search for: cmd_send_start

...R: send ioctl failed with %d: %s\n", ret, strerror(-ret)); + if (ret == -EINVAL && (!is_first_subvol || !is_last_subvol)) + fprintf(stderr, + "Try upgrading your kernel or don''t use -e.\n"); goto out; } if (g_verbose > 0) @@ -435,15 +443,19 @@ int cmd_send_start(int argc, char **argv) u64 root_id; u64 parent_root_id = 0; int full_send = 1; + int new_end_cmd_semantic = 0; memset(&send, 0, sizeof(send)); send.dump_fd = fileno(stdout); - while ((c = getopt(argc, argv, "vc:f:i:p:")) != -1) { + while ((c = getopt(argc, argv, "ve...

Signed-off-by: David Sterba <dsterba@suse.cz> --- cmds-send.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/cmds-send.c b/cmds-send.c index 9b47e70..e37f5d2 100644 --- a/cmds-send.c +++ b/cmds-send.c @@ -427,6 +427,11 @@ int cmd_send_start(int argc, char **argv) memset(&send, 0, sizeof(send)); send.dump_fd = fileno(stdout); + if (isatty(send.dump_fd)) { + fprintf(stderr, "ERROR: not dumping send stream into a terminal, redirect it into a file\n"); + return 1; + } + while ((c = getopt(argc, argv, "vf:i:p:&...

Most fixes are trivial. The one from Alex is fixing a real bug that several users have reported. Alex sent the patch half a year ago and it was not yet integrated. The patch "Use /proc/mounts instead of /etc/mtab" is a repost. The patch "btrfs-receive optionally honors the end-cmd" is a preparation step to allow backup tools to multiplex a single communication stream (e.g. a