2023-06-28 02:45:45

by Waiman Long

Subject: [PATCH v4 1/4] x86/speculation: Add __update_spec_ctrl() helper

Add a new __update_spec_ctrl() helper which is a variant of
update_spec_ctrl() that can be used in a noinstr function.

Suggested-by: Peter Zijlstra <[email protected]>
Signed-off-by: Waiman Long <[email protected]>
---
arch/x86/include/asm/nospec-branch.h | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h
index 55388c9f7601..1d363fcea207 100644
--- a/arch/x86/include/asm/nospec-branch.h
+++ b/arch/x86/include/asm/nospec-branch.h
@@ -9,7 +9,7 @@

#include <asm/alternative.h>
#include <asm/cpufeatures.h>
-#include <asm/msr-index.h>
+#include <asm/msr.h>
#include <asm/unwind_hints.h>
#include <asm/percpu.h>
#include <asm/current.h>
@@ -488,6 +488,15 @@ DECLARE_PER_CPU(u64, x86_spec_ctrl_current);
extern void update_spec_ctrl_cond(u64 val);
extern u64 spec_ctrl_current(void);

+/*
+ * This can be used in noinstr function.
+ */
+static __always_inline void __update_spec_ctrl(u64 val)
+{
+ __this_cpu_write(x86_spec_ctrl_current, val);
+ native_wrmsrl(MSR_IA32_SPEC_CTRL, val);
+}
+
/*
* With retpoline, we must use IBRS to restrict branch prediction
* before calling into firmware.
--
2.31.1

2023-07-03 10:31:22

by Peter Zijlstra

Subject: Re: [PATCH v4 1/4] x86/speculation: Add __update_spec_ctrl() helper

On Tue, Jun 27, 2023 at 10:25:51PM -0400, Waiman Long wrote:
> Add a new __update_spec_ctrl() helper which is a variant of
> update_spec_ctrl() that can be used in a noinstr function.
>
> Suggested-by: Peter Zijlstra <[email protected]>
> Signed-off-by: Waiman Long <[email protected]>
> ---
> arch/x86/include/asm/nospec-branch.h | 11 ++++++++++-
> 1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h
> index 55388c9f7601..1d363fcea207 100644
> --- a/arch/x86/include/asm/nospec-branch.h
> +++ b/arch/x86/include/asm/nospec-branch.h
> @@ -9,7 +9,7 @@
>
> #include <asm/alternative.h>
> #include <asm/cpufeatures.h>
> -#include <asm/msr-index.h>
> +#include <asm/msr.h>
> #include <asm/unwind_hints.h>
> #include <asm/percpu.h>
> #include <asm/current.h>
> @@ -488,6 +488,15 @@ DECLARE_PER_CPU(u64, x86_spec_ctrl_current);
> extern void update_spec_ctrl_cond(u64 val);
> extern u64 spec_ctrl_current(void);
>
> +/*
> + * This can be used in noinstr function.
> + */
> +static __always_inline void __update_spec_ctrl(u64 val)
> +{
> + __this_cpu_write(x86_spec_ctrl_current, val);
> + native_wrmsrl(MSR_IA32_SPEC_CTRL, val);
> +}

Should we also use this to implement update_spec_ctrl() ?

2023-07-03 15:25:15

by Waiman Long

Subject: Re: [PATCH v4 1/4] x86/speculation: Add __update_spec_ctrl() helper

On 7/3/23 06:28, Peter Zijlstra wrote:
> On Tue, Jun 27, 2023 at 10:25:51PM -0400, Waiman Long wrote:
>> Add a new __update_spec_ctrl() helper which is a variant of
>> update_spec_ctrl() that can be used in a noinstr function.
>>
>> Suggested-by: Peter Zijlstra <[email protected]>
>> Signed-off-by: Waiman Long <[email protected]>
>> ---
>> arch/x86/include/asm/nospec-branch.h | 11 ++++++++++-
>> 1 file changed, 10 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h
>> index 55388c9f7601..1d363fcea207 100644
>> --- a/arch/x86/include/asm/nospec-branch.h
>> +++ b/arch/x86/include/asm/nospec-branch.h
>> @@ -9,7 +9,7 @@
>>
>> #include <asm/alternative.h>
>> #include <asm/cpufeatures.h>
>> -#include <asm/msr-index.h>
>> +#include <asm/msr.h>
>> #include <asm/unwind_hints.h>
>> #include <asm/percpu.h>
>> #include <asm/current.h>
>> @@ -488,6 +488,15 @@ DECLARE_PER_CPU(u64, x86_spec_ctrl_current);
>> extern void update_spec_ctrl_cond(u64 val);
>> extern u64 spec_ctrl_current(void);
>>
>> +/*
>> + * This can be used in noinstr function.
>> + */
>> +static __always_inline void __update_spec_ctrl(u64 val)
>> +{
>> + __this_cpu_write(x86_spec_ctrl_current, val);
>> + native_wrmsrl(MSR_IA32_SPEC_CTRL, val);
>> +}
> Should we also use this to implement update_spec_ctrl() ?

I also thought about that. However, I am a bit worry about losing some
additional check done in this_cpu_write() and the paravirt support in
wrmsrl(). I have no problem making update_spec_ctrl() an instrument-able
wrapper on __update_spec_ctrl() if you think these are not valid concerns.

Cheers,
Longman