aboutsummaryrefslogtreecommitdiff
path: root/gdb/infcmd.c
diff options
context:
space:
mode:
authorPedro Alves <palves@redhat.com>2014-07-09 15:59:02 +0100
committerPedro Alves <palves@redhat.com>2014-07-09 15:59:02 +0100
commit7180e04a36d812bbea2c280f2db33a7e8ce6b07b (patch)
tree6faeef54a13b3a432c270d02b1cd60d7b2602916 /gdb/infcmd.c
parent9a9a76082919371f4ceb571f6c9892325b80a2e0 (diff)
downloadgdb-7180e04a36d812bbea2c280f2db33a7e8ce6b07b.zip
gdb-7180e04a36d812bbea2c280f2db33a7e8ce6b07b.tar.gz
gdb-7180e04a36d812bbea2c280f2db33a7e8ce6b07b.tar.bz2
Fix "attach" command vs user input race
On async targets, a synchronous attach is done like this: #1 - target_attach is called (PTRACE_ATTACH is issued) #2 - a continuation is installed #3 - we go back to the event loop #4 - target reports stop (SIGSTOP), event loop wakes up, and attach continuation is called #5 - among other things, the continuation calls target_terminal_inferior, which removes stdin from the event loop Note that in #3, GDB is still processing user input. If the user is fast enough, e.g., with something like: echo -e "attach PID\nset xxx=1" | gdb ... then the "set" command is processed before the attach completes. We get worse behavior even, if input is a tty and therefore readline/editing is enabled, with e.g.,: (gdb) attach PID\nset xxx=1 we then crash readline/gdb, with: Attaching to program: attach-wait-input, process 14537 readline: readline_callback_read_char() called with no handler! Aborted $ Fix this by calling target_terminal_inferior before #3 above. The test covers both scenarios by running with editing/readline forced to both on and off. gdb/ 2014-07-09 Pedro Alves <palves@redhat.com> * infcmd.c (attach_command_post_wait): Don't call target_terminal_inferior here. (attach_command): Call it here instead. gdb/testsuite/ 2014-07-09 Pedro Alves <palves@redhat.com> * gdb.base/attach-wait-input.exp: New file. * gdb.base/attach-wait-input.c: New file.
Diffstat (limited to 'gdb/infcmd.c')
-rw-r--r--gdb/infcmd.c19
1 files changed, 16 insertions, 3 deletions
diff --git a/gdb/infcmd.c b/gdb/infcmd.c
index 14736a5..021a587 100644
--- a/gdb/infcmd.c
+++ b/gdb/infcmd.c
@@ -2380,9 +2380,6 @@ attach_command_post_wait (char *args, int from_tty, int async_exec)
post_create_inferior (&current_target, from_tty);
- /* Install inferior's terminal modes. */
- target_terminal_inferior ();
-
if (async_exec)
{
/* The user requested an `attach&', so be sure to leave threads
@@ -2498,6 +2495,22 @@ attach_command (char *args, int from_tty)
based on what modes we are starting it with. */
target_terminal_init ();
+ /* Install inferior's terminal modes. This may look like a no-op,
+ as we've just saved them above, however, this does more than
+ restore terminal settings:
+
+ - installs a SIGINT handler that forwards SIGINT to the inferior.
+ Otherwise a Ctrl-C pressed just while waiting for the initial
+ stop would end up as a spurious Quit.
+
+ - removes stdin from the event loop, which we need if attaching
+ in the foreground, otherwise on targets that report an initial
+ stop on attach (which are most) we'd process input/commands
+ while we're in the event loop waiting for that stop. That is,
+ before the attach continuation runs and the command is really
+ finished. */
+ target_terminal_inferior ();
+
/* Set up execution context to know that we should return from
wait_for_inferior as soon as the target reports a stop. */
init_wait_for_inferior ();