x86/mm/pat: Fix typo in the Kconfig help text

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 5e8949953660..b6778575cf6f 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -1512,7 +1512,7 @@ config X86_CPA_STATISTICS
 	bool "Enable statistic for Change Page Attribute"
 	depends on DEBUG_FS
 	---help---
-	  Expose statistics about the Change Page Attribute mechanims, which
+	  Expose statistics about the Change Page Attribute mechanism, which
 	  helps to determine the effectiveness of preserving large and huge
 	  page mappings when mapping protections are changed.
 

From 4efb56649132687b5093d90aa62887595e65a09d Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Wed, 20 Nov 2019 17:38:39 +0100
Subject: [PATCH 14/17] x86/mm: Tabulate the page table encoding definitions

I got lost in trying to figure out which bits were enabled
in one of the PTE masks, so let's make it pretty
obvious at the definition site already:

 #define PAGE_NONE            __pg(   0|   0|   0|___A|   0|   0|   0|___G)
 #define PAGE_SHARED          __pg(__PP|__RW|_USR|___A|__NX|   0|   0|   0)
 #define PAGE_SHARED_EXEC     __pg(__PP|__RW|_USR|___A|   0|   0|   0|   0)
 #define PAGE_COPY_NOEXEC     __pg(__PP|   0|_USR|___A|__NX|   0|   0|   0)
 #define PAGE_COPY_EXEC       __pg(__PP|   0|_USR|___A|   0|   0|   0|   0)
 #define PAGE_COPY            __pg(__PP|   0|_USR|___A|__NX|   0|   0|   0)
 #define PAGE_READONLY        __pg(__PP|   0|_USR|___A|__NX|   0|   0|   0)
 #define PAGE_READONLY_EXEC   __pg(__PP|   0|_USR|___A|   0|   0|   0|   0)

 #define __PAGE_KERNEL            (__PP|__RW|   0|___A|__NX|___D|   0|___G)
 #define __PAGE_KERNEL_EXEC       (__PP|__RW|   0|___A|   0|___D|   0|___G)
 #define _KERNPG_TABLE_NOENC      (__PP|__RW|   0|___A|   0|___D|   0|   0)
 #define _KERNPG_TABLE            (__PP|__RW|   0|___A|   0|___D|   0|   0| _ENC)
 #define _PAGE_TABLE_NOENC        (__PP|__RW|_USR|___A|   0|___D|   0|   0)
 #define _PAGE_TABLE              (__PP|__RW|_USR|___A|   0|___D|   0|   0| _ENC)
 #define __PAGE_KERNEL_RO         (__PP|   0|   0|___A|__NX|___D|   0|___G)
 #define __PAGE_KERNEL_RX         (__PP|   0|   0|___A|   0|___D|   0|___G)
 #define __PAGE_KERNEL_NOCACHE    (__PP|__RW|   0|___A|__NX|___D|   0|___G| __NC)
 #define __PAGE_KERNEL_VVAR       (__PP|   0|_USR|___A|__NX|___D|   0|___G)
 #define __PAGE_KERNEL_LARGE      (__PP|__RW|   0|___A|__NX|___D|_PSE|___G)
 #define __PAGE_KERNEL_LARGE_EXEC (__PP|__RW|   0|___A|   0|___D|_PSE|___G)
 #define __PAGE_KERNEL_WP         (__PP|__RW|   0|___A|__NX|___D|   0|___G| __WP)

Especially security relevant bits like 'NX' or coherence related bits like 'G'
are now super easy to read based on a single grep.

We do the underscore gymnastics to not pollute the kernel's symbol namespace,
and the longest line still fits into 80 columns, so this should be readable
for everyone.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/pgtable_types.h | 135 ++++++++++++++-------------
 1 file changed, 70 insertions(+), 65 deletions(-)

diff --git a/arch/x86/include/asm/pgtable_types.h b/arch/x86/include/asm/pgtable_types.h
index b5e49e6bac63..ea7400726d7a 100644
--- a/arch/x86/include/asm/pgtable_types.h
+++ b/arch/x86/include/asm/pgtable_types.h
@@ -110,11 +110,6 @@
 
 #define _PAGE_PROTNONE	(_AT(pteval_t, 1) << _PAGE_BIT_PROTNONE)
 
-#define _PAGE_TABLE_NOENC	(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER |\
-				 _PAGE_ACCESSED | _PAGE_DIRTY)
-#define _KERNPG_TABLE_NOENC	(_PAGE_PRESENT | _PAGE_RW |		\
-				 _PAGE_ACCESSED | _PAGE_DIRTY)
-
 /*
  * Set of bits not changed in pte_modify.  The pte's
  * protection key is treated like _PAGE_RW, for
@@ -136,80 +131,93 @@
  */
 #ifndef __ASSEMBLY__
 enum page_cache_mode {
-	_PAGE_CACHE_MODE_WB = 0,
-	_PAGE_CACHE_MODE_WC = 1,
+	_PAGE_CACHE_MODE_WB       = 0,
+	_PAGE_CACHE_MODE_WC       = 1,
 	_PAGE_CACHE_MODE_UC_MINUS = 2,
-	_PAGE_CACHE_MODE_UC = 3,
-	_PAGE_CACHE_MODE_WT = 4,
-	_PAGE_CACHE_MODE_WP = 5,
-	_PAGE_CACHE_MODE_NUM = 8
+	_PAGE_CACHE_MODE_UC       = 3,
+	_PAGE_CACHE_MODE_WT       = 4,
+	_PAGE_CACHE_MODE_WP       = 5,
+
+	_PAGE_CACHE_MODE_NUM      = 8
 };
 #endif
 
-#define _PAGE_CACHE_MASK	(_PAGE_PAT | _PAGE_PCD | _PAGE_PWT)
+#define _PAGE_ENC		(_AT(pteval_t, sme_me_mask))
+
+#define _PAGE_CACHE_MASK	(_PAGE_PWT | _PAGE_PCD | _PAGE_PAT)
+
 #define _PAGE_NOCACHE		(cachemode2protval(_PAGE_CACHE_MODE_UC))
 #define _PAGE_CACHE_WP		(cachemode2protval(_PAGE_CACHE_MODE_WP))
 
-#define PAGE_NONE	__pgprot(_PAGE_PROTNONE | _PAGE_ACCESSED)
-#define PAGE_SHARED	__pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | \
-				 _PAGE_ACCESSED | _PAGE_NX)
+#define __PP _PAGE_PRESENT
+#define __RW _PAGE_RW
+#define _USR _PAGE_USER
+#define ___A _PAGE_ACCESSED
+#define ___D _PAGE_DIRTY
+#define ___G _PAGE_GLOBAL
+#define __NX _PAGE_NX
 
-#define PAGE_SHARED_EXEC	__pgprot(_PAGE_PRESENT | _PAGE_RW |	\
-					 _PAGE_USER | _PAGE_ACCESSED)
-#define PAGE_COPY_NOEXEC	__pgprot(_PAGE_PRESENT | _PAGE_USER |	\
-					 _PAGE_ACCESSED | _PAGE_NX)
-#define PAGE_COPY_EXEC		__pgprot(_PAGE_PRESENT | _PAGE_USER |	\
-					 _PAGE_ACCESSED)
-#define PAGE_COPY		PAGE_COPY_NOEXEC
-#define PAGE_READONLY		__pgprot(_PAGE_PRESENT | _PAGE_USER |	\
-					 _PAGE_ACCESSED | _PAGE_NX)
-#define PAGE_READONLY_EXEC	__pgprot(_PAGE_PRESENT | _PAGE_USER |	\
-					 _PAGE_ACCESSED)
+#define _ENC _PAGE_ENC
+#define __WP _PAGE_CACHE_WP
+#define __NC _PAGE_NOCACHE
+#define _PSE _PAGE_PSE
 
-#define __PAGE_KERNEL_EXEC						\
-	(_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_GLOBAL)
-#define __PAGE_KERNEL		(__PAGE_KERNEL_EXEC | _PAGE_NX)
+#define pgprot_val(x)		((x).pgprot)
+#define __pgprot(x)		((pgprot_t) { (x) } )
+#define __pg(x)			__pgprot(x)
 
-#define __PAGE_KERNEL_RO		(__PAGE_KERNEL & ~_PAGE_RW)
-#define __PAGE_KERNEL_RX		(__PAGE_KERNEL_EXEC & ~_PAGE_RW)
-#define __PAGE_KERNEL_NOCACHE		(__PAGE_KERNEL | _PAGE_NOCACHE)
-#define __PAGE_KERNEL_VVAR		(__PAGE_KERNEL_RO | _PAGE_USER)
-#define __PAGE_KERNEL_LARGE		(__PAGE_KERNEL | _PAGE_PSE)
-#define __PAGE_KERNEL_LARGE_EXEC	(__PAGE_KERNEL_EXEC | _PAGE_PSE)
-#define __PAGE_KERNEL_WP		(__PAGE_KERNEL | _PAGE_CACHE_WP)
+#define _PAGE_PAT_LARGE		(_AT(pteval_t, 1) << _PAGE_BIT_PAT_LARGE)
+
+#define PAGE_NONE	     __pg(   0|   0|   0|___A|   0|   0|   0|___G)
+#define PAGE_SHARED	     __pg(__PP|__RW|_USR|___A|__NX|   0|   0|   0)
+#define PAGE_SHARED_EXEC     __pg(__PP|__RW|_USR|___A|   0|   0|   0|   0)
+#define PAGE_COPY_NOEXEC     __pg(__PP|   0|_USR|___A|__NX|   0|   0|   0)
+#define PAGE_COPY_EXEC	     __pg(__PP|   0|_USR|___A|   0|   0|   0|   0)
+#define PAGE_COPY	     __pg(__PP|   0|_USR|___A|__NX|   0|   0|   0)
+#define PAGE_READONLY	     __pg(__PP|   0|_USR|___A|__NX|   0|   0|   0)
+#define PAGE_READONLY_EXEC   __pg(__PP|   0|_USR|___A|   0|   0|   0|   0)
+
+#define __PAGE_KERNEL		 (__PP|__RW|   0|___A|__NX|___D|   0|___G)
+#define __PAGE_KERNEL_EXEC	 (__PP|__RW|   0|___A|   0|___D|   0|___G)
+#define _KERNPG_TABLE_NOENC	 (__PP|__RW|   0|___A|   0|___D|   0|   0)
+#define _KERNPG_TABLE		 (__PP|__RW|   0|___A|   0|___D|   0|   0| _ENC)
+#define _PAGE_TABLE_NOENC	 (__PP|__RW|_USR|___A|   0|___D|   0|   0)
+#define _PAGE_TABLE		 (__PP|__RW|_USR|___A|   0|___D|   0|   0| _ENC)
+#define __PAGE_KERNEL_RO	 (__PP|   0|   0|___A|__NX|___D|   0|___G)
+#define __PAGE_KERNEL_RX	 (__PP|   0|   0|___A|   0|___D|   0|___G)
+#define __PAGE_KERNEL_NOCACHE	 (__PP|__RW|   0|___A|__NX|___D|   0|___G| __NC)
+#define __PAGE_KERNEL_VVAR	 (__PP|   0|_USR|___A|__NX|___D|   0|___G)
+#define __PAGE_KERNEL_LARGE	 (__PP|__RW|   0|___A|__NX|___D|_PSE|___G)
+#define __PAGE_KERNEL_LARGE_EXEC (__PP|__RW|   0|___A|   0|___D|_PSE|___G)
+#define __PAGE_KERNEL_WP	 (__PP|__RW|   0|___A|__NX|___D|   0|___G| __WP)
+
+
+#define __PAGE_KERNEL_IO		__PAGE_KERNEL
+#define __PAGE_KERNEL_IO_NOCACHE	__PAGE_KERNEL_NOCACHE
 
-#define __PAGE_KERNEL_IO		(__PAGE_KERNEL)
-#define __PAGE_KERNEL_IO_NOCACHE	(__PAGE_KERNEL_NOCACHE)
 
 #ifndef __ASSEMBLY__
 
-#define _PAGE_ENC	(_AT(pteval_t, sme_me_mask))
+#define __PAGE_KERNEL_ENC	(__PAGE_KERNEL    | _ENC)
+#define __PAGE_KERNEL_ENC_WP	(__PAGE_KERNEL_WP | _ENC)
+#define __PAGE_KERNEL_NOENC	(__PAGE_KERNEL    |    0)
+#define __PAGE_KERNEL_NOENC_WP	(__PAGE_KERNEL_WP |    0)
 
-#define _KERNPG_TABLE	(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED |	\
-			 _PAGE_DIRTY | _PAGE_ENC)
-#define _PAGE_TABLE	(_KERNPG_TABLE | _PAGE_USER)
+#define __pgprot_mask(x)	__pgprot((x) & __default_kernel_pte_mask)
 
-#define __PAGE_KERNEL_ENC	(__PAGE_KERNEL | _PAGE_ENC)
-#define __PAGE_KERNEL_ENC_WP	(__PAGE_KERNEL_WP | _PAGE_ENC)
+#define PAGE_KERNEL		__pgprot_mask(__PAGE_KERNEL            | _ENC)
+#define PAGE_KERNEL_NOENC	__pgprot_mask(__PAGE_KERNEL            |    0)
+#define PAGE_KERNEL_RO		__pgprot_mask(__PAGE_KERNEL_RO         | _ENC)
+#define PAGE_KERNEL_EXEC	__pgprot_mask(__PAGE_KERNEL_EXEC       | _ENC)
+#define PAGE_KERNEL_EXEC_NOENC	__pgprot_mask(__PAGE_KERNEL_EXEC       |    0)
+#define PAGE_KERNEL_RX		__pgprot_mask(__PAGE_KERNEL_RX         | _ENC)
+#define PAGE_KERNEL_NOCACHE	__pgprot_mask(__PAGE_KERNEL_NOCACHE    | _ENC)
+#define PAGE_KERNEL_LARGE	__pgprot_mask(__PAGE_KERNEL_LARGE      | _ENC)
+#define PAGE_KERNEL_LARGE_EXEC	__pgprot_mask(__PAGE_KERNEL_LARGE_EXEC | _ENC)
+#define PAGE_KERNEL_VVAR	__pgprot_mask(__PAGE_KERNEL_VVAR       | _ENC)
 
-#define __PAGE_KERNEL_NOENC	(__PAGE_KERNEL)
-#define __PAGE_KERNEL_NOENC_WP	(__PAGE_KERNEL_WP)
-
-#define default_pgprot(x)	__pgprot((x) & __default_kernel_pte_mask)
-
-#define PAGE_KERNEL		default_pgprot(__PAGE_KERNEL | _PAGE_ENC)
-#define PAGE_KERNEL_NOENC	default_pgprot(__PAGE_KERNEL)
-#define PAGE_KERNEL_RO		default_pgprot(__PAGE_KERNEL_RO | _PAGE_ENC)
-#define PAGE_KERNEL_EXEC	default_pgprot(__PAGE_KERNEL_EXEC | _PAGE_ENC)
-#define PAGE_KERNEL_EXEC_NOENC	default_pgprot(__PAGE_KERNEL_EXEC)
-#define PAGE_KERNEL_RX		default_pgprot(__PAGE_KERNEL_RX | _PAGE_ENC)
-#define PAGE_KERNEL_NOCACHE	default_pgprot(__PAGE_KERNEL_NOCACHE | _PAGE_ENC)
-#define PAGE_KERNEL_LARGE	default_pgprot(__PAGE_KERNEL_LARGE | _PAGE_ENC)
-#define PAGE_KERNEL_LARGE_EXEC	default_pgprot(__PAGE_KERNEL_LARGE_EXEC | _PAGE_ENC)
-#define PAGE_KERNEL_VVAR	default_pgprot(__PAGE_KERNEL_VVAR | _PAGE_ENC)
-
-#define PAGE_KERNEL_IO		default_pgprot(__PAGE_KERNEL_IO)
-#define PAGE_KERNEL_IO_NOCACHE	default_pgprot(__PAGE_KERNEL_IO_NOCACHE)
+#define PAGE_KERNEL_IO		__pgprot_mask(__PAGE_KERNEL_IO)
+#define PAGE_KERNEL_IO_NOCACHE	__pgprot_mask(__PAGE_KERNEL_IO_NOCACHE)
 
 #endif	/* __ASSEMBLY__ */
 
@@ -449,9 +457,6 @@ static inline pteval_t pte_flags(pte_t pte)
 	return native_pte_val(pte) & PTE_FLAGS_MASK;
 }
 
-#define pgprot_val(x)	((x).pgprot)
-#define __pgprot(x)	((pgprot_t) { (x) } )
-
 extern uint16_t __cachemode2pte_tbl[_PAGE_CACHE_MODE_NUM];
 extern uint8_t __pte2cachemode_tbl[8];
 

From da9144c5ad8943f9003ec4a6a0200637b4ba9ebd Mon Sep 17 00:00:00 2001
From: kbuild test robot <lkp@intel.com>
Date: Sat, 23 Nov 2019 23:30:23 +0800
Subject: [PATCH 15/17] x86/mm/pat: Mark __cpa_flush_tlb() as static

Signed-off-by: kbuild test robot <lkp@intel.com>
Link: https://lkml.kernel.org/r/20191123153023.bj6m66scjeubhbjg@4978f4969bb8
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/mm/pat/set_memory.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c
index d4ab493b1647..20823392f4f2 100644
--- a/arch/x86/mm/pat/set_memory.c
+++ b/arch/x86/mm/pat/set_memory.c
@@ -331,7 +331,7 @@ static void cpa_flush_all(unsigned long cache)
 	on_each_cpu(__cpa_flush_all, (void *) cache, 1);
 }
 
-void __cpa_flush_tlb(void *data)
+static void __cpa_flush_tlb(void *data)
 {
 	struct cpa_data *cpa = data;
 	unsigned int i;

From 1f059dfdf5d170dccbac92193be2fee3c1763384 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Thu, 28 Nov 2019 08:19:36 +0100
Subject: [PATCH 16/17] mm/vmalloc: Add empty <asm/vmalloc.h> headers and use
 them from <linux/vmalloc.h>

In the x86 MM code we'd like to untangle various types of historic
header dependency spaghetti, but for this we'd need to pass to
the generic vmalloc code various vmalloc related defines that
customarily come via the <asm/page.h> low level arch header.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/alpha/include/asm/vmalloc.h      | 4 ++++
 arch/arc/include/asm/vmalloc.h        | 4 ++++
 arch/arm/include/asm/vmalloc.h        | 4 ++++
 arch/arm64/include/asm/vmalloc.h      | 4 ++++
 arch/c6x/include/asm/vmalloc.h        | 4 ++++
 arch/csky/include/asm/vmalloc.h       | 4 ++++
 arch/h8300/include/asm/vmalloc.h      | 4 ++++
 arch/hexagon/include/asm/vmalloc.h    | 4 ++++
 arch/ia64/include/asm/vmalloc.h       | 4 ++++
 arch/m68k/include/asm/vmalloc.h       | 4 ++++
 arch/microblaze/include/asm/vmalloc.h | 4 ++++
 arch/mips/include/asm/vmalloc.h       | 4 ++++
 arch/nds32/include/asm/vmalloc.h      | 4 ++++
 arch/nios2/include/asm/vmalloc.h      | 4 ++++
 arch/openrisc/include/asm/vmalloc.h   | 4 ++++
 arch/parisc/include/asm/vmalloc.h     | 4 ++++
 arch/powerpc/include/asm/vmalloc.h    | 4 ++++
 arch/riscv/include/asm/vmalloc.h      | 4 ++++
 arch/s390/include/asm/vmalloc.h       | 4 ++++
 arch/sh/include/asm/vmalloc.h         | 4 ++++
 arch/sparc/include/asm/vmalloc.h      | 4 ++++
 arch/um/include/asm/vmalloc.h         | 4 ++++
 arch/unicore32/include/asm/vmalloc.h  | 4 ++++
 arch/x86/include/asm/vmalloc.h        | 4 ++++
 arch/xtensa/include/asm/vmalloc.h     | 4 ++++
 include/linux/vmalloc.h               | 2 ++
 26 files changed, 102 insertions(+)
 create mode 100644 arch/alpha/include/asm/vmalloc.h
 create mode 100644 arch/arc/include/asm/vmalloc.h
 create mode 100644 arch/arm/include/asm/vmalloc.h
 create mode 100644 arch/arm64/include/asm/vmalloc.h
 create mode 100644 arch/c6x/include/asm/vmalloc.h
 create mode 100644 arch/csky/include/asm/vmalloc.h
 create mode 100644 arch/h8300/include/asm/vmalloc.h
 create mode 100644 arch/hexagon/include/asm/vmalloc.h
 create mode 100644 arch/ia64/include/asm/vmalloc.h
 create mode 100644 arch/m68k/include/asm/vmalloc.h
 create mode 100644 arch/microblaze/include/asm/vmalloc.h
 create mode 100644 arch/mips/include/asm/vmalloc.h
 create mode 100644 arch/nds32/include/asm/vmalloc.h
 create mode 100644 arch/nios2/include/asm/vmalloc.h
 create mode 100644 arch/openrisc/include/asm/vmalloc.h
 create mode 100644 arch/parisc/include/asm/vmalloc.h
 create mode 100644 arch/powerpc/include/asm/vmalloc.h
 create mode 100644 arch/riscv/include/asm/vmalloc.h
 create mode 100644 arch/s390/include/asm/vmalloc.h
 create mode 100644 arch/sh/include/asm/vmalloc.h
 create mode 100644 arch/sparc/include/asm/vmalloc.h
 create mode 100644 arch/um/include/asm/vmalloc.h
 create mode 100644 arch/unicore32/include/asm/vmalloc.h
 create mode 100644 arch/x86/include/asm/vmalloc.h
 create mode 100644 arch/xtensa/include/asm/vmalloc.h

diff --git a/arch/alpha/include/asm/vmalloc.h b/arch/alpha/include/asm/vmalloc.h
new file mode 100644
index 000000000000..0a9a366a4d34
--- /dev/null
+++ b/arch/alpha/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_ALPHA_VMALLOC_H
+#define _ASM_ALPHA_VMALLOC_H
+
+#endif /* _ASM_ALPHA_VMALLOC_H */
diff --git a/arch/arc/include/asm/vmalloc.h b/arch/arc/include/asm/vmalloc.h
new file mode 100644
index 000000000000..973095aad665
--- /dev/null
+++ b/arch/arc/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_ARC_VMALLOC_H
+#define _ASM_ARC_VMALLOC_H
+
+#endif /* _ASM_ARC_VMALLOC_H */
diff --git a/arch/arm/include/asm/vmalloc.h b/arch/arm/include/asm/vmalloc.h
new file mode 100644
index 000000000000..a9b3718b8600
--- /dev/null
+++ b/arch/arm/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_ARM_VMALLOC_H
+#define _ASM_ARM_VMALLOC_H
+
+#endif /* _ASM_ARM_VMALLOC_H */
diff --git a/arch/arm64/include/asm/vmalloc.h b/arch/arm64/include/asm/vmalloc.h
new file mode 100644
index 000000000000..2ca708ab9b20
--- /dev/null
+++ b/arch/arm64/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_ARM64_VMALLOC_H
+#define _ASM_ARM64_VMALLOC_H
+
+#endif /* _ASM_ARM64_VMALLOC_H */
diff --git a/arch/c6x/include/asm/vmalloc.h b/arch/c6x/include/asm/vmalloc.h
new file mode 100644
index 000000000000..26c6c6696bbd
--- /dev/null
+++ b/arch/c6x/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_C6X_VMALLOC_H
+#define _ASM_C6X_VMALLOC_H
+
+#endif /* _ASM_C6X_VMALLOC_H */
diff --git a/arch/csky/include/asm/vmalloc.h b/arch/csky/include/asm/vmalloc.h
new file mode 100644
index 000000000000..43dca6336b4c
--- /dev/null
+++ b/arch/csky/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_CSKY_VMALLOC_H
+#define _ASM_CSKY_VMALLOC_H
+
+#endif /* _ASM_CSKY_VMALLOC_H */
diff --git a/arch/h8300/include/asm/vmalloc.h b/arch/h8300/include/asm/vmalloc.h
new file mode 100644
index 000000000000..08a55c1dfa23
--- /dev/null
+++ b/arch/h8300/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_H8300_VMALLOC_H
+#define _ASM_H8300_VMALLOC_H
+
+#endif /* _ASM_H8300_VMALLOC_H */
diff --git a/arch/hexagon/include/asm/vmalloc.h b/arch/hexagon/include/asm/vmalloc.h
new file mode 100644
index 000000000000..7b04609e525c
--- /dev/null
+++ b/arch/hexagon/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_HEXAGON_VMALLOC_H
+#define _ASM_HEXAGON_VMALLOC_H
+
+#endif /* _ASM_HEXAGON_VMALLOC_H */
diff --git a/arch/ia64/include/asm/vmalloc.h b/arch/ia64/include/asm/vmalloc.h
new file mode 100644
index 000000000000..a2b51141ad28
--- /dev/null
+++ b/arch/ia64/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_IA64_VMALLOC_H
+#define _ASM_IA64_VMALLOC_H
+
+#endif /* _ASM_IA64_VMALLOC_H */
diff --git a/arch/m68k/include/asm/vmalloc.h b/arch/m68k/include/asm/vmalloc.h
new file mode 100644
index 000000000000..bc1dca6cf134
--- /dev/null
+++ b/arch/m68k/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_M68K_VMALLOC_H
+#define _ASM_M68K_VMALLOC_H
+
+#endif /* _ASM_M68K_VMALLOC_H */
diff --git a/arch/microblaze/include/asm/vmalloc.h b/arch/microblaze/include/asm/vmalloc.h
new file mode 100644
index 000000000000..04013a42b0fe
--- /dev/null
+++ b/arch/microblaze/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_MICROBLAZE_VMALLOC_H
+#define _ASM_MICROBLAZE_VMALLOC_H
+
+#endif /* _ASM_MICROBLAZE_VMALLOC_H */
diff --git a/arch/mips/include/asm/vmalloc.h b/arch/mips/include/asm/vmalloc.h
new file mode 100644
index 000000000000..25dc09b25eaf
--- /dev/null
+++ b/arch/mips/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_MIPS_VMALLOC_H
+#define _ASM_MIPS_VMALLOC_H
+
+#endif /* _ASM_MIPS_VMALLOC_H */
diff --git a/arch/nds32/include/asm/vmalloc.h b/arch/nds32/include/asm/vmalloc.h
new file mode 100644
index 000000000000..caeed3898419
--- /dev/null
+++ b/arch/nds32/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_NDS32_VMALLOC_H
+#define _ASM_NDS32_VMALLOC_H
+
+#endif /* _ASM_NDS32_VMALLOC_H */
diff --git a/arch/nios2/include/asm/vmalloc.h b/arch/nios2/include/asm/vmalloc.h
new file mode 100644
index 000000000000..ec7a9260090b
--- /dev/null
+++ b/arch/nios2/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_NIOS2_VMALLOC_H
+#define _ASM_NIOS2_VMALLOC_H
+
+#endif /* _ASM_NIOS2_VMALLOC_H */
diff --git a/arch/openrisc/include/asm/vmalloc.h b/arch/openrisc/include/asm/vmalloc.h
new file mode 100644
index 000000000000..75435eceec32
--- /dev/null
+++ b/arch/openrisc/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_OPENRISC_VMALLOC_H
+#define _ASM_OPENRISC_VMALLOC_H
+
+#endif /* _ASM_OPENRISC_VMALLOC_H */
diff --git a/arch/parisc/include/asm/vmalloc.h b/arch/parisc/include/asm/vmalloc.h
new file mode 100644
index 000000000000..1088ae4e7af9
--- /dev/null
+++ b/arch/parisc/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_PARISC_VMALLOC_H
+#define _ASM_PARISC_VMALLOC_H
+
+#endif /* _ASM_PARISC_VMALLOC_H */
diff --git a/arch/powerpc/include/asm/vmalloc.h b/arch/powerpc/include/asm/vmalloc.h
new file mode 100644
index 000000000000..b992dfaaa161
--- /dev/null
+++ b/arch/powerpc/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_POWERPC_VMALLOC_H
+#define _ASM_POWERPC_VMALLOC_H
+
+#endif /* _ASM_POWERPC_VMALLOC_H */
diff --git a/arch/riscv/include/asm/vmalloc.h b/arch/riscv/include/asm/vmalloc.h
new file mode 100644
index 000000000000..ff9abc00d139
--- /dev/null
+++ b/arch/riscv/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_RISCV_VMALLOC_H
+#define _ASM_RISCV_VMALLOC_H
+
+#endif /* _ASM_RISCV_VMALLOC_H */
diff --git a/arch/s390/include/asm/vmalloc.h b/arch/s390/include/asm/vmalloc.h
new file mode 100644
index 000000000000..3ba3a6bdca25
--- /dev/null
+++ b/arch/s390/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_S390_VMALLOC_H
+#define _ASM_S390_VMALLOC_H
+
+#endif /* _ASM_S390_VMALLOC_H */
diff --git a/arch/sh/include/asm/vmalloc.h b/arch/sh/include/asm/vmalloc.h
new file mode 100644
index 000000000000..716b77472646
--- /dev/null
+++ b/arch/sh/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_SH_VMALLOC_H
+#define _ASM_SH_VMALLOC_H
+
+#endif /* _ASM_SH_VMALLOC_H */
diff --git a/arch/sparc/include/asm/vmalloc.h b/arch/sparc/include/asm/vmalloc.h
new file mode 100644
index 000000000000..04b8ab9518b8
--- /dev/null
+++ b/arch/sparc/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_SPARC_VMALLOC_H
+#define _ASM_SPARC_VMALLOC_H
+
+#endif /* _ASM_SPARC_VMALLOC_H */
diff --git a/arch/um/include/asm/vmalloc.h b/arch/um/include/asm/vmalloc.h
new file mode 100644
index 000000000000..9a7b9ed93733
--- /dev/null
+++ b/arch/um/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_UM_VMALLOC_H
+#define _ASM_UM_VMALLOC_H
+
+#endif /* _ASM_UM_VMALLOC_H */
diff --git a/arch/unicore32/include/asm/vmalloc.h b/arch/unicore32/include/asm/vmalloc.h
new file mode 100644
index 000000000000..054435818a14
--- /dev/null
+++ b/arch/unicore32/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_UNICORE32_VMALLOC_H
+#define _ASM_UNICORE32_VMALLOC_H
+
+#endif /* _ASM_UNICORE32_VMALLOC_H */
diff --git a/arch/x86/include/asm/vmalloc.h b/arch/x86/include/asm/vmalloc.h
new file mode 100644
index 000000000000..ba6295fa5876
--- /dev/null
+++ b/arch/x86/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_X86_VMALLOC_H
+#define _ASM_X86_VMALLOC_H
+
+#endif /* _ASM_X86_VMALLOC_H */
diff --git a/arch/xtensa/include/asm/vmalloc.h b/arch/xtensa/include/asm/vmalloc.h
new file mode 100644
index 000000000000..0eb94b70be55
--- /dev/null
+++ b/arch/xtensa/include/asm/vmalloc.h
@@ -0,0 +1,4 @@
+#ifndef _ASM_XTENSA_VMALLOC_H
+#define _ASM_XTENSA_VMALLOC_H
+
+#endif /* _ASM_XTENSA_VMALLOC_H */
diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h
index a4b241102771..ec3813236699 100644
--- a/include/linux/vmalloc.h
+++ b/include/linux/vmalloc.h
@@ -10,6 +10,8 @@
 #include <linux/rbtree.h>
 #include <linux/overflow.h>
 
+#include <asm/vmalloc.h>
+
 struct vm_area_struct;		/* vma defining user mapping in mm_types.h */
 struct notifier_block;		/* in notifier.h */
 

From 186525bd6b83efc592672e2d6185e4d7c810d2b4 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@kernel.org>
Date: Fri, 29 Nov 2019 08:17:25 +0100
Subject: [PATCH 17/17] mm, x86/mm: Untangle address space layout definitions
 from basic pgtable type definitions

- Untangle the somewhat incestous way of how VMALLOC_START is used all across the
  kernel, but is, on x86, defined deep inside one of the lowest level page table headers.
  It doesn't help that vmalloc.h only includes a single asm header:

     #include <asm/page.h>           /* pgprot_t */

  So there was no existing cross-arch way to decouple address layout
  definitions from page.h details. I used this:

   #ifndef VMALLOC_START
   # include <asm/vmalloc.h>
   #endif

  This way every architecture that wants to simplify page.h can do so.

- Also on x86 we had a couple of LDT related inline functions that used
  the late-stage address space layout positions - but these could be
  uninlined without real trouble - the end result is cleaner this way as
  well.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: linux-kernel@vger.kernel.org
Cc: linux-mm@kvack.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/cpu_entry_area.h   | 10 +--
 arch/x86/include/asm/mmu_context.h      | 86 ++-----------------------
 arch/x86/include/asm/pgtable_32_areas.h | 53 +++++++++++++++
 arch/x86/include/asm/pgtable_32_types.h | 57 +---------------
 arch/x86/include/asm/pgtable_areas.h    | 16 +++++
 arch/x86/include/asm/vmalloc.h          |  2 +
 arch/x86/kernel/ldt.c                   | 83 ++++++++++++++++++++++++
 arch/x86/kernel/setup.c                 |  1 +
 arch/x86/mm/fault.c                     |  1 +
 arch/x86/mm/init_32.c                   |  1 +
 arch/x86/mm/pgtable_32.c                |  1 +
 arch/x86/mm/physaddr.c                  |  1 +
 include/linux/mm.h                      | 15 ++---
 mm/highmem.c                            |  2 +-
 mm/vmalloc.c                            |  8 +++
 15 files changed, 183 insertions(+), 154 deletions(-)
 create mode 100644 arch/x86/include/asm/pgtable_32_areas.h
 create mode 100644 arch/x86/include/asm/pgtable_areas.h

diff --git a/arch/x86/include/asm/cpu_entry_area.h b/arch/x86/include/asm/cpu_entry_area.h
index 804734058c77..02c0078d3787 100644
--- a/arch/x86/include/asm/cpu_entry_area.h
+++ b/arch/x86/include/asm/cpu_entry_area.h
@@ -6,6 +6,7 @@
 #include <linux/percpu-defs.h>
 #include <asm/processor.h>
 #include <asm/intel_ds.h>
+#include <asm/pgtable_areas.h>
 
 #ifdef CONFIG_X86_64
 
@@ -134,15 +135,6 @@ DECLARE_PER_CPU(struct cea_exception_stacks *, cea_exception_stacks);
 extern void setup_cpu_entry_areas(void);
 extern void cea_set_pte(void *cea_vaddr, phys_addr_t pa, pgprot_t flags);
 
-/* Single page reserved for the readonly IDT mapping: */
-#define	CPU_ENTRY_AREA_RO_IDT		CPU_ENTRY_AREA_BASE
-#define CPU_ENTRY_AREA_PER_CPU		(CPU_ENTRY_AREA_RO_IDT + PAGE_SIZE)
-
-#define CPU_ENTRY_AREA_RO_IDT_VADDR	((void *)CPU_ENTRY_AREA_RO_IDT)
-
-#define CPU_ENTRY_AREA_MAP_SIZE			\
-	(CPU_ENTRY_AREA_PER_CPU + CPU_ENTRY_AREA_ARRAY_SIZE - CPU_ENTRY_AREA_BASE)
-
 extern struct cpu_entry_area *get_cpu_entry_area(int cpu);
 
 static inline struct entry_stack *cpu_entry_stack(int cpu)
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 5f33924e200f..b243234e90cb 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -69,14 +69,6 @@ struct ldt_struct {
 	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)
-{
-	return (void *)(LDT_BASE_ADDR + LDT_SLOT_STRIDE * slot);
-}
-
 /*
  * Used for LDT copy/destruction.
  */
@@ -99,87 +91,21 @@ static inline void destroy_context_ldt(struct mm_struct *mm) { }
 static inline void ldt_arch_exit_mmap(struct mm_struct *mm) { }
 #endif
 
+#ifdef CONFIG_MODIFY_LDT_SYSCALL
+extern void load_mm_ldt(struct mm_struct *mm);
+extern void switch_ldt(struct mm_struct *prev, struct mm_struct *next);
+#else
 static inline void load_mm_ldt(struct mm_struct *mm)
 {
-#ifdef CONFIG_MODIFY_LDT_SYSCALL
-	struct ldt_struct *ldt;
-
-	/* READ_ONCE synchronizes with smp_store_release */
-	ldt = READ_ONCE(mm->context.ldt);
-
-	/*
-	 * Any change to mm->context.ldt is followed by an IPI to all
-	 * CPUs with the mm active.  The LDT will not be freed until
-	 * after the IPI is handled by all such CPUs.  This means that,
-	 * if the ldt_struct changes before we return, the values we see
-	 * will be safe, and the new values will be loaded before we run
-	 * any user code.
-	 *
-	 * NB: don't try to convert this to use RCU without extreme care.
-	 * We would still need IRQs off, because we don't want to change
-	 * the local LDT after an IPI loaded a newer value than the one
-	 * that we can see.
-	 */
-
-	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
 }
-
 static inline void switch_ldt(struct mm_struct *prev, struct mm_struct *next)
 {
-#ifdef CONFIG_MODIFY_LDT_SYSCALL
-	/*
-	 * Load the LDT if either the old or new mm had an LDT.
-	 *
-	 * An mm will never go from having an LDT to not having an LDT.  Two
-	 * mms never share an LDT, so we don't gain anything by checking to
-	 * see whether the LDT changed.  There's also no guarantee that
-	 * prev->context.ldt actually matches LDTR, but, if LDTR is non-NULL,
-	 * then prev->context.ldt will also be non-NULL.
-	 *
-	 * If we really cared, we could optimize the case where prev == next
-	 * and we're exiting lazy mode.  Most of the time, if this happens,
-	 * we don't actually need to reload LDTR, but modify_ldt() is mostly
-	 * used by legacy code and emulators where we don't need this level of
-	 * performance.
-	 *
-	 * This uses | instead of || because it generates better code.
-	 */
-	if (unlikely((unsigned long)prev->context.ldt |
-		     (unsigned long)next->context.ldt))
-		load_mm_ldt(next);
-#endif
-
 	DEBUG_LOCKS_WARN_ON(preemptible());
 }
+#endif
 
-void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk);
+extern void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk);
 
 /*
  * Init a new mm.  Used on mm copies, like at fork()
diff --git a/arch/x86/include/asm/pgtable_32_areas.h b/arch/x86/include/asm/pgtable_32_areas.h
new file mode 100644
index 000000000000..b6355416a15a
--- /dev/null
+++ b/arch/x86/include/asm/pgtable_32_areas.h
@@ -0,0 +1,53 @@
+#ifndef _ASM_X86_PGTABLE_32_AREAS_H
+#define _ASM_X86_PGTABLE_32_AREAS_H
+
+#include <asm/cpu_entry_area.h>
+
+/*
+ * Just any arbitrary offset to the start of the vmalloc VM area: the
+ * current 8MB value just means that there will be a 8MB "hole" after the
+ * physical memory until the kernel virtual memory starts.  That means that
+ * any out-of-bounds memory accesses will hopefully be caught.
+ * The vmalloc() routines leaves a hole of 4kB between each vmalloced
+ * area for the same reason. ;)
+ */
+#define VMALLOC_OFFSET	(8 * 1024 * 1024)
+
+#ifndef __ASSEMBLY__
+extern bool __vmalloc_start_set; /* set once high_memory is set */
+#endif
+
+#define VMALLOC_START	((unsigned long)high_memory + VMALLOC_OFFSET)
+#ifdef CONFIG_X86_PAE
+#define LAST_PKMAP 512
+#else
+#define LAST_PKMAP 1024
+#endif
+
+#define CPU_ENTRY_AREA_PAGES		(NR_CPUS * DIV_ROUND_UP(sizeof(struct cpu_entry_area), PAGE_SIZE))
+
+/* The +1 is for the readonly IDT page: */
+#define CPU_ENTRY_AREA_BASE	\
+	((FIXADDR_TOT_START - PAGE_SIZE*(CPU_ENTRY_AREA_PAGES+1)) & PMD_MASK)
+
+#define LDT_BASE_ADDR		\
+	((CPU_ENTRY_AREA_BASE - PAGE_SIZE) & PMD_MASK)
+
+#define LDT_END_ADDR		(LDT_BASE_ADDR + PMD_SIZE)
+
+#define PKMAP_BASE		\
+	((LDT_BASE_ADDR - PAGE_SIZE) & PMD_MASK)
+
+#ifdef CONFIG_HIGHMEM
+# define VMALLOC_END	(PKMAP_BASE - 2 * PAGE_SIZE)
+#else
+# define VMALLOC_END	(LDT_BASE_ADDR - 2 * PAGE_SIZE)
+#endif
+
+#define MODULES_VADDR	VMALLOC_START
+#define MODULES_END	VMALLOC_END
+#define MODULES_LEN	(MODULES_VADDR - MODULES_END)
+
+#define MAXMEM	(VMALLOC_END - PAGE_OFFSET - __VMALLOC_RESERVE)
+
+#endif /* _ASM_X86_PGTABLE_32_AREAS_H */
diff --git a/arch/x86/include/asm/pgtable_32_types.h b/arch/x86/include/asm/pgtable_32_types.h
index 0416d42e5bdd..5356a46b0373 100644
--- a/arch/x86/include/asm/pgtable_32_types.h
+++ b/arch/x86/include/asm/pgtable_32_types.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0 */
-#ifndef _ASM_X86_PGTABLE_32_DEFS_H
-#define _ASM_X86_PGTABLE_32_DEFS_H
+#ifndef _ASM_X86_PGTABLE_32_TYPES_H
+#define _ASM_X86_PGTABLE_32_TYPES_H
 
 /*
  * The Linux x86 paging architecture is 'compile-time dual-mode', it
@@ -20,55 +20,4 @@
 #define PGDIR_SIZE	(1UL << PGDIR_SHIFT)
 #define PGDIR_MASK	(~(PGDIR_SIZE - 1))
 
-/* Just any arbitrary offset to the start of the vmalloc VM area: the
- * current 8MB value just means that there will be a 8MB "hole" after the
- * physical memory until the kernel virtual memory starts.  That means that
- * any out-of-bounds memory accesses will hopefully be caught.
- * The vmalloc() routines leaves a hole of 4kB between each vmalloced
- * area for the same reason. ;)
- */
-#define VMALLOC_OFFSET	(8 * 1024 * 1024)
-
-#ifndef __ASSEMBLY__
-extern bool __vmalloc_start_set; /* set once high_memory is set */
-#endif
-
-#define VMALLOC_START	((unsigned long)high_memory + VMALLOC_OFFSET)
-#ifdef CONFIG_X86_PAE
-#define LAST_PKMAP 512
-#else
-#define LAST_PKMAP 1024
-#endif
-
-/*
- * This is an upper bound on sizeof(struct cpu_entry_area) / PAGE_SIZE.
- * Define this here and validate with BUILD_BUG_ON() in cpu_entry_area.c
- * to avoid include recursion hell.
- */
-#define CPU_ENTRY_AREA_PAGES	(NR_CPUS * 43)
-
-/* The +1 is for the readonly IDT page: */
-#define CPU_ENTRY_AREA_BASE	\
-	((FIXADDR_TOT_START - PAGE_SIZE*(CPU_ENTRY_AREA_PAGES+1)) & PMD_MASK)
-
-#define LDT_BASE_ADDR		\
-	((CPU_ENTRY_AREA_BASE - PAGE_SIZE) & PMD_MASK)
-
-#define LDT_END_ADDR		(LDT_BASE_ADDR + PMD_SIZE)
-
-#define PKMAP_BASE		\
-	((LDT_BASE_ADDR - PAGE_SIZE) & PMD_MASK)
-
-#ifdef CONFIG_HIGHMEM
-# define VMALLOC_END	(PKMAP_BASE - 2 * PAGE_SIZE)
-#else
-# define VMALLOC_END	(LDT_BASE_ADDR - 2 * PAGE_SIZE)
-#endif
-
-#define MODULES_VADDR	VMALLOC_START
-#define MODULES_END	VMALLOC_END
-#define MODULES_LEN	(MODULES_VADDR - MODULES_END)
-
-#define MAXMEM	(VMALLOC_END - PAGE_OFFSET - __VMALLOC_RESERVE)
-
-#endif /* _ASM_X86_PGTABLE_32_DEFS_H */
+#endif /* _ASM_X86_PGTABLE_32_TYPES_H */
diff --git a/arch/x86/include/asm/pgtable_areas.h b/arch/x86/include/asm/pgtable_areas.h
new file mode 100644
index 000000000000..d34cce1b995c
--- /dev/null
+++ b/arch/x86/include/asm/pgtable_areas.h
@@ -0,0 +1,16 @@
+#ifndef _ASM_X86_PGTABLE_AREAS_H
+#define _ASM_X86_PGTABLE_AREAS_H
+
+#ifdef CONFIG_X86_32
+# include <asm/pgtable_32_areas.h>
+#endif
+
+/* Single page reserved for the readonly IDT mapping: */
+#define CPU_ENTRY_AREA_RO_IDT		CPU_ENTRY_AREA_BASE
+#define CPU_ENTRY_AREA_PER_CPU		(CPU_ENTRY_AREA_RO_IDT + PAGE_SIZE)
+
+#define CPU_ENTRY_AREA_RO_IDT_VADDR	((void *)CPU_ENTRY_AREA_RO_IDT)
+
+#define CPU_ENTRY_AREA_MAP_SIZE		(CPU_ENTRY_AREA_PER_CPU + CPU_ENTRY_AREA_ARRAY_SIZE - CPU_ENTRY_AREA_BASE)
+
+#endif /* _ASM_X86_PGTABLE_AREAS_H */
diff --git a/arch/x86/include/asm/vmalloc.h b/arch/x86/include/asm/vmalloc.h
index ba6295fa5876..29837740b520 100644
--- a/arch/x86/include/asm/vmalloc.h
+++ b/arch/x86/include/asm/vmalloc.h
@@ -1,4 +1,6 @@
 #ifndef _ASM_X86_VMALLOC_H
 #define _ASM_X86_VMALLOC_H
 
+#include <asm/pgtable_areas.h>
+
 #endif /* _ASM_X86_VMALLOC_H */
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index b2463fcb20a8..c57e1ca70fd1 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -28,6 +28,89 @@
 #include <asm/desc.h>
 #include <asm/mmu_context.h>
 #include <asm/syscalls.h>
+#include <asm/pgtable_areas.h>
+
+/* This is a multiple of PAGE_SIZE. */
+#define LDT_SLOT_STRIDE (LDT_ENTRIES * LDT_ENTRY_SIZE)
+
+static inline void *ldt_slot_va(int slot)
+{
+	return (void *)(LDT_BASE_ADDR + LDT_SLOT_STRIDE * slot);
+}
+
+void load_mm_ldt(struct mm_struct *mm)
+{
+	struct ldt_struct *ldt;
+
+	/* READ_ONCE synchronizes with smp_store_release */
+	ldt = READ_ONCE(mm->context.ldt);
+
+	/*
+	 * Any change to mm->context.ldt is followed by an IPI to all
+	 * CPUs with the mm active.  The LDT will not be freed until
+	 * after the IPI is handled by all such CPUs.  This means that,
+	 * if the ldt_struct changes before we return, the values we see
+	 * will be safe, and the new values will be loaded before we run
+	 * any user code.
+	 *
+	 * NB: don't try to convert this to use RCU without extreme care.
+	 * We would still need IRQs off, because we don't want to change
+	 * the local LDT after an IPI loaded a newer value than the one
+	 * that we can see.
+	 */
+
+	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();
+	}
+}
+
+void switch_ldt(struct mm_struct *prev, struct mm_struct *next)
+{
+	/*
+	 * Load the LDT if either the old or new mm had an LDT.
+	 *
+	 * An mm will never go from having an LDT to not having an LDT.  Two
+	 * mms never share an LDT, so we don't gain anything by checking to
+	 * see whether the LDT changed.  There's also no guarantee that
+	 * prev->context.ldt actually matches LDTR, but, if LDTR is non-NULL,
+	 * then prev->context.ldt will also be non-NULL.
+	 *
+	 * If we really cared, we could optimize the case where prev == next
+	 * and we're exiting lazy mode.  Most of the time, if this happens,
+	 * we don't actually need to reload LDTR, but modify_ldt() is mostly
+	 * used by legacy code and emulators where we don't need this level of
+	 * performance.
+	 *
+	 * This uses | instead of || because it generates better code.
+	 */
+	if (unlikely((unsigned long)prev->context.ldt |
+		     (unsigned long)next->context.ldt))
+		load_mm_ldt(next);
+
+	DEBUG_LOCKS_WARN_ON(preemptible());
+}
 
 static void refresh_ldt_segments(void)
 {
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index b5ac9932bcf6..90296a04e5ad 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -41,6 +41,7 @@
 #include <asm/proto.h>
 #include <asm/unwind.h>
 #include <asm/vsyscall.h>
+#include <linux/vmalloc.h>
 
 /*
  * max_low_pfn_mapped: highest directly mapped pfn < 4 GB
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 304d31d8cbbc..c9c8523a3a48 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -29,6 +29,7 @@
 #include <asm/efi.h>			/* efi_recover_from_page_fault()*/
 #include <asm/desc.h>			/* store_idt(), ...		*/
 #include <asm/cpu_entry_area.h>		/* exception stack		*/
+#include <asm/pgtable_areas.h>		/* VMALLOC_START, ...		*/
 
 #define CREATE_TRACE_POINTS
 #include <asm/trace/exceptions.h>
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index 930edeb41ec3..16274a3f751c 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -52,6 +52,7 @@
 #include <asm/page_types.h>
 #include <asm/cpu_entry_area.h>
 #include <asm/init.h>
+#include <asm/pgtable_areas.h>
 
 #include "mm_internal.h"
 
diff --git a/arch/x86/mm/pgtable_32.c b/arch/x86/mm/pgtable_32.c
index 9bb7f0ab9fe6..0e6700eaa4f9 100644
--- a/arch/x86/mm/pgtable_32.c
+++ b/arch/x86/mm/pgtable_32.c
@@ -18,6 +18,7 @@
 #include <asm/tlb.h>
 #include <asm/tlbflush.h>
 #include <asm/io.h>
+#include <linux/vmalloc.h>
 
 unsigned int __VMALLOC_RESERVE = 128 << 20;
 
diff --git a/arch/x86/mm/physaddr.c b/arch/x86/mm/physaddr.c
index bdc98150d4db..fc3f3d3e2ef2 100644
--- a/arch/x86/mm/physaddr.c
+++ b/arch/x86/mm/physaddr.c
@@ -5,6 +5,7 @@
 #include <linux/mm.h>
 
 #include <asm/page.h>
+#include <linux/vmalloc.h>
 
 #include "physaddr.h"
 
diff --git a/include/linux/mm.h b/include/linux/mm.h
index c97ea3b694e6..fb8f9412e2cf 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -625,24 +625,19 @@ unsigned long vmalloc_to_pfn(const void *addr);
  * On nommu, vmalloc/vfree wrap through kmalloc/kfree directly, so there
  * is no special casing required.
  */
-static inline bool is_vmalloc_addr(const void *x)
-{
-#ifdef CONFIG_MMU
-	unsigned long addr = (unsigned long)x;
-
-	return addr >= VMALLOC_START && addr < VMALLOC_END;
-#else
-	return false;
-#endif
-}
 
 #ifndef is_ioremap_addr
 #define is_ioremap_addr(x) is_vmalloc_addr(x)
 #endif
 
 #ifdef CONFIG_MMU
+extern bool is_vmalloc_addr(const void *x);
 extern int is_vmalloc_or_module_addr(const void *x);
 #else
+static inline bool is_vmalloc_addr(const void *x)
+{
+	return false;
+}
 static inline int is_vmalloc_or_module_addr(const void *x)
 {
 	return 0;
diff --git a/mm/highmem.c b/mm/highmem.c
index 107b10f9878e..64d8dea47dd1 100644
--- a/mm/highmem.c
+++ b/mm/highmem.c
@@ -29,7 +29,7 @@
 #include <linux/highmem.h>
 #include <linux/kgdb.h>
 #include <asm/tlbflush.h>
-
+#include <linux/vmalloc.h>
 
 #if defined(CONFIG_HIGHMEM) || defined(CONFIG_X86_32)
 DEFINE_PER_CPU(int, __kmap_atomic_idx);
diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 4d3b3d60d893..19cdbb11fe07 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -41,6 +41,14 @@
 
 #include "internal.h"
 
+bool is_vmalloc_addr(const void *x)
+{
+	unsigned long addr = (unsigned long)x;
+
+	return addr >= VMALLOC_START && addr < VMALLOC_END;
+}
+EXPORT_SYMBOL(is_vmalloc_addr);
+
 struct vfree_deferred {
 	struct llist_head list;
 	struct work_struct wq;