On Tue, 3 Jul 2001, Jan Petranek wrote:

> Breakpoint 1, main (argc=3, argv=0xbffffb2c) at setfiles.c:458
> 458 if (items < 2) {
> 2: *context = 0 '\000'
> 1: context = 0x8158a80 ""

Oops. I gave you incorrect guidance. Display 'type' at this breakpoint. Also, at line 472, display 'context'. Each line in file_contexts has a pathname regular expression field, an optional type field, and a security context field. We aren't currently using the optional type field, so immediately after the sscanf call, 'type' points to the context and 'context' points to the empty string. Then, between line 458 and 472, we see that there were only two fields, so we set 'context' to 'type' and clear 'type'. The purpose of the first breakpoint is to see the content of the field immediately after the sscanf, to see if it has the bogus 'i' suffix, which is not actually present in file_contexts.

> As we come to the second breakpoint, the content has read the context
> correctly from the file:
>
> Breakpoint 2, main (argc=3, argv=0xbffffb2c) at setfiles.c:553
> 553 len = strlen(context);
> 2: *context = 115 's'
> 1: context = 0x81586f0 "system_u:object_r:user_netscape_rw_ti"
> (gdb)

No - the context is not right here - it has the 'i' suffix, which is not in file_contexts.

> I am not using MLS, so where does the "i" come frome?
> 1: context = 0x81586f0 "system_u:object_r:user_netscape_rw_ti"

Right - that's the question. The 'i' isn't in the file_contexts file.

--
Stephen D. Smalley, NAI Labs
ssmalley@nai.com





--
You have received this message because you are subscribed to the selinux list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.