Aren
/
NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Now that the shell is protecting its internal fds properly, moving 

them whenever the user tries to step on one, we can change our behaviour
back to what the kernel considers to be that of a well behaved shell
(wrt file descriptor usage).  If our user causes problems, we will soon
move into recalcitrant process territory, but that should normally be
rare.  This should reduce kernel overheads a little.
This commit is contained in:
kre 2017-05-29 22:21:00 +00:00
parent 8ffd1099d3
commit 374c12e6d0
1 changed files with 12 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
bin/sh/redir.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: redir.c,v 1.56 2017/05/18 13:56:58 kre Exp $ */
/* $NetBSD: redir.c,v 1.57 2017/05/29 22:21:00 kre Exp $ */
/*-
* Copyright (c) 1991, 1993
@ -37,7 +37,7 @@
#if 0
static char sccsid[] = "@(#)redir.c 8.2 (Berkeley) 5/4/95";
#else
__RCSID("$NetBSD: redir.c,v 1.56 2017/05/18 13:56:58 kre Exp $");
__RCSID("$NetBSD: redir.c,v 1.57 2017/05/29 22:21:00 kre Exp $");
#endif
#endif /* not lint */
@ -518,7 +518,7 @@ STATIC void
find_big_fd(void)
{
int i, fd;
static int last_start = 6;
static int last_start = 3; /* aim to keep sh fd's under 20 */
if (last_start < 10)
last_start++;
@ -630,6 +630,15 @@ renumber_sh_fd(struct shell_fds *fp)
#define CLOEXEC(fd)
#endif
/*
* if we have had a collision, and the sh fd was a "big" one
* try moving the sh fd base to a higher number (if possible)
* so future sh fds are less likely to be in the user's sights
* (incl this one when moved)
*/
if (fp->fd >= big_sh_fd)
find_big_fd();
to = fcntl(fp->fd, F_DUPFD_CLOEXEC, big_sh_fd);
if (to == -1)
to = fcntl(fp->fd, F_DUPFD_CLOEXEC, big_sh_fd/2);