x86/crypto: aesni: Fix build with LLVM_IAS=1
Subsystems (2)
PATCH v2 1/1 x86/crypto: aesni: Fix build with LLVM_IAS=1
When building with LLVM_IAS=1 means using Clang's Integrated Assembly (IAS) from LLVM/Clang >= v10.0.1-rc1+ instead of GNU/as from GNU/binutils I see the following breakage in Debian/testing AMD64:
<instantiation>:15:74: error: too many positional arguments
PRECOMPUTE 8*3+8(%rsp), %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
^
arch/x86/crypto/aesni-intel_asm.S:1598:2: note: while in macro instantiation
GCM_INIT %r9, 8*3 +8(%rsp), 8*3 +16(%rsp), 8*3 +24(%rsp)
^
<instantiation>:47:2: error: unknown use of instruction mnemonic without a size suffix
GHASH_4_ENCRYPT_4_PARALLEL_dec %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14, %xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, %xmm8, enc ^arch/x86/crypto/aesni-intel_asm.S:1599:2: note: while in macro instantiation
GCM_ENC_DEC dec
^
<instantiation>:15:74: error: too many positional arguments
PRECOMPUTE 8*3+8(%rsp), %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
^
arch/x86/crypto/aesni-intel_asm.S:1686:2: note: while in macro instantiation
GCM_INIT %r9, 8*3 +8(%rsp), 8*3 +16(%rsp), 8*3 +24(%rsp)
^
<instantiation>:47:2: error: unknown use of instruction mnemonic without a size suffix
GHASH_4_ENCRYPT_4_PARALLEL_enc %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14, %xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, %xmm8, enc ^arch/x86/crypto/aesni-intel_asm.S:1687:2: note: while in macro instantiation
GCM_ENC_DEC enc
Craig Topper suggested me in ClangBuiltLinux issue #1050:
Quoted 7 lines
I think the "too many positional arguments" is because the parser isn't able to handle the trailing commas.
The "unknown use of instruction mnemonic" is because the macro was named
GHASH_4_ENCRYPT_4_PARALLEL_DECbut its being instantiated withGHASH_4_ENCRYPT_4_PARALLEL_decI guess gas ignores case on the macro instantiation, but llvm doesn't.
First, I removed the trailing comma in the PRECOMPUTE line.
Second, I substituted:
1. GHASH_4_ENCRYPT_4_PARALLEL_DEC -> GHASH_4_ENCRYPT_4_PARALLEL_dec
2. GHASH_4_ENCRYPT_4_PARALLEL_ENC -> GHASH_4_ENCRYPT_4_PARALLEL_enc
With these changes I was able to build with LLVM_IAS=1 and boot on bare metal.
I confirmed that this works with Linux-kernel v5.7.5 final.
NOTE: This patch is on top of Linux v5.7 final.
Thanks to Craig and especially Nick for double-checking and his comments.
--- arch/x86/crypto/aesni-intel_asm.S | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
arch/x86/crypto/aesni-intel_asm.S +3 -3
@@ -266,7 +266,7 @@ ALL_F: .octa 0xffffffffffffffffffffffffffffffff PSHUFB_XMM %xmm2, %xmm0 movdqu %xmm0, CurCount(%arg2) # ctx_data.current_counter = iv - PRECOMPUTE \SUBKEY, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, + PRECOMPUTE \SUBKEY, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7 movdqu HashKey(%arg2), %xmm13 CALC_AAD_HASH %xmm13, \AAD, \AADLEN, %xmm0, %xmm1, %xmm2, %xmm3, \
@@ -978,7 +978,7 @@ _initial_blocks_done\@: * arg1, %arg3, %arg4 are used as pointers only, not modified * %r11 is the data offset value */ -.macro GHASH_4_ENCRYPT_4_PARALLEL_ENC TMP1 TMP2 TMP3 TMP4 TMP5 \ +.macro GHASH_4_ENCRYPT_4_PARALLEL_enc TMP1 TMP2 TMP3 TMP4 TMP5 \ TMP6 XMM0 XMM1 XMM2 XMM3 XMM4 XMM5 XMM6 XMM7 XMM8 operation movdqa \XMM1, \XMM5
@@ -1186,7 +1186,7 @@ aes_loop_par_enc_done\@: * arg1, %arg3, %arg4 are used as pointers only, not modified * %r11 is the data offset value */ -.macro GHASH_4_ENCRYPT_4_PARALLEL_DEC TMP1 TMP2 TMP3 TMP4 TMP5 \ +.macro GHASH_4_ENCRYPT_4_PARALLEL_dec TMP1 TMP2 TMP3 TMP4 TMP5 \ TMP6 XMM0 XMM1 XMM2 XMM3 XMM4 XMM5 XMM6 XMM7 XMM8 operation movdqa \XMM1, \XMM5
Signature · 1 line
2.27.0
Quoted 45 lines · Sedat Dilek
When building with LLVM_IAS=1 means using Clang's Integrated Assembly (IAS) from LLVM/Clang >= v10.0.1-rc1+ instead of GNU/as from GNU/binutils I see the following breakage in Debian/testing AMD64:
<instantiation>:15:74: error: too many positional arguments
PRECOMPUTE 8*3+8(%rsp), %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,^
arch/x86/crypto/aesni-intel_asm.S:1598:2: note: while in macro instantiation
GCM_INIT%r9, 8*3 +8(%rsp), 8*3 +16(%rsp), 8*3 +24(%rsp)
^
<instantiation>:47:2: error: unknown use of instruction mnemonic without a size suffix
GHASH_4_ENCRYPT_4_PARALLEL_dec%xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14, %xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, %xmm8, enc ^arch/x86/crypto/aesni-intel_asm.S:1599:2: note: while in macro instantiation
GCM_ENC_DECdec
^
<instantiation>:15:74: error: too many positional arguments
PRECOMPUTE 8*3+8(%rsp), %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,^
arch/x86/crypto/aesni-intel_asm.S:1686:2: note: while in macro instantiation
GCM_INIT%r9, 8*3 +8(%rsp), 8*3 +16(%rsp), 8*3 +24(%rsp)
^
<instantiation>:47:2: error: unknown use of instruction mnemonic without a size suffix
GHASH_4_ENCRYPT_4_PARALLEL_enc%xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14, %xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, %xmm8, enc ^arch/x86/crypto/aesni-intel_asm.S:1687:2: note: while in macro instantiation
GCM_ENC_DECencCraig Topper suggested me in ClangBuiltLinux issue #1050:
Quoted 7 lines
I think the "too many positional arguments" is because the parser isn't able to handle the trailing commas.
The "unknown use of instruction mnemonic" is because the macro was named
GHASH_4_ENCRYPT_4_PARALLEL_DECbut its being instantiated withGHASH_4_ENCRYPT_4_PARALLEL_decI guess gas ignores case on the macro instantiation, but llvm doesn't.First, I removed the trailing comma in the PRECOMPUTE line.
Second, I substituted:
1.GHASH_4_ENCRYPT_4_PARALLEL_DEC->GHASH_4_ENCRYPT_4_PARALLEL_dec
2.GHASH_4_ENCRYPT_4_PARALLEL_ENC->GHASH_4_ENCRYPT_4_PARALLEL_enc
Quoted 5 lines
With these changes I was able to build with LLVM_IAS=1 and boot on bare metal.
I confirmed that this works with Linux-kernel v5.7.5 final.
NOTE: This patch is on top of Linux v5.7 final.
Thanks for the note, still applies cleanly on top of linux-next today for me.
Quoted 6 lines
Thanks to Craig and especially Nick for double-checking and his comments.
Suggested-by Craig Topper <[email protected]>Suggested-by Craig Topper <[email protected]>Suggested-by Nick Desaulniers [email protected]
^ oh, may have missed <> around email addr.
Quoted 3 lines
Cc "ClangBuiltLinux" <[email protected]>Signed-off-by Sedat Dilek <[email protected]>
Following the same testing methodology from V1
(https://lore.kernel.org/patchwork/comment/1456822/) I verified for GCC+GAS this is no functional change.
$ wget https://lore.kernel.org/patchwork/patch/1261340/mbox/ -O sedat_v3.patch $ git am sedat_v3.patch $ make -j71 arch/x86/crypto/aesni-intel_asm.o $ llvm-objdump -dr arch/x86/crypto/aesni-intel_asm.o > postpatch_v3.txt $ diff -u <(cat prepatch.txt | tr -s ' ' | cut -d ' ' -f 2-) <(cat postpatch_v3.txt| tr -s ' ' | cut -d ' ' -f 2-) | less (no output)
Quoted 36 lines
--- arch/x86/crypto/aesni-intel_asm.S | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)arch/x86/crypto/aesni-intel_asm.S +3
-4@@ -266,7 +266,7 @@ ALL_F: .octa 0xffffffffffffffffffffffffffffffff PSHUFB_XMM %xmm2, %xmm0 movdqu %xmm0, CurCount(%arg2) # ctx_data.current_counter = iv - PRECOMPUTE \SUBKEY, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, + PRECOMPUTE \SUBKEY, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7 movdqu HashKey(%arg2), %xmm13 CALC_AAD_HASH %xmm13, \AAD, \AADLEN, %xmm0, %xmm1, %xmm2, %xmm3, \@@ -978,7 +978,7 @@ _initial_blocks_done\@: * arg1, %arg3, %arg4 are used as pointers only, not modified * %r11 is the data offset value */ -.macro GHASH_4_ENCRYPT_4_PARALLEL_ENC TMP1 TMP2 TMP3 TMP4 TMP5 \ +.macro GHASH_4_ENCRYPT_4_PARALLEL_enc TMP1 TMP2 TMP3 TMP4 TMP5 \ TMP6 XMM0 XMM1 XMM2 XMM3 XMM4 XMM5 XMM6 XMM7 XMM8 operation movdqa \XMM1, \XMM5@@ -1186,7 +1186,7 @@ aes_loop_par_enc_done\@: * arg1, %arg3, %arg4 are used as pointers only, not modified * %r11 is the data offset value */ -.macro GHASH_4_ENCRYPT_4_PARALLEL_DEC TMP1 TMP2 TMP3 TMP4 TMP5 \ +.macro GHASH_4_ENCRYPT_4_PARALLEL_dec TMP1 TMP2 TMP3 TMP4 TMP5 \ TMP6 XMM0 XMM1 XMM2 XMM3 XMM4 XMM5 XMM6 XMM7 XMM8 operation movdqa \XMM1, \XMM5 --
Signature · 2 lines
Thanks, ~Nick Desaulniers
Quoted 80 lines · Nick Desaulniers
Quoted 45 lines · Sedat Dilek
When building with LLVM_IAS=1 means using Clang's Integrated Assembly (IAS) from LLVM/Clang >= v10.0.1-rc1+ instead of GNU/as from GNU/binutils I see the following breakage in Debian/testing AMD64:
<instantiation>:15:74: error: too many positional arguments
PRECOMPUTE 8*3+8(%rsp), %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,^
arch/x86/crypto/aesni-intel_asm.S:1598:2: note: while in macro instantiation
GCM_INIT%r9, 8*3 +8(%rsp), 8*3 +16(%rsp), 8*3 +24(%rsp)
^
<instantiation>:47:2: error: unknown use of instruction mnemonic without a size suffix
GHASH_4_ENCRYPT_4_PARALLEL_dec%xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14, %xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, %xmm8, enc ^arch/x86/crypto/aesni-intel_asm.S:1599:2: note: while in macro instantiation
GCM_ENC_DECdec
^
<instantiation>:15:74: error: too many positional arguments
PRECOMPUTE 8*3+8(%rsp), %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,^
arch/x86/crypto/aesni-intel_asm.S:1686:2: note: while in macro instantiation
GCM_INIT%r9, 8*3 +8(%rsp), 8*3 +16(%rsp), 8*3 +24(%rsp)
^
<instantiation>:47:2: error: unknown use of instruction mnemonic without a size suffix
GHASH_4_ENCRYPT_4_PARALLEL_enc%xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14, %xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, %xmm8, enc ^arch/x86/crypto/aesni-intel_asm.S:1687:2: note: while in macro instantiation
GCM_ENC_DECencCraig Topper suggested me in ClangBuiltLinux issue #1050:
Quoted 7 lines
I think the "too many positional arguments" is because the parser isn't able to handle the trailing commas.
The "unknown use of instruction mnemonic" is because the macro was named
GHASH_4_ENCRYPT_4_PARALLEL_DECbut its being instantiated withGHASH_4_ENCRYPT_4_PARALLEL_decI guess gas ignores case on the macro instantiation, but llvm doesn't.First, I removed the trailing comma in the PRECOMPUTE line.
Second, I substituted:
1.GHASH_4_ENCRYPT_4_PARALLEL_DEC->GHASH_4_ENCRYPT_4_PARALLEL_dec
2.GHASH_4_ENCRYPT_4_PARALLEL_ENC->GHASH_4_ENCRYPT_4_PARALLEL_encQuoted 5 lines
With these changes I was able to build with LLVM_IAS=1 and boot on bare metal.
I confirmed that this works with Linux-kernel v5.7.5 final.
NOTE: This patch is on top of Linux v5.7 final.
Thanks for the note, still applies cleanly on top of linux-next today for me.
Quoted 6 lines
Thanks to Craig and especially Nick for double-checking and his comments.
Suggested-by Craig Topper <[email protected]>Suggested-by Craig Topper <[email protected]>Suggested-by Nick Desaulniers [email protected]^ oh, may have missed <> around email addr.
Quoted 3 lines
Cc "ClangBuiltLinux" <[email protected]>Signed-off-by Sedat Dilek <[email protected]>Following the same testing methodology from V1
(https://lore.kernel.org/patchwork/comment/1456822/) I verified for GCC+GAS this is no functional change.$ wget https://lore.kernel.org/patchwork/patch/1261340/mbox/ -O sedat_v3.patch $ git am sedat_v3.patch $ make -j71 arch/x86/crypto/aesni-intel_asm.o $ llvm-objdump -dr arch/x86/crypto/aesni-intel_asm.o > postpatch_v3.txt $ diff -u <(cat prepatch.txt | tr -s ' ' | cut -d ' ' -f 2-) <(cat postpatch_v3.txt| tr -s ' ' | cut -d ' ' -f 2-) | less (no output)
Quoted 2 lines
Reviewed-by Nick Desaulniers <[email protected]>
Thanks a lot for your feedback.
I have sent a v4 with corrected email-address and your Reviewed-by.
- Sedat -
[1] https://lore.kernel.org/patchwork/patch/1263102/
[2] https://lore.kernel.org/lkml/[email protected]/
Quoted 40 lines
Quoted 27 lines
--- arch/x86/crypto/aesni-intel_asm.S | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)arch/x86/crypto/aesni-intel_asm.S +2
-2@@ -266,7 +266,7 @@ ALL_F: .octa 0xffffffffffffffffffffffffffffffff PSHUFB_XMM %xmm2, %xmm0 movdqu %xmm0, CurCount(%arg2) # ctx_data.current_counter = iv - PRECOMPUTE \SUBKEY, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7, + PRECOMPUTE \SUBKEY, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7 movdqu HashKey(%arg2), %xmm13 CALC_AAD_HASH %xmm13, \AAD, \AADLEN, %xmm0, %xmm1, %xmm2, %xmm3, \@@ -978,7 +978,7 @@ _initial_blocks_done\@: * arg1, %arg3, %arg4 are used as pointers only, not modified * %r11 is the data offset value */ -.macro GHASH_4_ENCRYPT_4_PARALLEL_ENC TMP1 TMP2 TMP3 TMP4 TMP5 \ +.macro GHASH_4_ENCRYPT_4_PARALLEL_enc TMP1 TMP2 TMP3 TMP4 TMP5 \ TMP6 XMM0 XMM1 XMM2 XMM3 XMM4 XMM5 XMM6 XMM7 XMM8 operation movdqa \XMM1, \XMM5arch/x86/crypto/aesni-intel_asm.S aes_loop_par_enc_done\@:
Quoted 9 lines
arch/x86/crypto/aesni-intel_asm.S +1
-2@@ -1186,7 +1186,7 @@ aes_loop_par_enc_done\@: * arg1, %arg3, %arg4 are used as pointers only, not modified * %r11 is the data offset value */ -.macro GHASH_4_ENCRYPT_4_PARALLEL_DEC TMP1 TMP2 TMP3 TMP4 TMP5 \ +.macro GHASH_4_ENCRYPT_4_PARALLEL_dec TMP1 TMP2 TMP3 TMP4 TMP5 \ TMP6 XMM0 XMM1 XMM2 XMM3 XMM4 XMM5 XMM6 XMM7 XMM8 operation movdqa \XMM1, \XMM5 ----
Thanks,
~Nick Desaulniers