Linux-ARM-Kernel Archive mirror
 help / color / mirror / Atom feed
* [PATCH v7 00/11] CFI for ARM32 using LLVM
@ 2024-04-21 18:28 Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 01/11] ARM: bugs: Check in the vtable instead of defined aliases Linus Walleij
                   ` (10 more replies)
  0 siblings, 11 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

This is a first patch set to support CLANG CFI (Control Flow
Integrity) on ARM32.

For information about what CFI is, see:
https://clang.llvm.org/docs/ControlFlowIntegrity.html

For the kernel KCFI flavor, see:
https://lwn.net/Articles/898040/

The base changes required to bring up KCFI on ARM32 was mostly
related to the use of custom vtables in the kernel, combined
with defines to call into these vtable members directly from
sites where they are used.

We annotate all assembly calls that are called directly from
C with SYM_TYPED_FUNC_START()/SYM_FUNC_END() so it is easy
to see while reading the assembly that these functions are
called from C and can have CFI prototype information prefixed
to them.

As protype prefix information is just some random bytes, it is
not possible to "fall through" into an assembly function that
is tagged with SYM_TYPED_FUNC_START(): there will be some
binary noise in front of the function so this design pattern
needs to be explicitly avoided at each site where it occurred.

The approach to binding the calls to C is two-fold:

- Either convert the affected vtable struct to C and provide
  per-CPU prototypes for all the calls (done for TLB, cache)
  or:

- Provide prototypes in a special files just for CFI and tag
  all these functions addressable.

The permissive mode handles the new breakpoint type (0x03) that
LLVM CLANG is emitting.

To runtime-test the patches:
- Enable CONFIG_LKDTM
- echo CFI_FORWARD_PROTO > /sys/kernel/debug/provoke-crash/DIRECT

The patch set has been booted to userspace on the following
test platforms:

- Arm Versatile (QEMU)
- Arm Versatile Express (QEMU)
- multi_v7 booted on Versatile Express (QEMU)
- Footbridge Netwinder (SA110 ARMv4)
- Ux500 (ARMv7 SMP)
- Gemini (FA526)

I am not saying there will not be corner cases that we need
to fix in addition to this, but it is enough to get started.
Looking at what was fixed for arm64 I am a bit weary that
e.g. BPF might need something to trampoline properly.

But hopefullt people can get to testing it and help me fix
remaining issues before the final version, or we can fix it
in-tree.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
Changes in v7:
- Use report_cfi_failure_noaddr() when reporting CFI faults.
- Leave a better comment on what needs to be done to get better
  target reporting.
- Link to v6: https://lore.kernel.org/r/20240417-arm32-cfi-v6-0-6486385eb136@linaro.org

Changes in v6:
- Add a separate patch adding aliases for some cache functions
  that were just branches to another function.
- Link to v5: https://lore.kernel.org/r/20240415-arm32-cfi-v5-0-ff11093eeccc@linaro.org

Changes in v5:
- I started to put the patches into the patch tracker and it
  rightfully complained that the patches tagging all assembly
  with CFI symbol type macros and adding C prototypes were
  too large.
- Split the two patches annotating assembly into one patch
  doing the annotation and one patch adding the C prototypes.
  This is a good split anyway.
- The first patches from the series are unchanged and in the
  patch tracker, I resend them anyway and will soon populate
  the patch tracker with the split patches from this
  series unless there are more comments.
- Link to v4: https://lore.kernel.org/r/20240328-arm32-cfi-v4-0-a11046139125@linaro.org

Changes in v4:
- Rebase on v6.9-rc1
- Use Ard's patch for converting TLB operation vtables to C
- Rewrite the cache vtables in C and use SYM_SYM_TYPED_FUNC in the
  assembly to make CFI work all the way down.
- Instead of tagging all the delay functions as __nocfi get to the
  root cause and annotate the loop delay code with SYM_TYPED_FUNC_START()
  and rewrite it using explicit branches so we get CFI all the way
  down.
- Drop the patch turning highmem page accesses into static inlines:
  this was probably a development artifact since this code does
  a lot of cache and TLB flusing, and that assembly is now properly
  annotated.
- Do not define static inlines tagged __nocfi for all the proc functions,
  instead provide proper C prototypes in a separate CFI-only file
  and make these explicitly addressable.
- Link to v3: https://lore.kernel.org/r/20240311-arm32-cfi-v3-0-224a0f0a45c2@linaro.org

Changes in v3:
- Use report_cfi_failure() like everyone else in the breakpoint
  handler.
- I think we cannot implement target and type for the report callback
  without operand bundling compiler extensions, so just leaving these as zero.
- Link to v2: https://lore.kernel.org/r/20240307-arm32-cfi-v2-0-cc74ea0306b3@linaro.org

Changes in v2:
- Add the missing ftrace graph tracer stub.
- Enable permissive mode using a breakpoint handler.
- Link to v1: https://lore.kernel.org/r/20240225-arm32-cfi-v1-0-6943306f065b@linaro.org

---
Ard Biesheuvel (1):
      ARM: mm: Make tlbflush routines CFI safe

Linus Walleij (10):
      ARM: bugs: Check in the vtable instead of defined aliases
      ARM: ftrace: Define ftrace_stub_graph
      ARM: mm: Type-annotate all cache assembly routines
      ARM: mm: Use symbol alias for two cache functions
      ARM: mm: Rewrite cacheflush vtables in CFI safe C
      ARM: mm: Type-annotate all per-processor assembly routines
      ARM: mm: Define prototypes for all per-processor calls
      ARM: lib: Annotate loop delay instructions for CFI
      ARM: hw_breakpoint: Handle CFI breakpoints
      ARM: Support CLANG CFI

 arch/arm/Kconfig                     |   1 +
 arch/arm/include/asm/glue-cache.h    |  28 +-
 arch/arm/include/asm/hw_breakpoint.h |   1 +
 arch/arm/kernel/bugs.c               |   2 +-
 arch/arm/kernel/entry-ftrace.S       |   4 +
 arch/arm/kernel/hw_breakpoint.c      |  35 ++
 arch/arm/lib/delay-loop.S            |  16 +-
 arch/arm/mm/Makefile                 |   3 +
 arch/arm/mm/cache-b15-rac.c          |   1 +
 arch/arm/mm/cache-fa.S               |  43 ++-
 arch/arm/mm/cache-nop.S              |  61 ++--
 arch/arm/mm/cache-v4.S               |  53 ++-
 arch/arm/mm/cache-v4wb.S             |  43 +--
 arch/arm/mm/cache-v4wt.S             |  51 ++-
 arch/arm/mm/cache-v6.S               |  47 ++-
 arch/arm/mm/cache-v7.S               |  72 ++--
 arch/arm/mm/cache-v7m.S              |  53 ++-
 arch/arm/mm/cache.c                  | 663 +++++++++++++++++++++++++++++++++++
 arch/arm/mm/proc-arm1020.S           |  65 ++--
 arch/arm/mm/proc-arm1020e.S          |  66 ++--
 arch/arm/mm/proc-arm1022.S           |  65 ++--
 arch/arm/mm/proc-arm1026.S           |  66 ++--
 arch/arm/mm/proc-arm720.S            |  25 +-
 arch/arm/mm/proc-arm740.S            |  26 +-
 arch/arm/mm/proc-arm7tdmi.S          |  34 +-
 arch/arm/mm/proc-arm920.S            |  72 ++--
 arch/arm/mm/proc-arm922.S            |  65 ++--
 arch/arm/mm/proc-arm925.S            |  62 ++--
 arch/arm/mm/proc-arm926.S            |  71 ++--
 arch/arm/mm/proc-arm940.S            |  65 ++--
 arch/arm/mm/proc-arm946.S            |  61 ++--
 arch/arm/mm/proc-arm9tdmi.S          |  26 +-
 arch/arm/mm/proc-fa526.S             |  24 +-
 arch/arm/mm/proc-feroceon.S          | 101 +++---
 arch/arm/mm/proc-macros.S            |  33 --
 arch/arm/mm/proc-mohawk.S            |  70 ++--
 arch/arm/mm/proc-sa110.S             |  23 +-
 arch/arm/mm/proc-sa1100.S            |  31 +-
 arch/arm/mm/proc-v6.S                |  31 +-
 arch/arm/mm/proc-v7-2level.S         |   8 +-
 arch/arm/mm/proc-v7-3level.S         |   8 +-
 arch/arm/mm/proc-v7-bugs.c           |   4 +-
 arch/arm/mm/proc-v7.S                |  66 ++--
 arch/arm/mm/proc-v7m.S               |  41 +--
 arch/arm/mm/proc-xsc3.S              |  71 ++--
 arch/arm/mm/proc-xscale.S            | 125 +++----
 arch/arm/mm/proc.c                   | 500 ++++++++++++++++++++++++++
 arch/arm/mm/tlb-fa.S                 |  12 +-
 arch/arm/mm/tlb-v4.S                 |  15 +-
 arch/arm/mm/tlb-v4wb.S               |  12 +-
 arch/arm/mm/tlb-v4wbi.S              |  12 +-
 arch/arm/mm/tlb-v6.S                 |  12 +-
 arch/arm/mm/tlb-v7.S                 |  14 +-
 arch/arm/mm/tlb.c                    |  84 +++++
 54 files changed, 2271 insertions(+), 972 deletions(-)
---
base-commit: 4cece764965020c22cff7665b18a012006359095
change-id: 20240115-arm32-cfi-65d60f201108

Best regards,
-- 
Linus Walleij <linus.walleij@linaro.org>


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH v7 01/11] ARM: bugs: Check in the vtable instead of defined aliases
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-22 21:27   ` Sami Tolvanen
  2024-04-21 18:28 ` [PATCH v7 02/11] ARM: ftrace: Define ftrace_stub_graph Linus Walleij
                   ` (9 subsequent siblings)
  10 siblings, 1 reply; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

Instead of checking if cpu_check_bugs() exist, check for this
callback directly in the CPU vtable: this is better because the
function is just a define to the vtable entry and this is why
the code works. But we want to be able to specify a proper
function for cpu_check_bugs() so look into the vtable instead.

In bugs.c assign PROC_VTABLE(switch_mm) instead of
assigning cpu_do_switch_mm where again this is just a define
into the vtable: this makes it possible to make
cpu_do_switch_mm() into a real function.

Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/kernel/bugs.c     | 2 +-
 arch/arm/mm/proc-v7-bugs.c | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/arch/arm/kernel/bugs.c b/arch/arm/kernel/bugs.c
index 087bce6ec8e9..35d39efb51ed 100644
--- a/arch/arm/kernel/bugs.c
+++ b/arch/arm/kernel/bugs.c
@@ -7,7 +7,7 @@
 void check_other_bugs(void)
 {
 #ifdef MULTI_CPU
-	if (cpu_check_bugs)
+	if (PROC_VTABLE(check_bugs))
 		cpu_check_bugs();
 #endif
 }
diff --git a/arch/arm/mm/proc-v7-bugs.c b/arch/arm/mm/proc-v7-bugs.c
index 8bc7a2d6d6c7..ea3ee2bd7b56 100644
--- a/arch/arm/mm/proc-v7-bugs.c
+++ b/arch/arm/mm/proc-v7-bugs.c
@@ -87,14 +87,14 @@ static unsigned int spectre_v2_install_workaround(unsigned int method)
 	case SPECTRE_V2_METHOD_HVC:
 		per_cpu(harden_branch_predictor_fn, cpu) =
 			call_hvc_arch_workaround_1;
-		cpu_do_switch_mm = cpu_v7_hvc_switch_mm;
+		PROC_VTABLE(switch_mm) = cpu_v7_hvc_switch_mm;
 		spectre_v2_method = "hypervisor";
 		break;
 
 	case SPECTRE_V2_METHOD_SMC:
 		per_cpu(harden_branch_predictor_fn, cpu) =
 			call_smc_arch_workaround_1;
-		cpu_do_switch_mm = cpu_v7_smc_switch_mm;
+		PROC_VTABLE(switch_mm) = cpu_v7_smc_switch_mm;
 		spectre_v2_method = "firmware";
 		break;
 	}

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 02/11] ARM: ftrace: Define ftrace_stub_graph
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 01/11] ARM: bugs: Check in the vtable instead of defined aliases Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 03/11] ARM: mm: Make tlbflush routines CFI safe Linus Walleij
                   ` (8 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

Several architectures defines this stub for the graph tracer,
and it is needed for CFI, as it needs a separate symbol for it.
The trick from include/asm-generic/vmlinux.lds.h to define
ftrace_stub_graph to ftrace_stub isn't working when using CFI.
Commit 883bbbffa5a4 contains the details.

Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/kernel/entry-ftrace.S | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm/kernel/entry-ftrace.S b/arch/arm/kernel/entry-ftrace.S
index 3e7bcaca5e07..bc598e3d8dd2 100644
--- a/arch/arm/kernel/entry-ftrace.S
+++ b/arch/arm/kernel/entry-ftrace.S
@@ -271,6 +271,10 @@ ENTRY(ftrace_stub)
 	ret	lr
 ENDPROC(ftrace_stub)
 
+ENTRY(ftrace_stub_graph)
+	ret	lr
+ENDPROC(ftrace_stub_graph)
+
 #ifdef CONFIG_DYNAMIC_FTRACE
 
 	__INIT

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 03/11] ARM: mm: Make tlbflush routines CFI safe
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 01/11] ARM: bugs: Check in the vtable instead of defined aliases Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 02/11] ARM: ftrace: Define ftrace_stub_graph Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 05/11] ARM: mm: Use symbol alias for two cache functions Linus Walleij
                   ` (7 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

From: Ard Biesheuvel <ardb@kernel.org>

Instead of avoiding CFI entirely on the TLB flush helpers, reorganize
the code so that the CFI machinery can deal with it. The important
things to take into account are:
- functions in asm called indirectly from C need to be defined using
  SYM_TYPED_FUNC_START()
- a reference to the asm function needs to be visible to the compiler,
  in order to get it to emit the typeid symbol.

The latter means that defining the cpu_tlb_fns structs is best done from
C code, so that the references in the static initializers will be
visible to the compiler.

Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/mm/Makefile      |  1 +
 arch/arm/mm/proc-macros.S | 15 ---------
 arch/arm/mm/tlb-fa.S      | 12 +++----
 arch/arm/mm/tlb-v4.S      | 15 +++++----
 arch/arm/mm/tlb-v4wb.S    | 12 +++----
 arch/arm/mm/tlb-v4wbi.S   | 12 +++----
 arch/arm/mm/tlb-v6.S      | 12 +++----
 arch/arm/mm/tlb-v7.S      | 14 +++-----
 arch/arm/mm/tlb.c         | 84 +++++++++++++++++++++++++++++++++++++++++++++++
 9 files changed, 119 insertions(+), 58 deletions(-)

diff --git a/arch/arm/mm/Makefile b/arch/arm/mm/Makefile
index 71b858c9b10c..cc8255fdf56e 100644
--- a/arch/arm/mm/Makefile
+++ b/arch/arm/mm/Makefile
@@ -62,6 +62,7 @@ obj-$(CONFIG_CPU_TLB_FEROCEON)	+= tlb-v4wbi.o	# reuse v4wbi TLB functions
 obj-$(CONFIG_CPU_TLB_V6)	+= tlb-v6.o
 obj-$(CONFIG_CPU_TLB_V7)	+= tlb-v7.o
 obj-$(CONFIG_CPU_TLB_FA)	+= tlb-fa.o
+obj-y				+= tlb.o
 
 obj-$(CONFIG_CPU_ARM7TDMI)	+= proc-arm7tdmi.o
 obj-$(CONFIG_CPU_ARM720T)	+= proc-arm720.o
diff --git a/arch/arm/mm/proc-macros.S b/arch/arm/mm/proc-macros.S
index e43f6d716b4b..c0acfeac3e84 100644
--- a/arch/arm/mm/proc-macros.S
+++ b/arch/arm/mm/proc-macros.S
@@ -338,21 +338,6 @@ ENTRY(\name\()_cache_fns)
 	.size	\name\()_cache_fns, . - \name\()_cache_fns
 .endm
 
-.macro define_tlb_functions name:req, flags_up:req, flags_smp
-	.type	\name\()_tlb_fns, #object
-	.align 2
-ENTRY(\name\()_tlb_fns)
-	.long	\name\()_flush_user_tlb_range
-	.long	\name\()_flush_kern_tlb_range
-	.ifnb \flags_smp
-		ALT_SMP(.long	\flags_smp )
-		ALT_UP(.long	\flags_up )
-	.else
-		.long	\flags_up
-	.endif
-	.size	\name\()_tlb_fns, . - \name\()_tlb_fns
-.endm
-
 .macro globl_equ x, y
 	.globl	\x
 	.equ	\x, \y
diff --git a/arch/arm/mm/tlb-fa.S b/arch/arm/mm/tlb-fa.S
index def6161ec452..85a6fe766b21 100644
--- a/arch/arm/mm/tlb-fa.S
+++ b/arch/arm/mm/tlb-fa.S
@@ -15,6 +15,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
 #include <asm/tlbflush.h>
@@ -31,7 +32,7 @@
  *	- mm    - mm_struct describing address space
  */
 	.align	4
-ENTRY(fa_flush_user_tlb_range)
+SYM_TYPED_FUNC_START(fa_flush_user_tlb_range)
 	vma_vm_mm ip, r2
 	act_mm	r3				@ get current->active_mm
 	eors	r3, ip, r3			@ == mm ?
@@ -46,9 +47,10 @@ ENTRY(fa_flush_user_tlb_range)
 	blo	1b
 	mcr	p15, 0, r3, c7, c10, 4		@ data write barrier
 	ret	lr
+SYM_FUNC_END(fa_flush_user_tlb_range)
 
 
-ENTRY(fa_flush_kern_tlb_range)
+SYM_TYPED_FUNC_START(fa_flush_kern_tlb_range)
 	mov	r3, #0
 	mcr	p15, 0, r3, c7, c10, 4		@ drain WB
 	bic	r0, r0, #0x0ff
@@ -60,8 +62,4 @@ ENTRY(fa_flush_kern_tlb_range)
 	mcr	p15, 0, r3, c7, c10, 4		@ data write barrier
 	mcr	p15, 0, r3, c7, c5, 4		@ prefetch flush (isb)
 	ret	lr
-
-	__INITDATA
-
-	/* define struct cpu_tlb_fns (see <asm/tlbflush.h> and proc-macros.S) */
-	define_tlb_functions fa, fa_tlb_flags
+SYM_FUNC_END(fa_flush_kern_tlb_range)
diff --git a/arch/arm/mm/tlb-v4.S b/arch/arm/mm/tlb-v4.S
index b962b4e75158..09ff69008d94 100644
--- a/arch/arm/mm/tlb-v4.S
+++ b/arch/arm/mm/tlb-v4.S
@@ -11,6 +11,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
 #include <asm/tlbflush.h>
@@ -27,7 +28,7 @@
  *	- mm    - mm_struct describing address space
  */
 	.align	5
-ENTRY(v4_flush_user_tlb_range)
+SYM_TYPED_FUNC_START(v4_flush_user_tlb_range)
 	vma_vm_mm ip, r2
 	act_mm	r3				@ get current->active_mm
 	eors	r3, ip, r3				@ == mm ?
@@ -40,6 +41,7 @@ ENTRY(v4_flush_user_tlb_range)
 	cmp	r0, r1
 	blo	1b
 	ret	lr
+SYM_FUNC_END(v4_flush_user_tlb_range)
 
 /*
  *	v4_flush_kern_tlb_range(start, end)
@@ -50,10 +52,11 @@ ENTRY(v4_flush_user_tlb_range)
  *	- start - virtual address (may not be aligned)
  *	- end   - virtual address (may not be aligned)
  */
+#ifdef CONFIG_CFI_CLANG
+SYM_TYPED_FUNC_START(v4_flush_kern_tlb_range)
+	b	.v4_flush_kern_tlb_range
+SYM_FUNC_END(v4_flush_kern_tlb_range)
+#else
 .globl v4_flush_kern_tlb_range
 .equ v4_flush_kern_tlb_range, .v4_flush_kern_tlb_range
-
-	__INITDATA
-
-	/* define struct cpu_tlb_fns (see <asm/tlbflush.h> and proc-macros.S) */
-	define_tlb_functions v4, v4_tlb_flags
+#endif
diff --git a/arch/arm/mm/tlb-v4wb.S b/arch/arm/mm/tlb-v4wb.S
index 9348bba7586a..04e46c359e75 100644
--- a/arch/arm/mm/tlb-v4wb.S
+++ b/arch/arm/mm/tlb-v4wb.S
@@ -11,6 +11,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
 #include <asm/tlbflush.h>
@@ -27,7 +28,7 @@
  *	- mm    - mm_struct describing address space
  */
 	.align	5
-ENTRY(v4wb_flush_user_tlb_range)
+SYM_TYPED_FUNC_START(v4wb_flush_user_tlb_range)
 	vma_vm_mm ip, r2
 	act_mm	r3				@ get current->active_mm
 	eors	r3, ip, r3				@ == mm ?
@@ -43,6 +44,7 @@ ENTRY(v4wb_flush_user_tlb_range)
 	cmp	r0, r1
 	blo	1b
 	ret	lr
+SYM_FUNC_END(v4wb_flush_user_tlb_range)
 
 /*
  *	v4_flush_kern_tlb_range(start, end)
@@ -53,7 +55,7 @@ ENTRY(v4wb_flush_user_tlb_range)
  *	- start - virtual address (may not be aligned)
  *	- end   - virtual address (may not be aligned)
  */
-ENTRY(v4wb_flush_kern_tlb_range)
+SYM_TYPED_FUNC_START(v4wb_flush_kern_tlb_range)
 	mov	r3, #0
 	mcr	p15, 0, r3, c7, c10, 4		@ drain WB
 	bic	r0, r0, #0x0ff
@@ -64,8 +66,4 @@ ENTRY(v4wb_flush_kern_tlb_range)
 	cmp	r0, r1
 	blo	1b
 	ret	lr
-
-	__INITDATA
-
-	/* define struct cpu_tlb_fns (see <asm/tlbflush.h> and proc-macros.S) */
-	define_tlb_functions v4wb, v4wb_tlb_flags
+SYM_FUNC_END(v4wb_flush_kern_tlb_range)
diff --git a/arch/arm/mm/tlb-v4wbi.S b/arch/arm/mm/tlb-v4wbi.S
index d4f9040a4111..502dfe5628a3 100644
--- a/arch/arm/mm/tlb-v4wbi.S
+++ b/arch/arm/mm/tlb-v4wbi.S
@@ -11,6 +11,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
 #include <asm/tlbflush.h>
@@ -26,7 +27,7 @@
  *	- mm    - mm_struct describing address space
  */
 	.align	5
-ENTRY(v4wbi_flush_user_tlb_range)
+SYM_TYPED_FUNC_START(v4wbi_flush_user_tlb_range)
 	vma_vm_mm ip, r2
 	act_mm	r3				@ get current->active_mm
 	eors	r3, ip, r3			@ == mm ?
@@ -43,8 +44,9 @@ ENTRY(v4wbi_flush_user_tlb_range)
 	cmp	r0, r1
 	blo	1b
 	ret	lr
+SYM_FUNC_END(v4wbi_flush_user_tlb_range)
 
-ENTRY(v4wbi_flush_kern_tlb_range)
+SYM_TYPED_FUNC_START(v4wbi_flush_kern_tlb_range)
 	mov	r3, #0
 	mcr	p15, 0, r3, c7, c10, 4		@ drain WB
 	bic	r0, r0, #0x0ff
@@ -55,8 +57,4 @@ ENTRY(v4wbi_flush_kern_tlb_range)
 	cmp	r0, r1
 	blo	1b
 	ret	lr
-
-	__INITDATA
-
-	/* define struct cpu_tlb_fns (see <asm/tlbflush.h> and proc-macros.S) */
-	define_tlb_functions v4wbi, v4wbi_tlb_flags
+SYM_FUNC_END(v4wbi_flush_kern_tlb_range)
diff --git a/arch/arm/mm/tlb-v6.S b/arch/arm/mm/tlb-v6.S
index 1d91e49b2c2d..8256a67ac654 100644
--- a/arch/arm/mm/tlb-v6.S
+++ b/arch/arm/mm/tlb-v6.S
@@ -9,6 +9,7 @@
  */
 #include <linux/init.h>
 #include <linux/linkage.h>
+#include <linux/cfi_types.h>
 #include <asm/asm-offsets.h>
 #include <asm/assembler.h>
 #include <asm/page.h>
@@ -32,7 +33,7 @@
  *	- the "Invalidate single entry" instruction will invalidate
  *	  both the I and the D TLBs on Harvard-style TLBs
  */
-ENTRY(v6wbi_flush_user_tlb_range)
+SYM_TYPED_FUNC_START(v6wbi_flush_user_tlb_range)
 	vma_vm_mm r3, r2			@ get vma->vm_mm
 	mov	ip, #0
 	mmid	r3, r3				@ get vm_mm->context.id
@@ -56,6 +57,7 @@ ENTRY(v6wbi_flush_user_tlb_range)
 	blo	1b
 	mcr	p15, 0, ip, c7, c10, 4		@ data synchronization barrier
 	ret	lr
+SYM_FUNC_END(v6wbi_flush_user_tlb_range)
 
 /*
  *	v6wbi_flush_kern_tlb_range(start,end)
@@ -65,7 +67,7 @@ ENTRY(v6wbi_flush_user_tlb_range)
  *	- start - start address (may not be aligned)
  *	- end   - end address (exclusive, may not be aligned)
  */
-ENTRY(v6wbi_flush_kern_tlb_range)
+SYM_TYPED_FUNC_START(v6wbi_flush_kern_tlb_range)
 	mov	r2, #0
 	mcr	p15, 0, r2, c7, c10, 4		@ drain write buffer
 	mov	r0, r0, lsr #PAGE_SHIFT		@ align address
@@ -85,8 +87,4 @@ ENTRY(v6wbi_flush_kern_tlb_range)
 	mcr	p15, 0, r2, c7, c10, 4		@ data synchronization barrier
 	mcr	p15, 0, r2, c7, c5, 4		@ prefetch flush (isb)
 	ret	lr
-
-	__INIT
-
-	/* define struct cpu_tlb_fns (see <asm/tlbflush.h> and proc-macros.S) */
-	define_tlb_functions v6wbi, v6wbi_tlb_flags
+SYM_FUNC_END(v6wbi_flush_kern_tlb_range)
diff --git a/arch/arm/mm/tlb-v7.S b/arch/arm/mm/tlb-v7.S
index 35fd6d4f0d03..f1aa0764a2cc 100644
--- a/arch/arm/mm/tlb-v7.S
+++ b/arch/arm/mm/tlb-v7.S
@@ -10,6 +10,7 @@
  */
 #include <linux/init.h>
 #include <linux/linkage.h>
+#include <linux/cfi_types.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
 #include <asm/page.h>
@@ -31,7 +32,7 @@
  *	- the "Invalidate single entry" instruction will invalidate
  *	  both the I and the D TLBs on Harvard-style TLBs
  */
-ENTRY(v7wbi_flush_user_tlb_range)
+SYM_TYPED_FUNC_START(v7wbi_flush_user_tlb_range)
 	vma_vm_mm r3, r2			@ get vma->vm_mm
 	mmid	r3, r3				@ get vm_mm->context.id
 	dsb	ish
@@ -57,7 +58,7 @@ ENTRY(v7wbi_flush_user_tlb_range)
 	blo	1b
 	dsb	ish
 	ret	lr
-ENDPROC(v7wbi_flush_user_tlb_range)
+SYM_FUNC_END(v7wbi_flush_user_tlb_range)
 
 /*
  *	v7wbi_flush_kern_tlb_range(start,end)
@@ -67,7 +68,7 @@ ENDPROC(v7wbi_flush_user_tlb_range)
  *	- start - start address (may not be aligned)
  *	- end   - end address (exclusive, may not be aligned)
  */
-ENTRY(v7wbi_flush_kern_tlb_range)
+SYM_TYPED_FUNC_START(v7wbi_flush_kern_tlb_range)
 	dsb	ish
 	mov	r0, r0, lsr #PAGE_SHIFT		@ align address
 	mov	r1, r1, lsr #PAGE_SHIFT
@@ -86,9 +87,4 @@ ENTRY(v7wbi_flush_kern_tlb_range)
 	dsb	ish
 	isb
 	ret	lr
-ENDPROC(v7wbi_flush_kern_tlb_range)
-
-	__INIT
-
-	/* define struct cpu_tlb_fns (see <asm/tlbflush.h> and proc-macros.S) */
-	define_tlb_functions v7wbi, v7wbi_tlb_flags_up, flags_smp=v7wbi_tlb_flags_smp
+SYM_FUNC_END(v7wbi_flush_kern_tlb_range)
diff --git a/arch/arm/mm/tlb.c b/arch/arm/mm/tlb.c
new file mode 100644
index 000000000000..42359793120b
--- /dev/null
+++ b/arch/arm/mm/tlb.c
@@ -0,0 +1,84 @@
+// SPDX-License-Identifier: GPL-2.0-only
+// Copyright 2024 Google LLC
+// Author: Ard Biesheuvel <ardb@google.com>
+
+#include <linux/types.h>
+#include <asm/tlbflush.h>
+
+#ifdef CONFIG_CPU_TLB_V4WT
+void v4_flush_user_tlb_range(unsigned long, unsigned long, struct vm_area_struct *);
+void v4_flush_kern_tlb_range(unsigned long, unsigned long);
+
+struct cpu_tlb_fns v4_tlb_fns __initconst = {
+	.flush_user_range	= v4_flush_user_tlb_range,
+	.flush_kern_range	= v4_flush_kern_tlb_range,
+	.tlb_flags		= v4_tlb_flags,
+};
+#endif
+
+#ifdef CONFIG_CPU_TLB_V4WB
+void v4wb_flush_user_tlb_range(unsigned long, unsigned long, struct vm_area_struct *);
+void v4wb_flush_kern_tlb_range(unsigned long, unsigned long);
+
+struct cpu_tlb_fns v4wb_tlb_fns __initconst = {
+	.flush_user_range	= v4wb_flush_user_tlb_range,
+	.flush_kern_range	= v4wb_flush_kern_tlb_range,
+	.tlb_flags		= v4wb_tlb_flags,
+};
+#endif
+
+#if defined(CONFIG_CPU_TLB_V4WBI) || defined(CONFIG_CPU_TLB_FEROCEON)
+void v4wbi_flush_user_tlb_range(unsigned long, unsigned long, struct vm_area_struct *);
+void v4wbi_flush_kern_tlb_range(unsigned long, unsigned long);
+
+struct cpu_tlb_fns v4wbi_tlb_fns __initconst = {
+	.flush_user_range	= v4wbi_flush_user_tlb_range,
+	.flush_kern_range	= v4wbi_flush_kern_tlb_range,
+	.tlb_flags		= v4wbi_tlb_flags,
+};
+#endif
+
+#ifdef CONFIG_CPU_TLB_V6
+void v6wbi_flush_user_tlb_range(unsigned long, unsigned long, struct vm_area_struct *);
+void v6wbi_flush_kern_tlb_range(unsigned long, unsigned long);
+
+struct cpu_tlb_fns v6wbi_tlb_fns __initconst = {
+	.flush_user_range	= v6wbi_flush_user_tlb_range,
+	.flush_kern_range	= v6wbi_flush_kern_tlb_range,
+	.tlb_flags		= v6wbi_tlb_flags,
+};
+#endif
+
+#ifdef CONFIG_CPU_TLB_V7
+void v7wbi_flush_user_tlb_range(unsigned long, unsigned long, struct vm_area_struct *);
+void v7wbi_flush_kern_tlb_range(unsigned long, unsigned long);
+
+struct cpu_tlb_fns v7wbi_tlb_fns __initconst = {
+	.flush_user_range	= v7wbi_flush_user_tlb_range,
+	.flush_kern_range	= v7wbi_flush_kern_tlb_range,
+	.tlb_flags		= IS_ENABLED(CONFIG_SMP) ? v7wbi_tlb_flags_smp
+							 : v7wbi_tlb_flags_up,
+};
+
+#ifdef CONFIG_SMP_ON_UP
+/* This will be run-time patched so the offset better be right */
+static_assert(offsetof(struct cpu_tlb_fns, tlb_flags) == 8);
+
+asm("	.pushsection	\".alt.smp.init\", \"a\"		\n" \
+    "	.align		2					\n" \
+    "	.long		v7wbi_tlb_fns + 8 - .			\n" \
+    "	.long "  	__stringify(v7wbi_tlb_flags_up) "	\n" \
+    "	.popsection						\n");
+#endif
+#endif
+
+#ifdef CONFIG_CPU_TLB_FA
+void fa_flush_user_tlb_range(unsigned long, unsigned long, struct vm_area_struct *);
+void fa_flush_kern_tlb_range(unsigned long, unsigned long);
+
+struct cpu_tlb_fns fa_tlb_fns __initconst = {
+	.flush_user_range	= fa_flush_user_tlb_range,
+	.flush_kern_range	= fa_flush_kern_tlb_range,
+	.tlb_flags		= fa_tlb_flags,
+};
+#endif

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 05/11] ARM: mm: Use symbol alias for two cache functions
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (2 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 03/11] ARM: mm: Make tlbflush routines CFI safe Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 06/11] ARM: mm: Rewrite cacheflush vtables in CFI safe C Linus Walleij
                   ` (6 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

The cache functions to flush user cache (*_flush_user_cache_all)
and coherent kernel range (*_coherent_kern_range) are in many
cases just a branch to the corresponfing userspace or kernelspace
function. These functions also have the same arguments.

Simplify these two by using SYM_FUNC_ALIAS() in all affected sites.

The NOP cache has very many similar calls which are just returns,
but it would be confusing to use aliases here, so leave all the
explicit returns and drop a comment on why we are not using aliases.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/mm/cache-fa.S      | 8 ++------
 arch/arm/mm/cache-nop.S     | 4 ++++
 arch/arm/mm/cache-v4.S      | 4 +---
 arch/arm/mm/cache-v4wb.S    | 8 ++------
 arch/arm/mm/cache-v4wt.S    | 8 ++------
 arch/arm/mm/cache-v6.S      | 4 +---
 arch/arm/mm/cache-v7.S      | 4 +---
 arch/arm/mm/proc-arm1020.S  | 8 ++------
 arch/arm/mm/proc-arm1020e.S | 8 ++------
 arch/arm/mm/proc-arm1022.S  | 8 ++------
 arch/arm/mm/proc-arm1026.S  | 8 ++------
 arch/arm/mm/proc-arm920.S   | 8 ++------
 arch/arm/mm/proc-arm922.S   | 8 ++------
 arch/arm/mm/proc-arm925.S   | 8 ++------
 arch/arm/mm/proc-arm926.S   | 8 ++------
 arch/arm/mm/proc-arm940.S   | 8 ++------
 arch/arm/mm/proc-arm946.S   | 8 ++------
 arch/arm/mm/proc-feroceon.S | 8 ++------
 arch/arm/mm/proc-mohawk.S   | 8 ++------
 arch/arm/mm/proc-xsc3.S     | 8 ++------
 arch/arm/mm/proc-xscale.S   | 4 +---
 21 files changed, 40 insertions(+), 108 deletions(-)

diff --git a/arch/arm/mm/cache-fa.S b/arch/arm/mm/cache-fa.S
index c3642d5daf38..6fe06608f34e 100644
--- a/arch/arm/mm/cache-fa.S
+++ b/arch/arm/mm/cache-fa.S
@@ -52,9 +52,7 @@ SYM_FUNC_END(fa_flush_icache_all)
  *	Clean and invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(fa_flush_user_cache_all)
-	b	fa_flush_kern_cache_all
-SYM_FUNC_END(fa_flush_user_cache_all)
+SYM_FUNC_ALIAS(fa_flush_user_cache_all, fa_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -113,9 +111,7 @@ SYM_FUNC_END(fa_flush_user_cache_range)
  *	- start  - virtual start address
  *	- end	 - virtual end address
  */
-SYM_TYPED_FUNC_START(fa_coherent_kern_range)
-	b	fa_coherent_user_range
-SYM_FUNC_END(fa_coherent_kern_range)
+SYM_FUNC_ALIAS(fa_coherent_kern_range, fa_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/cache-nop.S b/arch/arm/mm/cache-nop.S
index 56e94091a55f..cd191aa90313 100644
--- a/arch/arm/mm/cache-nop.S
+++ b/arch/arm/mm/cache-nop.S
@@ -6,6 +6,10 @@
 
 #include "proc-macros.S"
 
+/*
+ * These are all open-coded instead of aliased, to make clear
+ * what is going on here: all functions are stubbed out.
+ */
 SYM_TYPED_FUNC_START(nop_flush_icache_all)
 	ret	lr
 SYM_FUNC_END(nop_flush_icache_all)
diff --git a/arch/arm/mm/cache-v4.S b/arch/arm/mm/cache-v4.S
index 22d9c9d9e0d7..f7b7e498d3b6 100644
--- a/arch/arm/mm/cache-v4.S
+++ b/arch/arm/mm/cache-v4.S
@@ -28,9 +28,7 @@ SYM_FUNC_END(v4_flush_icache_all)
  *
  *	- mm	- mm_struct describing address space
  */
-SYM_TYPED_FUNC_START(v4_flush_user_cache_all)
-	b	v4_flush_kern_cache_all
-SYM_FUNC_END(v4_flush_user_cache_all)
+SYM_FUNC_ALIAS(v4_flush_user_cache_all, v4_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
diff --git a/arch/arm/mm/cache-v4wb.S b/arch/arm/mm/cache-v4wb.S
index 0d97b594e23f..19fae44b89cd 100644
--- a/arch/arm/mm/cache-v4wb.S
+++ b/arch/arm/mm/cache-v4wb.S
@@ -66,9 +66,7 @@ SYM_FUNC_END(v4wb_flush_icache_all)
  *	Clean and invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(v4wb_flush_user_cache_all)
-	b	v4wb_flush_kern_cache_all
-SYM_FUNC_END(v4wb_flush_user_cache_all)
+SYM_FUNC_ALIAS(v4wb_flush_user_cache_all, v4wb_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -151,9 +149,7 @@ SYM_FUNC_END(v4wb_flush_kern_dcache_area)
  *	- start  - virtual start address
  *	- end	 - virtual end address
  */
-SYM_TYPED_FUNC_START(v4wb_coherent_kern_range)
-	b	v4wb_coherent_user_range
-SYM_FUNC_END(v4wb_coherent_kern_range)
+SYM_FUNC_ALIAS(v4wb_coherent_kern_range, v4wb_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/cache-v4wt.S b/arch/arm/mm/cache-v4wt.S
index eee6d8f06b4d..5be76ff861d7 100644
--- a/arch/arm/mm/cache-v4wt.S
+++ b/arch/arm/mm/cache-v4wt.S
@@ -56,9 +56,7 @@ SYM_FUNC_END(v4wt_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(v4wt_flush_user_cache_all)
-	b	v4wt_flush_kern_cache_all
-SYM_FUNC_END(v4wt_flush_user_cache_all)
+SYM_FUNC_ALIAS(v4wt_flush_user_cache_all, v4wt_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -109,9 +107,7 @@ SYM_FUNC_END(v4wt_flush_user_cache_range)
  *	- start  - virtual start address
  *	- end	 - virtual end address
  */
-SYM_TYPED_FUNC_START(v4wt_coherent_kern_range)
-	b	v4wt_coherent_user_range
-SYM_FUNC_END(v4wt_coherent_kern_range)
+SYM_FUNC_ALIAS(v4wt_coherent_kern_range, v4wt_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/cache-v6.S b/arch/arm/mm/cache-v6.S
index 5c7549a49db5..a590044b7282 100644
--- a/arch/arm/mm/cache-v6.S
+++ b/arch/arm/mm/cache-v6.S
@@ -116,9 +116,7 @@ SYM_FUNC_END(v6_flush_user_cache_range)
  *	It is assumed that:
  *	- the Icache does not read data from the write buffer
  */
-SYM_TYPED_FUNC_START(v6_coherent_kern_range)
-	b	v6_coherent_user_range
-SYM_FUNC_END(v6_coherent_kern_range)
+SYM_FUNC_ALIAS(v6_coherent_kern_range, v6_coherent_user_range)
 
 /*
  *	v6_coherent_user_range(start,end)
diff --git a/arch/arm/mm/cache-v7.S b/arch/arm/mm/cache-v7.S
index 5908dd54de47..6c0bc756d29a 100644
--- a/arch/arm/mm/cache-v7.S
+++ b/arch/arm/mm/cache-v7.S
@@ -260,9 +260,7 @@ SYM_FUNC_END(v7_flush_user_cache_range)
  *	It is assumed that:
  *	- the Icache does not read data from the write buffer
  */
-SYM_TYPED_FUNC_START(v7_coherent_kern_range)
-	b	v7_coherent_user_range
-SYM_FUNC_END(v7_coherent_kern_range)
+SYM_FUNC_ALIAS(v7_coherent_kern_range, v7_coherent_user_range)
 
 /*
  *	v7_coherent_user_range(start,end)
diff --git a/arch/arm/mm/proc-arm1020.S b/arch/arm/mm/proc-arm1020.S
index a3f99e1c1186..379628e8ef4e 100644
--- a/arch/arm/mm/proc-arm1020.S
+++ b/arch/arm/mm/proc-arm1020.S
@@ -127,9 +127,7 @@ SYM_FUNC_END(arm1020_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(arm1020_flush_user_cache_all)
-	b	arm1020_flush_kern_cache_all
-SYM_FUNC_END(arm1020_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm1020_flush_user_cache_all, arm1020_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -201,9 +199,7 @@ SYM_FUNC_END(arm1020_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm1020_coherent_kern_range)
-	b	arm1020_coherent_user_range
-SYM_FUNC_END(arm1020_coherent_kern_range)
+SYM_FUNC_ALIAS(arm1020_coherent_kern_range, arm1020_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm1020e.S b/arch/arm/mm/proc-arm1020e.S
index 64c63eb5d830..b5846fbea040 100644
--- a/arch/arm/mm/proc-arm1020e.S
+++ b/arch/arm/mm/proc-arm1020e.S
@@ -127,9 +127,7 @@ SYM_FUNC_END(arm1020e_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(arm1020e_flush_user_cache_all)
-	b	arm1020e_flush_kern_cache_all
-SYM_FUNC_END(arm1020e_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm1020e_flush_user_cache_all, arm1020e_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -198,9 +196,7 @@ SYM_FUNC_END(arm1020e_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm1020e_coherent_kern_range)
-	b	arm1020e_coherent_user_range
-SYM_FUNC_END(arm1020e_coherent_kern_range)
+SYM_FUNC_ALIAS(arm1020e_coherent_kern_range, arm1020e_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm1022.S b/arch/arm/mm/proc-arm1022.S
index e170497353ae..c40b268cc274 100644
--- a/arch/arm/mm/proc-arm1022.S
+++ b/arch/arm/mm/proc-arm1022.S
@@ -127,9 +127,7 @@ SYM_FUNC_END(arm1022_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(arm1022_flush_user_cache_all)
-	b	arm1022_flush_kern_cache_all
-SYM_FUNC_END(arm1022_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm1022_flush_user_cache_all, arm1022_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -197,9 +195,7 @@ SYM_FUNC_END(arm1022_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm1022_coherent_kern_range)
-	b	arm1022_coherent_user_range
-SYM_FUNC_END(arm1022_coherent_kern_range)
+SYM_FUNC_ALIAS(arm1022_coherent_kern_range, arm1022_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm1026.S b/arch/arm/mm/proc-arm1026.S
index 4b5a4849ad85..7ef2c6d88dc0 100644
--- a/arch/arm/mm/proc-arm1026.S
+++ b/arch/arm/mm/proc-arm1026.S
@@ -127,9 +127,7 @@ SYM_FUNC_END(arm1026_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(arm1026_flush_user_cache_all)
-	b	arm1026_flush_kern_cache_all
-SYM_FUNC_END(arm1026_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm1026_flush_user_cache_all, arm1026_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -192,9 +190,7 @@ SYM_FUNC_END(arm1026_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm1026_coherent_kern_range)
-	b	arm1026_coherent_user_range
-SYM_FUNC_END(arm1026_coherent_kern_range)
+SYM_FUNC_ALIAS(arm1026_coherent_kern_range, arm1026_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm920.S b/arch/arm/mm/proc-arm920.S
index fbf8937eae85..eb89a322a534 100644
--- a/arch/arm/mm/proc-arm920.S
+++ b/arch/arm/mm/proc-arm920.S
@@ -116,9 +116,7 @@ SYM_FUNC_END(arm920_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(arm920_flush_user_cache_all)
-	b	arm920_flush_kern_cache_all
-SYM_FUNC_END(arm920_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm920_flush_user_cache_all, arm920_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -179,9 +177,7 @@ SYM_FUNC_END(arm920_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm920_coherent_kern_range)
-	b	arm920_coherent_user_range
-SYM_FUNC_END(arm920_coherent_kern_range)
+SYM_FUNC_ALIAS(arm920_coherent_kern_range, arm920_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm922.S b/arch/arm/mm/proc-arm922.S
index ccfff2b65f49..035a1d1a26b0 100644
--- a/arch/arm/mm/proc-arm922.S
+++ b/arch/arm/mm/proc-arm922.S
@@ -118,9 +118,7 @@ SYM_FUNC_END(arm922_flush_icache_all)
  *	Clean and invalidate all cache entries in a particular
  *	address space.
  */
-SYM_TYPED_FUNC_START(arm922_flush_user_cache_all)
-	b	arm922_flush_kern_cache_all
-SYM_FUNC_END(arm922_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm922_flush_user_cache_all, arm922_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -181,9 +179,7 @@ SYM_FUNC_END(arm922_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm922_coherent_kern_range)
-	b	arm922_coherent_user_range
-SYM_FUNC_END(arm922_coherent_kern_range)
+SYM_FUNC_ALIAS(arm922_coherent_kern_range, arm922_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm925.S b/arch/arm/mm/proc-arm925.S
index d0f73242f70a..2510722647b4 100644
--- a/arch/arm/mm/proc-arm925.S
+++ b/arch/arm/mm/proc-arm925.S
@@ -151,9 +151,7 @@ SYM_FUNC_END(arm925_flush_icache_all)
  *	Clean and invalidate all cache entries in a particular
  *	address space.
  */
-SYM_TYPED_FUNC_START(arm925_flush_user_cache_all)
-	b	arm925_flush_kern_cache_all
-SYM_FUNC_END(arm925_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm925_flush_user_cache_all, arm925_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -227,9 +225,7 @@ SYM_FUNC_END(arm925_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm925_coherent_kern_range)
-	b	arm925_coherent_user_range
-SYM_FUNC_END(arm925_coherent_kern_range)
+SYM_FUNC_ALIAS(arm925_coherent_kern_range, arm925_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm926.S b/arch/arm/mm/proc-arm926.S
index 00f953dee122..dac4a22369ba 100644
--- a/arch/arm/mm/proc-arm926.S
+++ b/arch/arm/mm/proc-arm926.S
@@ -117,9 +117,7 @@ SYM_FUNC_END(arm926_flush_icache_all)
  *	Clean and invalidate all cache entries in a particular
  *	address space.
  */
-SYM_TYPED_FUNC_START(arm926_flush_user_cache_all)
-	b	arm926_flush_kern_cache_all
-SYM_FUNC_END(arm926_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm926_flush_user_cache_all, arm926_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -190,9 +188,7 @@ SYM_FUNC_END(arm926_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm926_coherent_kern_range)
-	b	arm926_coherent_user_range
-SYM_FUNC_END(arm926_coherent_kern_range)
+SYM_FUNC_ALIAS(arm926_coherent_kern_range, arm926_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm940.S b/arch/arm/mm/proc-arm940.S
index 7e32ec271e8a..7c2268059536 100644
--- a/arch/arm/mm/proc-arm940.S
+++ b/arch/arm/mm/proc-arm940.S
@@ -81,9 +81,7 @@ SYM_FUNC_END(arm940_flush_icache_all)
 /*
  *	flush_user_cache_all()
  */
-SYM_TYPED_FUNC_START(arm940_flush_user_cache_all)
-	b	arm940_flush_kern_cache_all
-SYM_FUNC_END(arm940_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm940_flush_user_cache_all, arm940_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -134,9 +132,7 @@ SYM_FUNC_END(arm940_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm940_coherent_kern_range)
-	b	arm940_flush_kern_dcache_area
-SYM_FUNC_END(arm940_coherent_kern_range)
+SYM_FUNC_ALIAS(arm940_coherent_kern_range, arm940_flush_kern_dcache_area)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-arm946.S b/arch/arm/mm/proc-arm946.S
index 4fc883572e19..3955be1f4521 100644
--- a/arch/arm/mm/proc-arm946.S
+++ b/arch/arm/mm/proc-arm946.S
@@ -88,9 +88,7 @@ SYM_FUNC_END(arm946_flush_icache_all)
 /*
  *	flush_user_cache_all()
  */
-SYM_TYPED_FUNC_START(arm946_flush_user_cache_all)
-	b	arm946_flush_kern_cache_all
-SYM_FUNC_END(arm946_flush_user_cache_all)
+SYM_FUNC_ALIAS(arm946_flush_user_cache_all, arm946_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -168,9 +166,7 @@ SYM_FUNC_END(arm946_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(arm946_coherent_kern_range)
-	b	arm946_coherent_user_range
-SYM_FUNC_END(arm946_coherent_kern_range)
+SYM_FUNC_ALIAS(arm946_coherent_kern_range, arm946_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-feroceon.S b/arch/arm/mm/proc-feroceon.S
index ee936c23cac5..9b1570ea6858 100644
--- a/arch/arm/mm/proc-feroceon.S
+++ b/arch/arm/mm/proc-feroceon.S
@@ -136,9 +136,7 @@ SYM_FUNC_END(feroceon_flush_icache_all)
  *	address space.
  */
 	.align	5
-SYM_TYPED_FUNC_START(feroceon_flush_user_cache_all)
-	b	feroceon_flush_kern_cache_all
-SYM_FUNC_END(feroceon_flush_user_cache_all)
+SYM_FUNC_ALIAS(feroceon_flush_user_cache_all, feroceon_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -206,9 +204,7 @@ SYM_FUNC_END(feroceon_flush_user_cache_range)
  *	- end	- virtual end address
  */
 	.align	5
-SYM_TYPED_FUNC_START(feroceon_coherent_kern_range)
-	b	feroceon_coherent_user_range
-SYM_FUNC_END(feroceon_coherent_kern_range)
+SYM_FUNC_ALIAS(feroceon_coherent_kern_range, feroceon_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-mohawk.S b/arch/arm/mm/proc-mohawk.S
index 519b7ff2c589..0a94cb0464d8 100644
--- a/arch/arm/mm/proc-mohawk.S
+++ b/arch/arm/mm/proc-mohawk.S
@@ -100,9 +100,7 @@ SYM_FUNC_END(mohawk_flush_icache_all)
  *	Clean and invalidate all cache entries in a particular
  *	address space.
  */
-SYM_TYPED_FUNC_START(mohawk_flush_user_cache_all)
-	b	mohawk_flush_kern_cache_all
-SYM_FUNC_END(mohawk_flush_user_cache_all)
+SYM_FUNC_ALIAS(mohawk_flush_user_cache_all, mohawk_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -161,9 +159,7 @@ SYM_FUNC_END(mohawk_flush_user_cache_range)
  *	- start	- virtual start address
  *	- end	- virtual end address
  */
-SYM_TYPED_FUNC_START(mohawk_coherent_kern_range)
-	b	mohawk_coherent_user_range
-SYM_FUNC_END(mohawk_coherent_kern_range)
+SYM_FUNC_ALIAS(mohawk_coherent_kern_range, mohawk_coherent_user_range)
 
 /*
  *	coherent_user_range(start, end)
diff --git a/arch/arm/mm/proc-xsc3.S b/arch/arm/mm/proc-xsc3.S
index f08b3fce4c95..b2d907d748e9 100644
--- a/arch/arm/mm/proc-xsc3.S
+++ b/arch/arm/mm/proc-xsc3.S
@@ -157,9 +157,7 @@ SYM_FUNC_END(xsc3_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(xsc3_flush_user_cache_all)
-	b	xsc3_flush_kern_cache_all
-SYM_FUNC_END(xsc3_flush_user_cache_all)
+SYM_FUNC_ALIAS(xsc3_flush_user_cache_all, xsc3_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()
@@ -221,9 +219,7 @@ SYM_FUNC_END(xsc3_flush_user_cache_range)
  *	Note: single I-cache line invalidation isn't used here since
  *	it also trashes the mini I-cache used by JTAG debuggers.
  */
-SYM_TYPED_FUNC_START(xsc3_coherent_kern_range)
-	b	xsc3_coherent_user_range
-SYM_FUNC_END(xsc3_coherent_kern_range)
+SYM_FUNC_ALIAS(xsc3_coherent_kern_range, xsc3_coherent_user_range)
 
 SYM_TYPED_FUNC_START(xsc3_coherent_user_range)
 	bic	r0, r0, #CACHELINESIZE - 1
diff --git a/arch/arm/mm/proc-xscale.S b/arch/arm/mm/proc-xscale.S
index 3e427db18d5b..05d9ed952983 100644
--- a/arch/arm/mm/proc-xscale.S
+++ b/arch/arm/mm/proc-xscale.S
@@ -199,9 +199,7 @@ SYM_FUNC_END(xscale_flush_icache_all)
  *	Invalidate all cache entries in a particular address
  *	space.
  */
-SYM_TYPED_FUNC_START(xscale_flush_user_cache_all)
-	b	xscale_flush_kern_cache_all
-SYM_FUNC_END(xscale_flush_user_cache_all)
+SYM_FUNC_ALIAS(xscale_flush_user_cache_all, xscale_flush_kern_cache_all)
 
 /*
  *	flush_kern_cache_all()

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 06/11] ARM: mm: Rewrite cacheflush vtables in CFI safe C
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (3 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 05/11] ARM: mm: Use symbol alias for two cache functions Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 07/11] ARM: mm: Type-annotate all per-processor assembly routines Linus Walleij
                   ` (5 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

Instead of defining all cache flush operations with an assembly
macro in proc-macros.S, provide an explicit struct cpu_cache_fns
for each CPU cache type in mm/cache.c.

As a side effect from rewriting the vtables in C, we can
avoid the aliasing for the "louis" cache callback, instead we
can just assign the NN_flush_kern_cache_all() function to the
louis callback in the C vtable.

As the louis cache callback is called explicitly (not through the
vtable) if we only have one type of cache support compiled in, we
need an ifdef quirk for this in the !MULTI_CACHE case.

Feroceon and XScale have some dma mapping quirk, in this case we
can just define two structs and assign all but one callback to the
main implementation; since each of them invoked define_cache_functions
twice they require MULTI_CACHE by definition so the compiled-in
shortcut is not used on these variants.

Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/include/asm/glue-cache.h |  28 +-
 arch/arm/mm/Makefile              |   1 +
 arch/arm/mm/cache-b15-rac.c       |   1 +
 arch/arm/mm/cache-fa.S            |   8 -
 arch/arm/mm/cache-nop.S           |   8 -
 arch/arm/mm/cache-v4.S            |   8 -
 arch/arm/mm/cache-v4wb.S          |   8 -
 arch/arm/mm/cache-v4wt.S          |   8 -
 arch/arm/mm/cache-v6.S            |   8 -
 arch/arm/mm/cache-v7.S            |  25 --
 arch/arm/mm/cache-v7m.S           |   8 -
 arch/arm/mm/cache.c               | 663 ++++++++++++++++++++++++++++++++++++++
 arch/arm/mm/proc-arm1020.S        |   6 -
 arch/arm/mm/proc-arm1020e.S       |   6 -
 arch/arm/mm/proc-arm1022.S        |   6 -
 arch/arm/mm/proc-arm1026.S        |   6 -
 arch/arm/mm/proc-arm920.S         |   5 -
 arch/arm/mm/proc-arm922.S         |   6 -
 arch/arm/mm/proc-arm925.S         |   6 -
 arch/arm/mm/proc-arm926.S         |   6 -
 arch/arm/mm/proc-arm940.S         |   6 -
 arch/arm/mm/proc-arm946.S         |   6 -
 arch/arm/mm/proc-feroceon.S       |  27 --
 arch/arm/mm/proc-macros.S         |  18 --
 arch/arm/mm/proc-mohawk.S         |   6 -
 arch/arm/mm/proc-xsc3.S           |   6 -
 arch/arm/mm/proc-xscale.S         |  57 +---
 27 files changed, 688 insertions(+), 259 deletions(-)

diff --git a/arch/arm/include/asm/glue-cache.h b/arch/arm/include/asm/glue-cache.h
index 724f8dac1e5b..4186fbf7341f 100644
--- a/arch/arm/include/asm/glue-cache.h
+++ b/arch/arm/include/asm/glue-cache.h
@@ -118,6 +118,10 @@
 # define MULTI_CACHE 1
 #endif
 
+#ifdef CONFIG_CPU_CACHE_NOP
+#  define MULTI_CACHE 1
+#endif
+
 #if defined(CONFIG_CPU_V7M)
 #  define MULTI_CACHE 1
 #endif
@@ -126,29 +130,15 @@
 #error Unknown cache maintenance model
 #endif
 
-#ifndef __ASSEMBLER__
-static inline void nop_flush_icache_all(void) { }
-static inline void nop_flush_kern_cache_all(void) { }
-static inline void nop_flush_kern_cache_louis(void) { }
-static inline void nop_flush_user_cache_all(void) { }
-static inline void nop_flush_user_cache_range(unsigned long a,
-		unsigned long b, unsigned int c) { }
-
-static inline void nop_coherent_kern_range(unsigned long a, unsigned long b) { }
-static inline int nop_coherent_user_range(unsigned long a,
-		unsigned long b) { return 0; }
-static inline void nop_flush_kern_dcache_area(void *a, size_t s) { }
-
-static inline void nop_dma_flush_range(const void *a, const void *b) { }
-
-static inline void nop_dma_map_area(const void *s, size_t l, int f) { }
-static inline void nop_dma_unmap_area(const void *s, size_t l, int f) { }
-#endif
-
 #ifndef MULTI_CACHE
 #define __cpuc_flush_icache_all		__glue(_CACHE,_flush_icache_all)
 #define __cpuc_flush_kern_all		__glue(_CACHE,_flush_kern_cache_all)
+/* This function only has a dedicated assembly callback on the v7 cache */
+#ifdef CONFIG_CPU_CACHE_V7
 #define __cpuc_flush_kern_louis		__glue(_CACHE,_flush_kern_cache_louis)
+#else
+#define __cpuc_flush_kern_louis		__glue(_CACHE,_flush_kern_cache_all)
+#endif
 #define __cpuc_flush_user_all		__glue(_CACHE,_flush_user_cache_all)
 #define __cpuc_flush_user_range		__glue(_CACHE,_flush_user_cache_range)
 #define __cpuc_coherent_kern_range	__glue(_CACHE,_coherent_kern_range)
diff --git a/arch/arm/mm/Makefile b/arch/arm/mm/Makefile
index cc8255fdf56e..17665381be96 100644
--- a/arch/arm/mm/Makefile
+++ b/arch/arm/mm/Makefile
@@ -45,6 +45,7 @@ obj-$(CONFIG_CPU_CACHE_V7)	+= cache-v7.o
 obj-$(CONFIG_CPU_CACHE_FA)	+= cache-fa.o
 obj-$(CONFIG_CPU_CACHE_NOP)	+= cache-nop.o
 obj-$(CONFIG_CPU_CACHE_V7M)	+= cache-v7m.o
+obj-y				+= cache.o
 
 obj-$(CONFIG_CPU_COPY_V4WT)	+= copypage-v4wt.o
 obj-$(CONFIG_CPU_COPY_V4WB)	+= copypage-v4wb.o
diff --git a/arch/arm/mm/cache-b15-rac.c b/arch/arm/mm/cache-b15-rac.c
index 9c1172f26885..6f63b90f9e1a 100644
--- a/arch/arm/mm/cache-b15-rac.c
+++ b/arch/arm/mm/cache-b15-rac.c
@@ -5,6 +5,7 @@
  * Copyright (C) 2015-2016 Broadcom
  */
 
+#include <linux/cfi_types.h>
 #include <linux/err.h>
 #include <linux/spinlock.h>
 #include <linux/io.h>
diff --git a/arch/arm/mm/cache-fa.S b/arch/arm/mm/cache-fa.S
index 6fe06608f34e..4610105e058c 100644
--- a/arch/arm/mm/cache-fa.S
+++ b/arch/arm/mm/cache-fa.S
@@ -241,11 +241,3 @@ SYM_FUNC_END(fa_dma_map_area)
 SYM_TYPED_FUNC_START(fa_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(fa_dma_unmap_area)
-
-	.globl	fa_flush_kern_cache_louis
-	.equ	fa_flush_kern_cache_louis, fa_flush_kern_cache_all
-
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions fa
diff --git a/arch/arm/mm/cache-nop.S b/arch/arm/mm/cache-nop.S
index cd191aa90313..f68dde2014ee 100644
--- a/arch/arm/mm/cache-nop.S
+++ b/arch/arm/mm/cache-nop.S
@@ -18,9 +18,6 @@ SYM_TYPED_FUNC_START(nop_flush_kern_cache_all)
 	ret	lr
 SYM_FUNC_END(nop_flush_kern_cache_all)
 
-	.globl nop_flush_kern_cache_louis
-	.equ nop_flush_kern_cache_louis, nop_flush_icache_all
-
 SYM_TYPED_FUNC_START(nop_flush_user_cache_all)
 	ret	lr
 SYM_FUNC_END(nop_flush_user_cache_all)
@@ -50,11 +47,6 @@ SYM_TYPED_FUNC_START(nop_dma_map_area)
 	ret	lr
 SYM_FUNC_END(nop_dma_map_area)
 
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions nop
-
 SYM_TYPED_FUNC_START(nop_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(nop_dma_unmap_area)
diff --git a/arch/arm/mm/cache-v4.S b/arch/arm/mm/cache-v4.S
index f7b7e498d3b6..0df97a610026 100644
--- a/arch/arm/mm/cache-v4.S
+++ b/arch/arm/mm/cache-v4.S
@@ -144,11 +144,3 @@ SYM_FUNC_END(v4_dma_unmap_area)
 SYM_TYPED_FUNC_START(v4_dma_map_area)
 	ret	lr
 SYM_FUNC_END(v4_dma_map_area)
-
-	.globl	v4_flush_kern_cache_louis
-	.equ	v4_flush_kern_cache_louis, v4_flush_kern_cache_all
-
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions v4
diff --git a/arch/arm/mm/cache-v4wb.S b/arch/arm/mm/cache-v4wb.S
index 19fae44b89cd..945a7881cc94 100644
--- a/arch/arm/mm/cache-v4wb.S
+++ b/arch/arm/mm/cache-v4wb.S
@@ -251,11 +251,3 @@ SYM_FUNC_END(v4wb_dma_map_area)
 SYM_TYPED_FUNC_START(v4wb_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(v4wb_dma_unmap_area)
-
-	.globl	v4wb_flush_kern_cache_louis
-	.equ	v4wb_flush_kern_cache_louis, v4wb_flush_kern_cache_all
-
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions v4wb
diff --git a/arch/arm/mm/cache-v4wt.S b/arch/arm/mm/cache-v4wt.S
index 5be76ff861d7..d788962e2ed8 100644
--- a/arch/arm/mm/cache-v4wt.S
+++ b/arch/arm/mm/cache-v4wt.S
@@ -198,11 +198,3 @@ SYM_FUNC_END(v4wt_dma_unmap_area)
 SYM_TYPED_FUNC_START(v4wt_dma_map_area)
 	ret	lr
 SYM_FUNC_END(v4wt_dma_map_area)
-
-	.globl	v4wt_flush_kern_cache_louis
-	.equ	v4wt_flush_kern_cache_louis, v4wt_flush_kern_cache_all
-
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions v4wt
diff --git a/arch/arm/mm/cache-v6.S b/arch/arm/mm/cache-v6.S
index a590044b7282..ebe96d40907a 100644
--- a/arch/arm/mm/cache-v6.S
+++ b/arch/arm/mm/cache-v6.S
@@ -296,11 +296,3 @@ SYM_TYPED_FUNC_START(v6_dma_unmap_area)
 	bne	v6_dma_inv_range
 	ret	lr
 SYM_FUNC_END(v6_dma_unmap_area)
-
-	.globl	v6_flush_kern_cache_louis
-	.equ	v6_flush_kern_cache_louis, v6_flush_kern_cache_all
-
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions v6
diff --git a/arch/arm/mm/cache-v7.S b/arch/arm/mm/cache-v7.S
index 6c0bc756d29a..c0ebe1aa0f02 100644
--- a/arch/arm/mm/cache-v7.S
+++ b/arch/arm/mm/cache-v7.S
@@ -454,28 +454,3 @@ SYM_TYPED_FUNC_START(v7_dma_unmap_area)
 	bne	v7_dma_inv_range
 	ret	lr
 SYM_FUNC_END(v7_dma_unmap_area)
-
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions v7
-
-	/* The Broadcom Brahma-B15 read-ahead cache requires some modifications
-	 * to the v7_cache_fns, we only override the ones we need
-	 */
-#ifndef CONFIG_CACHE_B15_RAC
-	globl_equ	b15_flush_kern_cache_all,	v7_flush_kern_cache_all
-#endif
-	globl_equ	b15_flush_icache_all,		v7_flush_icache_all
-	globl_equ	b15_flush_kern_cache_louis,	v7_flush_kern_cache_louis
-	globl_equ	b15_flush_user_cache_all,	v7_flush_user_cache_all
-	globl_equ	b15_flush_user_cache_range,	v7_flush_user_cache_range
-	globl_equ	b15_coherent_kern_range,	v7_coherent_kern_range
-	globl_equ	b15_coherent_user_range,	v7_coherent_user_range
-	globl_equ	b15_flush_kern_dcache_area,	v7_flush_kern_dcache_area
-
-	globl_equ	b15_dma_map_area,		v7_dma_map_area
-	globl_equ	b15_dma_unmap_area,		v7_dma_unmap_area
-	globl_equ	b15_dma_flush_range,		v7_dma_flush_range
-
-	define_cache_functions b15
diff --git a/arch/arm/mm/cache-v7m.S b/arch/arm/mm/cache-v7m.S
index 5a62b9a224e1..4e670697eabc 100644
--- a/arch/arm/mm/cache-v7m.S
+++ b/arch/arm/mm/cache-v7m.S
@@ -447,11 +447,3 @@ SYM_TYPED_FUNC_START(v7m_dma_unmap_area)
 	bne	v7m_dma_inv_range
 	ret	lr
 SYM_FUNC_END(v7m_dma_unmap_area)
-
-	.globl	v7m_flush_kern_cache_louis
-	.equ	v7m_flush_kern_cache_louis, v7m_flush_kern_cache_all
-
-	__INITDATA
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions v7m
diff --git a/arch/arm/mm/cache.c b/arch/arm/mm/cache.c
new file mode 100644
index 000000000000..e6fbc599c9ed
--- /dev/null
+++ b/arch/arm/mm/cache.c
@@ -0,0 +1,663 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file defines C prototypes for the low-level cache assembly functions
+ * and populates a vtable for each selected ARM CPU cache type.
+ */
+
+#include <linux/types.h>
+#include <asm/cacheflush.h>
+
+#ifdef CONFIG_CPU_CACHE_V4
+void v4_flush_icache_all(void);
+void v4_flush_kern_cache_all(void);
+void v4_flush_user_cache_all(void);
+void v4_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void v4_coherent_kern_range(unsigned long, unsigned long);
+int v4_coherent_user_range(unsigned long, unsigned long);
+void v4_flush_kern_dcache_area(void *, size_t);
+void v4_dma_map_area(const void *, size_t, int);
+void v4_dma_unmap_area(const void *, size_t, int);
+void v4_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns v4_cache_fns __initconst = {
+	.flush_icache_all = v4_flush_icache_all,
+	.flush_kern_all = v4_flush_kern_cache_all,
+	.flush_kern_louis = v4_flush_kern_cache_all,
+	.flush_user_all = v4_flush_user_cache_all,
+	.flush_user_range = v4_flush_user_cache_range,
+	.coherent_kern_range = v4_coherent_kern_range,
+	.coherent_user_range = v4_coherent_user_range,
+	.flush_kern_dcache_area = v4_flush_kern_dcache_area,
+	.dma_map_area = v4_dma_map_area,
+	.dma_unmap_area = v4_dma_unmap_area,
+	.dma_flush_range = v4_dma_flush_range,
+};
+#endif
+
+/* V4 write-back cache "V4WB" */
+#ifdef CONFIG_CPU_CACHE_V4WB
+void v4wb_flush_icache_all(void);
+void v4wb_flush_kern_cache_all(void);
+void v4wb_flush_user_cache_all(void);
+void v4wb_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void v4wb_coherent_kern_range(unsigned long, unsigned long);
+int v4wb_coherent_user_range(unsigned long, unsigned long);
+void v4wb_flush_kern_dcache_area(void *, size_t);
+void v4wb_dma_map_area(const void *, size_t, int);
+void v4wb_dma_unmap_area(const void *, size_t, int);
+void v4wb_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns v4wb_cache_fns __initconst = {
+	.flush_icache_all = v4wb_flush_icache_all,
+	.flush_kern_all = v4wb_flush_kern_cache_all,
+	.flush_kern_louis = v4wb_flush_kern_cache_all,
+	.flush_user_all = v4wb_flush_user_cache_all,
+	.flush_user_range = v4wb_flush_user_cache_range,
+	.coherent_kern_range = v4wb_coherent_kern_range,
+	.coherent_user_range = v4wb_coherent_user_range,
+	.flush_kern_dcache_area = v4wb_flush_kern_dcache_area,
+	.dma_map_area = v4wb_dma_map_area,
+	.dma_unmap_area = v4wb_dma_unmap_area,
+	.dma_flush_range = v4wb_dma_flush_range,
+};
+#endif
+
+/* V4 write-through cache "V4WT" */
+#ifdef CONFIG_CPU_CACHE_V4WT
+void v4wt_flush_icache_all(void);
+void v4wt_flush_kern_cache_all(void);
+void v4wt_flush_user_cache_all(void);
+void v4wt_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void v4wt_coherent_kern_range(unsigned long, unsigned long);
+int v4wt_coherent_user_range(unsigned long, unsigned long);
+void v4wt_flush_kern_dcache_area(void *, size_t);
+void v4wt_dma_map_area(const void *, size_t, int);
+void v4wt_dma_unmap_area(const void *, size_t, int);
+void v4wt_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns v4wt_cache_fns __initconst = {
+	.flush_icache_all = v4wt_flush_icache_all,
+	.flush_kern_all = v4wt_flush_kern_cache_all,
+	.flush_kern_louis = v4wt_flush_kern_cache_all,
+	.flush_user_all = v4wt_flush_user_cache_all,
+	.flush_user_range = v4wt_flush_user_cache_range,
+	.coherent_kern_range = v4wt_coherent_kern_range,
+	.coherent_user_range = v4wt_coherent_user_range,
+	.flush_kern_dcache_area = v4wt_flush_kern_dcache_area,
+	.dma_map_area = v4wt_dma_map_area,
+	.dma_unmap_area = v4wt_dma_unmap_area,
+	.dma_flush_range = v4wt_dma_flush_range,
+};
+#endif
+
+/* Faraday FA526 cache */
+#ifdef CONFIG_CPU_CACHE_FA
+void fa_flush_icache_all(void);
+void fa_flush_kern_cache_all(void);
+void fa_flush_user_cache_all(void);
+void fa_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void fa_coherent_kern_range(unsigned long, unsigned long);
+int fa_coherent_user_range(unsigned long, unsigned long);
+void fa_flush_kern_dcache_area(void *, size_t);
+void fa_dma_map_area(const void *, size_t, int);
+void fa_dma_unmap_area(const void *, size_t, int);
+void fa_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns fa_cache_fns __initconst = {
+	.flush_icache_all = fa_flush_icache_all,
+	.flush_kern_all = fa_flush_kern_cache_all,
+	.flush_kern_louis = fa_flush_kern_cache_all,
+	.flush_user_all = fa_flush_user_cache_all,
+	.flush_user_range = fa_flush_user_cache_range,
+	.coherent_kern_range = fa_coherent_kern_range,
+	.coherent_user_range = fa_coherent_user_range,
+	.flush_kern_dcache_area = fa_flush_kern_dcache_area,
+	.dma_map_area = fa_dma_map_area,
+	.dma_unmap_area = fa_dma_unmap_area,
+	.dma_flush_range = fa_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_CACHE_V6
+void v6_flush_icache_all(void);
+void v6_flush_kern_cache_all(void);
+void v6_flush_user_cache_all(void);
+void v6_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void v6_coherent_kern_range(unsigned long, unsigned long);
+int v6_coherent_user_range(unsigned long, unsigned long);
+void v6_flush_kern_dcache_area(void *, size_t);
+void v6_dma_map_area(const void *, size_t, int);
+void v6_dma_unmap_area(const void *, size_t, int);
+void v6_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns v6_cache_fns __initconst = {
+	.flush_icache_all = v6_flush_icache_all,
+	.flush_kern_all = v6_flush_kern_cache_all,
+	.flush_kern_louis = v6_flush_kern_cache_all,
+	.flush_user_all = v6_flush_user_cache_all,
+	.flush_user_range = v6_flush_user_cache_range,
+	.coherent_kern_range = v6_coherent_kern_range,
+	.coherent_user_range = v6_coherent_user_range,
+	.flush_kern_dcache_area = v6_flush_kern_dcache_area,
+	.dma_map_area = v6_dma_map_area,
+	.dma_unmap_area = v6_dma_unmap_area,
+	.dma_flush_range = v6_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_CACHE_V7
+void v7_flush_icache_all(void);
+void v7_flush_kern_cache_all(void);
+void v7_flush_kern_cache_louis(void);
+void v7_flush_user_cache_all(void);
+void v7_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void v7_coherent_kern_range(unsigned long, unsigned long);
+int v7_coherent_user_range(unsigned long, unsigned long);
+void v7_flush_kern_dcache_area(void *, size_t);
+void v7_dma_map_area(const void *, size_t, int);
+void v7_dma_unmap_area(const void *, size_t, int);
+void v7_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns v7_cache_fns __initconst = {
+	.flush_icache_all = v7_flush_icache_all,
+	.flush_kern_all = v7_flush_kern_cache_all,
+	.flush_kern_louis = v7_flush_kern_cache_louis,
+	.flush_user_all = v7_flush_user_cache_all,
+	.flush_user_range = v7_flush_user_cache_range,
+	.coherent_kern_range = v7_coherent_kern_range,
+	.coherent_user_range = v7_coherent_user_range,
+	.flush_kern_dcache_area = v7_flush_kern_dcache_area,
+	.dma_map_area = v7_dma_map_area,
+	.dma_unmap_area = v7_dma_unmap_area,
+	.dma_flush_range = v7_dma_flush_range,
+};
+
+/* Special quirky cache flush function for Broadcom B15 v7 caches */
+void b15_flush_kern_cache_all(void);
+
+struct cpu_cache_fns b15_cache_fns __initconst = {
+	.flush_icache_all = v7_flush_icache_all,
+#ifdef CONFIG_CACHE_B15_RAC
+	.flush_kern_all = b15_flush_kern_cache_all,
+#else
+	.flush_kern_all = v7_flush_kern_cache_all,
+#endif
+	.flush_kern_louis = v7_flush_kern_cache_louis,
+	.flush_user_all = v7_flush_user_cache_all,
+	.flush_user_range = v7_flush_user_cache_range,
+	.coherent_kern_range = v7_coherent_kern_range,
+	.coherent_user_range = v7_coherent_user_range,
+	.flush_kern_dcache_area = v7_flush_kern_dcache_area,
+	.dma_map_area = v7_dma_map_area,
+	.dma_unmap_area = v7_dma_unmap_area,
+	.dma_flush_range = v7_dma_flush_range,
+};
+#endif
+
+/* The NOP cache is just a set of dummy stubs that by definition does nothing */
+#ifdef CONFIG_CPU_CACHE_NOP
+void nop_flush_icache_all(void);
+void nop_flush_kern_cache_all(void);
+void nop_flush_user_cache_all(void);
+void nop_flush_user_cache_range(unsigned long start, unsigned long end, unsigned int flags);
+void nop_coherent_kern_range(unsigned long start, unsigned long end);
+int nop_coherent_user_range(unsigned long, unsigned long);
+void nop_flush_kern_dcache_area(void *kaddr, size_t size);
+void nop_dma_map_area(const void *start, size_t size, int flags);
+void nop_dma_unmap_area(const void *start, size_t size, int flags);
+void nop_dma_flush_range(const void *start, const void *end);
+
+struct cpu_cache_fns nop_cache_fns __initconst = {
+	.flush_icache_all = nop_flush_icache_all,
+	.flush_kern_all = nop_flush_kern_cache_all,
+	.flush_kern_louis = nop_flush_kern_cache_all,
+	.flush_user_all = nop_flush_user_cache_all,
+	.flush_user_range = nop_flush_user_cache_range,
+	.coherent_kern_range = nop_coherent_kern_range,
+	.coherent_user_range = nop_coherent_user_range,
+	.flush_kern_dcache_area = nop_flush_kern_dcache_area,
+	.dma_map_area = nop_dma_map_area,
+	.dma_unmap_area = nop_dma_unmap_area,
+	.dma_flush_range = nop_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_CACHE_V7M
+void v7m_flush_icache_all(void);
+void v7m_flush_kern_cache_all(void);
+void v7m_flush_user_cache_all(void);
+void v7m_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void v7m_coherent_kern_range(unsigned long, unsigned long);
+int v7m_coherent_user_range(unsigned long, unsigned long);
+void v7m_flush_kern_dcache_area(void *, size_t);
+void v7m_dma_map_area(const void *, size_t, int);
+void v7m_dma_unmap_area(const void *, size_t, int);
+void v7m_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns v7m_cache_fns __initconst = {
+	.flush_icache_all = v7m_flush_icache_all,
+	.flush_kern_all = v7m_flush_kern_cache_all,
+	.flush_kern_louis = v7m_flush_kern_cache_all,
+	.flush_user_all = v7m_flush_user_cache_all,
+	.flush_user_range = v7m_flush_user_cache_range,
+	.coherent_kern_range = v7m_coherent_kern_range,
+	.coherent_user_range = v7m_coherent_user_range,
+	.flush_kern_dcache_area = v7m_flush_kern_dcache_area,
+	.dma_map_area = v7m_dma_map_area,
+	.dma_unmap_area = v7m_dma_unmap_area,
+	.dma_flush_range = v7m_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM1020
+void arm1020_flush_icache_all(void);
+void arm1020_flush_kern_cache_all(void);
+void arm1020_flush_user_cache_all(void);
+void arm1020_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm1020_coherent_kern_range(unsigned long, unsigned long);
+int arm1020_coherent_user_range(unsigned long, unsigned long);
+void arm1020_flush_kern_dcache_area(void *, size_t);
+void arm1020_dma_map_area(const void *, size_t, int);
+void arm1020_dma_unmap_area(const void *, size_t, int);
+void arm1020_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm1020_cache_fns __initconst = {
+	.flush_icache_all = arm1020_flush_icache_all,
+	.flush_kern_all = arm1020_flush_kern_cache_all,
+	.flush_kern_louis = arm1020_flush_kern_cache_all,
+	.flush_user_all = arm1020_flush_user_cache_all,
+	.flush_user_range = arm1020_flush_user_cache_range,
+	.coherent_kern_range = arm1020_coherent_kern_range,
+	.coherent_user_range = arm1020_coherent_user_range,
+	.flush_kern_dcache_area = arm1020_flush_kern_dcache_area,
+	.dma_map_area = arm1020_dma_map_area,
+	.dma_unmap_area = arm1020_dma_unmap_area,
+	.dma_flush_range = arm1020_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM1020E
+void arm1020e_flush_icache_all(void);
+void arm1020e_flush_kern_cache_all(void);
+void arm1020e_flush_user_cache_all(void);
+void arm1020e_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm1020e_coherent_kern_range(unsigned long, unsigned long);
+int arm1020e_coherent_user_range(unsigned long, unsigned long);
+void arm1020e_flush_kern_dcache_area(void *, size_t);
+void arm1020e_dma_map_area(const void *, size_t, int);
+void arm1020e_dma_unmap_area(const void *, size_t, int);
+void arm1020e_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm1020e_cache_fns __initconst = {
+	.flush_icache_all = arm1020e_flush_icache_all,
+	.flush_kern_all = arm1020e_flush_kern_cache_all,
+	.flush_kern_louis = arm1020e_flush_kern_cache_all,
+	.flush_user_all = arm1020e_flush_user_cache_all,
+	.flush_user_range = arm1020e_flush_user_cache_range,
+	.coherent_kern_range = arm1020e_coherent_kern_range,
+	.coherent_user_range = arm1020e_coherent_user_range,
+	.flush_kern_dcache_area = arm1020e_flush_kern_dcache_area,
+	.dma_map_area = arm1020e_dma_map_area,
+	.dma_unmap_area = arm1020e_dma_unmap_area,
+	.dma_flush_range = arm1020e_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM1022
+void arm1022_flush_icache_all(void);
+void arm1022_flush_kern_cache_all(void);
+void arm1022_flush_user_cache_all(void);
+void arm1022_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm1022_coherent_kern_range(unsigned long, unsigned long);
+int arm1022_coherent_user_range(unsigned long, unsigned long);
+void arm1022_flush_kern_dcache_area(void *, size_t);
+void arm1022_dma_map_area(const void *, size_t, int);
+void arm1022_dma_unmap_area(const void *, size_t, int);
+void arm1022_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm1022_cache_fns __initconst = {
+	.flush_icache_all = arm1022_flush_icache_all,
+	.flush_kern_all = arm1022_flush_kern_cache_all,
+	.flush_kern_louis = arm1022_flush_kern_cache_all,
+	.flush_user_all = arm1022_flush_user_cache_all,
+	.flush_user_range = arm1022_flush_user_cache_range,
+	.coherent_kern_range = arm1022_coherent_kern_range,
+	.coherent_user_range = arm1022_coherent_user_range,
+	.flush_kern_dcache_area = arm1022_flush_kern_dcache_area,
+	.dma_map_area = arm1022_dma_map_area,
+	.dma_unmap_area = arm1022_dma_unmap_area,
+	.dma_flush_range = arm1022_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM1026
+void arm1026_flush_icache_all(void);
+void arm1026_flush_kern_cache_all(void);
+void arm1026_flush_user_cache_all(void);
+void arm1026_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm1026_coherent_kern_range(unsigned long, unsigned long);
+int arm1026_coherent_user_range(unsigned long, unsigned long);
+void arm1026_flush_kern_dcache_area(void *, size_t);
+void arm1026_dma_map_area(const void *, size_t, int);
+void arm1026_dma_unmap_area(const void *, size_t, int);
+void arm1026_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm1026_cache_fns __initconst = {
+	.flush_icache_all = arm1026_flush_icache_all,
+	.flush_kern_all = arm1026_flush_kern_cache_all,
+	.flush_kern_louis = arm1026_flush_kern_cache_all,
+	.flush_user_all = arm1026_flush_user_cache_all,
+	.flush_user_range = arm1026_flush_user_cache_range,
+	.coherent_kern_range = arm1026_coherent_kern_range,
+	.coherent_user_range = arm1026_coherent_user_range,
+	.flush_kern_dcache_area = arm1026_flush_kern_dcache_area,
+	.dma_map_area = arm1026_dma_map_area,
+	.dma_unmap_area = arm1026_dma_unmap_area,
+	.dma_flush_range = arm1026_dma_flush_range,
+};
+#endif
+
+#if defined(CONFIG_CPU_ARM920T) && !defined(CONFIG_CPU_DCACHE_WRITETHROUGH)
+void arm920_flush_icache_all(void);
+void arm920_flush_kern_cache_all(void);
+void arm920_flush_user_cache_all(void);
+void arm920_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm920_coherent_kern_range(unsigned long, unsigned long);
+int arm920_coherent_user_range(unsigned long, unsigned long);
+void arm920_flush_kern_dcache_area(void *, size_t);
+void arm920_dma_map_area(const void *, size_t, int);
+void arm920_dma_unmap_area(const void *, size_t, int);
+void arm920_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm920_cache_fns __initconst = {
+	.flush_icache_all = arm920_flush_icache_all,
+	.flush_kern_all = arm920_flush_kern_cache_all,
+	.flush_kern_louis = arm920_flush_kern_cache_all,
+	.flush_user_all = arm920_flush_user_cache_all,
+	.flush_user_range = arm920_flush_user_cache_range,
+	.coherent_kern_range = arm920_coherent_kern_range,
+	.coherent_user_range = arm920_coherent_user_range,
+	.flush_kern_dcache_area = arm920_flush_kern_dcache_area,
+	.dma_map_area = arm920_dma_map_area,
+	.dma_unmap_area = arm920_dma_unmap_area,
+	.dma_flush_range = arm920_dma_flush_range,
+};
+#endif
+
+#if defined(CONFIG_CPU_ARM922T) && !defined(CONFIG_CPU_DCACHE_WRITETHROUGH)
+void arm922_flush_icache_all(void);
+void arm922_flush_kern_cache_all(void);
+void arm922_flush_user_cache_all(void);
+void arm922_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm922_coherent_kern_range(unsigned long, unsigned long);
+int arm922_coherent_user_range(unsigned long, unsigned long);
+void arm922_flush_kern_dcache_area(void *, size_t);
+void arm922_dma_map_area(const void *, size_t, int);
+void arm922_dma_unmap_area(const void *, size_t, int);
+void arm922_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm922_cache_fns __initconst = {
+	.flush_icache_all = arm922_flush_icache_all,
+	.flush_kern_all = arm922_flush_kern_cache_all,
+	.flush_kern_louis = arm922_flush_kern_cache_all,
+	.flush_user_all = arm922_flush_user_cache_all,
+	.flush_user_range = arm922_flush_user_cache_range,
+	.coherent_kern_range = arm922_coherent_kern_range,
+	.coherent_user_range = arm922_coherent_user_range,
+	.flush_kern_dcache_area = arm922_flush_kern_dcache_area,
+	.dma_map_area = arm922_dma_map_area,
+	.dma_unmap_area = arm922_dma_unmap_area,
+	.dma_flush_range = arm922_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM925T
+void arm925_flush_icache_all(void);
+void arm925_flush_kern_cache_all(void);
+void arm925_flush_user_cache_all(void);
+void arm925_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm925_coherent_kern_range(unsigned long, unsigned long);
+int arm925_coherent_user_range(unsigned long, unsigned long);
+void arm925_flush_kern_dcache_area(void *, size_t);
+void arm925_dma_map_area(const void *, size_t, int);
+void arm925_dma_unmap_area(const void *, size_t, int);
+void arm925_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm925_cache_fns __initconst = {
+	.flush_icache_all = arm925_flush_icache_all,
+	.flush_kern_all = arm925_flush_kern_cache_all,
+	.flush_kern_louis = arm925_flush_kern_cache_all,
+	.flush_user_all = arm925_flush_user_cache_all,
+	.flush_user_range = arm925_flush_user_cache_range,
+	.coherent_kern_range = arm925_coherent_kern_range,
+	.coherent_user_range = arm925_coherent_user_range,
+	.flush_kern_dcache_area = arm925_flush_kern_dcache_area,
+	.dma_map_area = arm925_dma_map_area,
+	.dma_unmap_area = arm925_dma_unmap_area,
+	.dma_flush_range = arm925_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM926T
+void arm926_flush_icache_all(void);
+void arm926_flush_kern_cache_all(void);
+void arm926_flush_user_cache_all(void);
+void arm926_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm926_coherent_kern_range(unsigned long, unsigned long);
+int arm926_coherent_user_range(unsigned long, unsigned long);
+void arm926_flush_kern_dcache_area(void *, size_t);
+void arm926_dma_map_area(const void *, size_t, int);
+void arm926_dma_unmap_area(const void *, size_t, int);
+void arm926_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm926_cache_fns __initconst = {
+	.flush_icache_all = arm926_flush_icache_all,
+	.flush_kern_all = arm926_flush_kern_cache_all,
+	.flush_kern_louis = arm926_flush_kern_cache_all,
+	.flush_user_all = arm926_flush_user_cache_all,
+	.flush_user_range = arm926_flush_user_cache_range,
+	.coherent_kern_range = arm926_coherent_kern_range,
+	.coherent_user_range = arm926_coherent_user_range,
+	.flush_kern_dcache_area = arm926_flush_kern_dcache_area,
+	.dma_map_area = arm926_dma_map_area,
+	.dma_unmap_area = arm926_dma_unmap_area,
+	.dma_flush_range = arm926_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM940T
+void arm940_flush_icache_all(void);
+void arm940_flush_kern_cache_all(void);
+void arm940_flush_user_cache_all(void);
+void arm940_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm940_coherent_kern_range(unsigned long, unsigned long);
+int arm940_coherent_user_range(unsigned long, unsigned long);
+void arm940_flush_kern_dcache_area(void *, size_t);
+void arm940_dma_map_area(const void *, size_t, int);
+void arm940_dma_unmap_area(const void *, size_t, int);
+void arm940_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm940_cache_fns __initconst = {
+	.flush_icache_all = arm940_flush_icache_all,
+	.flush_kern_all = arm940_flush_kern_cache_all,
+	.flush_kern_louis = arm940_flush_kern_cache_all,
+	.flush_user_all = arm940_flush_user_cache_all,
+	.flush_user_range = arm940_flush_user_cache_range,
+	.coherent_kern_range = arm940_coherent_kern_range,
+	.coherent_user_range = arm940_coherent_user_range,
+	.flush_kern_dcache_area = arm940_flush_kern_dcache_area,
+	.dma_map_area = arm940_dma_map_area,
+	.dma_unmap_area = arm940_dma_unmap_area,
+	.dma_flush_range = arm940_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_ARM946E
+void arm946_flush_icache_all(void);
+void arm946_flush_kern_cache_all(void);
+void arm946_flush_user_cache_all(void);
+void arm946_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void arm946_coherent_kern_range(unsigned long, unsigned long);
+int arm946_coherent_user_range(unsigned long, unsigned long);
+void arm946_flush_kern_dcache_area(void *, size_t);
+void arm946_dma_map_area(const void *, size_t, int);
+void arm946_dma_unmap_area(const void *, size_t, int);
+void arm946_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns arm946_cache_fns __initconst = {
+	.flush_icache_all = arm946_flush_icache_all,
+	.flush_kern_all = arm946_flush_kern_cache_all,
+	.flush_kern_louis = arm946_flush_kern_cache_all,
+	.flush_user_all = arm946_flush_user_cache_all,
+	.flush_user_range = arm946_flush_user_cache_range,
+	.coherent_kern_range = arm946_coherent_kern_range,
+	.coherent_user_range = arm946_coherent_user_range,
+	.flush_kern_dcache_area = arm946_flush_kern_dcache_area,
+	.dma_map_area = arm946_dma_map_area,
+	.dma_unmap_area = arm946_dma_unmap_area,
+	.dma_flush_range = arm946_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_XSCALE
+void xscale_flush_icache_all(void);
+void xscale_flush_kern_cache_all(void);
+void xscale_flush_user_cache_all(void);
+void xscale_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void xscale_coherent_kern_range(unsigned long, unsigned long);
+int xscale_coherent_user_range(unsigned long, unsigned long);
+void xscale_flush_kern_dcache_area(void *, size_t);
+void xscale_dma_map_area(const void *, size_t, int);
+void xscale_dma_unmap_area(const void *, size_t, int);
+void xscale_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns xscale_cache_fns __initconst = {
+	.flush_icache_all = xscale_flush_icache_all,
+	.flush_kern_all = xscale_flush_kern_cache_all,
+	.flush_kern_louis = xscale_flush_kern_cache_all,
+	.flush_user_all = xscale_flush_user_cache_all,
+	.flush_user_range = xscale_flush_user_cache_range,
+	.coherent_kern_range = xscale_coherent_kern_range,
+	.coherent_user_range = xscale_coherent_user_range,
+	.flush_kern_dcache_area = xscale_flush_kern_dcache_area,
+	.dma_map_area = xscale_dma_map_area,
+	.dma_unmap_area = xscale_dma_unmap_area,
+	.dma_flush_range = xscale_dma_flush_range,
+};
+
+/* The 80200 A0 and A1 need a special quirk for dma_map_area() */
+void xscale_80200_A0_A1_dma_map_area(const void *, size_t, int);
+
+struct cpu_cache_fns xscale_80200_A0_A1_cache_fns __initconst = {
+	.flush_icache_all = xscale_flush_icache_all,
+	.flush_kern_all = xscale_flush_kern_cache_all,
+	.flush_kern_louis = xscale_flush_kern_cache_all,
+	.flush_user_all = xscale_flush_user_cache_all,
+	.flush_user_range = xscale_flush_user_cache_range,
+	.coherent_kern_range = xscale_coherent_kern_range,
+	.coherent_user_range = xscale_coherent_user_range,
+	.flush_kern_dcache_area = xscale_flush_kern_dcache_area,
+	.dma_map_area = xscale_80200_A0_A1_dma_map_area,
+	.dma_unmap_area = xscale_dma_unmap_area,
+	.dma_flush_range = xscale_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_XSC3
+void xsc3_flush_icache_all(void);
+void xsc3_flush_kern_cache_all(void);
+void xsc3_flush_user_cache_all(void);
+void xsc3_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void xsc3_coherent_kern_range(unsigned long, unsigned long);
+int xsc3_coherent_user_range(unsigned long, unsigned long);
+void xsc3_flush_kern_dcache_area(void *, size_t);
+void xsc3_dma_map_area(const void *, size_t, int);
+void xsc3_dma_unmap_area(const void *, size_t, int);
+void xsc3_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns xsc3_cache_fns __initconst = {
+	.flush_icache_all = xsc3_flush_icache_all,
+	.flush_kern_all = xsc3_flush_kern_cache_all,
+	.flush_kern_louis = xsc3_flush_kern_cache_all,
+	.flush_user_all = xsc3_flush_user_cache_all,
+	.flush_user_range = xsc3_flush_user_cache_range,
+	.coherent_kern_range = xsc3_coherent_kern_range,
+	.coherent_user_range = xsc3_coherent_user_range,
+	.flush_kern_dcache_area = xsc3_flush_kern_dcache_area,
+	.dma_map_area = xsc3_dma_map_area,
+	.dma_unmap_area = xsc3_dma_unmap_area,
+	.dma_flush_range = xsc3_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_MOHAWK
+void mohawk_flush_icache_all(void);
+void mohawk_flush_kern_cache_all(void);
+void mohawk_flush_user_cache_all(void);
+void mohawk_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void mohawk_coherent_kern_range(unsigned long, unsigned long);
+int mohawk_coherent_user_range(unsigned long, unsigned long);
+void mohawk_flush_kern_dcache_area(void *, size_t);
+void mohawk_dma_map_area(const void *, size_t, int);
+void mohawk_dma_unmap_area(const void *, size_t, int);
+void mohawk_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns mohawk_cache_fns __initconst = {
+	.flush_icache_all = mohawk_flush_icache_all,
+	.flush_kern_all = mohawk_flush_kern_cache_all,
+	.flush_kern_louis = mohawk_flush_kern_cache_all,
+	.flush_user_all = mohawk_flush_user_cache_all,
+	.flush_user_range = mohawk_flush_user_cache_range,
+	.coherent_kern_range = mohawk_coherent_kern_range,
+	.coherent_user_range = mohawk_coherent_user_range,
+	.flush_kern_dcache_area = mohawk_flush_kern_dcache_area,
+	.dma_map_area = mohawk_dma_map_area,
+	.dma_unmap_area = mohawk_dma_unmap_area,
+	.dma_flush_range = mohawk_dma_flush_range,
+};
+#endif
+
+#ifdef CONFIG_CPU_FEROCEON
+void feroceon_flush_icache_all(void);
+void feroceon_flush_kern_cache_all(void);
+void feroceon_flush_user_cache_all(void);
+void feroceon_flush_user_cache_range(unsigned long, unsigned long, unsigned int);
+void feroceon_coherent_kern_range(unsigned long, unsigned long);
+int feroceon_coherent_user_range(unsigned long, unsigned long);
+void feroceon_flush_kern_dcache_area(void *, size_t);
+void feroceon_dma_map_area(const void *, size_t, int);
+void feroceon_dma_unmap_area(const void *, size_t, int);
+void feroceon_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns feroceon_cache_fns __initconst = {
+	.flush_icache_all = feroceon_flush_icache_all,
+	.flush_kern_all = feroceon_flush_kern_cache_all,
+	.flush_kern_louis = feroceon_flush_kern_cache_all,
+	.flush_user_all = feroceon_flush_user_cache_all,
+	.flush_user_range = feroceon_flush_user_cache_range,
+	.coherent_kern_range = feroceon_coherent_kern_range,
+	.coherent_user_range = feroceon_coherent_user_range,
+	.flush_kern_dcache_area = feroceon_flush_kern_dcache_area,
+	.dma_map_area = feroceon_dma_map_area,
+	.dma_unmap_area = feroceon_dma_unmap_area,
+	.dma_flush_range = feroceon_dma_flush_range,
+};
+
+void feroceon_range_flush_kern_dcache_area(void *, size_t);
+void feroceon_range_dma_map_area(const void *, size_t, int);
+void feroceon_range_dma_flush_range(const void *, const void *);
+
+struct cpu_cache_fns feroceon_range_cache_fns __initconst = {
+	.flush_icache_all = feroceon_flush_icache_all,
+	.flush_kern_all = feroceon_flush_kern_cache_all,
+	.flush_kern_louis = feroceon_flush_kern_cache_all,
+	.flush_user_all = feroceon_flush_user_cache_all,
+	.flush_user_range = feroceon_flush_user_cache_range,
+	.coherent_kern_range = feroceon_coherent_kern_range,
+	.coherent_user_range = feroceon_coherent_user_range,
+	.flush_kern_dcache_area = feroceon_range_flush_kern_dcache_area,
+	.dma_map_area = feroceon_range_dma_map_area,
+	.dma_unmap_area = feroceon_dma_unmap_area,
+	.dma_flush_range = feroceon_range_dma_flush_range,
+};
+#endif
diff --git a/arch/arm/mm/proc-arm1020.S b/arch/arm/mm/proc-arm1020.S
index 379628e8ef4e..1e014cc5b4d1 100644
--- a/arch/arm/mm/proc-arm1020.S
+++ b/arch/arm/mm/proc-arm1020.S
@@ -357,12 +357,6 @@ SYM_TYPED_FUNC_START(arm1020_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm1020_dma_unmap_area)
 
-	.globl	arm1020_flush_kern_cache_louis
-	.equ	arm1020_flush_kern_cache_louis, arm1020_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm1020
-
 	.align	5
 ENTRY(cpu_arm1020_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
diff --git a/arch/arm/mm/proc-arm1020e.S b/arch/arm/mm/proc-arm1020e.S
index b5846fbea040..7d80761f207a 100644
--- a/arch/arm/mm/proc-arm1020e.S
+++ b/arch/arm/mm/proc-arm1020e.S
@@ -344,12 +344,6 @@ SYM_TYPED_FUNC_START(arm1020e_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm1020e_dma_unmap_area)
 
-	.globl	arm1020e_flush_kern_cache_louis
-	.equ	arm1020e_flush_kern_cache_louis, arm1020e_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm1020e
-
 	.align	5
 ENTRY(cpu_arm1020e_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
diff --git a/arch/arm/mm/proc-arm1022.S b/arch/arm/mm/proc-arm1022.S
index c40b268cc274..53b1541c50d8 100644
--- a/arch/arm/mm/proc-arm1022.S
+++ b/arch/arm/mm/proc-arm1022.S
@@ -343,12 +343,6 @@ SYM_TYPED_FUNC_START(arm1022_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm1022_dma_unmap_area)
 
-	.globl	arm1022_flush_kern_cache_louis
-	.equ	arm1022_flush_kern_cache_louis, arm1022_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm1022
-
 	.align	5
 ENTRY(cpu_arm1022_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
diff --git a/arch/arm/mm/proc-arm1026.S b/arch/arm/mm/proc-arm1026.S
index 7ef2c6d88dc0..6c6ea0357a77 100644
--- a/arch/arm/mm/proc-arm1026.S
+++ b/arch/arm/mm/proc-arm1026.S
@@ -338,12 +338,6 @@ SYM_TYPED_FUNC_START(arm1026_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm1026_dma_unmap_area)
 
-	.globl	arm1026_flush_kern_cache_louis
-	.equ	arm1026_flush_kern_cache_louis, arm1026_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm1026
-
 	.align	5
 ENTRY(cpu_arm1026_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
diff --git a/arch/arm/mm/proc-arm920.S b/arch/arm/mm/proc-arm920.S
index eb89a322a534..08a5bac0d89d 100644
--- a/arch/arm/mm/proc-arm920.S
+++ b/arch/arm/mm/proc-arm920.S
@@ -309,11 +309,6 @@ SYM_TYPED_FUNC_START(arm920_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm920_dma_unmap_area)
 
-	.globl	arm920_flush_kern_cache_louis
-	.equ	arm920_flush_kern_cache_louis, arm920_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm920
 #endif /* !CONFIG_CPU_DCACHE_WRITETHROUGH */
 
 
diff --git a/arch/arm/mm/proc-arm922.S b/arch/arm/mm/proc-arm922.S
index 035a1d1a26b0..8bcc0b913ba0 100644
--- a/arch/arm/mm/proc-arm922.S
+++ b/arch/arm/mm/proc-arm922.S
@@ -311,12 +311,6 @@ SYM_TYPED_FUNC_START(arm922_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm922_dma_unmap_area)
 
-	.globl	arm922_flush_kern_cache_louis
-	.equ	arm922_flush_kern_cache_louis, arm922_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm922
-
 #endif /* !CONFIG_CPU_DCACHE_WRITETHROUGH */
 
 ENTRY(cpu_arm922_dcache_clean_area)
diff --git a/arch/arm/mm/proc-arm925.S b/arch/arm/mm/proc-arm925.S
index 2510722647b4..d0d87f9705d3 100644
--- a/arch/arm/mm/proc-arm925.S
+++ b/arch/arm/mm/proc-arm925.S
@@ -366,12 +366,6 @@ SYM_TYPED_FUNC_START(arm925_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm925_dma_unmap_area)
 
-	.globl	arm925_flush_kern_cache_louis
-	.equ	arm925_flush_kern_cache_louis, arm925_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm925
-
 ENTRY(cpu_arm925_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
diff --git a/arch/arm/mm/proc-arm926.S b/arch/arm/mm/proc-arm926.S
index dac4a22369ba..6cb98b7a0fee 100644
--- a/arch/arm/mm/proc-arm926.S
+++ b/arch/arm/mm/proc-arm926.S
@@ -329,12 +329,6 @@ SYM_TYPED_FUNC_START(arm926_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm926_dma_unmap_area)
 
-	.globl	arm926_flush_kern_cache_louis
-	.equ	arm926_flush_kern_cache_louis, arm926_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm926
-
 ENTRY(cpu_arm926_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
diff --git a/arch/arm/mm/proc-arm940.S b/arch/arm/mm/proc-arm940.S
index 7c2268059536..527f1c044683 100644
--- a/arch/arm/mm/proc-arm940.S
+++ b/arch/arm/mm/proc-arm940.S
@@ -267,12 +267,6 @@ SYM_TYPED_FUNC_START(arm940_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm940_dma_unmap_area)
 
-	.globl	arm940_flush_kern_cache_louis
-	.equ	arm940_flush_kern_cache_louis, arm940_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm940
-
 	.type	__arm940_setup, #function
 __arm940_setup:
 	mov	r0, #0
diff --git a/arch/arm/mm/proc-arm946.S b/arch/arm/mm/proc-arm946.S
index 3955be1f4521..3155e819ae5f 100644
--- a/arch/arm/mm/proc-arm946.S
+++ b/arch/arm/mm/proc-arm946.S
@@ -310,12 +310,6 @@ SYM_TYPED_FUNC_START(arm946_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm946_dma_unmap_area)
 
-	.globl	arm946_flush_kern_cache_louis
-	.equ	arm946_flush_kern_cache_louis, arm946_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions arm946
-
 ENTRY(cpu_arm946_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
diff --git a/arch/arm/mm/proc-feroceon.S b/arch/arm/mm/proc-feroceon.S
index 9b1570ea6858..af9482b07a4f 100644
--- a/arch/arm/mm/proc-feroceon.S
+++ b/arch/arm/mm/proc-feroceon.S
@@ -412,33 +412,6 @@ SYM_TYPED_FUNC_START(feroceon_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(feroceon_dma_unmap_area)
 
-	.globl	feroceon_flush_kern_cache_louis
-	.equ	feroceon_flush_kern_cache_louis, feroceon_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions feroceon
-
-.macro range_alias basename
-	.globl feroceon_range_\basename
-	.type feroceon_range_\basename , %function
-	.equ feroceon_range_\basename , feroceon_\basename
-.endm
-
-/*
- * Most of the cache functions are unchanged for this case.
- * Export suitable alias symbols for the unchanged functions:
- */
-	range_alias flush_icache_all
-	range_alias flush_user_cache_all
-	range_alias flush_kern_cache_all
-	range_alias flush_kern_cache_louis
-	range_alias flush_user_cache_range
-	range_alias coherent_kern_range
-	range_alias coherent_user_range
-	range_alias dma_unmap_area
-
-	define_cache_functions feroceon_range
-
 	.align	5
 ENTRY(cpu_feroceon_dcache_clean_area)
 #if defined(CONFIG_CACHE_FEROCEON_L2) && \
diff --git a/arch/arm/mm/proc-macros.S b/arch/arm/mm/proc-macros.S
index c0acfeac3e84..e388c4cc0c44 100644
--- a/arch/arm/mm/proc-macros.S
+++ b/arch/arm/mm/proc-macros.S
@@ -320,24 +320,6 @@ ENTRY(\name\()_processor_functions)
 #endif
 .endm
 
-.macro define_cache_functions name:req
-	.align 2
-	.type	\name\()_cache_fns, #object
-ENTRY(\name\()_cache_fns)
-	.long	\name\()_flush_icache_all
-	.long	\name\()_flush_kern_cache_all
-	.long   \name\()_flush_kern_cache_louis
-	.long	\name\()_flush_user_cache_all
-	.long	\name\()_flush_user_cache_range
-	.long	\name\()_coherent_kern_range
-	.long	\name\()_coherent_user_range
-	.long	\name\()_flush_kern_dcache_area
-	.long	\name\()_dma_map_area
-	.long	\name\()_dma_unmap_area
-	.long	\name\()_dma_flush_range
-	.size	\name\()_cache_fns, . - \name\()_cache_fns
-.endm
-
 .macro globl_equ x, y
 	.globl	\x
 	.equ	\x, \y
diff --git a/arch/arm/mm/proc-mohawk.S b/arch/arm/mm/proc-mohawk.S
index 0a94cb0464d8..be3a1a997838 100644
--- a/arch/arm/mm/proc-mohawk.S
+++ b/arch/arm/mm/proc-mohawk.S
@@ -294,12 +294,6 @@ SYM_TYPED_FUNC_START(mohawk_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(mohawk_dma_unmap_area)
 
-	.globl	mohawk_flush_kern_cache_louis
-	.equ	mohawk_flush_kern_cache_louis, mohawk_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions mohawk
-
 ENTRY(cpu_mohawk_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
diff --git a/arch/arm/mm/proc-xsc3.S b/arch/arm/mm/proc-xsc3.S
index b2d907d748e9..7975f93b1e14 100644
--- a/arch/arm/mm/proc-xsc3.S
+++ b/arch/arm/mm/proc-xsc3.S
@@ -339,12 +339,6 @@ SYM_TYPED_FUNC_START(xsc3_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(xsc3_dma_unmap_area)
 
-	.globl	xsc3_flush_kern_cache_louis
-	.equ	xsc3_flush_kern_cache_louis, xsc3_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions xsc3
-
 ENTRY(cpu_xsc3_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean L1 D line
 	add	r0, r0, #CACHELINESIZE
diff --git a/arch/arm/mm/proc-xscale.S b/arch/arm/mm/proc-xscale.S
index 05d9ed952983..bbf1e94ba554 100644
--- a/arch/arm/mm/proc-xscale.S
+++ b/arch/arm/mm/proc-xscale.S
@@ -391,6 +391,20 @@ SYM_TYPED_FUNC_START(xscale_dma_map_area)
 	b	xscale_dma_flush_range
 SYM_FUNC_END(xscale_dma_map_area)
 
+/*
+ * On stepping A0/A1 of the 80200, invalidating D-cache by line doesn't
+ * clear the dirty bits, which means that if we invalidate a dirty line,
+ * the dirty data can still be written back to external memory later on.
+ *
+ * The recommended workaround is to always do a clean D-cache line before
+ * doing an invalidate D-cache line, so on the affected processors,
+ * dma_inv_range() is implemented as dma_flush_range().
+ *
+ * See erratum #25 of "Intel 80200 Processor Specification Update",
+ * revision January 22, 2003, available at:
+ *     http://www.intel.com/design/iio/specupdt/273415.htm
+ */
+
 /*
  *	dma_map_area(start, size, dir)
  *	- start	- kernel virtual start address
@@ -414,49 +428,6 @@ SYM_TYPED_FUNC_START(xscale_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(xscale_dma_unmap_area)
 
-	.globl	xscale_flush_kern_cache_louis
-	.equ	xscale_flush_kern_cache_louis, xscale_flush_kern_cache_all
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions xscale
-
-/*
- * On stepping A0/A1 of the 80200, invalidating D-cache by line doesn't
- * clear the dirty bits, which means that if we invalidate a dirty line,
- * the dirty data can still be written back to external memory later on.
- *
- * The recommended workaround is to always do a clean D-cache line before
- * doing an invalidate D-cache line, so on the affected processors,
- * dma_inv_range() is implemented as dma_flush_range().
- *
- * See erratum #25 of "Intel 80200 Processor Specification Update",
- * revision January 22, 2003, available at:
- *     http://www.intel.com/design/iio/specupdt/273415.htm
- */
-.macro a0_alias basename
-	.globl xscale_80200_A0_A1_\basename
-	.type xscale_80200_A0_A1_\basename , %function
-	.equ xscale_80200_A0_A1_\basename , xscale_\basename
-.endm
-
-/*
- * Most of the cache functions are unchanged for these processor revisions.
- * Export suitable alias symbols for the unchanged functions:
- */
-	a0_alias flush_icache_all
-	a0_alias flush_user_cache_all
-	a0_alias flush_kern_cache_all
-	a0_alias flush_kern_cache_louis
-	a0_alias flush_user_cache_range
-	a0_alias coherent_kern_range
-	a0_alias coherent_user_range
-	a0_alias flush_kern_dcache_area
-	a0_alias dma_flush_range
-	a0_alias dma_unmap_area
-
-	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	define_cache_functions xscale_80200_A0_A1
-
 ENTRY(cpu_xscale_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHELINESIZE

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 07/11] ARM: mm: Type-annotate all per-processor assembly routines
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (4 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 06/11] ARM: mm: Rewrite cacheflush vtables in CFI safe C Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-22  6:00   ` Arnd Bergmann
  2024-04-21 18:28 ` [PATCH v7 08/11] ARM: mm: Define prototypes for all per-processor calls Linus Walleij
                   ` (4 subsequent siblings)
  10 siblings, 1 reply; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

Type tag the remaining per-processor assembly using the CFI
symbol macros, in addition to those that were previously tagged
for cache maintenance calls.

This will be used to finally provide proper C prototypes for
all these calls as well so that CFI can be made to work.

Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/mm/proc-arm1020.S   | 24 ++++++++++------
 arch/arm/mm/proc-arm1020e.S  | 24 ++++++++++------
 arch/arm/mm/proc-arm1022.S   | 24 ++++++++++------
 arch/arm/mm/proc-arm1026.S   | 24 ++++++++++------
 arch/arm/mm/proc-arm720.S    | 25 +++++++++++------
 arch/arm/mm/proc-arm740.S    | 26 ++++++++++++-----
 arch/arm/mm/proc-arm7tdmi.S  | 34 +++++++++++++++--------
 arch/arm/mm/proc-arm920.S    | 31 ++++++++++++---------
 arch/arm/mm/proc-arm922.S    | 23 +++++++++------
 arch/arm/mm/proc-arm925.S    | 22 +++++++++------
 arch/arm/mm/proc-arm926.S    | 31 +++++++++++++--------
 arch/arm/mm/proc-arm940.S    | 21 +++++++++-----
 arch/arm/mm/proc-arm946.S    | 21 +++++++++-----
 arch/arm/mm/proc-arm9tdmi.S  | 26 ++++++++++++-----
 arch/arm/mm/proc-fa526.S     | 24 ++++++++++------
 arch/arm/mm/proc-feroceon.S  | 30 ++++++++++++--------
 arch/arm/mm/proc-mohawk.S    | 30 ++++++++++++--------
 arch/arm/mm/proc-sa110.S     | 23 +++++++++------
 arch/arm/mm/proc-sa1100.S    | 31 +++++++++++++--------
 arch/arm/mm/proc-v6.S        | 31 +++++++++++++--------
 arch/arm/mm/proc-v7-2level.S |  8 +++---
 arch/arm/mm/proc-v7-3level.S |  8 +++---
 arch/arm/mm/proc-v7.S        | 66 +++++++++++++++++++++++---------------------
 arch/arm/mm/proc-v7m.S       | 41 +++++++++++++--------------
 arch/arm/mm/proc-xsc3.S      | 30 ++++++++++++--------
 arch/arm/mm/proc-xscale.S    | 30 ++++++++++++--------
 26 files changed, 434 insertions(+), 274 deletions(-)

diff --git a/arch/arm/mm/proc-arm1020.S b/arch/arm/mm/proc-arm1020.S
index 1e014cc5b4d1..e6944ecd23ab 100644
--- a/arch/arm/mm/proc-arm1020.S
+++ b/arch/arm/mm/proc-arm1020.S
@@ -57,18 +57,20 @@
 /*
  * cpu_arm1020_proc_init()
  */
-ENTRY(cpu_arm1020_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm1020_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm1020_proc_init)
 
 /*
  * cpu_arm1020_proc_fin()
  */
-ENTRY(cpu_arm1020_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm1020_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000 		@ ...i............
 	bic	r0, r0, #0x000e 		@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm1020_proc_fin)
 
 /*
  * cpu_arm1020_reset(loc)
@@ -81,7 +83,7 @@ ENTRY(cpu_arm1020_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm1020_reset)
+SYM_TYPED_FUNC_START(cpu_arm1020_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -93,16 +95,17 @@ ENTRY(cpu_arm1020_reset)
 	bic	ip, ip, #0x1100 		@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm1020_reset)
+SYM_FUNC_END(cpu_arm1020_reset)
 	.popsection
 
 /*
  * cpu_arm1020_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm1020_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm1020_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_arm1020_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -358,7 +361,7 @@ SYM_TYPED_FUNC_START(arm1020_dma_unmap_area)
 SYM_FUNC_END(arm1020_dma_unmap_area)
 
 	.align	5
-ENTRY(cpu_arm1020_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm1020_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
 	mov	ip, #0
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
@@ -368,6 +371,7 @@ ENTRY(cpu_arm1020_dcache_clean_area)
 	bhi	1b
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm1020_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -379,7 +383,7 @@ ENTRY(cpu_arm1020_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm1020_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm1020_switch_mm)
 #ifdef CONFIG_MMU
 #ifndef CONFIG_CPU_DCACHE_DISABLE
 	mcr	p15, 0, r3, c7, c10, 4
@@ -407,14 +411,15 @@ ENTRY(cpu_arm1020_switch_mm)
 	mcr	p15, 0, r1, c8, c7, 0		@ invalidate I & D TLBs
 #endif /* CONFIG_MMU */
 	ret	lr
-        
+SYM_FUNC_END(cpu_arm1020_switch_mm)
+
 /*
  * cpu_arm1020_set_pte(ptep, pte)
  *
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm1020_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm1020_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -425,6 +430,7 @@ ENTRY(cpu_arm1020_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif /* CONFIG_MMU */
 	ret	lr
+SYM_FUNC_END(cpu_arm1020_set_pte_ext)
 
 	.type	__arm1020_setup, #function
 __arm1020_setup:
diff --git a/arch/arm/mm/proc-arm1020e.S b/arch/arm/mm/proc-arm1020e.S
index 7d80761f207a..5fae6e28c7a3 100644
--- a/arch/arm/mm/proc-arm1020e.S
+++ b/arch/arm/mm/proc-arm1020e.S
@@ -57,18 +57,20 @@
 /*
  * cpu_arm1020e_proc_init()
  */
-ENTRY(cpu_arm1020e_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm1020e_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm1020e_proc_init)
 
 /*
  * cpu_arm1020e_proc_fin()
  */
-ENTRY(cpu_arm1020e_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm1020e_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000 		@ ...i............
 	bic	r0, r0, #0x000e 		@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm1020e_proc_fin)
 
 /*
  * cpu_arm1020e_reset(loc)
@@ -81,7 +83,7 @@ ENTRY(cpu_arm1020e_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm1020e_reset)
+SYM_TYPED_FUNC_START(cpu_arm1020e_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -93,16 +95,17 @@ ENTRY(cpu_arm1020e_reset)
 	bic	ip, ip, #0x1100 		@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm1020e_reset)
+SYM_FUNC_END(cpu_arm1020e_reset)
 	.popsection
 
 /*
  * cpu_arm1020e_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm1020e_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm1020e_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_arm1020e_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -345,7 +348,7 @@ SYM_TYPED_FUNC_START(arm1020e_dma_unmap_area)
 SYM_FUNC_END(arm1020e_dma_unmap_area)
 
 	.align	5
-ENTRY(cpu_arm1020e_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm1020e_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
 	mov	ip, #0
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
@@ -354,6 +357,7 @@ ENTRY(cpu_arm1020e_dcache_clean_area)
 	bhi	1b
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm1020e_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -365,7 +369,7 @@ ENTRY(cpu_arm1020e_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm1020e_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm1020e_switch_mm)
 #ifdef CONFIG_MMU
 #ifndef CONFIG_CPU_DCACHE_DISABLE
 	mcr	p15, 0, r3, c7, c10, 4
@@ -392,14 +396,15 @@ ENTRY(cpu_arm1020e_switch_mm)
 	mcr	p15, 0, r1, c8, c7, 0		@ invalidate I & D TLBs
 #endif
 	ret	lr
-        
+SYM_FUNC_END(cpu_arm1020e_switch_mm)
+
 /*
  * cpu_arm1020e_set_pte(ptep, pte)
  *
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm1020e_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm1020e_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -408,6 +413,7 @@ ENTRY(cpu_arm1020e_set_pte_ext)
 #endif
 #endif /* CONFIG_MMU */
 	ret	lr
+SYM_FUNC_END(cpu_arm1020e_set_pte_ext)
 
 	.type	__arm1020e_setup, #function
 __arm1020e_setup:
diff --git a/arch/arm/mm/proc-arm1022.S b/arch/arm/mm/proc-arm1022.S
index 53b1541c50d8..05a7f14b2751 100644
--- a/arch/arm/mm/proc-arm1022.S
+++ b/arch/arm/mm/proc-arm1022.S
@@ -57,18 +57,20 @@
 /*
  * cpu_arm1022_proc_init()
  */
-ENTRY(cpu_arm1022_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm1022_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm1022_proc_init)
 
 /*
  * cpu_arm1022_proc_fin()
  */
-ENTRY(cpu_arm1022_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm1022_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000 		@ ...i............
 	bic	r0, r0, #0x000e 		@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm1022_proc_fin)
 
 /*
  * cpu_arm1022_reset(loc)
@@ -81,7 +83,7 @@ ENTRY(cpu_arm1022_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm1022_reset)
+SYM_TYPED_FUNC_START(cpu_arm1022_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -93,16 +95,17 @@ ENTRY(cpu_arm1022_reset)
 	bic	ip, ip, #0x1100 		@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm1022_reset)
+SYM_FUNC_END(cpu_arm1022_reset)
 	.popsection
 
 /*
  * cpu_arm1022_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm1022_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm1022_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_arm1022_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -344,7 +347,7 @@ SYM_TYPED_FUNC_START(arm1022_dma_unmap_area)
 SYM_FUNC_END(arm1022_dma_unmap_area)
 
 	.align	5
-ENTRY(cpu_arm1022_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm1022_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
 	mov	ip, #0
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
@@ -353,6 +356,7 @@ ENTRY(cpu_arm1022_dcache_clean_area)
 	bhi	1b
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm1022_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -364,7 +368,7 @@ ENTRY(cpu_arm1022_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm1022_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm1022_switch_mm)
 #ifdef CONFIG_MMU
 #ifndef CONFIG_CPU_DCACHE_DISABLE
 	mov	r1, #(CACHE_DSEGMENTS - 1) << 5	@ 16 segments
@@ -384,14 +388,15 @@ ENTRY(cpu_arm1022_switch_mm)
 	mcr	p15, 0, r1, c8, c7, 0		@ invalidate I & D TLBs
 #endif
 	ret	lr
-        
+SYM_FUNC_END(cpu_arm1022_switch_mm)
+
 /*
  * cpu_arm1022_set_pte_ext(ptep, pte, ext)
  *
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm1022_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm1022_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -400,6 +405,7 @@ ENTRY(cpu_arm1022_set_pte_ext)
 #endif
 #endif /* CONFIG_MMU */
 	ret	lr
+SYM_FUNC_END(cpu_arm1022_set_pte_ext)
 
 	.type	__arm1022_setup, #function
 __arm1022_setup:
diff --git a/arch/arm/mm/proc-arm1026.S b/arch/arm/mm/proc-arm1026.S
index 6c6ea0357a77..6800dd7c73f8 100644
--- a/arch/arm/mm/proc-arm1026.S
+++ b/arch/arm/mm/proc-arm1026.S
@@ -57,18 +57,20 @@
 /*
  * cpu_arm1026_proc_init()
  */
-ENTRY(cpu_arm1026_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm1026_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm1026_proc_init)
 
 /*
  * cpu_arm1026_proc_fin()
  */
-ENTRY(cpu_arm1026_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm1026_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000 		@ ...i............
 	bic	r0, r0, #0x000e 		@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm1026_proc_fin)
 
 /*
  * cpu_arm1026_reset(loc)
@@ -81,7 +83,7 @@ ENTRY(cpu_arm1026_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm1026_reset)
+SYM_TYPED_FUNC_START(cpu_arm1026_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -93,16 +95,17 @@ ENTRY(cpu_arm1026_reset)
 	bic	ip, ip, #0x1100 		@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm1026_reset)
+SYM_FUNC_END(cpu_arm1026_reset)
 	.popsection
 
 /*
  * cpu_arm1026_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm1026_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm1026_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_arm1026_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -339,7 +342,7 @@ SYM_TYPED_FUNC_START(arm1026_dma_unmap_area)
 SYM_FUNC_END(arm1026_dma_unmap_area)
 
 	.align	5
-ENTRY(cpu_arm1026_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm1026_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_DISABLE
 	mov	ip, #0
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
@@ -348,6 +351,7 @@ ENTRY(cpu_arm1026_dcache_clean_area)
 	bhi	1b
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm1026_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -359,7 +363,7 @@ ENTRY(cpu_arm1026_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm1026_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm1026_switch_mm)
 #ifdef CONFIG_MMU
 	mov	r1, #0
 #ifndef CONFIG_CPU_DCACHE_DISABLE
@@ -374,14 +378,15 @@ ENTRY(cpu_arm1026_switch_mm)
 	mcr	p15, 0, r1, c8, c7, 0		@ invalidate I & D TLBs
 #endif
 	ret	lr
-        
+SYM_FUNC_END(cpu_arm1026_switch_mm)
+
 /*
  * cpu_arm1026_set_pte_ext(ptep, pte, ext)
  *
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm1026_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm1026_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -390,6 +395,7 @@ ENTRY(cpu_arm1026_set_pte_ext)
 #endif
 #endif /* CONFIG_MMU */
 	ret	lr
+SYM_FUNC_END(cpu_arm1026_set_pte_ext)
 
 	.type	__arm1026_setup, #function
 __arm1026_setup:
diff --git a/arch/arm/mm/proc-arm720.S b/arch/arm/mm/proc-arm720.S
index 3b687e6dd9fd..59732c334e1d 100644
--- a/arch/arm/mm/proc-arm720.S
+++ b/arch/arm/mm/proc-arm720.S
@@ -20,6 +20,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
@@ -35,24 +36,30 @@
  *
  * Notes   : This processor does not require these
  */
-ENTRY(cpu_arm720_dcache_clean_area)
-ENTRY(cpu_arm720_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm720_dcache_clean_area)
 		ret	lr
+SYM_FUNC_END(cpu_arm720_dcache_clean_area)
 
-ENTRY(cpu_arm720_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm720_proc_init)
+		ret	lr
+SYM_FUNC_END(cpu_arm720_proc_init)
+
+SYM_TYPED_FUNC_START(cpu_arm720_proc_fin)
 		mrc	p15, 0, r0, c1, c0, 0
 		bic	r0, r0, #0x1000			@ ...i............
 		bic	r0, r0, #0x000e			@ ............wca.
 		mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 		ret	lr
+SYM_FUNC_END(cpu_arm720_proc_fin)
 
 /*
  * Function: arm720_proc_do_idle(void)
  * Params  : r0 = unused
  * Purpose : put the processor in proper idle mode
  */
-ENTRY(cpu_arm720_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm720_do_idle)
 		ret	lr
+SYM_FUNC_END(cpu_arm720_do_idle)
 
 /*
  * Function: arm720_switch_mm(unsigned long pgd_phys)
@@ -60,7 +67,7 @@ ENTRY(cpu_arm720_do_idle)
  * Purpose : Perform a task switch, saving the old process' state and restoring
  *	     the new.
  */
-ENTRY(cpu_arm720_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm720_switch_mm)
 #ifdef CONFIG_MMU
 		mov	r1, #0
 		mcr	p15, 0, r1, c7, c7, 0		@ invalidate cache
@@ -68,6 +75,7 @@ ENTRY(cpu_arm720_switch_mm)
 		mcr	p15, 0, r1, c8, c7, 0		@ flush TLB (v4)
 #endif
 		ret	lr
+SYM_FUNC_END(cpu_arm720_switch_mm)
 
 /*
  * Function: arm720_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext)
@@ -76,11 +84,12 @@ ENTRY(cpu_arm720_switch_mm)
  * Purpose : Set a PTE and flush it out of any WB cache
  */
 	.align	5
-ENTRY(cpu_arm720_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm720_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext wc_disable=0
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm720_set_pte_ext)
 
 /*
  * Function: arm720_reset
@@ -88,7 +97,7 @@ ENTRY(cpu_arm720_set_pte_ext)
  * Notes   : This sets up everything for a reset
  */
 		.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm720_reset)
+SYM_TYPED_FUNC_START(cpu_arm720_reset)
 		mov	ip, #0
 		mcr	p15, 0, ip, c7, c7, 0		@ invalidate cache
 #ifdef CONFIG_MMU
@@ -99,7 +108,7 @@ ENTRY(cpu_arm720_reset)
 		bic	ip, ip, #0x2100			@ ..v....s........
 		mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 		ret	r0
-ENDPROC(cpu_arm720_reset)
+SYM_FUNC_END(cpu_arm720_reset)
 		.popsection
 
 	.type	__arm710_setup, #function
diff --git a/arch/arm/mm/proc-arm740.S b/arch/arm/mm/proc-arm740.S
index f2ec3bc60874..78854df63964 100644
--- a/arch/arm/mm/proc-arm740.S
+++ b/arch/arm/mm/proc-arm740.S
@@ -6,6 +6,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
@@ -24,21 +25,32 @@
  *
  * These are not required.
  */
-ENTRY(cpu_arm740_proc_init)
-ENTRY(cpu_arm740_do_idle)
-ENTRY(cpu_arm740_dcache_clean_area)
-ENTRY(cpu_arm740_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm740_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm740_proc_init)
+
+SYM_TYPED_FUNC_START(cpu_arm740_do_idle)
+	ret	lr
+SYM_FUNC_END(cpu_arm740_do_idle)
+
+SYM_TYPED_FUNC_START(cpu_arm740_dcache_clean_area)
+	ret	lr
+SYM_FUNC_END(cpu_arm740_dcache_clean_area)
+
+SYM_TYPED_FUNC_START(cpu_arm740_switch_mm)
+	ret	lr
+SYM_FUNC_END(cpu_arm740_switch_mm)
 
 /*
  * cpu_arm740_proc_fin()
  */
-ENTRY(cpu_arm740_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm740_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0
 	bic	r0, r0, #0x3f000000		@ bank/f/lock/s
 	bic	r0, r0, #0x0000000c		@ w-buffer/cache
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm740_proc_fin)
 
 /*
  * cpu_arm740_reset(loc)
@@ -46,14 +58,14 @@ ENTRY(cpu_arm740_proc_fin)
  * Notes   : This sets up everything for a reset
  */
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm740_reset)
+SYM_TYPED_FUNC_START(cpu_arm740_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c0, 0		@ invalidate cache
 	mrc	p15, 0, ip, c1, c0, 0		@ get ctrl register
 	bic	ip, ip, #0x0000000c		@ ............wc..
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm740_reset)
+SYM_FUNC_END(cpu_arm740_reset)
 	.popsection
 
 	.type	__arm740_setup, #function
diff --git a/arch/arm/mm/proc-arm7tdmi.S b/arch/arm/mm/proc-arm7tdmi.S
index 01bbe7576c1c..baa3d4472147 100644
--- a/arch/arm/mm/proc-arm7tdmi.S
+++ b/arch/arm/mm/proc-arm7tdmi.S
@@ -6,6 +6,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
@@ -23,18 +24,29 @@
  * cpu_arm7tdmi_switch_mm()
  *
  * These are not required.
- */
-ENTRY(cpu_arm7tdmi_proc_init)
-ENTRY(cpu_arm7tdmi_do_idle)
-ENTRY(cpu_arm7tdmi_dcache_clean_area)
-ENTRY(cpu_arm7tdmi_switch_mm)
-		ret	lr
+*/
+SYM_TYPED_FUNC_START(cpu_arm7tdmi_proc_init)
+	ret lr
+SYM_FUNC_END(cpu_arm7tdmi_proc_init)
+
+SYM_TYPED_FUNC_START(cpu_arm7tdmi_do_idle)
+	ret lr
+SYM_FUNC_END(cpu_arm7tdmi_do_idle)
+
+SYM_TYPED_FUNC_START(cpu_arm7tdmi_dcache_clean_area)
+	ret lr
+SYM_FUNC_END(cpu_arm7tdmi_dcache_clean_area)
+
+SYM_TYPED_FUNC_START(cpu_arm7tdmi_switch_mm)
+	ret	lr
+SYM_FUNC_END(cpu_arm7tdmi_switch_mm)
 
 /*
  * cpu_arm7tdmi_proc_fin()
- */
-ENTRY(cpu_arm7tdmi_proc_fin)
-		ret	lr
+*/
+SYM_TYPED_FUNC_START(cpu_arm7tdmi_proc_fin)
+	ret	lr
+SYM_FUNC_END(cpu_arm7tdmi_proc_fin)
 
 /*
  * Function: cpu_arm7tdmi_reset(loc)
@@ -42,9 +54,9 @@ ENTRY(cpu_arm7tdmi_proc_fin)
  * Purpose : Sets up everything for a reset and jump to the location for soft reset.
  */
 		.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm7tdmi_reset)
+SYM_TYPED_FUNC_START(cpu_arm7tdmi_reset)
 		ret	r0
-ENDPROC(cpu_arm7tdmi_reset)
+SYM_FUNC_END(cpu_arm7tdmi_reset)
 		.popsection
 
 		.type	__arm7tdmi_setup, #function
diff --git a/arch/arm/mm/proc-arm920.S b/arch/arm/mm/proc-arm920.S
index 08a5bac0d89d..a1eec82070e5 100644
--- a/arch/arm/mm/proc-arm920.S
+++ b/arch/arm/mm/proc-arm920.S
@@ -49,18 +49,20 @@
 /*
  * cpu_arm920_proc_init()
  */
-ENTRY(cpu_arm920_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm920_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm920_proc_init)
 
 /*
  * cpu_arm920_proc_fin()
  */
-ENTRY(cpu_arm920_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm920_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x000e			@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm920_proc_fin)
 
 /*
  * cpu_arm920_reset(loc)
@@ -73,7 +75,7 @@ ENTRY(cpu_arm920_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm920_reset)
+SYM_TYPED_FUNC_START(cpu_arm920_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -85,17 +87,17 @@ ENTRY(cpu_arm920_reset)
 	bic	ip, ip, #0x1100			@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm920_reset)
+SYM_FUNC_END(cpu_arm920_reset)
 	.popsection
 
 /*
  * cpu_arm920_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm920_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm920_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
-
+SYM_FUNC_END(cpu_arm920_do_idle)
 
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 
@@ -312,12 +314,13 @@ SYM_FUNC_END(arm920_dma_unmap_area)
 #endif /* !CONFIG_CPU_DCACHE_WRITETHROUGH */
 
 
-ENTRY(cpu_arm920_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm920_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	subs	r1, r1, #CACHE_DLINESIZE
 	bhi	1b
 	ret	lr
+SYM_FUNC_END(cpu_arm920_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -329,7 +332,7 @@ ENTRY(cpu_arm920_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm920_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm920_switch_mm)
 #ifdef CONFIG_MMU
 	mov	ip, #0
 #ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -353,6 +356,7 @@ ENTRY(cpu_arm920_switch_mm)
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I & D TLBs
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm920_switch_mm)
 
 /*
  * cpu_arm920_set_pte(ptep, pte, ext)
@@ -360,7 +364,7 @@ ENTRY(cpu_arm920_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm920_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm920_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -368,21 +372,22 @@ ENTRY(cpu_arm920_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm920_set_pte_ext)
 
 /* Suspend/resume support: taken from arch/arm/plat-s3c24xx/sleep.S */
 .globl	cpu_arm920_suspend_size
 .equ	cpu_arm920_suspend_size, 4 * 3
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_arm920_do_suspend)
+SYM_TYPED_FUNC_START(cpu_arm920_do_suspend)
 	stmfd	sp!, {r4 - r6, lr}
 	mrc	p15, 0, r4, c13, c0, 0	@ PID
 	mrc	p15, 0, r5, c3, c0, 0	@ Domain ID
 	mrc	p15, 0, r6, c1, c0, 0	@ Control register
 	stmia	r0, {r4 - r6}
 	ldmfd	sp!, {r4 - r6, pc}
-ENDPROC(cpu_arm920_do_suspend)
+SYM_FUNC_END(cpu_arm920_do_suspend)
 
-ENTRY(cpu_arm920_do_resume)
+SYM_TYPED_FUNC_START(cpu_arm920_do_resume)
 	mov	ip, #0
 	mcr	p15, 0, ip, c8, c7, 0	@ invalidate I+D TLBs
 	mcr	p15, 0, ip, c7, c7, 0	@ invalidate I+D caches
@@ -392,7 +397,7 @@ ENTRY(cpu_arm920_do_resume)
 	mcr	p15, 0, r1, c2, c0, 0	@ TTB address
 	mov	r0, r6			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_arm920_do_resume)
+SYM_FUNC_END(cpu_arm920_do_resume)
 #endif
 
 	.type	__arm920_setup, #function
diff --git a/arch/arm/mm/proc-arm922.S b/arch/arm/mm/proc-arm922.S
index 8bcc0b913ba0..aeafac5143f6 100644
--- a/arch/arm/mm/proc-arm922.S
+++ b/arch/arm/mm/proc-arm922.S
@@ -51,18 +51,20 @@
 /*
  * cpu_arm922_proc_init()
  */
-ENTRY(cpu_arm922_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm922_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm922_proc_init)
 
 /*
  * cpu_arm922_proc_fin()
  */
-ENTRY(cpu_arm922_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm922_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x000e			@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm922_proc_fin)
 
 /*
  * cpu_arm922_reset(loc)
@@ -75,7 +77,7 @@ ENTRY(cpu_arm922_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm922_reset)
+SYM_TYPED_FUNC_START(cpu_arm922_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -87,17 +89,17 @@ ENTRY(cpu_arm922_reset)
 	bic	ip, ip, #0x1100			@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm922_reset)
+SYM_FUNC_END(cpu_arm922_reset)
 	.popsection
 
 /*
  * cpu_arm922_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm922_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm922_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
-
+SYM_FUNC_END(cpu_arm922_do_idle)
 
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 
@@ -313,7 +315,7 @@ SYM_FUNC_END(arm922_dma_unmap_area)
 
 #endif /* !CONFIG_CPU_DCACHE_WRITETHROUGH */
 
-ENTRY(cpu_arm922_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm922_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
@@ -321,6 +323,7 @@ ENTRY(cpu_arm922_dcache_clean_area)
 	bhi	1b
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm922_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -332,7 +335,7 @@ ENTRY(cpu_arm922_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm922_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm922_switch_mm)
 #ifdef CONFIG_MMU
 	mov	ip, #0
 #ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -356,6 +359,7 @@ ENTRY(cpu_arm922_switch_mm)
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I & D TLBs
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm922_switch_mm)
 
 /*
  * cpu_arm922_set_pte_ext(ptep, pte, ext)
@@ -363,7 +367,7 @@ ENTRY(cpu_arm922_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm922_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm922_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -371,6 +375,7 @@ ENTRY(cpu_arm922_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif /* CONFIG_MMU */
 	ret	lr
+SYM_FUNC_END(cpu_arm922_set_pte_ext)
 
 	.type	__arm922_setup, #function
 __arm922_setup:
diff --git a/arch/arm/mm/proc-arm925.S b/arch/arm/mm/proc-arm925.S
index d0d87f9705d3..191f4fa606c7 100644
--- a/arch/arm/mm/proc-arm925.S
+++ b/arch/arm/mm/proc-arm925.S
@@ -72,18 +72,20 @@
 /*
  * cpu_arm925_proc_init()
  */
-ENTRY(cpu_arm925_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm925_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm925_proc_init)
 
 /*
  * cpu_arm925_proc_fin()
  */
-ENTRY(cpu_arm925_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm925_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x000e			@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm925_proc_fin)
 
 /*
  * cpu_arm925_reset(loc)
@@ -96,14 +98,14 @@ ENTRY(cpu_arm925_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm925_reset)
+SYM_TYPED_FUNC_START(cpu_arm925_reset)
 	/* Send software reset to MPU and DSP */
 	mov	ip, #0xff000000
 	orr	ip, ip, #0x00fe0000
 	orr	ip, ip, #0x0000ce00
 	mov	r4, #1
 	strh	r4, [ip, #0x10]
-ENDPROC(cpu_arm925_reset)
+SYM_FUNC_END(cpu_arm925_reset)
 	.popsection
 
 	mov	ip, #0
@@ -124,7 +126,7 @@ ENDPROC(cpu_arm925_reset)
  * Called with IRQs disabled
  */
 	.align	10
-ENTRY(cpu_arm925_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm925_do_idle)
 	mov	r0, #0
 	mrc	p15, 0, r1, c1, c0, 0		@ Read control register
 	mcr	p15, 0, r0, c7, c10, 4		@ Drain write buffer
@@ -133,6 +135,7 @@ ENTRY(cpu_arm925_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	mcr	p15, 0, r1, c1, c0, 0		@ Restore ICache enable
 	ret	lr
+SYM_FUNC_END(cpu_arm925_do_idle)
 
 /*
  *	flush_icache_all()
@@ -366,7 +369,7 @@ SYM_TYPED_FUNC_START(arm925_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm925_dma_unmap_area)
 
-ENTRY(cpu_arm925_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm925_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
@@ -375,6 +378,7 @@ ENTRY(cpu_arm925_dcache_clean_area)
 #endif
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 	ret	lr
+SYM_FUNC_END(cpu_arm925_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -386,7 +390,7 @@ ENTRY(cpu_arm925_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm925_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm925_switch_mm)
 #ifdef CONFIG_MMU
 	mov	ip, #0
 #ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -404,6 +408,7 @@ ENTRY(cpu_arm925_switch_mm)
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I & D TLBs
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm925_switch_mm)
 
 /*
  * cpu_arm925_set_pte_ext(ptep, pte, ext)
@@ -411,7 +416,7 @@ ENTRY(cpu_arm925_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm925_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm925_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -421,6 +426,7 @@ ENTRY(cpu_arm925_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif /* CONFIG_MMU */
 	ret	lr
+SYM_FUNC_END(cpu_arm925_set_pte_ext)
 
 	.type	__arm925_setup, #function
 __arm925_setup:
diff --git a/arch/arm/mm/proc-arm926.S b/arch/arm/mm/proc-arm926.S
index 6cb98b7a0fee..3bf1d4072283 100644
--- a/arch/arm/mm/proc-arm926.S
+++ b/arch/arm/mm/proc-arm926.S
@@ -41,18 +41,20 @@
 /*
  * cpu_arm926_proc_init()
  */
-ENTRY(cpu_arm926_proc_init)
+SYM_TYPED_FUNC_START(cpu_arm926_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm926_proc_init)
 
 /*
  * cpu_arm926_proc_fin()
  */
-ENTRY(cpu_arm926_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm926_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x000e			@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm926_proc_fin)
 
 /*
  * cpu_arm926_reset(loc)
@@ -65,7 +67,7 @@ ENTRY(cpu_arm926_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm926_reset)
+SYM_TYPED_FUNC_START(cpu_arm926_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -77,7 +79,7 @@ ENTRY(cpu_arm926_reset)
 	bic	ip, ip, #0x1100			@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm926_reset)
+SYM_FUNC_END(cpu_arm926_reset)
 	.popsection
 
 /*
@@ -86,7 +88,7 @@ ENDPROC(cpu_arm926_reset)
  * Called with IRQs disabled
  */
 	.align	10
-ENTRY(cpu_arm926_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm926_do_idle)
 	mov	r0, #0
 	mrc	p15, 0, r1, c1, c0, 0		@ Read control register
 	mcr	p15, 0, r0, c7, c10, 4		@ Drain write buffer
@@ -99,6 +101,7 @@ ENTRY(cpu_arm926_do_idle)
 	mcr	p15, 0, r1, c1, c0, 0		@ Restore ICache enable
 	msr	cpsr_c, r3			@ Restore FIQ state
 	ret	lr
+SYM_FUNC_END(cpu_arm926_do_idle)
 
 /*
  *	flush_icache_all()
@@ -329,7 +332,7 @@ SYM_TYPED_FUNC_START(arm926_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm926_dma_unmap_area)
 
-ENTRY(cpu_arm926_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm926_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
@@ -338,6 +341,7 @@ ENTRY(cpu_arm926_dcache_clean_area)
 #endif
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 	ret	lr
+SYM_FUNC_END(cpu_arm926_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -349,7 +353,8 @@ ENTRY(cpu_arm926_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_arm926_switch_mm)
+
+SYM_TYPED_FUNC_START(cpu_arm926_switch_mm)
 #ifdef CONFIG_MMU
 	mov	ip, #0
 #ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -365,6 +370,7 @@ ENTRY(cpu_arm926_switch_mm)
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I & D TLBs
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm926_switch_mm)
 
 /*
  * cpu_arm926_set_pte_ext(ptep, pte, ext)
@@ -372,7 +378,7 @@ ENTRY(cpu_arm926_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_arm926_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_arm926_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -382,21 +388,22 @@ ENTRY(cpu_arm926_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_arm926_set_pte_ext)
 
 /* Suspend/resume support: taken from arch/arm/plat-s3c24xx/sleep.S */
 .globl	cpu_arm926_suspend_size
 .equ	cpu_arm926_suspend_size, 4 * 3
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_arm926_do_suspend)
+SYM_TYPED_FUNC_START(cpu_arm926_do_suspend)
 	stmfd	sp!, {r4 - r6, lr}
 	mrc	p15, 0, r4, c13, c0, 0	@ PID
 	mrc	p15, 0, r5, c3, c0, 0	@ Domain ID
 	mrc	p15, 0, r6, c1, c0, 0	@ Control register
 	stmia	r0, {r4 - r6}
 	ldmfd	sp!, {r4 - r6, pc}
-ENDPROC(cpu_arm926_do_suspend)
+SYM_FUNC_END(cpu_arm926_do_suspend)
 
-ENTRY(cpu_arm926_do_resume)
+SYM_TYPED_FUNC_START(cpu_arm926_do_resume)
 	mov	ip, #0
 	mcr	p15, 0, ip, c8, c7, 0	@ invalidate I+D TLBs
 	mcr	p15, 0, ip, c7, c7, 0	@ invalidate I+D caches
@@ -406,7 +413,7 @@ ENTRY(cpu_arm926_do_resume)
 	mcr	p15, 0, r1, c2, c0, 0	@ TTB address
 	mov	r0, r6			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_arm926_do_resume)
+SYM_FUNC_END(cpu_arm926_do_resume)
 #endif
 
 	.type	__arm926_setup, #function
diff --git a/arch/arm/mm/proc-arm940.S b/arch/arm/mm/proc-arm940.S
index 527f1c044683..cd95fca4656f 100644
--- a/arch/arm/mm/proc-arm940.S
+++ b/arch/arm/mm/proc-arm940.S
@@ -26,19 +26,24 @@
  *
  * These are not required.
  */
-ENTRY(cpu_arm940_proc_init)
-ENTRY(cpu_arm940_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm940_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm940_proc_init)
+
+SYM_TYPED_FUNC_START(cpu_arm940_switch_mm)
+	ret	lr
+SYM_FUNC_END(cpu_arm940_switch_mm)
 
 /*
  * cpu_arm940_proc_fin()
  */
-ENTRY(cpu_arm940_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm940_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x00001000		@ i-cache
 	bic	r0, r0, #0x00000004		@ d-cache
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm940_proc_fin)
 
 /*
  * cpu_arm940_reset(loc)
@@ -46,7 +51,7 @@ ENTRY(cpu_arm940_proc_fin)
  * Notes   : This sets up everything for a reset
  */
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm940_reset)
+SYM_TYPED_FUNC_START(cpu_arm940_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c5, 0		@ flush I cache
 	mcr	p15, 0, ip, c7, c6, 0		@ flush D cache
@@ -56,16 +61,17 @@ ENTRY(cpu_arm940_reset)
 	bic	ip, ip, #0x00001000		@ i-cache
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm940_reset)
+SYM_FUNC_END(cpu_arm940_reset)
 	.popsection
 
 /*
  * cpu_arm940_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm940_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm940_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_arm940_do_idle)
 
 /*
  *	flush_icache_all()
@@ -202,7 +208,7 @@ arm940_dma_inv_range:
  *	- end	- virtual end address
  */
 arm940_dma_clean_range:
-ENTRY(cpu_arm940_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm940_dcache_clean_area)
 	mov	ip, #0
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 	mov	r1, #(CACHE_DSEGMENTS - 1) << 4	@ 4 segments
@@ -215,6 +221,7 @@ ENTRY(cpu_arm940_dcache_clean_area)
 #endif
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
 	ret	lr
+SYM_FUNC_END(cpu_arm940_dcache_clean_area)
 
 /*
  *	dma_flush_range(start, end)
diff --git a/arch/arm/mm/proc-arm946.S b/arch/arm/mm/proc-arm946.S
index 3155e819ae5f..7df7c6e5598a 100644
--- a/arch/arm/mm/proc-arm946.S
+++ b/arch/arm/mm/proc-arm946.S
@@ -33,19 +33,24 @@
  *
  * These are not required.
  */
-ENTRY(cpu_arm946_proc_init)
-ENTRY(cpu_arm946_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm946_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_arm946_proc_init)
+
+SYM_TYPED_FUNC_START(cpu_arm946_switch_mm)
+	ret	lr
+SYM_FUNC_END(cpu_arm946_switch_mm)
 
 /*
  * cpu_arm946_proc_fin()
  */
-ENTRY(cpu_arm946_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm946_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x00001000		@ i-cache
 	bic	r0, r0, #0x00000004		@ d-cache
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_arm946_proc_fin)
 
 /*
  * cpu_arm946_reset(loc)
@@ -53,7 +58,7 @@ ENTRY(cpu_arm946_proc_fin)
  * Notes   : This sets up everything for a reset
  */
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm946_reset)
+SYM_TYPED_FUNC_START(cpu_arm946_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c5, 0		@ flush I cache
 	mcr	p15, 0, ip, c7, c6, 0		@ flush D cache
@@ -63,16 +68,17 @@ ENTRY(cpu_arm946_reset)
 	bic	ip, ip, #0x00001000		@ i-cache
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_arm946_reset)
+SYM_FUNC_END(cpu_arm946_reset)
 	.popsection
 
 /*
  * cpu_arm946_do_idle()
  */
 	.align	5
-ENTRY(cpu_arm946_do_idle)
+SYM_TYPED_FUNC_START(cpu_arm946_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_arm946_do_idle)
 
 /*
  *	flush_icache_all()
@@ -310,7 +316,7 @@ SYM_TYPED_FUNC_START(arm946_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(arm946_dma_unmap_area)
 
-ENTRY(cpu_arm946_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_arm946_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
@@ -319,6 +325,7 @@ ENTRY(cpu_arm946_dcache_clean_area)
 #endif
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 	ret	lr
+SYM_FUNC_END(cpu_arm946_dcache_clean_area)
 
 	.type	__arm946_setup, #function
 __arm946_setup:
diff --git a/arch/arm/mm/proc-arm9tdmi.S b/arch/arm/mm/proc-arm9tdmi.S
index a054c0e9c034..c480a8400eff 100644
--- a/arch/arm/mm/proc-arm9tdmi.S
+++ b/arch/arm/mm/proc-arm9tdmi.S
@@ -6,6 +6,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
@@ -24,17 +25,28 @@
  *
  * These are not required.
  */
-ENTRY(cpu_arm9tdmi_proc_init)
-ENTRY(cpu_arm9tdmi_do_idle)
-ENTRY(cpu_arm9tdmi_dcache_clean_area)
-ENTRY(cpu_arm9tdmi_switch_mm)
+SYM_TYPED_FUNC_START(cpu_arm9tdmi_proc_init)
 		ret	lr
+SYM_FUNC_END(cpu_arm9tdmi_proc_init)
+
+SYM_TYPED_FUNC_START(cpu_arm9tdmi_do_idle)
+		ret	lr
+SYM_FUNC_END(cpu_arm9tdmi_do_idle)
+
+SYM_TYPED_FUNC_START(cpu_arm9tdmi_dcache_clean_area)
+		ret	lr
+SYM_FUNC_END(cpu_arm9tdmi_dcache_clean_area)
+
+SYM_TYPED_FUNC_START(cpu_arm9tdmi_switch_mm)
+		ret	lr
+SYM_FUNC_END(cpu_arm9tdmi_switch_mm)
 
 /*
  * cpu_arm9tdmi_proc_fin()
  */
-ENTRY(cpu_arm9tdmi_proc_fin)
+SYM_TYPED_FUNC_START(cpu_arm9tdmi_proc_fin)
 		ret	lr
+SYM_FUNC_END(cpu_arm9tdmi_proc_fin)
 
 /*
  * Function: cpu_arm9tdmi_reset(loc)
@@ -42,9 +54,9 @@ ENTRY(cpu_arm9tdmi_proc_fin)
  * Purpose : Sets up everything for a reset and jump to the location for soft reset.
  */
 		.pushsection	.idmap.text, "ax"
-ENTRY(cpu_arm9tdmi_reset)
+SYM_TYPED_FUNC_START(cpu_arm9tdmi_reset)
 		ret	r0
-ENDPROC(cpu_arm9tdmi_reset)
+SYM_FUNC_END(cpu_arm9tdmi_reset)
 		.popsection
 
 		.type	__arm9tdmi_setup, #function
diff --git a/arch/arm/mm/proc-fa526.S b/arch/arm/mm/proc-fa526.S
index 2c73e0d47d08..7c16ccac8a05 100644
--- a/arch/arm/mm/proc-fa526.S
+++ b/arch/arm/mm/proc-fa526.S
@@ -11,6 +11,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
 #include <asm/hwcap.h>
@@ -26,13 +27,14 @@
 /*
  * cpu_fa526_proc_init()
  */
-ENTRY(cpu_fa526_proc_init)
+SYM_TYPED_FUNC_START(cpu_fa526_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_fa526_proc_init)
 
 /*
  * cpu_fa526_proc_fin()
  */
-ENTRY(cpu_fa526_proc_fin)
+SYM_TYPED_FUNC_START(cpu_fa526_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x000e			@ ............wca.
@@ -40,6 +42,7 @@ ENTRY(cpu_fa526_proc_fin)
 	nop
 	nop
 	ret	lr
+SYM_FUNC_END(cpu_fa526_proc_fin)
 
 /*
  * cpu_fa526_reset(loc)
@@ -52,7 +55,7 @@ ENTRY(cpu_fa526_proc_fin)
  */
 	.align	4
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_fa526_reset)
+SYM_TYPED_FUNC_START(cpu_fa526_reset)
 /* TODO: Use CP8 if possible... */
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
@@ -68,24 +71,25 @@ ENTRY(cpu_fa526_reset)
 	nop
 	nop
 	ret	r0
-ENDPROC(cpu_fa526_reset)
+SYM_FUNC_END(cpu_fa526_reset)
 	.popsection
 
 /*
  * cpu_fa526_do_idle()
  */
 	.align	4
-ENTRY(cpu_fa526_do_idle)
+SYM_TYPED_FUNC_START(cpu_fa526_do_idle)
 	ret	lr
+SYM_FUNC_END(cpu_fa526_do_idle)
 
-
-ENTRY(cpu_fa526_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_fa526_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	subs	r1, r1, #CACHE_DLINESIZE
 	bhi	1b
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 	ret	lr
+SYM_FUNC_END(cpu_fa526_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -97,7 +101,7 @@ ENTRY(cpu_fa526_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	4
-ENTRY(cpu_fa526_switch_mm)
+SYM_TYPED_FUNC_START(cpu_fa526_switch_mm)
 #ifdef CONFIG_MMU
 	mov	ip, #0
 #ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -113,6 +117,7 @@ ENTRY(cpu_fa526_switch_mm)
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate UTLB
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_fa526_switch_mm)
 
 /*
  * cpu_fa526_set_pte_ext(ptep, pte, ext)
@@ -120,7 +125,7 @@ ENTRY(cpu_fa526_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	4
-ENTRY(cpu_fa526_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_fa526_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -129,6 +134,7 @@ ENTRY(cpu_fa526_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_fa526_set_pte_ext)
 
 	.type	__fa526_setup, #function
 __fa526_setup:
diff --git a/arch/arm/mm/proc-feroceon.S b/arch/arm/mm/proc-feroceon.S
index af9482b07a4f..4c70eb0cc0d5 100644
--- a/arch/arm/mm/proc-feroceon.S
+++ b/arch/arm/mm/proc-feroceon.S
@@ -44,7 +44,7 @@ __cache_params:
 /*
  * cpu_feroceon_proc_init()
  */
-ENTRY(cpu_feroceon_proc_init)
+SYM_TYPED_FUNC_START(cpu_feroceon_proc_init)
 	mrc	p15, 0, r0, c0, c0, 1		@ read cache type register
 	ldr	r1, __cache_params
 	mov	r2, #(16 << 5)
@@ -62,11 +62,12 @@ ENTRY(cpu_feroceon_proc_init)
 	str_l	r1, VFP_arch_feroceon, r2
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_feroceon_proc_init)
 
 /*
  * cpu_feroceon_proc_fin()
  */
-ENTRY(cpu_feroceon_proc_fin)
+SYM_TYPED_FUNC_START(cpu_feroceon_proc_fin)
 #if defined(CONFIG_CACHE_FEROCEON_L2) && \
 	!defined(CONFIG_CACHE_FEROCEON_L2_WRITETHROUGH)
 	mov	r0, #0
@@ -79,6 +80,7 @@ ENTRY(cpu_feroceon_proc_fin)
 	bic	r0, r0, #0x000e			@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_feroceon_proc_fin)
 
 /*
  * cpu_feroceon_reset(loc)
@@ -91,7 +93,7 @@ ENTRY(cpu_feroceon_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_feroceon_reset)
+SYM_TYPED_FUNC_START(cpu_feroceon_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -103,7 +105,7 @@ ENTRY(cpu_feroceon_reset)
 	bic	ip, ip, #0x1100			@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_feroceon_reset)
+SYM_FUNC_END(cpu_feroceon_reset)
 	.popsection
 
 /*
@@ -112,11 +114,12 @@ ENDPROC(cpu_feroceon_reset)
  * Called with IRQs disabled
  */
 	.align	5
-ENTRY(cpu_feroceon_do_idle)
+SYM_TYPED_FUNC_START(cpu_feroceon_do_idle)
 	mov	r0, #0
 	mcr	p15, 0, r0, c7, c10, 4		@ Drain write buffer
 	mcr	p15, 0, r0, c7, c0, 4		@ Wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_feroceon_do_idle)
 
 /*
  *	flush_icache_all()
@@ -413,7 +416,7 @@ SYM_TYPED_FUNC_START(feroceon_dma_unmap_area)
 SYM_FUNC_END(feroceon_dma_unmap_area)
 
 	.align	5
-ENTRY(cpu_feroceon_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_feroceon_dcache_clean_area)
 #if defined(CONFIG_CACHE_FEROCEON_L2) && \
 	!defined(CONFIG_CACHE_FEROCEON_L2_WRITETHROUGH)
 	mov	r2, r0
@@ -432,6 +435,7 @@ ENTRY(cpu_feroceon_dcache_clean_area)
 #endif
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 	ret	lr
+SYM_FUNC_END(cpu_feroceon_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -443,7 +447,7 @@ ENTRY(cpu_feroceon_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_feroceon_switch_mm)
+SYM_TYPED_FUNC_START(cpu_feroceon_switch_mm)
 #ifdef CONFIG_MMU
 	/*
 	 * Note: we wish to call __flush_whole_cache but we need to preserve
@@ -464,6 +468,7 @@ ENTRY(cpu_feroceon_switch_mm)
 #else
 	ret	lr
 #endif
+SYM_FUNC_END(cpu_feroceon_switch_mm)
 
 /*
  * cpu_feroceon_set_pte_ext(ptep, pte, ext)
@@ -471,7 +476,7 @@ ENTRY(cpu_feroceon_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_feroceon_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_feroceon_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext wc_disable=0
 	mov	r0, r0
@@ -483,21 +488,22 @@ ENTRY(cpu_feroceon_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_feroceon_set_pte_ext)
 
 /* Suspend/resume support: taken from arch/arm/mm/proc-arm926.S */
 .globl	cpu_feroceon_suspend_size
 .equ	cpu_feroceon_suspend_size, 4 * 3
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_feroceon_do_suspend)
+SYM_TYPED_FUNC_START(cpu_feroceon_do_suspend)
 	stmfd	sp!, {r4 - r6, lr}
 	mrc	p15, 0, r4, c13, c0, 0	@ PID
 	mrc	p15, 0, r5, c3, c0, 0	@ Domain ID
 	mrc	p15, 0, r6, c1, c0, 0	@ Control register
 	stmia	r0, {r4 - r6}
 	ldmfd	sp!, {r4 - r6, pc}
-ENDPROC(cpu_feroceon_do_suspend)
+SYM_FUNC_END(cpu_feroceon_do_suspend)
 
-ENTRY(cpu_feroceon_do_resume)
+SYM_TYPED_FUNC_START(cpu_feroceon_do_resume)
 	mov	ip, #0
 	mcr	p15, 0, ip, c8, c7, 0	@ invalidate I+D TLBs
 	mcr	p15, 0, ip, c7, c7, 0	@ invalidate I+D caches
@@ -507,7 +513,7 @@ ENTRY(cpu_feroceon_do_resume)
 	mcr	p15, 0, r1, c2, c0, 0	@ TTB address
 	mov	r0, r6			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_feroceon_do_resume)
+SYM_FUNC_END(cpu_feroceon_do_resume)
 #endif
 
 	.type	__feroceon_setup, #function
diff --git a/arch/arm/mm/proc-mohawk.S b/arch/arm/mm/proc-mohawk.S
index be3a1a997838..6871395958ce 100644
--- a/arch/arm/mm/proc-mohawk.S
+++ b/arch/arm/mm/proc-mohawk.S
@@ -32,18 +32,20 @@
 /*
  * cpu_mohawk_proc_init()
  */
-ENTRY(cpu_mohawk_proc_init)
+SYM_TYPED_FUNC_START(cpu_mohawk_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_mohawk_proc_init)
 
 /*
  * cpu_mohawk_proc_fin()
  */
-ENTRY(cpu_mohawk_proc_fin)
+SYM_TYPED_FUNC_START(cpu_mohawk_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1800			@ ...iz...........
 	bic	r0, r0, #0x0006			@ .............ca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_mohawk_proc_fin)
 
 /*
  * cpu_mohawk_reset(loc)
@@ -58,7 +60,7 @@ ENTRY(cpu_mohawk_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_mohawk_reset)
+SYM_TYPED_FUNC_START(cpu_mohawk_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -68,7 +70,7 @@ ENTRY(cpu_mohawk_reset)
 	bic	ip, ip, #0x1100			@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_mohawk_reset)
+SYM_FUNC_END(cpu_mohawk_reset)
 	.popsection
 
 /*
@@ -77,11 +79,12 @@ ENDPROC(cpu_mohawk_reset)
  * Called with IRQs disabled
  */
 	.align	5
-ENTRY(cpu_mohawk_do_idle)
+SYM_TYPED_FUNC_START(cpu_mohawk_do_idle)
 	mov	r0, #0
 	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer
 	mcr	p15, 0, r0, c7, c0, 4		@ wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_mohawk_do_idle)
 
 /*
  *	flush_icache_all()
@@ -294,13 +297,14 @@ SYM_TYPED_FUNC_START(mohawk_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(mohawk_dma_unmap_area)
 
-ENTRY(cpu_mohawk_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_mohawk_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	subs	r1, r1, #CACHE_DLINESIZE
 	bhi	1b
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 	ret	lr
+SYM_FUNC_END(cpu_mohawk_dcache_clean_area)
 
 /*
  * cpu_mohawk_switch_mm(pgd)
@@ -310,7 +314,7 @@ ENTRY(cpu_mohawk_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_mohawk_switch_mm)
+SYM_TYPED_FUNC_START(cpu_mohawk_switch_mm)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c14, 0		@ clean & invalidate all D cache
 	mcr	p15, 0, ip, c7, c5, 0		@ invalidate I cache
@@ -319,6 +323,7 @@ ENTRY(cpu_mohawk_switch_mm)
 	mcr	p15, 0, r0, c2, c0, 0		@ load page table pointer
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I & D TLBs
 	ret	lr
+SYM_FUNC_END(cpu_mohawk_switch_mm)
 
 /*
  * cpu_mohawk_set_pte_ext(ptep, pte, ext)
@@ -326,7 +331,7 @@ ENTRY(cpu_mohawk_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_mohawk_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_mohawk_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext
 	mov	r0, r0
@@ -334,11 +339,12 @@ ENTRY(cpu_mohawk_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 	ret	lr
 #endif
+SYM_FUNC_END(cpu_mohawk_set_pte_ext)
 
 .globl	cpu_mohawk_suspend_size
 .equ	cpu_mohawk_suspend_size, 4 * 6
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_mohawk_do_suspend)
+SYM_TYPED_FUNC_START(cpu_mohawk_do_suspend)
 	stmfd	sp!, {r4 - r9, lr}
 	mrc	p14, 0, r4, c6, c0, 0	@ clock configuration, for turbo mode
 	mrc	p15, 0, r5, c15, c1, 0	@ CP access reg
@@ -349,9 +355,9 @@ ENTRY(cpu_mohawk_do_suspend)
 	bic	r4, r4, #2		@ clear frequency change bit
 	stmia	r0, {r4 - r9}		@ store cp regs
 	ldmia	sp!, {r4 - r9, pc}
-ENDPROC(cpu_mohawk_do_suspend)
+SYM_FUNC_END(cpu_mohawk_do_suspend)
 
-ENTRY(cpu_mohawk_do_resume)
+SYM_TYPED_FUNC_START(cpu_mohawk_do_resume)
 	ldmia	r0, {r4 - r9}		@ load cp regs
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0	@ invalidate I & D caches, BTB
@@ -367,7 +373,7 @@ ENTRY(cpu_mohawk_do_resume)
 	mcr	p15, 0, r8, c1, c0, 1	@ auxiliary control reg
 	mov	r0, r9			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_mohawk_do_resume)
+SYM_FUNC_END(cpu_mohawk_do_resume)
 #endif
 
 	.type	__mohawk_setup, #function
diff --git a/arch/arm/mm/proc-sa110.S b/arch/arm/mm/proc-sa110.S
index 4071f7a61cb6..3da76fab8ac3 100644
--- a/arch/arm/mm/proc-sa110.S
+++ b/arch/arm/mm/proc-sa110.S
@@ -12,6 +12,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
@@ -32,15 +33,16 @@
 /*
  * cpu_sa110_proc_init()
  */
-ENTRY(cpu_sa110_proc_init)
+SYM_TYPED_FUNC_START(cpu_sa110_proc_init)
 	mov	r0, #0
 	mcr	p15, 0, r0, c15, c1, 2		@ Enable clock switching
 	ret	lr
+SYM_FUNC_END(cpu_sa110_proc_init)
 
 /*
  * cpu_sa110_proc_fin()
  */
-ENTRY(cpu_sa110_proc_fin)
+SYM_TYPED_FUNC_START(cpu_sa110_proc_fin)
 	mov	r0, #0
 	mcr	p15, 0, r0, c15, c2, 2		@ Disable clock switching
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
@@ -48,6 +50,7 @@ ENTRY(cpu_sa110_proc_fin)
 	bic	r0, r0, #0x000e			@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_sa110_proc_fin)
 
 /*
  * cpu_sa110_reset(loc)
@@ -60,7 +63,7 @@ ENTRY(cpu_sa110_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_sa110_reset)
+SYM_TYPED_FUNC_START(cpu_sa110_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -72,7 +75,7 @@ ENTRY(cpu_sa110_reset)
 	bic	ip, ip, #0x1100			@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_sa110_reset)
+SYM_FUNC_END(cpu_sa110_reset)
 	.popsection
 
 /*
@@ -88,7 +91,7 @@ ENDPROC(cpu_sa110_reset)
  */
 	.align	5
 
-ENTRY(cpu_sa110_do_idle)
+SYM_TYPED_FUNC_START(cpu_sa110_do_idle)
 	mcr	p15, 0, ip, c15, c2, 2		@ disable clock switching
 	ldr	r1, =UNCACHEABLE_ADDR		@ load from uncacheable loc
 	ldr	r1, [r1, #0]			@ force switch to MCLK
@@ -101,6 +104,7 @@ ENTRY(cpu_sa110_do_idle)
 	mov	r0, r0				@ safety
 	mcr	p15, 0, r0, c15, c1, 2		@ enable clock switching
 	ret	lr
+SYM_FUNC_END(cpu_sa110_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -113,12 +117,13 @@ ENTRY(cpu_sa110_do_idle)
  * addr: cache-unaligned virtual address
  */
 	.align	5
-ENTRY(cpu_sa110_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_sa110_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #DCACHELINESIZE
 	subs	r1, r1, #DCACHELINESIZE
 	bhi	1b
 	ret	lr
+SYM_FUNC_END(cpu_sa110_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -130,7 +135,7 @@ ENTRY(cpu_sa110_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_sa110_switch_mm)
+SYM_TYPED_FUNC_START(cpu_sa110_switch_mm)
 #ifdef CONFIG_MMU
 	str	lr, [sp, #-4]!
 	bl	v4wb_flush_kern_cache_all	@ clears IP
@@ -140,6 +145,7 @@ ENTRY(cpu_sa110_switch_mm)
 #else
 	ret	lr
 #endif
+SYM_FUNC_END(cpu_sa110_switch_mm)
 
 /*
  * cpu_sa110_set_pte_ext(ptep, pte, ext)
@@ -147,7 +153,7 @@ ENTRY(cpu_sa110_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_sa110_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_sa110_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext wc_disable=0
 	mov	r0, r0
@@ -155,6 +161,7 @@ ENTRY(cpu_sa110_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_sa110_set_pte_ext)
 
 	.type	__sa110_setup, #function
 __sa110_setup:
diff --git a/arch/arm/mm/proc-sa1100.S b/arch/arm/mm/proc-sa1100.S
index e723bd4119d3..7c496195e440 100644
--- a/arch/arm/mm/proc-sa1100.S
+++ b/arch/arm/mm/proc-sa1100.S
@@ -17,6 +17,7 @@
  */
 #include <linux/linkage.h>
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
 #include <asm/asm-offsets.h>
@@ -36,11 +37,12 @@
 /*
  * cpu_sa1100_proc_init()
  */
-ENTRY(cpu_sa1100_proc_init)
+SYM_TYPED_FUNC_START(cpu_sa1100_proc_init)
 	mov	r0, #0
 	mcr	p15, 0, r0, c15, c1, 2		@ Enable clock switching
 	mcr	p15, 0, r0, c9, c0, 5		@ Allow read-buffer operations from userland
 	ret	lr
+SYM_FUNC_END(cpu_sa1100_proc_init)
 
 /*
  * cpu_sa1100_proc_fin()
@@ -49,13 +51,14 @@ ENTRY(cpu_sa1100_proc_init)
  *  - Disable interrupts
  *  - Clean and turn off caches.
  */
-ENTRY(cpu_sa1100_proc_fin)
+SYM_TYPED_FUNC_START(cpu_sa1100_proc_fin)
 	mcr	p15, 0, ip, c15, c2, 2		@ Disable clock switching
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x000e			@ ............wca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_sa1100_proc_fin)
 
 /*
  * cpu_sa1100_reset(loc)
@@ -68,7 +71,7 @@ ENTRY(cpu_sa1100_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_sa1100_reset)
+SYM_TYPED_FUNC_START(cpu_sa1100_reset)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0		@ invalidate I,D caches
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
@@ -80,7 +83,7 @@ ENTRY(cpu_sa1100_reset)
 	bic	ip, ip, #0x1100			@ ...i...s........
 	mcr	p15, 0, ip, c1, c0, 0		@ ctrl register
 	ret	r0
-ENDPROC(cpu_sa1100_reset)
+SYM_FUNC_END(cpu_sa1100_reset)
 	.popsection
 
 /*
@@ -95,7 +98,7 @@ ENDPROC(cpu_sa1100_reset)
  *   3 = switch to fast processor clock
  */
 	.align	5
-ENTRY(cpu_sa1100_do_idle)
+SYM_TYPED_FUNC_START(cpu_sa1100_do_idle)
 	mov	r0, r0				@ 4 nop padding
 	mov	r0, r0
 	mov	r0, r0
@@ -111,6 +114,7 @@ ENTRY(cpu_sa1100_do_idle)
 	mov	r0, r0				@ safety
 	mcr	p15, 0, r0, c15, c1, 2		@ enable clock switching
 	ret	lr
+SYM_FUNC_END(cpu_sa1100_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -123,12 +127,13 @@ ENTRY(cpu_sa1100_do_idle)
  * addr: cache-unaligned virtual address
  */
 	.align	5
-ENTRY(cpu_sa1100_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_sa1100_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #DCACHELINESIZE
 	subs	r1, r1, #DCACHELINESIZE
 	bhi	1b
 	ret	lr
+SYM_FUNC_END(cpu_sa1100_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -140,7 +145,7 @@ ENTRY(cpu_sa1100_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_sa1100_switch_mm)
+SYM_TYPED_FUNC_START(cpu_sa1100_switch_mm)
 #ifdef CONFIG_MMU
 	str	lr, [sp, #-4]!
 	bl	v4wb_flush_kern_cache_all	@ clears IP
@@ -151,6 +156,7 @@ ENTRY(cpu_sa1100_switch_mm)
 #else
 	ret	lr
 #endif
+SYM_FUNC_END(cpu_sa1100_switch_mm)
 
 /*
  * cpu_sa1100_set_pte_ext(ptep, pte, ext)
@@ -158,7 +164,7 @@ ENTRY(cpu_sa1100_switch_mm)
  * Set a PTE and flush it out
  */
 	.align	5
-ENTRY(cpu_sa1100_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_sa1100_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv3_set_pte_ext wc_disable=0
 	mov	r0, r0
@@ -166,20 +172,21 @@ ENTRY(cpu_sa1100_set_pte_ext)
 	mcr	p15, 0, r0, c7, c10, 4		@ drain WB
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_sa1100_set_pte_ext)
 
 .globl	cpu_sa1100_suspend_size
 .equ	cpu_sa1100_suspend_size, 4 * 3
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_sa1100_do_suspend)
+SYM_TYPED_FUNC_START(cpu_sa1100_do_suspend)
 	stmfd	sp!, {r4 - r6, lr}
 	mrc	p15, 0, r4, c3, c0, 0		@ domain ID
 	mrc	p15, 0, r5, c13, c0, 0		@ PID
 	mrc	p15, 0, r6, c1, c0, 0		@ control reg
 	stmia	r0, {r4 - r6}			@ store cp regs
 	ldmfd	sp!, {r4 - r6, pc}
-ENDPROC(cpu_sa1100_do_suspend)
+SYM_FUNC_END(cpu_sa1100_do_suspend)
 
-ENTRY(cpu_sa1100_do_resume)
+SYM_TYPED_FUNC_START(cpu_sa1100_do_resume)
 	ldmia	r0, {r4 - r6}			@ load cp regs
 	mov	ip, #0
 	mcr	p15, 0, ip, c8, c7, 0		@ flush I+D TLBs
@@ -192,7 +199,7 @@ ENTRY(cpu_sa1100_do_resume)
 	mcr	p15, 0, r5, c13, c0, 0		@ PID
 	mov	r0, r6				@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_sa1100_do_resume)
+SYM_FUNC_END(cpu_sa1100_do_resume)
 #endif
 
 	.type	__sa1100_setup, #function
diff --git a/arch/arm/mm/proc-v6.S b/arch/arm/mm/proc-v6.S
index 203dff89ab1a..90a01f5950b9 100644
--- a/arch/arm/mm/proc-v6.S
+++ b/arch/arm/mm/proc-v6.S
@@ -8,6 +8,7 @@
  *  This is the "shell" of the ARMv6 processor support.
  */
 #include <linux/init.h>
+#include <linux/cfi_types.h>
 #include <linux/linkage.h>
 #include <linux/pgtable.h>
 #include <asm/assembler.h>
@@ -34,15 +35,17 @@
 
 .arch armv6
 
-ENTRY(cpu_v6_proc_init)
+SYM_TYPED_FUNC_START(cpu_v6_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_v6_proc_init)
 
-ENTRY(cpu_v6_proc_fin)
+SYM_TYPED_FUNC_START(cpu_v6_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x0006			@ .............ca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_v6_proc_fin)
 
 /*
  *	cpu_v6_reset(loc)
@@ -55,14 +58,14 @@ ENTRY(cpu_v6_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_v6_reset)
+SYM_TYPED_FUNC_START(cpu_v6_reset)
 	mrc	p15, 0, r1, c1, c0, 0		@ ctrl register
 	bic	r1, r1, #0x1			@ ...............m
 	mcr	p15, 0, r1, c1, c0, 0		@ disable MMU
 	mov	r1, #0
 	mcr	p15, 0, r1, c7, c5, 4		@ ISB
 	ret	r0
-ENDPROC(cpu_v6_reset)
+SYM_FUNC_END(cpu_v6_reset)
 	.popsection
 
 /*
@@ -72,18 +75,20 @@ ENDPROC(cpu_v6_reset)
  *
  *	IRQs are already disabled.
  */
-ENTRY(cpu_v6_do_idle)
+SYM_TYPED_FUNC_START(cpu_v6_do_idle)
 	mov	r1, #0
 	mcr	p15, 0, r1, c7, c10, 4		@ DWB - WFI may enter a low-power mode
 	mcr	p15, 0, r1, c7, c0, 4		@ wait for interrupt
 	ret	lr
+SYM_FUNC_END(cpu_v6_do_idle)
 
-ENTRY(cpu_v6_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_v6_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #D_CACHE_LINE_SIZE
 	subs	r1, r1, #D_CACHE_LINE_SIZE
 	bhi	1b
 	ret	lr
+SYM_FUNC_END(cpu_v6_dcache_clean_area)
 
 /*
  *	cpu_v6_switch_mm(pgd_phys, tsk)
@@ -95,7 +100,7 @@ ENTRY(cpu_v6_dcache_clean_area)
  *	It is assumed that:
  *	- we are not using split page tables
  */
-ENTRY(cpu_v6_switch_mm)
+SYM_TYPED_FUNC_START(cpu_v6_switch_mm)
 #ifdef CONFIG_MMU
 	mov	r2, #0
 	mmid	r1, r1				@ get mm->context.id
@@ -113,6 +118,7 @@ ENTRY(cpu_v6_switch_mm)
 	mcr	p15, 0, r1, c13, c0, 1		@ set context ID
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_v6_switch_mm)
 
 /*
  *	cpu_v6_set_pte_ext(ptep, pte, ext)
@@ -126,17 +132,18 @@ ENTRY(cpu_v6_switch_mm)
  */
 	armv6_mt_table cpu_v6
 
-ENTRY(cpu_v6_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_v6_set_pte_ext)
 #ifdef CONFIG_MMU
 	armv6_set_pte_ext cpu_v6
 #endif
 	ret	lr
+SYM_FUNC_END(cpu_v6_set_pte_ext)
 
 /* Suspend/resume support: taken from arch/arm/mach-s3c64xx/sleep.S */
 .globl	cpu_v6_suspend_size
 .equ	cpu_v6_suspend_size, 4 * 6
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_v6_do_suspend)
+SYM_TYPED_FUNC_START(cpu_v6_do_suspend)
 	stmfd	sp!, {r4 - r9, lr}
 	mrc	p15, 0, r4, c13, c0, 0	@ FCSE/PID
 #ifdef CONFIG_MMU
@@ -148,9 +155,9 @@ ENTRY(cpu_v6_do_suspend)
 	mrc	p15, 0, r9, c1, c0, 0	@ control register
 	stmia	r0, {r4 - r9}
 	ldmfd	sp!, {r4- r9, pc}
-ENDPROC(cpu_v6_do_suspend)
+SYM_FUNC_END(cpu_v6_do_suspend)
 
-ENTRY(cpu_v6_do_resume)
+SYM_TYPED_FUNC_START(cpu_v6_do_resume)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c14, 0	@ clean+invalidate D cache
 	mcr	p15, 0, ip, c7, c5, 0	@ invalidate I cache
@@ -172,7 +179,7 @@ ENTRY(cpu_v6_do_resume)
 	mcr	p15, 0, ip, c7, c5, 4	@ ISB
 	mov	r0, r9			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_v6_do_resume)
+SYM_FUNC_END(cpu_v6_do_resume)
 #endif
 
 	string	cpu_v6_name, "ARMv6-compatible processor"
diff --git a/arch/arm/mm/proc-v7-2level.S b/arch/arm/mm/proc-v7-2level.S
index 0a3083ad19c2..1007702fcaf3 100644
--- a/arch/arm/mm/proc-v7-2level.S
+++ b/arch/arm/mm/proc-v7-2level.S
@@ -40,7 +40,7 @@
  *	even on Cortex-A8 revisions not affected by 430973.
  *	If IBE is not set, the flush BTAC/BTB won't do anything.
  */
-ENTRY(cpu_v7_switch_mm)
+SYM_TYPED_FUNC_START(cpu_v7_switch_mm)
 #ifdef CONFIG_MMU
 	mmid	r1, r1				@ get mm->context.id
 	ALT_SMP(orr	r0, r0, #TTB_FLAGS_SMP)
@@ -59,7 +59,7 @@ ENTRY(cpu_v7_switch_mm)
 	isb
 #endif
 	bx	lr
-ENDPROC(cpu_v7_switch_mm)
+SYM_FUNC_END(cpu_v7_switch_mm)
 
 /*
  *	cpu_v7_set_pte_ext(ptep, pte)
@@ -71,7 +71,7 @@ ENDPROC(cpu_v7_switch_mm)
  *	- pte   - PTE value to store
  *	- ext	- value for extended PTE bits
  */
-ENTRY(cpu_v7_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_v7_set_pte_ext)
 #ifdef CONFIG_MMU
 	str	r1, [r0]			@ linux version
 
@@ -106,7 +106,7 @@ ENTRY(cpu_v7_set_pte_ext)
 	ALT_UP (mcr	p15, 0, r0, c7, c10, 1)		@ flush_pte
 #endif
 	bx	lr
-ENDPROC(cpu_v7_set_pte_ext)
+SYM_FUNC_END(cpu_v7_set_pte_ext)
 
 	/*
 	 * Memory region attributes with SCTLR.TRE=1
diff --git a/arch/arm/mm/proc-v7-3level.S b/arch/arm/mm/proc-v7-3level.S
index 131984462d0d..bdabc15cde56 100644
--- a/arch/arm/mm/proc-v7-3level.S
+++ b/arch/arm/mm/proc-v7-3level.S
@@ -42,7 +42,7 @@
  * Set the translation table base pointer to be pgd_phys (physical address of
  * the new TTB).
  */
-ENTRY(cpu_v7_switch_mm)
+SYM_TYPED_FUNC_START(cpu_v7_switch_mm)
 #ifdef CONFIG_MMU
 	mmid	r2, r2
 	asid	r2, r2
@@ -51,7 +51,7 @@ ENTRY(cpu_v7_switch_mm)
 	isb
 #endif
 	ret	lr
-ENDPROC(cpu_v7_switch_mm)
+SYM_FUNC_END(cpu_v7_switch_mm)
 
 #ifdef __ARMEB__
 #define rl r3
@@ -68,7 +68,7 @@ ENDPROC(cpu_v7_switch_mm)
  * - ptep - pointer to level 3 translation table entry
  * - pte - PTE value to store (64-bit in r2 and r3)
  */
-ENTRY(cpu_v7_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_v7_set_pte_ext)
 #ifdef CONFIG_MMU
 	tst	rl, #L_PTE_VALID
 	beq	1f
@@ -87,7 +87,7 @@ ENTRY(cpu_v7_set_pte_ext)
 	ALT_UP (mcr	p15, 0, r0, c7, c10, 1)		@ flush_pte
 #endif
 	ret	lr
-ENDPROC(cpu_v7_set_pte_ext)
+SYM_FUNC_END(cpu_v7_set_pte_ext)
 
 	/*
 	 * Memory region attributes for LPAE (defined in pgtable-3level.h):
diff --git a/arch/arm/mm/proc-v7.S b/arch/arm/mm/proc-v7.S
index 193c7aeb6703..5fb9a6aecb00 100644
--- a/arch/arm/mm/proc-v7.S
+++ b/arch/arm/mm/proc-v7.S
@@ -7,6 +7,7 @@
  *  This is the "shell" of the ARMv7 processor support.
  */
 #include <linux/arm-smccc.h>
+#include <linux/cfi_types.h>
 #include <linux/init.h>
 #include <linux/linkage.h>
 #include <linux/pgtable.h>
@@ -26,17 +27,17 @@
 
 .arch armv7-a
 
-ENTRY(cpu_v7_proc_init)
+SYM_TYPED_FUNC_START(cpu_v7_proc_init)
 	ret	lr
-ENDPROC(cpu_v7_proc_init)
+SYM_FUNC_END(cpu_v7_proc_init)
 
-ENTRY(cpu_v7_proc_fin)
+SYM_TYPED_FUNC_START(cpu_v7_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1000			@ ...i............
 	bic	r0, r0, #0x0006			@ .............ca.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
-ENDPROC(cpu_v7_proc_fin)
+SYM_FUNC_END(cpu_v7_proc_fin)
 
 /*
  *	cpu_v7_reset(loc, hyp)
@@ -53,7 +54,7 @@ ENDPROC(cpu_v7_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_v7_reset)
+SYM_TYPED_FUNC_START(cpu_v7_reset)
 	mrc	p15, 0, r2, c1, c0, 0		@ ctrl register
 	bic	r2, r2, #0x1			@ ...............m
  THUMB(	bic	r2, r2, #1 << 30 )		@ SCTLR.TE (Thumb exceptions)
@@ -64,7 +65,7 @@ ENTRY(cpu_v7_reset)
 	bne	__hyp_soft_restart
 #endif
 	bx	r0
-ENDPROC(cpu_v7_reset)
+SYM_FUNC_END(cpu_v7_reset)
 	.popsection
 
 /*
@@ -74,13 +75,13 @@ ENDPROC(cpu_v7_reset)
  *
  *	IRQs are already disabled.
  */
-ENTRY(cpu_v7_do_idle)
+SYM_TYPED_FUNC_START(cpu_v7_do_idle)
 	dsb					@ WFI may enter a low-power mode
 	wfi
 	ret	lr
-ENDPROC(cpu_v7_do_idle)
+SYM_FUNC_END(cpu_v7_do_idle)
 
-ENTRY(cpu_v7_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_v7_dcache_clean_area)
 	ALT_SMP(W(nop))			@ MP extensions imply L1 PTW
 	ALT_UP_B(1f)
 	ret	lr
@@ -91,38 +92,39 @@ ENTRY(cpu_v7_dcache_clean_area)
 	bhi	2b
 	dsb	ishst
 	ret	lr
-ENDPROC(cpu_v7_dcache_clean_area)
+SYM_FUNC_END(cpu_v7_dcache_clean_area)
 
 #ifdef CONFIG_ARM_PSCI
 	.arch_extension sec
-ENTRY(cpu_v7_smc_switch_mm)
+SYM_TYPED_FUNC_START(cpu_v7_smc_switch_mm)
 	stmfd	sp!, {r0 - r3}
 	movw	r0, #:lower16:ARM_SMCCC_ARCH_WORKAROUND_1
 	movt	r0, #:upper16:ARM_SMCCC_ARCH_WORKAROUND_1
 	smc	#0
 	ldmfd	sp!, {r0 - r3}
 	b	cpu_v7_switch_mm
-ENDPROC(cpu_v7_smc_switch_mm)
+SYM_FUNC_END(cpu_v7_smc_switch_mm)
 	.arch_extension virt
-ENTRY(cpu_v7_hvc_switch_mm)
+SYM_TYPED_FUNC_START(cpu_v7_hvc_switch_mm)
 	stmfd	sp!, {r0 - r3}
 	movw	r0, #:lower16:ARM_SMCCC_ARCH_WORKAROUND_1
 	movt	r0, #:upper16:ARM_SMCCC_ARCH_WORKAROUND_1
 	hvc	#0
 	ldmfd	sp!, {r0 - r3}
 	b	cpu_v7_switch_mm
-ENDPROC(cpu_v7_hvc_switch_mm)
+SYM_FUNC_END(cpu_v7_hvc_switch_mm)
 #endif
-ENTRY(cpu_v7_iciallu_switch_mm)
+
+SYM_TYPED_FUNC_START(cpu_v7_iciallu_switch_mm)
 	mov	r3, #0
 	mcr	p15, 0, r3, c7, c5, 0		@ ICIALLU
 	b	cpu_v7_switch_mm
-ENDPROC(cpu_v7_iciallu_switch_mm)
-ENTRY(cpu_v7_bpiall_switch_mm)
+SYM_FUNC_END(cpu_v7_iciallu_switch_mm)
+SYM_TYPED_FUNC_START(cpu_v7_bpiall_switch_mm)
 	mov	r3, #0
 	mcr	p15, 0, r3, c7, c5, 6		@ flush BTAC/BTB
 	b	cpu_v7_switch_mm
-ENDPROC(cpu_v7_bpiall_switch_mm)
+SYM_FUNC_END(cpu_v7_bpiall_switch_mm)
 
 	string	cpu_v7_name, "ARMv7 Processor"
 	.align
@@ -131,7 +133,7 @@ ENDPROC(cpu_v7_bpiall_switch_mm)
 .globl	cpu_v7_suspend_size
 .equ	cpu_v7_suspend_size, 4 * 9
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_v7_do_suspend)
+SYM_TYPED_FUNC_START(cpu_v7_do_suspend)
 	stmfd	sp!, {r4 - r11, lr}
 	mrc	p15, 0, r4, c13, c0, 0	@ FCSE/PID
 	mrc	p15, 0, r5, c13, c0, 3	@ User r/o thread ID
@@ -150,9 +152,9 @@ ENTRY(cpu_v7_do_suspend)
 	mrc	p15, 0, r10, c1, c0, 2	@ Co-processor access control
 	stmia	r0, {r5 - r11}
 	ldmfd	sp!, {r4 - r11, pc}
-ENDPROC(cpu_v7_do_suspend)
+SYM_FUNC_END(cpu_v7_do_suspend)
 
-ENTRY(cpu_v7_do_resume)
+SYM_TYPED_FUNC_START(cpu_v7_do_resume)
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c5, 0	@ invalidate I cache
 	mcr	p15, 0, ip, c13, c0, 1	@ set reserved context ID
@@ -186,22 +188,22 @@ ENTRY(cpu_v7_do_resume)
 	dsb
 	mov	r0, r8			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_v7_do_resume)
+SYM_FUNC_END(cpu_v7_do_resume)
 #endif
 
 .globl	cpu_ca9mp_suspend_size
 .equ	cpu_ca9mp_suspend_size, cpu_v7_suspend_size + 4 * 2
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_ca9mp_do_suspend)
+SYM_TYPED_FUNC_START(cpu_ca9mp_do_suspend)
 	stmfd	sp!, {r4 - r5}
 	mrc	p15, 0, r4, c15, c0, 1		@ Diagnostic register
 	mrc	p15, 0, r5, c15, c0, 0		@ Power register
 	stmia	r0!, {r4 - r5}
 	ldmfd	sp!, {r4 - r5}
 	b	cpu_v7_do_suspend
-ENDPROC(cpu_ca9mp_do_suspend)
+SYM_FUNC_END(cpu_ca9mp_do_suspend)
 
-ENTRY(cpu_ca9mp_do_resume)
+SYM_TYPED_FUNC_START(cpu_ca9mp_do_resume)
 	ldmia	r0!, {r4 - r5}
 	mrc	p15, 0, r10, c15, c0, 1		@ Read Diagnostic register
 	teq	r4, r10				@ Already restored?
@@ -210,7 +212,7 @@ ENTRY(cpu_ca9mp_do_resume)
 	teq	r5, r10				@ Already restored?
 	mcrne	p15, 0, r5, c15, c0, 0		@ No, so restore it
 	b	cpu_v7_do_resume
-ENDPROC(cpu_ca9mp_do_resume)
+SYM_FUNC_END(cpu_ca9mp_do_resume)
 #endif
 
 #ifdef CONFIG_CPU_PJ4B
@@ -220,18 +222,18 @@ ENDPROC(cpu_ca9mp_do_resume)
 	globl_equ	cpu_pj4b_proc_fin, 	cpu_v7_proc_fin
 	globl_equ	cpu_pj4b_reset,	   	cpu_v7_reset
 #ifdef CONFIG_PJ4B_ERRATA_4742
-ENTRY(cpu_pj4b_do_idle)
+SYM_TYPED_FUNC_START(cpu_pj4b_do_idle)
 	dsb					@ WFI may enter a low-power mode
 	wfi
 	dsb					@barrier
 	ret	lr
-ENDPROC(cpu_pj4b_do_idle)
+SYM_FUNC_END(cpu_pj4b_do_idle)
 #else
 	globl_equ	cpu_pj4b_do_idle,  	cpu_v7_do_idle
 #endif
 	globl_equ	cpu_pj4b_dcache_clean_area,	cpu_v7_dcache_clean_area
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_pj4b_do_suspend)
+SYM_TYPED_FUNC_START(cpu_pj4b_do_suspend)
 	stmfd	sp!, {r6 - r10}
 	mrc	p15, 1, r6, c15, c1, 0  @ save CP15 - extra features
 	mrc	p15, 1, r7, c15, c2, 0	@ save CP15 - Aux Func Modes Ctrl 0
@@ -241,9 +243,9 @@ ENTRY(cpu_pj4b_do_suspend)
 	stmia	r0!, {r6 - r10}
 	ldmfd	sp!, {r6 - r10}
 	b cpu_v7_do_suspend
-ENDPROC(cpu_pj4b_do_suspend)
+SYM_FUNC_END(cpu_pj4b_do_suspend)
 
-ENTRY(cpu_pj4b_do_resume)
+SYM_TYPED_FUNC_START(cpu_pj4b_do_resume)
 	ldmia	r0!, {r6 - r10}
 	mcr	p15, 1, r6, c15, c1, 0  @ restore CP15 - extra features
 	mcr	p15, 1, r7, c15, c2, 0	@ restore CP15 - Aux Func Modes Ctrl 0
@@ -251,7 +253,7 @@ ENTRY(cpu_pj4b_do_resume)
 	mcr	p15, 1, r9, c15, c1, 1  @ restore CP15 - Aux Debug Modes Ctrl 1
 	mcr	p15, 0, r10, c9, c14, 0  @ restore CP15 - PMC
 	b cpu_v7_do_resume
-ENDPROC(cpu_pj4b_do_resume)
+SYM_FUNC_END(cpu_pj4b_do_resume)
 #endif
 .globl	cpu_pj4b_suspend_size
 .equ	cpu_pj4b_suspend_size, cpu_v7_suspend_size + 4 * 5
diff --git a/arch/arm/mm/proc-v7m.S b/arch/arm/mm/proc-v7m.S
index d65a12f851a9..d4675603593b 100644
--- a/arch/arm/mm/proc-v7m.S
+++ b/arch/arm/mm/proc-v7m.S
@@ -8,18 +8,19 @@
  *  This is the "shell" of the ARMv7-M processor support.
  */
 #include <linux/linkage.h>
+#include <linux/cfi_types.h>
 #include <asm/assembler.h>
 #include <asm/page.h>
 #include <asm/v7m.h>
 #include "proc-macros.S"
 
-ENTRY(cpu_v7m_proc_init)
+SYM_TYPED_FUNC_START(cpu_v7m_proc_init)
 	ret	lr
-ENDPROC(cpu_v7m_proc_init)
+SYM_FUNC_END(cpu_v7m_proc_init)
 
-ENTRY(cpu_v7m_proc_fin)
+SYM_TYPED_FUNC_START(cpu_v7m_proc_fin)
 	ret	lr
-ENDPROC(cpu_v7m_proc_fin)
+SYM_FUNC_END(cpu_v7m_proc_fin)
 
 /*
  *	cpu_v7m_reset(loc)
@@ -31,9 +32,9 @@ ENDPROC(cpu_v7m_proc_fin)
  *	- loc   - location to jump to for soft reset
  */
 	.align	5
-ENTRY(cpu_v7m_reset)
+SYM_TYPED_FUNC_START(cpu_v7m_reset)
 	ret	r0
-ENDPROC(cpu_v7m_reset)
+SYM_FUNC_END(cpu_v7m_reset)
 
 /*
  *	cpu_v7m_do_idle()
@@ -42,36 +43,36 @@ ENDPROC(cpu_v7m_reset)
  *
  *	IRQs are already disabled.
  */
-ENTRY(cpu_v7m_do_idle)
+SYM_TYPED_FUNC_START(cpu_v7m_do_idle)
 	wfi
 	ret	lr
-ENDPROC(cpu_v7m_do_idle)
+SYM_FUNC_END(cpu_v7m_do_idle)
 
-ENTRY(cpu_v7m_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_v7m_dcache_clean_area)
 	ret	lr
-ENDPROC(cpu_v7m_dcache_clean_area)
+SYM_FUNC_END(cpu_v7m_dcache_clean_area)
 
 /*
  * There is no MMU, so here is nothing to do.
  */
-ENTRY(cpu_v7m_switch_mm)
+SYM_TYPED_FUNC_START(cpu_v7m_switch_mm)
 	ret	lr
-ENDPROC(cpu_v7m_switch_mm)
+SYM_FUNC_END(cpu_v7m_switch_mm)
 
 .globl	cpu_v7m_suspend_size
 .equ	cpu_v7m_suspend_size, 0
 
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_v7m_do_suspend)
+SYM_TYPED_FUNC_START(cpu_v7m_do_suspend)
 	ret	lr
-ENDPROC(cpu_v7m_do_suspend)
+SYM_FUNC_END(cpu_v7m_do_suspend)
 
-ENTRY(cpu_v7m_do_resume)
+SYM_TYPED_FUNC_START(cpu_v7m_do_resume)
 	ret	lr
-ENDPROC(cpu_v7m_do_resume)
+SYM_FUNC_END(cpu_v7m_do_resume)
 #endif
 
-ENTRY(cpu_cm7_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_cm7_dcache_clean_area)
 	dcache_line_size r2, r3
 	movw	r3, #:lower16:BASEADDR_V7M_SCB + V7M_SCB_DCCMVAC
 	movt	r3, #:upper16:BASEADDR_V7M_SCB + V7M_SCB_DCCMVAC
@@ -82,16 +83,16 @@ ENTRY(cpu_cm7_dcache_clean_area)
 	bhi	1b
 	dsb
 	ret	lr
-ENDPROC(cpu_cm7_dcache_clean_area)
+SYM_FUNC_END(cpu_cm7_dcache_clean_area)
 
-ENTRY(cpu_cm7_proc_fin)
+SYM_TYPED_FUNC_START(cpu_cm7_proc_fin)
 	movw	r2, #:lower16:(BASEADDR_V7M_SCB + V7M_SCB_CCR)
 	movt	r2, #:upper16:(BASEADDR_V7M_SCB + V7M_SCB_CCR)
 	ldr	r0, [r2]
 	bic	r0, r0, #(V7M_SCB_CCR_DC | V7M_SCB_CCR_IC)
 	str	r0, [r2]
 	ret	lr
-ENDPROC(cpu_cm7_proc_fin)
+SYM_FUNC_END(cpu_cm7_proc_fin)
 
 	.section ".init.text", "ax"
 
diff --git a/arch/arm/mm/proc-xsc3.S b/arch/arm/mm/proc-xsc3.S
index 7975f93b1e14..0e3d8e76376a 100644
--- a/arch/arm/mm/proc-xsc3.S
+++ b/arch/arm/mm/proc-xsc3.S
@@ -80,18 +80,20 @@
  *
  * Nothing too exciting at the moment
  */
-ENTRY(cpu_xsc3_proc_init)
+SYM_TYPED_FUNC_START(cpu_xsc3_proc_init)
 	ret	lr
+SYM_FUNC_END(cpu_xsc3_proc_init)
 
 /*
  * cpu_xsc3_proc_fin()
  */
-ENTRY(cpu_xsc3_proc_fin)
+SYM_TYPED_FUNC_START(cpu_xsc3_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1800			@ ...IZ...........
 	bic	r0, r0, #0x0006			@ .............CA.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_xsc3_proc_fin)
 
 /*
  * cpu_xsc3_reset(loc)
@@ -104,7 +106,7 @@ ENTRY(cpu_xsc3_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_xsc3_reset)
+SYM_TYPED_FUNC_START(cpu_xsc3_reset)
 	mov	r1, #PSR_F_BIT|PSR_I_BIT|SVC_MODE
 	msr	cpsr_c, r1			@ reset CPSR
 	mrc	p15, 0, r1, c1, c0, 0		@ ctrl register
@@ -118,7 +120,7 @@ ENTRY(cpu_xsc3_reset)
 	@ already containing those two last instructions to survive.
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I and D TLBs
 	ret	r0
-ENDPROC(cpu_xsc3_reset)
+SYM_FUNC_END(cpu_xsc3_reset)
 	.popsection
 
 /*
@@ -133,10 +135,11 @@ ENDPROC(cpu_xsc3_reset)
  */
 	.align	5
 
-ENTRY(cpu_xsc3_do_idle)
+SYM_TYPED_FUNC_START(cpu_xsc3_do_idle)
 	mov	r0, #1
 	mcr	p14, 0, r0, c7, c0, 0		@ go to idle
 	ret	lr
+SYM_FUNC_END(cpu_xsc3_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -339,12 +342,13 @@ SYM_TYPED_FUNC_START(xsc3_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(xsc3_dma_unmap_area)
 
-ENTRY(cpu_xsc3_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_xsc3_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean L1 D line
 	add	r0, r0, #CACHELINESIZE
 	subs	r1, r1, #CACHELINESIZE
 	bhi	1b
 	ret	lr
+SYM_FUNC_END(cpu_xsc3_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -356,7 +360,7 @@ ENTRY(cpu_xsc3_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_xsc3_switch_mm)
+SYM_TYPED_FUNC_START(cpu_xsc3_switch_mm)
 	clean_d_cache r1, r2
 	mcr	p15, 0, ip, c7, c5, 0		@ invalidate L1 I cache and BTB
 	mcr	p15, 0, ip, c7, c10, 4		@ data write barrier
@@ -365,6 +369,7 @@ ENTRY(cpu_xsc3_switch_mm)
 	mcr	p15, 0, r0, c2, c0, 0		@ load page table pointer
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I and D TLBs
 	cpwait_ret lr, ip
+SYM_FUNC_END(cpu_xsc3_switch_mm)
 
 /*
  * cpu_xsc3_set_pte_ext(ptep, pte, ext)
@@ -390,7 +395,7 @@ cpu_xsc3_mt_table:
 	.long	0x00						@ unused
 
 	.align	5
-ENTRY(cpu_xsc3_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_xsc3_set_pte_ext)
 	xscale_set_pte_ext_prologue
 
 	tst	r1, #L_PTE_SHARED		@ shared?
@@ -403,6 +408,7 @@ ENTRY(cpu_xsc3_set_pte_ext)
 
 	xscale_set_pte_ext_epilogue
 	ret	lr
+SYM_FUNC_END(cpu_xsc3_set_pte_ext)
 
 	.ltorg
 	.align
@@ -410,7 +416,7 @@ ENTRY(cpu_xsc3_set_pte_ext)
 .globl	cpu_xsc3_suspend_size
 .equ	cpu_xsc3_suspend_size, 4 * 6
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_xsc3_do_suspend)
+SYM_TYPED_FUNC_START(cpu_xsc3_do_suspend)
 	stmfd	sp!, {r4 - r9, lr}
 	mrc	p14, 0, r4, c6, c0, 0	@ clock configuration, for turbo mode
 	mrc	p15, 0, r5, c15, c1, 0	@ CP access reg
@@ -421,9 +427,9 @@ ENTRY(cpu_xsc3_do_suspend)
 	bic	r4, r4, #2		@ clear frequency change bit
 	stmia	r0, {r4 - r9}		@ store cp regs
 	ldmia	sp!, {r4 - r9, pc}
-ENDPROC(cpu_xsc3_do_suspend)
+SYM_FUNC_END(cpu_xsc3_do_suspend)
 
-ENTRY(cpu_xsc3_do_resume)
+SYM_TYPED_FUNC_START(cpu_xsc3_do_resume)
 	ldmia	r0, {r4 - r9}		@ load cp regs
 	mov	ip, #0
 	mcr	p15, 0, ip, c7, c7, 0	@ invalidate I & D caches, BTB
@@ -439,7 +445,7 @@ ENTRY(cpu_xsc3_do_resume)
 	mcr	p15, 0, r8, c1, c0, 1	@ auxiliary control reg
 	mov	r0, r9			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_xsc3_do_resume)
+SYM_FUNC_END(cpu_xsc3_do_resume)
 #endif
 
 	.type	__xsc3_setup, #function
diff --git a/arch/arm/mm/proc-xscale.S b/arch/arm/mm/proc-xscale.S
index bbf1e94ba554..d8462df8020b 100644
--- a/arch/arm/mm/proc-xscale.S
+++ b/arch/arm/mm/proc-xscale.S
@@ -112,22 +112,24 @@ clean_addr:	.word	CLEAN_ADDR
  *
  * Nothing too exciting at the moment
  */
-ENTRY(cpu_xscale_proc_init)
+SYM_TYPED_FUNC_START(cpu_xscale_proc_init)
 	@ enable write buffer coalescing. Some bootloader disable it
 	mrc	p15, 0, r1, c1, c0, 1
 	bic	r1, r1, #1
 	mcr	p15, 0, r1, c1, c0, 1
 	ret	lr
+SYM_FUNC_END(cpu_xscale_proc_init)
 
 /*
  * cpu_xscale_proc_fin()
  */
-ENTRY(cpu_xscale_proc_fin)
+SYM_TYPED_FUNC_START(cpu_xscale_proc_fin)
 	mrc	p15, 0, r0, c1, c0, 0		@ ctrl register
 	bic	r0, r0, #0x1800			@ ...IZ...........
 	bic	r0, r0, #0x0006			@ .............CA.
 	mcr	p15, 0, r0, c1, c0, 0		@ disable caches
 	ret	lr
+SYM_FUNC_END(cpu_xscale_proc_fin)
 
 /*
  * cpu_xscale_reset(loc)
@@ -142,7 +144,7 @@ ENTRY(cpu_xscale_proc_fin)
  */
 	.align	5
 	.pushsection	.idmap.text, "ax"
-ENTRY(cpu_xscale_reset)
+SYM_TYPED_FUNC_START(cpu_xscale_reset)
 	mov	r1, #PSR_F_BIT|PSR_I_BIT|SVC_MODE
 	msr	cpsr_c, r1			@ reset CPSR
 	mcr	p15, 0, r1, c10, c4, 1		@ unlock I-TLB
@@ -160,7 +162,7 @@ ENTRY(cpu_xscale_reset)
 	@ already containing those two last instructions to survive.
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I & D TLBs
 	ret	r0
-ENDPROC(cpu_xscale_reset)
+SYM_FUNC_END(cpu_xscale_reset)
 	.popsection
 
 /*
@@ -175,10 +177,11 @@ ENDPROC(cpu_xscale_reset)
  */
 	.align	5
 
-ENTRY(cpu_xscale_do_idle)
+SYM_TYPED_FUNC_START(cpu_xscale_do_idle)
 	mov	r0, #1
 	mcr	p14, 0, r0, c7, c0, 0		@ Go to IDLE
 	ret	lr
+SYM_FUNC_END(cpu_xscale_do_idle)
 
 /* ================================= CACHE ================================ */
 
@@ -428,12 +431,13 @@ SYM_TYPED_FUNC_START(xscale_dma_unmap_area)
 	ret	lr
 SYM_FUNC_END(xscale_dma_unmap_area)
 
-ENTRY(cpu_xscale_dcache_clean_area)
+SYM_TYPED_FUNC_START(cpu_xscale_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	add	r0, r0, #CACHELINESIZE
 	subs	r1, r1, #CACHELINESIZE
 	bhi	1b
 	ret	lr
+SYM_FUNC_END(cpu_xscale_dcache_clean_area)
 
 /* =============================== PageTable ============================== */
 
@@ -445,13 +449,14 @@ ENTRY(cpu_xscale_dcache_clean_area)
  * pgd: new page tables
  */
 	.align	5
-ENTRY(cpu_xscale_switch_mm)
+SYM_TYPED_FUNC_START(cpu_xscale_switch_mm)
 	clean_d_cache r1, r2
 	mcr	p15, 0, ip, c7, c5, 0		@ Invalidate I cache & BTB
 	mcr	p15, 0, ip, c7, c10, 4		@ Drain Write (& Fill) Buffer
 	mcr	p15, 0, r0, c2, c0, 0		@ load page table pointer
 	mcr	p15, 0, ip, c8, c7, 0		@ invalidate I & D TLBs
 	cpwait_ret lr, ip
+SYM_FUNC_END(cpu_xscale_switch_mm)
 
 /*
  * cpu_xscale_set_pte_ext(ptep, pte, ext)
@@ -479,7 +484,7 @@ cpu_xscale_mt_table:
 	.long	0x00						@ unused
 
 	.align	5
-ENTRY(cpu_xscale_set_pte_ext)
+SYM_TYPED_FUNC_START(cpu_xscale_set_pte_ext)
 	xscale_set_pte_ext_prologue
 
 	@
@@ -497,6 +502,7 @@ ENTRY(cpu_xscale_set_pte_ext)
 
 	xscale_set_pte_ext_epilogue
 	ret	lr
+SYM_FUNC_END(cpu_xscale_set_pte_ext)
 
 	.ltorg
 	.align
@@ -504,7 +510,7 @@ ENTRY(cpu_xscale_set_pte_ext)
 .globl	cpu_xscale_suspend_size
 .equ	cpu_xscale_suspend_size, 4 * 6
 #ifdef CONFIG_ARM_CPU_SUSPEND
-ENTRY(cpu_xscale_do_suspend)
+SYM_TYPED_FUNC_START(cpu_xscale_do_suspend)
 	stmfd	sp!, {r4 - r9, lr}
 	mrc	p14, 0, r4, c6, c0, 0	@ clock configuration, for turbo mode
 	mrc	p15, 0, r5, c15, c1, 0	@ CP access reg
@@ -515,9 +521,9 @@ ENTRY(cpu_xscale_do_suspend)
 	bic	r4, r4, #2		@ clear frequency change bit
 	stmia	r0, {r4 - r9}		@ store cp regs
 	ldmfd	sp!, {r4 - r9, pc}
-ENDPROC(cpu_xscale_do_suspend)
+SYM_FUNC_END(cpu_xscale_do_suspend)
 
-ENTRY(cpu_xscale_do_resume)
+SYM_TYPED_FUNC_START(cpu_xscale_do_resume)
 	ldmia	r0, {r4 - r9}		@ load cp regs
 	mov	ip, #0
 	mcr	p15, 0, ip, c8, c7, 0	@ invalidate I & D TLBs
@@ -530,7 +536,7 @@ ENTRY(cpu_xscale_do_resume)
 	mcr	p15, 0, r8, c1, c0, 1	@ auxiliary control reg
 	mov	r0, r9			@ control register
 	b	cpu_resume_mmu
-ENDPROC(cpu_xscale_do_resume)
+SYM_FUNC_END(cpu_xscale_do_resume)
 #endif
 
 	.type	__xscale_setup, #function

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 08/11] ARM: mm: Define prototypes for all per-processor calls
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (5 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 07/11] ARM: mm: Type-annotate all per-processor assembly routines Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 09/11] ARM: lib: Annotate loop delay instructions for CFI Linus Walleij
                   ` (3 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

Each CPU type ("proc") has assembly calls for initializing and
setting up the MM context, idle and so forth.

These calls have the C form of e.g.:

void cpu_arm920_init(void);

However this prototype is not really specified, instead it is
generated by the glue code in <asm/glue-proc.h> and the prototype
is implicit from the generic prototype defined in <asm/proc-fns.h>
such as cpu_proc_init() in this case. (This is a bit similar to
the "interface" or inheritance concept in other languages.)

To be able to annotate these assembly calls for CFI, they all need
to have a proper C prototype per CPU call.

Define these in a new C file that is only compiled when we use
CFI, and add __ADDRESSABLE() to each so the compiler knows that
these will be addressed (they are not explicitly called in C, they
are called by way of cpu_proc_init() etc).

It is a bit of definitions, but we do not expect new ARM32 CPUs
to appear very much so it should be pretty static.

Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/mm/Makefile |   1 +
 arch/arm/mm/proc.c   | 500 +++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 501 insertions(+)

diff --git a/arch/arm/mm/Makefile b/arch/arm/mm/Makefile
index 17665381be96..f1f231f20ff9 100644
--- a/arch/arm/mm/Makefile
+++ b/arch/arm/mm/Makefile
@@ -90,6 +90,7 @@ obj-$(CONFIG_CPU_V6)		+= proc-v6.o
 obj-$(CONFIG_CPU_V6K)		+= proc-v6.o
 obj-$(CONFIG_CPU_V7)		+= proc-v7.o proc-v7-bugs.o
 obj-$(CONFIG_CPU_V7M)		+= proc-v7m.o
+obj-$(CONFIG_CFI_CLANG)		+= proc.o
 
 obj-$(CONFIG_OUTER_CACHE)	+= l2c-common.o
 obj-$(CONFIG_CACHE_B15_RAC)	+= cache-b15-rac.o
diff --git a/arch/arm/mm/proc.c b/arch/arm/mm/proc.c
new file mode 100644
index 000000000000..bdbbf65d1b36
--- /dev/null
+++ b/arch/arm/mm/proc.c
@@ -0,0 +1,500 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file defines C prototypes for the low-level processor assembly functions
+ * and creates a reference for CFI. This needs to be done for every assembly
+ * processor ("proc") function that is called from C but does not have a
+ * corresponding C implementation.
+ *
+ * Processors are listed in the order they appear in the Makefile.
+ *
+ * Functions are listed if and only if they see use on the target CPU, and in
+ * the order they are defined in struct processor.
+ */
+#include <asm/proc-fns.h>
+
+#ifdef CONFIG_CPU_ARM7TDMI
+void cpu_arm7tdmi_proc_init(void);
+__ADDRESSABLE(cpu_arm7tdmi_proc_init);
+void cpu_arm7tdmi_proc_fin(void);
+__ADDRESSABLE(cpu_arm7tdmi_proc_fin);
+void cpu_arm7tdmi_reset(void);
+__ADDRESSABLE(cpu_arm7tdmi_reset);
+int cpu_arm7tdmi_do_idle(void);
+__ADDRESSABLE(cpu_arm7tdmi_do_idle);
+void cpu_arm7tdmi_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm7tdmi_dcache_clean_area);
+void cpu_arm7tdmi_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm7tdmi_switch_mm);
+#endif
+
+#ifdef CONFIG_CPU_ARM720T
+void cpu_arm720_proc_init(void);
+__ADDRESSABLE(cpu_arm720_proc_init);
+void cpu_arm720_proc_fin(void);
+__ADDRESSABLE(cpu_arm720_proc_fin);
+void cpu_arm720_reset(void);
+__ADDRESSABLE(cpu_arm720_reset);
+int cpu_arm720_do_idle(void);
+__ADDRESSABLE(cpu_arm720_do_idle);
+void cpu_arm720_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm720_dcache_clean_area);
+void cpu_arm720_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm720_switch_mm);
+void cpu_arm720_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm720_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_ARM740T
+void cpu_arm740_proc_init(void);
+__ADDRESSABLE(cpu_arm740_proc_init);
+void cpu_arm740_proc_fin(void);
+__ADDRESSABLE(cpu_arm740_proc_fin);
+void cpu_arm740_reset(void);
+__ADDRESSABLE(cpu_arm740_reset);
+int cpu_arm740_do_idle(void);
+__ADDRESSABLE(cpu_arm740_do_idle);
+void cpu_arm740_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm740_dcache_clean_area);
+void cpu_arm740_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm740_switch_mm);
+#endif
+
+#ifdef CONFIG_CPU_ARM9TDMI
+void cpu_arm9tdmi_proc_init(void);
+__ADDRESSABLE(cpu_arm9tdmi_proc_init);
+void cpu_arm9tdmi_proc_fin(void);
+__ADDRESSABLE(cpu_arm9tdmi_proc_fin);
+void cpu_arm9tdmi_reset(void);
+__ADDRESSABLE(cpu_arm9tdmi_reset);
+int cpu_arm9tdmi_do_idle(void);
+__ADDRESSABLE(cpu_arm9tdmi_do_idle);
+void cpu_arm9tdmi_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm9tdmi_dcache_clean_area);
+void cpu_arm9tdmi_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm9tdmi_switch_mm);
+#endif
+
+#ifdef CONFIG_CPU_ARM920T
+void cpu_arm920_proc_init(void);
+__ADDRESSABLE(cpu_arm920_proc_init);
+void cpu_arm920_proc_fin(void);
+__ADDRESSABLE(cpu_arm920_proc_fin);
+void cpu_arm920_reset(void);
+__ADDRESSABLE(cpu_arm920_reset);
+int cpu_arm920_do_idle(void);
+__ADDRESSABLE(cpu_arm920_do_idle);
+void cpu_arm920_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm920_dcache_clean_area);
+void cpu_arm920_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm920_switch_mm);
+void cpu_arm920_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm920_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_arm920_do_suspend(void *);
+__ADDRESSABLE(cpu_arm920_do_suspend);
+void cpu_arm920_do_resume(void *);
+__ADDRESSABLE(cpu_arm920_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_ARM920T */
+
+#ifdef CONFIG_CPU_ARM922T
+void cpu_arm922_proc_init(void);
+__ADDRESSABLE(cpu_arm922_proc_init);
+void cpu_arm922_proc_fin(void);
+__ADDRESSABLE(cpu_arm922_proc_fin);
+void cpu_arm922_reset(void);
+__ADDRESSABLE(cpu_arm922_reset);
+int cpu_arm922_do_idle(void);
+__ADDRESSABLE(cpu_arm922_do_idle);
+void cpu_arm922_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm922_dcache_clean_area);
+void cpu_arm922_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm922_switch_mm);
+void cpu_arm922_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm922_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_ARM925T
+void cpu_arm925_proc_init(void);
+__ADDRESSABLE(cpu_arm925_proc_init);
+void cpu_arm925_proc_fin(void);
+__ADDRESSABLE(cpu_arm925_proc_fin);
+void cpu_arm925_reset(void);
+__ADDRESSABLE(cpu_arm925_reset);
+int cpu_arm925_do_idle(void);
+__ADDRESSABLE(cpu_arm925_do_idle);
+void cpu_arm925_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm925_dcache_clean_area);
+void cpu_arm925_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm925_switch_mm);
+void cpu_arm925_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm925_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_ARM926T
+void cpu_arm926_proc_init(void);
+__ADDRESSABLE(cpu_arm926_proc_init);
+void cpu_arm926_proc_fin(void);
+__ADDRESSABLE(cpu_arm926_proc_fin);
+void cpu_arm926_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_arm926_reset);
+int cpu_arm926_do_idle(void);
+__ADDRESSABLE(cpu_arm926_do_idle);
+void cpu_arm926_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm926_dcache_clean_area);
+void cpu_arm926_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm926_switch_mm);
+void cpu_arm926_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm926_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_arm926_do_suspend(void *);
+__ADDRESSABLE(cpu_arm926_do_suspend);
+void cpu_arm926_do_resume(void *);
+__ADDRESSABLE(cpu_arm926_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_ARM926T */
+
+#ifdef CONFIG_CPU_ARM940T
+void cpu_arm940_proc_init(void);
+__ADDRESSABLE(cpu_arm940_proc_init);
+void cpu_arm940_proc_fin(void);
+__ADDRESSABLE(cpu_arm940_proc_fin);
+void cpu_arm940_reset(void);
+__ADDRESSABLE(cpu_arm940_reset);
+int cpu_arm940_do_idle(void);
+__ADDRESSABLE(cpu_arm940_do_idle);
+void cpu_arm940_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm940_dcache_clean_area);
+void cpu_arm940_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm940_switch_mm);
+#endif
+
+#ifdef CONFIG_CPU_ARM946E
+void cpu_arm946_proc_init(void);
+__ADDRESSABLE(cpu_arm946_proc_init);
+void cpu_arm946_proc_fin(void);
+__ADDRESSABLE(cpu_arm946_proc_fin);
+void cpu_arm946_reset(void);
+__ADDRESSABLE(cpu_arm946_reset);
+int cpu_arm946_do_idle(void);
+__ADDRESSABLE(cpu_arm946_do_idle);
+void cpu_arm946_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm946_dcache_clean_area);
+void cpu_arm946_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm946_switch_mm);
+#endif
+
+#ifdef CONFIG_CPU_FA526
+void cpu_fa526_proc_init(void);
+__ADDRESSABLE(cpu_fa526_proc_init);
+void cpu_fa526_proc_fin(void);
+__ADDRESSABLE(cpu_fa526_proc_fin);
+void cpu_fa526_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_fa526_reset);
+int cpu_fa526_do_idle(void);
+__ADDRESSABLE(cpu_fa526_do_idle);
+void cpu_fa526_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_fa526_dcache_clean_area);
+void cpu_fa526_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_fa526_switch_mm);
+void cpu_fa526_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_fa526_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_ARM1020
+void cpu_arm1020_proc_init(void);
+__ADDRESSABLE(cpu_arm1020_proc_init);
+void cpu_arm1020_proc_fin(void);
+__ADDRESSABLE(cpu_arm1020_proc_fin);
+void cpu_arm1020_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_arm1020_reset);
+int cpu_arm1020_do_idle(void);
+__ADDRESSABLE(cpu_arm1020_do_idle);
+void cpu_arm1020_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm1020_dcache_clean_area);
+void cpu_arm1020_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm1020_switch_mm);
+void cpu_arm1020_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm1020_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_ARM1020E
+void cpu_arm1020e_proc_init(void);
+__ADDRESSABLE(cpu_arm1020e_proc_init);
+void cpu_arm1020e_proc_fin(void);
+__ADDRESSABLE(cpu_arm1020e_proc_fin);
+void cpu_arm1020e_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_arm1020e_reset);
+int cpu_arm1020e_do_idle(void);
+__ADDRESSABLE(cpu_arm1020e_do_idle);
+void cpu_arm1020e_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm1020e_dcache_clean_area);
+void cpu_arm1020e_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm1020e_switch_mm);
+void cpu_arm1020e_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm1020e_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_ARM1022
+void cpu_arm1022_proc_init(void);
+__ADDRESSABLE(cpu_arm1022_proc_init);
+void cpu_arm1022_proc_fin(void);
+__ADDRESSABLE(cpu_arm1022_proc_fin);
+void cpu_arm1022_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_arm1022_reset);
+int cpu_arm1022_do_idle(void);
+__ADDRESSABLE(cpu_arm1022_do_idle);
+void cpu_arm1022_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm1022_dcache_clean_area);
+void cpu_arm1022_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm1022_switch_mm);
+void cpu_arm1022_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm1022_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_ARM1026
+void cpu_arm1026_proc_init(void);
+__ADDRESSABLE(cpu_arm1026_proc_init);
+void cpu_arm1026_proc_fin(void);
+__ADDRESSABLE(cpu_arm1026_proc_fin);
+void cpu_arm1026_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_arm1026_reset);
+int cpu_arm1026_do_idle(void);
+__ADDRESSABLE(cpu_arm1026_do_idle);
+void cpu_arm1026_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_arm1026_dcache_clean_area);
+void cpu_arm1026_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_arm1026_switch_mm);
+void cpu_arm1026_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_arm1026_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_SA110
+void cpu_sa110_proc_init(void);
+__ADDRESSABLE(cpu_sa110_proc_init);
+void cpu_sa110_proc_fin(void);
+__ADDRESSABLE(cpu_sa110_proc_fin);
+void cpu_sa110_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_sa110_reset);
+int cpu_sa110_do_idle(void);
+__ADDRESSABLE(cpu_sa110_do_idle);
+void cpu_sa110_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_sa110_dcache_clean_area);
+void cpu_sa110_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_sa110_switch_mm);
+void cpu_sa110_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_sa110_set_pte_ext);
+#endif
+
+#ifdef CONFIG_CPU_SA1100
+void cpu_sa1100_proc_init(void);
+__ADDRESSABLE(cpu_sa1100_proc_init);
+void cpu_sa1100_proc_fin(void);
+__ADDRESSABLE(cpu_sa1100_proc_fin);
+void cpu_sa1100_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_sa1100_reset);
+int cpu_sa1100_do_idle(void);
+__ADDRESSABLE(cpu_sa1100_do_idle);
+void cpu_sa1100_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_sa1100_dcache_clean_area);
+void cpu_sa1100_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_sa1100_switch_mm);
+void cpu_sa1100_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_sa1100_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_sa1100_do_suspend(void *);
+__ADDRESSABLE(cpu_sa1100_do_suspend);
+void cpu_sa1100_do_resume(void *);
+__ADDRESSABLE(cpu_sa1100_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_SA1100 */
+
+#ifdef CONFIG_CPU_XSCALE
+void cpu_xscale_proc_init(void);
+__ADDRESSABLE(cpu_xscale_proc_init);
+void cpu_xscale_proc_fin(void);
+__ADDRESSABLE(cpu_xscale_proc_fin);
+void cpu_xscale_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_xscale_reset);
+int cpu_xscale_do_idle(void);
+__ADDRESSABLE(cpu_xscale_do_idle);
+void cpu_xscale_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_xscale_dcache_clean_area);
+void cpu_xscale_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_xscale_switch_mm);
+void cpu_xscale_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_xscale_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_xscale_do_suspend(void *);
+__ADDRESSABLE(cpu_xscale_do_suspend);
+void cpu_xscale_do_resume(void *);
+__ADDRESSABLE(cpu_xscale_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_XSCALE */
+
+#ifdef CONFIG_CPU_XSC3
+void cpu_xsc3_proc_init(void);
+__ADDRESSABLE(cpu_xsc3_proc_init);
+void cpu_xsc3_proc_fin(void);
+__ADDRESSABLE(cpu_xsc3_proc_fin);
+void cpu_xsc3_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_xsc3_reset);
+int cpu_xsc3_do_idle(void);
+__ADDRESSABLE(cpu_xsc3_do_idle);
+void cpu_xsc3_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_xsc3_dcache_clean_area);
+void cpu_xsc3_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_xsc3_switch_mm);
+void cpu_xsc3_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_xsc3_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_xsc3_do_suspend(void *);
+__ADDRESSABLE(cpu_xsc3_do_suspend);
+void cpu_xsc3_do_resume(void *);
+__ADDRESSABLE(cpu_xsc3_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_XSC3 */
+
+#ifdef CONFIG_CPU_MOHAWK
+void cpu_mohawk_proc_init(void);
+__ADDRESSABLE(cpu_mohawk_proc_init);
+void cpu_mohawk_proc_fin(void);
+__ADDRESSABLE(cpu_mohawk_proc_fin);
+void cpu_mohawk_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_mohawk_reset);
+int cpu_mohawk_do_idle(void);
+__ADDRESSABLE(cpu_mohawk_do_idle);
+void cpu_mohawk_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_mohawk_dcache_clean_area);
+void cpu_mohawk_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_mohawk_switch_mm);
+void cpu_mohawk_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_mohawk_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_mohawk_do_suspend(void *);
+__ADDRESSABLE(cpu_mohawk_do_suspend);
+void cpu_mohawk_do_resume(void *);
+__ADDRESSABLE(cpu_mohawk_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_MOHAWK */
+
+#ifdef CONFIG_CPU_FEROCEON
+void cpu_feroceon_proc_init(void);
+__ADDRESSABLE(cpu_feroceon_proc_init);
+void cpu_feroceon_proc_fin(void);
+__ADDRESSABLE(cpu_feroceon_proc_fin);
+void cpu_feroceon_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_feroceon_reset);
+int cpu_feroceon_do_idle(void);
+__ADDRESSABLE(cpu_feroceon_do_idle);
+void cpu_feroceon_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_feroceon_dcache_clean_area);
+void cpu_feroceon_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_feroceon_switch_mm);
+void cpu_feroceon_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_feroceon_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_feroceon_do_suspend(void *);
+__ADDRESSABLE(cpu_feroceon_do_suspend);
+void cpu_feroceon_do_resume(void *);
+__ADDRESSABLE(cpu_feroceon_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_FEROCEON */
+
+#if defined(CONFIG_CPU_V6) || defined(CONFIG_CPU_V6K)
+void cpu_v6_proc_init(void);
+__ADDRESSABLE(cpu_v6_proc_init);
+void cpu_v6_proc_fin(void);
+__ADDRESSABLE(cpu_v6_proc_fin);
+void cpu_v6_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_v6_reset);
+int cpu_v6_do_idle(void);
+__ADDRESSABLE(cpu_v6_do_idle);
+void cpu_v6_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_v6_dcache_clean_area);
+void cpu_v6_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_v6_switch_mm);
+void cpu_v6_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_v6_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_v6_do_suspend(void *);
+__ADDRESSABLE(cpu_v6_do_suspend);
+void cpu_v6_do_resume(void *);
+__ADDRESSABLE(cpu_v6_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CPU_V6 */
+
+#ifdef CONFIG_CPU_V7
+void cpu_v7_proc_init(void);
+__ADDRESSABLE(cpu_v7_proc_init);
+void cpu_v7_proc_fin(void);
+__ADDRESSABLE(cpu_v7_proc_fin);
+void cpu_v7_reset(void);
+__ADDRESSABLE(cpu_v7_reset);
+int cpu_v7_do_idle(void);
+__ADDRESSABLE(cpu_v7_do_idle);
+#ifdef CONFIG_PJ4B_ERRATA_4742
+int cpu_pj4b_do_idle(void);
+__ADDRESSABLE(cpu_pj4b_do_idle);
+#endif
+void cpu_v7_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_v7_dcache_clean_area);
+void cpu_v7_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+/* Special switch_mm() callbacks to work around bugs in v7 */
+__ADDRESSABLE(cpu_v7_switch_mm);
+void cpu_v7_iciallu_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_v7_iciallu_switch_mm);
+void cpu_v7_bpiall_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_v7_bpiall_switch_mm);
+#ifdef CONFIG_ARM_LPAE
+void cpu_v7_set_pte_ext(pte_t *ptep, pte_t pte);
+#else
+void cpu_v7_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+#endif
+__ADDRESSABLE(cpu_v7_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_v7_do_suspend(void *);
+__ADDRESSABLE(cpu_v7_do_suspend);
+void cpu_v7_do_resume(void *);
+__ADDRESSABLE(cpu_v7_do_resume);
+/* Special versions of suspend and resume for the CA9MP cores */
+void cpu_ca9mp_do_suspend(void *);
+__ADDRESSABLE(cpu_ca9mp_do_suspend);
+void cpu_ca9mp_do_resume(void *);
+__ADDRESSABLE(cpu_ca9mp_do_resume);
+/* Special versions of suspend and resume for the Marvell PJ4B cores */
+#ifdef CONFIG_CPU_PJ4B
+void cpu_pj4b_do_suspend(void *);
+__ADDRESSABLE(cpu_pj4b_do_suspend);
+void cpu_pj4b_do_resume(void *);
+__ADDRESSABLE(cpu_pj4b_do_resume);
+#endif /* CONFIG_CPU_PJ4B */
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+#endif /* CONFIG_CPU_V7 */
+
+#ifdef CONFIG_CPU_V7M
+void cpu_v7m_proc_init(void);
+__ADDRESSABLE(cpu_v7m_proc_init);
+void cpu_v7m_proc_fin(void);
+__ADDRESSABLE(cpu_v7m_proc_fin);
+void cpu_v7m_reset(unsigned long addr, bool hvc);
+__ADDRESSABLE(cpu_v7m_reset);
+int cpu_v7m_do_idle(void);
+__ADDRESSABLE(cpu_v7m_do_idle);
+void cpu_v7m_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_v7m_dcache_clean_area);
+void cpu_v7m_switch_mm(phys_addr_t pgd_phys, struct mm_struct *mm);
+__ADDRESSABLE(cpu_v7m_switch_mm);
+void cpu_v7m_set_pte_ext(pte_t *ptep, pte_t pte, unsigned int ext);
+__ADDRESSABLE(cpu_v7m_set_pte_ext);
+#ifdef CONFIG_ARM_CPU_SUSPEND
+void cpu_v7m_do_suspend(void *);
+__ADDRESSABLE(cpu_v7m_do_suspend);
+void cpu_v7m_do_resume(void *);
+__ADDRESSABLE(cpu_v7m_do_resume);
+#endif /* CONFIG_ARM_CPU_SUSPEND */
+void cpu_cm7_proc_fin(void);
+__ADDRESSABLE(cpu_cm7_proc_fin);
+void cpu_cm7_dcache_clean_area(void *addr, int size);
+__ADDRESSABLE(cpu_cm7_dcache_clean_area);
+#endif /* CONFIG_CPU_V7M */

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 09/11] ARM: lib: Annotate loop delay instructions for CFI
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (6 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 08/11] ARM: mm: Define prototypes for all per-processor calls Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 10/11] ARM: hw_breakpoint: Handle CFI breakpoints Linus Walleij
                   ` (2 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

When we annotate the loop delay code with SYM_TYPED_FUNC_START()
a function prototype signature will be emitted into the object
file above each site called from C, and the delay loop code is
using "fallthroughs" from the different assembly callbacks. This
will not work as the execution flow will run into the prototype
signatures.

Rewrite the code to use explicit branches to the other code
segments and annotate the code using SYM_TYPED_FUNC_START().

Tested on the ARM Versatile which uses the calibrated loop delay.

Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/lib/delay-loop.S | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/arch/arm/lib/delay-loop.S b/arch/arm/lib/delay-loop.S
index 3ac05177d097..33b08ca1c242 100644
--- a/arch/arm/lib/delay-loop.S
+++ b/arch/arm/lib/delay-loop.S
@@ -5,6 +5,7 @@
  *  Copyright (C) 1995, 1996 Russell King
  */
 #include <linux/linkage.h>
+#include <linux/cfi_types.h>
 #include <asm/assembler.h>
 #include <asm/delay.h>
 
@@ -24,21 +25,26 @@
  * HZ  <= 1000
  */
 
-ENTRY(__loop_udelay)
+SYM_TYPED_FUNC_START(__loop_udelay)
 		ldr	r2, .LC1
 		mul	r0, r2, r0		@ r0 = delay_us * UDELAY_MULT
-ENTRY(__loop_const_udelay)			@ 0 <= r0 <= 0xfffffaf0
+		b	__loop_const_udelay
+SYM_FUNC_END(__loop_udelay)
+
+SYM_TYPED_FUNC_START(__loop_const_udelay)	@ 0 <= r0 <= 0xfffffaf0
 		ldr	r2, .LC0
 		ldr	r2, [r2]
 		umull	r1, r0, r2, r0		@ r0-r1 = r0 * loops_per_jiffy
 		adds	r1, r1, #0xffffffff	@ rounding up ...
 		adcs	r0, r0, r0		@ and right shift by 31
 		reteq	lr
+		b	__loop_delay
+SYM_FUNC_END(__loop_const_udelay)
 
 		.align 3
 
 @ Delay routine
-ENTRY(__loop_delay)
+SYM_TYPED_FUNC_START(__loop_delay)
 		subs	r0, r0, #1
 #if 0
 		retls	lr
@@ -58,6 +64,4 @@ ENTRY(__loop_delay)
 #endif
 		bhi	__loop_delay
 		ret	lr
-ENDPROC(__loop_udelay)
-ENDPROC(__loop_const_udelay)
-ENDPROC(__loop_delay)
+SYM_FUNC_END(__loop_delay)

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 10/11] ARM: hw_breakpoint: Handle CFI breakpoints
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (7 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 09/11] ARM: lib: Annotate loop delay instructions for CFI Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-21 18:28 ` [PATCH v7 11/11] ARM: Support CLANG CFI Linus Walleij
  2024-04-22 21:50 ` [PATCH v7 00/11] CFI for ARM32 using LLVM Sami Tolvanen
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

This registers a breakpoint handler for the new breakpoint type
(0x03) inserted by LLVM CLANG for CFI breakpoints.

If we are in permissive mode, just print a backtrace and continue.

Example with CONFIG_CFI_PERMISSIVE enabled:

> echo CFI_FORWARD_PROTO > /sys/kernel/debug/provoke-crash/DIRECT
lkdtm: Performing direct entry CFI_FORWARD_PROTO
lkdtm: Calling matched prototype ...
lkdtm: Calling mismatched prototype ...
CFI failure at lkdtm_indirect_call+0x40/0x4c (target: 0x0; expected type: 0x00000000)
WARNING: CPU: 1 PID: 112 at lkdtm_indirect_call+0x40/0x4c
CPU: 1 PID: 112 Comm: sh Not tainted 6.8.0-rc1+ #150
Hardware name: ARM-Versatile Express
(...)
lkdtm: FAIL: survived mismatched prototype function call!
lkdtm: Unexpected! This kernel (6.8.0-rc1+ armv7l) was built with CONFIG_CFI_CLANG=y

As you can see the LKDTM test fails, but I expect that this would be
expected behaviour in the permissive mode.

We are currently not implementing target and type for the CFI
breakpoint as this requires additional operand bundling compiler
extensions.

CPUs without breakpoint support cannot handle breakpoints naturally,
in these cases the permissive mode will not work, CFI will fall over
on an undefined instruction:

Internal error: Oops - undefined instruction: 0 [#1] PREEMPT ARM
CPU: 0 PID: 186 Comm: ash Tainted: G        W          6.9.0-rc1+ #7
Hardware name: Gemini (Device Tree)
PC is at lkdtm_indirect_call+0x38/0x4c
LR is at lkdtm_CFI_FORWARD_PROTO+0x30/0x6c

This is reasonable I think: it's the best CFI can do to ascertain
the the control flow is not broken on these CPUs.

Reviewed-by: Kees Cook <keescook@chromium.org>
Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/include/asm/hw_breakpoint.h |  1 +
 arch/arm/kernel/hw_breakpoint.c      | 35 +++++++++++++++++++++++++++++++++++
 2 files changed, 36 insertions(+)

diff --git a/arch/arm/include/asm/hw_breakpoint.h b/arch/arm/include/asm/hw_breakpoint.h
index 62358d3ca0a8..e7f9961c53b2 100644
--- a/arch/arm/include/asm/hw_breakpoint.h
+++ b/arch/arm/include/asm/hw_breakpoint.h
@@ -84,6 +84,7 @@ static inline void decode_ctrl_reg(u32 reg,
 #define ARM_DSCR_MOE(x)			((x >> 2) & 0xf)
 #define ARM_ENTRY_BREAKPOINT		0x1
 #define ARM_ENTRY_ASYNC_WATCHPOINT	0x2
+#define ARM_ENTRY_CFI_BREAKPOINT	0x3
 #define ARM_ENTRY_SYNC_WATCHPOINT	0xa
 
 /* DSCR monitor/halting bits. */
diff --git a/arch/arm/kernel/hw_breakpoint.c b/arch/arm/kernel/hw_breakpoint.c
index dc0fb7a81371..93f9034e4e72 100644
--- a/arch/arm/kernel/hw_breakpoint.c
+++ b/arch/arm/kernel/hw_breakpoint.c
@@ -17,6 +17,7 @@
 #include <linux/perf_event.h>
 #include <linux/hw_breakpoint.h>
 #include <linux/smp.h>
+#include <linux/cfi.h>
 #include <linux/cpu_pm.h>
 #include <linux/coresight.h>
 
@@ -903,6 +904,37 @@ static void breakpoint_handler(unsigned long unknown, struct pt_regs *regs)
 	watchpoint_single_step_handler(addr);
 }
 
+#ifdef CONFIG_CFI_CLANG
+static void hw_breakpoint_cfi_handler(struct pt_regs *regs)
+{
+	/*
+	 * TODO: implementing target and type to pass to CFI using the more
+	 * elaborate report_cfi_failure() requires compiler work. To be able
+	 * to properly extract target information the compiler needs to
+	 * emit a stable instructions sequence for the CFI checks so we can
+	 * decode the instructions preceding the trap and figure out which
+	 * registers were used.
+	 */
+
+	switch (report_cfi_failure_noaddr(regs, instruction_pointer(regs))) {
+	case BUG_TRAP_TYPE_BUG:
+		die("Oops - CFI", regs, 0);
+		break;
+	case BUG_TRAP_TYPE_WARN:
+		/* Skip the breaking instruction */
+		instruction_pointer(regs) += 4;
+		break;
+	default:
+		die("Unknown CFI error", regs, 0);
+		break;
+	}
+}
+#else
+static void hw_breakpoint_cfi_handler(struct pt_regs *regs)
+{
+}
+#endif
+
 /*
  * Called from either the Data Abort Handler [watchpoint] or the
  * Prefetch Abort Handler [breakpoint] with interrupts disabled.
@@ -932,6 +964,9 @@ static int hw_breakpoint_pending(unsigned long addr, unsigned int fsr,
 	case ARM_ENTRY_SYNC_WATCHPOINT:
 		watchpoint_handler(addr, fsr, regs);
 		break;
+	case ARM_ENTRY_CFI_BREAKPOINT:
+		hw_breakpoint_cfi_handler(regs);
+		break;
 	default:
 		ret = 1; /* Unhandled fault. */
 	}

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* [PATCH v7 11/11] ARM: Support CLANG CFI
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (8 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 10/11] ARM: hw_breakpoint: Handle CFI breakpoints Linus Walleij
@ 2024-04-21 18:28 ` Linus Walleij
  2024-04-22 21:50 ` [PATCH v7 00/11] CFI for ARM32 using LLVM Sami Tolvanen
  10 siblings, 0 replies; 14+ messages in thread
From: Linus Walleij @ 2024-04-21 18:28 UTC (permalink / raw
  To: Russell King, Sami Tolvanen, Kees Cook, Nathan Chancellor,
	Nick Desaulniers, Ard Biesheuvel, Arnd Bergmann
  Cc: linux-arm-kernel, llvm, Linus Walleij

Support Control Flow Integrity (CFI) when compiling with
CLANG.

In the as-of-writing LLVM CLANG implementation (v17)
the 32-bit ARM platform is supported by the generic CFI
implementation, which isn't tailored specifically for ARM32
but works well enough to enable the feature.

Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index b14aed3a17ab..df7bd07ad0d4 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -35,6 +35,7 @@ config ARM
 	select ARCH_OPTIONAL_KERNEL_RWX if ARCH_HAS_STRICT_KERNEL_RWX
 	select ARCH_OPTIONAL_KERNEL_RWX_DEFAULT if CPU_V7
 	select ARCH_SUPPORTS_ATOMIC_RMW
+	select ARCH_SUPPORTS_CFI_CLANG
 	select ARCH_SUPPORTS_HUGETLBFS if ARM_LPAE
 	select ARCH_SUPPORTS_PER_VMA_LOCK
 	select ARCH_USE_BUILTIN_BSWAP

-- 
2.44.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* Re: [PATCH v7 07/11] ARM: mm: Type-annotate all per-processor assembly routines
  2024-04-21 18:28 ` [PATCH v7 07/11] ARM: mm: Type-annotate all per-processor assembly routines Linus Walleij
@ 2024-04-22  6:00   ` Arnd Bergmann
  0 siblings, 0 replies; 14+ messages in thread
From: Arnd Bergmann @ 2024-04-22  6:00 UTC (permalink / raw
  To: Linus Walleij, Russell King, Sami Tolvanen, Kees Cook,
	Nathan Chancellor, Nick Desaulniers, Ard Biesheuvel
  Cc: linux-arm-kernel, llvm

On Sun, Apr 21, 2024, at 20:28, Linus Walleij wrote:
> Type tag the remaining per-processor assembly using the CFI
> symbol macros, in addition to those that were previously tagged
> for cache maintenance calls.
>
> This will be used to finally provide proper C prototypes for
> all these calls as well so that CFI can be made to work.
>
> Tested-by: Kees Cook <keescook@chromium.org>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

The patch looks fine to me:

Acked-by: Arnd Bergmann <arnd@arndb.de>

> ---
>  arch/arm/mm/proc-arm1020.S   | 24 ++++++++++------
>  arch/arm/mm/proc-arm1020e.S  | 24 ++++++++++------
>  arch/arm/mm/proc-arm1022.S   | 24 ++++++++++------
>  arch/arm/mm/proc-arm1026.S   | 24 ++++++++++------
>  arch/arm/mm/proc-arm720.S    | 25 +++++++++++------
>  arch/arm/mm/proc-arm740.S    | 26 ++++++++++++-----
>  arch/arm/mm/proc-arm7tdmi.S  | 34 +++++++++++++++--------
>  arch/arm/mm/proc-arm920.S    | 31 ++++++++++++---------
>  arch/arm/mm/proc-arm922.S    | 23 +++++++++------
>  arch/arm/mm/proc-arm925.S    | 22 +++++++++------
>  arch/arm/mm/proc-arm926.S    | 31 +++++++++++++--------
>  arch/arm/mm/proc-arm940.S    | 21 +++++++++-----
>  arch/arm/mm/proc-arm946.S    | 21 +++++++++-----
>  arch/arm/mm/proc-arm9tdmi.S  | 26 ++++++++++++-----

I wonder if we should use the opportunity to clean out the
unused cores. Most of the above are not part of any SoC
in the kernel, the only ones that are actually used are
arm720 and arm92x, the arm10 variants are supported by
integrator but that is not really useful by itself.

       Arnd

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH v7 01/11] ARM: bugs: Check in the vtable instead of defined aliases
  2024-04-21 18:28 ` [PATCH v7 01/11] ARM: bugs: Check in the vtable instead of defined aliases Linus Walleij
@ 2024-04-22 21:27   ` Sami Tolvanen
  0 siblings, 0 replies; 14+ messages in thread
From: Sami Tolvanen @ 2024-04-22 21:27 UTC (permalink / raw
  To: Linus Walleij
  Cc: Russell King, Kees Cook, Nathan Chancellor, Nick Desaulniers,
	Ard Biesheuvel, Arnd Bergmann, linux-arm-kernel, llvm

On Sun, Apr 21, 2024 at 11:28 AM Linus Walleij <linus.walleij@linaro.org> wrote:
>
> Instead of checking if cpu_check_bugs() exist, check for this
> callback directly in the CPU vtable: this is better because the
> function is just a define to the vtable entry and this is why
> the code works. But we want to be able to specify a proper
> function for cpu_check_bugs() so look into the vtable instead.
>
> In bugs.c assign PROC_VTABLE(switch_mm) instead of
> assigning cpu_do_switch_mm where again this is just a define
> into the vtable: this makes it possible to make
> cpu_do_switch_mm() into a real function.

Is this patch still needed? None of the remaining patches in this
series define these as functions anymore.

Sami

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH v7 00/11] CFI for ARM32 using LLVM
  2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
                   ` (9 preceding siblings ...)
  2024-04-21 18:28 ` [PATCH v7 11/11] ARM: Support CLANG CFI Linus Walleij
@ 2024-04-22 21:50 ` Sami Tolvanen
  10 siblings, 0 replies; 14+ messages in thread
From: Sami Tolvanen @ 2024-04-22 21:50 UTC (permalink / raw
  To: Linus Walleij
  Cc: Russell King, Kees Cook, Nathan Chancellor, Nick Desaulniers,
	Ard Biesheuvel, Arnd Bergmann, linux-arm-kernel, llvm

On Sun, Apr 21, 2024 at 11:28 AM Linus Walleij <linus.walleij@linaro.org> wrote:
>
> This is a first patch set to support CLANG CFI (Control Flow
> Integrity) on ARM32.
>
> For information about what CFI is, see:
> https://clang.llvm.org/docs/ControlFlowIntegrity.html
>
> For the kernel KCFI flavor, see:
> https://lwn.net/Articles/898040/
>
> The base changes required to bring up KCFI on ARM32 was mostly
> related to the use of custom vtables in the kernel, combined
> with defines to call into these vtable members directly from
> sites where they are used.
>
> We annotate all assembly calls that are called directly from
> C with SYM_TYPED_FUNC_START()/SYM_FUNC_END() so it is easy
> to see while reading the assembly that these functions are
> called from C and can have CFI prototype information prefixed
> to them.
>
> As protype prefix information is just some random bytes, it is
> not possible to "fall through" into an assembly function that
> is tagged with SYM_TYPED_FUNC_START(): there will be some
> binary noise in front of the function so this design pattern
> needs to be explicitly avoided at each site where it occurred.
>
> The approach to binding the calls to C is two-fold:
>
> - Either convert the affected vtable struct to C and provide
>   per-CPU prototypes for all the calls (done for TLB, cache)
>   or:
>
> - Provide prototypes in a special files just for CFI and tag
>   all these functions addressable.
>
> The permissive mode handles the new breakpoint type (0x03) that
> LLVM CLANG is emitting.
>
> To runtime-test the patches:
> - Enable CONFIG_LKDTM
> - echo CFI_FORWARD_PROTO > /sys/kernel/debug/provoke-crash/DIRECT
>
> The patch set has been booted to userspace on the following
> test platforms:
>
> - Arm Versatile (QEMU)
> - Arm Versatile Express (QEMU)
> - multi_v7 booted on Versatile Express (QEMU)
> - Footbridge Netwinder (SA110 ARMv4)
> - Ux500 (ARMv7 SMP)
> - Gemini (FA526)
>
> I am not saying there will not be corner cases that we need
> to fix in addition to this, but it is enough to get started.
> Looking at what was fixed for arm64 I am a bit weary that
> e.g. BPF might need something to trampoline properly.

CFI is disabled for BPF calls by default on architectures that don't
define __bpfcall, so it should mostly work. However, it should be
possible to add CFI support to the 32-bit arm BPF JIT following the
riscv and arm64 examples.

> But hopefullt people can get to testing it and help me fix
> remaining issues before the final version, or we can fix it
> in-tree.

Overall these patches look pretty good to me now, although the first
one doesn't seem necessary anymore. I also suspect some of the
function declaration lists could be made slightly less verbose with
macros, but that makes it harder to grep for them, so the current
format works for me too. For the series:

Reviewed-by: Sami Tolvanen <samitolvanen@google.com>

Sami

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 14+ messages in thread

end of thread, other threads:[~2024-04-22 21:51 UTC | newest]

Thread overview: 14+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-04-21 18:28 [PATCH v7 00/11] CFI for ARM32 using LLVM Linus Walleij
2024-04-21 18:28 ` [PATCH v7 01/11] ARM: bugs: Check in the vtable instead of defined aliases Linus Walleij
2024-04-22 21:27   ` Sami Tolvanen
2024-04-21 18:28 ` [PATCH v7 02/11] ARM: ftrace: Define ftrace_stub_graph Linus Walleij
2024-04-21 18:28 ` [PATCH v7 03/11] ARM: mm: Make tlbflush routines CFI safe Linus Walleij
2024-04-21 18:28 ` [PATCH v7 05/11] ARM: mm: Use symbol alias for two cache functions Linus Walleij
2024-04-21 18:28 ` [PATCH v7 06/11] ARM: mm: Rewrite cacheflush vtables in CFI safe C Linus Walleij
2024-04-21 18:28 ` [PATCH v7 07/11] ARM: mm: Type-annotate all per-processor assembly routines Linus Walleij
2024-04-22  6:00   ` Arnd Bergmann
2024-04-21 18:28 ` [PATCH v7 08/11] ARM: mm: Define prototypes for all per-processor calls Linus Walleij
2024-04-21 18:28 ` [PATCH v7 09/11] ARM: lib: Annotate loop delay instructions for CFI Linus Walleij
2024-04-21 18:28 ` [PATCH v7 10/11] ARM: hw_breakpoint: Handle CFI breakpoints Linus Walleij
2024-04-21 18:28 ` [PATCH v7 11/11] ARM: Support CLANG CFI Linus Walleij
2024-04-22 21:50 ` [PATCH v7 00/11] CFI for ARM32 using LLVM Sami Tolvanen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).