linux-kernel - [PATCH 4.9 198/264] random: help compiler out with fast

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20220623164349.673626206@linuxfoundation.org>
Date:   Thu, 23 Jun 2022 18:43:11 +0200
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, "Jason A. Donenfeld" <Jason@...c4.com>
Subject: [PATCH 4.9 198/264] random: help compiler out with fast_mix() by using simpler arguments

From: "Jason A. Donenfeld" <Jason@...c4.com>

commit 791332b3cbb080510954a4c152ce02af8832eac9 upstream.

Now that fast_mix() has more than one caller, gcc no longer inlines it.
That's fine. But it also doesn't handle the compound literal argument we
pass it very efficiently, nor does it handle the loop as well as it
could. So just expand the code to spell out this function so that it
generates the same code as it did before. Performance-wise, this now
behaves as it did before the last commit. The difference in actual code
size on x86 is 45 bytes, which is less than a cache line.

Signed-off-by: Jason A. Donenfeld <Jason@...c4.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
---
 drivers/char/random.c |   44 +++++++++++++++++++++++---------------------
 1 file changed, 23 insertions(+), 21 deletions(-)

--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -1027,25 +1027,30 @@ static DEFINE_PER_CPU(struct fast_pool,
  * and therefore this has no security on its own. s represents the
  * four-word SipHash state, while v represents a two-word input.
  */
-static void fast_mix(unsigned long s[4], const unsigned long v[2])
+static void fast_mix(unsigned long s[4], unsigned long v1, unsigned long v2)
 {
-	size_t i;
-
-	for (i = 0; i < 2; ++i) {
-		s[3] ^= v[i];
 #ifdef CONFIG_64BIT
-		s[0] += s[1]; s[1] = rol64(s[1], 13); s[1] ^= s[0]; s[0] = rol64(s[0], 32);
-		s[2] += s[3]; s[3] = rol64(s[3], 16); s[3] ^= s[2];
-		s[0] += s[3]; s[3] = rol64(s[3], 21); s[3] ^= s[0];
-		s[2] += s[1]; s[1] = rol64(s[1], 17); s[1] ^= s[2]; s[2] = rol64(s[2], 32);
+#define PERM() do { \
+	s[0] += s[1]; s[1] = rol64(s[1], 13); s[1] ^= s[0]; s[0] = rol64(s[0], 32); \
+	s[2] += s[3]; s[3] = rol64(s[3], 16); s[3] ^= s[2]; \
+	s[0] += s[3]; s[3] = rol64(s[3], 21); s[3] ^= s[0]; \
+	s[2] += s[1]; s[1] = rol64(s[1], 17); s[1] ^= s[2]; s[2] = rol64(s[2], 32); \
+} while (0)
 #else
-		s[0] += s[1]; s[1] = rol32(s[1],  5); s[1] ^= s[0]; s[0] = rol32(s[0], 16);
-		s[2] += s[3]; s[3] = rol32(s[3],  8); s[3] ^= s[2];
-		s[0] += s[3]; s[3] = rol32(s[3],  7); s[3] ^= s[0];
-		s[2] += s[1]; s[1] = rol32(s[1], 13); s[1] ^= s[2]; s[2] = rol32(s[2], 16);
+#define PERM() do { \
+	s[0] += s[1]; s[1] = rol32(s[1],  5); s[1] ^= s[0]; s[0] = rol32(s[0], 16); \
+	s[2] += s[3]; s[3] = rol32(s[3],  8); s[3] ^= s[2]; \
+	s[0] += s[3]; s[3] = rol32(s[3],  7); s[3] ^= s[0]; \
+	s[2] += s[1]; s[1] = rol32(s[1], 13); s[1] ^= s[2]; s[2] = rol32(s[2], 16); \
+} while (0)
 #endif
-		s[0] ^= v[i];
-	}
+
+	s[3] ^= v1;
+	PERM();
+	s[0] ^= v1;
+	s[3] ^= v2;
+	PERM();
+	s[0] ^= v2;
 }
 
 #ifdef CONFIG_SMP
@@ -1115,10 +1120,8 @@ void add_interrupt_randomness(int irq)
 	struct pt_regs *regs = get_irq_regs();
 	unsigned int new_count;
 
-	fast_mix(fast_pool->pool, (unsigned long[2]){
-		entropy,
-		(regs ? instruction_pointer(regs) : _RET_IP_) ^ swab(irq)
-	});
+	fast_mix(fast_pool->pool, entropy,
+		 (regs ? instruction_pointer(regs) : _RET_IP_) ^ swab(irq));
 	new_count = ++fast_pool->count;
 
 	if (new_count & MIX_INFLIGHT)
@@ -1158,8 +1161,7 @@ static void add_timer_randomness(struct
 	 * sometime after, so mix into the fast pool.
 	 */
 	if (in_irq()) {
-		fast_mix(this_cpu_ptr(&irq_randomness)->pool,
-			 (unsigned long[2]){ entropy, num });
+		fast_mix(this_cpu_ptr(&irq_randomness)->pool, entropy, num);
 	} else {
 		spin_lock_irqsave(&input_pool.lock, flags);
 		_mix_pool_bytes(&entropy, sizeof(entropy));