From: Peter Stephenson <p.stephenson@samsung.com>
To: Zsh hackers list <zsh-workers@zsh.org>
Subject: Re: "echo | ps -j $(:) | cat | cat | cat" runs components in different process groups
Date: Thu, 12 Apr 2018 17:23:42 +0100 [thread overview]
Message-ID: <20180412172342.52df6b10@camnpupstephen.cam.scsc.local> (raw)
In-Reply-To: <180411151025.ZM19332@torch.brasslantern.com>
On Wed, 11 Apr 2018 15:10:24 -0700
Bart Schaefer <schaefer@brasslantern.com> wrote:
> On Apr 10, 2:59pm, Peter Stephenson wrote:
> } Subject: Re: "echo | ps -j $(:) | cat | cat | cat" runs components
> in diff }
> } + if (pn->pid == jn->gleader) {
> } + jn->gleader = 0;
> } + /* HERE: check this was actually in
> foreground... */ } + attachtty(mypgrp);
> } + }
>
> I think attachtty() will simply fail here if mypgrp is not already in
> the foreground, but I confess not to be 100% certain.
The man page suggests it's SIGTTOU time, though this isn't my natural
territory either.
I think the right test is for STAT_NOSTTY, which is the flag associated
with running in the background. (Paranoid checK: we set this with
"bg" as well as "&".)
Moving the unqueuing inside readoutput() both catches other similar
cases and allows us to continue protecting memory allocation while
allowing signals around the blocking read.
So I think this is a definite improvement, though it's probably not safe
enough to sneak into 5.5.1.
pws
diff --git a/Src/exec.c b/Src/exec.c
index e154d12..65ce348 100644
--- a/Src/exec.c
+++ b/Src/exec.c
@@ -4576,10 +4576,20 @@ readoutput(int in, int qt, int *readerror)
char *buf, *ptr;
int bsiz, c, cnt = 0;
FILE *fin;
+ int q = queue_signal_level();
fin = fdopen(in, "r");
ret = newlinklist();
ptr = buf = (char *) hcalloc(bsiz = 64);
+ /*
+ * We need to be sensitive to SIGCHLD else we can be
+ * stuck forever with important processes unreaped.
+ * The case that triggered this was where the exiting
+ * process is group leader of the foreground process and we need
+ * to reclaim the terminal else ^C doesn't work.
+ */
+ dont_queue_signals();
+ child_unblock();
while ((c = fgetc(fin)) != EOF || errno == EINTR) {
if (c == EOF) {
errno = 0;
@@ -4592,13 +4602,18 @@ readoutput(int in, int qt, int *readerror)
cnt++;
}
if (++cnt >= bsiz) {
- char *pp = (char *) hcalloc(bsiz *= 2);
+ char *pp;
+ queue_signals();
+ pp = (char *) hcalloc(bsiz *= 2);
+ dont_queue_signals();
memcpy(pp, buf, cnt - 1);
ptr = (buf = pp) + cnt - 1;
}
*ptr++ = c;
}
+ child_block();
+ restore_queue_signals(q);
if (readerror)
*readerror = ferror(fin) ? errno : 0;
fclose(fin);
diff --git a/Src/signals.c b/Src/signals.c
index 94f379e..2a6aa3f 100644
--- a/Src/signals.c
+++ b/Src/signals.c
@@ -537,6 +537,19 @@ wait_for_processes(void)
#else
update_process(pn, status);
#endif
+ if (pn->pid == jn->gleader) {
+ jn->gleader = 0;
+ if (!(jn->stat & STAT_NOSTTY)) {
+ /*
+ * This PID was in control of the terminal;
+ * reclaim terminal now it has exited.
+ * It's still possible some future forked
+ * process of this job will become group
+ * leader, however.
+ */
+ attachtty(mypgrp);
+ }
+ }
}
update_job(jn);
} else if (findproc(pid, &jn, &pn, 1)) {
next prev parent reply other threads:[~2018-04-12 16:30 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-03-23 16:16 Stephane Chazelas
2018-03-23 23:36 ` Bart Schaefer
2018-03-24 5:19 ` Bart Schaefer
2018-03-24 8:05 ` Joey Pabalinas
2018-03-24 17:34 ` Stephane Chazelas
2018-03-24 22:23 ` Bart Schaefer
2018-03-25 7:00 ` Stephane Chazelas
2018-03-25 8:26 ` Stephane Chazelas
2018-03-25 7:48 ` Stephane Chazelas
2018-03-25 18:09 ` Stephane Chazelas
2018-03-27 10:17 ` Stephane Chazelas
2018-03-24 22:09 ` Bart Schaefer
2018-04-10 11:45 ` Peter Stephenson
2018-04-10 13:59 ` Peter Stephenson
2018-04-11 22:10 ` Bart Schaefer
2018-04-12 16:23 ` Peter Stephenson [this message]
2018-04-15 16:23 ` Stephane Chazelas
2018-04-15 17:38 ` Bart Schaefer
2018-04-15 18:58 ` Stephane Chazelas
2018-04-17 5:39 ` Bart Schaefer
2018-04-17 9:19 ` Peter Stephenson
2018-04-17 16:09 ` Bart Schaefer
2018-04-17 16:27 ` Bart Schaefer
2018-04-17 16:35 ` Peter Stephenson
2018-04-17 17:52 ` Bart Schaefer
2018-04-19 9:40 ` Peter Stephenson
2018-04-20 9:28 ` Forking earlier for background code Peter Stephenson
2018-04-20 10:01 ` Peter Stephenson
2018-04-20 12:54 ` Vin Shelton
[not found] ` <CGME20180420132008eucas1p1c297624e870975cd892c74254970faab@eucas1p1.samsung.com>
2018-04-20 13:20 ` Peter Stephenson
2018-04-20 16:01 ` Vin Shelton
2018-04-23 8:29 ` Peter Stephenson
2018-05-01 9:23 ` Peter Stephenson
[not found] ` <F3A62E38-24E2-4A62-8E19-F54C9B81E9E5@kba.biglobe.ne.jp>
2018-04-23 13:52 ` "echo | ps -j $(:) | cat | cat | cat" runs components in different process groups Peter Stephenson
2018-04-23 14:03 ` Peter Stephenson
[not found] ` <CGME20180423140859eucas1p2591bf1422614209979d4890383268c37@eucas1p2.samsung.com>
2018-04-23 14:08 ` Peter Stephenson
2018-04-23 15:29 ` Jun T.
2018-04-18 6:29 ` Stephane Chazelas
2018-04-18 16:33 ` Bart Schaefer
2018-04-10 9:53 ` Peter Stephenson
2018-03-25 7:38 ` Stephane Chazelas
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180412172342.52df6b10@camnpupstephen.cam.scsc.local \
--to=p.stephenson@samsung.com \
--cc=zsh-workers@zsh.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://git.vuxu.org/mirror/zsh/
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).