This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: Include putpkt in TRY_CATCH. PR gdb/15275
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Gareth McMullin <gareth at blacksphere dot co dot nz>
- Cc: gdb-patches at sourceware dot org, Pedro Alves <palves at redhat dot com>
- Date: Mon, 25 Mar 2013 20:58:32 +0100
- Subject: Re: Include putpkt in TRY_CATCH. PR gdb/15275
- References: <CAL8qUbrD=fgMP7nE0O8tX=AXifUpQXas25o_4SfK4p79rfoUpw at mail dot gmail dot com> <CAL8qUbqjtBKJFJZ6dPS78Zh8Eb3b33U9JXTmn1pS3Le93xt7Rw at mail dot gmail dot com>
On Thu, 14 Mar 2013 03:02:29 +0100, Gareth McMullin wrote:
> If debugging with a remote connection over a serial link and the link is
> disconnected, say by disconnecting USB serial port, the GDB quit command no
> longer works.
[...]
> 2013-03-14 Gareth McMullin <gareth@blacksphere.co.nz>
>
> PR gdb/15275
> * remote.c (remote_get_trace_status): Include putpkt in TRY_CATCH.
I am OK with the patch.
Ccing also Pedro, the TARGET_CLOSE_ERROR was probably meant to handle it.
I will check it in.
Used this patch to simulate the problem with TCP gdbserver:
(gdb) q
A debugging session is active.
Inferior 1 [process 30674] will be killed.
Quit anyway? (y or n) y
Sending packet: $qTStatus#49...putpkt: write failed: Input/output error.
(gdb) q
->
(gdb) q
A debugging session is active.
Inferior 1 [process 31864] will be killed.
Quit anyway? (y or n) y
Sending packet: $qTStatus#49...qTStatus: putpkt: write failed: Input/output error.
Sending packet: $k#6b...putpkt: write failed: Input/output error.
$ _
#0 net_write_prim (scb=0x41f4b90, buf=0x7fff617af320, count=12) at ser-tcp.c:352
#1 in ser_base_write (scb=0x41f4b90, str=0x7fff617af320 "$qTStatus#49", len=12) at ser-base.c:449
#2 in serial_write (scb=0x41f4b90, str=0x7fff617af320 "$qTStatus#49", len=12) at serial.c:427
#3 in putpkt_binary (buf=0xf617b7 "qTStatus", cnt=8) at remote.c:7183
#4 in putpkt (buf=0xf617b7 "qTStatus") at remote.c:7114
#5 in remote_get_trace_status (ts=0x1ebd860 <trace_status>) at remote.c:10695
#6 in disconnect_tracing (from_tty=1) at tracepoint.c:2224
#7 in quit_command (args=0x0, from_tty=1) at ./cli/cli-cmds.c:325
#--- a/gdb/ser-tcp.c
#+++ b/gdb/ser-tcp.c
#@@ -341,9 +341,19 @@ net_read_prim (struct serial *scb, size_t count)
# return recv (scb->fd, scb->buf, count, 0);
# }
#
#+static volatile int hacked;
#+static void hackusr2(int signo) {
#+ hacked=1;
#+}
#+
# int
# net_write_prim (struct serial *scb, const void *buf, size_t count)
# {
#+signal(SIGUSR2,hackusr2);
#+if (hacked) {
#+ errno=EIO;
#+ return -1;
#+}
# return send (scb->fd, buf, count, 0);
# }
#
Thanks,
Jan
> ---
> diff --git a/gdb/remote.c b/gdb/remote.c
> index 21d86f7..29c64c0 100644
> --- a/gdb/remote.c
> +++ b/gdb/remote.c
> @@ -10695,10 +10695,9 @@ remote_get_trace_status (struct trace_status *ts)
>
> trace_regblock_size = get_remote_arch_state ()->sizeof_g_packet;
>
> - putpkt ("qTStatus");
> -
> TRY_CATCH (ex, RETURN_MASK_ERROR)
> {
> + putpkt ("qTStatus");
> p = remote_get_noisy_reply (&target_buf, &target_buf_size);
> }
> if (ex.reason < 0)