#include <linux/audit.h>
#include <linux/seccomp.h>
#include <linux/signal.h>
+#include <linux/workqueue.h>
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/prctl.h>
#include <asm/proto.h>
#include <asm/ds.h>
-
-#include <trace/syscall.h>
+#include <asm/hw_breakpoint.h>
#include "tls.h"
+#define CREATE_TRACE_POINTS
+#include <trace/events/syscalls.h>
+
enum x86_regset {
REGSET_GENERAL,
REGSET_FP,
return 0;
}
-static unsigned long debugreg_addr_limit(struct task_struct *task)
-{
- return TASK_SIZE - 3;
-}
-
#else /* CONFIG_X86_64 */
#define FLAG_MASK (FLAG_MASK_32 | X86_EFLAGS_NT)
return 0;
}
-static unsigned long debugreg_addr_limit(struct task_struct *task)
-{
-#ifdef CONFIG_IA32_EMULATION
- if (test_tsk_thread_flag(task, TIF_IA32))
- return IA32_PAGE_OFFSET - 3;
-#endif
- return TASK_SIZE_MAX - 7;
-}
-
#endif /* CONFIG_X86_32 */
static unsigned long get_flags(struct task_struct *task)
}
/*
- * This function is trivial and will be inlined by the compiler.
- * Having it separates the implementation details of debug
- * registers from the interface details of ptrace.
+ * Decode the length and type bits for a particular breakpoint as
+ * stored in debug register 7. Return the "enabled" status.
*/
-static unsigned long ptrace_get_debugreg(struct task_struct *child, int n)
+static int decode_dr7(unsigned long dr7, int bpnum, unsigned *len,
+ unsigned *type)
{
- switch (n) {
- case 0: return child->thread.debugreg0;
- case 1: return child->thread.debugreg1;
- case 2: return child->thread.debugreg2;
- case 3: return child->thread.debugreg3;
- case 6: return child->thread.debugreg6;
- case 7: return child->thread.debugreg7;
- }
- return 0;
+ int bp_info = dr7 >> (DR_CONTROL_SHIFT + bpnum * DR_CONTROL_SIZE);
+
+ *len = (bp_info & 0xc) | 0x40;
+ *type = (bp_info & 0x3) | 0x80;
+ return (dr7 >> (bpnum * DR_ENABLE_SIZE)) & 0x3;
}
-static int ptrace_set_debugreg(struct task_struct *child,
- int n, unsigned long data)
+static void ptrace_triggered(struct hw_breakpoint *bp, struct pt_regs *regs)
{
+ struct thread_struct *thread = &(current->thread);
int i;
- if (unlikely(n == 4 || n == 5))
- return -EIO;
+ /*
+ * Store in the virtual DR6 register the fact that the breakpoint
+ * was hit so the thread's debugger will see it.
+ */
+ for (i = 0; i < hbp_kernel_pos; i++)
+ /*
+ * We will check bp->info.address against the address stored in
+ * thread's hbp structure and not debugreg[i]. This is to ensure
+ * that the corresponding bit for 'i' in DR7 register is enabled
+ */
+ if (bp->info.address == thread->hbp[i]->info.address)
+ break;
- if (n < 4 && unlikely(data >= debugreg_addr_limit(child)))
- return -EIO;
+ thread->debugreg6 |= (DR_TRAP0 << i);
+}
- switch (n) {
- case 0: child->thread.debugreg0 = data; break;
- case 1: child->thread.debugreg1 = data; break;
- case 2: child->thread.debugreg2 = data; break;
- case 3: child->thread.debugreg3 = data; break;
+/*
+ * Handle ptrace writes to debug register 7.
+ */
+static int ptrace_write_dr7(struct task_struct *tsk, unsigned long data)
+{
+ struct thread_struct *thread = &(tsk->thread);
+ unsigned long old_dr7 = thread->debugreg7;
+ int i, orig_ret = 0, rc = 0;
+ int enabled, second_pass = 0;
+ unsigned len, type;
+ struct hw_breakpoint *bp;
+
+ data &= ~DR_CONTROL_RESERVED;
+restore:
+ /*
+ * Loop through all the hardware breakpoints, making the
+ * appropriate changes to each.
+ */
+ for (i = 0; i < HBP_NUM; i++) {
+ enabled = decode_dr7(data, i, &len, &type);
+ bp = thread->hbp[i];
+
+ if (!enabled) {
+ if (bp) {
+ /* Don't unregister the breakpoints right-away,
+ * unless all register_user_hw_breakpoint()
+ * requests have succeeded. This prevents
+ * any window of opportunity for debug
+ * register grabbing by other users.
+ */
+ if (!second_pass)
+ continue;
+ unregister_user_hw_breakpoint(tsk, bp);
+ kfree(bp);
+ }
+ continue;
+ }
+ if (!bp) {
+ rc = -ENOMEM;
+ bp = kzalloc(sizeof(struct hw_breakpoint), GFP_KERNEL);
+ if (bp) {
+ bp->info.address = thread->debugreg[i];
+ bp->triggered = ptrace_triggered;
+ bp->info.len = len;
+ bp->info.type = type;
+ rc = register_user_hw_breakpoint(tsk, bp);
+ if (rc)
+ kfree(bp);
+ }
+ } else
+ rc = modify_user_hw_breakpoint(tsk, bp);
+ if (rc)
+ break;
+ }
+ /*
+ * Make a second pass to free the remaining unused breakpoints
+ * or to restore the original breakpoints if an error occurred.
+ */
+ if (!second_pass) {
+ second_pass = 1;
+ if (rc < 0) {
+ orig_ret = rc;
+ data = old_dr7;
+ }
+ goto restore;
+ }
+ return ((orig_ret < 0) ? orig_ret : rc);
+}
- case 6:
- if ((data & ~0xffffffffUL) != 0)
- return -EIO;
- child->thread.debugreg6 = data;
- break;
+/*
+ * Handle PTRACE_PEEKUSR calls for the debug register area.
+ */
+static unsigned long ptrace_get_debugreg(struct task_struct *tsk, int n)
+{
+ struct thread_struct *thread = &(tsk->thread);
+ unsigned long val = 0;
+
+ if (n < HBP_NUM)
+ val = thread->debugreg[n];
+ else if (n == 6)
+ val = thread->debugreg6;
+ else if (n == 7)
+ val = thread->debugreg7;
+ return val;
+}
- case 7:
- /*
- * Sanity-check data. Take one half-byte at once with
- * check = (val >> (16 + 4*i)) & 0xf. It contains the
- * R/Wi and LENi bits; bits 0 and 1 are R/Wi, and bits
- * 2 and 3 are LENi. Given a list of invalid values,
- * we do mask |= 1 << invalid_value, so that
- * (mask >> check) & 1 is a correct test for invalid
- * values.
- *
- * R/Wi contains the type of the breakpoint /
- * watchpoint, LENi contains the length of the watched
- * data in the watchpoint case.
- *
- * The invalid values are:
- * - LENi == 0x10 (undefined), so mask |= 0x0f00. [32-bit]
- * - R/Wi == 0x10 (break on I/O reads or writes), so
- * mask |= 0x4444.
- * - R/Wi == 0x00 && LENi != 0x00, so we have mask |=
- * 0x1110.
- *
- * Finally, mask = 0x0f00 | 0x4444 | 0x1110 == 0x5f54.
- *
- * See the Intel Manual "System Programming Guide",
- * 15.2.4
- *
- * Note that LENi == 0x10 is defined on x86_64 in long
- * mode (i.e. even for 32-bit userspace software, but
- * 64-bit kernel), so the x86_64 mask value is 0x5454.
- * See the AMD manual no. 24593 (AMD64 System Programming)
- */
-#ifdef CONFIG_X86_32
-#define DR7_MASK 0x5f54
-#else
-#define DR7_MASK 0x5554
-#endif
- data &= ~DR_CONTROL_RESERVED;
- for (i = 0; i < 4; i++)
- if ((DR7_MASK >> ((data >> (16 + 4*i)) & 0xf)) & 1)
- return -EIO;
- child->thread.debugreg7 = data;
- if (data)
- set_tsk_thread_flag(child, TIF_DEBUG);
- else
- clear_tsk_thread_flag(child, TIF_DEBUG);
- break;
+/*
+ * Handle PTRACE_POKEUSR calls for the debug register area.
+ */
+int ptrace_set_debugreg(struct task_struct *tsk, int n, unsigned long val)
+{
+ struct thread_struct *thread = &(tsk->thread);
+ int rc = 0;
+
+ /* There are no DR4 or DR5 registers */
+ if (n == 4 || n == 5)
+ return -EIO;
+
+ if (n == 6) {
+ tsk->thread.debugreg6 = val;
+ goto ret_path;
}
+ if (n < HBP_NUM) {
+ if (thread->hbp[n]) {
+ if (arch_check_va_in_userspace(val,
+ thread->hbp[n]->info.len) == 0) {
+ rc = -EIO;
+ goto ret_path;
+ }
+ thread->hbp[n]->info.address = val;
+ }
+ thread->debugreg[n] = val;
+ }
+ /* All that's left is DR7 */
+ if (n == 7)
+ rc = ptrace_write_dr7(tsk, val);
- return 0;
+ret_path:
+ return rc;
}
/*
}
#ifdef CONFIG_X86_PTRACE_BTS
+/*
+ * A branch trace store context.
+ *
+ * Contexts may only be installed by ptrace_bts_config() and only for
+ * ptraced tasks.
+ *
+ * Contexts are destroyed when the tracee is detached from the tracer.
+ * The actual destruction work requires interrupts enabled, so the
+ * work is deferred and will be scheduled during __ptrace_unlink().
+ *
+ * Contexts hold an additional task_struct reference on the traced
+ * task, as well as a reference on the tracer's mm.
+ *
+ * Ptrace already holds a task_struct for the duration of ptrace operations,
+ * but since destruction is deferred, it may be executed after both
+ * tracer and tracee exited.
+ */
+struct bts_context {
+ /* The branch trace handle. */
+ struct bts_tracer *tracer;
+
+ /* The buffer used to store the branch trace and its size. */
+ void *buffer;
+ unsigned int size;
+
+ /* The mm that paid for the above buffer. */
+ struct mm_struct *mm;
+
+ /* The task this context belongs to. */
+ struct task_struct *task;
+
+ /* The signal to send on a bts buffer overflow. */
+ unsigned int bts_ovfl_signal;
+
+ /* The work struct to destroy a context. */
+ struct work_struct work;
+};
+
+static int alloc_bts_buffer(struct bts_context *context, unsigned int size)
+{
+ void *buffer = NULL;
+ int err = -ENOMEM;
+
+ err = account_locked_memory(current->mm, current->signal->rlim, size);
+ if (err < 0)
+ return err;
+
+ buffer = kzalloc(size, GFP_KERNEL);
+ if (!buffer)
+ goto out_refund;
+
+ context->buffer = buffer;
+ context->size = size;
+ context->mm = get_task_mm(current);
+
+ return 0;
+
+ out_refund:
+ refund_locked_memory(current->mm, size);
+ return err;
+}
+
+static inline void free_bts_buffer(struct bts_context *context)
+{
+ if (!context->buffer)
+ return;
+
+ kfree(context->buffer);
+ context->buffer = NULL;
+
+ refund_locked_memory(context->mm, context->size);
+ context->size = 0;
+
+ mmput(context->mm);
+ context->mm = NULL;
+}
+
+static void free_bts_context_work(struct work_struct *w)
+{
+ struct bts_context *context;
+
+ context = container_of(w, struct bts_context, work);
+
+ ds_release_bts(context->tracer);
+ put_task_struct(context->task);
+ free_bts_buffer(context);
+ kfree(context);
+}
+
+static inline void free_bts_context(struct bts_context *context)
+{
+ INIT_WORK(&context->work, free_bts_context_work);
+ schedule_work(&context->work);
+}
+
+static inline struct bts_context *alloc_bts_context(struct task_struct *task)
+{
+ struct bts_context *context = kzalloc(sizeof(*context), GFP_KERNEL);
+ if (context) {
+ context->task = task;
+ task->bts = context;
+
+ get_task_struct(task);
+ }
+
+ return context;
+}
+
static int ptrace_bts_read_record(struct task_struct *child, size_t index,
struct bts_struct __user *out)
{
+ struct bts_context *context;
const struct bts_trace *trace;
struct bts_struct bts;
const unsigned char *at;
int error;
- trace = ds_read_bts(child->bts);
+ context = child->bts;
+ if (!context)
+ return -ESRCH;
+
+ trace = ds_read_bts(context->tracer);
if (!trace)
- return -EPERM;
+ return -ESRCH;
at = trace->ds.top - ((index + 1) * trace->ds.size);
if ((void *)at < trace->ds.begin)
if (!trace->read)
return -EOPNOTSUPP;
- error = trace->read(child->bts, at, &bts);
+ error = trace->read(context->tracer, at, &bts);
if (error < 0)
return error;
long size,
struct bts_struct __user *out)
{
+ struct bts_context *context;
const struct bts_trace *trace;
const unsigned char *at;
int error, drained = 0;
- trace = ds_read_bts(child->bts);
+ context = child->bts;
+ if (!context)
+ return -ESRCH;
+
+ trace = ds_read_bts(context->tracer);
if (!trace)
- return -EPERM;
+ return -ESRCH;
if (!trace->read)
return -EOPNOTSUPP;
for (at = trace->ds.begin; (void *)at < trace->ds.top;
out++, drained++, at += trace->ds.size) {
struct bts_struct bts;
- int error;
- error = trace->read(child->bts, at, &bts);
+ error = trace->read(context->tracer, at, &bts);
if (error < 0)
return error;
memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size);
- error = ds_reset_bts(child->bts);
+ error = ds_reset_bts(context->tracer);
if (error < 0)
return error;
return drained;
}
-static int ptrace_bts_allocate_buffer(struct task_struct *child, size_t size)
-{
- child->bts_buffer = alloc_locked_buffer(size);
- if (!child->bts_buffer)
- return -ENOMEM;
-
- child->bts_size = size;
-
- return 0;
-}
-
-static void ptrace_bts_free_buffer(struct task_struct *child)
-{
- free_locked_buffer(child->bts_buffer, child->bts_size);
- child->bts_buffer = NULL;
- child->bts_size = 0;
-}
-
static int ptrace_bts_config(struct task_struct *child,
long cfg_size,
const struct ptrace_bts_config __user *ucfg)
{
+ struct bts_context *context;
struct ptrace_bts_config cfg;
unsigned int flags = 0;
if (copy_from_user(&cfg, ucfg, sizeof(cfg)))
return -EFAULT;
- if (child->bts) {
- ds_release_bts(child->bts);
- child->bts = NULL;
- }
+ context = child->bts;
+ if (!context)
+ context = alloc_bts_context(child);
+ if (!context)
+ return -ENOMEM;
if (cfg.flags & PTRACE_BTS_O_SIGNAL) {
if (!cfg.signal)
return -EINVAL;
- child->thread.bts_ovfl_signal = cfg.signal;
return -EOPNOTSUPP;
+ context->bts_ovfl_signal = cfg.signal;
}
- if ((cfg.flags & PTRACE_BTS_O_ALLOC) &&
- (cfg.size != child->bts_size)) {
- int error;
+ ds_release_bts(context->tracer);
+ context->tracer = NULL;
- ptrace_bts_free_buffer(child);
+ if ((cfg.flags & PTRACE_BTS_O_ALLOC) && (cfg.size != context->size)) {
+ int err;
- error = ptrace_bts_allocate_buffer(child, cfg.size);
- if (error < 0)
- return error;
+ free_bts_buffer(context);
+ if (!cfg.size)
+ return 0;
+
+ err = alloc_bts_buffer(context, cfg.size);
+ if (err < 0)
+ return err;
}
if (cfg.flags & PTRACE_BTS_O_TRACE)
if (cfg.flags & PTRACE_BTS_O_SCHED)
flags |= BTS_TIMESTAMPS;
- child->bts = ds_request_bts(child, child->bts_buffer, child->bts_size,
- /* ovfl = */ NULL, /* th = */ (size_t)-1,
- flags);
- if (IS_ERR(child->bts)) {
- int error = PTR_ERR(child->bts);
-
- ptrace_bts_free_buffer(child);
- child->bts = NULL;
+ context->tracer =
+ ds_request_bts_task(child, context->buffer, context->size,
+ NULL, (size_t)-1, flags);
+ if (unlikely(IS_ERR(context->tracer))) {
+ int error = PTR_ERR(context->tracer);
+ free_bts_buffer(context);
+ context->tracer = NULL;
return error;
}
long cfg_size,
struct ptrace_bts_config __user *ucfg)
{
+ struct bts_context *context;
const struct bts_trace *trace;
struct ptrace_bts_config cfg;
+ context = child->bts;
+ if (!context)
+ return -ESRCH;
+
if (cfg_size < sizeof(cfg))
return -EIO;
- trace = ds_read_bts(child->bts);
+ trace = ds_read_bts(context->tracer);
if (!trace)
- return -EPERM;
+ return -ESRCH;
memset(&cfg, 0, sizeof(cfg));
- cfg.size = trace->ds.end - trace->ds.begin;
- cfg.signal = child->thread.bts_ovfl_signal;
- cfg.bts_size = sizeof(struct bts_struct);
+ cfg.size = trace->ds.end - trace->ds.begin;
+ cfg.signal = context->bts_ovfl_signal;
+ cfg.bts_size = sizeof(struct bts_struct);
if (cfg.signal)
cfg.flags |= PTRACE_BTS_O_SIGNAL;
static int ptrace_bts_clear(struct task_struct *child)
{
+ struct bts_context *context;
const struct bts_trace *trace;
- trace = ds_read_bts(child->bts);
+ context = child->bts;
+ if (!context)
+ return -ESRCH;
+
+ trace = ds_read_bts(context->tracer);
if (!trace)
- return -EPERM;
+ return -ESRCH;
memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size);
- return ds_reset_bts(child->bts);
+ return ds_reset_bts(context->tracer);
}
static int ptrace_bts_size(struct task_struct *child)
{
+ struct bts_context *context;
const struct bts_trace *trace;
- trace = ds_read_bts(child->bts);
+ context = child->bts;
+ if (!context)
+ return -ESRCH;
+
+ trace = ds_read_bts(context->tracer);
if (!trace)
- return -EPERM;
+ return -ESRCH;
return (trace->ds.top - trace->ds.begin) / trace->ds.size;
}
-static void ptrace_bts_fork(struct task_struct *tsk)
-{
- tsk->bts = NULL;
- tsk->bts_buffer = NULL;
- tsk->bts_size = 0;
- tsk->thread.bts_ovfl_signal = 0;
-}
-
-static void ptrace_bts_untrace(struct task_struct *child)
+/*
+ * Called from __ptrace_unlink() after the child has been moved back
+ * to its original parent.
+ */
+void ptrace_bts_untrace(struct task_struct *child)
{
if (unlikely(child->bts)) {
- ds_release_bts(child->bts);
+ free_bts_context(child->bts);
child->bts = NULL;
-
- /* We cannot update total_vm and locked_vm since
- child's mm is already gone. But we can reclaim the
- memory. */
- kfree(child->bts_buffer);
- child->bts_buffer = NULL;
- child->bts_size = 0;
}
}
-
-static void ptrace_bts_detach(struct task_struct *child)
-{
- /*
- * Ptrace_detach() races with ptrace_untrace() in case
- * the child dies and is reaped by another thread.
- *
- * We only do the memory accounting at this point and
- * leave the buffer deallocation and the bts tracer
- * release to ptrace_bts_untrace() which will be called
- * later on with tasklist_lock held.
- */
- release_locked_buffer(child->bts_buffer, child->bts_size);
-}
-#else
-static inline void ptrace_bts_fork(struct task_struct *tsk) {}
-static inline void ptrace_bts_detach(struct task_struct *child) {}
-static inline void ptrace_bts_untrace(struct task_struct *child) {}
#endif /* CONFIG_X86_PTRACE_BTS */
-void x86_ptrace_fork(struct task_struct *child, unsigned long clone_flags)
-{
- ptrace_bts_fork(child);
-}
-
-void x86_ptrace_untrace(struct task_struct *child)
-{
- ptrace_bts_untrace(child);
-}
-
/*
* Called by kernel/ptrace.c when detaching..
*
#ifdef TIF_SYSCALL_EMU
clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
#endif
- ptrace_bts_detach(child);
}
#if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION
tracehook_report_syscall_entry(regs))
ret = -1L;
- if (unlikely(test_thread_flag(TIF_SYSCALL_FTRACE)))
- ftrace_syscall_enter(regs);
+ if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
+ trace_sys_enter(regs, regs->orig_ax);
if (unlikely(current->audit_context)) {
if (IS_IA32)
if (unlikely(current->audit_context))
audit_syscall_exit(AUDITSC_RESULT(regs->ax), regs->ax);
- if (unlikely(test_thread_flag(TIF_SYSCALL_FTRACE)))
- ftrace_syscall_exit(regs);
+ if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
+ trace_sys_exit(regs, regs->ax);
if (test_thread_flag(TIF_SYSCALL_TRACE))
tracehook_report_syscall_exit(regs, 0);