From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754648Ab2HMWE0 (ORCPT ); Mon, 13 Aug 2012 18:04:26 -0400 Received: from mail-yx0-f174.google.com ([209.85.213.174]:35533 "EHLO mail-yx0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754241Ab2HMWEW (ORCPT ); Mon, 13 Aug 2012 18:04:22 -0400 From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org, stable@vger.kernel.org Cc: Greg KH , torvalds@linux-foundation.org, akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk, "Theodore Tso" Subject: [ 21/44] random: use lockless techniques in the interrupt path Date: Mon, 13 Aug 2012 15:02:28 -0700 Message-Id: <20120813220144.007275420@linuxfoundation.org> X-Mailer: git-send-email 1.7.10.1.362.g242cab3 In-Reply-To: <20120813220142.113186818@linuxfoundation.org> References: <20120813220142.113186818@linuxfoundation.org> User-Agent: quilt/0.60-20.5 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Greg KH 3.0-stable review patch. If anyone has any objections, please let me know. ------------------ From: Theodore Ts'o commit 902c098a3663de3fa18639efbb71b6080f0bcd3c upstream. The real-time Linux folks don't like add_interrupt_randomness() taking a spinlock since it is called in the low-level interrupt routine. This also allows us to reduce the overhead in the fast path, for the random driver, which is the interrupt collection path. Signed-off-by: "Theodore Ts'o" Signed-off-by: Greg Kroah-Hartman --- drivers/char/random.c | 78 +++++++++++++++++++++++++------------------------- 1 file changed, 39 insertions(+), 39 deletions(-) --- a/drivers/char/random.c +++ b/drivers/char/random.c @@ -418,9 +418,9 @@ struct entropy_store { /* read-write data: */ spinlock_t lock; unsigned add_ptr; + unsigned input_rotate; int entropy_count; int entropy_total; - int input_rotate; unsigned int initialized:1; __u8 last_data[EXTRACT_SIZE]; }; @@ -468,26 +468,24 @@ static __u32 const twist_table[8] = { * it's cheap to do so and helps slightly in the expected case where * the entropy is concentrated in the low-order bits. */ -static void mix_pool_bytes_extract(struct entropy_store *r, const void *in, - int nbytes, __u8 out[64]) +static void __mix_pool_bytes(struct entropy_store *r, const void *in, + int nbytes, __u8 out[64]) { unsigned long i, j, tap1, tap2, tap3, tap4, tap5; int input_rotate; int wordmask = r->poolinfo->poolwords - 1; const char *bytes = in; __u32 w; - unsigned long flags; - /* Taps are constant, so we can load them without holding r->lock. */ tap1 = r->poolinfo->tap1; tap2 = r->poolinfo->tap2; tap3 = r->poolinfo->tap3; tap4 = r->poolinfo->tap4; tap5 = r->poolinfo->tap5; - spin_lock_irqsave(&r->lock, flags); - input_rotate = r->input_rotate; - i = r->add_ptr; + smp_rmb(); + input_rotate = ACCESS_ONCE(r->input_rotate); + i = ACCESS_ONCE(r->add_ptr); /* mix one byte at a time to simplify size handling and churn faster */ while (nbytes--) { @@ -514,19 +512,23 @@ static void mix_pool_bytes_extract(struc input_rotate += i ? 7 : 14; } - r->input_rotate = input_rotate; - r->add_ptr = i; + ACCESS_ONCE(r->input_rotate) = input_rotate; + ACCESS_ONCE(r->add_ptr) = i; + smp_wmb(); if (out) for (j = 0; j < 16; j++) ((__u32 *)out)[j] = r->pool[(i - j) & wordmask]; - - spin_unlock_irqrestore(&r->lock, flags); } -static void mix_pool_bytes(struct entropy_store *r, const void *in, int bytes) +static void mix_pool_bytes(struct entropy_store *r, const void *in, + int nbytes, __u8 out[64]) { - mix_pool_bytes_extract(r, in, bytes, NULL); + unsigned long flags; + + spin_lock_irqsave(&r->lock, flags); + __mix_pool_bytes(r, in, nbytes, out); + spin_unlock_irqrestore(&r->lock, flags); } struct fast_pool { @@ -564,23 +566,22 @@ static void fast_mix(struct fast_pool *f */ static void credit_entropy_bits(struct entropy_store *r, int nbits) { - unsigned long flags; - int entropy_count; + int entropy_count, orig; if (!nbits) return; - spin_lock_irqsave(&r->lock, flags); - DEBUG_ENT("added %d entropy credits to %s\n", nbits, r->name); - entropy_count = r->entropy_count; +retry: + entropy_count = orig = ACCESS_ONCE(r->entropy_count); entropy_count += nbits; if (entropy_count < 0) { DEBUG_ENT("negative entropy/overflow\n"); entropy_count = 0; } else if (entropy_count > r->poolinfo->POOLBITS) entropy_count = r->poolinfo->POOLBITS; - r->entropy_count = entropy_count; + if (cmpxchg(&r->entropy_count, orig, entropy_count) != orig) + goto retry; if (!r->initialized && nbits > 0) { r->entropy_total += nbits; @@ -593,7 +594,6 @@ static void credit_entropy_bits(struct e wake_up_interruptible(&random_read_wait); kill_fasync(&fasync, SIGIO, POLL_IN); } - spin_unlock_irqrestore(&r->lock, flags); } /********************************************************************* @@ -680,7 +680,7 @@ static void add_timer_randomness(struct sample.cycles = get_cycles(); sample.num = num; - mix_pool_bytes(&input_pool, &sample, sizeof(sample)); + mix_pool_bytes(&input_pool, &sample, sizeof(sample), NULL); /* * Calculate number of bits of randomness we probably added. @@ -764,7 +764,7 @@ void add_interrupt_randomness(int irq, i fast_pool->last = now; r = nonblocking_pool.initialized ? &input_pool : &nonblocking_pool; - mix_pool_bytes(r, &fast_pool->pool, sizeof(fast_pool->pool)); + __mix_pool_bytes(r, &fast_pool->pool, sizeof(fast_pool->pool), NULL); /* * If we don't have a valid cycle counter, and we see * back-to-back timer interrupts, then skip giving credit for @@ -829,7 +829,7 @@ static void xfer_secondary_pool(struct e bytes = extract_entropy(r->pull, tmp, bytes, random_read_wakeup_thresh / 8, rsvd); - mix_pool_bytes(r, tmp, bytes); + mix_pool_bytes(r, tmp, bytes, NULL); credit_entropy_bits(r, bytes*8); } } @@ -890,9 +890,11 @@ static void extract_buf(struct entropy_s int i; __u32 hash[5], workspace[SHA_WORKSPACE_WORDS]; __u8 extract[64]; + unsigned long flags; /* Generate a hash across the pool, 16 words (512 bits) at a time */ sha_init(hash); + spin_lock_irqsave(&r->lock, flags); for (i = 0; i < r->poolinfo->poolwords; i += 16) sha_transform(hash, (__u8 *)(r->pool + i), workspace); @@ -905,7 +907,8 @@ static void extract_buf(struct entropy_s * brute-forcing the feedback as hard as brute-forcing the * hash. */ - mix_pool_bytes_extract(r, hash, sizeof(hash), extract); + __mix_pool_bytes(r, hash, sizeof(hash), extract); + spin_unlock_irqrestore(&r->lock, flags); /* * To avoid duplicates, we atomically extract a portion of the @@ -928,11 +931,10 @@ static void extract_buf(struct entropy_s } static ssize_t extract_entropy(struct entropy_store *r, void *buf, - size_t nbytes, int min, int reserved) + size_t nbytes, int min, int reserved) { ssize_t ret = 0, i; __u8 tmp[EXTRACT_SIZE]; - unsigned long flags; xfer_secondary_pool(r, nbytes); nbytes = account(r, nbytes, min, reserved); @@ -941,6 +943,8 @@ static ssize_t extract_entropy(struct en extract_buf(r, tmp); if (fips_enabled) { + unsigned long flags; + spin_lock_irqsave(&r->lock, flags); if (!memcmp(tmp, r->last_data, EXTRACT_SIZE)) panic("Hardware RNG duplicated output!\n"); @@ -1034,22 +1038,18 @@ EXPORT_SYMBOL(get_random_bytes); static void init_std_data(struct entropy_store *r) { int i; - ktime_t now; - unsigned long flags; + ktime_t now = ktime_get_real(); + unsigned long rv; - spin_lock_irqsave(&r->lock, flags); r->entropy_count = 0; r->entropy_total = 0; - spin_unlock_irqrestore(&r->lock, flags); - - now = ktime_get_real(); - mix_pool_bytes(r, &now, sizeof(now)); - for (i = r->poolinfo->POOLBYTES; i > 0; i -= sizeof flags) { - if (!arch_get_random_long(&flags)) + mix_pool_bytes(r, &now, sizeof(now), NULL); + for (i = r->poolinfo->POOLBYTES; i > 0; i -= sizeof(rv)) { + if (!arch_get_random_long(&rv)) break; - mix_pool_bytes(r, &flags, sizeof(flags)); + mix_pool_bytes(r, &rv, sizeof(rv), NULL); } - mix_pool_bytes(r, utsname(), sizeof(*(utsname()))); + mix_pool_bytes(r, utsname(), sizeof(*(utsname())), NULL); } static int rand_initialize(void) @@ -1186,7 +1186,7 @@ write_pool(struct entropy_store *r, cons count -= bytes; p += bytes; - mix_pool_bytes(r, buf, bytes); + mix_pool_bytes(r, buf, bytes, NULL); cond_resched(); }