Skip to content

[compiler-rt][AArch64] Allow platform-specific mangling of SME routines. #119864

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Dec 16, 2024
Merged

[compiler-rt][AArch64] Allow platform-specific mangling of SME routines. #119864

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 14 additions & 12 deletions compiler-rt/lib/builtins/aarch64/sme-abi.S
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,7 +40,7 @@ DEFINE_COMPILERRT_PRIVATE_FUNCTION(do_abort)
.cfi_offset w30, -24
.cfi_offset w29, -32
.cfi_offset 46, -16
bl __arm_sme_state
bl SYMBOL_NAME(__arm_sme_state)
tbz x0, #0, 2f
1:
smstop sm
Expand All @@ -54,7 +54,7 @@ END_COMPILERRT_FUNCTION(do_abort)
// __arm_sme_state fills the result registers based on a local
// that is set as part of the compiler-rt startup code.
// __aarch64_has_sme_and_tpidr2_el0
DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_sme_state)
DEFINE_COMPILERRT_FUNCTION(__arm_sme_state)
.variant_pcs __arm_sme_state
BTI_C
mov x0, xzr
Expand All @@ -70,9 +70,9 @@ DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_sme_state)
mrs x1, TPIDR2_EL0
1:
ret
END_COMPILERRT_OUTLINE_FUNCTION(__arm_sme_state)
END_COMPILERRT_FUNCTION(__arm_sme_state)

DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_tpidr2_restore)
DEFINE_COMPILERRT_FUNCTION(__arm_tpidr2_restore)
.variant_pcs __arm_tpidr2_restore
BTI_C
// If TPIDR2_EL0 is nonnull, the subroutine aborts in some platform-specific
Expand Down Expand Up @@ -106,9 +106,9 @@ DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_tpidr2_restore)
ret
2:
b SYMBOL_NAME(do_abort)
END_COMPILERRT_OUTLINE_FUNCTION(__arm_tpidr2_restore)
END_COMPILERRT_FUNCTION(__arm_tpidr2_restore)

DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_tpidr2_save)
DEFINE_COMPILERRT_FUNCTION(__arm_tpidr2_save)
.variant_pcs __arm_tpidr2_save
BTI_C
// If the current thread does not have access to TPIDR2_EL0, the subroutine
Expand Down Expand Up @@ -147,9 +147,10 @@ DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_tpidr2_save)
ret
2:
b SYMBOL_NAME(do_abort)
END_COMPILERRT_OUTLINE_FUNCTION(__arm_tpidr2_save)
END_COMPILERRT_FUNCTION(__arm_tpidr2_save)

DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_za_disable)
DEFINE_COMPILERRT_FUNCTION(__arm_za_disable)
.cfi_startproc
.variant_pcs __arm_za_disable
BTI_C
// If the current thread does not have access to SME, the subroutine does
Expand All @@ -166,7 +167,7 @@ DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_za_disable)
.cfi_def_cfa w29, 16
.cfi_offset w30, -8
.cfi_offset w29, -16
bl __arm_tpidr2_save
bl SYMBOL_NAME(__arm_tpidr2_save)

// * Set TPIDR2_EL0 to null.
msr TPIDR2_EL0, xzr
Expand All @@ -181,9 +182,10 @@ DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_za_disable)
.cfi_restore w29
0:
ret
END_COMPILERRT_OUTLINE_FUNCTION(__arm_za_disable)
.cfi_endproc
END_COMPILERRT_FUNCTION(__arm_za_disable)

DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_get_current_vg)
DEFINE_COMPILERRT_FUNCTION(__arm_get_current_vg)
.variant_pcs __arm_get_current_vg
BTI_C

Expand All @@ -200,7 +202,7 @@ DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_get_current_vg)
2:
mov x0, xzr
ret
END_COMPILERRT_OUTLINE_FUNCTION(__arm_get_current_vg)
END_COMPILERRT_FUNCTION(__arm_get_current_vg)

NO_EXEC_STACK_DIRECTIVE

Expand Down
8 changes: 4 additions & 4 deletions compiler-rt/lib/builtins/aarch64/sme-libc-mem-routines.S
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,7 +52,7 @@
The loop tail is handled by always copying 64 bytes from the end.
*/

DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_sc_memcpy)
DEFINE_COMPILERRT_FUNCTION(__arm_sc_memcpy)
add srcend1, src, count
add dstend1, dstin, count
cmp count, 128
Expand Down Expand Up @@ -232,7 +232,7 @@ L(copy64_from_start):
stp B_l, B_h, [dstin, 16]
stp C_l, C_h, [dstin]
ret
END_COMPILERRT_OUTLINE_FUNCTION(__arm_sc_memcpy)
END_COMPILERRT_FUNCTION(__arm_sc_memcpy)

DEFINE_COMPILERRT_FUNCTION_ALIAS(__arm_sc_memmove, __arm_sc_memcpy)

Expand All @@ -250,7 +250,7 @@ DEFINE_COMPILERRT_FUNCTION_ALIAS(__arm_sc_memmove, __arm_sc_memcpy)
#define dstend2 x4
#define zva_val x5

DEFINE_COMPILERRT_OUTLINE_FUNCTION_UNMANGLED(__arm_sc_memset)
DEFINE_COMPILERRT_FUNCTION(__arm_sc_memset)
#ifdef __ARM_FEATURE_SVE
mov z0.b, valw
#else
Expand Down Expand Up @@ -346,6 +346,6 @@ L(no_zva_loop):
stp q0, q0, [dstend2, -64]
stp q0, q0, [dstend2, -32]
ret
END_COMPILERRT_OUTLINE_FUNCTION(__arm_sc_memset)
END_COMPILERRT_FUNCTION(__arm_sc_memset)

#endif // __aarch64__
Loading