too many positional arguments in arm64 kvm #939

nickdesaulniers · 2020-03-19T23:37:19Z

arch/arm64/kvm/hyp/entry.S:112:87: error: too many positional arguments
 alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1
                                                                                      ^

The text was updated successfully, but these errors were encountered:

ihalip · 2020-03-20T16:00:11Z

LLVM ASM parser has trouble parsing that expression. I don't like it, but this workaround does the trick:

diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h
index b91570ff9db1..6dda7aa650e1 100644
--- a/arch/arm64/include/asm/sysreg.h
+++ b/arch/arm64/include/asm/sysreg.h
@@ -49,9 +49,9 @@
 #ifndef CONFIG_BROKEN_GAS_INST
 
 #ifdef __ASSEMBLY__
-#define __emit_inst(x)                 .inst (x)
+#define __emit_inst(x)                 .inst(x)
 #else
-#define __emit_inst(x)                 ".inst " __stringify((x)) "\n\t"
+#define __emit_inst(x)                 ".inst" __stringify((x)) "\n\t"
 #endif
 
 #else  /* CONFIG_BROKEN_GAS_INST */

jcai19 · 2020-03-26T21:34:15Z

Reduced reproducer:

$ cat bad.s
.macro alternative_insn insn1
.endm

alternative_insn .inst (0x0)

$ llvm-mc -triple=aarch64 bad.s -o /dev/null
bad.s:4:25: error: too many positional arguments
alternative_insn .inst (0x0)

$ aarch64-linux-gnu-gcc -c bad.s
$ echo $?
0

jcai19 · 2020-03-26T23:49:09Z

Took a look at LLVM's code, and it turned out the parsing of macro argument (AsmParser::parseMacroArgument) stopped after it read any space after .inst directive. Interestingly, on x86, gcc also does not support such syntax.

$ cat bad.s
.macro alternative_insn insn
\insn
.endm

alternative_insn .space 8

$ gcc -c bad.s
bad.s: Assembler messages:
bad.s:5: Error: too many positional arguments

jcai19 · 2020-03-27T00:54:32Z

Having the integrated assembler to fully support directive calls as macro arguments seems to be a non-trivial task as we would have to parse all the directives with (sometimes optional) arguments. Unless such cases prevail in Linux code base, the proposed workaround may be a better choice instead of having such support in LLVM IMO.

jcai19 · 2020-03-28T00:43:43Z

On second thoughts, the fix in the integrated assembler may be simpler than I thought. Send a patch for review at https://reviews.llvm.org/D76962.

MaskRay · 2020-03-28T00:59:42Z

It is worth mentioning that you may need to add psmith instead of peter.smith due to https://reviews.llvm.org/D73542#1871004 .

jcai19 · 2020-03-28T01:02:24Z

Thanks! Will use the new username.

MaskRay · 2020-03-28T07:32:29Z

I investigated GNU as' behavior. I think we need to work around it in the kernel https://reviews.llvm.org/D76962#1947556 It is not clear to me that this (parsing .inst (x) as a whole actual parameter) is intentionally supported. The behavior may not be reasonably implemented in LLVM MC.

LLVM ASM parser has trouble parsing that expression. I don't like it, but this workaround does the trick:

diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h
index b91570ff9db1..6dda7aa650e1 100644
--- a/arch/arm64/include/asm/sysreg.h
+++ b/arch/arm64/include/asm/sysreg.h
@@ -49,9 +49,9 @@
 #ifndef CONFIG_BROKEN_GAS_INST
 
 #ifdef __ASSEMBLY__
-#define __emit_inst(x)                 .inst (x)
+#define __emit_inst(x)                 .inst(x)
 #else
-#define __emit_inst(x)                 ".inst " __stringify((x)) "\n\t"
+#define __emit_inst(x)                 ".inst" __stringify((x)) "\n\t"
 #endif
 
 #else  /* CONFIG_BROKEN_GAS_INST */

--- i/arch/arm64/include/asm/sysreg.h
+++ w/arch/arm64/include/asm/sysreg.h
@@ -51,3 +51,3 @@
 #ifdef __ASSEMBLY__
-#define __emit_inst(x)                 .inst (x)
+#define __emit_inst(x)                 .inst(x)
 #else

may be sufficient.

nickdesaulniers · 2020-03-30T18:09:27Z

Ok, with the additional information:

https://reviews.llvm.org/D76962#1948922
too many positional arguments in arm64 kvm #939 (comment)
and additional vote for @ihalip 's sugguested change, I'm more supportive of fixing this on the Linux side. Thanks @jcai19 and @MaskRay for digging into this more; sometimes the issues are quite nuanced, but I think @jcai19 's comment (link 1) is enough for me to think that we shouldn't imitate inconsistent behavior, as it's probably unintentional. It might be worthwhile @jcai19 to file a bug against binutils with some examples demonstrating the inconsistency, and ask for clarity. We might be able to come to a mutual solution with them.

jcai19 · 2020-03-30T20:28:05Z

Thanks for the clarification. I have reported a bug against gas at https://sourceware.org/bugzilla/show_bug.cgi?id=25750, but don't know who I should cc so please feel free to add the right people.

nickdesaulniers · 2020-04-01T18:12:35Z

Thanks for reporting. Nick Clifton suggests a :vararg suffix.

does that solve the issue here when using -no-integrated-as?
does IA recognize :vararg suffix?

MaskRay · 2020-04-01T19:06:17Z

:vararg does not work for more than 1 argument. As I mentioned in the review, properly quoting the argument (yeah, "working around" it in the kernel) with " is preferred.

The existing use case just makes use of a subtle rule.

jcai19 · 2020-04-01T21:27:55Z

does that solve the issue here when using -no-integrated-as?

Like @MaskRay mentioned, I could not figure how to make :vararg work when there are multiple directives with arguments with gnu as.

does IA recognize :vararg suffix?

Yes it does.

$ cat good.s
.macro foo insn1:vararg
\insn1
.endm

foo .section .foo

$llvm-mc -triple=i386 -filetype=obj good.s -o good.o
$llvm-readelf -S good.o

There are 5 section headers, starting at offset 0x60:

Section Headers:
[Nr] Name Type Address Off Size ES Flg Lk Inf Al
[ 0] NULL 00000000 000000 000000 00 0 0 0
[ 1] .strtab STRTAB 00000000 000044 00001c 00 0 0 1
[ 2] .text PROGBITS 00000000 000034 000000 00 AX 0 0 4
[ 3] .foo PROGBITS 00000000 000034 000000 00 0 0 1
[ 4] .symtab SYMTAB 00000000 000034 000010 10 1 1 4
Key to Flags:
W (write), A (alloc), X (execute), M (merge), S (strings), I (info),
L (link order), O (extra OS processing required), G (group), T (TLS),
C (compressed), x (unknown), o (OS specific), E (exclude),
p (processor specific)

Many instances of __emit_inst(x) expand to a directive. In a few places it is used as a macro argument, e.g. arch/arm64/include/asm/sysreg.h #define __emit_inst(x) .inst (x) arch/arm64/include/asm/sysreg.h #define SET_PSTATE_PAN(x) __emit_inst(0xd500401f | PSTATE_PAN | ((!!x) << PSTATE_Imm_shift)) arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) Clang integrated assembler parses `.inst (x)` as two arguments passing to a macro. We delete the space separator so that `.inst(x)` will be parsed as one argument. Note, GNU as parsing `.inst (x)` as one argument is unintentional (for example the x86 backend will parse the construct as two arguments). See https://sourceware.org/bugzilla/show_bug.cgi?id=25750#c10 Link: ClangBuiltLinux#939 Cc: clang-built-linux@googlegroups.com Signed-off-by: Fangrui Song <maskray@google.com>

MaskRay · 2020-04-13T04:22:00Z

I have studied the integrated assembler behavior and GNU as'. Requested a documentation improvement for GNU as (https://sourceware.org/bugzilla/show_bug.cgi?id=25750#c11) and sent [PATCH] arm64: Delete the space separator in __emit_inst to linux-arm-kernel

In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux#939

nickdesaulniers · 2020-04-15T16:33:05Z

accepted: https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/commit/?h=for-next/fixes&id=c9a4ef66450145a356a626c833d3d7b1668b3ded

In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>

MaskRay · 2020-04-21T16:42:23Z

Merged into mainline: http://git.kernel.org/linus/c9a4ef66450145a356a626c833d3d7b1668b3ded

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Change-Id: Ia6619829be82fb04b950fa39330a97c36fed5c01 Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: TogoFire <italomellopereira@gmail.com> Signed-off-by: Dmitrii <bankersenator@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: John Vincent <git@tensevntysevn.cf> Signed-off-by: John Vincent <git@tenseventyseven.cf>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: Omar Mohamed <omarcoptan9@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: GhostMaster69-dev <rathore6375@gmail.com> Signed-off-by: Omar Mohamed <omarcoptan9@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: John Vincent <git@tensevntysevn.cf> Signed-off-by: John Vincent <git@tenseventyseven.cf>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Change-Id: Ia6619829be82fb04b950fa39330a97c36fed5c01 Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: TogoFire <italomellopereira@gmail.com> Signed-off-by: Dmitrii <bankersenator@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: dotkit <ewprjkt@proton.me> Signed-off-by: strongreasons <abenkenari@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: John Vincent <git@tensevntysevn.cf> Signed-off-by: John Vincent <git@tenseventyseven.cf>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: RyuujiX <saputradenny712@gmail.com> Signed-off-by: strongreasons <strongreasons@users.noreply.github.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: dotkit <dotkit@electrowizard.me>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: RyuujiX <saputradenny712@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Change-Id: Ia6619829be82fb04b950fa39330a97c36fed5c01 Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: TogoFire <italomellopereira@gmail.com> Signed-off-by: Dmitrii <bankersenator@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: John Vincent <git@tensevntysevn.cf> Signed-off-by: John Vincent <git@tenseventyseven.cf>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Change-Id: Ia6619829be82fb04b950fa39330a97c36fed5c01 Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: TogoFire <italomellopereira@gmail.com> Signed-off-by: Dmitrii <bankersenator@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: RyuujiX <saputradenny712@gmail.com>

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581

[ Upstream commit c9a4ef6 ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581

[ Upstream commit c9a4ef66450145a356a626c833d3d7b1668b3ded ] In assembly, many instances of __emit_inst(x) expand to a directive. In a few places __emit_inst(x) is used as an assembler macro argument. For example, in arch/arm64/kvm/hyp/entry.S ALTERNATIVE(nop, SET_PSTATE_PAN(1), ARM64_HAS_PAN, CONFIG_ARM64_PAN) expands to the following by the C preprocessor: alternative_insn nop, .inst (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 Both comma and space are separators, with an exception that content inside a pair of parentheses/quotes is not split, so the clang integrated assembler splits the arguments to: nop, .inst, (0xd500401f | ((0) << 16 | (4) << 5) | ((!!1) << 8)), 4, 1 GNU as preprocesses the input with do_scrub_chars(). Its arm64 backend (along with many other non-x86 backends) sees: alternative_insn nop,.inst(0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # .inst(...) is parsed as one argument while its x86 backend sees: alternative_insn nop,.inst (0xd500401f|((0)<<16|(4)<<5)|((!!1)<<8)),4,1 # The extra space before '(' makes the whole .inst (...) parsed as two arguments The non-x86 backend's behavior is considered unintentional (https://sourceware.org/bugzilla/show_bug.cgi?id=25750). So drop the space separator inside `.inst (...)` to make the clang integrated assembler work. Suggested-by: Ilie Halip <ilie.halip@gmail.com> Signed-off-by: Fangrui Song <maskray@google.com> Reviewed-by: Mark Rutland <mark.rutland@arm.com> Link: ClangBuiltLinux/linux#939 Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Change-Id: I6644ce237fcfe594fb6147f5d3df9359dc192581 Signed-off-by: RyuujiX <saputradenny712@gmail.com>

nickdesaulniers added [BUG] Untriaged Something isn't working [TOOL] integrated-as The issue is relevant to LLVM integrated assembler labels Mar 19, 2020

tpimh mentioned this issue Mar 21, 2020

investigate cross-compilation with LLVM tools and without GNU binutils ClangBuiltLinux/continuous-integration#73

Open

nickdesaulniers mentioned this issue Mar 24, 2020

LLD + ThinLTO fails with "SHF_MERGE section size must be a multiple of sh_entsize" #510

Closed

nickdesaulniers added the [PATCH] Submitted A patch has been submitted for review label Apr 13, 2020

nickdesaulniers assigned MaskRay Apr 13, 2020

nickdesaulniers added [PATCH] Accepted A submitted patch has been accepted upstream [BUG] linux A bug that should be fixed in the mainline kernel. and removed [BUG] Untriaged Something isn't working [PATCH] Submitted A patch has been submitted for review labels Apr 15, 2020

MaskRay closed this as completed Apr 21, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

too many positional arguments in arm64 kvm #939

too many positional arguments in arm64 kvm #939

nickdesaulniers commented Mar 19, 2020

ihalip commented Mar 20, 2020 •

edited

Loading

jcai19 commented Mar 26, 2020 •

edited

Loading

jcai19 commented Mar 26, 2020 •

edited

Loading

jcai19 commented Mar 27, 2020

jcai19 commented Mar 28, 2020

MaskRay commented Mar 28, 2020 •

edited

Loading

jcai19 commented Mar 28, 2020

MaskRay commented Mar 28, 2020 •

edited

Loading

nickdesaulniers commented Mar 30, 2020

jcai19 commented Mar 30, 2020

nickdesaulniers commented Apr 1, 2020

MaskRay commented Apr 1, 2020 •

edited

Loading

jcai19 commented Apr 1, 2020

MaskRay commented Apr 13, 2020

nickdesaulniers commented Apr 15, 2020

MaskRay commented Apr 21, 2020

too many positional arguments in arm64 kvm #939

too many positional arguments in arm64 kvm #939

Comments

nickdesaulniers commented Mar 19, 2020

ihalip commented Mar 20, 2020 • edited Loading

jcai19 commented Mar 26, 2020 • edited Loading

jcai19 commented Mar 26, 2020 • edited Loading

jcai19 commented Mar 27, 2020

jcai19 commented Mar 28, 2020

MaskRay commented Mar 28, 2020 • edited Loading

jcai19 commented Mar 28, 2020

MaskRay commented Mar 28, 2020 • edited Loading

nickdesaulniers commented Mar 30, 2020

jcai19 commented Mar 30, 2020

nickdesaulniers commented Apr 1, 2020

MaskRay commented Apr 1, 2020 • edited Loading

jcai19 commented Apr 1, 2020

MaskRay commented Apr 13, 2020

nickdesaulniers commented Apr 15, 2020

MaskRay commented Apr 21, 2020

ihalip commented Mar 20, 2020 •

edited

Loading

jcai19 commented Mar 26, 2020 •

edited

Loading

jcai19 commented Mar 26, 2020 •

edited

Loading

MaskRay commented Mar 28, 2020 •

edited

Loading

MaskRay commented Mar 28, 2020 •

edited

Loading

MaskRay commented Apr 1, 2020 •

edited

Loading