MIPS: Add uprobes support.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

5 files changed, 446 insertions, 0 deletions

diff --git a/arch/mips/kernel/Makefile b/arch/mips/kernel/Makefile
index 3f5cf8a..a61435b 100644
--- a/arch/mips/kernel/Makefile
+++ b/arch/mips/kernel/Makefile
@@ -100,6 +100,7 @@ obj-$(CONFIG_PERF_EVENTS)	+= perf_event.o
 obj-$(CONFIG_HW_PERF_EVENTS)	+= perf_event_mipsxx.o
 
 obj-$(CONFIG_JUMP_LABEL)	+= jump_label.o
+obj-$(CONFIG_UPROBES)		+= uprobes.o
 
 obj-$(CONFIG_MIPS_CM)		+= mips-cm.o
 obj-$(CONFIG_MIPS_CPC)		+= mips-cpc.o
diff --git a/arch/mips/kernel/ptrace.c b/arch/mips/kernel/ptrace.c
index e933a30..4f0ac78 100644
--- a/arch/mips/kernel/ptrace.c
+++ b/arch/mips/kernel/ptrace.c
@@ -25,6 +25,7 @@
 #include <linux/regset.h>
 #include <linux/smp.h>
 #include <linux/security.h>
+#include <linux/stddef.h>
 #include <linux/tracehook.h>
 #include <linux/audit.h>
 #include <linux/seccomp.h>
@@ -490,6 +491,93 @@ enum mips_regset {
 	REGSET_FPR,
 };
 
+struct pt_regs_offset {
+	const char *name;
+	int offset;
+};
+
+#define REG_OFFSET_NAME(reg, r) {					\
+	.name = #reg,							\
+	.offset = offsetof(struct pt_regs, r)				\
+}
+
+#define REG_OFFSET_END {						\
+	.name = NULL,							\
+	.offset = 0							\
+}
+
+static const struct pt_regs_offset regoffset_table[] = {
+	REG_OFFSET_NAME(r0, regs[0]),
+	REG_OFFSET_NAME(r1, regs[1]),
+	REG_OFFSET_NAME(r2, regs[2]),
+	REG_OFFSET_NAME(r3, regs[3]),
+	REG_OFFSET_NAME(r4, regs[4]),
+	REG_OFFSET_NAME(r5, regs[5]),
+	REG_OFFSET_NAME(r6, regs[6]),
+	REG_OFFSET_NAME(r7, regs[7]),
+	REG_OFFSET_NAME(r8, regs[8]),
+	REG_OFFSET_NAME(r9, regs[9]),
+	REG_OFFSET_NAME(r10, regs[10]),
+	REG_OFFSET_NAME(r11, regs[11]),
+	REG_OFFSET_NAME(r12, regs[12]),
+	REG_OFFSET_NAME(r13, regs[13]),
+	REG_OFFSET_NAME(r14, regs[14]),
+	REG_OFFSET_NAME(r15, regs[15]),
+	REG_OFFSET_NAME(r16, regs[16]),
+	REG_OFFSET_NAME(r17, regs[17]),
+	REG_OFFSET_NAME(r18, regs[18]),
+	REG_OFFSET_NAME(r19, regs[19]),
+	REG_OFFSET_NAME(r20, regs[20]),
+	REG_OFFSET_NAME(r21, regs[21]),
+	REG_OFFSET_NAME(r22, regs[22]),
+	REG_OFFSET_NAME(r23, regs[23]),
+	REG_OFFSET_NAME(r24, regs[24]),
+	REG_OFFSET_NAME(r25, regs[25]),
+	REG_OFFSET_NAME(r26, regs[26]),
+	REG_OFFSET_NAME(r27, regs[27]),
+	REG_OFFSET_NAME(r28, regs[28]),
+	REG_OFFSET_NAME(r29, regs[29]),
+	REG_OFFSET_NAME(r30, regs[30]),
+	REG_OFFSET_NAME(r31, regs[31]),
+	REG_OFFSET_NAME(c0_status, cp0_status),
+	REG_OFFSET_NAME(hi, hi),
+	REG_OFFSET_NAME(lo, lo),
+#ifdef CONFIG_CPU_HAS_SMARTMIPS
+	REG_OFFSET_NAME(acx, acx),
+#endif
+	REG_OFFSET_NAME(c0_badvaddr, cp0_badvaddr),
+	REG_OFFSET_NAME(c0_cause, cp0_cause),
+	REG_OFFSET_NAME(c0_epc, cp0_epc),
+#ifdef CONFIG_MIPS_MT_SMTC
+	REG_OFFSET_NAME(c0_tcstatus, cp0_tcstatus),
+#endif
+#ifdef CONFIG_CPU_CAVIUM_OCTEON
+	REG_OFFSET_NAME(mpl0, mpl[0]),
+	REG_OFFSET_NAME(mpl1, mpl[1]),
+	REG_OFFSET_NAME(mpl2, mpl[2]),
+	REG_OFFSET_NAME(mtp0, mtp[0]),
+	REG_OFFSET_NAME(mtp1, mtp[1]),
+	REG_OFFSET_NAME(mtp2, mtp[2]),
+#endif
+	REG_OFFSET_END,
+};
+
+/**
+ * regs_query_register_offset() - query register offset from its name
+ * @name:       the name of a register
+ *
+ * regs_query_register_offset() returns the offset of a register in struct
+ * pt_regs from its name. If the name is invalid, this returns -EINVAL;
+ */
+int regs_query_register_offset(const char *name)
+{
+        const struct pt_regs_offset *roff;
+        for (roff = regoffset_table; roff->name != NULL; roff++)
+                if (!strcmp(roff->name, name))
+                        return roff->offset;
+        return -EINVAL;
+}
+
 #if defined(CONFIG_32BIT) || defined(CONFIG_MIPS32_O32)
 
 static const struct user_regset mips_regsets[] = {
diff --git a/arch/mips/kernel/signal.c b/arch/mips/kernel/signal.c
index fa13a52..2fec67b 100644
--- a/arch/mips/kernel/signal.c
+++ b/arch/mips/kernel/signal.c
@@ -21,6 +21,7 @@
 #include <linux/wait.h>
 #include <linux/ptrace.h>
 #include <linux/unistd.h>
+#include <linux/uprobes.h>
 #include <linux/compiler.h>
 #include <linux/syscalls.h>
 #include <linux/uaccess.h>
@@ -856,6 +857,9 @@ asmlinkage void do_notify_resume(struct pt_regs *regs, void *unused,
 
 	user_exit();
 
+	if (thread_info_flags & _TIF_UPROBE)
+		uprobe_notify_resume(regs);
+
 	/* deal with pending signal delivery */
 	if (thread_info_flags & _TIF_SIGPENDING)
 		do_signal(regs);
diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
index cea964d..fdb392b 100644
--- a/arch/mips/kernel/traps.c
+++ b/arch/mips/kernel/traps.c
@@ -984,6 +984,18 @@ asmlinkage void do_bp(struct pt_regs *regs)
 	 * pertain to them.
 	 */
 	switch (bcode) {
+	case BRK_UPROBE:
+		if (notify_die(DIE_UPROBE, "uprobe", regs, bcode,
+			       current->thread.trap_nr, SIGTRAP) == NOTIFY_STOP)
+			goto out;
+		else
+			break;
+	case BRK_UPROBE_XOL:
+		if (notify_die(DIE_UPROBE_XOL, "uprobe_xol", regs, bcode,
+			       current->thread.trap_nr, SIGTRAP) == NOTIFY_STOP)
+			goto out;
+		else
+			break;
 	case BRK_KPROBE_BP:
 		if (notify_die(DIE_BREAK, "debug", regs, bcode,
 			       current->thread.trap_nr, SIGTRAP) == NOTIFY_STOP)
diff --git a/arch/mips/kernel/uprobes.c b/arch/mips/kernel/uprobes.c
new file mode 100644
index 0000000..8452d93
--- /dev/null
+++ b/arch/mips/kernel/uprobes.c
@@ -0,0 +1,341 @@
+#include <linux/highmem.h>
+#include <linux/kdebug.h>
+#include <linux/types.h>
+#include <linux/notifier.h>
+#include <linux/sched.h>
+#include <linux/uprobes.h>
+
+#include <asm/branch.h>
+#include <asm/cpu-features.h>
+#include <asm/ptrace.h>
+#include <asm/inst.h>
+
+static inline int insn_has_delay_slot(const union mips_instruction insn)
+{
+	switch (insn.i_format.opcode) {
+	/*
+	 * jr and jalr are in r_format format.
+	 */
+	case spec_op:
+		switch (insn.r_format.func) {
+		case jalr_op:
+		case jr_op:
+			return 1;
+		}
+		break;
+
+	/*
+	 * This group contains:
+	 * bltz_op, bgez_op, bltzl_op, bgezl_op,
+	 * bltzal_op, bgezal_op, bltzall_op, bgezall_op.
+	 */
+	case bcond_op:
+		switch (insn.i_format.rt) {
+		case bltz_op:
+		case bltzl_op:
+		case bgez_op:
+		case bgezl_op:
+		case bltzal_op:
+		case bltzall_op:
+		case bgezal_op:
+		case bgezall_op:
+		case bposge32_op:
+			return 1;
+		}
+		break;
+
+	/*
+	 * These are unconditional and in j_format.
+	 */
+	case jal_op:
+	case j_op:
+	case beq_op:
+	case beql_op:
+	case bne_op:
+	case bnel_op:
+	case blez_op: /* not really i_format */
+	case blezl_op:
+	case bgtz_op:
+	case bgtzl_op:
+		return 1;
+
+	/*
+	 * And now the FPA/cp1 branch instructions.
+	 */
+	case cop1_op:
+#ifdef CONFIG_CPU_CAVIUM_OCTEON
+	case lwc2_op: /* This is bbit0 on Octeon */
+	case ldc2_op: /* This is bbit032 on Octeon */
+	case swc2_op: /* This is bbit1 on Octeon */
+	case sdc2_op: /* This is bbit132 on Octeon */
+#endif
+		return 1;
+	}
+
+	return 0;
+}
+
+/**
+ * arch_uprobe_analyze_insn - instruction analysis including validity and fixups.
+ * @mm: the probed address space.
+ * @arch_uprobe: the probepoint information.
+ * @addr: virtual address at which to install the probepoint
+ * Return 0 on success or a -ve number on error.
+ */
+int arch_uprobe_analyze_insn(struct arch_uprobe *aup,
+	struct mm_struct *mm, unsigned long addr)
+{
+	union mips_instruction inst;
+
+	/*
+	 * For the time being this also blocks attempts to use uprobes with
+	 * MIPS16 and microMIPS.
+	 */
+	if (addr & 0x03)
+		return -EINVAL;
+
+	inst.word = aup->insn[0];
+	aup->ixol[0] = aup->insn[insn_has_delay_slot(inst)];
+	aup->ixol[1] = UPROBE_BRK_UPROBE_XOL;		/* NOP  */
+
+	return 0;
+}
+
+/**
+ * is_trap_insn - check if the instruction is a trap variant
+ * @insn: instruction to be checked.
+ * Returns true if @insn is a trap variant.
+ *
+ * This definition overrides the weak definition in kernel/events/uprobes.c.
+ * and is needed for the case where an architecture has multiple trap
+ * instructions (like PowerPC or MIPS).  We treat BREAK just like the more
+ * modern conditional trap instructions.
+ */
+bool is_trap_insn(uprobe_opcode_t *insn)
+{
+	union mips_instruction inst;
+
+	inst.word = *insn;
+
+	switch (inst.i_format.opcode) {
+	case spec_op:
+		switch (inst.r_format.func) {
+		case break_op:
+		case teq_op:
+		case tge_op:
+		case tgeu_op:
+		case tlt_op:
+		case tltu_op:
+		case tne_op:
+			return 1;
+		}
+		break;
+
+	case bcond_op:	/* Yes, really ...  */
+		switch (inst.u_format.rt) {
+		case teqi_op:
+		case tgei_op:
+		case tgeiu_op:
+		case tlti_op:
+		case tltiu_op:
+		case tnei_op:
+			return 1;
+		}
+		break;
+	}
+
+	return 0;
+}
+
+#define UPROBE_TRAP_NR	ULONG_MAX
+
+/*
+ * arch_uprobe_pre_xol - prepare to execute out of line.
+ * @auprobe: the probepoint information.
+ * @regs: reflects the saved user state of current task.
+ */
+int arch_uprobe_pre_xol(struct arch_uprobe *aup, struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+	union mips_instruction insn;
+
+	/*
+	 * Now find the EPC where to resume after the breakpoint has been
+	 * dealt with.  This may require emulation of a branch.
+	 */
+	aup->resume_epc = regs->cp0_epc + 4;
+	if (insn_has_delay_slot((union mips_instruction) aup->insn[0])) {
+		unsigned long epc;
+
+		epc = regs->cp0_epc;
+		__compute_return_epc_for_insn(regs, insn);
+		aup->resume_epc = regs->cp0_epc;
+	}
+
+	utask->autask.saved_trap_nr = current->thread.trap_nr;
+	current->thread.trap_nr = UPROBE_TRAP_NR;
+	regs->cp0_epc = current->utask->xol_vaddr;
+
+	return 0;
+}
+
+int arch_uprobe_post_xol(struct arch_uprobe *aup, struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+
+	current->thread.trap_nr = utask->autask.saved_trap_nr;
+	regs->cp0_epc = aup->resume_epc;
+
+	return 0;
+}
+
+/*
+ * If xol insn itself traps and generates a signal(Say,
+ * SIGILL/SIGSEGV/etc), then detect the case where a singlestepped
+ * instruction jumps back to its own address. It is assumed that anything
+ * like do_page_fault/do_trap/etc sets thread.trap_nr != -1.
+ *
+ * arch_uprobe_pre_xol/arch_uprobe_post_xol save/restore thread.trap_nr,
+ * arch_uprobe_xol_was_trapped() simply checks that ->trap_nr is not equal to
+ * UPROBE_TRAP_NR == -1 set by arch_uprobe_pre_xol().
+ */
+bool arch_uprobe_xol_was_trapped(struct task_struct *tsk)
+{
+	if (tsk->thread.trap_nr != UPROBE_TRAP_NR)
+		return true;
+
+	return false;
+}
+
+int arch_uprobe_exception_notify(struct notifier_block *self,
+	unsigned long val, void *data)
+{
+	struct die_args *args = data;
+	struct pt_regs *regs = args->regs;
+
+	/* regs == NULL is a kernel bug */
+	if (WARN_ON(!regs))
+		return NOTIFY_DONE;
+
+	/* We are only interested in userspace traps */
+	if (!user_mode(regs))
+		return NOTIFY_DONE;
+
+	switch (val) {
+	case DIE_BREAK:
+		if (uprobe_pre_sstep_notifier(regs))
+			return NOTIFY_STOP;
+		break;
+	case DIE_UPROBE_XOL:
+		if (uprobe_post_sstep_notifier(regs))
+			return NOTIFY_STOP;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+/*
+ * This function gets called when XOL instruction either gets trapped or
+ * the thread has a fatal signal. Reset the instruction pointer to its
+ * probed address for the potential restart or for post mortem analysis.
+ */
+void arch_uprobe_abort_xol(struct arch_uprobe *aup,
+	struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+
+	instruction_pointer_set(regs, utask->vaddr);
+}
+
+unsigned long arch_uretprobe_hijack_return_addr(
+	unsigned long trampoline_vaddr, struct pt_regs *regs)
+{
+	unsigned long ra;
+
+	ra = regs->regs[31];
+
+	/* Replace the return address with the trampoline address */
+	regs->regs[31] = ra;
+
+	return ra;
+}
+
+/**
+ * set_swbp - store breakpoint at a given address.
+ * @auprobe: arch specific probepoint information.
+ * @mm: the probed process address space.
+ * @vaddr: the virtual address to insert the opcode.
+ *
+ * For mm @mm, store the breakpoint instruction at @vaddr.
+ * Return 0 (success) or a negative errno.
+ *
+ * This version overrides the weak version in kernel/events/uprobes.c.
+ * It is required to handle MIPS16 and microMIPS.
+ */
+int __weak set_swbp(struct arch_uprobe *auprobe, struct mm_struct *mm,
+	unsigned long vaddr)
+{
+	return uprobe_write_opcode(mm, vaddr, UPROBE_SWBP_INSN);
+}
+
+/**
+ * set_orig_insn - Restore the original instruction.
+ * @mm: the probed process address space.
+ * @auprobe: arch specific probepoint information.
+ * @vaddr: the virtual address to insert the opcode.
+ *
+ * For mm @mm, restore the original opcode (opcode) at @vaddr.
+ * Return 0 (success) or a negative errno.
+ *
+ * This overrides the weak version in kernel/events/uprobes.c.
+ */
+int set_orig_insn(struct arch_uprobe *auprobe, struct mm_struct *mm,
+		 unsigned long vaddr)
+{
+	return uprobe_write_opcode(mm, vaddr,
+			*(uprobe_opcode_t *)&auprobe->orig_inst[0].word);
+}
+
+void __weak arch_uprobe_copy_ixol(struct page *page, unsigned long vaddr,
+				  void *src, unsigned long len)
+{
+	void *kaddr;
+
+	/* Initialize the slot */
+	kaddr = kmap_atomic(page);
+	memcpy(kaddr + (vaddr & ~PAGE_MASK), src, len);
+	kunmap_atomic(kaddr);
+
+	/*
+	 * The MIPS version of flush_icache_range will operate safely on
+	 * user space addresses and more importantly, it doesn't require a
+	 * VMA argument.
+	 */
+	flush_icache_range(vaddr, vaddr + len);
+}
+
+/**
+ * uprobe_get_swbp_addr - compute address of swbp given post-swbp regs
+ * @regs: Reflects the saved state of the task after it has hit a breakpoint
+ * instruction.
+ * Return the address of the breakpoint instruction.
+ *
+ * This overrides the weak version in kernel/events/uprobes.c.
+ */
+unsigned long uprobe_get_swbp_addr(struct pt_regs *regs)
+{
+	return instruction_pointer(regs);
+}
+
+/*
+ * See if the instruction can be emulated.
+ * Returns true if instruction was emulated, false otherwise.
+ *
+ * For now we always emulate so this function just returns 0.
+ */
+bool arch_uprobe_skip_sstep(struct arch_uprobe *auprobe, struct pt_regs *regs)
+{
+	return 0;
+}