[PATCH] L1TF support in PowerPC SYSFS.

[PATCH] L1TF support in PowerPC SYSFS.

[asteinhauser]

From: Anthony Steinhauser <asteinhauser@google.com>

PowerPC CPUs are vulnerable to L1TF to the same extent as to Meltdown.
It is also mitigated by flushing the L1D on privilege transition.
Currently the SYSFS gives a false negative on L1TF on CPUs that I verified
to be vulnerable.
https://www.ibm.com/blogs/psirt/potential-impact-processors-power-family/
https://github.com/google/safeside/pull/52
Signed-off-by: Anthony Steinhauser <asteinhauser@google.com>
---
 arch/powerpc/kernel/security.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/arch/powerpc/kernel/security.c b/arch/powerpc/kernel/security.c
index 7cfcb294b11c..a1a5017c5170 100644
--- a/arch/powerpc/kernel/security.c
+++ b/arch/powerpc/kernel/security.c
@@ -135,7 +135,8 @@ void setup_spectre_v2(void)
 #endif /* CONFIG_PPC_FSL_BOOK3E */
 
 #ifdef CONFIG_PPC_BOOK3S_64
-ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, char *buf)
+static ssize_t cpu_show_delayed_fault_common(
+		struct device *dev, struct device_attribute *attr, char *buf)
 {
 	bool thread_priv;
 
@@ -167,6 +168,16 @@ ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, cha
 
 	return sprintf(buf, "Vulnerable\n");
 }
+
+ssize_t cpu_show_l1tf(struct device *dev, struct device_attribute *attr, char *buf)
+{
+	return cpu_show_delayed_fault_common(dev, attr, buf);
+}
+
+ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, char *buf)
+{
+	return cpu_show_delayed_fault_common(dev, attr, buf);
+}
 #endif
 
 ssize_t cpu_show_spectre_v1(struct device *dev, struct device_attribute *attr, char *buf)
-- 
2.24.0.rc0.303.g954a862665-goog

Re: [PATCH] L1TF support in PowerPC SYSFS.

[Michael Ellerman]

Hi Anthony,

Thanks for the patch.

asteinhauser@google.com writes:
> From: Anthony Steinhauser <asteinhauser@google.com>
>
> PowerPC CPUs are vulnerable to L1TF to the same extent as to Meltdown.
> It is also mitigated by flushing the L1D on privilege transition.
> Currently the SYSFS gives a false negative on L1TF on CPUs that I verified
> to be vulnerable.

Can you include the details of which CPUs you tested?

> https://www.ibm.com/blogs/psirt/potential-impact-processors-power-family/
> https://github.com/google/safeside/pull/52
> Signed-off-by: Anthony Steinhauser <asteinhauser@google.com>
> ---

I don't usually retain links in the change log, because they bit rot, so
I'll drop those, but they will remain in the email archive.

> diff --git a/arch/powerpc/kernel/security.c b/arch/powerpc/kernel/security.c
> index 7cfcb294b11c..a1a5017c5170 100644
> --- a/arch/powerpc/kernel/security.c
> +++ b/arch/powerpc/kernel/security.c
> @@ -135,7 +135,8 @@ void setup_spectre_v2(void)
>  #endif /* CONFIG_PPC_FSL_BOOK3E */
>  
>  #ifdef CONFIG_PPC_BOOK3S_64
> -ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, char *buf)
> +static ssize_t cpu_show_delayed_fault_common(

I guess that's an accurate name.

I probably would have just had cpu_show_l1tf() call cpu_show_meltdown(),
but I guess either approach is fine.

cheers


> @@ -167,6 +168,16 @@ ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, cha
>  
>  	return sprintf(buf, "Vulnerable\n");
>  }
> +
> +ssize_t cpu_show_l1tf(struct device *dev, struct device_attribute *attr, char *buf)
> +{
> +	return cpu_show_delayed_fault_common(dev, attr, buf);
> +}
> +
> +ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, char *buf)
> +{
> +	return cpu_show_delayed_fault_common(dev, attr, buf);
> +}
>  #endif
>  
>  ssize_t cpu_show_spectre_v1(struct device *dev, struct device_attribute *attr, char *buf)
> -- 
> 2.24.0.rc0.303.g954a862665-goog

Re: [PATCH] L1TF support in PowerPC SYSFS.

[Anthony Steinhauser]

Hi Michael,

I tested in on Power9 Talos II Boston 004e 1202, PowerNV T2P9D01.
Links can be naturally dropped. They're just for explanation.

Best,


On Sat, Nov 2, 2019 at 12:11 PM Michael Ellerman <mpe@ellerman.id.au> wrote:
>
> Hi Anthony,
>
> Thanks for the patch.
>
> asteinhauser@google.com writes:
> > From: Anthony Steinhauser <asteinhauser@google.com>
> >
> > PowerPC CPUs are vulnerable to L1TF to the same extent as to Meltdown.
> > It is also mitigated by flushing the L1D on privilege transition.
> > Currently the SYSFS gives a false negative on L1TF on CPUs that I verified
> > to be vulnerable.
>
> Can you include the details of which CPUs you tested?
>
> > https://www.ibm.com/blogs/psirt/potential-impact-processors-power-family/
> > https://github.com/google/safeside/pull/52
> > Signed-off-by: Anthony Steinhauser <asteinhauser@google.com>
> > ---
>
> I don't usually retain links in the change log, because they bit rot, so
> I'll drop those, but they will remain in the email archive.
>
> > diff --git a/arch/powerpc/kernel/security.c b/arch/powerpc/kernel/security.c
> > index 7cfcb294b11c..a1a5017c5170 100644
> > --- a/arch/powerpc/kernel/security.c
> > +++ b/arch/powerpc/kernel/security.c
> > @@ -135,7 +135,8 @@ void setup_spectre_v2(void)
> >  #endif /* CONFIG_PPC_FSL_BOOK3E */
> >
> >  #ifdef CONFIG_PPC_BOOK3S_64
> > -ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, char *buf)
> > +static ssize_t cpu_show_delayed_fault_common(
>
> I guess that's an accurate name.
>
> I probably would have just had cpu_show_l1tf() call cpu_show_meltdown(),
> but I guess either approach is fine.
>
> cheers
>
>
> > @@ -167,6 +168,16 @@ ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, cha
> >
> >       return sprintf(buf, "Vulnerable\n");
> >  }
> > +
> > +ssize_t cpu_show_l1tf(struct device *dev, struct device_attribute *attr, char *buf)
> > +{
> > +     return cpu_show_delayed_fault_common(dev, attr, buf);
> > +}
> > +
> > +ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr, char *buf)
> > +{
> > +     return cpu_show_delayed_fault_common(dev, attr, buf);
> > +}
> >  #endif
> >
> >  ssize_t cpu_show_spectre_v1(struct device *dev, struct device_attribute *attr, char *buf)
> > --
> > 2.24.0.rc0.303.g954a862665-goog

Re: [PATCH] L1TF support in PowerPC SYSFS.

[Michael Ellerman]

On Tue, 2019-10-29 at 19:07:59 UTC, asteinhauser@google.com wrote:
> From: Anthony Steinhauser <asteinhauser@google.com>
> 
> PowerPC CPUs are vulnerable to L1TF to the same extent as to Meltdown.
> It is also mitigated by flushing the L1D on privilege transition.
> Currently the SYSFS gives a false negative on L1TF on CPUs that I verified
> to be vulnerable.
> https://www.ibm.com/blogs/psirt/potential-impact-processors-power-family/
> https://github.com/google/safeside/pull/52
> Signed-off-by: Anthony Steinhauser <asteinhauser@google.com>

Applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/8e6b6da91ac9b9ec5a925b6cb13f287a54bd547d

cheers