1 files changed, 179 insertions, 28 deletions

diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index 8712ce3c6a0e..d00d1a8f1f26 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -2,7 +2,7 @@
 #include <trace/events/syscalls.h>
 #include <linux/kernel.h>
 #include <linux/ftrace.h>
-#include <linux/perf_counter.h>
+#include <linux/perf_event.h>
 #include <asm/syscall.h>
 
 #include "trace_output.h"
@@ -14,6 +14,69 @@ static int sys_refcount_exit;
 static DECLARE_BITMAP(enabled_enter_syscalls, NR_syscalls);
 static DECLARE_BITMAP(enabled_exit_syscalls, NR_syscalls);
 
+extern unsigned long __start_syscalls_metadata[];
+extern unsigned long __stop_syscalls_metadata[];
+
+static struct syscall_metadata **syscalls_metadata;
+
+static struct syscall_metadata *find_syscall_meta(unsigned long syscall)
+{
+        struct syscall_metadata *start;
+        struct syscall_metadata *stop;
+        char str[KSYM_SYMBOL_LEN];
+
+
+        start = (struct syscall_metadata *)__start_syscalls_metadata;
+        stop = (struct syscall_metadata *)__stop_syscalls_metadata;
+        kallsyms_lookup(syscall, NULL, NULL, NULL, str);
+
+        for ( ; start < stop; start++) {
+                /*
+                 * Only compare after the "sys" prefix. Archs that use
+                 * syscall wrappers may have syscalls symbols aliases prefixed
+                 * with "SyS" instead of "sys", leading to an unwanted
+                 * mismatch.
+                 */
+                if (start->name && !strcmp(start->name + 3, str + 3))
+                        return start;
+        }
+        return NULL;
+}
+
+static struct syscall_metadata *syscall_nr_to_meta(int nr)
+{
+        if (!syscalls_metadata || nr >= NR_syscalls || nr < 0)
+                return NULL;
+
+        return syscalls_metadata[nr];
+}
+
+int syscall_name_to_nr(char *name)
+{
+        int i;
+
+        if (!syscalls_metadata)
+                return -1;
+
+        for (i = 0; i < NR_syscalls; i++) {
+                if (syscalls_metadata[i]) {
+                        if (!strcmp(syscalls_metadata[i]->name, name))
+                                return i;
+                }
+        }
+        return -1;
+}
+
+void set_syscall_enter_id(int num, int id)
+{
+        syscalls_metadata[num]->enter_id = id;
+}
+
+void set_syscall_exit_id(int num, int id)
+{
+        syscalls_metadata[num]->exit_id = id;
+}
+
 enum print_line_t
 print_syscall_enter(struct trace_iterator *iter, int flags)
 {
@@ -103,7 +166,8 @@ extern char *__bad_type_size(void);
 #define SYSCALL_FIELD(type, name)                                       \
         sizeof(type) != sizeof(trace.name) ?                            \
                 __bad_type_size() :                                     \
-                #type, #name, offsetof(typeof(trace), name), sizeof(trace.name)
+                #type, #name, offsetof(typeof(trace), name),            \
+                sizeof(trace.name), is_signed_type(type)
 
 int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
 {
@@ -120,7 +184,8 @@ int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
         if (!entry)
                 return 0;
 
-        ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+        ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
+                               "\tsigned:%u;\n",
                                SYSCALL_FIELD(int, nr));
         if (!ret)
                 return 0;
@@ -130,8 +195,10 @@ int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
                                         entry->args[i]);
                 if (!ret)
                         return 0;
-                ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;\n", offset,
-                                       sizeof(unsigned long));
+                ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;"
+                                       "\tsigned:%u;\n", offset,
+                                       sizeof(unsigned long),
+                                       is_signed_type(unsigned long));
                 if (!ret)
                         return 0;
                 offset += sizeof(unsigned long);
@@ -163,10 +230,12 @@ int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
         struct syscall_trace_exit trace;
 
         ret = trace_seq_printf(s,
-                               "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n"
-                               "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+                               "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
+                               "\tsigned:%u;\n"
+                               "\tfield:%s %s;\toffset:%zu;\tsize:%zu;"
+                               "\tsigned:%u;\n",
                                SYSCALL_FIELD(int, nr),
-                               SYSCALL_FIELD(unsigned long, ret));
+                               SYSCALL_FIELD(long, ret));
         if (!ret)
                 return 0;
 
@@ -212,7 +281,7 @@ int syscall_exit_define_fields(struct ftrace_event_call *call)
         if (ret)
                 return ret;
 
-        ret = trace_define_field(call, SYSCALL_FIELD(unsigned long, ret), 0,
+        ret = trace_define_field(call, SYSCALL_FIELD(long, ret),
                                  FILTER_OTHER);
 
         return ret;
@@ -375,6 +444,29 @@ struct trace_event event_syscall_exit = {
         .trace                  = print_syscall_exit,
 };
 
+int __init init_ftrace_syscalls(void)
+{
+        struct syscall_metadata *meta;
+        unsigned long addr;
+        int i;
+
+        syscalls_metadata = kzalloc(sizeof(*syscalls_metadata) *
+                                        NR_syscalls, GFP_KERNEL);
+        if (!syscalls_metadata) {
+                WARN_ON(1);
+                return -ENOMEM;
+        }
+
+        for (i = 0; i < NR_syscalls; i++) {
+                addr = arch_syscall_addr(i);
+                meta = find_syscall_meta(addr);
+                syscalls_metadata[i] = meta;
+        }
+
+        return 0;
+}
+core_initcall(init_ftrace_syscalls);
+
 #ifdef CONFIG_EVENT_PROFILE
 
 static DECLARE_BITMAP(enabled_prof_enter_syscalls, NR_syscalls);
@@ -384,10 +476,13 @@ static int sys_prof_refcount_exit;
 
 static void prof_syscall_enter(struct pt_regs *regs, long id)
 {
-        struct syscall_trace_enter *rec;
         struct syscall_metadata *sys_data;
+        struct syscall_trace_enter *rec;
+        unsigned long flags;
+        char *raw_data;
         int syscall_nr;
         int size;
+        int cpu;
 
         syscall_nr = syscall_get_nr(current, regs);
         if (!test_bit(syscall_nr, enabled_prof_enter_syscalls))
@@ -402,20 +497,38 @@ static void prof_syscall_enter(struct pt_regs *regs, long id)
         size = ALIGN(size + sizeof(u32), sizeof(u64));
         size -= sizeof(u32);
 
-        do {
-                char raw_data[size];
+        if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE,
+                      "profile buffer not large enough"))
+                return;
+
+        /* Protect the per cpu buffer, begin the rcu read side */
+        local_irq_save(flags);
 
-                /* zero the dead bytes from align to not leak stack to user */
-                *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;
+        cpu = smp_processor_id();
 
-                rec = (struct syscall_trace_enter *) raw_data;
-                tracing_generic_entry_update(&rec->ent, 0, 0);
-                rec->ent.type = sys_data->enter_id;
-                rec->nr = syscall_nr;
-                syscall_get_arguments(current, regs, 0, sys_data->nb_args,
-                                       (unsigned long *)&rec->args);
-                perf_tpcounter_event(sys_data->enter_id, 0, 1, rec, size);
-        } while(0);
+        if (in_nmi())
+                raw_data = rcu_dereference(trace_profile_buf_nmi);
+        else
+                raw_data = rcu_dereference(trace_profile_buf);
+
+        if (!raw_data)
+                goto end;
+
+        raw_data = per_cpu_ptr(raw_data, cpu);
+
+        /* zero the dead bytes from align to not leak stack to user */
+        *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;
+
+        rec = (struct syscall_trace_enter *) raw_data;
+        tracing_generic_entry_update(&rec->ent, 0, 0);
+        rec->ent.type = sys_data->enter_id;
+        rec->nr = syscall_nr;
+        syscall_get_arguments(current, regs, 0, sys_data->nb_args,
+                               (unsigned long *)&rec->args);
+        perf_tp_event(sys_data->enter_id, 0, 1, rec, size);
+
+end:
+        local_irq_restore(flags);
 }
 
 int reg_prof_syscall_enter(char *name)
@@ -460,8 +573,12 @@ void unreg_prof_syscall_enter(char *name)
 static void prof_syscall_exit(struct pt_regs *regs, long ret)
 {
         struct syscall_metadata *sys_data;
-        struct syscall_trace_exit rec;
+        struct syscall_trace_exit *rec;
+        unsigned long flags;
         int syscall_nr;
+        char *raw_data;
+        int size;
+        int cpu;
 
         syscall_nr = syscall_get_nr(current, regs);
         if (!test_bit(syscall_nr, enabled_prof_exit_syscalls))
@@ -471,12 +588,46 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret)
         if (!sys_data)
                 return;
 
-        tracing_generic_entry_update(&rec.ent, 0, 0);
-        rec.ent.type = sys_data->exit_id;
-        rec.nr = syscall_nr;
-        rec.ret = syscall_get_return_value(current, regs);
+        /* We can probably do that at build time */
+        size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64));
+        size -= sizeof(u32);
+
+        /*
+         * Impossible, but be paranoid with the future
+         * How to put this check outside runtime?
+         */
+        if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE,
+                "exit event has grown above profile buffer size"))
+                return;
+
+        /* Protect the per cpu buffer, begin the rcu read side */
+        local_irq_save(flags);
+        cpu = smp_processor_id();
+
+        if (in_nmi())
+                raw_data = rcu_dereference(trace_profile_buf_nmi);
+        else
+                raw_data = rcu_dereference(trace_profile_buf);
+
+        if (!raw_data)
+                goto end;
+
+        raw_data = per_cpu_ptr(raw_data, cpu);
+
+        /* zero the dead bytes from align to not leak stack to user */
+        *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;
+
+        rec = (struct syscall_trace_exit *)raw_data;
 
-        perf_tpcounter_event(sys_data->exit_id, 0, 1, &rec, sizeof(rec));
+        tracing_generic_entry_update(&rec->ent, 0, 0);
+        rec->ent.type = sys_data->exit_id;
+        rec->nr = syscall_nr;
+        rec->ret = syscall_get_return_value(current, regs);
+
+        perf_tp_event(sys_data->exit_id, 0, 1, rec, size);
+
+end:
+        local_irq_restore(flags);
 }
 
 int reg_prof_syscall_exit(char *name)