[PATCH] x86/kexec: Fix bug with call depth tracking

[PATCH] x86/kexec: Fix bug with call depth tracking

[Borislav Petkov]

From: David Kaplan <david.kaplan@amd.com>

The call to cc_platform_has() triggers a fault and system crash if call depth
tracking is active because the GS segment has been reset by load_segments() and
GS_BASE is now 0 but call depth tracking uses per-CPU variables to operate.

Call cc_platform_has() earlier in the function when GS is still valid.

  [ bp: Massage. ]

Signed-off-by: David Kaplan <david.kaplan@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
---
 arch/x86/kernel/machine_kexec_64.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
index b180d8e497c3..90b1946c06e5 100644
--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -295,8 +295,15 @@ void machine_kexec_cleanup(struct kimage *image)
 void machine_kexec(struct kimage *image)
 {
 	unsigned long page_list[PAGES_NR];
-	void *control_page;
+        unsigned int host_mem_enc_active;
 	int save_ftrace_enabled;
+	void *control_page;
+
+	/*
+	 * This must be done before load_segments() since if call depth tracking
+	 * is used then GS must be valid to make any function calls.
+	 */
+	host_mem_enc_active = cc_platform_has(CC_ATTR_HOST_MEM_ENCRYPT);
 
 #ifdef CONFIG_KEXEC_JUMP
 	if (image->preserve_context)
@@ -358,7 +365,7 @@ void machine_kexec(struct kimage *image)
 				       (unsigned long)page_list,
 				       image->start,
 				       image->preserve_context,
-				       cc_platform_has(CC_ATTR_HOST_MEM_ENCRYPT));
+				       host_mem_enc_active);
 
 #ifdef CONFIG_KEXEC_JUMP
 	if (image->preserve_context)
-- 
2.43.0

Re: [PATCH] x86/kexec: Fix bug with call depth tracking

[Tom Lendacky]

On 6/3/24 03:30, Borislav Petkov wrote:
> From: David Kaplan <david.kaplan@amd.com>
> 
> The call to cc_platform_has() triggers a fault and system crash if call depth
> tracking is active because the GS segment has been reset by load_segments() and
> GS_BASE is now 0 but call depth tracking uses per-CPU variables to operate.
> 
> Call cc_platform_has() earlier in the function when GS is still valid.
> 
>    [ bp: Massage. ]
> 
> Signed-off-by: David Kaplan <david.kaplan@amd.com>
> Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>

Minor nit below, otherwise:

Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com>

Does it need a Fixes: tag for call depth tracking (before the change to 
cc_platform_has() it was a call to sme_active())?

> ---
>   arch/x86/kernel/machine_kexec_64.c | 11 +++++++++--
>   1 file changed, 9 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
> index b180d8e497c3..90b1946c06e5 100644
> --- a/arch/x86/kernel/machine_kexec_64.c
> +++ b/arch/x86/kernel/machine_kexec_64.c
> @@ -295,8 +295,15 @@ void machine_kexec_cleanup(struct kimage *image)
>   void machine_kexec(struct kimage *image)
>   {
>   	unsigned long page_list[PAGES_NR];
> -	void *control_page;
> +        unsigned int host_mem_enc_active;

Looks like spaces used here instead of tabs.

Thanks,
Tom

>   	int save_ftrace_enabled;
> +	void *control_page;
> +
> +	/*
> +	 * This must be done before load_segments() since if call depth tracking
> +	 * is used then GS must be valid to make any function calls.
> +	 */
> +	host_mem_enc_active = cc_platform_has(CC_ATTR_HOST_MEM_ENCRYPT);
>   
>   #ifdef CONFIG_KEXEC_JUMP
>   	if (image->preserve_context)
> @@ -358,7 +365,7 @@ void machine_kexec(struct kimage *image)
>   				       (unsigned long)page_list,
>   				       image->start,
>   				       image->preserve_context,
> -				       cc_platform_has(CC_ATTR_HOST_MEM_ENCRYPT));
> +				       host_mem_enc_active);
>   
>   #ifdef CONFIG_KEXEC_JUMP
>   	if (image->preserve_context)

Re: [PATCH] x86/kexec: Fix bug with call depth tracking

[Borislav Petkov]

On Mon, Jun 03, 2024 at 08:30:26AM -0500, Tom Lendacky wrote:
> Does it need a Fixes: tag for call depth tracking (before the change to
> cc_platform_has() it was a call to sme_active())?

Yeah, something like

Fixes: 5d8213864ade ("x86/retbleed: Add SKL return thunk")

I guess.

> Looks like spaces used here instead of tabs.

I was fixing those and forgot to refresh before sending - it is fixed
here locally.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

[tip: x86/urgent] x86/kexec: Fix bug with call depth tracking

[tip-bot2 for David Kaplan]

The following commit has been merged into the x86/urgent branch of tip:

Commit-ID:     93c1800b3799f17375989b0daf76497dd3e80922
Gitweb:        https://git.kernel.org/tip/93c1800b3799f17375989b0daf76497dd3e80922
Author:        David Kaplan <david.kaplan@amd.com>
AuthorDate:    Sun, 02 Jun 2024 13:19:09 -05:00
Committer:     Borislav Petkov (AMD) <bp@alien8.de>
CommitterDate: Mon, 03 Jun 2024 17:19:03 +02:00

x86/kexec: Fix bug with call depth tracking

The call to cc_platform_has() triggers a fault and system crash if call depth
tracking is active because the GS segment has been reset by load_segments() and
GS_BASE is now 0 but call depth tracking uses per-CPU variables to operate.

Call cc_platform_has() earlier in the function when GS is still valid.

  [ bp: Massage. ]

Fixes: 5d8213864ade ("x86/retbleed: Add SKL return thunk")
Signed-off-by: David Kaplan <david.kaplan@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com>
Cc: <stable@kernel.org>
Link: https://lore.kernel.org/r/20240603083036.637-1-bp@kernel.org
---
 arch/x86/kernel/machine_kexec_64.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
index b180d8e..cc0f7f7 100644
--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -295,8 +295,15 @@ void machine_kexec_cleanup(struct kimage *image)
 void machine_kexec(struct kimage *image)
 {
 	unsigned long page_list[PAGES_NR];
-	void *control_page;
+	unsigned int host_mem_enc_active;
 	int save_ftrace_enabled;
+	void *control_page;
+
+	/*
+	 * This must be done before load_segments() since if call depth tracking
+	 * is used then GS must be valid to make any function calls.
+	 */
+	host_mem_enc_active = cc_platform_has(CC_ATTR_HOST_MEM_ENCRYPT);
 
 #ifdef CONFIG_KEXEC_JUMP
 	if (image->preserve_context)
@@ -358,7 +365,7 @@ void machine_kexec(struct kimage *image)
 				       (unsigned long)page_list,
 				       image->start,
 				       image->preserve_context,
-				       cc_platform_has(CC_ATTR_HOST_MEM_ENCRYPT));
+				       host_mem_enc_active);
 
 #ifdef CONFIG_KEXEC_JUMP
 	if (image->preserve_context)