Skip to contentSkip to conversation

mainline build failure for x86_64 and arm64 with clang

DISCUSSION Sudip Mukherjee (Codethink) 3 participants 3 messages Sep 26, 2022 Kbuild

lore
Share & tools
lore
#0 lore raw

Hi All,

Yesterday I updated my clang to:
clang version 16.0.0 (https://github.com/llvm/llvm-project.git bcb1397bda667e75200ae0be9a65fd17dd0763d4)

And with that I see new build failures of mainline.

I am not copying all the errors but others look similar to these. From x86_64 almodconfig:

In file included from scripts/mod/devicetable-offsets.c:3:
In file included from ./include/linux/mod_devicetable.h:13:
In file included from ./include/linux/uuid.h:12:
In file included from ./include/linux/string.h:253:
./include/linux/fortify-string.h:159:10: error: ISO C does not allow indirection on operand of type 'void *' [-Werror,-Wvoid-ptr-dereference]

        q_len = strlen(q);
                ^~~~~~~~~

From x86_64 defconfig:

In file included from arch/x86/kernel/asm-offsets.c:9:
In file included from ./include/linux/crypto.h:20:
In file included from ./include/linux/slab.h:15:
In file included from ./include/linux/gfp.h:7:
In file included from ./include/linux/mmzone.h:8:
In file included from ./include/linux/spinlock.h:55:
In file included from ./include/linux/preempt.h:78:
In file included from ./arch/x86/include/asm/preempt.h:7:
In file included from ./include/linux/thread_info.h:60:
In file included from ./arch/x86/include/asm/thread_info.h:53:
In file included from ./arch/x86/include/asm/cpufeature.h:5:
In file included from ./arch/x86/include/asm/processor.h:22:
In file included from ./arch/x86/include/asm/msr.h:11:
In file included from ./arch/x86/include/asm/cpumask.h:5:
In file included from ./include/linux/cpumask.h:12:
In file included from ./include/linux/bitmap.h:9:
./include/linux/find.h:119:31: error: ISO C does not allow indirection on operand of type 'void *' [-Werror,-Wvoid-ptr-dereference]

                unsigned long val = *addr & GENMASK(size - 1, 0);
                                            ^~~~~~~~~~~~~~~~~~~~
./include/linux/bits.h:38:3: note: expanded from macro 'GENMASK'
        (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
         ^~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/bits.h:25:3: note: expanded from macro 'GENMASK_INPUT_CHECK'
                __is_constexpr((l) > (h)), (l) > (h), 0)))
                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/const.h:12:25: note: expanded from macro '__is_constexpr'
        (sizeof(int) == sizeof(*(8 ? ((void *)((long)(x) * 0l)) : (int *)8)))
                               ^

./include/linux/build_bug.h:16:62: note: expanded from macro 'BUILD_BUG_ON_ZERO' #define BUILD_BUG_ON_ZERO(e) ((int)(sizeof(struct { int:(-!!(e)); })))

From arm64 allmodconfig:

In file included from scripts/mod/devicetable-offsets.c:3:
In file included from ./include/linux/mod_devicetable.h:13:
In file included from ./include/linux/uuid.h:12:
In file included from ./include/linux/string.h:253:
./include/linux/fortify-string.h:159:10: error: ISO C does not allow indirection on operand of type 'void *' [-Werror,-Wvoid-ptr-dereference]

        q_len = strlen(q);
                ^~~~~~~~~
./include/linux/fortify-string.h:131:24: note: expanded from macro 'strlen'
        __builtin_choose_expr(__is_constexpr(__builtin_strlen(p)),      \
                              ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/const.h:12:25: note: expanded from macro '__is_constexpr'
        (sizeof(int) == sizeof(*(8 ? ((void *)((long)(x) * 0l)) : (int *)8)))
                               ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I dont know if its some regression introduced in clang, or really a kernel issue.

I will go back to my previous clang version (329b972d416a) for now untill I know more.

Signature · 2 lines
Regards
Sudip
#1 lore raw

Hi Sudip,

Quoted 68 lines · Sudip Mukherjee (Codethink)

Hi All,

Yesterday I updated my clang to:
clang version 16.0.0 (https://github.com/llvm/llvm-project.git bcb1397bda667e75200ae0be9a65fd17dd0763d4)

And with that I see new build failures of mainline.

I am not copying all the errors but others look similar to these. From x86_64 almodconfig:

In file included from scripts/mod/devicetable-offsets.c:3:
In file included from ./include/linux/mod_devicetable.h:13:
In file included from ./include/linux/uuid.h:12:
In file included from ./include/linux/string.h:253:
./include/linux/fortify-string.h:159:10: error: ISO C does not allow indirection on operand of type 'void *' [-Werror,-Wvoid-ptr-dereference]

        q_len = strlen(q);
                ^~~~~~~~~

From x86_64 defconfig:

In file included from arch/x86/kernel/asm-offsets.c:9:
In file included from ./include/linux/crypto.h:20:
In file included from ./include/linux/slab.h:15:
In file included from ./include/linux/gfp.h:7:
In file included from ./include/linux/mmzone.h:8:
In file included from ./include/linux/spinlock.h:55:
In file included from ./include/linux/preempt.h:78:
In file included from ./arch/x86/include/asm/preempt.h:7:
In file included from ./include/linux/thread_info.h:60:
In file included from ./arch/x86/include/asm/thread_info.h:53:
In file included from ./arch/x86/include/asm/cpufeature.h:5:
In file included from ./arch/x86/include/asm/processor.h:22:
In file included from ./arch/x86/include/asm/msr.h:11:
In file included from ./arch/x86/include/asm/cpumask.h:5:
In file included from ./include/linux/cpumask.h:12:
In file included from ./include/linux/bitmap.h:9:
./include/linux/find.h:119:31: error: ISO C does not allow indirection on operand of type 'void *' [-Werror,-Wvoid-ptr-dereference]

                unsigned long val = *addr & GENMASK(size - 1, 0);
                                            ^~~~~~~~~~~~~~~~~~~~
./include/linux/bits.h:38:3: note: expanded from macro 'GENMASK'
        (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
         ^~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/bits.h:25:3: note: expanded from macro 'GENMASK_INPUT_CHECK'
                __is_constexpr((l) > (h)), (l) > (h), 0)))
                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/const.h:12:25: note: expanded from macro '__is_constexpr'
        (sizeof(int) == sizeof(*(8 ? ((void *)((long)(x) * 0l)) : (int *)8)))
                               ^

./include/linux/build_bug.h:16:62: note: expanded from macro 'BUILD_BUG_ON_ZERO' #define BUILD_BUG_ON_ZERO(e) ((int)(sizeof(struct { int:(-!!(e)); })))

From arm64 allmodconfig:

In file included from scripts/mod/devicetable-offsets.c:3:
In file included from ./include/linux/mod_devicetable.h:13:
In file included from ./include/linux/uuid.h:12:
In file included from ./include/linux/string.h:253:
./include/linux/fortify-string.h:159:10: error: ISO C does not allow indirection on operand of type 'void *' [-Werror,-Wvoid-ptr-dereference]

        q_len = strlen(q);
                ^~~~~~~~~
./include/linux/fortify-string.h:131:24: note: expanded from macro 'strlen'
        __builtin_choose_expr(__is_constexpr(__builtin_strlen(p)),      \
                              ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/const.h:12:25: note: expanded from macro '__is_constexpr'
        (sizeof(int) == sizeof(*(8 ? ((void *)((long)(x) * 0l)) : (int *)8)))
                               ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Quoted 3 lines

I dont know if its some regression introduced in clang, or really a kernel issue.

I will go back to my previous clang version (329b972d416a) for now untill I know more.

Thanks for testing and the report! I noticed this over the weekend, it comes from a new warning in clang:

https://github.com/llvm/llvm-project/commit/e07ead85a368173a56e96a21d6841aa497ad80f8

It was brought up on GitHub as well:

https://github.com/ClangBuiltLinux/linux/issues/1720

I have reported this to the clang developers upstream to see if there is any way that the diagnostic could be improved:

https://reviews.llvm.org/D134461#3815298

Otherwise, we will just have to disable it, which I will send a patch for later if the warning is not going to be adjusted in any way.

If you wanted to continue testing with upstream LLVM, you can either just revert that commit or explicitly disable -Wvoid-ptr-dereference in the main Makefile like so, which is ultimately what I will send if I need to.

Cheers,
Nathan

Makefile +3 -0
index 647a42a1f800..0bf60134f13c 100644
@@ -967,6 +967,9 @@ KBUILD_CFLAGS += $(call cc-disable-warning, stringop-overflow)
 # Another good warning that we'll want to enable eventually
 KBUILD_CFLAGS += $(call cc-disable-warning, restrict)
 
+# New warning from clang that is quite noisy due to '__is_constexpr'
+KBUILD_CFLAGS += $(call cc-disable-warning, void-ptr-dereference)
+
 # Enabled with W=2, disabled by default as noisy
 ifdef CONFIG_CC_IS_GCC
 KBUILD_CFLAGS += -Wno-maybe-uninitialized
#2 lore raw
Quoted 18 lines · Nathan Chancellor

Thanks for testing and the report! I noticed this over the weekend, it comes from a new warning in clang:

https://github.com/llvm/llvm-project/commit/e07ead85a368173a56e96a21d6841aa497ad80f8

It was brought up on GitHub as well:

https://github.com/ClangBuiltLinux/linux/issues/1720

I have reported this to the clang developers upstream to see if there is any way that the diagnostic could be improved:

https://reviews.llvm.org/D134461#3815298

Otherwise, we will just have to disable it, which I will send a patch for later if the warning is not going to be adjusted in any way.

Quoted 7 lines

If you wanted to continue testing with upstream LLVM, you can either just revert that commit or explicitly disable -Wvoid-ptr-dereference in the main Makefile like so, which is ultimately what I will send if I need to.

Cheers,
Nathan

Thanks for the reports. It sounds like the Clang developers will pare back the warning for unevaluated contexts (sizeof, typeof). If it takes longer than a few days for that to land, let's send your below patch to help keep CI systems green.

Consider putting the below into
scripts/Makefile.extrawarn

Quoted 15 lines
Makefile +3 -0
index 647a42a1f800..0bf60134f13c 100644
@@ -967,6 +967,9 @@ KBUILD_CFLAGS += $(call cc-disable-warning, stringop-overflow)
 # Another good warning that we'll want to enable eventually
 KBUILD_CFLAGS += $(call cc-disable-warning, restrict)
 
+# New warning from clang that is quite noisy due to '__is_constexpr'
+KBUILD_CFLAGS += $(call cc-disable-warning, void-ptr-dereference)
+
 # Enabled with W=2, disabled by default as noisy
 ifdef CONFIG_CC_IS_GCC
 KBUILD_CFLAGS += -Wno-maybe-uninitialized
Signature · 2 lines
Thanks,
~Nick Desaulniers