[PATCH 1/5] powerpc/asm: Add a patch_site macro & helpers for patching instructions

@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright 2018, Michael Ellerman, IBM Corporation.
+ */
+#ifndef _ASM_POWERPC_CODE_PATCHING_ASM_H
+#define _ASM_POWERPC_CODE_PATCHING_ASM_H
+
+/* Define a "site" that can be patched */
+.macro patch_site label name
+	.pushsection ".rodata"
+	.balign 4
+	.global \name
+\name:
+	.4byte	\label - .
+	.popsection
+.endm
+
+#endif /* _ASM_POWERPC_CODE_PATCHING_ASM_H */

@@ -32,6 +32,8 @@ unsigned int create_cond_branch(const unsigned int *addr,
 int patch_branch(unsigned int *addr, unsigned long target, int flags);
 int patch_instruction(unsigned int *addr, unsigned int instr);
 int raw_patch_instruction(unsigned int *addr, unsigned int instr);
+int patch_instruction_site(s32 *addr, unsigned int instr);
+int patch_branch_site(s32 *site, unsigned long target, int flags);
 
 int instr_is_relative_branch(unsigned int instr);
 int instr_is_relative_link_branch(unsigned int instr);

@@ -195,6 +195,22 @@ int patch_branch(unsigned int *addr, unsigned long target, int flags)
 	return patch_instruction(addr, create_branch(addr, target, flags));
 }
 
+int patch_branch_site(s32 *site, unsigned long target, int flags)
+{
+	unsigned int *addr;
+
+	addr = (unsigned int *)((unsigned long)site + *site);
+	return patch_instruction(addr, create_branch(addr, target, flags));
+}
+
+int patch_instruction_site(s32 *site, unsigned int instr)
+{
+	unsigned int *addr;
+
+	addr = (unsigned int *)((unsigned long)site + *site);
+	return patch_instruction(addr, instr);
+}
+
 bool is_offset_in_branch_range(long offset)
 {
 	/*

@@ -59,6 +59,9 @@ static inline bool security_ftr_enabled(unsigned long feature)
 // Indirect branch prediction cache disabled
 #define SEC_FTR_COUNT_CACHE_DISABLED	0x0000000000000020ull
 
+// bcctr 2,0,0 triggers a hardware assisted count cache flush
+#define SEC_FTR_BCCTR_FLUSH_ASSIST	0x0000000000000800ull
+
 
 // Features indicating need for Spectre/Meltdown mitigations
 

@@ -74,6 +77,9 @@ static inline bool security_ftr_enabled(unsigned long feature)
 // Firmware configuration indicates user favours security over performance
 #define SEC_FTR_FAVOUR_SECURITY		0x0000000000000200ull
 
+// Software required to flush count cache on context switch
+#define SEC_FTR_FLUSH_COUNT_CACHE	0x0000000000000400ull
+
 
 // Features enabled by default
 #define SEC_FTR_DEFAULT \

@@ -143,4 +143,10 @@ struct kvm_vcpu;
 void _kvmppc_restore_tm_pr(struct kvm_vcpu *vcpu, u64 guest_msr);
 void _kvmppc_save_tm_pr(struct kvm_vcpu *vcpu, u64 guest_msr);
 
+/* Patch sites */
+extern s32 patch__call_flush_count_cache;
+extern s32 patch__flush_count_cache_return;
+
+extern long flush_count_cache;
+
 #endif /* _ASM_POWERPC_ASM_PROTOTYPES_H */

@@ -22,6 +22,7 @@ enum stf_barrier_type {
 
 void setup_stf_barrier(void);
 void do_stf_barrier_fixups(enum stf_barrier_type types);
+void setup_count_cache_flush(void);
 
 static inline void security_ftr_set(unsigned long feature)
 {

@@ -25,6 +25,7 @@
 #include <asm/page.h>
 #include <asm/mmu.h>
 #include <asm/thread_info.h>
+#include <asm/code-patching-asm.h>
 #include <asm/ppc_asm.h>
 #include <asm/asm-offsets.h>
 #include <asm/cputable.h>

@@ -504,6 +505,57 @@ _GLOBAL(ret_from_kernel_thread)
 	li	r3,0
 	b	.Lsyscall_exit
 
+#ifdef CONFIG_PPC_BOOK3S_64
+
+#define FLUSH_COUNT_CACHE	\
+1:	nop;			\
+	patch_site 1b, patch__call_flush_count_cache
+
+
+#define BCCTR_FLUSH	.long 0x4c400420
+
+.macro nops number
+	.rept \number
+	nop
+	.endr
+.endm
+
+.balign 32
+.global flush_count_cache
+flush_count_cache:
+	/* Save LR into r9 */
+	mflr	r9
+
+	.rept 64
+	bl	.+4
+	.endr
+	b	1f
+	nops	6
+
+	.balign 32
+	/* Restore LR */
+1:	mtlr	r9
+	li	r9,0x7fff
+	mtctr	r9
+
+	BCCTR_FLUSH
+
+2:	nop
+	patch_site 2b patch__flush_count_cache_return
+
+	nops	3
+
+	.rept 278
+	.balign 32
+	BCCTR_FLUSH
+	nops	7
+	.endr
+
+	blr
+#else
+#define FLUSH_COUNT_CACHE
+#endif /* CONFIG_PPC_BOOK3S_64 */
+
 /*
  * This routine switches between two different tasks.  The process
  * state of one is saved on its kernel stack.  Then the state

@@ -535,6 +587,8 @@ _GLOBAL(_switch)
 	std	r23,_CCR(r1)
 	std	r1,KSP(r3)	/* Set old stack pointer */
 
+	FLUSH_COUNT_CACHE
+
 	/*
 	 * On SMP kernels, care must be taken because a task may be
 	 * scheduled off CPUx and on to CPUy. Memory ordering must be

@@ -8,6 +8,8 @@
 #include <linux/device.h>
 #include <linux/seq_buf.h>
 
+#include <asm/asm-prototypes.h>
+#include <asm/code-patching.h>
 #include <asm/debugfs.h>
 #include <asm/security_features.h>
 #include <asm/setup.h>

@@ -15,6 +17,13 @@
 
 unsigned long powerpc_security_features __read_mostly = SEC_FTR_DEFAULT;
 
+enum count_cache_flush_type {
+	COUNT_CACHE_FLUSH_NONE	= 0x1,
+	COUNT_CACHE_FLUSH_SW	= 0x2,
+	COUNT_CACHE_FLUSH_HW	= 0x4,
+};
+static enum count_cache_flush_type count_cache_flush_type;
+
 bool barrier_nospec_enabled;
 
 static void enable_barrier_nospec(bool enable)

@@ -147,17 +156,29 @@ ssize_t cpu_show_spectre_v2(struct device *dev, struct device_attribute *attr, c
 	bcs = security_ftr_enabled(SEC_FTR_BCCTRL_SERIALISED);
 	ccd = security_ftr_enabled(SEC_FTR_COUNT_CACHE_DISABLED);
 
-	if (bcs || ccd) {
+	if (bcs || ccd || count_cache_flush_type != COUNT_CACHE_FLUSH_NONE) {
+		bool comma = false;
 		seq_buf_printf(&s, "Mitigation: ");
 
-		if (bcs)
+		if (bcs) {
 			seq_buf_printf(&s, "Indirect branch serialisation (kernel only)");
+			comma = true;
+		}
+
+		if (ccd) {
+			if (comma)
+				seq_buf_printf(&s, ", ");
+			seq_buf_printf(&s, "Indirect branch cache disabled");
+			comma = true;
+		}
 
-		if (bcs && ccd)
+		if (comma)
 			seq_buf_printf(&s, ", ");
 
-		if (ccd)
-			seq_buf_printf(&s, "Indirect branch cache disabled");
+		seq_buf_printf(&s, "Software count cache flush");
+
+		if (count_cache_flush_type == COUNT_CACHE_FLUSH_HW)
+			seq_buf_printf(&s, "(hardware accelerated)");
 	} else
 		seq_buf_printf(&s, "Vulnerable");
 

@@ -313,3 +334,68 @@ static __init int stf_barrier_debugfs_init(void)
 }
 device_initcall(stf_barrier_debugfs_init);
 #endif /* CONFIG_DEBUG_FS */
+
+static void toggle_count_cache_flush(bool enable)
+{
+	if (!enable || !security_ftr_enabled(SEC_FTR_FLUSH_COUNT_CACHE)) {
+		patch_instruction_site(&patch__call_flush_count_cache, PPC_INST_NOP);
+		count_cache_flush_type = COUNT_CACHE_FLUSH_NONE;
+		pr_info("count-cache-flush: software flush disabled.\n");
+		return;
+	}
+
+	patch_branch_site(&patch__call_flush_count_cache,
+			  (u64)&flush_count_cache, BRANCH_SET_LINK);
+
+	if (!security_ftr_enabled(SEC_FTR_BCCTR_FLUSH_ASSIST)) {
+		count_cache_flush_type = COUNT_CACHE_FLUSH_SW;
+		pr_info("count-cache-flush: full software flush sequence enabled.\n");
+		return;
+	}
+
+	patch_instruction_site(&patch__flush_count_cache_return, PPC_INST_BLR);
+	count_cache_flush_type = COUNT_CACHE_FLUSH_HW;
+	pr_info("count-cache-flush: hardware assisted flush sequence enabled\n");
+}
+
+void setup_count_cache_flush(void)
+{
+	toggle_count_cache_flush(true);
+}
+
+#ifdef CONFIG_DEBUG_FS
+static int count_cache_flush_set(void *data, u64 val)
+{
+	bool enable;
+
+	if (val == 1)
+		enable = true;
+	else if (val == 0)
+		enable = false;
+	else
+		return -EINVAL;
+
+	toggle_count_cache_flush(enable);
+
+	return 0;
+}
+
+static int count_cache_flush_get(void *data, u64 *val)
+{
+	if (count_cache_flush_type == COUNT_CACHE_FLUSH_NONE)
+		*val = 0;
+	else
+		*val = 1;
+
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(fops_count_cache_flush, count_cache_flush_get, count_cache_flush_set, "%llu\n");
+
+static __init int count_cache_flush_debugfs_init(void)
+{
+	debugfs_create_file("count_cache_flush", 0600, powerpc_debugfs_root, NULL, &fops_count_cache_flush);
+	return 0;
+}
+device_initcall(count_cache_flush_debugfs_init);
+#endif /* CONFIG_DEBUG_FS */

@@ -342,10 +342,12 @@
 #define H_CPU_CHAR_BRANCH_HINTS_HONORED	(1ull << 58) // IBM bit 5
 #define H_CPU_CHAR_THREAD_RECONFIG_CTRL	(1ull << 57) // IBM bit 6
 #define H_CPU_CHAR_COUNT_CACHE_DISABLED	(1ull << 56) // IBM bit 7
+#define H_CPU_CHAR_BCCTR_FLUSH_ASSIST	(1ull << 54) // IBM bit 9
 
 #define H_CPU_BEHAV_FAVOUR_SECURITY	(1ull << 63) // IBM bit 0
 #define H_CPU_BEHAV_L1D_FLUSH_PR	(1ull << 62) // IBM bit 1
 #define H_CPU_BEHAV_BNDS_CHK_SPEC_BAR	(1ull << 61) // IBM bit 2
+#define H_CPU_BEHAV_FLUSH_COUNT_CACHE	(1ull << 58) // IBM bit 5
 
 /* Flag values used in H_REGISTER_PROC_TBL hcall */
 #define PROC_TABLE_OP_MASK	0x18

@@ -484,6 +484,12 @@ static void init_cpu_char_feature_flags(struct h_cpu_char_result *result)
 	if (result->character & H_CPU_CHAR_COUNT_CACHE_DISABLED)
 		security_ftr_set(SEC_FTR_COUNT_CACHE_DISABLED);
 
+	if (result->character & H_CPU_CHAR_BCCTR_FLUSH_ASSIST)
+		security_ftr_set(SEC_FTR_BCCTR_FLUSH_ASSIST);
+
+	if (result->behaviour & H_CPU_BEHAV_FLUSH_COUNT_CACHE)
+		security_ftr_set(SEC_FTR_FLUSH_COUNT_CACHE);
+
 	/*
 	 * The features below are enabled by default, so we instead look to see
 	 * if firmware has *disabled* them, and clear them if so.

@@ -535,6 +541,7 @@ void pseries_setup_rfi_flush(void)
 
 	setup_rfi_flush(types, enable);
 	setup_barrier_nospec();
+	setup_count_cache_flush();
 }
 
 #ifdef CONFIG_PCI_IOV

@@ -78,6 +78,12 @@ static void init_fw_feat_flags(struct device_node *np)
 	if (fw_feature_is("enabled", "fw-count-cache-disabled", np))
 		security_ftr_set(SEC_FTR_COUNT_CACHE_DISABLED);
 
+	if (fw_feature_is("enabled", "fw-count-cache-flush-bcctr2,0,0", np))
+		security_ftr_set(SEC_FTR_BCCTR_FLUSH_ASSIST);
+
+	if (fw_feature_is("enabled", "needs-count-cache-flush-on-context-switch", np))
+		security_ftr_set(SEC_FTR_FLUSH_COUNT_CACHE);
+
 	/*
 	 * The features below are enabled by default, so we instead look to see
 	 * if firmware has *disabled* them, and clear them if so.

@@ -125,6 +131,7 @@ static void pnv_setup_rfi_flush(void)
 
 	setup_rfi_flush(type, enable);
 	setup_barrier_nospec();
+	setup_count_cache_flush();
 }
 
 static void __init pnv_setup_arch(void)