Something like this then?
If it's too much burden to keep these local patches I can drop it, no problem.
Index: bdftopcf.c
===================================================================
RCS file: /cvs/xenocara/app/bdftopcf/bdftopcf.c,v
retrieving revision 1.5
diff -u -p -u -r1.5 bdftopcf.c
--- bdftopcf.c 29 Mar 2018 20:34:30 -0000 1.5
+++ bdftopcf.c 25 Oct 2018 07:00:50 -0000
@@ -39,6 +39,7 @@ from The Open Group.
#include "bdfint.h"
#include "pcf.h"
#include <stdio.h>
+#include <unistd.h>
#include <X11/Xos.h>
int
@@ -158,6 +159,38 @@ main(int argc, char *argv[])
}
argv++;
}
+
+ if (input_name) {
+ if (unveil(input_name, "r") == -1) {
+ fprintf(stderr, "%s: could not unveil %s\n",
+ program_name, input_name);
+ exit(1);
+ }
+ }
+ if (output_name) {
+ if (unveil(output_name, "rwc") == -1) {
+ fprintf(stderr, "%s: could not unveil %s\n",
+ program_name, output_name);
+ exit(1);
+ }
+ if (pledge("stdio rpath wpath cpath", NULL) == -1) {
+ fprintf(stderr, "%s: could not pledge", program_name);
+ exit(1);
+ }
+ }
+ if (input_name && !output_name) {
+ if (pledge("stdio rpath", NULL) == -1) {
+ fprintf(stderr, "%s: could not pledge", program_name);
+ exit(1);
+ }
+ }
+ if (!input_name && !output_name) {
+ if (pledge("stdio", NULL) == -1) {
+ fprintf(stderr, "%s: could not pledge", program_name);
+ exit(1);
+ }
+ }
+
if (input_name) {
input = FontFileOpen(input_name);
if (!input) {
On 10:41 Wed 24 Oct , Theo de Raadt wrote:
> Matthieu Herrb <[email protected]> wrote:
>
> > Generally, I'm not too found of pledging/unveiling random X client
> > programs. There are a lot of "hidden" features in X libraries that
> > will probably break with too strict pledges and/or unveils.
>
> Well eventually we want to see if something can be done about xterm.
> Especially if the lessons learned (I suspect some hoisting will occur)
> can be pushed back upstream, and maybe allow others to apply their
> own system call limiter mechanism. Perhaps not possible...
>
> > Also since this is OpenBSD-specific, it will be difficult to get it
> > upstreams, especially if you don't provide the autoconf goo to make
> > the code still build/work on Linux. And when not upstreaming it
> > creates more burden to merge new versions of the applications.
>
> Well, I doubt it will create too much burden, generally these unveil
> or pledge chunks are a small set of + lines, without changing other
> logic.
>
> Anyways, bdftopcf is not running near a security boundary.
>
