Skip to contentSkip to conversation

EFI changes for v4.4

GIT PULL New Matt Fleming 2 participants 13 messages Oct 12, 2015 – Oct 14, 2015

Subsystems (7)
lore
Share & tools
lore
#0 lore raw

From: Matt Fleming <[email protected]>

Folks, please pull the below patches. The largest thing is probably the introduction of Taku's "efi_fake_mem" kernel option which allows the EFI memory map passed from the firmware to the kernel to be modified with additional memory map attributes.

There's also the ground work from Ard for supporting the EFI_PROPERTIES_TABLE feature, which allows us to use more strict page mapping attributes for the EFI runtime services regions (such as RO for code and NX for data). Actually doing that will come in the future.

Other than that, it's small changes and cleanups.

The following changes since commit 0ce423b6492a02be11662bfaa837dd16945aad3e:

efi: Use the generic efi.memmap instead of 'memmap' (2015-10-11 11:04:18 +0200)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/efi.git tags/efi-next

for you to fetch changes up to 0f96a99dab366333439e110d6ad253bc7c557c09:

efi: Add "efi_fake_mem" boot option (2015-10-12 14:20:09 +0100)

----------------------------------------------------------------
* Make the EFI System Resource Table (ESRT) driver explicitly non-modular by ripping out the module_* code since Kconfig doesn't allow it to be built as a module anyway - Paul Gortmaker

* Make the x86 efi=debug kernel parameter, which enables EFI debug code and output, generic and usable by arm64 - Leif Lindholm

* Add support to the x86 EFI boot stub for 64-bit Graphics Output Protocol frame buffer addresses - Matt Fleming

* Detect when the UEFI v2.5 EFI_PROPERTIES_TABLE feature is enabled in the firmware and set an efi.flags bit so the kernel knows when it can apply more strict runtime mapping attributes - Ard Biesheuvel

* Auto-load the efi-pstore module on EFI systems, just like we currently do for the efivars module - Ben Hutchings

* Add "efi_fake_mem" kernel parameter which allows the system's EFI memory map to be updated with additional attributes for specific memory ranges. This is useful for testing the kernel code that handles the EFI_MEMORY_MORE_RELIABLE memmap bit even if your firmware doesn't include support - Taku Izumi

----------------------------------------------------------------
Ard Biesheuvel (2):
efi: Add support for UEFIv2.5 Properties table
efi: Introduce EFI_NX_PE_DATA bit and set it from properties table

Ben Hutchings (1):
efi: Auto-load the efi-pstore module

Leif Lindholm (3):
efi/x86: Move efi=debug option parsing to core
arm64: Use core efi=debug instead of uefi_debug command line parameter efi/arm64: Clean up efi_get_fdt_params() interface

Matt Fleming (1):
efifb: Add support for 64-bit frame buffer addresses

Paul Gortmaker (1):
drivers/firmware: Make efi/esrt.c driver explicitly non-modular

Taku Izumi (3):
efi: Add EFI_MEMORY_MORE_RELIABLE support to efi_md_typeattr_format() x86/efi: Rename print_efi_memmap() to efi_print_memmap()
efi: Add "efi_fake_mem" boot option

 Documentation/arm/uefi.txt          |   2 -
 Documentation/kernel-parameters.txt |  15 +++
 arch/arm64/kernel/efi.c             |  19 +--
 arch/x86/boot/compressed/eboot.c    |  24 +++-
 arch/x86/include/asm/efi.h          |   1 +
 arch/x86/kernel/setup.c             |   4 +-
 arch/x86/platform/efi/efi.c         |   6 +-
 drivers/firmware/efi/Kconfig        |  22 ++++
 drivers/firmware/efi/Makefile       |   1 +
 drivers/firmware/efi/efi-pstore.c   |   1 +
 drivers/firmware/efi/efi.c          |  62 +++++++---
 drivers/firmware/efi/esrt.c         |  19 +--
 drivers/firmware/efi/fake_mem.c     | 238 ++++++++++++++++++++++++++++++++++++
 drivers/video/fbdev/efifb.c         |  24 +++-
 include/linux/efi.h                 |  22 +++-
 include/uapi/linux/screen_info.h    |   5 +-
 16 files changed, 400 insertions(+), 65 deletions(-)
 create mode 100644 drivers/firmware/efi/fake_mem.c
#1 lore raw

PATCH 1/11 drivers/firmware: Make efi/esrt.c driver explicitly non-modular

From: Paul Gortmaker <[email protected]>

The Kconfig for this driver is currently hidden with:

config EFI_ESRT
bool

...meaning that it currently is not being built as a module by anyone. Lets remove the modular code that is essentially orphaned, so that when reading the driver there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular case, the init ordering remains unchanged with this commit.

We leave some tags like MODULE_AUTHOR for documentation purposes.

We don't replace module.h with init.h since the file already has that.

---
 drivers/firmware/efi/esrt.c | 19 +++----------------
 1 file changed, 3 insertions(+), 16 deletions(-)
drivers/firmware/efi/esrt.c +3 -16
index a5b95d61ae71..22c5285f7705 100644
@@ -20,7 +20,6 @@
 #include <linux/kobject.h>
 #include <linux/list.h>
 #include <linux/memblock.h>
-#include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/types.h>
 
@@ -450,22 +449,10 @@ err:
 	esrt = NULL;
 	return error;
 }
+device_initcall(esrt_sysfs_init);
 
-static void __exit esrt_sysfs_exit(void)
-{
-	pr_debug("esrt-sysfs: unloading.\n");
-	cleanup_entry_list();
-	kset_unregister(esrt_kset);
-	sysfs_remove_group(esrt_kobj, &esrt_attr_group);
-	kfree(esrt);
-	esrt = NULL;
-	kobject_del(esrt_kobj);
-	kobject_put(esrt_kobj);
-}
-
-module_init(esrt_sysfs_init);
-module_exit(esrt_sysfs_exit);
-
+/*
 MODULE_AUTHOR("Peter Jones <[email protected]>");
 MODULE_DESCRIPTION("EFI System Resource Table support");
 MODULE_LICENSE("GPL");
+*/
Signature · 1 line
2.1.0
#2 lore raw

PATCH 2/11 efi/x86: Move efi=debug option parsing to core

In reply to Matt Fleming

From: Leif Lindholm <[email protected]>

fed6cefe3b6e ("x86/efi: Add a "debug" option to the efi= cmdline") adds the DBG flag, but does so for x86 only. Move this early param parsing to core code.

---
 arch/x86/platform/efi/efi.c | 2 --
 drivers/firmware/efi/efi.c  | 3 +++
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/x86/platform/efi/efi.c +0 -2
index 2f61fcddcddb..4d0ff0f27a99 100644
@@ -961,8 +961,6 @@ static int __init arch_parse_efi_cmdline(char *str)
 
 	if (parse_option_str(str, "old_map"))
 		set_bit(EFI_OLD_MEMMAP, &efi.flags);
-	if (parse_option_str(str, "debug"))
-		set_bit(EFI_DBG, &efi.flags);
 
 	return 0;
 }
drivers/firmware/efi/efi.c +3 -0
index 16c4928e36af..f0372a022c86 100644
@@ -63,6 +63,9 @@ static int __init parse_efi_cmdline(char *str)
 		return -EINVAL;
 	}
 
+	if (parse_option_str(str, "debug"))
+		set_bit(EFI_DBG, &efi.flags);
+
 	if (parse_option_str(str, "noruntime"))
 		disable_runtime = true;
 
Signature · 1 line
2.1.0
#3 lore raw

PATCH 3/11 arm64: Use core efi=debug instead of uefi_debug command line parameter

In reply to Matt Fleming

From: Leif Lindholm <[email protected]>

Now that we have an efi=debug command line option in the core code, use this instead of the arm64-specific uefi_debug option.

---
 Documentation/arm/uefi.txt |  2 --
 arch/arm64/kernel/efi.c    | 19 +++++--------------
 2 files changed, 5 insertions(+), 16 deletions(-)
Documentation/arm/uefi.txt +0 -2
index d60030a1b909..7b3fdfe0f7ba 100644
@@ -60,5 +60,3 @@ linux,uefi-mmap-desc-ver  | 32-bit | Version of the mmap descriptor format.
 --------------------------------------------------------------------------------
 linux,uefi-stub-kern-ver  | string | Copy of linux_banner from build.
 --------------------------------------------------------------------------------
-
-For verbose debug messages, specify 'uefi_debug' on the kernel command line.
arch/arm64/kernel/efi.c +5 -14
index e8ca6eaedd02..612ad5ec1d2e 100644
@@ -51,15 +51,6 @@ static struct mm_struct efi_mm = {
 	INIT_MM_CONTEXT(efi_mm)
 };
 
-static int uefi_debug __initdata;
-static int __init uefi_debug_setup(char *str)
-{
-	uefi_debug = 1;
-
-	return 0;
-}
-early_param("uefi_debug", uefi_debug_setup);
-
 static int __init is_normal_ram(efi_memory_desc_t *md)
 {
 	if (md->attribute & EFI_MEMORY_WB)
@@ -171,14 +162,14 @@ static __init void reserve_regions(void)
 	efi_memory_desc_t *md;
 	u64 paddr, npages, size;
 
-	if (uefi_debug)
+	if (efi_enabled(EFI_DBG))
 		pr_info("Processing EFI memory map:\n");
 
 	for_each_efi_memory_desc(&memmap, md) {
 		paddr = md->phys_addr;
 		npages = md->num_pages;
 
-		if (uefi_debug) {
+		if (efi_enabled(EFI_DBG)) {
 			char buf[64];
 
 			pr_info("  0x%012llx-0x%012llx %s",
@@ -194,11 +185,11 @@ static __init void reserve_regions(void)
 
 		if (is_reserve_region(md)) {
 			memblock_reserve(paddr, size);
-			if (uefi_debug)
+			if (efi_enabled(EFI_DBG))
 				pr_cont("*");
 		}
 
-		if (uefi_debug)
+		if (efi_enabled(EFI_DBG))
 			pr_cont("\n");
 	}
 
@@ -210,7 +201,7 @@ void __init efi_init(void)
 	struct efi_fdt_params params;
 
 	/* Grab UEFI information placed in FDT by stub */
-	if (!efi_get_fdt_params(&params, uefi_debug))
+	if (!efi_get_fdt_params(&params, efi_enabled(EFI_DBG)))
 		return;
 
 	efi_system_table = params.system_table;
Signature · 1 line
2.1.0
#4 lore raw

PATCH 4/11 efi/arm64: Clean up efi_get_fdt_params() interface

In reply to Matt Fleming

From: Leif Lindholm <[email protected]>

As we now have a common debug infrastructure between core and arm64 efi, drop the bit of the interface passing verbose output flags around.

---
 arch/arm64/kernel/efi.c    | 2 +-
 drivers/firmware/efi/efi.c | 6 ++----
 include/linux/efi.h        | 2 +-
 3 files changed, 4 insertions(+), 6 deletions(-)
arch/arm64/kernel/efi.c +1 -1
index 612ad5ec1d2e..ab5eeb63e2ca 100644
@@ -201,7 +201,7 @@ void __init efi_init(void)
 	struct efi_fdt_params params;
 
 	/* Grab UEFI information placed in FDT by stub */
-	if (!efi_get_fdt_params(&params, efi_enabled(EFI_DBG)))
+	if (!efi_get_fdt_params(&params))
 		return;
 
 	efi_system_table = params.system_table;
drivers/firmware/efi/efi.c +2 -4
index f0372a022c86..a0a0469e2869 100644
@@ -492,7 +492,6 @@ static __initdata struct {
 };
 
 struct param_info {
-	int verbose;
 	int found;
 	void *params;
 };
@@ -523,21 +522,20 @@ static int __init fdt_find_uefi_params(unsigned long node, const char *uname,
 		else
 			*(u64 *)dest = val;
 
-		if (info->verbose)
+		if (efi_enabled(EFI_DBG))
 			pr_info("  %s: 0x%0*llx\n", dt_params[i].name,
 				dt_params[i].size * 2, val);
 	}
 	return 1;
 }
 
-int __init efi_get_fdt_params(struct efi_fdt_params *params, int verbose)
+int __init efi_get_fdt_params(struct efi_fdt_params *params)
 {
 	struct param_info info;
 	int ret;
 
 	pr_info("Getting EFI parameters from FDT:\n");
 
-	info.verbose = verbose;
 	info.found = 0;
 	info.params = params;
 
include/linux/efi.h +1 -1
index 26ca9e2fd30e..4677d8a1bfd0 100644
@@ -902,7 +902,7 @@ extern void efi_initialize_iomem_resources(struct resource *code_resource,
 		struct resource *data_resource, struct resource *bss_resource);
 extern void efi_get_time(struct timespec *now);
 extern void efi_reserve_boot_services(void);
-extern int efi_get_fdt_params(struct efi_fdt_params *params, int verbose);
+extern int efi_get_fdt_params(struct efi_fdt_params *params);
 extern struct efi_memory_map memmap;
 extern struct kobject *efi_kobj;
 
Signature · 1 line
2.1.0
#5 lore raw

PATCH 5/11 efifb: Add support for 64-bit frame buffer addresses

In reply to Matt Fleming

From: Matt Fleming <[email protected]>

The EFI Graphics Output Protocol uses 64-bit frame buffer addresses but these get truncated to 32-bit by the EFI boot stub when storing the address in the 'lfb_base' field of 'struct screen_info'.

Add a 'ext_lfb_base' field for the upper 32-bits of the frame buffer address and set VIDEO_TYPE_CAPABILITY_64BIT_BASE when the field is useable.

It turns out that the reason no one has required this support so far is that there's actually code in tianocore to "downgrade" PCI resources that have option ROMs and 64-bit BARS from 64-bit to 32-bit to cope with legacy option ROMs that can't handle 64-bit addresses. The upshot is that basically all GOP devices in the wild use a 32-bit frame buffer address.

Still, it is possible to build firmware that uses a full 64-bit GOP frame buffer address. Chad did, which led to him reporting this issue.

Add support in anticipation of GOP devices using 64-bit addresses more widely, and so that efifb works out of the box when that happens.

---
 arch/x86/boot/compressed/eboot.c | 24 ++++++++++++++++++++----
 drivers/video/fbdev/efifb.c      | 24 +++++++++++++++++++++++-
 include/uapi/linux/screen_info.h |  5 +++--
 3 files changed, 46 insertions(+), 7 deletions(-)
arch/x86/boot/compressed/eboot.c +20 -4
index 7d69afd8b6fa..43473bb86ac0 100644
@@ -624,7 +624,7 @@ setup_pixel_info(struct screen_info *si, u32 pixels_per_scan_line,
 static efi_status_t
 __gop_query32(struct efi_graphics_output_protocol_32 *gop32,
 	      struct efi_graphics_output_mode_info **info,
-	      unsigned long *size, u32 *fb_base)
+	      unsigned long *size, u64 *fb_base)
 {
 	struct efi_graphics_output_protocol_mode_32 *mode;
 	efi_status_t status;
@@ -650,7 +650,8 @@ setup_gop32(struct screen_info *si, efi_guid_t *proto,
 	unsigned long nr_gops;
 	u16 width, height;
 	u32 pixels_per_scan_line;
-	u32 fb_base;
+	u32 ext_lfb_base;
+	u64 fb_base;
 	struct efi_pixel_bitmask pixel_info;
 	int pixel_format;
 	efi_status_t status;
@@ -713,6 +714,13 @@ setup_gop32(struct screen_info *si, efi_guid_t *proto,
 	si->lfb_width = width;
 	si->lfb_height = height;
 	si->lfb_base = fb_base;
+
+	ext_lfb_base = (u64)(unsigned long)fb_base >> 32;
+	if (ext_lfb_base) {
+		si->capabilities |= VIDEO_CAPABILITY_64BIT_BASE;
+		si->ext_lfb_base = ext_lfb_base;
+	}
+
 	si->pages = 1;
 
 	setup_pixel_info(si, pixels_per_scan_line, pixel_info, pixel_format);
@@ -727,7 +735,7 @@ out:
 static efi_status_t
 __gop_query64(struct efi_graphics_output_protocol_64 *gop64,
 	      struct efi_graphics_output_mode_info **info,
-	      unsigned long *size, u32 *fb_base)
+	      unsigned long *size, u64 *fb_base)
 {
 	struct efi_graphics_output_protocol_mode_64 *mode;
 	efi_status_t status;
@@ -753,7 +761,8 @@ setup_gop64(struct screen_info *si, efi_guid_t *proto,
 	unsigned long nr_gops;
 	u16 width, height;
 	u32 pixels_per_scan_line;
-	u32 fb_base;
+	u32 ext_lfb_base;
+	u64 fb_base;
 	struct efi_pixel_bitmask pixel_info;
 	int pixel_format;
 	efi_status_t status;
@@ -816,6 +825,13 @@ setup_gop64(struct screen_info *si, efi_guid_t *proto,
 	si->lfb_width = width;
 	si->lfb_height = height;
 	si->lfb_base = fb_base;
+
+	ext_lfb_base = (u64)(unsigned long)fb_base >> 32;
+	if (ext_lfb_base) {
+		si->capabilities |= VIDEO_CAPABILITY_64BIT_BASE;
+		si->ext_lfb_base = ext_lfb_base;
+	}
+
 	si->pages = 1;
 
 	setup_pixel_info(si, pixels_per_scan_line, pixel_info, pixel_format);
drivers/video/fbdev/efifb.c +23 -1
index 4bfff349b1fb..95d293b7445a 100644
@@ -114,6 +114,20 @@ static int efifb_setup(char *options)
 	return 0;
 }
 
+static inline bool fb_base_is_valid(void)
+{
+	if (screen_info.lfb_base)
+		return true;
+
+	if (!(screen_info.capabilities & VIDEO_CAPABILITY_64BIT_BASE))
+		return false;
+
+	if (screen_info.ext_lfb_base)
+		return true;
+
+	return false;
+}
+
 static int efifb_probe(struct platform_device *dev)
 {
 	struct fb_info *info;
@@ -141,7 +155,7 @@ static int efifb_probe(struct platform_device *dev)
 		screen_info.lfb_depth = 32;
 	if (!screen_info.pages)
 		screen_info.pages = 1;
-	if (!screen_info.lfb_base) {
+	if (!fb_base_is_valid()) {
 		printk(KERN_DEBUG "efifb: invalid framebuffer address\n");
 		return -ENODEV;
 	}
@@ -160,6 +174,14 @@ static int efifb_probe(struct platform_device *dev)
 	}
 
 	efifb_fix.smem_start = screen_info.lfb_base;
+
+	if (screen_info.capabilities & VIDEO_CAPABILITY_64BIT_BASE) {
+		u64 ext_lfb_base;
+
+		ext_lfb_base = (u64)(unsigned long)screen_info.ext_lfb_base << 32;
+		efifb_fix.smem_start |= ext_lfb_base;
+	}
+
 	efifb_defined.bits_per_pixel = screen_info.lfb_depth;
 	efifb_defined.xres = screen_info.lfb_width;
 	efifb_defined.yres = screen_info.lfb_height;
include/uapi/linux/screen_info.h +3 -2
index 7530e7447620..8b8d39dfb67f 100644
@@ -43,7 +43,8 @@ struct screen_info {
 	__u16 pages;		/* 0x32 */
 	__u16 vesa_attributes;	/* 0x34 */
 	__u32 capabilities;     /* 0x36 */
-	__u8  _reserved[6];	/* 0x3a */
+	__u32 ext_lfb_base;	/* 0x3a */
+	__u8  _reserved[2];	/* 0x3e */
 } __attribute__((packed));
 
 #define VIDEO_TYPE_MDA		0x10	/* Monochrome Text Display	*/
@@ -69,6 +70,6 @@ struct screen_info {
 #define VIDEO_FLAGS_NOCURSOR	(1 << 0) /* The video mode has no cursor set */
 
 #define VIDEO_CAPABILITY_SKIP_QUIRKS	(1 << 0)
-
+#define VIDEO_CAPABILITY_64BIT_BASE	(1 << 1)	/* Frame buffer base is 64-bit */
 
 #endif /* _UAPI_SCREEN_INFO_H */
Signature · 1 line
2.1.0
#6 lore raw

PATCH 6/11 efi: Add EFI_MEMORY_MORE_RELIABLE support to efi_md_typeattr_format()

In reply to Matt Fleming

From: Taku Izumi <[email protected]>

UEFI spec 2.5 introduces new Memory Attribute Definition named EFI_MEMORY_MORE_RELIABLE. This patch adds this new attribute support to efi_md_typeattr_format().

Cc Tony Luck <[email protected]>
Cc Ard Biesheuvel <[email protected]>
Cc Xishi Qiu <[email protected]>
Cc KAMEZAWA Hiroyuki <[email protected]>
---
 drivers/firmware/efi/efi.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/firmware/efi/efi.c +3 -2
index a0a0469e2869..278d4d88d250 100644
@@ -591,12 +591,13 @@ char * __init efi_md_typeattr_format(char *buf, size_t size,
 	if (attr & ~(EFI_MEMORY_UC | EFI_MEMORY_WC | EFI_MEMORY_WT |
 		     EFI_MEMORY_WB | EFI_MEMORY_UCE | EFI_MEMORY_RO |
 		     EFI_MEMORY_WP | EFI_MEMORY_RP | EFI_MEMORY_XP |
-		     EFI_MEMORY_RUNTIME))
+		     EFI_MEMORY_RUNTIME | EFI_MEMORY_MORE_RELIABLE))
 		snprintf(pos, size, "|attr=0x%016llx]",
 			 (unsigned long long)attr);
 	else
-		snprintf(pos, size, "|%3s|%2s|%2s|%2s|%2s|%3s|%2s|%2s|%2s|%2s]",
+		snprintf(pos, size, "|%3s|%2s|%2s|%2s|%2s|%2s|%3s|%2s|%2s|%2s|%2s]",
 			 attr & EFI_MEMORY_RUNTIME ? "RUN" : "",
+			 attr & EFI_MEMORY_MORE_RELIABLE ? "MR" : "",
 			 attr & EFI_MEMORY_XP      ? "XP"  : "",
 			 attr & EFI_MEMORY_RP      ? "RP"  : "",
 			 attr & EFI_MEMORY_WP      ? "WP"  : "",
Signature · 1 line
2.1.0
#7 lore raw

PATCH 7/11 efi: Add support for UEFIv2.5 Properties table

In reply to Matt Fleming

From: Ard Biesheuvel <[email protected]>

Version 2.5 of the UEFI spec introduces a new configuration table called the 'EFI Properties table'. Currently, it is only used to convey whether the Memory Protection feature is enabled, which splits PE/COFF images into separate code and data memory regions.

---
 drivers/firmware/efi/efi.c | 30 ++++++++++++++++--------------
 include/linux/efi.h        | 13 +++++++++++++
 2 files changed, 29 insertions(+), 14 deletions(-)
drivers/firmware/efi/efi.c +16 -14
index 278d4d88d250..c297d78f50fd 100644
@@ -26,20 +26,21 @@
 #include <linux/platform_device.h>
 
 struct efi __read_mostly efi = {
-	.mps        = EFI_INVALID_TABLE_ADDR,
-	.acpi       = EFI_INVALID_TABLE_ADDR,
-	.acpi20     = EFI_INVALID_TABLE_ADDR,
-	.smbios     = EFI_INVALID_TABLE_ADDR,
-	.smbios3    = EFI_INVALID_TABLE_ADDR,
-	.sal_systab = EFI_INVALID_TABLE_ADDR,
-	.boot_info  = EFI_INVALID_TABLE_ADDR,
-	.hcdp       = EFI_INVALID_TABLE_ADDR,
-	.uga        = EFI_INVALID_TABLE_ADDR,
-	.uv_systab  = EFI_INVALID_TABLE_ADDR,
-	.fw_vendor  = EFI_INVALID_TABLE_ADDR,
-	.runtime    = EFI_INVALID_TABLE_ADDR,
-	.config_table  = EFI_INVALID_TABLE_ADDR,
-	.esrt       = EFI_INVALID_TABLE_ADDR,
+	.mps			= EFI_INVALID_TABLE_ADDR,
+	.acpi			= EFI_INVALID_TABLE_ADDR,
+	.acpi20			= EFI_INVALID_TABLE_ADDR,
+	.smbios			= EFI_INVALID_TABLE_ADDR,
+	.smbios3		= EFI_INVALID_TABLE_ADDR,
+	.sal_systab		= EFI_INVALID_TABLE_ADDR,
+	.boot_info		= EFI_INVALID_TABLE_ADDR,
+	.hcdp			= EFI_INVALID_TABLE_ADDR,
+	.uga			= EFI_INVALID_TABLE_ADDR,
+	.uv_systab		= EFI_INVALID_TABLE_ADDR,
+	.fw_vendor		= EFI_INVALID_TABLE_ADDR,
+	.runtime		= EFI_INVALID_TABLE_ADDR,
+	.config_table		= EFI_INVALID_TABLE_ADDR,
+	.esrt			= EFI_INVALID_TABLE_ADDR,
+	.properties_table	= EFI_INVALID_TABLE_ADDR,
 };
 EXPORT_SYMBOL(efi);
 
@@ -365,6 +366,7 @@ static __initdata efi_config_table_type_t common_tables[] = {
 	{SMBIOS3_TABLE_GUID, "SMBIOS 3.0", &efi.smbios3},
 	{UGA_IO_PROTOCOL_GUID, "UGA", &efi.uga},
 	{EFI_SYSTEM_RESOURCE_TABLE_GUID, "ESRT", &efi.esrt},
+	{EFI_PROPERTIES_TABLE_GUID, "PROP", &efi.properties_table},
 	{NULL_GUID, NULL, NULL},
 };
 
include/linux/efi.h +13 -0
index 4677d8a1bfd0..d6a9bee755f2 100644
@@ -596,6 +596,9 @@ void efi_native_runtime_setup(void);
 #define DEVICE_TREE_GUID \
     EFI_GUID(  0xb1b621d5, 0xf19c, 0x41a5, 0x83, 0x0b, 0xd9, 0x15, 0x2c, 0x69, 0xaa, 0xe0 )
 
+#define EFI_PROPERTIES_TABLE_GUID \
+    EFI_GUID(  0x880aaca3, 0x4adc, 0x4a04, 0x90, 0x79, 0xb7, 0x47, 0x34, 0x08, 0x25, 0xe5 )
+
 typedef struct {
 	efi_guid_t guid;
 	u64 table;
@@ -809,6 +812,15 @@ typedef struct _efi_file_io_interface {
 #define EFI_FILE_MODE_WRITE	0x0000000000000002
 #define EFI_FILE_MODE_CREATE	0x8000000000000000
 
+typedef struct {
+	u32 version;
+	u32 length;
+	u64 memory_protection_attribute;
+} efi_properties_table_t;
+
+#define EFI_PROPERTIES_TABLE_VERSION	0x00010000
+#define EFI_PROPERTIES_RUNTIME_MEMORY_PROTECTION_NON_EXECUTABLE_PE_DATA	0x1
+
 #define EFI_INVALID_TABLE_ADDR		(~0UL)
 
 /*
@@ -831,6 +843,7 @@ extern struct efi {
 	unsigned long runtime;		/* runtime table */
 	unsigned long config_table;	/* config tables */
 	unsigned long esrt;		/* ESRT table */
+	unsigned long properties_table;	/* properties table */
 	efi_get_time_t *get_time;
 	efi_set_time_t *set_time;
 	efi_get_wakeup_time_t *get_wakeup_time;
Signature · 1 line
2.1.0
#8 lore raw

PATCH 8/11 efi: Introduce EFI_NX_PE_DATA bit and set it from properties table

In reply to Matt Fleming

From: Ard Biesheuvel <[email protected]>

UEFI v2.5 introduces a runtime memory protection feature that splits PE/COFF runtime images into separate code and data regions. Since this may require special handling by the OS, allocate a EFI_xxx bit to keep track of whether this feature is currently active or not.

---
 drivers/firmware/efi/efi.c | 18 ++++++++++++++++++
 include/linux/efi.h        |  1 +
 2 files changed, 19 insertions(+)
drivers/firmware/efi/efi.c +18 -0
index c297d78f50fd..31fc864eb037 100644
@@ -426,6 +426,24 @@ int __init efi_config_parse_tables(void *config_tables, int count, int sz,
 	}
 	pr_cont("\n");
 	set_bit(EFI_CONFIG_TABLES, &efi.flags);
+
+	/* Parse the EFI Properties table if it exists */
+	if (efi.properties_table != EFI_INVALID_TABLE_ADDR) {
+		efi_properties_table_t *tbl;
+
+		tbl = early_memremap(efi.properties_table, sizeof(*tbl));
+		if (tbl == NULL) {
+			pr_err("Could not map Properties table!\n");
+			return -ENOMEM;
+		}
+
+		if (tbl->memory_protection_attribute &
+		    EFI_PROPERTIES_RUNTIME_MEMORY_PROTECTION_NON_EXECUTABLE_PE_DATA)
+			set_bit(EFI_NX_PE_DATA, &efi.flags);
+
+		early_memunmap(tbl, sizeof(*tbl));
+	}
+
 	return 0;
 }
 
include/linux/efi.h +1 -0
index d6a9bee755f2..fa5106c2f9f5 100644
@@ -973,6 +973,7 @@ extern int __init efi_setup_pcdp_console(char *);
 #define EFI_PARAVIRT		6	/* Access is via a paravirt interface */
 #define EFI_ARCH_1		7	/* First arch-specific bit */
 #define EFI_DBG			8	/* Print additional debug info at runtime */
+#define EFI_NX_PE_DATA		9	/* Can runtime data regions be mapped non-executable? */
 
 #ifdef CONFIG_EFI
 /*
Signature · 1 line
2.1.0
#9 lore raw

PATCH 9/11 efi: Auto-load the efi-pstore module

In reply to Matt Fleming

From: Ben Hutchings <[email protected]>

efi-pstore should be auto-loaded on EFI systems, same as efivars.

Cc Matthew Garrett <[email protected]>
Cc "Lee, Chun-Yi" <[email protected]>
Cc Ard Biesheuvel <[email protected]>
---
 drivers/firmware/efi/efi-pstore.c | 1 +
 1 file changed, 1 insertion(+)
drivers/firmware/efi/efi-pstore.c +1 -0
index e992abc5ef26..c8d794c58479 100644
@@ -400,3 +400,4 @@ module_exit(efivars_pstore_exit);
 
 MODULE_DESCRIPTION("EFI variable backend for pstore");
 MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:efivars");
Signature · 1 line
2.1.0
#10 lore raw

PATCH 10/11 x86/efi: Rename print_efi_memmap() to efi_print_memmap()

In reply to Matt Fleming

From: Taku Izumi <[email protected]>

This patch renames print_efi_memmap() to efi_print_memmap() and make it global function so that we can invoke it outside of arch/x86/platform/efi/efi.c

Cc Tony Luck <[email protected]>
Cc Xishi Qiu <[email protected]>
Cc Kamezawa Hiroyuki <[email protected]>
Cc Ard Biesheuvel <[email protected]>
---
 arch/x86/include/asm/efi.h  | 1 +
 arch/x86/platform/efi/efi.c | 4 ++--
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/x86/include/asm/efi.h +1 -0
index 155162ea0e00..cfee9d4b02af 100644
@@ -93,6 +93,7 @@ extern void __init efi_set_executable(efi_memory_desc_t *md, bool executable);
 extern int __init efi_memblock_x86_reserve_range(void);
 extern pgd_t * __init efi_call_phys_prolog(void);
 extern void __init efi_call_phys_epilog(pgd_t *save_pgd);
+extern void __init efi_print_memmap(void);
 extern void __init efi_unmap_memmap(void);
 extern void __init efi_memory_uc(u64 addr, unsigned long size);
 extern void __init efi_map_region(efi_memory_desc_t *md);
arch/x86/platform/efi/efi.c +2 -2
index 4d0ff0f27a99..c69e58fb7f19 100644
@@ -222,7 +222,7 @@ int __init efi_memblock_x86_reserve_range(void)
 	return 0;
 }
 
-static void __init print_efi_memmap(void)
+void __init efi_print_memmap(void)
 {
 #ifdef EFI_DEBUG
 	efi_memory_desc_t *md;
@@ -524,7 +524,7 @@ void __init efi_init(void)
 		return;
 
 	if (efi_enabled(EFI_DBG))
-		print_efi_memmap();
+		efi_print_memmap();
 
 	efi_esrt_init();
 }
Signature · 1 line
2.1.0
#11 lore raw

PATCH 11/11 efi: Add "efi_fake_mem" boot option

In reply to Matt Fleming

From: Taku Izumi <[email protected]>

This patch introduces new boot option named "efi_fake_mem". By specifying this parameter, you can add arbitrary attribute to specific memory range.
This is useful for debugging of Address Range Mirroring feature.

For example, if "efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000" is specified, the original (firmware provided) EFI memmap will be updated so that the specified memory regions have EFI_MEMORY_MORE_RELIABLE attribute (0x10000):

<original>
efi: mem36: [Conventional Memory| | | | | | |WB|WT|WC|UC] range=[0x0000000100000000-0x00000020a0000000) (129536MB)

<updated>

   efi: mem36: [Conventional Memory|  |MR|  |  |  |   |WB|WT|WC|UC] range=[0x0000000100000000-0x0000000180000000) (2048MB)
   efi: mem37: [Conventional Memory|  |  |  |  |  |   |WB|WT|WC|UC] range=[0x0000000180000000-0x00000010a0000000) (61952MB)
   efi: mem38: [Conventional Memory|  |MR|  |  |  |   |WB|WT|WC|UC] range=[0x00000010a0000000-0x0000001120000000) (2048MB)
   efi: mem39: [Conventional Memory|  |  |  |  |  |   |WB|WT|WC|UC] range=[0x0000001120000000-0x00000020a0000000) (63488MB)

And you will find that the following message is output:

efi: Memory: 4096M/131455M mirrored memory

Cc Tony Luck <[email protected]>
Cc Xishi Qiu <[email protected]>
Cc Kamezawa Hiroyuki <[email protected]>
Cc Ard Biesheuvel <[email protected]>
---
 Documentation/kernel-parameters.txt |  15 +++
 arch/x86/kernel/setup.c             |   4 +-
 drivers/firmware/efi/Kconfig        |  22 ++++
 drivers/firmware/efi/Makefile       |   1 +
 drivers/firmware/efi/fake_mem.c     | 238 ++++++++++++++++++++++++++++++++++++
 include/linux/efi.h                 |   6 +
 6 files changed, 285 insertions(+), 1 deletion(-)
 create mode 100644 drivers/firmware/efi/fake_mem.c
Documentation/kernel-parameters.txt +15 -0
index 1d6f0459cd7b..cd5312f24981 100644
@@ -1092,6 +1092,21 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 			you are really sure that your UEFI does sane gc and
 			fulfills the spec otherwise your board may brick.
 
+	efi_fake_mem=	nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
+			Add arbitrary attribute to specific memory range by
+			updating original EFI memory map.
+			Region of memory which aa attribute is added to is
+			from ss to ss+nn.
+			If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
+			is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
+			attribute is added to range 0x100000000-0x180000000 and
+			0x10a0000000-0x1120000000.
+
+			Using this parameter you can do debugging of EFI memmap
+			related feature. For example, you can do debugging of
+			Address Range Mirroring feature even if your box
+			doesn't support it.
+
 	eisa_irq_edge=	[PARISC,HW]
 			See header of drivers/parisc/eisa.c.
 
arch/x86/kernel/setup.c +3 -1
index 80f874bf999e..e3ed628f7db4 100644
@@ -1104,8 +1104,10 @@ void __init setup_arch(char **cmdline_p)
 	memblock_set_current_limit(ISA_END_ADDRESS);
 	memblock_x86_fill();
 
-	if (efi_enabled(EFI_BOOT))
+	if (efi_enabled(EFI_BOOT)) {
+		efi_fake_memmap();
 		efi_find_mirror();
+	}
 
 	/*
 	 * The EFI specification says that boot service code won't be called
drivers/firmware/efi/Kconfig +22 -0
index 54071c148340..1de6f0ed5077 100644
@@ -52,6 +52,28 @@ config EFI_RUNTIME_MAP
 
 	  See also Documentation/ABI/testing/sysfs-firmware-efi-runtime-map.
 
+config EFI_FAKE_MEMMAP
+	bool "Enable EFI fake memory map"
+	depends on EFI && X86
+	default n
+	help
+	  Saying Y here will enable "efi_fake_mem" boot option.
+	  By specifying this parameter, you can add arbitrary attribute
+	  to specific memory range by updating original (firmware provided)
+	  EFI memmap.
+	  This is useful for debugging of EFI memmap related feature.
+	  e.g. Address Range Mirroring feature.
+
+config EFI_MAX_FAKE_MEM
+	int "maximum allowable number of ranges in efi_fake_mem boot option"
+	depends on EFI_FAKE_MEMMAP
+	range 1 128
+	default 8
+	help
+	  Maximum allowable number of ranges in efi_fake_mem boot option.
+	  Ranges can be set up to this value using comma-separated list.
+	  The default value is 8.
+
 config EFI_PARAMS_FROM_FDT
 	bool
 	help
drivers/firmware/efi/Makefile +1 -0
index 6fd3da938717..c24f00569acb 100644
@@ -9,3 +9,4 @@ obj-$(CONFIG_UEFI_CPER)			+= cper.o
 obj-$(CONFIG_EFI_RUNTIME_MAP)		+= runtime-map.o
 obj-$(CONFIG_EFI_RUNTIME_WRAPPERS)	+= runtime-wrappers.o
 obj-$(CONFIG_EFI_STUB)			+= libstub/
+obj-$(CONFIG_EFI_FAKE_MEMMAP)		+= fake_mem.o
drivers/firmware/efi/fake_mem.c new file +238 -0

238 lines not on this page · raw message

include/linux/efi.h +6 -0
index fa5106c2f9f5..4d01c1033fce 100644
@@ -922,6 +922,12 @@ extern struct kobject *efi_kobj;
 extern int efi_reboot_quirk_mode;
 extern bool efi_poweroff_required(void);
 
+#ifdef CONFIG_EFI_FAKE_MEMMAP
+extern void __init efi_fake_memmap(void);
+#else
+static inline void efi_fake_memmap(void) { }
+#endif
+
 /* Iterate through an efi_memory_map */
 #define for_each_efi_memory_desc(m, md)					   \
 	for ((md) = (m)->map;						   \
Signature · 1 line
2.1.0
#12 lore raw
In reply to Matt Fleming
Quoted 85 lines · Matt Fleming

From: Matt Fleming <[email protected]>

Folks, please pull the below patches. The largest thing is probably the introduction of Taku's "efi_fake_mem" kernel option which allows the EFI memory map passed from the firmware to the kernel to be modified with additional memory map attributes.

There's also the ground work from Ard for supporting the EFI_PROPERTIES_TABLE feature, which allows us to use more strict page mapping attributes for the EFI runtime services regions (such as RO for code and NX for data). Actually doing that will come in the future.

Other than that, it's small changes and cleanups.

The following changes since commit 0ce423b6492a02be11662bfaa837dd16945aad3e:

efi: Use the generic efi.memmap instead of 'memmap' (2015-10-11 11:04:18 +0200)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/efi.git tags/efi-next

for you to fetch changes up to 0f96a99dab366333439e110d6ad253bc7c557c09:

efi: Add "efi_fake_mem" boot option (2015-10-12 14:20:09 +0100)

----------------------------------------------------------------
* Make the EFI System Resource Table (ESRT) driver explicitly non-modular by ripping out the module_* code since Kconfig doesn't allow it to be built as a module anyway - Paul Gortmaker

* Make the x86 efi=debug kernel parameter, which enables EFI debug code and output, generic and usable by arm64 - Leif Lindholm

* Add support to the x86 EFI boot stub for 64-bit Graphics Output Protocol frame buffer addresses - Matt Fleming

* Detect when the UEFI v2.5 EFI_PROPERTIES_TABLE feature is enabled in the firmware and set an efi.flags bit so the kernel knows when it can apply more strict runtime mapping attributes - Ard Biesheuvel

* Auto-load the efi-pstore module on EFI systems, just like we currently do for the efivars module - Ben Hutchings

* Add "efi_fake_mem" kernel parameter which allows the system's EFI memory map to be updated with additional attributes for specific memory ranges. This is useful for testing the kernel code that handles the EFI_MEMORY_MORE_RELIABLE memmap bit even if your firmware doesn't include support - Taku Izumi

----------------------------------------------------------------
Ard Biesheuvel (2):
efi: Add support for UEFIv2.5 Properties table
efi: Introduce EFI_NX_PE_DATA bit and set it from properties table

Ben Hutchings (1):
efi: Auto-load the efi-pstore module

Leif Lindholm (3):
efi/x86: Move efi=debug option parsing to core
arm64: Use core efi=debug instead of uefi_debug command line parameter efi/arm64: Clean up efi_get_fdt_params() interface

Matt Fleming (1):
efifb: Add support for 64-bit frame buffer addresses

Paul Gortmaker (1):
drivers/firmware: Make efi/esrt.c driver explicitly non-modular

Taku Izumi (3):
efi: Add EFI_MEMORY_MORE_RELIABLE support to efi_md_typeattr_format() x86/efi: Rename print_efi_memmap() to efi_print_memmap()
efi: Add "efi_fake_mem" boot option

 Documentation/arm/uefi.txt          |   2 -
 Documentation/kernel-parameters.txt |  15 +++
 arch/arm64/kernel/efi.c             |  19 +--
 arch/x86/boot/compressed/eboot.c    |  24 +++-
 arch/x86/include/asm/efi.h          |   1 +
 arch/x86/kernel/setup.c             |   4 +-
 arch/x86/platform/efi/efi.c         |   6 +-
 drivers/firmware/efi/Kconfig        |  22 ++++
 drivers/firmware/efi/Makefile       |   1 +
 drivers/firmware/efi/efi-pstore.c   |   1 +
Quoted 8 lines
 drivers/firmware/efi/efi.c          |  62 +++++++---
 drivers/firmware/efi/esrt.c         |  19 +--
 drivers/firmware/efi/fake_mem.c     | 238 ++++++++++++++++++++++++++++++++++++
 drivers/video/fbdev/efifb.c         |  24 +++-
 include/linux/efi.h                 |  22 +++-
 include/uapi/linux/screen_info.h    |   5 +-
 16 files changed, 400 insertions(+), 65 deletions(-)
 create mode 100644 drivers/firmware/efi/fake_mem.c

Pulled, thanks Matt!

Ingo