commit edf2d546bfd6 ("riscv: patch: Flush the icache right after
patching to avoid illegal insns") mistakenly removed the global icache
flush in patch_text_nosync() and patch_text_set_nosync() functions, so
reintroduce them.

Fixes: edf2d546bfd6 ("riscv: patch: Flush the icache right after patching to avoid illegal insns")
Reported-by: Samuel Holland <samuel.holland@sifive.com>
Closes: https://lore.kernel.org/linux-riscv/CAHVXubh8Adb4=-vN4cSh0FrZ16TeOKJbLj4AF09QC241bRk1Jg@mail.gmail.com/T/#m800757c26f72a1d45c240cb815650430166c82ea
Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com>
---
 arch/riscv/kernel/patch.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/arch/riscv/kernel/patch.c b/arch/riscv/kernel/patch.c
index ab03732d06c4..91edfd764ed9 100644
--- a/arch/riscv/kernel/patch.c
+++ b/arch/riscv/kernel/patch.c
@@ -205,6 +205,9 @@ int patch_text_set_nosync(void *addr, u8 c, size_t len)
 
 	ret = patch_insn_set(tp, c, len);
 
+	if (!ret)
+		flush_icache_range((uintptr_t)tp, (uintptr_t)tp + len);
+
 	return ret;
 }
 NOKPROBE_SYMBOL(patch_text_set_nosync);
@@ -237,6 +240,9 @@ int patch_text_nosync(void *addr, const void *insns, size_t len)
 
 	ret = patch_insn_write(tp, insns, len);
 
+	if (!ret)
+		flush_icache_range((uintptr_t) tp, (uintptr_t) tp + len);
+
 	return ret;
 }
 NOKPROBE_SYMBOL(patch_text_nosync);
-- 
2.39.2

Hi Alex,

On 2024-07-30 8:59 AM, Alexandre Ghiti wrote:
> commit edf2d546bfd6 ("riscv: patch: Flush the icache right after
> patching to avoid illegal insns") mistakenly removed the global icache
> flush in patch_text_nosync() and patch_text_set_nosync() functions, so
> reintroduce them.
> 
> Fixes: edf2d546bfd6 ("riscv: patch: Flush the icache right after patching to avoid illegal insns")
> Reported-by: Samuel Holland <samuel.holland@sifive.com>
> Closes: https://lore.kernel.org/linux-riscv/CAHVXubh8Adb4=-vN4cSh0FrZ16TeOKJbLj4AF09QC241bRk1Jg@mail.gmail.com/T/#m800757c26f72a1d45c240cb815650430166c82ea

Shouldn't this use the permalink for the specific message, not the thread?

> Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com>
> ---
>  arch/riscv/kernel/patch.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/arch/riscv/kernel/patch.c b/arch/riscv/kernel/patch.c
> index ab03732d06c4..91edfd764ed9 100644
> --- a/arch/riscv/kernel/patch.c
> +++ b/arch/riscv/kernel/patch.c
> @@ -205,6 +205,9 @@ int patch_text_set_nosync(void *addr, u8 c, size_t len)
>  
>  	ret = patch_insn_set(tp, c, len);
>  
> +	if (!ret)
> +		flush_icache_range((uintptr_t)tp, (uintptr_t)tp + len);

This patch was based on an old tree from before
https://git.kernel.org/riscv/c/47742484ee16 removed the "tp" variable. While it
still compiles because flush_icache_range() is a macro that discards its
arguments, it will be confusing to anyone reading the code.

Regards,
Samuel

> +
>  	return ret;
>  }
>  NOKPROBE_SYMBOL(patch_text_set_nosync);
> @@ -237,6 +240,9 @@ int patch_text_nosync(void *addr, const void *insns, size_t len)
>  
>  	ret = patch_insn_write(tp, insns, len);
>  
> +	if (!ret)
> +		flush_icache_range((uintptr_t) tp, (uintptr_t) tp + len);
> +
>  	return ret;
>  }
>  NOKPROBE_SYMBOL(patch_text_nosync);

Hi Samuel,

On 01/08/2024 19:32, Samuel Holland wrote:
> Hi Alex,
>
> On 2024-07-30 8:59 AM, Alexandre Ghiti wrote:
>> commit edf2d546bfd6 ("riscv: patch: Flush the icache right after
>> patching to avoid illegal insns") mistakenly removed the global icache
>> flush in patch_text_nosync() and patch_text_set_nosync() functions, so
>> reintroduce them.
>>
>> Fixes: edf2d546bfd6 ("riscv: patch: Flush the icache right after patching to avoid illegal insns")
>> Reported-by: Samuel Holland <samuel.holland@sifive.com>
>> Closes: https://lore.kernel.org/linux-riscv/CAHVXubh8Adb4=-vN4cSh0FrZ16TeOKJbLj4AF09QC241bRk1Jg@mail.gmail.com/T/#m800757c26f72a1d45c240cb815650430166c82ea
> Shouldn't this use the permalink for the specific message, not the thread?


I never paid attention to that, but the permalink is shorter so I'll 
change that, thanks


>
>> Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com>
>> ---
>>   arch/riscv/kernel/patch.c | 6 ++++++
>>   1 file changed, 6 insertions(+)
>>
>> diff --git a/arch/riscv/kernel/patch.c b/arch/riscv/kernel/patch.c
>> index ab03732d06c4..91edfd764ed9 100644
>> --- a/arch/riscv/kernel/patch.c
>> +++ b/arch/riscv/kernel/patch.c
>> @@ -205,6 +205,9 @@ int patch_text_set_nosync(void *addr, u8 c, size_t len)
>>   
>>   	ret = patch_insn_set(tp, c, len);
>>   
>> +	if (!ret)
>> +		flush_icache_range((uintptr_t)tp, (uintptr_t)tp + len);
> This patch was based on an old tree from before
> https://git.kernel.org/riscv/c/47742484ee16 removed the "tp" variable. While it
> still compiles because flush_icache_range() is a macro that discards its
> arguments, it will be confusing to anyone reading the code.


Hmmm not the first time I hear this :) Sorry, I prepared the patch a 
couple of weeks ago and the rebase on 6.11-rc1 did not complain, I do 
that right now and send a new version tomorrow.

Thanks,

Alex


>
> Regards,
> Samuel
>
>> +
>>   	return ret;
>>   }
>>   NOKPROBE_SYMBOL(patch_text_set_nosync);
>> @@ -237,6 +240,9 @@ int patch_text_nosync(void *addr, const void *insns, size_t len)
>>   
>>   	ret = patch_insn_write(tp, insns, len);
>>   
>> +	if (!ret)
>> +		flush_icache_range((uintptr_t) tp, (uintptr_t) tp + len);
>> +
>>   	return ret;
>>   }
>>   NOKPROBE_SYMBOL(patch_text_nosync);
>
> _______________________________________________
> linux-riscv mailing list
> linux-riscv@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-riscv

On Thu, 01 Aug 2024 10:32:28 PDT (-0700), samuel.holland@sifive.com wrote:
> Hi Alex,
>
> On 2024-07-30 8:59 AM, Alexandre Ghiti wrote:
>> commit edf2d546bfd6 ("riscv: patch: Flush the icache right after
>> patching to avoid illegal insns") mistakenly removed the global icache
>> flush in patch_text_nosync() and patch_text_set_nosync() functions, so
>> reintroduce them.
>>
>> Fixes: edf2d546bfd6 ("riscv: patch: Flush the icache right after patching to avoid illegal insns")
>> Reported-by: Samuel Holland <samuel.holland@sifive.com>
>> Closes: https://lore.kernel.org/linux-riscv/CAHVXubh8Adb4=-vN4cSh0FrZ16TeOKJbLj4AF09QC241bRk1Jg@mail.gmail.com/T/#m800757c26f72a1d45c240cb815650430166c82ea
>
> Shouldn't this use the permalink for the specific message, not the thread?
>
>> Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com>
>> ---
>>  arch/riscv/kernel/patch.c | 6 ++++++
>>  1 file changed, 6 insertions(+)
>>
>> diff --git a/arch/riscv/kernel/patch.c b/arch/riscv/kernel/patch.c
>> index ab03732d06c4..91edfd764ed9 100644
>> --- a/arch/riscv/kernel/patch.c
>> +++ b/arch/riscv/kernel/patch.c
>> @@ -205,6 +205,9 @@ int patch_text_set_nosync(void *addr, u8 c, size_t len)
>>
>>  	ret = patch_insn_set(tp, c, len);
>>
>> +	if (!ret)
>> +		flush_icache_range((uintptr_t)tp, (uintptr_t)tp + len);
>
> This patch was based on an old tree from before
> https://git.kernel.org/riscv/c/47742484ee16 removed the "tp" variable. While it
> still compiles because flush_icache_range() is a macro that discards its
> arguments, it will be confusing to anyone reading the code.

Thanks.  Alex is going to spin a new one, so I dropped this.

>
> Regards,
> Samuel
>
>> +
>>  	return ret;
>>  }
>>  NOKPROBE_SYMBOL(patch_text_set_nosync);
>> @@ -237,6 +240,9 @@ int patch_text_nosync(void *addr, const void *insns, size_t len)
>>
>>  	ret = patch_insn_write(tp, insns, len);
>>
>> +	if (!ret)
>> +		flush_icache_range((uintptr_t) tp, (uintptr_t) tp + len);
>> +
>>  	return ret;
>>  }
>>  NOKPROBE_SYMBOL(patch_text_nosync);