Linux-2.6.12-rc2
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
diff --git a/arch/um/kernel/tt/tracer.c b/arch/um/kernel/tt/tracer.c
new file mode 100644
index 0000000..7b5d937
--- /dev/null
+++ b/arch/um/kernel/tt/tracer.c
@@ -0,0 +1,480 @@
+/*
+ * Copyright (C) 2002 Jeff Dike (jdike@karaya.com)
+ * Licensed under the GPL
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#include <unistd.h>
+#include <signal.h>
+#include <errno.h>
+#include <sched.h>
+#include <string.h>
+#include <sys/mman.h>
+#include <sys/time.h>
+#include <sys/wait.h>
+#include "user.h"
+#include "sysdep/ptrace.h"
+#include "sigcontext.h"
+#include "sysdep/sigcontext.h"
+#include "os.h"
+#include "signal_user.h"
+#include "user_util.h"
+#include "mem_user.h"
+#include "process.h"
+#include "kern_util.h"
+#include "chan_user.h"
+#include "ptrace_user.h"
+#include "mode.h"
+#include "tt.h"
+
+static int tracer_winch[2];
+
+int is_tracer_winch(int pid, int fd, void *data)
+{
+ if(pid != tracing_pid)
+ return(0);
+
+ register_winch_irq(tracer_winch[0], fd, -1, data);
+ return(1);
+}
+
+static void tracer_winch_handler(int sig)
+{
+ int n;
+ char c = 1;
+
+ n = os_write_file(tracer_winch[1], &c, sizeof(c));
+ if(n != sizeof(c))
+ printk("tracer_winch_handler - write failed, err = %d\n", -n);
+}
+
+/* Called only by the tracing thread during initialization */
+
+static void setup_tracer_winch(void)
+{
+ int err;
+
+ err = os_pipe(tracer_winch, 1, 1);
+ if(err < 0){
+ printk("setup_tracer_winch : os_pipe failed, err = %d\n", -err);
+ return;
+ }
+ signal(SIGWINCH, tracer_winch_handler);
+}
+
+void attach_process(int pid)
+{
+ if((ptrace(PTRACE_ATTACH, pid, 0, 0) < 0) ||
+ (ptrace(PTRACE_CONT, pid, 0, 0) < 0))
+ tracer_panic("OP_FORK failed to attach pid");
+ wait_for_stop(pid, SIGSTOP, PTRACE_CONT, NULL);
+ if (ptrace(PTRACE_OLDSETOPTIONS, pid, 0, (void *)PTRACE_O_TRACESYSGOOD) < 0)
+ tracer_panic("OP_FORK: PTRACE_SETOPTIONS failed, errno = %d", errno);
+ if(ptrace(PTRACE_CONT, pid, 0, 0) < 0)
+ tracer_panic("OP_FORK failed to continue process");
+}
+
+void tracer_panic(char *format, ...)
+{
+ va_list ap;
+
+ va_start(ap, format);
+ vprintf(format, ap);
+ va_end(ap);
+ printf("\n");
+ while(1) pause();
+}
+
+static void tracer_segv(int sig, struct sigcontext sc)
+{
+ printf("Tracing thread segfault at address 0x%lx, ip 0x%lx\n",
+ SC_FAULT_ADDR(&sc), SC_IP(&sc));
+ while(1)
+ pause();
+}
+
+/* Changed early in boot, and then only read */
+int debug = 0;
+int debug_stop = 1;
+int debug_parent = 0;
+int honeypot = 0;
+
+static int signal_tramp(void *arg)
+{
+ int (*proc)(void *);
+
+ if(honeypot && munmap((void *) (host_task_size - 0x10000000),
+ 0x10000000))
+ panic("Unmapping stack failed");
+ if(ptrace(PTRACE_TRACEME, 0, 0, 0) < 0)
+ panic("ptrace PTRACE_TRACEME failed");
+ os_stop_process(os_getpid());
+ change_sig(SIGWINCH, 0);
+ signal(SIGUSR1, SIG_IGN);
+ change_sig(SIGCHLD, 0);
+ signal(SIGSEGV, (__sighandler_t) sig_handler);
+ set_cmdline("(idle thread)");
+ set_init_pid(os_getpid());
+ proc = arg;
+ return((*proc)(NULL));
+}
+
+static void sleeping_process_signal(int pid, int sig)
+{
+ switch(sig){
+ /* These two result from UML being ^Z-ed and bg-ed. PTRACE_CONT is
+ * right because the process must be in the kernel already.
+ */
+ case SIGCONT:
+ case SIGTSTP:
+ if(ptrace(PTRACE_CONT, pid, 0, sig) < 0)
+ tracer_panic("sleeping_process_signal : Failed to "
+ "continue pid %d, signal = %d, "
+ "errno = %d\n", pid, sig, errno);
+ break;
+
+ /* This happens when the debugger (e.g. strace) is doing system call
+ * tracing on the kernel. During a context switch, the current task
+ * will be set to the incoming process and the outgoing process will
+ * hop into write and then read. Since it's not the current process
+ * any more, the trace of those will land here. So, we need to just
+ * PTRACE_SYSCALL it.
+ */
+ case (SIGTRAP + 0x80):
+ if(ptrace(PTRACE_SYSCALL, pid, 0, 0) < 0)
+ tracer_panic("sleeping_process_signal : Failed to "
+ "PTRACE_SYSCALL pid %d, errno = %d\n",
+ pid, errno);
+ break;
+ case SIGSTOP:
+ break;
+ default:
+ tracer_panic("sleeping process %d got unexpected "
+ "signal : %d\n", pid, sig);
+ break;
+ }
+}
+
+/* Accessed only by the tracing thread */
+int debugger_pid = -1;
+int debugger_parent = -1;
+int debugger_fd = -1;
+int gdb_pid = -1;
+
+struct {
+ int pid;
+ int signal;
+ unsigned long addr;
+ struct timeval time;
+} signal_record[1024][32];
+
+int signal_index[32];
+int nsignals = 0;
+int debug_trace = 0;
+extern int io_nsignals, io_count, intr_count;
+
+extern void signal_usr1(int sig);
+
+int tracing_pid = -1;
+
+int tracer(int (*init_proc)(void *), void *sp)
+{
+ void *task = NULL;
+ int status, pid = 0, sig = 0, cont_type, tracing = 0, op = 0;
+ int proc_id = 0, n, err, old_tracing = 0, strace = 0;
+ int local_using_sysemu = 0;
+#ifdef UML_CONFIG_SYSCALL_DEBUG
+ unsigned long eip = 0;
+ int last_index;
+#endif
+ signal(SIGPIPE, SIG_IGN);
+ setup_tracer_winch();
+ tracing_pid = os_getpid();
+ printf("tracing thread pid = %d\n", tracing_pid);
+
+ pid = clone(signal_tramp, sp, CLONE_FILES | SIGCHLD, init_proc);
+ CATCH_EINTR(n = waitpid(pid, &status, WUNTRACED));
+ if(n < 0){
+ printf("waitpid on idle thread failed, errno = %d\n", errno);
+ exit(1);
+ }
+ if (ptrace(PTRACE_OLDSETOPTIONS, pid, 0, (void *)PTRACE_O_TRACESYSGOOD) < 0) {
+ printf("Failed to PTRACE_SETOPTIONS for idle thread, errno = %d\n", errno);
+ exit(1);
+ }
+ if((ptrace(PTRACE_CONT, pid, 0, 0) < 0)){
+ printf("Failed to continue idle thread, errno = %d\n", errno);
+ exit(1);
+ }
+
+ signal(SIGSEGV, (sighandler_t) tracer_segv);
+ signal(SIGUSR1, signal_usr1);
+ if(debug_trace){
+ printf("Tracing thread pausing to be attached\n");
+ stop();
+ }
+ if(debug){
+ if(gdb_pid != -1)
+ debugger_pid = attach_debugger(pid, gdb_pid, 1);
+ else debugger_pid = init_ptrace_proxy(pid, 1, debug_stop);
+ if(debug_parent){
+ debugger_parent = os_process_parent(debugger_pid);
+ init_parent_proxy(debugger_parent);
+ err = attach(debugger_parent);
+ if(err){
+ printf("Failed to attach debugger parent %d, "
+ "errno = %d\n", debugger_parent, -err);
+ debugger_parent = -1;
+ }
+ else {
+ if(ptrace(PTRACE_SYSCALL, debugger_parent,
+ 0, 0) < 0){
+ printf("Failed to continue debugger "
+ "parent, errno = %d\n", errno);
+ debugger_parent = -1;
+ }
+ }
+ }
+ }
+ set_cmdline("(tracing thread)");
+ while(1){
+ CATCH_EINTR(pid = waitpid(-1, &status, WUNTRACED));
+ if(pid <= 0){
+ if(errno != ECHILD){
+ printf("wait failed - errno = %d\n", errno);
+ }
+ continue;
+ }
+ if(pid == debugger_pid){
+ int cont = 0;
+
+ if(WIFEXITED(status) || WIFSIGNALED(status))
+ debugger_pid = -1;
+ /* XXX Figure out how to deal with gdb and SMP */
+ else cont = debugger_signal(status, cpu_tasks[0].pid);
+ if(cont == PTRACE_SYSCALL) strace = 1;
+ continue;
+ }
+ else if(pid == debugger_parent){
+ debugger_parent_signal(status, pid);
+ continue;
+ }
+ nsignals++;
+ if(WIFEXITED(status)) ;
+#ifdef notdef
+ {
+ printf("Child %d exited with status %d\n", pid,
+ WEXITSTATUS(status));
+ }
+#endif
+ else if(WIFSIGNALED(status)){
+ sig = WTERMSIG(status);
+ if(sig != 9){
+ printf("Child %d exited with signal %d\n", pid,
+ sig);
+ }
+ }
+ else if(WIFSTOPPED(status)){
+ proc_id = pid_to_processor_id(pid);
+ sig = WSTOPSIG(status);
+#ifdef UML_CONFIG_SYSCALL_DEBUG
+ if(signal_index[proc_id] == 1024){
+ signal_index[proc_id] = 0;
+ last_index = 1023;
+ }
+ else last_index = signal_index[proc_id] - 1;
+ if(((sig == SIGPROF) || (sig == SIGVTALRM) ||
+ (sig == SIGALRM)) &&
+ (signal_record[proc_id][last_index].signal == sig)&&
+ (signal_record[proc_id][last_index].pid == pid))
+ signal_index[proc_id] = last_index;
+ signal_record[proc_id][signal_index[proc_id]].pid = pid;
+ gettimeofday(&signal_record[proc_id][signal_index[proc_id]].time, NULL);
+ eip = ptrace(PTRACE_PEEKUSR, pid, PT_IP_OFFSET, 0);
+ signal_record[proc_id][signal_index[proc_id]].addr = eip;
+ signal_record[proc_id][signal_index[proc_id]++].signal = sig;
+#endif
+ if(proc_id == -1){
+ sleeping_process_signal(pid, sig);
+ continue;
+ }
+
+ task = cpu_tasks[proc_id].task;
+ tracing = is_tracing(task);
+ old_tracing = tracing;
+
+ /* Assume: no syscall, when coming from user */
+ if ( tracing )
+ do_sigtrap(task);
+
+ switch(sig){
+ case SIGUSR1:
+ sig = 0;
+ op = do_proc_op(task, proc_id);
+ switch(op){
+ /*
+ * This is called when entering user mode; after
+ * this, we start intercepting syscalls.
+ *
+ * In fact, a process is started in kernel mode,
+ * so with is_tracing() == 0 (and that is reset
+ * when executing syscalls, since UML kernel has
+ * the right to do syscalls);
+ */
+ case OP_TRACE_ON:
+ arch_leave_kernel(task, pid);
+ tracing = 1;
+ break;
+ case OP_REBOOT:
+ case OP_HALT:
+ unmap_physmem();
+ kmalloc_ok = 0;
+ os_kill_ptraced_process(pid, 0);
+ /* Now let's reap remaining zombies */
+ errno = 0;
+ do {
+ waitpid(-1, &status,
+ WUNTRACED);
+ } while (errno != ECHILD);
+ return(op == OP_REBOOT);
+ case OP_NONE:
+ printf("Detaching pid %d\n", pid);
+ detach(pid, SIGSTOP);
+ continue;
+ default:
+ break;
+ }
+ /* OP_EXEC switches host processes on us,
+ * we want to continue the new one.
+ */
+ pid = cpu_tasks[proc_id].pid;
+ break;
+ case (SIGTRAP + 0x80):
+ if(!tracing && (debugger_pid != -1)){
+ child_signal(pid, status & 0x7fff);
+ continue;
+ }
+ tracing = 0;
+ /* local_using_sysemu has been already set
+ * below, since if we are here, is_tracing() on
+ * the traced task was 1, i.e. the process had
+ * already run through one iteration of the
+ * loop which executed a OP_TRACE_ON request.*/
+ do_syscall(task, pid, local_using_sysemu);
+ sig = SIGUSR2;
+ break;
+ case SIGTRAP:
+ if(!tracing && (debugger_pid != -1)){
+ child_signal(pid, status);
+ continue;
+ }
+ tracing = 0;
+ break;
+ case SIGPROF:
+ if(tracing) sig = 0;
+ break;
+ case SIGCHLD:
+ case SIGHUP:
+ sig = 0;
+ break;
+ case SIGSEGV:
+ case SIGIO:
+ case SIGALRM:
+ case SIGVTALRM:
+ case SIGFPE:
+ case SIGBUS:
+ case SIGILL:
+ case SIGWINCH:
+
+ default:
+ tracing = 0;
+ break;
+ }
+ set_tracing(task, tracing);
+
+ if(!tracing && old_tracing)
+ arch_enter_kernel(task, pid);
+
+ if(!tracing && (debugger_pid != -1) && (sig != 0) &&
+ (sig != SIGALRM) && (sig != SIGVTALRM) &&
+ (sig != SIGSEGV) && (sig != SIGTRAP) &&
+ (sig != SIGUSR2) && (sig != SIGIO) &&
+ (sig != SIGFPE)){
+ child_signal(pid, status);
+ continue;
+ }
+
+ local_using_sysemu = get_using_sysemu();
+
+ if(tracing)
+ cont_type = SELECT_PTRACE_OPERATION(local_using_sysemu,
+ singlestepping(task));
+ else if((debugger_pid != -1) && strace)
+ cont_type = PTRACE_SYSCALL;
+ else
+ cont_type = PTRACE_CONT;
+
+ if(ptrace(cont_type, pid, 0, sig) != 0){
+ tracer_panic("ptrace failed to continue "
+ "process - errno = %d\n",
+ errno);
+ }
+ }
+ }
+ return(0);
+}
+
+static int __init uml_debug_setup(char *line, int *add)
+{
+ char *next;
+
+ debug = 1;
+ *add = 0;
+ if(*line != '=') return(0);
+ line++;
+
+ while(line != NULL){
+ next = strchr(line, ',');
+ if(next) *next++ = '\0';
+
+ if(!strcmp(line, "go")) debug_stop = 0;
+ else if(!strcmp(line, "parent")) debug_parent = 1;
+ else printf("Unknown debug option : '%s'\n", line);
+
+ line = next;
+ }
+ return(0);
+}
+
+__uml_setup("debug", uml_debug_setup,
+"debug\n"
+" Starts up the kernel under the control of gdb. See the \n"
+" kernel debugging tutorial and the debugging session pages\n"
+" at http://user-mode-linux.sourceforge.net/ for more information.\n\n"
+);
+
+static int __init uml_debugtrace_setup(char *line, int *add)
+{
+ debug_trace = 1;
+ return 0;
+}
+__uml_setup("debugtrace", uml_debugtrace_setup,
+"debugtrace\n"
+" Causes the tracing thread to pause until it is attached by a\n"
+" debugger and continued. This is mostly for debugging crashes\n"
+" early during boot, and should be pretty much obsoleted by\n"
+" the debug switch.\n\n"
+);
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only. This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-file-style: "linux"
+ * End:
+ */