x86/speculation: Fix user-mode spectre-v2 protection with KERNEL_IBRS

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

Quoted 21 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

Quoted 3 lines

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

thanks,

greg k-h

Quoted 27 lines · Greg KH

Quoted 21 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

Quoted 3 lines

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

Quoted 2 lines

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

Sure, it's okay. Please do note in my first patch, I did follow
https://www.kernel.org/doc/Documentation/admin-guide/security-bugs.rst,
if you want folks to explicitly Cc maintainers with their fix or report, I think it's worth mentioning in the guidelines there as the current language seems to imply that the maintainers will be pulled in by the security team:

"It is possible that the security team will bring in extra help from area maintainers to understand and fix the security vulnerability."

Quoted 38 lines · KP Singh

Quoted 27 lines · Greg KH

Quoted 21 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

Quoted 3 lines

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

Quoted 2 lines

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

Sure, it's okay. Please do note in my first patch, I did follow
https://www.kernel.org/doc/Documentation/admin-guide/security-bugs.rst,
if you want folks to explicitly Cc maintainers with their fix or report, I think it's worth mentioning in the guidelines there as the current language seems to imply that the maintainers will be pulled in by the security team:

Quoted 2 lines

"It is possible that the security team will bring in extra help from area maintainers to understand and fix the security vulnerability."

Yes, but you already have a patch here, what "help" do you need? You didn't specify any help, you just sent us a patch with no context. This wasn't any sort of a "report" or "hey, I think we found a problem over here, does this change look correct", right?

So please be specific as to what you are asking for, otherwise we have to guess (i.e. you cc:ed a seemingly random set of people but not the

x86 maintainers). And then you resent it to a public list, so there's no need for [email protected] to get involved in at all as it's a public issue now.

thanks,

greg k-h

Quoted 48 lines · Greg KH

Quoted 38 lines · KP Singh

Quoted 27 lines · Greg KH

Quoted 21 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

Quoted 3 lines

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

Quoted 2 lines

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

Sure, it's okay. Please do note in my first patch, I did follow
https://www.kernel.org/doc/Documentation/admin-guide/security-bugs.rst,
if you want folks to explicitly Cc maintainers with their fix or report, I think it's worth mentioning in the guidelines there as the current language seems to imply that the maintainers will be pulled in by the security team:

Quoted 2 lines

"It is possible that the security team will bring in extra help from area maintainers to understand and fix the security vulnerability."

Yes, but you already have a patch here, what "help" do you need? You didn't specify any help, you just sent us a patch with no context. This wasn't any sort of a "report" or "hey, I think we found a problem over here, does this change look correct", right?

Quoted 2 lines

So please be specific as to what you are asking for, otherwise we have to guess (i.e. you cc:ed a seemingly random set of people but not the

I don't see how it matters who I cc on the list. Anyways, I am still

not clear on what one is supposed to do in the case when one has a patch for an issue already. Should this not be send it to security@?

Quoted 3 lines

x86 maintainers). And then you resent it to a public list, so there's no need for [email protected] to get involved in at all as it's a public issue now.

Agreed.

Quoted 4 lines

thanks,

greg k-h

Quoted 48 lines · Greg KH

Quoted 38 lines · KP Singh

Quoted 27 lines · Greg KH

Quoted 21 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

Quoted 3 lines

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

Quoted 2 lines

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

Sure, it's okay. Please do note in my first patch, I did follow
https://www.kernel.org/doc/Documentation/admin-guide/security-bugs.rst,
if you want folks to explicitly Cc maintainers with their fix or report, I think it's worth mentioning in the guidelines there as the current language seems to imply that the maintainers will be pulled in by the security team:

Quoted 2 lines

"It is possible that the security team will bring in extra help from area maintainers to understand and fix the security vulnerability."

Yes, but you already have a patch here, what "help" do you need? You didn't specify any help, you just sent us a patch with no context. This wasn't any sort of a "report" or "hey, I think we found a problem over here, does this change look correct", right?

Quoted 2 lines

So please be specific as to what you are asking for, otherwise we have to guess (i.e. you cc:ed a seemingly random set of people but not the

I don't see how it matters who I cc on the list.

Anyways, I am still
not clear on what one is supposed to do in the case when one has a patch for an issue already. Should this not be send it to security@?

Quoted 49 lines · KP Singh

Quoted 48 lines · Greg KH

Quoted 38 lines · KP Singh

Quoted 27 lines · Greg KH

Quoted 21 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

Quoted 3 lines

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

Quoted 2 lines

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

Sure, it's okay. Please do note in my first patch, I did follow
https://www.kernel.org/doc/Documentation/admin-guide/security-bugs.rst,
if you want folks to explicitly Cc maintainers with their fix or report, I think it's worth mentioning in the guidelines there as the current language seems to imply that the maintainers will be pulled in by the security team:

Quoted 2 lines

"It is possible that the security team will bring in extra help from area maintainers to understand and fix the security vulnerability."

Yes, but you already have a patch here, what "help" do you need? You didn't specify any help, you just sent us a patch with no context. This wasn't any sort of a "report" or "hey, I think we found a problem over here, does this change look correct", right?

Quoted 4 lines

Quoted 2 lines

So please be specific as to what you are asking for, otherwise we have to guess (i.e. you cc:ed a seemingly random set of people but not the

I don't see how it matters who I cc on the list.

It gives us a hint as to who you are leaving out, right?

Quoted 3 lines

Anyways, I am still
not clear on what one is supposed to do in the case when one has a patch for an issue already. Should this not be send it to security@?

security@ is to take reports of potential security problems, triage them, and drag in the respective people to fix the problem as soon as possible by creating a patch and getting it merged.

You already had a patch, so you did all of the work that security@ would normally do, so what did you need us to do here?

You also did not ask or request anything, you just sent a patch with no context other than the changelog text.

So if you have a fix for a potential problem already, you either just send it to get it merged as soon as possible, in which case security@ is not needed. Or if you want to ask questions "is this really an issue and is this the fix", then send the patch and ask that question.

Again, as it is, this looks like any other normal patch sent to subsystems for review, and there was no request for help or context at all. Then you sent the patch to a public mailing list, so security@ is not needed at al, the normal development process applies as you determined it's not a secret by doing so.

If you have questions, ask them, we can't read minds.

thanks,

greg k-h

until I was nudged to read
https://people.kernel.org/tglx/notes-about-netiquette.

Quoted 21 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

Quoted 3 lines

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

Quoted 26 lines · Greg KH

Quoted 24 lines · KP Singh

With the introduction of KERNEL_IBRS, STIBP is no longer needed to prevent cross thread training in the kernel space. When KERNEL_IBRS was added, it also disabled the user-mode protections for spectre_v2. KERNEL_IBRS does not mitigate cross thread training in the userspace.

In order to demonstrate the issue, one needs to avoid syscalls in the victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the attacker may have done.

Allow users to select a spectre_v2_user mitigation (STIBP always on, opt-in via prctl) when KERNEL_IBRS is enabled.

Reported-by José Oliveira <[email protected]>

Reported-by Rodrigo Branco <[email protected]>

Reviewed-by Alexandra Sandulescu <[email protected]>

Reviewed-by Jim Mattson <[email protected]>

Fixes 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")

Cc [email protected]

Signed-off-by KP Singh <[email protected]>

---
 arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

Quoted 2 lines

As this is posted publicly, there's no need to send it to [email protected], it doesn't need to be involved.

Also, since this seems intended to be public, please add lkml to Cc on the next revision.

--
Josh

+++ b/arch/x86/kernel/cpu/bugs.c

arch/x86/kernel/cpu/bugs.c +2 -0

@@ -1132,6 +1132,19 @@ static inline bool spectre_v2_in_ibrs_mode(enum spectre_v2_mitigation mode)
 	       mode == SPECTRE_V2_EIBRS_LFENCE;
 }
 
+static inline bool spectre_v2_user_no_stibp(enum spectre_v2_mitigation mode)
+{

Quoted 7 lines · KP Singh

+++ b/arch/x86/kernel/cpu/bugs.c

arch/x86/kernel/cpu/bugs.c +2 -0

@@ -1132,6 +1132,19 @@ static inline bool spectre_v2_in_ibrs_mode(enum spectre_v2_mitigation mode)
 	       mode == SPECTRE_V2_EIBRS_LFENCE;
 }
 
+static inline bool spectre_v2_user_no_stibp(enum spectre_v2_mitigation mode)
+{

arch/x86/kernel/cpu/bugs.c static inline bool spectre_v2_in_ibrs_mode(enum spectre_v2_mitigation mode) · line 1146

Quoted 10 lines

arch/x86/kernel/cpu/bugs.c +10 -0

+	/* When IBRS or enhanced IBRS is enabled, STIBP is not needed.
+	 *
+	 * However, With KERNEL_IBRS, the IBRS bit is cleared on return
+	 * to user and the user-mode code needs to be able to enable protection
+	 * from cross-thread training, either by always enabling STIBP or
+	 * by enabling it via prctl.
+	 */
+	return (spectre_v2_in_ibrs_mode(mode) &&
+		!cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS));
+}

The comments and code confused me, they both seem to imply some distinction between IBRS and KERNEL_IBRS, but in the kernel those are functionally the same thing. e.g., the kernel doesn't have a user IBRS mode.

And, unless I'm missing some subtlety here, it seems to be a convoluted way of saying that eIBRS doesn't need STIBP in user space.

It would be simpler to just call it spectre_v2_in_eibrs_mode().

static inline bool spectre_v2_in_eibrs_mode(enum spectre_v2_mitigation mode)

{
	return mode == SPECTRE_V2_EIBRS ||
	       mode == SPECTRE_V2_EIBRS_RETPOLINE ||
	       mode == SPECTRE_V2_EIBRS_LFENCE;
}

And then spectre_v2_in_ibrs_mode() could be changed to call that:

static inline bool spectre_v2_in_eibrs_mode(enum spectre_v2_mitigation mode)

{
	return spectre_v2_in_eibrs_mode(mode) || mode == SPECTRE_V2_IBRS;
}

arch/x86/kernel/cpu/bugs.c static void __init spectre_v2_select_mitigation(void) · line 1507

Quoted 5 lines

arch/x86/kernel/cpu/bugs.c +1 -0

@@ -1496,6 +1504,7 @@ static void __init spectre_v2_select_mitigation(void)
 		break;
 
 	case SPECTRE_V2_IBRS:
+		pr_err("enabling KERNEL_IBRS");

Why?

arch/x86/kernel/cpu/bugs.c static ssize_t mmio_stale_data_show_state(char *buf) · line 2340

Quoted 10 lines

arch/x86/kernel/cpu/bugs.c +1 -1

 		setup_force_cpu_cap(X86_FEATURE_KERNEL_IBRS);
 		if (boot_cpu_has(X86_FEATURE_IBRS_ENHANCED))
 			pr_warn(SPECTRE_V2_IBRS_PERF_MSG);

@@ -2327,7 +2336,7 @@ static ssize_t mmio_stale_data_show_state(char *buf)
 
 static char *stibp_state(void)
 {
-	if (spectre_v2_in_ibrs_mode(spectre_v2_enabled))
+	if (spectre_v2_user_no_stibp(spectre_v2_enabled))
 		return "";

This seems like old cruft, can we just remove this check altogether? In the eIBRS case, spectre_v2_user_stibp will already have its default of SPECTRE_V2_USER_NONE.

Signature · 1 line

Josh