hush: reinstate `cmd` handling for NOMMU (with fat big warning).

hush: fix a case where none of pipe members could be started
  because of fork failure
hush: rename functions: xxx_real -> xxx
hush: try to add a bit more of vfork-friendliness
hush: add rudimentary design docs
hush: add TODO (newly discovered bug with globbing)

4 files changed, 163 insertions, 92 deletions

diff --git a/shell/hush.c b/shell/hush.c
index a754076..f3ea1a2 100644
--- a/shell/hush.c
+++ b/shell/hush.c
@@ -84,12 +84,24 @@
 #include "busybox.h" /* for APPLET_IS_NOFORK/NOEXEC */
 
 
-#if !BB_MMU
-/* A bit drastic. Can allow some simpler commands
- * by analysing command in generate_stream_from_list()
- */
-#undef ENABLE_HUSH_TICK
-#define ENABLE_HUSH_TICK 0
+#if !BB_MMU && ENABLE_HUSH_TICK
+//#undef ENABLE_HUSH_TICK
+//#define ENABLE_HUSH_TICK 0
+#warning On NOMMU, hush command substitution is dangerous.
+#warning Dont use it for commands which produce lots of output.
+#warning For more info see shell/hush.c, generate_stream_from_list().
+#endif
+
+#if !BB_MMU && ENABLE_HUSH_JOB
+#undef ENABLE_HUSH_JOB
+#define ENABLE_HUSH_JOB 0
+#endif
+
+#if !ENABLE_HUSH_INTERACTIVE
+#undef ENABLE_FEATURE_EDITING
+#define ENABLE_FEATURE_EDITING 0
+#undef ENABLE_FEATURE_EDITING_FANCY_PROMPT
+#define ENABLE_FEATURE_EDITING_FANCY_PROMPT 0
 #endif
 
 
@@ -176,13 +188,6 @@ void xxfree(void *ptr)
 #endif
 
 
-#if !ENABLE_HUSH_INTERACTIVE
-#undef ENABLE_FEATURE_EDITING
-#define ENABLE_FEATURE_EDITING 0
-#undef ENABLE_FEATURE_EDITING_FANCY_PROMPT
-#define ENABLE_FEATURE_EDITING_FANCY_PROMPT 0
-#endif
-
 #define SPECIAL_VAR_SYMBOL   3
 
 #define PARSEFLAG_EXIT_FROM_LOOP 1
@@ -508,10 +513,10 @@ static int free_pipe_list(struct pipe *head, int indent);
 static int free_pipe(struct pipe *pi, int indent);
 /*  really run the final data structures: */
 static int setup_redirects(struct child_prog *prog, int squirrel[]);
-static int run_list_real(struct pipe *pi);
+static int run_list(struct pipe *pi);
 static void pseudo_exec_argv(char **argv) ATTRIBUTE_NORETURN;
 static void pseudo_exec(struct child_prog *child) ATTRIBUTE_NORETURN;
-static int run_pipe_real(struct pipe *pi);
+static int run_pipe(struct pipe *pi);
 /*   extended glob support: */
 static char **globhack(const char *src, char **strings);
 static int glob_needed(const char *s);
@@ -1416,7 +1421,7 @@ static void restore_redirects(int squirrel[])
 	}
 }
 
-/* Called after [v]fork() in run_pipe_real(), or from builtin_exec().
+/* Called after [v]fork() in run_pipe(), or from builtin_exec().
  * Never returns.
  * XXX no exit() here.  If you don't exec, use _exit instead.
  * The at_exit handlers apparently confuse the calling process,
@@ -1438,9 +1443,8 @@ static void pseudo_exec_argv(char **argv)
 	/* If a variable is assigned in a forest, and nobody listens,
 	 * was it ever really set?
 	 */
-	if (argv[0] == NULL) {
+	if (!argv[0])
 		_exit(EXIT_SUCCESS);
-	}
 
 	argv = expand_strvec_to_strvec(argv);
 
@@ -1484,15 +1488,15 @@ static void pseudo_exec_argv(char **argv)
 	_exit(1);
 }
 
-/* Called after [v]fork() in run_pipe_real()
+/* Called after [v]fork() in run_pipe()
  */
 static void pseudo_exec(struct child_prog *child)
 {
 // FIXME: buggy wrt NOMMU! Must not modify any global data
-// until it does exec/_exit, but currently it does.
-	if (child->argv) {
+// until it does exec/_exit, but currently it does
+// (puts malloc'ed stuff into environment)
+	if (child->argv)
 		pseudo_exec_argv(child->argv);
-	}
 
 	if (child->group) {
 #if !BB_MMU
@@ -1501,11 +1505,12 @@ static void pseudo_exec(struct child_prog *child)
 		int rcode;
 
 #if ENABLE_HUSH_INTERACTIVE
-		debug_printf_exec("pseudo_exec: setting interactive_fd=0\n");
-		interactive_fd = 0;    /* crucial!!!! */
+// run_list_level now takes care of it?
+//		debug_printf_exec("pseudo_exec: setting interactive_fd=0\n");
+//		interactive_fd = 0;    /* crucial!!!! */
 #endif
-		debug_printf_exec("pseudo_exec: run_list_real\n");
-		rcode = run_list_real(child->group);
+		debug_printf_exec("pseudo_exec: run_list\n");
+		rcode = run_list(child->group);
 		/* OK to leak memory by not calling free_pipe_list,
 		 * since this process is about to exit */
 		_exit(rcode);
@@ -1672,7 +1677,7 @@ static int checkjobs(struct pipe* fg_pipe)
 					if (dead) {
 						fg_pipe->progs[i].pid = 0;
 						fg_pipe->running_progs--;
-						if (i == fg_pipe->num_progs-1)
+						if (i == fg_pipe->num_progs - 1)
 							/* last process gives overall exitstatus */
 							rcode = WEXITSTATUS(status);
 					} else {
@@ -1753,13 +1758,13 @@ static int checkjobs_and_fg_shell(struct pipe* fg_pipe)
 }
 #endif
 
-/* run_pipe_real() starts all the jobs, but doesn't wait for anything
+/* run_pipe() starts all the jobs, but doesn't wait for anything
  * to finish.  See checkjobs().
  *
  * return code is normally -1, when the caller has to wait for children
  * to finish to determine the exit status of the pipe.  If the pipe
  * is a simple builtin command, however, the action is done by the
- * time run_pipe_real returns, and the exit code is provided as the
+ * time run_pipe returns, and the exit code is provided as the
  * return value.
  *
  * The input of the pipe is always stdin, the output is always
@@ -1772,7 +1777,7 @@ static int checkjobs_and_fg_shell(struct pipe* fg_pipe)
  * Returns -1 only if started some children. IOW: we have to
  * mask out retvals of builtins etc with 0xff!
  */
-static int run_pipe_real(struct pipe *pi)
+static int run_pipe(struct pipe *pi)
 {
 	int i;
 	int nextin;
@@ -1785,7 +1790,7 @@ static int run_pipe_real(struct pipe *pi)
 	int rcode;
 	const int single_fg = (pi->num_progs == 1 && pi->followup != PIPE_BG);
 
-	debug_printf_exec("run_pipe_real start: single_fg=%d\n", single_fg);
+	debug_printf_exec("run_pipe start: single_fg=%d\n", single_fg);
 
 #if ENABLE_HUSH_JOB
 	pi->pgrp = -1;
@@ -1801,11 +1806,11 @@ static int run_pipe_real(struct pipe *pi)
 	if (single_fg && child->group && child->subshell == 0) {
 		debug_printf("non-subshell grouping\n");
 		setup_redirects(child, squirrel);
-		debug_printf_exec(": run_list_real\n");
-		rcode = run_list_real(child->group);
+		debug_printf_exec(": run_list\n");
+		rcode = run_list(child->group) & 0xff;
 		restore_redirects(squirrel);
-		debug_printf_exec("run_pipe_real return %d\n", rcode);
-		return rcode; // do we need to add '... & 0xff' ?
+		debug_printf_exec("run_pipe return %d\n", rcode);
+		return rcode;
 	}
 
 	if (single_fg && child->argv != NULL) {
@@ -1847,7 +1852,7 @@ static int run_pipe_real(struct pipe *pi)
 				rcode = x->function(argv_expanded) & 0xff;
 				free(argv_expanded);
 				restore_redirects(squirrel);
-				debug_printf_exec("run_pipe_real return %d\n", rcode);
+				debug_printf_exec("run_pipe return %d\n", rcode);
 				return rcode;
 			}
 		}
@@ -1864,7 +1869,7 @@ static int run_pipe_real(struct pipe *pi)
 				rcode = run_nofork_applet_prime(&nofork_save, a, argv_expanded) & 0xff;
 				free(argv_expanded);
 				restore_redirects(squirrel);
-				debug_printf_exec("run_pipe_real return %d\n", rcode);
+				debug_printf_exec("run_pipe return %d\n", rcode);
 				return rcode;
 			}
 		}
@@ -1892,24 +1897,22 @@ static int run_pipe_real(struct pipe *pi)
 		if ((i + 1) < pi->num_progs)
 			xpipe(pipefds);
 
-#if BB_MMU
-		child->pid = fork();
-#else
-		child->pid = vfork();
-#endif
+		child->pid = BB_MMU ? fork() : vfork();
 		if (!child->pid) { /* child */
 #if ENABLE_HUSH_JOB
 			/* Every child adds itself to new process group
-			 * with pgid == pid of first child in pipe */
+			 * with pgid == pid_of_first_child_in_pipe */
 			if (run_list_level == 1 && interactive_fd) {
+				pid_t pgrp;
 				/* Don't do pgrp restore anymore on fatal signals */
 				set_fatal_sighandler(SIG_DFL);
-				if (pi->pgrp < 0) /* true for 1st process only */
-					pi->pgrp = getpid();
-				if (setpgid(0, pi->pgrp) == 0 && pi->followup != PIPE_BG) {
+				pgrp = pi->pgrp;
+				if (pgrp < 0) /* true for 1st process only */
+					pgrp = getpid();
+				if (setpgid(0, pgrp) == 0 && pi->followup != PIPE_BG) {
 					/* We do it in *every* child, not just first,
 					 * to avoid races */
-					tcsetpgrp(interactive_fd, pi->pgrp);
+					tcsetpgrp(interactive_fd, pgrp);
 				}
 			}
 #endif
@@ -1930,7 +1933,7 @@ static int run_pipe_real(struct pipe *pi)
 
 		if (child->pid < 0) { /* [v]fork failed */
 			/* Clearly indicate, was it fork or vfork */
-			bb_perror_msg(BB_MMU ? "cannot fork" : "cannot vfork");
+			bb_perror_msg(BB_MMU ? "fork" : "vfork");
 		} else {
 			pi->running_progs++;
 #if ENABLE_HUSH_JOB
@@ -1948,7 +1951,12 @@ static int run_pipe_real(struct pipe *pi)
 		nextin = pipefds[0];
 	}
 
-	debug_printf_exec("run_pipe_real return -1\n");
+	if (!pi->running_progs) {
+		debug_printf_exec("run_pipe return 1 (all forks failed, no children)\n");
+		return 1;
+	}
+
+	debug_printf_exec("run_pipe return -1 (%u children started)\n", pi->running_progs);
 	return -1;
 }
 
@@ -2017,7 +2025,7 @@ static void debug_print_tree(struct pipe *pi, int lvl)
 
 /* NB: called by pseudo_exec, and therefore must not modify any
  * global data until exec/_exit (we can be a child after vfork!) */
-static int run_list_real(struct pipe *pi)
+static int run_list(struct pipe *pi)
 {
 	struct pipe *rpipe;
 #if ENABLE_HUSH_LOOPS
@@ -2026,7 +2034,6 @@ static int run_list_real(struct pipe *pi)
 	char **for_list = NULL;
 	int flag_rep = 0;
 #endif
-	int save_num_progs;
 	int flag_skip = 1;
 	int rcode = 0; /* probably for gcc only */
 	int flag_restore = 0;
@@ -2038,7 +2045,7 @@ static int run_list_real(struct pipe *pi)
 	reserved_style rword;
 	reserved_style skip_more_for_this_rword = RES_XXXX;
 
-	debug_printf_exec("run_list_real start lvl %d\n", run_list_level + 1);
+	debug_printf_exec("run_list start lvl %d\n", run_list_level + 1);
 
 #if ENABLE_HUSH_LOOPS
 	/* check syntax for "for" */
@@ -2047,7 +2054,7 @@ static int run_list_real(struct pipe *pi)
 		 && (rpipe->next == NULL)
 		) {
 			syntax("malformed for"); /* no IN or no commands after IN */
-			debug_printf_exec("run_list_real lvl %d return 1\n", run_list_level);
+			debug_printf_exec("run_list lvl %d return 1\n", run_list_level);
 			return 1;
 		}
 		if ((rpipe->res_word == RES_IN && rpipe->next->res_word == RES_IN && rpipe->next->progs[0].argv != NULL)
@@ -2055,7 +2062,7 @@ static int run_list_real(struct pipe *pi)
 		) {
 			/* TODO: what is tested in the first condition? */
 			syntax("malformed for"); /* 2nd condition: not followed by IN */
-			debug_printf_exec("run_list_real lvl %d return 1\n", run_list_level);
+			debug_printf_exec("run_list lvl %d return 1\n", run_list_level);
 			return 1;
 		}
 	}
@@ -2103,9 +2110,10 @@ static int run_list_real(struct pipe *pi)
 		signal_SA_RESTART(SIGTSTP, handler_ctrl_z);
 		signal(SIGINT, handler_ctrl_c);
 	}
-#endif
+#endif /* JOB */
 
 	for (; pi; pi = flag_restore ? rpipe : pi->next) {
+//why?		int save_num_progs;
 		rword = pi->res_word;
 #if ENABLE_HUSH_LOOPS
 		if (rword == RES_WHILE || rword == RES_UNTIL || rword == RES_FOR) {
@@ -2178,12 +2186,12 @@ static int run_list_real(struct pipe *pi)
 #endif
 		if (pi->num_progs == 0)
 			continue;
-		save_num_progs = pi->num_progs; /* save number of programs */
-		debug_printf_exec(": run_pipe_real with %d members\n", pi->num_progs);
-		rcode = run_pipe_real(pi);
+//why?		save_num_progs = pi->num_progs;
+		debug_printf_exec(": run_pipe with %d members\n", pi->num_progs);
+		rcode = run_pipe(pi);
 		if (rcode != -1) {
 			/* We only ran a builtin: rcode was set by the return value
-			 * of run_pipe_real(), and we don't need to wait for anything. */
+			 * of run_pipe(), and we don't need to wait for anything. */
 		} else if (pi->followup == PIPE_BG) {
 			/* What does bash do with attempts to background builtins? */
 			/* Even bash 3.2 doesn't do that well with nested bg:
@@ -2196,7 +2204,6 @@ static int run_list_real(struct pipe *pi)
 			rcode = EXIT_SUCCESS;
 		} else {
 #if ENABLE_HUSH_JOB
-			/* Paranoia, just "interactive_fd" should be enough? */
 			if (run_list_level == 1 && interactive_fd) {
 				/* waits for completion, then fg's main shell */
 				rcode = checkjobs_and_fg_shell(pi);
@@ -2210,7 +2217,7 @@ static int run_list_real(struct pipe *pi)
 		}
 		debug_printf_exec(": setting last_return_code=%d\n", rcode);
 		last_return_code = rcode;
-		pi->num_progs = save_num_progs; /* restore number of programs */
+//why?		pi->num_progs = save_num_progs;
 #if ENABLE_HUSH_IF
 		if (rword == RES_IF || rword == RES_ELIF)
 			next_if_code = rcode;  /* can be overwritten a number of times */
@@ -2241,7 +2248,7 @@ static int run_list_real(struct pipe *pi)
 		signal(SIGINT, SIG_IGN);
 	}
 #endif
-	debug_printf_exec("run_list_real lvl %d return %d\n", run_list_level + 1, rcode);
+	debug_printf_exec("run_list lvl %d return %d\n", run_list_level + 1, rcode);
 	return rcode;
 }
 
@@ -2315,19 +2322,19 @@ static int free_pipe_list(struct pipe *head, int indent)
 }
 
 /* Select which version we will use */
-static int run_list(struct pipe *pi)
+static int run_and_free_list(struct pipe *pi)
 {
 	int rcode = 0;
-	debug_printf_exec("run_list entered\n");
-	if (fake_mode == 0) {
-		debug_printf_exec(": run_list_real with %d members\n", pi->num_progs);
-		rcode = run_list_real(pi);
+	debug_printf_exec("run_and_free_list entered\n");
+	if (!fake_mode) {
+		debug_printf_exec(": run_list with %d members\n", pi->num_progs);
+		rcode = run_list(pi);
 	}
 	/* free_pipe_list has the side effect of clearing memory.
-	 * In the long run that function can be merged with run_list_real,
+	 * In the long run that function can be merged with run_list,
 	 * but doing that now would hobble the debugging effort. */
-	free_pipe_list(pi, 0);
-	debug_printf_exec("run_list return %d\n", rcode);
+	free_pipe_list(pi, /* indent: */ 0);
+	debug_printf_exec("run_nad_free_list return %d\n", rcode);
 	return rcode;
 }
 
@@ -3221,15 +3228,17 @@ static FILE *generate_stream_from_list(struct pipe *head)
 	int pid, channel[2];
 
 	xpipe(channel);
-	pid = fork();
-	if (pid < 0) {
-		bb_perror_msg_and_die("fork");
-	} else if (pid == 0) {
+/* *** NOMMU WARNING *** */
+/* By using vfork here, we suspend parent till child exits or execs.
+ * If child will not do it before it fills the pipe, it can block forever
+ * in write(STDOUT_FILENO), and parent (shell) will be also stuck.
+ */
+	pid = BB_MMU ? fork() : vfork();
+	if (pid < 0)
+		bb_perror_msg_and_die(BB_MMU ? "fork" : "vfork");
+	if (pid == 0) { /* child */
 		close(channel[0]);
-		if (channel[1] != 1) {
-			dup2(channel[1], 1);
-			close(channel[1]);
-		}
+		xmove_fd(channel[1], 1);
 		/* Prevent it from trying to handle ctrl-z etc */
 #if ENABLE_HUSH_JOB
 		run_list_level = 1;
@@ -3241,11 +3250,12 @@ static FILE *generate_stream_from_list(struct pipe *head)
 		 * everywhere outside actual command execution. */
 		/*set_jobctrl_sighandler(SIG_IGN);*/
 		set_misc_sighandler(SIG_DFL);
-		_exit(run_list_real(head));   /* leaks memory */
+		_exit(run_list(head));   /* leaks memory */
 	}
 	close(channel[1]);
 	pf = fdopen(channel[0], "r");
 	return pf;
+	/* head is freed by the caller */
 }
 
 /* Return code is exit status of the process that is run. */
@@ -3269,7 +3279,8 @@ static int process_command_subs(o_string *dest, struct p_context *ctx,
 	b_free(&result);
 
 	p = generate_stream_from_list(inner.list_head);
-	if (p == NULL) return 1;
+	if (p == NULL)
+		return 1;
 	close_on_exec_on(fileno(p));
 	setup_file_in_str(&pipe_str, p);
 
@@ -3294,7 +3305,7 @@ static int process_command_subs(o_string *dest, struct p_context *ctx,
 	 * at the same time.  That would be a lot of work, and contrary
 	 * to the KISS philosophy of this program. */
 	retcode = fclose(p);
-	free_pipe_list(inner.list_head, 0);
+	free_pipe_list(inner.list_head, /* indent: */ 0);
 	debug_printf("closed FILE from child, retcode=%d\n", retcode);
 	return retcode;
 }
@@ -3674,8 +3685,8 @@ static int parse_and_run_stream(struct in_str *inp, int parse_flag)
 			done_word(&temp, &ctx);
 			done_pipe(&ctx, PIPE_SEQ);
 			debug_print_tree(ctx.list_head, 0);
-			debug_printf_exec("parse_stream_outer: run_list\n");
-			run_list(ctx.list_head);
+			debug_printf_exec("parse_stream_outer: run_and_free_list\n");
+			run_and_free_list(ctx.list_head);
 		} else {
 			if (ctx.old_flag != 0) {
 				free(ctx.stack);
@@ -3684,7 +3695,7 @@ static int parse_and_run_stream(struct in_str *inp, int parse_flag)
 			temp.nonnull = 0;
 			temp.o_quote = 0;
 			inp->p = NULL;
-			free_pipe_list(ctx.list_head, 0);
+			free_pipe_list(ctx.list_head, /* indent: */ 0);
 		}
 		b_free(&temp);
 	} while (rcode != -1 && !(parse_flag & PARSEFLAG_EXIT_FROM_LOOP));   /* loop on syntax errors, return on EOF */
@@ -3898,15 +3909,14 @@ int hush_main(int argc, char **argv)
 
 	if (argv[optind] == NULL) {
 		opt = parse_and_run_file(stdin);
-		goto final_return;
+	} else {
+		debug_printf("\nrunning script '%s'\n", argv[optind]);
+		global_argv = argv + optind;
+		global_argc = argc - optind;
+		input = xfopen(argv[optind], "r");
+		opt = parse_and_run_file(input);
 	}
 
-	debug_printf("\nrunning script '%s'\n", argv[optind]);
-	global_argv = argv + optind;
-	global_argc = argc - optind;
-	input = xfopen(argv[optind], "r");
-	opt = parse_and_run_file(input);
-
  final_return:
 
 #if ENABLE_FEATURE_CLEAN_UP
diff --git a/shell/hush_doc.txt b/shell/hush_doc.txt
new file mode 100644
index 0000000..a3ead59
--- /dev/null
+++ b/shell/hush_doc.txt
@@ -0,0 +1,39 @@
+	This is how hush runs commands:
+
+/* callsite: process_command_subs */
+generate_stream_from_list(struct pipe *head) - handles `cmds`
+  create UNIX pipe
+  [v]fork
+  child:
+  redirect pipe output to stdout
+  _exit(run_list(head));   /* leaks memory */
+  parent:
+  return UNIX pipe's output fd
+  /* head is freed by the caller */
+
+/* callsite: parse_and_run_stream */
+run_and_free_list(struct pipe *)
+  run_list(struct pipe *)
+  free_pipe_list(struct pipe *)
+
+/* callsites: generate_stream_from_list, run_and_free_list, pseudo_exec, run_pipe */
+run_list(struct pipe *) - handles "cmd; cmd2 && cmd3", while/for/do loops
+  run_pipe - for every pipe in list
+
+/* callsite: run_list */
+run_pipe - runs "cmd1 | cmd2 | cmd3 [&]"
+  run_list - used if only one cmd and it is of the form "{ cmd4; cmd5 && cmd6; }"
+  forks for every cmd if more than one cmd or if & is there
+  pseudo_exec - runs each "cmdN" (handles builtins etc)
+
+/* callsite: run_pipe_real */
+pseudo_exec - runs "cmd" (handles builtins etc)
+  exec - execs external programs
+  run_list - used if cmdN is "(cmds)" or "{cmds;}"
+  /* problem: putenv's malloced strings into environ -
+  ** with vfork they will leak into parent process
+  */
+  /* problem with ENABLE_FEATURE_SH_STANDALONE:
+  ** run_applet_no_and_exit(a, argv) uses exit - this can interfere
+  ** with vfork - switch to _exit there?
+  */
diff --git a/shell/hush_test/run-all b/shell/hush_test/run-all
index c75d81e..805f75a 100755
--- a/shell/hush_test/run-all
+++ b/shell/hush_test/run-all
@@ -1,6 +1,9 @@
 #!/bin/sh
 
-test -x hush || { echo "No ./hush?!"; exit; }
+test -x hush || {
+    echo "No ./hush?! Perhaps you want to run 'ln -s ../../busybox hush'"
+    exit
+}
 
 PATH="$PWD:$PATH" # for hush and recho/zecho/printenv
 export PATH
diff --git a/shell/hush_test/zbad2 b/shell/hush_test/zbad2
new file mode 100644
index 0000000..c30fa85
--- /dev/null
+++ b/shell/hush_test/zbad2
@@ -0,0 +1,19 @@
+## TODO: fix and add to testsuite
+
+## # bash zbad2
+## ZVAR=z.map
+## *.map
+## # hush zbad2
+## ZVAR=z.map
+## z.map  <====== !!!
+
+## hush does globbing for "VAR=val" too!
+## it should do it only for non-assignments.
+## even if word looks like assignment, it can be non-assignemnt:
+## ZVAR=*.map /bin/echo ZVAR=*.map
+## ^dont_glob           ^glob
+
+>ZVAR=z.map
+ZVAR=*.map /bin/echo ZVAR=*.map
+ZVAR=*.map
+echo "$ZVAR"