[OS-BUILD PATCHv2 0/4] efi: x86: Set the NX-compatibility flag in the PE header

[Justin M. Forbes (via Email Bridge)]

From: Justin M. Forbes on gitlab.com
Merge Request: https://gitlab.com/cki-project/kernel-ark/-/merge_requests/1771

Following Baskov Evgeniy's "Handle UEFI NX-restricted page tables"
patches, it's safe to set this compatibility flag to let loaders know
they don't need to make special accommodations for kernel to load if
pre-boot NX is enabled.

This MR includes the "Handle UEFI NX-restricted page tables" patches as well
as the patch to actually set the flag in the header. They are queued for the
5.19 merge window [1], but will be needed in 5.18 for a shim build.


[1] https://git.kernel.org/pub/scm/linux/kernel/git/efi/efi.git/log/

---
 arch/x86/boot/header.S                                          |    4 +
 arch/x86/include/asm/efi.h                                      |    5 +
 drivers/firmware/efi/libstub/efistub.h                          |   74 ++++++
 drivers/firmware/efi/libstub/x86-stub.c                         |  119 +++++++++-
 drivers/firmware/efi/Kconfig                                    |   12 +
 include/linux/efi.h                                             |    2 +
 redhat/configs/common/generic/x86/CONFIG_EFI_DXE_MEM_ATTRIBUTES |    1 +
 7 files changed, 212 insertions(+), 5 deletions(-)

[OS-BUILD PATCHv2 1/4] efi: libstub: declare DXE services table

[Justin M. Forbes (via Email Bridge)]

From: Baskov Evgeniy <baskov(a)ispras.ru&gt;

efi: libstub: declare DXE services table

UEFI DXE services are not yet used in kernel code
but are required to manipulate page table memory
protection flags.

Add required declarations to use DXE services functions.

Signed-off-by: Baskov Evgeniy <baskov(a)ispras.ru&gt;
Link: https://lore.kernel.org/r/[e-mail]
[ardb: ignore absent DXE table but warn if the signature check fails]
Signed-off-by: Ard Biesheuvel <ardb(a)kernel.org&gt;

diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h
index blahblah..blahblah 100644
--- a/arch/x86/include/asm/efi.h
+++ b/arch/x86/include/asm/efi.h
@@ -357,6 +357,11 @@ static inline u32 efi64_convert_status(efi_status_t status)
 						   runtime),		\
 				    func, __VA_ARGS__))
 
+#define efi_dxe_call(func, ...)						\
+	(efi_is_native()						\
+		? efi_dxe_table->func(__VA_ARGS__)			\
+		: __efi64_thunk_map(efi_dxe_table, func, __VA_ARGS__))
+
 #else /* CONFIG_EFI_MIXED */
 
 static inline bool efi_is_64bit(void)
diff --git a/drivers/firmware/efi/libstub/efistub.h
b/drivers/firmware/efi/libstub/efistub.h
index blahblah..blahblah 100644
--- a/drivers/firmware/efi/libstub/efistub.h
+++ b/drivers/firmware/efi/libstub/efistub.h
@@ -36,6 +36,9 @@ extern bool efi_novamap;
 
 extern const efi_system_table_t *efi_system_table;
 
+typedef union efi_dxe_services_table efi_dxe_services_table_t;
+extern const efi_dxe_services_table_t *efi_dxe_table;
+
 efi_status_t __efiapi efi_pe_entry(efi_handle_t handle,
 				   efi_system_table_t *sys_table_arg);
 
@@ -44,6 +47,7 @@ efi_status_t __efiapi efi_pe_entry(efi_handle_t handle,
 #define efi_is_native()		(true)
 #define efi_bs_call(func, ...)	efi_system_table->boottime->func(__VA_ARGS__)
 #define efi_rt_call(func, ...)	efi_system_table->runtime->func(__VA_ARGS__)
+#define efi_dxe_call(func, ...)	efi_dxe_table->func(__VA_ARGS__)
 #define efi_table_attr(inst, attr)	(inst->attr)
 #define efi_call_proto(inst, func, ...) inst->func(inst, ##__VA_ARGS__)
 
@@ -329,6 +333,76 @@ union efi_boot_services {
 	} mixed_mode;
 };
 
+typedef enum {
+	EfiGcdMemoryTypeNonExistent,
+	EfiGcdMemoryTypeReserved,
+	EfiGcdMemoryTypeSystemMemory,
+	EfiGcdMemoryTypeMemoryMappedIo,
+	EfiGcdMemoryTypePersistent,
+	EfiGcdMemoryTypeMoreReliable,
+	EfiGcdMemoryTypeMaximum
+} efi_gcd_memory_type_t;
+
+typedef struct {
+	efi_physical_addr_t base_address;
+	u64 length;
+	u64 capabilities;
+	u64 attributes;
+	efi_gcd_memory_type_t gcd_memory_type;
+	void *image_handle;
+	void *device_handle;
+} efi_gcd_memory_space_desc_t;
+
+/*
+ * EFI DXE Services table
+ */
+union efi_dxe_services_table {
+	struct {
+		efi_table_hdr_t hdr;
+		void *add_memory_space;
+		void *allocate_memory_space;
+		void *free_memory_space;
+		void *remove_memory_space;
+		efi_status_t (__efiapi *get_memory_space_descriptor)(efi_physical_addr_t,
+								     efi_gcd_memory_space_desc_t *);
+		efi_status_t (__efiapi *set_memory_space_attributes)(efi_physical_addr_t,
+								     u64, u64);
+		void *get_memory_space_map;
+		void *add_io_space;
+		void *allocate_io_space;
+		void *free_io_space;
+		void *remove_io_space;
+		void *get_io_space_descriptor;
+		void *get_io_space_map;
+		void *dispatch;
+		void *schedule;
+		void *trust;
+		void *process_firmware_volume;
+		void *set_memory_space_capabilities;
+	};
+	struct {
+		efi_table_hdr_t hdr;
+		u32 add_memory_space;
+		u32 allocate_memory_space;
+		u32 free_memory_space;
+		u32 remove_memory_space;
+		u32 get_memory_space_descriptor;
+		u32 set_memory_space_attributes;
+		u32 get_memory_space_map;
+		u32 add_io_space;
+		u32 allocate_io_space;
+		u32 free_io_space;
+		u32 remove_io_space;
+		u32 get_io_space_descriptor;
+		u32 get_io_space_map;
+		u32 dispatch;
+		u32 schedule;
+		u32 trust;
+		u32 process_firmware_volume;
+		u32 set_memory_space_capabilities;
+	} mixed_mode;
+};
+
 typedef union efi_uga_draw_protocol efi_uga_draw_protocol_t;
 
 union efi_uga_draw_protocol {
diff --git a/drivers/firmware/efi/libstub/x86-stub.c
b/drivers/firmware/efi/libstub/x86-stub.c
index blahblah..blahblah 100644
--- a/drivers/firmware/efi/libstub/x86-stub.c
+++ b/drivers/firmware/efi/libstub/x86-stub.c
@@ -22,6 +22,7 @@
 #define MAXMEM_X86_64_4LEVEL (1ull << 46)
 
 const efi_system_table_t *efi_system_table;
+const efi_dxe_services_table_t *efi_dxe_table;
 extern u32 image_offset;
 static efi_loaded_image_t *image = NULL;
 
@@ -677,11 +678,17 @@ unsigned long efi_main(efi_handle_t handle,
 	efi_status_t status;
 
 	efi_system_table = sys_table_arg;
-
 	/* Check if we were booted by the EFI firmware */
 	if (efi_system_table->hdr.signature != EFI_SYSTEM_TABLE_SIGNATURE)
 		efi_exit(handle, EFI_INVALID_PARAMETER);
 
+	efi_dxe_table = get_efi_config_table(EFI_DXE_SERVICES_TABLE_GUID);
+	if (efi_dxe_table &&
+	    efi_dxe_table->hdr.signature != EFI_DXE_SERVICES_TABLE_SIGNATURE) {
+		efi_warn("Ignoring DXE services table: invalid signature\n");
+		efi_dxe_table = NULL;
+	}
+
 	/*
 	 * If the kernel isn't already loaded at a suitable address,
 	 * relocate it.
diff --git a/include/linux/efi.h b/include/linux/efi.h
index blahblah..blahblah 100644
--- a/include/linux/efi.h
+++ b/include/linux/efi.h
@@ -385,6 +385,7 @@ void efi_native_runtime_setup(void);
 #define EFI_LOAD_FILE_PROTOCOL_GUID		EFI_GUID(0x56ec3091, 0x954c, 0x11d2,  0x8e, 0x3f,
0x00, 0xa0, 0xc9, 0x69, 0x72, 0x3b)
 #define EFI_LOAD_FILE2_PROTOCOL_GUID		EFI_GUID(0x4006c0c1, 0xfcb3, 0x403e,  0x99, 0x6d,
0x4a, 0x6c, 0x87, 0x24, 0xe0, 0x6d)
 #define EFI_RT_PROPERTIES_TABLE_GUID		EFI_GUID(0xeb66918a, 0x7eef, 0x402a,  0x84, 0x2e,
0x93, 0x1d, 0x21, 0xc3, 0x8a, 0xe9)
+#define EFI_DXE_SERVICES_TABLE_GUID		EFI_GUID(0x05ad34ba, 0x6f02, 0x4214,  0x95, 0x2e,
0x4d, 0xa0, 0x39, 0x8e, 0x2b, 0xb9)
 
 #define EFI_IMAGE_SECURITY_DATABASE_GUID	EFI_GUID(0xd719b2cb, 0x3d3a, 0x4596,  0xa3,
0xbc, 0xda, 0xd0, 0x0e, 0x67, 0x65, 0x6f)
 #define EFI_SHIM_LOCK_GUID			EFI_GUID(0x605dab50, 0xe046, 0x4300,  0xab, 0xb6, 0x3d,
0xd8, 0x10, 0xdd, 0x8b, 0x23)
@@ -437,6 +438,7 @@ typedef struct {
 } efi_config_table_type_t;
 
 #define EFI_SYSTEM_TABLE_SIGNATURE ((u64)0x5453595320494249ULL)
+#define EFI_DXE_SERVICES_TABLE_SIGNATURE ((u64)0x565245535f455844ULL)
 
 #define EFI_2_30_SYSTEM_TABLE_REVISION  ((2 << 16) | (30))
 #define EFI_2_20_SYSTEM_TABLE_REVISION  ((2 << 16) | (20))

--
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/1771

[OS-BUILD PATCHv2 2/4] efi: libstub: ensure allocated memory to be executable

[Justin M. Forbes (via Email Bridge)]

From: Baskov Evgeniy <baskov(a)ispras.ru&gt;

efi: libstub: ensure allocated memory to be executable

There are UEFI versions that restrict execution of memory regions,
preventing the kernel from booting. Parts that needs to be executable
are:

* Area used for trampoline placement.
* All memory regions that the kernel may be relocated before
  and during extraction.

Use DXE services to ensure aforementioned address ranges
to be executable. Only modify attributes that does not
have appropriate attributes.

Signed-off-by: Baskov Evgeniy <baskov(a)ispras.ru&gt;
Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Ard Biesheuvel <ardb(a)kernel.org&gt;

diff --git a/drivers/firmware/efi/Kconfig b/drivers/firmware/efi/Kconfig
index blahblah..blahblah 100644
--- a/drivers/firmware/efi/Kconfig
+++ b/drivers/firmware/efi/Kconfig
@@ -91,6 +91,18 @@ config EFI_SOFT_RESERVE
 
 	  If unsure, say Y.
 
+config EFI_DXE_MEM_ATTRIBUTES
+	bool "Adjust memory attributes in EFISTUB"
+	depends on EFI && EFI_STUB && X86
+	default y
+	help
+	  UEFI specification does not guarantee all memory to be
+	  accessible for both write and execute as the kernel expects
+	  it to be.
+	  Use DXE services to check and alter memory protection
+	  attributes during boot via EFISTUB to ensure that memory
+	  ranges used by the kernel are writable and executable.
+
 config EFI_PARAMS_FROM_FDT
 	bool
 	help
diff --git a/drivers/firmware/efi/libstub/x86-stub.c
b/drivers/firmware/efi/libstub/x86-stub.c
index blahblah..blahblah 100644
--- a/drivers/firmware/efi/libstub/x86-stub.c
+++ b/drivers/firmware/efi/libstub/x86-stub.c
@@ -212,9 +212,110 @@ static void retrieve_apple_device_properties(struct boot_params
*boot_params)
 	}
 }
 
+static void
+adjust_memory_range_protection(unsigned long start, unsigned long size)
+{
+	efi_status_t status;
+	efi_gcd_memory_space_desc_t desc;
+	unsigned long end, next;
+	unsigned long rounded_start, rounded_end;
+	unsigned long unprotect_start, unprotect_size;
+	int has_system_memory = 0;
+
+	if (efi_dxe_table == NULL)
+		return;
+
+	rounded_start = rounddown(start, EFI_PAGE_SIZE);
+	rounded_end = roundup(start + size, EFI_PAGE_SIZE);
+
+	/*
+	 * Don't modify memory region attributes, they are
+	 * already suitable, to lower the possibility to
+	 * encounter firmware bugs.
+	 */
+
+	for (end = start + size; start < end; start = next) {
+
+		status = efi_dxe_call(get_memory_space_descriptor, start, &desc);
+
+		if (status != EFI_SUCCESS)
+			return;
+
+		next = desc.base_address + desc.length;
+
+		/*
+		 * Only system memory is suitable for trampoline/kernel image placement,
+		 * so only this type of memory needs its attributes to be modified.
+		 */
+
+		if (desc.gcd_memory_type != EfiGcdMemoryTypeSystemMemory ||
+		    (desc.attributes & (EFI_MEMORY_RO | EFI_MEMORY_XP)) == 0)
+			continue;
+
+		unprotect_start = max(rounded_start, (unsigned long)desc.base_address);
+		unprotect_size = min(rounded_end, next) - unprotect_start;
+
+		status = efi_dxe_call(set_memory_space_attributes,
+				      unprotect_start, unprotect_size,
+				      EFI_MEMORY_WB);
+
+		if (status != EFI_SUCCESS) {
+			efi_warn("Unable to unprotect memory range [%08lx,%08lx]: %d\n",
+				 unprotect_start,
+				 unprotect_start + unprotect_size,
+				 (int)status);
+		}
+	}
+}
+
+/*
+ * Trampoline takes 2 pages and can be loaded in first megabyte of memory
+ * with its end placed between 128k and 640k where BIOS might start.
+ * (see arch/x86/boot/compressed/pgtable_64.c)
+ *
+ * We cannot find exact trampoline placement since memory map
+ * can be modified by UEFI, and it can alter the computed address.
+ */
+
+#define TRAMPOLINE_PLACEMENT_BASE ((128 - 8)*1024)
+#define TRAMPOLINE_PLACEMENT_SIZE (640*1024 - (128 - 8)*1024)
+
+void startup_32(struct boot_params *boot_params);
+
+static void
+setup_memory_protection(unsigned long image_base, unsigned long image_size)
+{
+	/*
+	 * Allow execution of possible trampoline used
+	 * for switching between 4- and 5-level page tables
+	 * and relocated kernel image.
+	 */
+
+	adjust_memory_range_protection(TRAMPOLINE_PLACEMENT_BASE,
+				       TRAMPOLINE_PLACEMENT_SIZE);
+
+#ifdef CONFIG_64BIT
+	if (image_base != (unsigned long)startup_32)
+		adjust_memory_range_protection(image_base, image_size);
+#else
+	/*
+	 * Clear protection flags on a whole range of possible
+	 * addresses used for KASLR. We don't need to do that
+	 * on x86_64, since KASLR/extraction is performed after
+	 * dedicated identity page tables are built and we only
+	 * need to remove possible protection on relocated image
+	 * itself disregarding further relocations.
+	 */
+	adjust_memory_range_protection(LOAD_PHYSICAL_ADDR,
+				       KERNEL_IMAGE_SIZE - LOAD_PHYSICAL_ADDR);
+#endif
+}
+
 static const efi_char16_t apple[] = L"Apple";
 
-static void setup_quirks(struct boot_params *boot_params)
+static void setup_quirks(struct boot_params *boot_params,
+			 unsigned long image_base,
+			 unsigned long image_size)
 {
 	efi_char16_t *fw_vendor = (efi_char16_t *)(unsigned long)
 		efi_table_attr(efi_system_table, fw_vendor);
@@ -223,6 +324,9 @@ static void setup_quirks(struct boot_params *boot_params)
 		if (IS_ENABLED(CONFIG_APPLE_PROPERTIES))
 			retrieve_apple_device_properties(boot_params);
 	}
+
+	if (IS_ENABLED(CONFIG_EFI_DXE_MEM_ATTRIBUTES))
+		setup_memory_protection(image_base, image_size);
 }
 
 /*
@@ -342,8 +446,6 @@ static void __noreturn efi_exit(efi_handle_t handle, efi_status_t
status)
 		asm("hlt");
 }
 
-void startup_32(struct boot_params *boot_params);
-
 void __noreturn efi_stub_entry(efi_handle_t handle,
 			       efi_system_table_t *sys_table_arg,
 			       struct boot_params *boot_params);
@@ -798,7 +900,7 @@ unsigned long efi_main(efi_handle_t handle,
 
 	setup_efi_pci(boot_params);
 
-	setup_quirks(boot_params);
+	setup_quirks(boot_params, bzimage_addr, buffer_end - buffer_start);
 
 	status = exit_boot(boot_params, handle);
 	if (status != EFI_SUCCESS) {

--
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/1771

[OS-BUILD PATCHv2 3/4] efi: x86: Set the NX-compatibility flag in the PE header

[Justin M. Forbes (via Email Bridge)]

From: Peter Jones <pjones(a)redhat.com&gt;

efi: x86: Set the NX-compatibility flag in the PE header

Following Baskov Evgeniy's "Handle UEFI NX-restricted page tables"
patches, it's safe to set this compatibility flag to let loaders know
they don't need to make special accommodations for kernel to load if
pre-boot NX is enabled.

Signed-off-by: Peter Jones <pjones(a)redhat.com&gt;
Link: https://lore.kernel.org/all/[e-mail]/
Signed-off-by: Ard Biesheuvel <ardb(a)kernel.org&gt;

diff --git a/arch/x86/boot/header.S b/arch/x86/boot/header.S
index blahblah..blahblah 100644
--- a/arch/x86/boot/header.S
+++ b/arch/x86/boot/header.S
@@ -163,7 +163,11 @@ extra_header_fields:
 	.long	0x200				# SizeOfHeaders
 	.long	0				# CheckSum
 	.word	IMAGE_SUBSYSTEM_EFI_APPLICATION	# Subsystem (EFI application)
+#ifdef CONFIG_DXE_MEM_ATTRIBUTES
+	.word	IMAGE_DLL_CHARACTERISTICS_NX_COMPAT	# DllCharacteristics
+#else
 	.word	0				# DllCharacteristics
+#endif
 #ifdef CONFIG_X86_32
 	.long	0				# SizeOfStackReserve
 	.long	0				# SizeOfStackCommit

--
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/1771

[OS-BUILD PATCHv2 4/4] Add CONFIG_EFI_DXE_MEM_ATTRIBUTES

[Justin M. Forbes (via Email Bridge)]

From: Justin M. Forbes <jforbes(a)fedoraproject.org&gt;

Add CONFIG_EFI_DXE_MEM_ATTRIBUTES

UEFI specification does not guarantee all memory to be
accessible for both write and execute as the kernel expects
it to be.
Use DXE services to check and alter memory protection
attributes during boot via EFISTUB to ensure that memory
ranges used by the kernel are writable and executable.

Signed-off-by: Justin M. Forbes <jforbes(a)fedoraproject.org&gt;

diff --git a/redhat/configs/common/generic/x86/CONFIG_EFI_DXE_MEM_ATTRIBUTES
b/redhat/configs/common/generic/x86/CONFIG_EFI_DXE_MEM_ATTRIBUTES
new file mode 100644
index blahblah..blahblah 100644
--- /dev/null
+++ b/redhat/configs/common/generic/x86/CONFIG_EFI_DXE_MEM_ATTRIBUTES
@@ -0,0 +1 @@
+CONFIG_EFI_DXE_MEM_ATTRIBUTES=y

--
https://gitlab.com/cki-project/kernel-ark/-/merge_requests/1771