Commit ab1a517d55b0 ("powerpc/syscall: Rename syscall_64.c into
interrupt.c") missed to update these three lines:
GCOV_PROFILE_syscall_64.o := n
KCOV_INSTRUMENT_syscall_64.o := n
UBSAN_SANITIZE_syscall_64.o := n
To restore the original behavior, we could replace them with:
GCOV_PROFILE_interrupt.o := n
KCOV_INSTRUMENT_interrupt.o := n
UBSAN_SANITIZE_interrupt.o := n
However, nobody has noticed the functional change in the past three
years, so they were unneeded.
Signed-off-by: Masahiro Yamada <[email protected]>
---
arch/powerpc/kernel/Makefile | 3 ---
1 file changed, 3 deletions(-)
diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index 2919433be355..72d1cd6443bc 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -191,9 +191,6 @@ GCOV_PROFILE_kprobes-ftrace.o := n
KCOV_INSTRUMENT_kprobes-ftrace.o := n
KCSAN_SANITIZE_kprobes-ftrace.o := n
UBSAN_SANITIZE_kprobes-ftrace.o := n
-GCOV_PROFILE_syscall_64.o := n
-KCOV_INSTRUMENT_syscall_64.o := n
-UBSAN_SANITIZE_syscall_64.o := n
UBSAN_SANITIZE_vdso.o := n
# Necessary for booting with kcov enabled on book3e machines
--
2.40.1
+To: Daniel Axtens
Maybe, we should check if the issue fixed by
2f26ed1764b42a8c40d9c48441c73a70d805decf
came back.
On Fri, Feb 16, 2024 at 10:55 PM Masahiro Yamada <[email protected]> wrote:
>
> Commit ab1a517d55b0 ("powerpc/syscall: Rename syscall_64.c into
> interrupt.c") missed to update these three lines:
>
> GCOV_PROFILE_syscall_64.o := n
> KCOV_INSTRUMENT_syscall_64.o := n
> UBSAN_SANITIZE_syscall_64.o := n
>
> To restore the original behavior, we could replace them with:
>
> GCOV_PROFILE_interrupt.o := n
> KCOV_INSTRUMENT_interrupt.o := n
> UBSAN_SANITIZE_interrupt.o := n
>
> However, nobody has noticed the functional change in the past three
> years, so they were unneeded.
>
> Signed-off-by: Masahiro Yamada <[email protected]>
> ---
>
> arch/powerpc/kernel/Makefile | 3 ---
> 1 file changed, 3 deletions(-)
>
> diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
> index 2919433be355..72d1cd6443bc 100644
> --- a/arch/powerpc/kernel/Makefile
> +++ b/arch/powerpc/kernel/Makefile
> @@ -191,9 +191,6 @@ GCOV_PROFILE_kprobes-ftrace.o := n
> KCOV_INSTRUMENT_kprobes-ftrace.o := n
> KCSAN_SANITIZE_kprobes-ftrace.o := n
> UBSAN_SANITIZE_kprobes-ftrace.o := n
> -GCOV_PROFILE_syscall_64.o := n
> -KCOV_INSTRUMENT_syscall_64.o := n
> -UBSAN_SANITIZE_syscall_64.o := n
> UBSAN_SANITIZE_vdso.o := n
>
> # Necessary for booting with kcov enabled on book3e machines
> --
> 2.40.1
>
--
Best Regards
Masahiro Yamada
Masahiro Yamada <[email protected]> writes:
> +To: Daniel Axtens
Unfortunately dja doesn't work on the kernel anymore.
> Maybe, we should check if the issue fixed by
> 2f26ed1764b42a8c40d9c48441c73a70d805decf
> came back.
That crash is specific to Power8, which probably no one is running
syzkaller on anymore.
If I enable KCOV and boot with 1T segments disabled (to make it easier
to trigger the bug) it crashes similarly to the report in that commit.
> On Fri, Feb 16, 2024 at 10:55 PM Masahiro Yamada <[email protected]> wrote:
..
>> To restore the original behavior, we could replace them with:
>>
>> GCOV_PROFILE_interrupt.o := n
>> KCOV_INSTRUMENT_interrupt.o := n
>> UBSAN_SANITIZE_interrupt.o := n
But just putting those back isn't actually enough to fix it, the code
has changed and there are other places that need KCOV disabled on P8.
So I'm not sure how to handle this one. I guess I might just take it and
then make a todo to fix the KCOV problems later.
cheers