Merge tag 'x86_urgent_for_v6.5_rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 fixes from Borislav Petkov:
 "Extraordinary embargoed times call for extraordinary measures. That's
  why this week's x86/urgent branch is larger than usual, containing all
  the known fallout fixes after the SRSO mitigation got merged.

  I know, it is a bit late in the game but everyone who has reported a
  bug stemming from the SRSO pile, has tested that branch and has
  confirmed that it fixes their bug.

  Also, I've run it on every possible hardware I have and it is looking
  good. It is running on this very machine while I'm typing, for 2 days
  now without an issue. Famous last words...

   - Use LEA ...%rsp instead of ADD %rsp in the Zen1/2 SRSO return
     sequence as latter clobbers flags which interferes with fastop
     emulation in KVM, leading to guests freezing during boot

   - A fix for the DIV(0) quotient data leak on Zen1 to clear the
     divider buffers at the right time

   - Disable the SRSO mitigation on unaffected configurations as it got
     enabled there unnecessarily

   - Change .text section name to fix CONFIG_LTO_CLANG builds

   - Improve the optprobe indirect jmp check so that certain
     configurations can still be able to use optprobes at all

   - A serious and good scrubbing of the untraining routines by PeterZ:
      - Add proper speculation stopping traps so that objtool is happy
      - Adjust objtool to handle the new thunks
      - Make the thunk pointer assignable to the different untraining
        sequences at runtime, thus avoiding the alternative at the
        return thunk. It simplifies the code a bit too.
      - Add a entry_untrain_ret() main entry point which selects the
        respective untraining sequence
      - Rename things so that they're more clear
      - Fix stack validation with FRAME_POINTER=y builds

   - Fix static call patching to handle when a JMP to the return thunk
     is the last insn on the very last module memory page

   - Add more documentation about what each untraining routine does and
     why"

* tag 'x86_urgent_for_v6.5_rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/srso: Correct the mitigation status when SMT is disabled
  x86/static_call: Fix __static_call_fixup()
  objtool/x86: Fixup frame-pointer vs rethunk
  x86/srso: Explain the untraining sequences a bit more
  x86/cpu/kvm: Provide UNTRAIN_RET_VM
  x86/cpu: Cleanup the untrain mess
  x86/cpu: Rename srso_(.*)_alias to srso_alias_\1
  x86/cpu: Rename original retbleed methods
  x86/cpu: Clean up SRSO return thunk mess
  x86/alternative: Make custom return thunk unconditional
  objtool/x86: Fix SRSO mess
  x86/cpu: Fix up srso_safe_ret() and __x86_return_thunk()
  x86/cpu: Fix __x86_return_thunk symbol type
  x86/retpoline,kprobes: Skip optprobe check for indirect jumps with retpolines and IBT
  x86/retpoline,kprobes: Fix position of thunk sections with CONFIG_LTO_CLANG
  x86/srso: Disable the mitigation on unaffected configurations
  x86/CPU/AMD: Fix the DIV(0) initial fix attempt
  x86/retpoline: Don't clobber RFLAGS during srso_safe_ret()

12 files changed, 189 insertions, 114 deletions

diff --git a/arch/x86/include/asm/entry-common.h b/arch/x86/include/asm/entry-common.h
index 117903881fe4..ce8f50192ae3 100644
--- a/arch/x86/include/asm/entry-common.h
+++ b/arch/x86/include/asm/entry-common.h
@@ -92,6 +92,7 @@ static inline void arch_exit_to_user_mode_prepare(struct pt_regs *regs,
 static __always_inline void arch_exit_to_user_mode(void)
 {
 	mds_user_clear_cpu_buffers();
+	amd_clear_divider();
 }
 #define arch_exit_to_user_mode arch_exit_to_user_mode
 
diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h
index 3faf044569a5..c55cc243592e 100644
--- a/arch/x86/include/asm/nospec-branch.h
+++ b/arch/x86/include/asm/nospec-branch.h
@@ -272,9 +272,9 @@
 .endm
 
 #ifdef CONFIG_CPU_UNRET_ENTRY
-#define CALL_ZEN_UNTRAIN_RET	"call zen_untrain_ret"
+#define CALL_UNTRAIN_RET	"call entry_untrain_ret"
 #else
-#define CALL_ZEN_UNTRAIN_RET	""
+#define CALL_UNTRAIN_RET	""
 #endif
 
 /*
@@ -282,7 +282,7 @@
  * return thunk isn't mapped into the userspace tables (then again, AMD
  * typically has NO_MELTDOWN).
  *
- * While zen_untrain_ret() doesn't clobber anything but requires stack,
+ * While retbleed_untrain_ret() doesn't clobber anything but requires stack,
  * entry_ibpb() will clobber AX, CX, DX.
  *
  * As such, this must be placed after every *SWITCH_TO_KERNEL_CR3 at a point
@@ -293,14 +293,20 @@
 	defined(CONFIG_CALL_DEPTH_TRACKING) || defined(CONFIG_CPU_SRSO)
 	VALIDATE_UNRET_END
 	ALTERNATIVE_3 "",						\
-		      CALL_ZEN_UNTRAIN_RET, X86_FEATURE_UNRET,		\
+		      CALL_UNTRAIN_RET, X86_FEATURE_UNRET,		\
 		      "call entry_ibpb", X86_FEATURE_ENTRY_IBPB,	\
 		      __stringify(RESET_CALL_DEPTH), X86_FEATURE_CALL_DEPTH
 #endif
+.endm
 
-#ifdef CONFIG_CPU_SRSO
-	ALTERNATIVE_2 "", "call srso_untrain_ret", X86_FEATURE_SRSO, \
-			  "call srso_untrain_ret_alias", X86_FEATURE_SRSO_ALIAS
+.macro UNTRAIN_RET_VM
+#if defined(CONFIG_CPU_UNRET_ENTRY) || defined(CONFIG_CPU_IBPB_ENTRY) || \
+	defined(CONFIG_CALL_DEPTH_TRACKING) || defined(CONFIG_CPU_SRSO)
+	VALIDATE_UNRET_END
+	ALTERNATIVE_3 "",						\
+		      CALL_UNTRAIN_RET, X86_FEATURE_UNRET,		\
+		      "call entry_ibpb", X86_FEATURE_IBPB_ON_VMEXIT,	\
+		      __stringify(RESET_CALL_DEPTH), X86_FEATURE_CALL_DEPTH
 #endif
 .endm
 
@@ -309,15 +315,10 @@
 	defined(CONFIG_CALL_DEPTH_TRACKING)
 	VALIDATE_UNRET_END
 	ALTERNATIVE_3 "",						\
-		      CALL_ZEN_UNTRAIN_RET, X86_FEATURE_UNRET,		\
+		      CALL_UNTRAIN_RET, X86_FEATURE_UNRET,		\
 		      "call entry_ibpb", X86_FEATURE_ENTRY_IBPB,	\
 		      __stringify(RESET_CALL_DEPTH_FROM_CALL), X86_FEATURE_CALL_DEPTH
 #endif
-
-#ifdef CONFIG_CPU_SRSO
-	ALTERNATIVE_2 "", "call srso_untrain_ret", X86_FEATURE_SRSO, \
-			  "call srso_untrain_ret_alias", X86_FEATURE_SRSO_ALIAS
-#endif
 .endm
 
 
@@ -341,17 +342,24 @@ extern retpoline_thunk_t __x86_indirect_thunk_array[];
 extern retpoline_thunk_t __x86_indirect_call_thunk_array[];
 extern retpoline_thunk_t __x86_indirect_jump_thunk_array[];
 
+#ifdef CONFIG_RETHUNK
 extern void __x86_return_thunk(void);
-extern void zen_untrain_ret(void);
+#else
+static inline void __x86_return_thunk(void) {}
+#endif
+
+extern void retbleed_return_thunk(void);
+extern void srso_return_thunk(void);
+extern void srso_alias_return_thunk(void);
+
+extern void retbleed_untrain_ret(void);
 extern void srso_untrain_ret(void);
-extern void srso_untrain_ret_alias(void);
+extern void srso_alias_untrain_ret(void);
+
+extern void entry_untrain_ret(void);
 extern void entry_ibpb(void);
 
-#ifdef CONFIG_CALL_THUNKS
 extern void (*x86_return_thunk)(void);
-#else
-#define x86_return_thunk	(&__x86_return_thunk)
-#endif
 
 #ifdef CONFIG_CALL_DEPTH_TRACKING
 extern void __x86_return_skl(void);
@@ -478,9 +486,6 @@ enum ssb_mitigation {
 	SPEC_STORE_BYPASS_SECCOMP,
 };
 
-extern char __indirect_thunk_start[];
-extern char __indirect_thunk_end[];
-
 static __always_inline
 void alternative_msr_write(unsigned int msr, u64 val, unsigned int feature)
 {
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 2dcf3a06af09..099d58d02a26 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -687,10 +687,6 @@ void __init_or_module noinline apply_retpolines(s32 *start, s32 *end)
 
 #ifdef CONFIG_RETHUNK
 
-#ifdef CONFIG_CALL_THUNKS
-void (*x86_return_thunk)(void) __ro_after_init = &__x86_return_thunk;
-#endif
-
 /*
  * Rewrite the compiler generated return thunk tail-calls.
  *
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 70f9d56f9305..7eca6a8abbb1 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -1329,3 +1329,4 @@ void noinstr amd_clear_divider(void)
 	asm volatile(ALTERNATIVE("", "div %2\n\t", X86_BUG_DIV0)
 		     :: "a" (0), "d" (0), "r" (1));
 }
+EXPORT_SYMBOL_GPL(amd_clear_divider);
diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
index d02f73c5339d..f081d26616ac 100644
--- a/arch/x86/kernel/cpu/bugs.c
+++ b/arch/x86/kernel/cpu/bugs.c
@@ -63,6 +63,8 @@ EXPORT_SYMBOL_GPL(x86_pred_cmd);
 
 static DEFINE_MUTEX(spec_ctrl_mutex);
 
+void (*x86_return_thunk)(void) __ro_after_init = &__x86_return_thunk;
+
 /* Update SPEC_CTRL MSR and its cached copy unconditionally */
 static void update_spec_ctrl(u64 val)
 {
@@ -165,6 +167,11 @@ void __init cpu_select_mitigations(void)
 	md_clear_select_mitigation();
 	srbds_select_mitigation();
 	l1d_flush_select_mitigation();
+
+	/*
+	 * srso_select_mitigation() depends and must run after
+	 * retbleed_select_mitigation().
+	 */
 	srso_select_mitigation();
 	gds_select_mitigation();
 }
@@ -1035,6 +1042,9 @@ do_cmd_auto:
 		setup_force_cpu_cap(X86_FEATURE_RETHUNK);
 		setup_force_cpu_cap(X86_FEATURE_UNRET);
 
+		if (IS_ENABLED(CONFIG_RETHUNK))
+			x86_return_thunk = retbleed_return_thunk;
+
 		if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD &&
 		    boot_cpu_data.x86_vendor != X86_VENDOR_HYGON)
 			pr_err(RETBLEED_UNTRAIN_MSG);
@@ -1044,6 +1054,7 @@ do_cmd_auto:
 
 	case RETBLEED_MITIGATION_IBPB:
 		setup_force_cpu_cap(X86_FEATURE_ENTRY_IBPB);
+		setup_force_cpu_cap(X86_FEATURE_IBPB_ON_VMEXIT);
 		mitigate_smt = true;
 		break;
 
@@ -2417,9 +2428,10 @@ static void __init srso_select_mitigation(void)
 		 * Zen1/2 with SMT off aren't vulnerable after the right
 		 * IBPB microcode has been applied.
 		 */
-		if ((boot_cpu_data.x86 < 0x19) &&
-		    (!cpu_smt_possible() || (cpu_smt_control == CPU_SMT_DISABLED)))
+		if (boot_cpu_data.x86 < 0x19 && !cpu_smt_possible()) {
 			setup_force_cpu_cap(X86_FEATURE_SRSO_NO);
+			return;
+		}
 	}
 
 	if (retbleed_mitigation == RETBLEED_MITIGATION_IBPB) {
@@ -2448,11 +2460,15 @@ static void __init srso_select_mitigation(void)
 			 * like ftrace, static_call, etc.
 			 */
 			setup_force_cpu_cap(X86_FEATURE_RETHUNK);
+			setup_force_cpu_cap(X86_FEATURE_UNRET);
 
-			if (boot_cpu_data.x86 == 0x19)
+			if (boot_cpu_data.x86 == 0x19) {
 				setup_force_cpu_cap(X86_FEATURE_SRSO_ALIAS);
-			else
+				x86_return_thunk = srso_alias_return_thunk;
+			} else {
 				setup_force_cpu_cap(X86_FEATURE_SRSO);
+				x86_return_thunk = srso_return_thunk;
+			}
 			srso_mitigation = SRSO_MITIGATION_SAFE_RET;
 		} else {
 			pr_err("WARNING: kernel not compiled with CPU_SRSO.\n");
@@ -2696,6 +2712,9 @@ static ssize_t retbleed_show_state(char *buf)
 
 static ssize_t srso_show_state(char *buf)
 {
+	if (boot_cpu_has(X86_FEATURE_SRSO_NO))
+		return sysfs_emit(buf, "Mitigation: SMT disabled\n");
+
 	return sysfs_emit(buf, "%s%s\n",
 			  srso_strings[srso_mitigation],
 			  (cpu_has_ibpb_brtype_microcode() ? "" : ", no microcode"));
diff --git a/arch/x86/kernel/kprobes/opt.c b/arch/x86/kernel/kprobes/opt.c
index 57b0037d0a99..517821b48391 100644
--- a/arch/x86/kernel/kprobes/opt.c
+++ b/arch/x86/kernel/kprobes/opt.c
@@ -226,7 +226,7 @@ static int copy_optimized_instructions(u8 *dest, u8 *src, u8 *real)
 }
 
 /* Check whether insn is indirect jump */
-static int __insn_is_indirect_jump(struct insn *insn)
+static int insn_is_indirect_jump(struct insn *insn)
 {
 	return ((insn->opcode.bytes[0] == 0xff &&
 		(X86_MODRM_REG(insn->modrm.value) & 6) == 4) || /* Jump */
@@ -260,26 +260,6 @@ static int insn_jump_into_range(struct insn *insn, unsigned long start, int len)
 	return (start <= target && target <= start + len);
 }
 
-static int insn_is_indirect_jump(struct insn *insn)
-{
-	int ret = __insn_is_indirect_jump(insn);
-
-#ifdef CONFIG_RETPOLINE
-	/*
-	 * Jump to x86_indirect_thunk_* is treated as an indirect jump.
-	 * Note that even with CONFIG_RETPOLINE=y, the kernel compiled with
-	 * older gcc may use indirect jump. So we add this check instead of
-	 * replace indirect-jump check.
-	 */
-	if (!ret)
-		ret = insn_jump_into_range(insn,
-				(unsigned long)__indirect_thunk_start,
-				(unsigned long)__indirect_thunk_end -
-				(unsigned long)__indirect_thunk_start);
-#endif
-	return ret;
-}
-
 /* Decode whole function to ensure any instructions don't jump into target */
 static int can_optimize(unsigned long paddr)
 {
@@ -334,9 +314,21 @@ static int can_optimize(unsigned long paddr)
 		/* Recover address */
 		insn.kaddr = (void *)addr;
 		insn.next_byte = (void *)(addr + insn.length);
-		/* Check any instructions don't jump into target */
-		if (insn_is_indirect_jump(&insn) ||
-		    insn_jump_into_range(&insn, paddr + INT3_INSN_SIZE,
+		/*
+		 * Check any instructions don't jump into target, indirectly or
+		 * directly.
+		 *
+		 * The indirect case is present to handle a code with jump
+		 * tables. When the kernel uses retpolines, the check should in
+		 * theory additionally look for jumps to indirect thunks.
+		 * However, the kernel built with retpolines or IBT has jump
+		 * tables disabled so the check can be skipped altogether.
+		 */
+		if (!IS_ENABLED(CONFIG_RETPOLINE) &&
+		    !IS_ENABLED(CONFIG_X86_KERNEL_IBT) &&
+		    insn_is_indirect_jump(&insn))
+			return 0;
+		if (insn_jump_into_range(&insn, paddr + INT3_INSN_SIZE,
 					 DISP32_SIZE))
 			return 0;
 		addr += insn.length;
diff --git a/arch/x86/kernel/static_call.c b/arch/x86/kernel/static_call.c
index b70670a98597..77a9316da435 100644
--- a/arch/x86/kernel/static_call.c
+++ b/arch/x86/kernel/static_call.c
@@ -186,6 +186,19 @@ EXPORT_SYMBOL_GPL(arch_static_call_transform);
  */
 bool __static_call_fixup(void *tramp, u8 op, void *dest)
 {
+	unsigned long addr = (unsigned long)tramp;
+	/*
+	 * Not all .return_sites are a static_call trampoline (most are not).
+	 * Check if the 3 bytes after the return are still kernel text, if not,
+	 * then this definitely is not a trampoline and we need not worry
+	 * further.
+	 *
+	 * This avoids the memcmp() below tripping over pagefaults etc..
+	 */
+	if (((addr >> PAGE_SHIFT) != ((addr + 7) >> PAGE_SHIFT)) &&
+	    !kernel_text_address(addr + 7))
+		return false;
+
 	if (memcmp(tramp+5, tramp_ud, 3)) {
 		/* Not a trampoline site, not our problem. */
 		return false;
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 1885326a8f65..4a817d20ce3b 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -206,8 +206,6 @@ DEFINE_IDTENTRY(exc_divide_error)
 {
 	do_error_trap(regs, 0, "divide error", X86_TRAP_DE, SIGFPE,
 		      FPE_INTDIV, error_get_trap_addr(regs));
-
-	amd_clear_divider();
 }
 
 DEFINE_IDTENTRY(exc_overflow)
diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
index ef06211bae4c..83d41c2601d7 100644
--- a/arch/x86/kernel/vmlinux.lds.S
+++ b/arch/x86/kernel/vmlinux.lds.S
@@ -133,27 +133,25 @@ SECTIONS
 		KPROBES_TEXT
 		SOFTIRQENTRY_TEXT
 #ifdef CONFIG_RETPOLINE
-		__indirect_thunk_start = .;
-		*(.text.__x86.indirect_thunk)
-		*(.text.__x86.return_thunk)
-		__indirect_thunk_end = .;
+		*(.text..__x86.indirect_thunk)
+		*(.text..__x86.return_thunk)
 #endif
 		STATIC_CALL_TEXT
 
 		ALIGN_ENTRY_TEXT_BEGIN
 #ifdef CONFIG_CPU_SRSO
-		*(.text.__x86.rethunk_untrain)
+		*(.text..__x86.rethunk_untrain)
 #endif
 
 		ENTRY_TEXT
 
 #ifdef CONFIG_CPU_SRSO
 		/*
-		 * See the comment above srso_untrain_ret_alias()'s
+		 * See the comment above srso_alias_untrain_ret()'s
 		 * definition.
 		 */
-		. = srso_untrain_ret_alias | (1 << 2) | (1 << 8) | (1 << 14) | (1 << 20);
-		*(.text.__x86.rethunk_safe)
+		. = srso_alias_untrain_ret | (1 << 2) | (1 << 8) | (1 << 14) | (1 << 20);
+		*(.text..__x86.rethunk_safe)
 #endif
 		ALIGN_ENTRY_TEXT_END
 		*(.gnu.warning)
@@ -523,7 +521,7 @@ INIT_PER_CPU(irq_stack_backing_store);
 #endif
 
 #ifdef CONFIG_RETHUNK
-. = ASSERT((__ret & 0x3f) == 0, "__ret not cacheline-aligned");
+. = ASSERT((retbleed_return_thunk & 0x3f) == 0, "retbleed_return_thunk not cacheline-aligned");
 . = ASSERT((srso_safe_ret & 0x3f) == 0, "srso_safe_ret not cacheline-aligned");
 #endif
 
@@ -538,8 +536,8 @@ INIT_PER_CPU(irq_stack_backing_store);
  * Instead do: (A | B) - (A & B) in order to compute the XOR
  * of the two function addresses:
  */
-. = ASSERT(((ABSOLUTE(srso_untrain_ret_alias) | srso_safe_ret_alias) -
-		(ABSOLUTE(srso_untrain_ret_alias) & srso_safe_ret_alias)) == ((1 << 2) | (1 << 8) | (1 << 14) | (1 << 20)),
+. = ASSERT(((ABSOLUTE(srso_alias_untrain_ret) | srso_alias_safe_ret) -
+		(ABSOLUTE(srso_alias_untrain_ret) & srso_alias_safe_ret)) == ((1 << 2) | (1 << 8) | (1 << 14) | (1 << 20)),
 		"SRSO function pair won't alias");
 #endif
 
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 03e852dedcc1..d4bfdc607fe7 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -4006,6 +4006,8 @@ static noinstr void svm_vcpu_enter_exit(struct kvm_vcpu *vcpu, bool spec_ctrl_in
 
 	guest_state_enter_irqoff();
 
+	amd_clear_divider();
+
 	if (sev_es_guest(vcpu->kvm))
 		__svm_sev_es_vcpu_run(svm, spec_ctrl_intercepted);
 	else
diff --git a/arch/x86/kvm/svm/vmenter.S b/arch/x86/kvm/svm/vmenter.S
index 265452fc9ebe..ef2ebabb059c 100644
--- a/arch/x86/kvm/svm/vmenter.S
+++ b/arch/x86/kvm/svm/vmenter.S
@@ -222,10 +222,7 @@ SYM_FUNC_START(__svm_vcpu_run)
 	 * because interrupt handlers won't sanitize 'ret' if the return is
 	 * from the kernel.
 	 */
-	UNTRAIN_RET
-
-	/* SRSO */
-	ALTERNATIVE "", "call entry_ibpb", X86_FEATURE_IBPB_ON_VMEXIT
+	UNTRAIN_RET_VM
 
 	/*
 	 * Clear all general purpose registers except RSP and RAX to prevent
@@ -362,7 +359,7 @@ SYM_FUNC_START(__svm_sev_es_vcpu_run)
 	 * because interrupt handlers won't sanitize RET if the return is
 	 * from the kernel.
 	 */
-	UNTRAIN_RET
+	UNTRAIN_RET_VM
 
 	/* "Pop" @spec_ctrl_intercepted.  */
 	pop %_ASM_BX
diff --git a/arch/x86/lib/retpoline.S b/arch/x86/lib/retpoline.S
index 2cff585f22f2..cd86aeb5fdd3 100644
--- a/arch/x86/lib/retpoline.S
+++ b/arch/x86/lib/retpoline.S
@@ -13,7 +13,7 @@
 #include <asm/frame.h>
 #include <asm/nops.h>
 
-	.section .text.__x86.indirect_thunk
+	.section .text..__x86.indirect_thunk
 
 
 .macro POLINE reg
@@ -133,75 +133,106 @@ SYM_CODE_END(__x86_indirect_jump_thunk_array)
 #ifdef CONFIG_RETHUNK
 
 /*
- * srso_untrain_ret_alias() and srso_safe_ret_alias() are placed at
+ * srso_alias_untrain_ret() and srso_alias_safe_ret() are placed at
  * special addresses:
  *
- * - srso_untrain_ret_alias() is 2M aligned
- * - srso_safe_ret_alias() is also in the same 2M page but bits 2, 8, 14
+ * - srso_alias_untrain_ret() is 2M aligned
+ * - srso_alias_safe_ret() is also in the same 2M page but bits 2, 8, 14
  * and 20 in its virtual address are set (while those bits in the
- * srso_untrain_ret_alias() function are cleared).
+ * srso_alias_untrain_ret() function are cleared).
  *
  * This guarantees that those two addresses will alias in the branch
  * target buffer of Zen3/4 generations, leading to any potential
  * poisoned entries at that BTB slot to get evicted.
  *
- * As a result, srso_safe_ret_alias() becomes a safe return.
+ * As a result, srso_alias_safe_ret() becomes a safe return.
  */
 #ifdef CONFIG_CPU_SRSO
-	.section .text.__x86.rethunk_untrain
+	.section .text..__x86.rethunk_untrain
 
-SYM_START(srso_untrain_ret_alias, SYM_L_GLOBAL, SYM_A_NONE)
+SYM_START(srso_alias_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
+	UNWIND_HINT_FUNC
 	ANNOTATE_NOENDBR
 	ASM_NOP2
 	lfence
-	jmp __x86_return_thunk
-SYM_FUNC_END(srso_untrain_ret_alias)
-__EXPORT_THUNK(srso_untrain_ret_alias)
-
-	.section .text.__x86.rethunk_safe
+	jmp srso_alias_return_thunk
+SYM_FUNC_END(srso_alias_untrain_ret)
+__EXPORT_THUNK(srso_alias_untrain_ret)
+
+	.section .text..__x86.rethunk_safe
+#else
+/* dummy definition for alternatives */
+SYM_START(srso_alias_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
+	ANNOTATE_UNRET_SAFE
+	ret
+	int3
+SYM_FUNC_END(srso_alias_untrain_ret)
 #endif
 
-/* Needs a definition for the __x86_return_thunk alternative below. */
-SYM_START(srso_safe_ret_alias, SYM_L_GLOBAL, SYM_A_NONE)
-#ifdef CONFIG_CPU_SRSO
-	add $8, %_ASM_SP
+SYM_START(srso_alias_safe_ret, SYM_L_GLOBAL, SYM_A_NONE)
+	lea 8(%_ASM_SP), %_ASM_SP
 	UNWIND_HINT_FUNC
-#endif
 	ANNOTATE_UNRET_SAFE
 	ret
 	int3
-SYM_FUNC_END(srso_safe_ret_alias)
+SYM_FUNC_END(srso_alias_safe_ret)
 
-	.section .text.__x86.return_thunk
+	.section .text..__x86.return_thunk
+
+SYM_CODE_START(srso_alias_return_thunk)
+	UNWIND_HINT_FUNC
+	ANNOTATE_NOENDBR
+	call srso_alias_safe_ret
+	ud2
+SYM_CODE_END(srso_alias_return_thunk)
+
+/*
+ * Some generic notes on the untraining sequences:
+ *
+ * They are interchangeable when it comes to flushing potentially wrong
+ * RET predictions from the BTB.
+ *
+ * The SRSO Zen1/2 (MOVABS) untraining sequence is longer than the
+ * Retbleed sequence because the return sequence done there
+ * (srso_safe_ret()) is longer and the return sequence must fully nest
+ * (end before) the untraining sequence. Therefore, the untraining
+ * sequence must fully overlap the return sequence.
+ *
+ * Regarding alignment - the instructions which need to be untrained,
+ * must all start at a cacheline boundary for Zen1/2 generations. That
+ * is, instruction sequences starting at srso_safe_ret() and
+ * the respective instruction sequences at retbleed_return_thunk()
+ * must start at a cacheline boundary.
+ */
 
 /*
  * Safety details here pertain to the AMD Zen{1,2} microarchitecture:
- * 1) The RET at __x86_return_thunk must be on a 64 byte boundary, for
+ * 1) The RET at retbleed_return_thunk must be on a 64 byte boundary, for
  *    alignment within the BTB.
- * 2) The instruction at zen_untrain_ret must contain, and not
+ * 2) The instruction at retbleed_untrain_ret must contain, and not
  *    end with, the 0xc3 byte of the RET.
  * 3) STIBP must be enabled, or SMT disabled, to prevent the sibling thread
  *    from re-poisioning the BTB prediction.
  */
 	.align 64
-	.skip 64 - (__ret - zen_untrain_ret), 0xcc
-SYM_START(zen_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
+	.skip 64 - (retbleed_return_thunk - retbleed_untrain_ret), 0xcc
+SYM_START(retbleed_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
 	ANNOTATE_NOENDBR
 	/*
-	 * As executed from zen_untrain_ret, this is:
+	 * As executed from retbleed_untrain_ret, this is:
 	 *
 	 *   TEST $0xcc, %bl
 	 *   LFENCE
-	 *   JMP __x86_return_thunk
+	 *   JMP retbleed_return_thunk
 	 *
 	 * Executing the TEST instruction has a side effect of evicting any BTB
 	 * prediction (potentially attacker controlled) attached to the RET, as
-	 * __x86_return_thunk + 1 isn't an instruction boundary at the moment.
+	 * retbleed_return_thunk + 1 isn't an instruction boundary at the moment.
 	 */
 	.byte	0xf6
 
 	/*
-	 * As executed from __x86_return_thunk, this is a plain RET.
+	 * As executed from retbleed_return_thunk, this is a plain RET.
 	 *
 	 * As part of the TEST above, RET is the ModRM byte, and INT3 the imm8.
 	 *
@@ -213,13 +244,13 @@ SYM_START(zen_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
 	 * With SMT enabled and STIBP active, a sibling thread cannot poison
 	 * RET's prediction to a type of its choice, but can evict the
 	 * prediction due to competitive sharing. If the prediction is
-	 * evicted, __x86_return_thunk will suffer Straight Line Speculation
+	 * evicted, retbleed_return_thunk will suffer Straight Line Speculation
 	 * which will be contained safely by the INT3.
 	 */
-SYM_INNER_LABEL(__ret, SYM_L_GLOBAL)
+SYM_INNER_LABEL(retbleed_return_thunk, SYM_L_GLOBAL)
 	ret
 	int3
-SYM_CODE_END(__ret)
+SYM_CODE_END(retbleed_return_thunk)
 
 	/*
 	 * Ensure the TEST decoding / BTB invalidation is complete.
@@ -230,16 +261,16 @@ SYM_CODE_END(__ret)
 	 * Jump back and execute the RET in the middle of the TEST instruction.
 	 * INT3 is for SLS protection.
 	 */
-	jmp __ret
+	jmp retbleed_return_thunk
 	int3
-SYM_FUNC_END(zen_untrain_ret)
-__EXPORT_THUNK(zen_untrain_ret)
+SYM_FUNC_END(retbleed_untrain_ret)
+__EXPORT_THUNK(retbleed_untrain_ret)
 
 /*
- * SRSO untraining sequence for Zen1/2, similar to zen_untrain_ret()
+ * SRSO untraining sequence for Zen1/2, similar to retbleed_untrain_ret()
  * above. On kernel entry, srso_untrain_ret() is executed which is a
  *
- * movabs $0xccccccc308c48348,%rax
+ * movabs $0xccccc30824648d48,%rax
  *
  * and when the return thunk executes the inner label srso_safe_ret()
  * later, it is a stack manipulation and a RET which is mispredicted and
@@ -251,22 +282,44 @@ SYM_START(srso_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
 	ANNOTATE_NOENDBR
 	.byte 0x48, 0xb8
 
+/*
+ * This forces the function return instruction to speculate into a trap
+ * (UD2 in srso_return_thunk() below).  This RET will then mispredict
+ * and execution will continue at the return site read from the top of
+ * the stack.
+ */
 SYM_INNER_LABEL(srso_safe_ret, SYM_L_GLOBAL)
-	add $8, %_ASM_SP
+	lea 8(%_ASM_SP), %_ASM_SP
 	ret
 	int3
 	int3
-	int3
+	/* end of movabs */
 	lfence
 	call srso_safe_ret
-	int3
+	ud2
 SYM_CODE_END(srso_safe_ret)
 SYM_FUNC_END(srso_untrain_ret)
 __EXPORT_THUNK(srso_untrain_ret)
 
-SYM_FUNC_START(__x86_return_thunk)
-	ALTERNATIVE_2 "jmp __ret", "call srso_safe_ret", X86_FEATURE_SRSO, \
-			"call srso_safe_ret_alias", X86_FEATURE_SRSO_ALIAS
+SYM_CODE_START(srso_return_thunk)
+	UNWIND_HINT_FUNC
+	ANNOTATE_NOENDBR
+	call srso_safe_ret
+	ud2
+SYM_CODE_END(srso_return_thunk)
+
+SYM_FUNC_START(entry_untrain_ret)
+	ALTERNATIVE_2 "jmp retbleed_untrain_ret", \
+		      "jmp srso_untrain_ret", X86_FEATURE_SRSO, \
+		      "jmp srso_alias_untrain_ret", X86_FEATURE_SRSO_ALIAS
+SYM_FUNC_END(entry_untrain_ret)
+__EXPORT_THUNK(entry_untrain_ret)
+
+SYM_CODE_START(__x86_return_thunk)
+	UNWIND_HINT_FUNC
+	ANNOTATE_NOENDBR
+	ANNOTATE_UNRET_SAFE
+	ret
 	int3
 SYM_CODE_END(__x86_return_thunk)
 EXPORT_SYMBOL(__x86_return_thunk)