Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To:     linux-kernel@vger.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        stable@vger.kernel.org, Andrew Cooper <Andrew.Cooper3@citrix.com>,
        "Peter Zijlstra (Intel)" <peterz@infradead.org>,
        Borislav Petkov <bp@suse.de>,
        Josh Poimboeuf <jpoimboe@kernel.org>,
        Thadeu Lima de Souza Cascardo <cascardo@canonical.com>,
        Ben Hutchings <ben@decadent.org.uk>
Subject: [PATCH 5.10 108/130] x86/bugs: Add retbleed=ibpb
Date:   Tue, 12 Jul 2022 20:39:14 +0200
Message-Id: <20220712183251.457354010@linuxfoundation.org>
In-Reply-To: <20220712183246.394947160@linuxfoundation.org>
References: <20220712183246.394947160@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Peter Zijlstra <peterz@infradead.org>

commit 3ebc170068885b6fc7bedda6c667bb2c4d533159 upstream.

jmp2ret mitigates the easy-to-attack case at relatively low overhead.
It mitigates the long speculation windows after a mispredicted RET, but
it does not mitigate the short speculation window from arbitrary
instruction boundaries.

On Zen2, there is a chicken bit which needs setting, which mitigates
"arbitrary instruction boundaries" down to just "basic block boundaries".

But there is no fix for the short speculation window on basic block
boundaries, other than to flush the entire BTB to evict all attacker
predictions.

On the spectrum of "fast & blurry" -> "safe", there is (on top of STIBP
or no-SMT):

  1) Nothing		System wide open
  2) jmp2ret		May stop a script kiddy
  3) jmp2ret+chickenbit  Raises the bar rather further
  4) IBPB		Only thing which can count as "safe".

Tentative numbers put IBPB-on-entry at a 2.5x hit on Zen2, and a 10x hit
on Zen1 according to lmbench.

  [ bp: Fixup feature bit comments, document option, 32-bit build fix. ]

Suggested-by: Andrew Cooper <Andrew.Cooper3@citrix.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
[bwh: Backported to 5.10: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 Documentation/admin-guide/kernel-parameters.txt |    3 +
 arch/x86/entry/Makefile                         |    2 -
 arch/x86/entry/entry.S                          |   22 ++++++++++++
 arch/x86/include/asm/cpufeatures.h              |    2 -
 arch/x86/include/asm/nospec-branch.h            |    8 +++-
 arch/x86/kernel/cpu/bugs.c                      |   43 ++++++++++++++++++------
 6 files changed, 67 insertions(+), 13 deletions(-)
 create mode 100644 arch/x86/entry/entry.S

--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -4666,6 +4666,9 @@
 				       disabling SMT if necessary for
 				       the full mitigation (only on Zen1
 				       and older without STIBP).
+			ibpb	     - mitigate short speculation windows on
+				       basic block boundaries too. Safe, highest
+				       perf impact.
 			unret        - force enable untrained return thunks,
 				       only effective on AMD f15h-f17h
 				       based systems.
--- a/arch/x86/entry/Makefile
+++ b/arch/x86/entry/Makefile
@@ -21,7 +21,7 @@ CFLAGS_syscall_64.o		+= $(call cc-option
 CFLAGS_syscall_32.o		+= $(call cc-option,-Wno-override-init,)
 CFLAGS_syscall_x32.o		+= $(call cc-option,-Wno-override-init,)
 
-obj-y				:= entry_$(BITS).o thunk_$(BITS).o syscall_$(BITS).o
+obj-y				:= entry.o entry_$(BITS).o thunk_$(BITS).o syscall_$(BITS).o
 obj-y				+= common.o
 
 obj-y				+= vdso/
--- /dev/null
+++ b/arch/x86/entry/entry.S
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Common place for both 32- and 64-bit entry routines.
+ */
+
+#include <linux/linkage.h>
+#include <asm/export.h>
+#include <asm/msr-index.h>
+
+.pushsection .noinstr.text, "ax"
+
+SYM_FUNC_START(entry_ibpb)
+	movl	$MSR_IA32_PRED_CMD, %ecx
+	movl	$PRED_CMD_IBPB, %eax
+	xorl	%edx, %edx
+	wrmsr
+	RET
+SYM_FUNC_END(entry_ibpb)
+/* For KVM */
+EXPORT_SYMBOL_GPL(entry_ibpb);
+
+.popsection
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -292,7 +292,7 @@
 #define X86_FEATURE_PER_THREAD_MBA	(11*32+ 7) /* "" Per-thread Memory Bandwidth Allocation */
 /* FREE!				(11*32+ 8) */
 /* FREE!				(11*32+ 9) */
-/* FREE!				(11*32+10) */
+#define X86_FEATURE_ENTRY_IBPB		(11*32+10) /* "" Issue an IBPB on kernel entry */
 /* FREE!				(11*32+11) */
 #define X86_FEATURE_RETPOLINE		(11*32+12) /* "" Generic Retpoline mitigation for Spectre variant 2 */
 #define X86_FEATURE_RETPOLINE_LFENCE	(11*32+13) /* "" Use LFENCE for Spectre variant 2 */
--- a/arch/x86/include/asm/nospec-branch.h
+++ b/arch/x86/include/asm/nospec-branch.h
@@ -123,14 +123,17 @@
  * return thunk isn't mapped into the userspace tables (then again, AMD
  * typically has NO_MELTDOWN).
  *
- * Doesn't clobber any registers but does require a stable stack.
+ * While zen_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
  * where we have a stack but before any RET instruction.
  */
 .macro UNTRAIN_RET
 #ifdef CONFIG_RETPOLINE
-	ALTERNATIVE "", "call zen_untrain_ret", X86_FEATURE_UNRET
+	ALTERNATIVE_2 "",						\
+	              "call zen_untrain_ret", X86_FEATURE_UNRET,	\
+		      "call entry_ibpb", X86_FEATURE_ENTRY_IBPB
 #endif
 .endm
 
@@ -144,6 +147,7 @@
 
 extern void __x86_return_thunk(void);
 extern void zen_untrain_ret(void);
+extern void entry_ibpb(void);
 
 #ifdef CONFIG_RETPOLINE
 
--- a/arch/x86/kernel/cpu/bugs.c
+++ b/arch/x86/kernel/cpu/bugs.c
@@ -761,6 +761,7 @@ static enum spectre_v2_mitigation spectr
 enum retbleed_mitigation {
 	RETBLEED_MITIGATION_NONE,
 	RETBLEED_MITIGATION_UNRET,
+	RETBLEED_MITIGATION_IBPB,
 	RETBLEED_MITIGATION_IBRS,
 	RETBLEED_MITIGATION_EIBRS,
 };
@@ -769,11 +770,13 @@ enum retbleed_mitigation_cmd {
 	RETBLEED_CMD_OFF,
 	RETBLEED_CMD_AUTO,
 	RETBLEED_CMD_UNRET,
+	RETBLEED_CMD_IBPB,
 };
 
 const char * const retbleed_strings[] = {
 	[RETBLEED_MITIGATION_NONE]	= "Vulnerable",
 	[RETBLEED_MITIGATION_UNRET]	= "Mitigation: untrained return thunk",
+	[RETBLEED_MITIGATION_IBPB]	= "Mitigation: IBPB",
 	[RETBLEED_MITIGATION_IBRS]	= "Mitigation: IBRS",
 	[RETBLEED_MITIGATION_EIBRS]	= "Mitigation: Enhanced IBRS",
 };
@@ -803,6 +806,8 @@ static int __init retbleed_parse_cmdline
 			retbleed_cmd = RETBLEED_CMD_AUTO;
 		} else if (!strcmp(str, "unret")) {
 			retbleed_cmd = RETBLEED_CMD_UNRET;
+		} else if (!strcmp(str, "ibpb")) {
+			retbleed_cmd = RETBLEED_CMD_IBPB;
 		} else if (!strcmp(str, "nosmt")) {
 			retbleed_nosmt = true;
 		} else {
@@ -817,11 +822,13 @@ static int __init retbleed_parse_cmdline
 early_param("retbleed", retbleed_parse_cmdline);
 
 #define RETBLEED_UNTRAIN_MSG "WARNING: BTB untrained return thunk mitigation is only effective on AMD/Hygon!\n"
-#define RETBLEED_COMPILER_MSG "WARNING: kernel not compiled with RETPOLINE or -mfunction-return capable compiler!\n"
+#define RETBLEED_COMPILER_MSG "WARNING: kernel not compiled with RETPOLINE or -mfunction-return capable compiler; falling back to IBPB!\n"
 #define RETBLEED_INTEL_MSG "WARNING: Spectre v2 mitigation leaves CPU vulnerable to RETBleed attacks, data leaks possible!\n"
 
 static void __init retbleed_select_mitigation(void)
 {
+	bool mitigate_smt = false;
+
 	if (!boot_cpu_has_bug(X86_BUG_RETBLEED) || cpu_mitigations_off())
 		return;
 
@@ -833,11 +840,21 @@ static void __init retbleed_select_mitig
 		retbleed_mitigation = RETBLEED_MITIGATION_UNRET;
 		break;
 
+	case RETBLEED_CMD_IBPB:
+		retbleed_mitigation = RETBLEED_MITIGATION_IBPB;
+		break;
+
 	case RETBLEED_CMD_AUTO:
 	default:
 		if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD ||
-		    boot_cpu_data.x86_vendor == X86_VENDOR_HYGON)
-			retbleed_mitigation = RETBLEED_MITIGATION_UNRET;
+		    boot_cpu_data.x86_vendor == X86_VENDOR_HYGON) {
+
+			if (IS_ENABLED(CONFIG_RETPOLINE) &&
+			    IS_ENABLED(CONFIG_CC_HAS_RETURN_THUNK))
+				retbleed_mitigation = RETBLEED_MITIGATION_UNRET;
+			else
+				retbleed_mitigation = RETBLEED_MITIGATION_IBPB;
+		}
 
 		/*
 		 * The Intel mitigation (IBRS) was already selected in
@@ -853,26 +870,34 @@ static void __init retbleed_select_mitig
 		if (!IS_ENABLED(CONFIG_RETPOLINE) ||
 		    !IS_ENABLED(CONFIG_CC_HAS_RETURN_THUNK)) {
 			pr_err(RETBLEED_COMPILER_MSG);
-			retbleed_mitigation = RETBLEED_MITIGATION_NONE;
-			break;
+			retbleed_mitigation = RETBLEED_MITIGATION_IBPB;
+			goto retbleed_force_ibpb;
 		}
 
 		setup_force_cpu_cap(X86_FEATURE_RETHUNK);
 		setup_force_cpu_cap(X86_FEATURE_UNRET);
 
-		if (!boot_cpu_has(X86_FEATURE_STIBP) &&
-		    (retbleed_nosmt || cpu_mitigations_auto_nosmt()))
-			cpu_smt_disable(false);
-
 		if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD &&
 		    boot_cpu_data.x86_vendor != X86_VENDOR_HYGON)
 			pr_err(RETBLEED_UNTRAIN_MSG);
+
+		mitigate_smt = true;
+		break;
+
+	case RETBLEED_MITIGATION_IBPB:
+retbleed_force_ibpb:
+		setup_force_cpu_cap(X86_FEATURE_ENTRY_IBPB);
+		mitigate_smt = true;
 		break;
 
 	default:
 		break;
 	}
 
+	if (mitigate_smt && !boot_cpu_has(X86_FEATURE_STIBP) &&
+	    (retbleed_nosmt || cpu_mitigations_auto_nosmt()))
+		cpu_smt_disable(false);
+
 	/*
 	 * Let IBRS trump all on Intel without affecting the effects of the
 	 * retbleed= cmdline option.