FAILED: patch "[PATCH] x86/sev: Make vTPM SVSM calls preemption-safe" failed to apply to 6.18-stable tree

PATCH New [email protected] 2 participants 6 messages Sep 29

Subsystems (2)

Stable branch x86

lore

Share & tools

lore

[email protected]

author maintainer

2026-09-29 06:40 UTC #0 lore raw

The patch below does not apply to the 6.18-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.18.y
git checkout FETCH_HEAD
git cherry-pick -x 6c43c72748fffd29dec15cd1f31e9a32949bc437
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026092939-ecard-twirl-6094@gregkh' --subject-prefix 'PATCH 6.18.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 6c43c72748fffd29dec15cd1f31e9a32949bc437 Mon Sep 17 00:00:00 2001
From: Melody Wang <[email protected]>
Date: Mon, 14 Sep 2026 00:57:17 +0000
Subject: [PATCH] x86/sev: Make vTPM SVSM calls preemption-safe

Two functions in the SVSM vTPM guest implementation do not disable preemption when fetching the SVSM Calling Area Address (CAA).

The SVSM CAA is a per-CPU structure. When a thread is preempted and migrated to a different CPU after fetching the per-CPU CAA, the SVSM call will execute on the new CPU with the original CPU's CAA. Which is wrong.

Move the CAA fetching operation inside svsm_perform_call_protocol() which disables interrupts around the SVSM call and thus runs preemption-safe.

Fixes 770de678bc28 ("x86/sev: Add SVSM vTPM probe/send_command functions")

Signed-off-by Melody Wang <[email protected]>

Signed-off-by Borislav Petkov (AMD) <[email protected]>

Reviewed-by Stefano Garzarella <[email protected]>

Cc [email protected]

Link https://patch.msgid.link/a5bc0d4a2c462a0089109e145c21626b244b2ff0.1789345277.git.huibo.wang@amd.com open here

arch/x86/coco/sev/svsm.c +8 -2

index 916d62cd17dc..2d493d55ff2e 100644
@@ -74,6 +74,14 @@ int svsm_perform_call_protocol(struct svsm_call *call)
 
 	flags = native_local_irq_save();
 
+	/*
+	 * 'caa' is a per-CPU variable. To avoid using a stale or incorrect
+	 * 'caa' if the task is preempted or migrated to another CPU after it
+	 * is fetched, always fetch 'caa' and then issue the SVSM call with
+	 * interrupts disabled. This ensures the correct 'caa' is used.
+	 */
+	call->caa = svsm_get_caa();
+
 	ghcb = __sev_get_ghcb(&state);
 
 	do {

@@ -321,7 +329,6 @@ int snp_svsm_vtpm_send_command(u8 *buffer)
 {
 	struct svsm_call call = {};
 
-	call.caa = svsm_get_caa();
 	call.rax = SVSM_VTPM_CALL(SVSM_VTPM_CMD);
 	call.rcx = __pa(buffer);

@@ -345,7 +352,6 @@ bool snp_svsm_vtpm_probe(void)
 	if (!snp_vmpl)
 		return false;
 
-	call.caa = svsm_get_caa();
 	call.rax = SVSM_VTPM_CALL(SVSM_VTPM_QUERY);
 
 	if (svsm_perform_call_protocol(&call))

Sasha Levin

maintainer

2026-09-29 13:31 UTC #1 lore raw

PATCH 1/5 x86/sev: Move the internal header

From: "Borislav Petkov (AMD)" <[email protected]>

[ Upstream commit c1e8980fabf5d0106992a430284fac28bba053a6 ]

Move the internal header out of the usual include/asm/ include path because having an "internal" header there doesn't really make it internal - quite the opposite - that's the normal arch include path.

So move where it belongs and make it really internal.

No functional changes.

Reviewed-by Tom Lendacky <[email protected]>

Signed-off-by Borislav Petkov (AMD) <[email protected]>

Link https://lore.kernel.org/r/20251204145716.GDaTGhTEHNOtSdTkEe@fat_crate.local open here

Stable-dep-of: 6c43c72748ff ("x86/sev: Make vTPM SVSM calls preemption-safe")

Signed-off-by Sasha Levin <[email protected]>

---
 arch/x86/boot/startup/sev-startup.c                          | 3 ++-
 arch/x86/coco/sev/core.c                                     | 3 ++-
 arch/x86/{include/asm/sev-internal.h => coco/sev/internal.h} | 0
 arch/x86/coco/sev/noinstr.c                                  | 3 ++-
 arch/x86/coco/sev/vc-handle.c                                | 3 ++-
 5 files changed, 8 insertions(+), 4 deletions(-)
 rename arch/x86/{include/asm/sev-internal.h => coco/sev/internal.h} (100%)

arch/x86/boot/startup/sev-startup.c +2 -1

index 09725428d3e65..789e99d38d17b 100644
@@ -27,7 +27,6 @@
 #include <asm/cpu_entry_area.h>
 #include <asm/stacktrace.h>
 #include <asm/sev.h>
-#include <asm/sev-internal.h>
 #include <asm/insn-eval.h>
 #include <asm/fpu/xcr.h>
 #include <asm/processor.h>

@@ -41,6 +40,8 @@
 #include <asm/cpuid/api.h>
 #include <asm/cmdline.h>
 
+#include "../../coco/sev/internal.h"
+
 /* Include code shared with pre-decompression boot stage */
 #include "sev-shared.c"

arch/x86/coco/sev/core.c +2 -1

index d20e9cc065a87..19adb17bd09d1 100644
@@ -31,7 +31,6 @@
 #include <asm/cpu_entry_area.h>
 #include <asm/stacktrace.h>
 #include <asm/sev.h>
-#include <asm/sev-internal.h>
 #include <asm/insn-eval.h>
 #include <asm/fpu/xcr.h>
 #include <asm/processor.h>

@@ -46,6 +45,8 @@
 #include <asm/cmdline.h>
 #include <asm/msr.h>
 
+#include "internal.h"
+
 /* Bitmap of SEV features supported by the hypervisor */
 u64 sev_hv_features __ro_after_init;
 SYM_PIC_ALIAS(sev_hv_features);

arch/x86/coco/sev/internal.h renamed from arch/x86/include/asm/sev-internal.h +0 -0

similarity index 100%
rename from arch/x86/include/asm/sev-internal.h
rename to arch/x86/coco/sev/internal.h

arch/x86/coco/sev/noinstr.c +2 -1

index 422ce72f51d58..5afd663a1c217 100644
@@ -16,7 +16,8 @@
 #include <asm/msr.h>
 #include <asm/ptrace.h>
 #include <asm/sev.h>
-#include <asm/sev-internal.h>
+
+#include "internal.h"
 
 static __always_inline bool on_vc_stack(struct pt_regs *regs)
 {

arch/x86/coco/sev/vc-handle.c +2 -1

index 7fc136a353347..c54f2a032e392 100644
@@ -23,7 +23,6 @@
 #include <asm/init.h>
 #include <asm/stacktrace.h>
 #include <asm/sev.h>
-#include <asm/sev-internal.h>
 #include <asm/insn-eval.h>
 #include <asm/fpu/xcr.h>
 #include <asm/processor.h>

@@ -35,6 +34,8 @@
 #include <asm/apic.h>
 #include <asm/cpuid/api.h>
 
+#include "internal.h"
+
 static enum es_result vc_slow_virt_to_phys(struct ghcb *ghcb, struct es_em_ctxt *ctxt,
 					   unsigned long vaddr, phys_addr_t *paddr)
 {

Signature · 1 line

2.53.0

Sasha Levin

maintainer

2026-09-29 13:31 UTC #2 lore raw

PATCH 2/5 x86/sev: Add internal header guards

From: "Borislav Petkov (AMD)" <[email protected]>

[ Upstream commit f01c6489ad6ceb8d06eae0c5123fc6cf39276ff1 ]

All headers need guards ifdeffery.

Reviewed-by Tom Lendacky <[email protected]>

Signed-off-by Borislav Petkov (AMD) <[email protected]>

Link https://patch.msgid.link/[email protected] open here

Stable-dep-of: 6c43c72748ff ("x86/sev: Make vTPM SVSM calls preemption-safe")

Signed-off-by Sasha Levin <[email protected]>

---
 arch/x86/coco/sev/internal.h | 3 +++
 1 file changed, 3 insertions(+)

arch/x86/coco/sev/internal.h +3 -0

index c58c47c68ab6f..af991f1da0957 100644
@@ -1,4 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __X86_COCO_SEV_INTERNAL_H__
+#define __X86_COCO_SEV_INTERNAL_H__
 
 #define DR7_RESET_VALUE        0x400

@@ -85,3 +87,4 @@ enum es_result sev_es_ghcb_handle_msr(struct ghcb *ghcb, struct es_em_ctxt *ctxt
 u64 get_hv_features(void);
 
 const struct snp_cpuid_table *snp_cpuid_get_table(void);
+#endif /* __X86_COCO_SEV_INTERNAL_H__ */

Signature · 1 line

2.53.0

Sasha Levin

maintainer

2026-09-29 13:31 UTC #3 lore raw

PATCH 3/5 x86/sev: Carve out the SVSM code into a separate compilation unit

In reply to Sasha Levin

From: "Borislav Petkov (AMD)" <[email protected]>

[ Upstream commit e21279b73ef6c7d27237912c914f2db0c5a74786 ]

Move the SVSM-related machinery into a separate compilation unit in order to keep sev/core.c slim and "on-topic".

No functional changes.

Reviewed-by Tom Lendacky <[email protected]>

Signed-off-by Borislav Petkov (AMD) <[email protected]>

Link https://patch.msgid.link/[email protected] open here

Stable-dep-of: 6c43c72748ff ("x86/sev: Make vTPM SVSM calls preemption-safe")

Signed-off-by Sasha Levin <[email protected]>

---
 arch/x86/coco/sev/Makefile   |   2 +-
 arch/x86/coco/sev/core.c     | 377 -----------------------------------
 arch/x86/coco/sev/internal.h |  29 +++
 arch/x86/coco/sev/svsm.c     | 362 +++++++++++++++++++++++++++++++++
 4 files changed, 392 insertions(+), 378 deletions(-)
 create mode 100644 arch/x86/coco/sev/svsm.c

arch/x86/coco/sev/Makefile +1 -1

index b2e9ec2f69014..3111a2ee67221 100644
@@ -1,6 +1,6 @@
 # SPDX-License-Identifier: GPL-2.0
 
-obj-y += core.o noinstr.o vc-handle.o
+obj-y += core.o noinstr.o vc-handle.o svsm.o
 
 # Clang 14 and older may fail to respect __no_sanitize_undefined when inlining
 UBSAN_SANITIZE_noinstr.o	:= n

arch/x86/coco/sev/core.c +0 ~~-377~~

407 lines not on this page · raw message

arch/x86/coco/sev/internal.h +29 -0

index af991f1da0957..039326b5c799e 100644
@@ -66,6 +66,9 @@ extern u64 boot_svsm_caa_pa;
 
 enum es_result verify_exception_info(struct ghcb *ghcb, struct es_em_ctxt *ctxt);
 void vc_forward_exception(struct es_em_ctxt *ctxt);
+void svsm_pval_pages(struct snp_psc_desc *desc);
+int svsm_perform_call_protocol(struct svsm_call *call);
+bool snp_svsm_vtpm_probe(void);
 
 static inline u64 sev_es_rd_ghcb_msr(void)
 {

@@ -87,4 +90,30 @@ enum es_result sev_es_ghcb_handle_msr(struct ghcb *ghcb, struct es_em_ctxt *ctxt
 u64 get_hv_features(void);
 
 const struct snp_cpuid_table *snp_cpuid_get_table(void);
+
+static inline struct svsm_ca *svsm_get_caa(void)
+{
+	if (sev_cfg.use_cas)
+		return this_cpu_read(svsm_caa);
+	else
+		return rip_rel_ptr(&boot_svsm_ca_page);
+}
+
+static inline u64 svsm_get_caa_pa(void)
+{
+	if (sev_cfg.use_cas)
+		return this_cpu_read(svsm_caa_pa);
+	else
+		return boot_svsm_caa_pa;
+}
+
+static inline void __pval_terminate(u64 pfn, bool action, unsigned int page_size,
+				    int ret, u64 svsm_ret)
+{
+	WARN(1, "PVALIDATE failure: pfn: 0x%llx, action: %u, size: %u, ret: %d, svsm_ret: 0x%llx\n",
+	     pfn, action, page_size, ret, svsm_ret);
+
+	sev_es_terminate(SEV_TERM_SET_LINUX, GHCB_TERM_PVALIDATE);
+}
+
 #endif /* __X86_COCO_SEV_INTERNAL_H__ */

arch/x86/coco/sev/svsm.c new file +362 -0

362 lines not on this page · raw message

Signature · 1 line

2.53.0

Sasha Levin

maintainer

2026-09-29 13:31 UTC #4 lore raw

PATCH 4/5 x86/sev: Remove redundant ghcbs_initialized checks around __sev_{get,put}_ghcb()

In reply to Sasha Levin

From: Nikunj A Dadhania <[email protected]>

[ Upstream commit 9d8460a1c7a6b0f2dc6302e5d0f31d4e8c2a7913 ]

After

3645eb7e3915 ("x86/fred: Fix early boot failures on SEV-ES/SNP guests"),

__sev_{get,put}_ghcb() handle the early-boot GHCB fallback internally, making the ghcbs_initialized guards in __set_pages_state() and svsm_perform_call_protocol() redundant.

Remove them.

Also initialize state->ghcb to NULL in the early-boot path of __sev_get_ghcb() so that the ghcb_state is well-defined for all callers, even though __sev_put_ghcb() currently returns early before reading it.

No functional change intended.

Signed-off-by Nikunj A Dadhania <[email protected]>

Signed-off-by Borislav Petkov (AMD) <[email protected]>

Reviewed-by Tom Lendacky <[email protected]>

Link https://patch.msgid.link/[email protected] open here

Stable-dep-of: 6c43c72748ff ("x86/sev: Make vTPM SVSM calls preemption-safe")

Signed-off-by Sasha Levin <[email protected]>

---
 arch/x86/coco/sev/core.c    |  8 ++------
 arch/x86/coco/sev/noinstr.c |  4 +++-
 arch/x86/coco/sev/svsm.c    | 10 ++--------
 3 files changed, 7 insertions(+), 15 deletions(-)

arch/x86/coco/sev/core.c +2 -6

index 2415410e898ec..fefd7031be020 100644
@@ -367,17 +367,13 @@ static unsigned long __set_pages_state(struct snp_psc_desc *data, unsigned long
 
 	local_irq_save(flags);
 
-	if (sev_cfg.ghcbs_initialized)
-		ghcb = __sev_get_ghcb(&state);
-	else
-		ghcb = boot_ghcb;
+	ghcb = __sev_get_ghcb(&state);
 
 	/* Invoke the hypervisor to perform the page state changes */
 	if (!ghcb || vmgexit_psc(ghcb, data))
 		sev_es_terminate(SEV_TERM_SET_LINUX, GHCB_TERM_PSC);
 
-	if (sev_cfg.ghcbs_initialized)
-		__sev_put_ghcb(&state);
+	__sev_put_ghcb(&state);
 
 	local_irq_restore(flags);

arch/x86/coco/sev/noinstr.c +3 -1

index 5afd663a1c217..e1e03f12fc7bb 100644
@@ -121,8 +121,10 @@ noinstr struct ghcb *__sev_get_ghcb(struct ghcb_state *state)
 
 	WARN_ON(!irqs_disabled());
 
-	if (!sev_cfg.ghcbs_initialized)
+	if (!sev_cfg.ghcbs_initialized) {
+		state->ghcb = NULL;
 		return boot_ghcb;
+	}
 
 	data = this_cpu_read(runtime_data);
 	ghcb = &data->ghcb_page;

arch/x86/coco/sev/svsm.c +2 -8

index 2acf4a76afe7a..916d62cd17dc7 100644
@@ -74,20 +74,14 @@ int svsm_perform_call_protocol(struct svsm_call *call)
 
 	flags = native_local_irq_save();
 
-	if (sev_cfg.ghcbs_initialized)
-		ghcb = __sev_get_ghcb(&state);
-	else if (boot_ghcb)
-		ghcb = boot_ghcb;
-	else
-		ghcb = NULL;
+	ghcb = __sev_get_ghcb(&state);
 
 	do {
 		ret = ghcb ? svsm_perform_ghcb_protocol(ghcb, call)
 			   : __pi_svsm_perform_msr_protocol(call);
 	} while (ret == -EAGAIN);
 
-	if (sev_cfg.ghcbs_initialized)
-		__sev_put_ghcb(&state);
+	__sev_put_ghcb(&state);
 
 	native_local_irq_restore(flags);

Signature · 1 line

2.53.0

Sasha Levin

maintainer

2026-09-29 13:31 UTC #5 lore raw

PATCH 5/5 x86/sev: Make vTPM SVSM calls preemption-safe

In reply to Sasha Levin

From: Melody Wang <[email protected]>

[ Upstream commit 6c43c72748fffd29dec15cd1f31e9a32949bc437 ]

Two functions in the SVSM vTPM guest implementation do not disable preemption when fetching the SVSM Calling Area Address (CAA).

Move the CAA fetching operation inside svsm_perform_call_protocol() which disables interrupts around the SVSM call and thus runs preemption-safe.

Fixes 770de678bc28 ("x86/sev: Add SVSM vTPM probe/send_command functions")

Signed-off-by Melody Wang <[email protected]>

Signed-off-by Borislav Petkov (AMD) <[email protected]>

Reviewed-by Stefano Garzarella <[email protected]>

Cc [email protected]

Link https://patch.msgid.link/a5bc0d4a2c462a0089109e145c21626b244b2ff0.1789345277.git.huibo.wang@amd.com open here

Signed-off-by Sasha Levin <[email protected]>

---
 arch/x86/coco/sev/svsm.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

arch/x86/coco/sev/svsm.c +8 -2

index 916d62cd17dc7..2d493d55ff2e8 100644
@@ -74,6 +74,14 @@ int svsm_perform_call_protocol(struct svsm_call *call)
 
 	flags = native_local_irq_save();
 
+	/*
+	 * 'caa' is a per-CPU variable. To avoid using a stale or incorrect
+	 * 'caa' if the task is preempted or migrated to another CPU after it
+	 * is fetched, always fetch 'caa' and then issue the SVSM call with
+	 * interrupts disabled. This ensures the correct 'caa' is used.
+	 */
+	call->caa = svsm_get_caa();
+
 	ghcb = __sev_get_ghcb(&state);
 
 	do {

@@ -321,7 +329,6 @@ int snp_svsm_vtpm_send_command(u8 *buffer)
 {
 	struct svsm_call call = {};
 
-	call.caa = svsm_get_caa();
 	call.rax = SVSM_VTPM_CALL(SVSM_VTPM_CMD);
 	call.rcx = __pa(buffer);

@@ -345,7 +352,6 @@ bool snp_svsm_vtpm_probe(void)
 	if (!snp_vmpl)
 		return false;
 
-	call.caa = svsm_get_caa();
 	call.rax = SVSM_VTPM_CALL(SVSM_VTPM_QUERY);
 
 	if (svsm_perform_call_protocol(&call))

Signature · 1 line

2.53.0