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

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

   - reduce the x86/32 PAE per task PGD allocation overhead from 4K to
     0.032k (Fenghua Yu)

   - early_ioremap/memunmap() usage cleanups (Juergen Gross)

   - gbpages support cleanups (Luis R Rodriguez)

   - improve AMD Bulldozer (family 0x15) ASLR I$ aliasing workaround to
     increase randomization by 3 bits (per bootup) (Hector
     Marco-Gisbert)

   - misc fixlets"

* 'x86-mm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/mm: Improve AMD Bulldozer ASLR workaround
  x86/mm/pat: Initialize __cachemode2pte_tbl[] and __pte2cachemode_tbl[] in a bit more readable fashion
  init.h: Clean up the __setup()/early_param() macros
  x86/mm: Simplify probe_page_size_mask()
  x86/mm: Further simplify 1 GB kernel linear mappings handling
  x86/mm: Use early_param_on_off() for direct_gbpages
  init.h: Add early_param_on_off()
  x86/mm: Simplify enabling direct_gbpages
  x86/mm: Use IS_ENABLED() for direct_gbpages
  x86/mm: Unexport set_memory_ro() and set_memory_rw()
  x86/mm, efi: Use early_ioremap() in arch/x86/platform/efi/efi-bgrt.c
  x86/mm: Use early_memunmap() instead of early_iounmap()
  x86/mm/pat: Ensure different messages in STRICT_DEVMEM and PAT cases
  x86/mm: Reduce PAE-mode per task pgd allocation overhead from 4K to 32 bytes

5 files changed, 38 insertions, 10 deletions

diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index dd9e50500297..fd470ebf924e 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -5,6 +5,7 @@
 
 #include <linux/io.h>
 #include <linux/sched.h>
+#include <linux/random.h>
 #include <asm/processor.h>
 #include <asm/apic.h>
 #include <asm/cpu.h>
@@ -488,6 +489,9 @@ static void bsp_init_amd(struct cpuinfo_x86 *c)
 
 		va_align.mask	  = (upperbit - 1) & PAGE_MASK;
 		va_align.flags    = ALIGN_VA_32 | ALIGN_VA_64;
+
+		/* A random value per boot for bit slice [12:upper_bit) */
+		va_align.bits = get_random_int() & va_align.mask;
 	}
 }
 
diff --git a/arch/x86/kernel/devicetree.c b/arch/x86/kernel/devicetree.c
index 3d3503351242..6367a780cc8c 100644
--- a/arch/x86/kernel/devicetree.c
+++ b/arch/x86/kernel/devicetree.c
@@ -286,13 +286,13 @@ static void __init x86_flattree_get_config(void)
 	initial_boot_params = dt = early_memremap(initial_dtb, map_len);
 	size = of_get_flat_dt_size();
 	if (map_len < size) {
-		early_iounmap(dt, map_len);
+		early_memunmap(dt, map_len);
 		initial_boot_params = dt = early_memremap(initial_dtb, size);
 		map_len = size;
 	}
 
 	unflatten_and_copy_device_tree();
-	early_iounmap(dt, map_len);
+	early_memunmap(dt, map_len);
 }
 #else
 static inline void x86_flattree_get_config(void) { }
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 46201deee923..7d46bb260334 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -661,7 +661,7 @@ void __init parse_e820_ext(u64 phys_addr, u32 data_len)
 	extmap = (struct e820entry *)(sdata->data);
 	__append_e820_map(extmap, entries);
 	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
-	early_iounmap(sdata, data_len);
+	early_memunmap(sdata, data_len);
 	printk(KERN_INFO "e820: extended physical RAM map:\n");
 	e820_print_map("extended");
 }
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 014466b152b5..d74ac33290ae 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -354,7 +354,7 @@ static void __init relocate_initrd(void)
 		mapaddr = ramdisk_image & PAGE_MASK;
 		p = early_memremap(mapaddr, clen+slop);
 		memcpy(q, p+slop, clen);
-		early_iounmap(p, clen+slop);
+		early_memunmap(p, clen+slop);
 		q += clen;
 		ramdisk_image += clen;
 		ramdisk_size  -= clen;
@@ -438,7 +438,7 @@ static void __init parse_setup_data(void)
 		data_len = data->len + sizeof(struct setup_data);
 		data_type = data->type;
 		pa_next = data->next;
-		early_iounmap(data, sizeof(*data));
+		early_memunmap(data, sizeof(*data));
 
 		switch (data_type) {
 		case SETUP_E820_EXT:
@@ -470,7 +470,7 @@ static void __init e820_reserve_setup_data(void)
 			 E820_RAM, E820_RESERVED_KERN);
 		found = 1;
 		pa_data = data->next;
-		early_iounmap(data, sizeof(*data));
+		early_memunmap(data, sizeof(*data));
 	}
 	if (!found)
 		return;
@@ -491,7 +491,7 @@ static void __init memblock_x86_reserve_range_setup_data(void)
 		data = early_memremap(pa_data, sizeof(*data));
 		memblock_reserve(pa_data, sizeof(*data) + data->len);
 		pa_data = data->next;
-		early_iounmap(data, sizeof(*data));
+		early_memunmap(data, sizeof(*data));
 	}
 }
 
diff --git a/arch/x86/kernel/sys_x86_64.c b/arch/x86/kernel/sys_x86_64.c
index 30277e27431a..10e0272d789a 100644
--- a/arch/x86/kernel/sys_x86_64.c
+++ b/arch/x86/kernel/sys_x86_64.c
@@ -34,10 +34,26 @@ static unsigned long get_align_mask(void)
 	return va_align.mask;
 }
 
+/*
+ * To avoid aliasing in the I$ on AMD F15h, the bits defined by the
+ * va_align.bits, [12:upper_bit), are set to a random value instead of
+ * zeroing them. This random value is computed once per boot. This form
+ * of ASLR is known as "per-boot ASLR".
+ *
+ * To achieve this, the random value is added to the info.align_offset
+ * value before calling vm_unmapped_area() or ORed directly to the
+ * address.
+ */
+static unsigned long get_align_bits(void)
+{
+	return va_align.bits & get_align_mask();
+}
+
 unsigned long align_vdso_addr(unsigned long addr)
 {
 	unsigned long align_mask = get_align_mask();
-	return (addr + align_mask) & ~align_mask;
+	addr = (addr + align_mask) & ~align_mask;
+	return addr | get_align_bits();
 }
 
 static int __init control_va_addr_alignment(char *str)
@@ -135,8 +151,12 @@ arch_get_unmapped_area(struct file *filp, unsigned long addr,
 	info.length = len;
 	info.low_limit = begin;
 	info.high_limit = end;
-	info.align_mask = filp ? get_align_mask() : 0;
+	info.align_mask = 0;
 	info.align_offset = pgoff << PAGE_SHIFT;
+	if (filp) {
+		info.align_mask = get_align_mask();
+		info.align_offset += get_align_bits();
+	}
 	return vm_unmapped_area(&info);
 }
 
@@ -174,8 +194,12 @@ arch_get_unmapped_area_topdown(struct file *filp, const unsigned long addr0,
 	info.length = len;
 	info.low_limit = PAGE_SIZE;
 	info.high_limit = mm->mmap_base;
-	info.align_mask = filp ? get_align_mask() : 0;
+	info.align_mask = 0;
 	info.align_offset = pgoff << PAGE_SHIFT;
+	if (filp) {
+		info.align_mask = get_align_mask();
+		info.align_offset += get_align_bits();
+	}
 	addr = vm_unmapped_area(&info);
 	if (!(addr & ~PAGE_MASK))
 		return addr;