[PATCH -next] riscv: Enable ARCH_HAVE_NMI_SAFE_CMPXCHG

[PATCH -next] riscv: Enable ARCH_HAVE_NMI_SAFE_CMPXCHG

[Pu Lehui]

From: Pu Lehui <pulehui@huawei.com>

The implement of cmpxchg() in riscv is based on atomic primitives and
has NMI-safe features, so it can be used safely in the in_nmi context.
ftrace's ringbuffer relies on NMI-safe cmpxchg() in the NMI context.
Currently, in_nmi() is true when riscv kprobe is in trap-based mode, so
this config needs to be selected, otherwise kprobetrace will not be
available.

Signed-off-by: Pu Lehui <pulehui@huawei.com>
---
 arch/riscv/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index 36061f4732b7..0e2dfa98338a 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -23,6 +23,7 @@ config RISCV
 	select ARCH_ENABLE_MEMORY_HOTREMOVE if MEMORY_HOTPLUG
 	select ARCH_ENABLE_SPLIT_PMD_PTLOCK if PGTABLE_LEVELS > 2
 	select ARCH_ENABLE_THP_MIGRATION if TRANSPARENT_HUGEPAGE
+	select ARCH_HAVE_NMI_SAFE_CMPXCHG
 	select ARCH_HAS_BINFMT_FLAT
 	select ARCH_HAS_CRC32 if RISCV_ISA_ZBC
 	select ARCH_HAS_CRC64 if 64BIT && RISCV_ISA_ZBC
-- 
2.34.1

Re: [PATCH -next] riscv: Enable ARCH_HAVE_NMI_SAFE_CMPXCHG

[Pu Lehui]

On 2025/7/11 17:04, Pu Lehui wrote:
> From: Pu Lehui <pulehui@huawei.com>
> 
> The implement of cmpxchg() in riscv is based on atomic primitives and
> has NMI-safe features, so it can be used safely in the in_nmi context.
> ftrace's ringbuffer relies on NMI-safe cmpxchg() in the NMI context.
> Currently, in_nmi() is true when riscv kprobe is in trap-based mode, so
> this config needs to be selected, otherwise kprobetrace will not be

Gentle ping~

The issue can be reproduced by the following steps:

cd /sys/kernel/debug/tracing/
echo 'p do_nanosleep' > kprobe_events
echo 1 > events/kprobes/enable
sleep 1
cat trace # trace log will not show without this patch

> available.
> 
> Signed-off-by: Pu Lehui <pulehui@huawei.com>
> ---
>   arch/riscv/Kconfig | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> index 36061f4732b7..0e2dfa98338a 100644
> --- a/arch/riscv/Kconfig
> +++ b/arch/riscv/Kconfig
> @@ -23,6 +23,7 @@ config RISCV
>   	select ARCH_ENABLE_MEMORY_HOTREMOVE if MEMORY_HOTPLUG
>   	select ARCH_ENABLE_SPLIT_PMD_PTLOCK if PGTABLE_LEVELS > 2
>   	select ARCH_ENABLE_THP_MIGRATION if TRANSPARENT_HUGEPAGE
> +	select ARCH_HAVE_NMI_SAFE_CMPXCHG
>   	select ARCH_HAS_BINFMT_FLAT
>   	select ARCH_HAS_CRC32 if RISCV_ISA_ZBC
>   	select ARCH_HAS_CRC64 if 64BIT && RISCV_ISA_ZBC

Re: [PATCH -next] riscv: Enable ARCH_HAVE_NMI_SAFE_CMPXCHG

[Alexandre Ghiti]

Hi Pu,

On 7/11/25 11:04, Pu Lehui wrote:
> From: Pu Lehui <pulehui@huawei.com>
>
> The implement of cmpxchg() in riscv is based on atomic primitives and
> has NMI-safe features, so it can be used safely in the in_nmi context.
> ftrace's ringbuffer relies on NMI-safe cmpxchg() in the NMI context.
> Currently, in_nmi() is true when riscv kprobe is in trap-based mode, so
> this config needs to be selected, otherwise kprobetrace will not be
> available.
>
> Signed-off-by: Pu Lehui <pulehui@huawei.com>
> ---
>   arch/riscv/Kconfig | 1 +
>   1 file changed, 1 insertion(+)
>
> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> index 36061f4732b7..0e2dfa98338a 100644
> --- a/arch/riscv/Kconfig
> +++ b/arch/riscv/Kconfig
> @@ -23,6 +23,7 @@ config RISCV
>   	select ARCH_ENABLE_MEMORY_HOTREMOVE if MEMORY_HOTPLUG
>   	select ARCH_ENABLE_SPLIT_PMD_PTLOCK if PGTABLE_LEVELS > 2
>   	select ARCH_ENABLE_THP_MIGRATION if TRANSPARENT_HUGEPAGE
> +	select ARCH_HAVE_NMI_SAFE_CMPXCHG
>   	select ARCH_HAS_BINFMT_FLAT
>   	select ARCH_HAS_CRC32 if RISCV_ISA_ZBC
>   	select ARCH_HAS_CRC64 if 64BIT && RISCV_ISA_ZBC


Given the definition of ARCH_HAVE_NMI_SAFE_CMPXCHG in commit 
df013ffb8119 ("Add Kconfig option ARCH_HAVE_NMI_SAFE_CMPXCHG"), we can 
indeed enable this config, so:

Reviewed-by: Alexandre Ghiti <alexghiti@rivosinc.com>

And I followed your instructions in your answer, it works as you 
presented so:

Tested-by: Alexandre Ghiti <alexghiti@rivosinc.com>

I'll send this as part of my 6.17 PR.

Thanks,

Alex

Re: [PATCH -next] riscv: Enable ARCH_HAVE_NMI_SAFE_CMPXCHG

[patchwork-bot+linux-riscv]

Hello:

This patch was applied to riscv/linux.git (for-next)
by Alexandre Ghiti <alexghiti@rivosinc.com>:

On Fri, 11 Jul 2025 09:04:43 +0000 you wrote:
> From: Pu Lehui <pulehui@huawei.com>
> 
> The implement of cmpxchg() in riscv is based on atomic primitives and
> has NMI-safe features, so it can be used safely in the in_nmi context.
> ftrace's ringbuffer relies on NMI-safe cmpxchg() in the NMI context.
> Currently, in_nmi() is true when riscv kprobe is in trap-based mode, so
> this config needs to be selected, otherwise kprobetrace will not be
> available.
> 
> [...]

Here is the summary with links:
  - [-next] riscv: Enable ARCH_HAVE_NMI_SAFE_CMPXCHG
    https://git.kernel.org/riscv/c/9ad032ba418c

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html