Re: ftp(1): pledge smaller subset in SMALL version

[Jérémie Courrèges-Anglas]

j...@wxcvbn.org (Jérémie Courrèges-Anglas) writes:

> Frederic Nowak  writes:
>
>> Hi there,
>
> Hi,
>
>> at the moment ftp pledges "proc exec" in its SMALL version, but not
>> otherwise. This seems wrong, because the SMALL version does not support
>> interactive mode (which needs "proc exec" for e.g. the page command),
>> while the !SMALL version does.
>> The patch below switches the two pledges, so that the SMALL version
>> pledges a smaller subset.
>
> I can confirm that it is possible to have ftp(1) killed by pledge in the
> non-SMALL build.
>
> The diff looks correct and I doubt that there's a risk for it to
> introduce further regressions.  However there's a popen() call in
> recvrequest() that could trigger pledge protection.  If anyone wants to
> make sense of that code, be my guest...

Since no one stepped up to investigate this issue, I committed your diff
as is.  Thanks!

-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

Re: ftp(1): pledge smaller subset in SMALL version

[Jérémie Courrèges-Anglas]

Frederic Nowak  writes:

> Hi there,

Hi,

> at the moment ftp pledges "proc exec" in its SMALL version, but not
> otherwise. This seems wrong, because the SMALL version does not support
> interactive mode (which needs "proc exec" for e.g. the page command),
> while the !SMALL version does.
> The patch below switches the two pledges, so that the SMALL version
> pledges a smaller subset.

I can confirm that it is possible to have ftp(1) killed by pledge in the
non-SMALL build.

The diff looks correct and I doubt that there's a risk for it to
introduce further regressions.  However there's a popen() call in
recvrequest() that could trigger pledge protection.  If anyone wants to
make sense of that code, be my guest...

-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

ftp(1): pledge smaller subset in SMALL version

[Frederic Nowak]

Hi there,

at the moment ftp pledges "proc exec" in its SMALL version, but not
otherwise. This seems wrong, because the SMALL version does not support
interactive mode (which needs "proc exec" for e.g. the page command),
while the !SMALL version does.
The patch below switches the two pledges, so that the SMALL version
pledges a smaller subset.

Cheers,
Frederic


Index: usr.bin/ftp/main.c
===
RCS file: /cvs/src/usr.bin/ftp/main.c,v
retrieving revision 1.104
diff -u -p -r1.104 main.c
--- usr.bin/ftp/main.c  18 Oct 2015 03:39:37 -  1.104
+++ usr.bin/ftp/main.c  2 Nov 2015 05:29:39 -
@@ -444,21 +444,21 @@ main(volatile int argc, char *argv[])
if (isurl(argv[0])) {
if (pipeout) {
 #ifndef SMALL
-   if (pledge("stdio rpath dns tty inet fattr",
+   if (pledge("stdio rpath dns tty inet proc exec 
fattr",
NULL) == -1)
err(1, "pledge");
 #else
-   if (pledge("stdio rpath dns tty inet proc exec 
fattr",
+   if (pledge("stdio rpath dns tty inet fattr",
NULL) == -1)
err(1, "pledge");
 #endif
} else {
 #ifndef SMALL
-   if (pledge("stdio rpath wpath cpath dns tty 
inet fattr",
+   if (pledge("stdio rpath wpath cpath dns tty 
inet proc exec fattr",
NULL) == -1)
err(1, "pledge");
 #else
-   if (pledge("stdio rpath wpath cpath dns tty 
inet proc exec fattr",
+   if (pledge("stdio rpath wpath cpath dns tty 
inet fattr",
NULL) == -1)
err(1, "pledge");
 #endif