highperfocused/ffmpeg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

arm64: Add Armv8.3-A PAC support to assembly files

Browse Source 
This patch adds optional support for Arm Pointer Authentication Codes.

PAC support is turned on or off at compile time using additional
compiler flags. Unless any of these is enabled explicitly, no additional
code will be emitted at all.

Signed-off-by: André Kempe <andre.kempe@arm.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
This commit is contained in:
Andre Kempe
2022-02-14 12:28:04 +00:00
committed by Martin Storsjö
parent 74117abf0c
commit 248986a0db
4 changed files with 113 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
libavcodec/aarch64/fft_neon.S
View File

@@ -340,7 +340,7 @@ endfunc
.macro def_fft n, n2, n4 .macro def_fft n, n2, n4
function fft\n\()_neon, align=6 function fft\n\()_neon, align=6
AARCH64_VALID_JUMP_TARGET AARCH64_SIGN_LINK_REGISTER
sub sp, sp, #16 sub sp, sp, #16
stp x28, x30, [sp] stp x28, x30, [sp]
add x28, x0, #\n4*2*8 add x28, x0, #\n4*2*8
@@ -351,6 +351,7 @@ function fft\n\()_neon, align=6
bl fft\n4\()_neon bl fft\n4\()_neon
sub x0, x28, #\n4*2*8 sub x0, x28, #\n4*2*8
ldp x28, x30, [sp], #16 ldp x28, x30, [sp], #16
AARCH64_VALIDATE_LINK_REGISTER
movrel x4, X(ff_cos_\n) movrel x4, X(ff_cos_\n)
mov x2, #\n4>>1 mov x2, #\n4>>1
b fft_pass_neon b fft_pass_neon

17
libavcodec/aarch64/mdct_neon.S
View File

@@ -25,6 +25,7 @@
function ff_imdct_half_neon, export=1 function ff_imdct_half_neon, export=1
sub sp, sp, #32 sub sp, sp, #32
stp x19, x20, [sp] stp x19, x20, [sp]
AARCH64_SIGN_LINK_REGISTER
str x30, [sp, #16] str x30, [sp, #16]
mov x12, #1 mov x12, #1
ldr w14, [x0, #28] // mdct_bits ldr w14, [x0, #28] // mdct_bits
@@ -121,6 +122,7 @@ function ff_imdct_half_neon, export=1
ldp x19, x20, [sp] ldp x19, x20, [sp]
ldr x30, [sp, #16] ldr x30, [sp, #16]
AARCH64_VALIDATE_LINK_REGISTER
add sp, sp, #32 add sp, sp, #32
ret ret
@@ -129,6 +131,7 @@ endfunc
function ff_imdct_calc_neon, export=1 function ff_imdct_calc_neon, export=1
sub sp, sp, #32 sub sp, sp, #32
stp x19, x20, [sp] stp x19, x20, [sp]
AARCH64_SIGN_LINK_REGISTER
str x30, [sp, #16] str x30, [sp, #16]
ldr w3, [x0, #28] // mdct_bits ldr w3, [x0, #28] // mdct_bits
mov x19, #1 mov x19, #1
@@ -160,8 +163,10 @@ function ff_imdct_calc_neon, export=1
subs x19, x19, #16 subs x19, x19, #16
b.gt 1b b.gt 1b
ldp x19, x20, [sp], #16 ldp x19, x20, [sp]
ldr x30, [sp], #16 ldr x30, [sp, #16]
AARCH64_VALIDATE_LINK_REGISTER
add sp, sp, #32
ret ret
endfunc endfunc
@@ -170,6 +175,7 @@ endfunc
function ff_mdct_calc_neon, export=1 function ff_mdct_calc_neon, export=1
sub sp, sp, #32 sub sp, sp, #32
stp x19, x20, [sp] stp x19, x20, [sp]
AARCH64_SIGN_LINK_REGISTER
str x30, [sp, #16] str x30, [sp, #16]
mov x12, #1 mov x12, #1
@@ -317,7 +323,10 @@ function ff_mdct_calc_neon, export=1
st2 {v4.2s,v5.2s}, [x0] st2 {v4.2s,v5.2s}, [x0]
st2 {v6.2s,v7.2s}, [x8] st2 {v6.2s,v7.2s}, [x8]
ldp x19, x20, [sp], #16 ldp x19, x20, [sp]
ldr x30, [sp], #16 ldr x30, [sp, #16]
AARCH64_VALIDATE_LINK_REGISTER
add sp, sp, #32
ret ret
endfunc endfunc

2
libavcodec/aarch64/synth_filter_neon.S
View File

@@ -52,6 +52,7 @@ function ff_synth_filter_float_neon, export=1
stp x5, x1, [sp, #16] stp x5, x1, [sp, #16]
and x7, x7, #~63 and x7, x7, #~63
and w8, w8, #511 and w8, w8, #511
AARCH64_SIGN_LINK_REGISTER
stp x7, x30, [sp, #32] stp x7, x30, [sp, #32]
str w8, [x2] str w8, [x2]
str s0, [sp, #48] str s0, [sp, #48]
@@ -63,6 +64,7 @@ function ff_synth_filter_float_neon, export=1
ldp x2, x4, [sp] // synct_buf_2, window ldp x2, x4, [sp] // synct_buf_2, window
ldp x13, x9, [sp, #16] // out, synth_buf ldp x13, x9, [sp, #16] // out, synth_buf
ldp x0, x30, [sp, #32] // *synth_buf_offset ldp x0, x30, [sp, #32] // *synth_buf_offset
AARCH64_VALIDATE_LINK_REGISTER
ldr s0, [sp, #48] ldr s0, [sp, #48]
add x3, x2, #16*4 // synct_buf_2 + 16 add x3, x2, #16*4 // synct_buf_2 + 16

102
libavutil/aarch64/asm.S
View File

@@ -36,10 +36,79 @@
# define __has_feature(x) 0 # define __has_feature(x) 0
#endif #endif
/* Support macros for the Armv8.5-A Branch Target Identification feature which
* requires emitting a .note.gnu.property section with the appropriate /* Support macros for
* architecture-dependent feature bits set. * - Armv8.3-A Pointer Authentication and
* Read more: "ELF for the Arm® 64-bit Architecture" * - Armv8.5-A Branch Target Identification
* features which require emitting a .note.gnu.property section with the
* appropriate architecture-dependent feature bits set.
*
* |AARCH64_SIGN_LINK_REGISTER| and |AARCH64_VALIDATE_LINK_REGISTER| expand to
* PACIxSP and AUTIxSP, respectively. |AARCH64_SIGN_LINK_REGISTER| should be
* used immediately before saving the LR register (x30) to the stack.
* |AARCH64_VALIDATE_LINK_REGISTER| should be used immediately after restoring
* it. Note |AARCH64_SIGN_LINK_REGISTER|'s modifications to LR must be undone
* with |AARCH64_VALIDATE_LINK_REGISTER| before RET. The SP register must also
* have the same value at the two points. For example:
*
* .global f
* f:
* AARCH64_SIGN_LINK_REGISTER
* stp x29, x30, [sp, #-96]!
* mov x29, sp
* ...
* ldp x29, x30, [sp], #96
* AARCH64_VALIDATE_LINK_REGISTER
* ret
*
* |AARCH64_VALID_CALL_TARGET| expands to BTI 'c'. Either it, or
* |AARCH64_SIGN_LINK_REGISTER|, must be used at every point that may be an
* indirect call target. In particular, all symbols exported from a file must
* begin with one of these macros. For example, a leaf function that does not
* save LR can instead use |AARCH64_VALID_CALL_TARGET|:
*
* .globl return_zero
* return_zero:
* AARCH64_VALID_CALL_TARGET
* mov x0, #0
* ret
*
* A non-leaf function which does not immediately save LR may need both macros
* because |AARCH64_SIGN_LINK_REGISTER| appears late. For example, the function
* may jump to an alternate implementation before setting up the stack:
*
* .globl with_early_jump
* with_early_jump:
* AARCH64_VALID_CALL_TARGET
* cmp x0, #128
* b.lt .Lwith_early_jump_128
* AARCH64_SIGN_LINK_REGISTER
* stp x29, x30, [sp, #-96]!
* mov x29, sp
* ...
* ldp x29, x30, [sp], #96
* AARCH64_VALIDATE_LINK_REGISTER
* ret
*
* .Lwith_early_jump_128:
* ...
* ret
*
* These annotations are only required with indirect calls. Private symbols that
* are only the target of direct calls do not require annotations. Also note
* that |AARCH64_VALID_CALL_TARGET| is only valid for indirect calls (BLR), not
* indirect jumps (BR). Indirect jumps in assembly are supported through
* |AARCH64_VALID_JUMP_TARGET|. Landing Pads which shall serve for jumps and
* calls can be created using |AARCH64_VALID_JUMP_CALL_TARGET|.
*
* Although not necessary, it is safe to use these macros in 32-bit ARM
* assembly. This may be used to simplify dual 32-bit and 64-bit files.
*
* References:
* - "ELF for the Arm® 64-bit Architecture"
* https: *github.com/ARM-software/abi-aa/blob/master/aaelf64/aaelf64.rst
* - "Providing protection for complex software"
* https://developer.arm.com/architectures/learn-the-architecture/providing-protection-for-complex-software
*/ */
#if defined(__ARM_FEATURE_BTI_DEFAULT) && (__ARM_FEATURE_BTI_DEFAULT == 1) #if defined(__ARM_FEATURE_BTI_DEFAULT) && (__ARM_FEATURE_BTI_DEFAULT == 1)
# define GNU_PROPERTY_AARCH64_BTI (1 << 0) // Has BTI # define GNU_PROPERTY_AARCH64_BTI (1 << 0) // Has BTI
@@ -51,7 +120,28 @@
# define AARCH64_VALID_JUMP_TARGET # define AARCH64_VALID_JUMP_TARGET
#endif #endif
#if (GNU_PROPERTY_AARCH64_BTI != 0) #if defined(__ARM_FEATURE_PAC_DEFAULT)
# if ((__ARM_FEATURE_PAC_DEFAULT & (1 << 0)) != 0) // authentication using key A
# define AARCH64_SIGN_LINK_REGISTER paciasp
# define AARCH64_VALIDATE_LINK_REGISTER autiasp
# elif ((__ARM_FEATURE_PAC_DEFAULT & (1 << 1)) != 0) // authentication using key B
# define AARCH64_SIGN_LINK_REGISTER pacibsp
# define AARCH64_VALIDATE_LINK_REGISTER autibsp
# else
# error Pointer authentication defines no valid key!
# endif
# if ((__ARM_FEATURE_PAC_DEFAULT & (1 << 2)) != 0)
# error Authentication of leaf functions is enabled but not supported in FFmpeg!
# endif
# define GNU_PROPERTY_AARCH64_PAC (1 << 1)
#else
# define GNU_PROPERTY_AARCH64_PAC 0
# define AARCH64_SIGN_LINK_REGISTER
# define AARCH64_VALIDATE_LINK_REGISTER
#endif
#if (GNU_PROPERTY_AARCH64_BTI != 0 || GNU_PROPERTY_AARCH64_PAC != 0)
.pushsection .note.gnu.property, "a" .pushsection .note.gnu.property, "a"
.balign 8 .balign 8
.long 4 .long 4
@@ -60,7 +150,7 @@
.asciz "GNU" .asciz "GNU"
.long 0xc0000000 /* GNU_PROPERTY_AARCH64_FEATURE_1_AND */ .long 0xc0000000 /* GNU_PROPERTY_AARCH64_FEATURE_1_AND */
.long 4 .long 4
.long GNU_PROPERTY_AARCH64_BTI .long (GNU_PROPERTY_AARCH64_BTI | GNU_PROPERTY_AARCH64_PAC)
.long 0 .long 0
.popsection .popsection
#endif #endif