[PATCH v7 43/72] x86/sev-es: Adjust #VC IST Stack on entering NMI handler

[Joerg Roedel <joro@8bytes.org>]

From: Joerg Roedel <jroedel@suse.de>

When an NMI hits in the #VC handler entry code before it switched to
another stack, any subsequent #VC exception in the NMI code-path will
overwrite the interrupted #VC handlers stack.

Make sure this doesn't happen by  explicitly adjusting the #VC IST entry
in the NMI handler for the time in can cause #VC exceptions.

Signed-off-by: Joerg Roedel <jroedel@suse.de>
---
 arch/x86/include/asm/sev-es.h | 19 +++++++++++++
 arch/x86/kernel/nmi.c         |  9 ++++++
 arch/x86/kernel/sev-es.c      | 53 +++++++++++++++++++++++++++++++++++
 3 files changed, 81 insertions(+)

diff --git a/arch/x86/include/asm/sev-es.h b/arch/x86/include/asm/sev-es.h
index 9fbeedaa66ee..59176e8c6b81 100644
--- a/arch/x86/include/asm/sev-es.h
+++ b/arch/x86/include/asm/sev-es.h
@@ -78,4 +78,23 @@ extern void vc_no_ghcb(void);
 extern void vc_boot_ghcb(void);
 extern bool handle_vc_boot_ghcb(struct pt_regs *regs);
 
+#ifdef CONFIG_AMD_MEM_ENCRYPT
+extern struct static_key_false sev_es_enable_key;
+extern void __sev_es_ist_enter(struct pt_regs *regs);
+extern void __sev_es_ist_exit(void);
+static __always_inline void sev_es_ist_enter(struct pt_regs *regs)
+{
+	if (static_branch_unlikely(&sev_es_enable_key))
+		__sev_es_ist_enter(regs);
+}
+static __always_inline void sev_es_ist_exit(void)
+{
+	if (static_branch_unlikely(&sev_es_enable_key))
+		__sev_es_ist_exit();
+}
+#else
+static inline void sev_es_ist_enter(struct pt_regs *regs) { }
+static inline void sev_es_ist_exit(void) { }
+#endif
+
 #endif
diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c
index 4fc9954a9560..5859cec774a4 100644
--- a/arch/x86/kernel/nmi.c
+++ b/arch/x86/kernel/nmi.c
@@ -33,6 +33,7 @@
 #include <asm/reboot.h>
 #include <asm/cache.h>
 #include <asm/nospec-branch.h>
+#include <asm/sev-es.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/nmi.h>
@@ -488,6 +489,12 @@ DEFINE_IDTENTRY_RAW(exc_nmi)
 	this_cpu_write(nmi_cr2, read_cr2());
 nmi_restart:
 
+	/*
+	 * Needs to happen before DR7 is accessed, because the hypervisor can
+	 * intercept DR7 reads/writes, turings those into #VC exceptions.
+	 */
+	sev_es_ist_enter(regs);
+
 	this_cpu_write(nmi_dr7, local_db_save());
 
 	irq_state = idtentry_enter_nmi(regs);
@@ -501,6 +508,8 @@ DEFINE_IDTENTRY_RAW(exc_nmi)
 
 	local_db_restore(this_cpu_read(nmi_dr7));
 
+	sev_es_ist_exit();
+
 	if (unlikely(this_cpu_read(nmi_cr2) != read_cr2()))
 		write_cr2(this_cpu_read(nmi_cr2));
 	if (this_cpu_dec_return(nmi_state))
diff --git a/arch/x86/kernel/sev-es.c b/arch/x86/kernel/sev-es.c
index 5541788420ce..69c55f0fdf6a 100644
--- a/arch/x86/kernel/sev-es.c
+++ b/arch/x86/kernel/sev-es.c
@@ -51,6 +51,7 @@ struct sev_es_runtime_data {
 };
 
 static DEFINE_PER_CPU(struct sev_es_runtime_data*, runtime_data);
+DEFINE_STATIC_KEY_FALSE(sev_es_enable_key);
 
 static void __init setup_vc_stacks(int cpu)
 {
@@ -73,6 +74,55 @@ static void __init setup_vc_stacks(int cpu)
 	cea_set_pte((void *)vaddr, pa, PAGE_KERNEL);
 }
 
+static __always_inline bool on_vc_stack(unsigned long sp)
+{
+	return ((sp >= __this_cpu_ist_bottom_va(VC)) && (sp < __this_cpu_ist_top_va(VC)));
+}
+
+/*
+ * This function handles the case when an NM is raised in the #VC exception
+ * handler entry code. In this case the IST entry for #VC must be adjusted, so
+ * that any subsequent #VC exception will not overwrite the stack contents of the
+ * interrupted #VC handler.
+ *
+ * The IST entry is adjusted unconditionally so that it can be also be
+ * unconditionally adjusted back in sev_es_ist_exit(). Otherwise a nested
+ * sev_es_ist_exit() call may adjust back the IST entry too early.
+ */
+void noinstr __sev_es_ist_enter(struct pt_regs *regs)
+{
+	unsigned long old_ist, new_ist;
+
+	/* Read old IST entry */
+	old_ist = __this_cpu_read(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC]);
+
+	/* Make room on the IST stack */
+	if (on_vc_stack(regs->sp))
+		new_ist = ALIGN_DOWN(regs->sp, 8) - sizeof(old_ist);
+	else
+		new_ist = old_ist - sizeof(old_ist);
+
+	/* Store old IST entry */
+	*(unsigned long *)new_ist = old_ist;
+
+	/* Set new IST entry */
+	this_cpu_write(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC], new_ist);
+}
+
+void noinstr __sev_es_ist_exit(void)
+{
+	unsigned long ist;
+
+	/* Read IST entry */
+	ist = __this_cpu_read(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC]);
+
+	if (WARN_ON(ist == __this_cpu_ist_top_va(VC)))
+		return;
+
+	/* Read back old IST entry and write it to the TSS */
+	this_cpu_write(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC], *(unsigned long *)ist);
+}
+
 /* Needed in vc_early_forward_exception */
 void do_early_exception(struct pt_regs *regs, int trapnr);
 
@@ -277,6 +327,9 @@ void __init sev_es_init_vc_handling(void)
 	if (!sev_es_active())
 		return;
 
+	/* Enable SEV-ES special handling */
+	static_branch_enable(&sev_es_enable_key);
+
 	/* Initialize per-cpu GHCB pages */
 	for_each_possible_cpu(cpu) {
 		alloc_runtime_data(cpu);
-- 
2.28.0