stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Jason A. Donenfeld" <Jason@zx2c4.com>
To: Kees Cook <keescook@chromium.org>,
	linux-hardening@vger.kernel.org, linux-kernel@vger.kernel.org,
	PaX Team <pageexec@freemail.hu>
Cc: "Jason A. Donenfeld" <Jason@zx2c4.com>, stable@vger.kernel.org
Subject: [PATCH v4] gcc-plugins: latent_entropy: use /dev/urandom
Date: Wed,  6 Apr 2022 00:28:15 +0200	[thread overview]
Message-ID: <20220405222815.21155-1-Jason@zx2c4.com> (raw)
In-Reply-To: <CAHmME9pV4SdoSyMq4kax3w3Vu1nPxjO3faCZKq8d0RDo8t731g@mail.gmail.com>

While the latent entropy plugin mostly doesn't derive entropy from
get_random_const() for measuring the call graph, when __latent_entropy is
applied to a constant, then it's initialized statically to output from
get_random_const(). In that case, this data is derived from a 64-bit
seed, which means a buffer of 512 bits doesn't really have that amount
of compile-time entropy.

This patch fixes that shortcoming by just buffering chunks of
/dev/urandom output and doling it out as requested.

At the same time, it's important that we don't break the use of
-frandom-seed, for people who want the runtime benefits of the latent
entropy plugin, while still having compile-time determinism. In that
case, we detect whether gcc's set_random_seed() has been called by
making a call to get_random_seed(noinit=true) in the plugin init
function, which is called after set_random_seed() is called but before
anything that calls get_random_seed(noinit=false), and seeing if it's
zero or not. If it's not zero, we're in deterministic mode, and so we
just generate numbers with a basic xorshift prng.

[Note that we don't detect if -frandom-seed is being used using the
 documented local_tick variable, because it's assigned via:
   local_tick = (unsigned) tv.tv_sec * 1000 + tv.tv_usec / 1000;
 which may well overflow and become -1 on its own, and so isn't
 reliable.]

Fixes: 38addce8b600 ("gcc-plugins: Add latent_entropy plugin")
Cc: stable@vger.kernel.org
Cc: PaX Team <pageexec@freemail.hu>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
---
Changes v2->v4:
- v4 is based on v2, not on v3, so:
  * We're keeping the version string change, since Kees is going to
    remove that anyway from everything all in one swoop, so better to
    increment it as is the custom now, and let that patch later change
    the way things are done.
  * The xorshift prng is retained, because why not. It's faster and
    produces less code.
  * The get_random_seed(noinit=true) technique is used to detect whether
    -frandom-seed is being used, rather than local_tick, because of the
    overflow bug in local_tick.
- The size of the buffer is reduced to 256 bytes, which not only is the
  size which the kernel guarantees will never fail due to signals, but
  also more closely fits the histogram of usage, according to an
  allmodconfig.
- Pipacs pointed out that gcc uses gcc_assert()/gcc_unreachable(), rather
  than abort(), so switch to using that.

 scripts/gcc-plugins/latent_entropy_plugin.c | 46 +++++++++++++--------
 1 file changed, 28 insertions(+), 18 deletions(-)

diff --git a/scripts/gcc-plugins/latent_entropy_plugin.c b/scripts/gcc-plugins/latent_entropy_plugin.c
index 589454bce930..0639fa4d48bd 100644
--- a/scripts/gcc-plugins/latent_entropy_plugin.c
+++ b/scripts/gcc-plugins/latent_entropy_plugin.c
@@ -82,29 +82,35 @@ __visible int plugin_is_GPL_compatible;
 static GTY(()) tree latent_entropy_decl;
 
 static struct plugin_info latent_entropy_plugin_info = {
-	.version	= "201606141920vanilla",
+	.version	= "202203311920vanilla",
 	.help		= "disable\tturn off latent entropy instrumentation\n",
 };
 
-static unsigned HOST_WIDE_INT seed;
-/*
- * get_random_seed() (this is a GCC function) generates the seed.
- * This is a simple random generator without any cryptographic security because
- * the entropy doesn't come from here.
- */
+static unsigned HOST_WIDE_INT deterministic_seed;
+static unsigned HOST_WIDE_INT rnd_buf[32];
+static size_t rnd_idx = ARRAY_SIZE(rnd_buf);
+static int urandom_fd = -1;
+
 static unsigned HOST_WIDE_INT get_random_const(void)
 {
-	unsigned int i;
-	unsigned HOST_WIDE_INT ret = 0;
-
-	for (i = 0; i < 8 * sizeof(ret); i++) {
-		ret = (ret << 1) | (seed & 1);
-		seed >>= 1;
-		if (ret & 1)
-			seed ^= 0xD800000000000000ULL;
+	if (deterministic_seed) {
+		unsigned HOST_WIDE_INT w = deterministic_seed;
+		w ^= w << 13;
+		w ^= w >> 7;
+		w ^= w << 17;
+		deterministic_seed = w;
+		return deterministic_seed;
 	}
 
-	return ret;
+	if (urandom_fd < 0) {
+		urandom_fd = open("/dev/urandom", O_RDONLY);
+		gcc_assert(urandom_fd >= 0);
+	}
+	if (rnd_idx >= ARRAY_SIZE(rnd_buf)) {
+		gcc_assert(read(urandom_fd, rnd_buf, sizeof(rnd_buf)) == sizeof(rnd_buf));
+		rnd_idx = 0;
+	}
+	return rnd_buf[rnd_idx++];
 }
 
 static tree tree_get_random_const(tree type)
@@ -537,8 +543,6 @@ static void latent_entropy_start_unit(void *gcc_data __unused,
 	tree type, id;
 	int quals;
 
-	seed = get_random_seed(false);
-
 	if (in_lto_p)
 		return;
 
@@ -573,6 +577,12 @@ __visible int plugin_init(struct plugin_name_args *plugin_info,
 	const struct plugin_argument * const argv = plugin_info->argv;
 	int i;
 
+	/*
+	 * Call get_random_seed() with noinit=true, so that this returns
+	 * 0 in the case where no seed has been passed via -frandom-seed.
+	 */
+	deterministic_seed = get_random_seed(true);
+
 	static const struct ggc_root_tab gt_ggc_r_gt_latent_entropy[] = {
 		{
 			.base = &latent_entropy_decl,
-- 
2.35.1


  reply	other threads:[~2022-04-06  4:46 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CAHmME9otYi4pCzZwSGnK40dp1QMRVPxp+DBysVuLXUKkXinAxg@mail.gmail.com>
2022-04-03 20:40 ` [PATCH v2] gcc-plugins: latent_entropy: use /dev/urandom Jason A. Donenfeld
2022-04-04 18:49   ` Kees Cook
2022-04-04 22:47     ` Jason A. Donenfeld
2022-04-04 23:06       ` [PATCH] " Jason A. Donenfeld
2022-04-04 23:07       ` [PATCH v3] " Jason A. Donenfeld
2022-04-05  3:01       ` [PATCH v2] " Kees Cook
2022-04-05 12:38         ` Jason A. Donenfeld
2022-04-05 17:17           ` Kees Cook
2022-04-05 17:40             ` Jason A. Donenfeld
2022-04-05 22:28               ` Jason A. Donenfeld [this message]
2022-04-12 18:32                 ` [PATCH v4] " Kees Cook

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=20220405222815.21155-1-Jason@zx2c4.com \
    --to=jason@zx2c4.com \
    --cc=keescook@chromium.org \
    --cc=linux-hardening@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pageexec@freemail.hu \
    --cc=stable@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).