[PATCHv7 13/16] x86/traps: Handle LASS thrown #SS

Kirill A. Shutemov posted 16 patches 3 months, 2 weeks ago
There is a newer version of this series
[PATCHv7 13/16] x86/traps: Handle LASS thrown #SS
Posted by Kirill A. Shutemov 3 months, 2 weeks ago
LASS throws a #GP for any violations except for stack register accesses,
in which case it throws a #SS instead. Handle this similarly to how other
LASS violations are handled.

In case of FRED, before handling #SS as LASS violation, kernel has to
check if there's a fixup for the exception. It can address #SS due to
invalid user context on ERETU[1]. See 5105e7687ad3 ("x86/fred: Fixup
fault on ERETU by jumping to fred_entrypoint_user") for more details.

Co-developed-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 arch/x86/kernel/traps.c | 39 +++++++++++++++++++++++++++++++++------
 1 file changed, 33 insertions(+), 6 deletions(-)

diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index e2ad760b17ea..f1f92e1ba524 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -418,12 +418,6 @@ DEFINE_IDTENTRY_ERRORCODE(exc_segment_not_present)
 		      SIGBUS, 0, NULL);
 }
 
-DEFINE_IDTENTRY_ERRORCODE(exc_stack_segment)
-{
-	do_error_trap(regs, error_code, "stack segment", X86_TRAP_SS, SIGBUS,
-		      0, NULL);
-}
-
 DEFINE_IDTENTRY_ERRORCODE(exc_alignment_check)
 {
 	char *str = "alignment check";
@@ -866,6 +860,39 @@ DEFINE_IDTENTRY_ERRORCODE(exc_general_protection)
 	cond_local_irq_disable(regs);
 }
 
+#define SSFSTR "stack segment fault"
+
+DEFINE_IDTENTRY_ERRORCODE(exc_stack_segment)
+{
+	if (user_mode(regs))
+		goto error_trap;
+
+	if (cpu_feature_enabled(X86_FEATURE_FRED) &&
+	    fixup_exception(regs, X86_TRAP_SS, error_code, 0))
+		return;
+
+	if (cpu_feature_enabled(X86_FEATURE_LASS)) {
+		enum kernel_gp_hint hint;
+		unsigned long gp_addr;
+
+		hint = get_kernel_gp_address(regs, &gp_addr);
+		if (hint != GP_NO_HINT) {
+			printk(SSFSTR ", %s 0x%lx", kernel_gp_hint_help[hint],
+			       gp_addr);
+		}
+
+		if (hint != GP_NON_CANONICAL)
+			gp_addr = 0;
+
+		die_addr(SSFSTR, regs, error_code, gp_addr);
+		return;
+	}
+
+error_trap:
+	do_error_trap(regs, error_code, "stack segment", X86_TRAP_SS, SIGBUS,
+		0, NULL);
+}
+
 static bool do_int3(struct pt_regs *regs)
 {
 	int res;
-- 
2.47.2
Re: [PATCHv7 13/16] x86/traps: Handle LASS thrown #SS
Posted by Xin Li 3 months, 2 weeks ago
On 6/25/2025 5:51 AM, Kirill A. Shutemov wrote:
> LASS throws a #GP for any violations except for stack register accesses,
> in which case it throws a #SS instead. Handle this similarly to how other
> LASS violations are handled.
> 
> In case of FRED, before handling #SS as LASS violation, kernel has to
> check if there's a fixup for the exception. It can address #SS due to
> invalid user context on ERETU[1]. See 5105e7687ad3 ("x86/fred: Fixup

Forgot to put the link to [1]?  Maybe just remove "[1]"?

> fault on ERETU by jumping to fred_entrypoint_user") for more details.
> 
> Co-developed-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
> Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> ---
>   arch/x86/kernel/traps.c | 39 +++++++++++++++++++++++++++++++++------
>   1 file changed, 33 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
> index e2ad760b17ea..f1f92e1ba524 100644
> --- a/arch/x86/kernel/traps.c
> +++ b/arch/x86/kernel/traps.c
> @@ -418,12 +418,6 @@ DEFINE_IDTENTRY_ERRORCODE(exc_segment_not_present)
>   		      SIGBUS, 0, NULL);
>   }
>   
> -DEFINE_IDTENTRY_ERRORCODE(exc_stack_segment)
> -{
> -	do_error_trap(regs, error_code, "stack segment", X86_TRAP_SS, SIGBUS,
> -		      0, NULL);
> -}
> -
>   DEFINE_IDTENTRY_ERRORCODE(exc_alignment_check)
>   {
>   	char *str = "alignment check";
> @@ -866,6 +860,39 @@ DEFINE_IDTENTRY_ERRORCODE(exc_general_protection)
>   	cond_local_irq_disable(regs);
>   }
>   
> +#define SSFSTR "stack segment fault"
> +
> +DEFINE_IDTENTRY_ERRORCODE(exc_stack_segment)
> +{
> +	if (user_mode(regs))
> +		goto error_trap;
> +
> +	if (cpu_feature_enabled(X86_FEATURE_FRED) &&
> +	    fixup_exception(regs, X86_TRAP_SS, error_code, 0))
> +		return;
> +

Thanks for making the change for FRED.

> +	if (cpu_feature_enabled(X86_FEATURE_LASS)) {
> +		enum kernel_gp_hint hint;
> +		unsigned long gp_addr;
> +
> +		hint = get_kernel_gp_address(regs, &gp_addr);
> +		if (hint != GP_NO_HINT) {
> +			printk(SSFSTR ", %s 0x%lx", kernel_gp_hint_help[hint],
> +			       gp_addr);
> +		}
> +
> +		if (hint != GP_NON_CANONICAL)
> +			gp_addr = 0;

Nit: GP/gp don't seem fit here, maybe we need a more generic name?

Sorry I don't have a recommendation.

> +
> +		die_addr(SSFSTR, regs, error_code, gp_addr);
> +		return;
> +	}
> +
> +error_trap:
> +	do_error_trap(regs, error_code, "stack segment", X86_TRAP_SS, SIGBUS,
> +		0, NULL);

The indentation has changed; I believe the original formatting is
preferable.

> +}
> +
>   static bool do_int3(struct pt_regs *regs)
>   {
>   	int res;

Just minor comments, so

Reviewed-by: Xin Li (Intel) <xin@zytor.com>
Re: [PATCHv7 13/16] x86/traps: Handle LASS thrown #SS
Posted by Kirill A. Shutemov 3 months, 1 week ago
On Thu, Jun 26, 2025 at 10:57:47AM -0700, Xin Li wrote:
> On 6/25/2025 5:51 AM, Kirill A. Shutemov wrote:
> > LASS throws a #GP for any violations except for stack register accesses,
> > in which case it throws a #SS instead. Handle this similarly to how other
> > LASS violations are handled.
> > 
> > In case of FRED, before handling #SS as LASS violation, kernel has to
> > check if there's a fixup for the exception. It can address #SS due to
> > invalid user context on ERETU[1]. See 5105e7687ad3 ("x86/fred: Fixup
> 
> Forgot to put the link to [1]?  Maybe just remove "[1]"?

I will add the link. It is important context.

> > fault on ERETU by jumping to fred_entrypoint_user") for more details.
> > 
> > Co-developed-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
> > Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
> > Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> > ---
> >   arch/x86/kernel/traps.c | 39 +++++++++++++++++++++++++++++++++------
> >   1 file changed, 33 insertions(+), 6 deletions(-)
> > 
> > diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
> > index e2ad760b17ea..f1f92e1ba524 100644
> > --- a/arch/x86/kernel/traps.c
> > +++ b/arch/x86/kernel/traps.c
> > @@ -418,12 +418,6 @@ DEFINE_IDTENTRY_ERRORCODE(exc_segment_not_present)
> >   		      SIGBUS, 0, NULL);
> >   }
> > -DEFINE_IDTENTRY_ERRORCODE(exc_stack_segment)
> > -{
> > -	do_error_trap(regs, error_code, "stack segment", X86_TRAP_SS, SIGBUS,
> > -		      0, NULL);
> > -}
> > -
> >   DEFINE_IDTENTRY_ERRORCODE(exc_alignment_check)
> >   {
> >   	char *str = "alignment check";
> > @@ -866,6 +860,39 @@ DEFINE_IDTENTRY_ERRORCODE(exc_general_protection)
> >   	cond_local_irq_disable(regs);
> >   }
> > +#define SSFSTR "stack segment fault"
> > +
> > +DEFINE_IDTENTRY_ERRORCODE(exc_stack_segment)
> > +{
> > +	if (user_mode(regs))
> > +		goto error_trap;
> > +
> > +	if (cpu_feature_enabled(X86_FEATURE_FRED) &&
> > +	    fixup_exception(regs, X86_TRAP_SS, error_code, 0))
> > +		return;
> > +
> 
> Thanks for making the change for FRED.
> 
> > +	if (cpu_feature_enabled(X86_FEATURE_LASS)) {
> > +		enum kernel_gp_hint hint;
> > +		unsigned long gp_addr;
> > +
> > +		hint = get_kernel_gp_address(regs, &gp_addr);
> > +		if (hint != GP_NO_HINT) {
> > +			printk(SSFSTR ", %s 0x%lx", kernel_gp_hint_help[hint],
> > +			       gp_addr);
> > +		}
> > +
> > +		if (hint != GP_NON_CANONICAL)
> > +			gp_addr = 0;
> 
> Nit: GP/gp don't seem fit here, maybe we need a more generic name?
> 
> Sorry I don't have a recommendation.

Naming is hard.

Maybe get_kernel_exc_address()/kernel_exc_hint_help/EXC_NO_HINT/... ?

> > +
> > +		die_addr(SSFSTR, regs, error_code, gp_addr);
> > +		return;
> > +	}
> > +
> > +error_trap:
> > +	do_error_trap(regs, error_code, "stack segment", X86_TRAP_SS, SIGBUS,
> > +		0, NULL);
> 
> The indentation has changed; I believe the original formatting is
> preferable.
> 
> > +}
> > +
> >   static bool do_int3(struct pt_regs *regs)
> >   {
> >   	int res;
> 
> Just minor comments, so
> 
> Reviewed-by: Xin Li (Intel) <xin@zytor.com>

Thanks.

-- 
  Kiryl Shutsemau / Kirill A. Shutemov
Re: [PATCHv7 13/16] x86/traps: Handle LASS thrown #SS
Posted by David Laight 3 months, 1 week ago
On Fri, 27 Jun 2025 13:31:44 +0300
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> wrote:

> On Thu, Jun 26, 2025 at 10:57:47AM -0700, Xin Li wrote:
> > On 6/25/2025 5:51 AM, Kirill A. Shutemov wrote:  
> > > LASS throws a #GP for any violations except for stack register accesses,
> > > in which case it throws a #SS instead. Handle this similarly to how other
> > > LASS violations are handled.
> > > 
> > > In case of FRED, before handling #SS as LASS violation, kernel has to
> > > check if there's a fixup for the exception. It can address #SS due to
> > > invalid user context on ERETU[1]. See 5105e7687ad3 ("x86/fred: Fixup  
> > 
> > Forgot to put the link to [1]?  Maybe just remove "[1]"?  
> 
> I will add the link. It is important context.

Will the link still be valid in 5 years time when someone
is looking back at the changes?

	David
Re: [PATCHv7 13/16] x86/traps: Handle LASS thrown #SS
Posted by Kirill A. Shutemov 3 months, 1 week ago
On Mon, Jun 30, 2025 at 09:30:27AM +0100, David Laight wrote:
> On Fri, 27 Jun 2025 13:31:44 +0300
> "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> wrote:
> 
> > On Thu, Jun 26, 2025 at 10:57:47AM -0700, Xin Li wrote:
> > > On 6/25/2025 5:51 AM, Kirill A. Shutemov wrote:  
> > > > LASS throws a #GP for any violations except for stack register accesses,
> > > > in which case it throws a #SS instead. Handle this similarly to how other
> > > > LASS violations are handled.
> > > > 
> > > > In case of FRED, before handling #SS as LASS violation, kernel has to
> > > > check if there's a fixup for the exception. It can address #SS due to
> > > > invalid user context on ERETU[1]. See 5105e7687ad3 ("x86/fred: Fixup  
> > > 
> > > Forgot to put the link to [1]?  Maybe just remove "[1]"?  
> > 
> > I will add the link. It is important context.
> 
> Will the link still be valid in 5 years time when someone
> is looking back at the changes?

Re-reading the commit message I wrote, it is obvious that I reconsidered
putting the link and referenced commit instead.

I will drop [1];

-- 
  Kiryl Shutsemau / Kirill A. Shutemov