From: Cyril Hrubis <chrubis@suse.cz>
To: ltp@lists.linux.it
Subject: [LTP] [PATCH 4/4] syscalls/add_key03: new test for forging user keyrings
Date: Wed, 11 Oct 2017 15:47:48 +0200 [thread overview]
Message-ID: <20171011134748.GC15968@rei> (raw)
In-Reply-To: <20171010175120.90586-5-ebiggers3@gmail.com>
Hi!
> +static key_serial_t create_keyring(const char *description)
> +{
> + TEST(add_key("keyring", description, NULL, 0,
> + KEY_SPEC_PROCESS_KEYRING));
> + if (TEST_RETURN < 0) {
> + tst_brk(TBROK | TTERRNO,
> + "unable to create keyring '%s'", description);
> + }
> + return TEST_RETURN;
> +}
> +
> +static key_serial_t get_keyring_id(key_serial_t special_id)
> +{
> + TEST(keyctl(KEYCTL_GET_KEYRING_ID, special_id, 1));
> + if (TEST_RETURN < 0) {
> + tst_brk(TBROK | TTERRNO,
> + "unable to get ID of keyring %d", special_id);
> + }
> + return TEST_RETURN;
> +}
> +
> +static void unlink_keyring(key_serial_t id)
> +{
> + TEST(keyctl(KEYCTL_UNLINK, id, KEY_SPEC_PROCESS_KEYRING));
> + if (TEST_RETURN < 0) {
> + tst_brk(TBROK | TTERRNO,
> + "unable to unlink the keyring we created");
> + }
> +}
> +
> +static void do_test(void)
> +{
> + int i;
> +
> + /*
> + * Try with multiple user IDs before reporting success. By chance, some
> + * users may already have an existing user keyring; the bug will not be
> + * reproducible for them.
> + */
> + for (i = 0; i < 10; i++) {
> + char description[32];
> + uid_t uid;
> + key_serial_t fake_user_keyring;
> + key_serial_t fake_user_session_keyring;
> +
> + uid = rand();
> + if (uid == 0)
> + continue;
We have testcases that look for unused uid with this loop:
for (i = 1; i < 1000; i++) {
if (!getpwuid(i))
return i;
}
What about using this instead of doing 10 random tries?
> + sprintf(description, "_uid.%u", uid);
> + fake_user_keyring = create_keyring(description);
> + sprintf(description, "_uid_ses.%u", uid);
> + fake_user_session_keyring = create_keyring(description);
> +
> + TEST(setreuid(uid, 0));
> + if (TEST_RETURN < 0) {
> + tst_brk(TBROK | TTERRNO,
> + "unable to set real uid to %u", uid);
> + }
I guess that we should add SAFE_SETREUID() to the tst_safe_macros.h
library. We do have SAFE_SETRESUID() though, so we may as well use
SAFE_SETRESUID(uid, -1, -1) here.
> + if (fake_user_keyring == get_keyring_id(KEY_SPEC_USER_KEYRING)) {
> + tst_brk(TFAIL,
> + "created user keyring for another user");
> + }
> +
> + if (fake_user_session_keyring ==
> + get_keyring_id(KEY_SPEC_USER_SESSION_KEYRING)) {
> + tst_brk(TFAIL,
> + "created user session keyring for another user");
> + }
> +
> + TEST(setreuid(0, 0));
> + if (TEST_RETURN < 0)
> + tst_brk(TBROK | TTERRNO, "unable to reset real uid");
> + uid++;
> +
> + unlink_keyring(fake_user_keyring);
> + unlink_keyring(fake_user_session_keyring);
> + }
> + tst_res(TPASS, "expectedly could not create another user's keyrings");
> +}
> +
> +static struct tst_test test = {
> + .test_all = do_test,
> + .needs_root = 1,
> +};
> --
> 2.14.2.920.gcf0c67979c-goog
>
>
> --
> Mailing list info: https://lists.linux.it/listinfo/ltp
--
Cyril Hrubis
chrubis@suse.cz
next prev parent reply other threads:[~2017-10-11 13:47 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-10 17:51 [LTP] [PATCH 0/4] ltp: add tests for some recently-fixed keyrings bugs Eric Biggers
2017-10-10 17:51 ` [LTP] [PATCH 1/4] lapi/keyctl.h: add a few missing definitions Eric Biggers
2017-10-10 17:51 ` [LTP] [PATCH 2/4] syscalls/keyctl06: new test for keyring_read() buffer overrun Eric Biggers
2017-10-10 17:51 ` [LTP] [PATCH 3/4] syscalls/keyctl07: new test for oops when reading negative key Eric Biggers
2017-10-10 17:51 ` [LTP] [PATCH 4/4] syscalls/add_key03: new test for forging user keyrings Eric Biggers
2017-10-11 13:47 ` Cyril Hrubis [this message]
2017-10-11 13:53 ` Cyril Hrubis
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=20171011134748.GC15968@rei \
--to=chrubis@suse.cz \
--cc=ltp@lists.linux.it \
/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.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox