Merge tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-tracegrafted

Pull tracing fixes from Steven Rostedt:

 - Don't destroy user event fields when removal fails

   User event fields are destroyed before the event is removed from
   visibility. But that can fail leaving the still visible event with no
   fields. Move the destroying of the fields to after the event is
   successfully removed from visibility.

 - Initialize function graph state is fork before calling
   copy_exec_state()

   For non-CLONE_VM forks, copy_exec_state() allocates a new
   task_exec_state. If that allocation fails, ftrace_graph_exit_task()
   will free the tasks ret_stack pointer. Since that pointer is still
   using the parent's ret_stack, it mistakenly frees the parent's
   pointer too.

   Call ftrace_graph_init() on the task first which will NULL out the
   new tasks's ret_stack and if the copy fails, it will not free
   anything.

 - Remove FGRAPH_MAX_INDEX

   The macro FGRAPH_MAX_INDEX was added but never used. Remove it.

 - Save ent_size in function graph printing of nested functions

   The function graph tracer needs to look at the next event to see if
   the next event is the return of the current function entry. If it is,
   it prints a single line:

	ktime_get();

   Otherwise it prints it like a nested function:

	tick_nohz_irq_exit() {
	    ktime_get();
	    kcpustat_irq_exit();
	}

   In order to look at the next event, it must save the current event so
   that it has the information to print from it. It saves the event in
   the iterator descriptor called "ent". What it doesn't save is the
   ent_size of the event which is now used to know if the function graph
   arguments are to be printed. The peek doesn't save the size so the
   size used happens to be that of the size of the last event that was
   seen.

   Save the entry event size in the iterator descriptor so that the
   correct size is used.

 - Fix several errors with freeing data in the histogram code

   The histogram code had a lot of leaked or or incorrect accounting
   when failures happen. Correct them.

 - Fix histogram regression of .percent and .graph modifiers

   Up until 6.3 histogram values could have "percent" or "graph"
   modifiers that changed how they were printed. But a change that added
   restricting histograms values from being strings, stack traces and
   other modifiers inadvertently prevented them from using the percent
   and graph modifiers, which were legal use cases for values.

   Put back the percent and graph modifiers.

 - Fix various typos in the comments

 - Set the trace_clock before initializing a histogram with clock
   argument

   The histogram API allows the user to specific which trace clock to
   use via a "clock=" string. The histogram is set up first before the
   clock is checked. If the passed in clock is not valid, it exits
   without fully fixing up the histogram leaving it on the list and a
   use-after-free can trigger.

   Update the clock argument first and if it fails then exit gracefully
   before the histogram trigger is placed on any lists.

 - Restore :mod: trailer after parsing in ftrace_set_clr_event

   The function ftrace_set_clr_event() modifies the parse string and
   needs to put it back to what was passed in. It searches for ":mod:"
   via a strsep() but fails to put back the first ':' in the string.

   Add back the ':' in the passed in string.

 - Take trace_array reference when opening a tracer options file

   The options files are dynamically created and some tracers add their
   own options. When a tracer adds their own list of options, the
   trace_array holding them has an array to hold the list of options for
   each tracer. This array increases in size via a krealloc(), and the
   new entry gets a newly allocated array to hold the options of the new
   tracer being added.

   The element in each entry of the tracer's option array holds a
   pointer back to the trace_array, a pointer to the tracer it is
   associated to, a pointer to the flags of the option.

   The issue is that these arrays are freed when the trace_array is
   freed when its instance it represents is removed from the instances
   directory. There's a race that an open of one of these options files
   can happen when the instance is being removed.

   Add a new helper function to be called by the open function of the
   options file to iterate all existing trace_arrays under a lock and
   find the one that has the given option element in one of it's tracer
   arrays. If found, then update the associated trace_array's reference
   counter to keep it from being freed. If not found, have the open call
   return -ENODEV.

 - Disable interrupts when acquiring the lock in rb_wake_up_waiters()

   The function rb_wake_up_waiters() assumes it will be called in
   interrupt context and does not disable irqs when taking
   cpu_buffer->reader_lock, which can be called in hard interrupt
   context. The issue is in PREEMPT_RT, this function is called in
   thread context leaving this lock open to a deadlock.

   Take the lock with interrupts disabled.

 - Use rcu_assign_pointer() for tmp_ops filter hash

   The tmp_ops used in update_ftrace_direct_mod() assigns its
   filter_hash field directly, but that field is annotated as __rcu and
   sparse complains. Assign it with rcu_assign_pointer()

 - Fix use-after-free in enable_trigger_private_data_free()

   The trace_event_call is accessed through the event_trigger_data's
   trace_event_file pointer to put the trace_event_call on freeing. The
   issue is that the trace_event_file data may have been freed already
   causing a use-after-free. Add a field to the event_trigger_data that
   points directly to the trace_event_call so that it can decrement its
   reference directly without needing to go through the
   trace_event_file.

 - Fix accounting of buffer data remote headers

   trace_buffer_desc_size() and trace_remote_alloc_buffer() undercount
   the number of pages is needed for the asked for size as it doesn't
   take into account the meta data on each page. Add a helper function
   to do the calculation properly and use that in these functions.

 - Catch nr_page_va overflow in ring_buffer_desc sizing

   The number of pages per remote ring buffer is capped by
   ring_buffer_desc::nr_page_va (32 bits). A buffer_size large enough to
   overflow that field would silently allocate a descriptor smaller than
   what was asked for.

 - Do not resize the subbuf order if any per_cpu buffer is disabled

   The mmapping of ring buffers disables resizing the subbuffers, but it
   is done per-cpu whereas the subbuf size change is done for all the
   per_cpu buffers under the buffer->mutex. It could change the size of
   some while the mapping is happening on others. Have the resize of the
   subbuf order check all the per_cpu buffers under the lock to see if
   any of them is disabled before starting and causing an inconsistency
   between buffers that are being mapped.

* tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (25 commits)
  ring-buffer: Check resize_disabled before publishing the new subbuf order
  tracing/remotes: Catch nr_page_va overflow in ring_buffer_desc sizing
  tracing/remotes: Account for ring buffer page header in size calculation
  tracing: Don't dereference trace_event_file in deferred trigger free
  ftrace: Use rcu_assign_pointer() for tmp_ops filter hash
  ring-buffer: Acquire the lock with irqsave in rb_wake_up_waiters()
  tracing: Take trace_array reference when opening a tracer options file
  tracing: Fix ring_buffer_read_page_size() kernel-doc
  tracing: Restore :mod: trailer after parsing in ftrace_set_clr_event()
  tracing: Fix memory corruption from a "STACKTRACE" histogram key
  tracing: Fix memory corruption from the histogram stacktrace modifier
  tracing: Undo the registration when enabling the histogram trigger fails
  tracing: Take the reference before publishing the named histogram trigger
  tracing: Set the trace clock before registering the histogram trigger
  tracing: Fix typo "preceeded" in comment
  tracing: Fix typo "availabe" in comment
  tracing: Let histogram values keep the percent and graph modifiers
  tracing: Keep the entry count when the histogram stats allocation fails
  tracing: Free histogram the field rejected for a bad modifier
  tracing: Free histogram the var ref when its initialization fails
  ...

1 files changed, 1736 insertions, 0 deletions

diff --git a/crypto/ecc.c b/crypto/ecc.c
new file mode 100644
index 000000000..3250a464b
--- /dev/null
+++ b/crypto/ecc.c
@@ -0,0 +1,1736 @@
+/*
+ * Copyright (c) 2013, 2014 Kenneth MacKay. All rights reserved.
+ * Copyright (c) 2019 Vitaly Chikunov <vt@altlinux.org>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met:
+ *  * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ *  * Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <crypto/ecc_curve.h>
+#include <linux/module.h>
+#include <linux/random.h>
+#include <linux/slab.h>
+#include <linux/swab.h>
+#include <linux/fips.h>
+#include <crypto/ecdh.h>
+#include <crypto/rng.h>
+#include <crypto/internal/ecc.h>
+#include <linux/unaligned.h>
+#include <linux/ratelimit.h>
+
+#include "ecc_curve_defs.h"
+
+typedef struct {
+	u64 m_low;
+	u64 m_high;
+} uint128_t;
+
+/* Returns curv25519 curve param */
+const struct ecc_curve *ecc_get_curve25519(void)
+{
+	return &ecc_25519;
+}
+EXPORT_SYMBOL(ecc_get_curve25519);
+
+const struct ecc_curve *ecc_get_curve(unsigned int curve_id)
+{
+	switch (curve_id) {
+	/* In FIPS mode only allow P256 and higher */
+	case ECC_CURVE_NIST_P192:
+		return fips_enabled ? NULL : &nist_p192;
+	case ECC_CURVE_NIST_P256:
+		return &nist_p256;
+	case ECC_CURVE_NIST_P384:
+		return &nist_p384;
+	case ECC_CURVE_NIST_P521:
+		return &nist_p521;
+	default:
+		return NULL;
+	}
+}
+EXPORT_SYMBOL(ecc_get_curve);
+
+void ecc_digits_from_bytes(const u8 *in, unsigned int nbytes,
+			   u64 *out, unsigned int ndigits)
+{
+	int diff = ndigits - DIV_ROUND_UP_POW2(nbytes, sizeof(u64));
+	unsigned int o = nbytes & 7;
+	__be64 msd = 0;
+
+	/* diff > 0: not enough input bytes: set most significant digits to 0 */
+	if (diff > 0) {
+		ndigits -= diff;
+		memset(&out[ndigits], 0, diff * sizeof(u64));
+	}
+
+	if (o) {
+		memcpy((u8 *)&msd + sizeof(msd) - o, in, o);
+		out[--ndigits] = be64_to_cpu(msd);
+		in += o;
+	}
+	ecc_swap_digits(in, out, ndigits);
+}
+EXPORT_SYMBOL(ecc_digits_from_bytes);
+
+struct ecc_point *ecc_alloc_point(unsigned int ndigits)
+{
+	struct ecc_point *p;
+	size_t ndigits_sz;
+
+	if (!ndigits)
+		return NULL;
+
+	p = kmalloc_obj(*p);
+	if (!p)
+		return NULL;
+
+	ndigits_sz = ndigits * sizeof(u64);
+	p->x = kmalloc(ndigits_sz, GFP_KERNEL);
+	if (!p->x)
+		goto err_alloc_x;
+
+	p->y = kmalloc(ndigits_sz, GFP_KERNEL);
+	if (!p->y)
+		goto err_alloc_y;
+
+	p->ndigits = ndigits;
+
+	return p;
+
+err_alloc_y:
+	kfree(p->x);
+err_alloc_x:
+	kfree(p);
+	return NULL;
+}
+EXPORT_SYMBOL(ecc_alloc_point);
+
+void ecc_free_point(struct ecc_point *p)
+{
+	if (!p)
+		return;
+
+	kfree_sensitive(p->x);
+	kfree_sensitive(p->y);
+	kfree_sensitive(p);
+}
+EXPORT_SYMBOL(ecc_free_point);
+
+static void vli_clear(u64 *vli, unsigned int ndigits)
+{
+	int i;
+
+	for (i = 0; i < ndigits; i++)
+		vli[i] = 0;
+}
+
+/* Returns true if vli == 0, false otherwise. */
+bool vli_is_zero(const u64 *vli, unsigned int ndigits)
+{
+	int i;
+
+	for (i = 0; i < ndigits; i++) {
+		if (vli[i])
+			return false;
+	}
+
+	return true;
+}
+EXPORT_SYMBOL(vli_is_zero);
+
+/* Returns nonzero if bit of vli is set. */
+static u64 vli_test_bit(const u64 *vli, unsigned int bit)
+{
+	return (vli[bit / 64] & ((u64)1 << (bit % 64)));
+}
+
+static bool vli_is_negative(const u64 *vli, unsigned int ndigits)
+{
+	return vli_test_bit(vli, ndigits * 64 - 1);
+}
+
+/* Counts the number of 64-bit "digits" in vli. */
+static unsigned int vli_num_digits(const u64 *vli, unsigned int ndigits)
+{
+	int i;
+
+	/* Search from the end until we find a non-zero digit.
+	 * We do it in reverse because we expect that most digits will
+	 * be nonzero.
+	 */
+	for (i = ndigits - 1; i >= 0 && vli[i] == 0; i--);
+
+	return (i + 1);
+}
+
+/* Counts the number of bits required for vli. */
+unsigned int vli_num_bits(const u64 *vli, unsigned int ndigits)
+{
+	unsigned int i, num_digits;
+	u64 digit;
+
+	num_digits = vli_num_digits(vli, ndigits);
+	if (num_digits == 0)
+		return 0;
+
+	digit = vli[num_digits - 1];
+	for (i = 0; digit; i++)
+		digit >>= 1;
+
+	return ((num_digits - 1) * 64 + i);
+}
+EXPORT_SYMBOL(vli_num_bits);
+
+/* Set dest from unaligned bit string src. */
+void vli_from_be64(u64 *dest, const void *src, unsigned int ndigits)
+{
+	int i;
+	const u64 *from = src;
+
+	for (i = 0; i < ndigits; i++)
+		dest[i] = get_unaligned_be64(&from[ndigits - 1 - i]);
+}
+EXPORT_SYMBOL(vli_from_be64);
+
+void vli_from_le64(u64 *dest, const void *src, unsigned int ndigits)
+{
+	int i;
+	const u64 *from = src;
+
+	for (i = 0; i < ndigits; i++)
+		dest[i] = get_unaligned_le64(&from[i]);
+}
+EXPORT_SYMBOL(vli_from_le64);
+
+/* Sets dest = src. */
+static void vli_set(u64 *dest, const u64 *src, unsigned int ndigits)
+{
+	int i;
+
+	for (i = 0; i < ndigits; i++)
+		dest[i] = src[i];
+}
+
+/* Returns sign of left - right. */
+int vli_cmp(const u64 *left, const u64 *right, unsigned int ndigits)
+{
+	int i;
+
+	for (i = ndigits - 1; i >= 0; i--) {
+		if (left[i] > right[i])
+			return 1;
+		else if (left[i] < right[i])
+			return -1;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(vli_cmp);
+
+/* Computes result = in << c, returning carry. Can modify in place
+ * (if result == in). 0 < shift < 64.
+ */
+static u64 vli_lshift(u64 *result, const u64 *in, unsigned int shift,
+		      unsigned int ndigits)
+{
+	u64 carry = 0;
+	int i;
+
+	for (i = 0; i < ndigits; i++) {
+		u64 temp = in[i];
+
+		result[i] = (temp << shift) | carry;
+		carry = temp >> (64 - shift);
+	}
+
+	return carry;
+}
+
+/* Computes vli = vli >> 1. */
+static void vli_rshift1(u64 *vli, unsigned int ndigits)
+{
+	u64 *end = vli;
+	u64 carry = 0;
+
+	vli += ndigits;
+
+	while (vli-- > end) {
+		u64 temp = *vli;
+		*vli = (temp >> 1) | carry;
+		carry = temp << 63;
+	}
+}
+
+#ifdef __has_builtin
+#if __has_builtin(__builtin_addcll)
+#define USE_BUILTIN_ADDC
+#endif
+#endif
+
+/* Computes result = left + right + carry_in and updates carry_out */
+static inline void add_carry(u64 left, u64 right, u64 *result, u64 carry_in,
+			     u64 *carry_out)
+{
+#ifdef USE_BUILTIN_ADDC
+	*result = __builtin_addcll(left, right, carry_in, carry_out);
+#else
+	u64 sum1, sum2;
+	u64 c1 = __builtin_uaddll_overflow(left, right, &sum1);
+	u64 c2 = __builtin_uaddll_overflow(sum1, carry_in, &sum2);
+	*result = sum2;
+	*carry_out = c1 | c2;
+#endif
+}
+
+#ifdef __has_builtin
+#if __has_builtin(__builtin_subcll)
+#define USE_BUILTIN_SUBC
+#endif
+#endif
+
+/* Computes result = left - right - borrow_in and updates borrow_out */
+static inline void sub_borrow(u64 left, u64 right, u64 *result, u64 borrow_in,
+			      u64 *borrow_out)
+{
+#ifdef USE_BUILTIN_SUBC
+	*result = __builtin_subcll(left, right, borrow_in, borrow_out);
+#else
+	u64 diff1, diff2;
+	u64 b1 = __builtin_usubll_overflow(left, right, &diff1);
+	u64 b2 = __builtin_usubll_overflow(diff1, borrow_in, &diff2);
+	*result = diff2;
+	*borrow_out = b1 | b2;
+#endif
+}
+
+/* Computes result = left + right, returning carry. Can modify in place. */
+static u64 vli_add(u64 *result, const u64 *left, const u64 *right,
+		   unsigned int ndigits)
+{
+	u64 carry = 0;
+	int i;
+
+	for (i = 0; i < ndigits; i++)
+		add_carry(left[i], right[i], &result[i], carry, &carry);
+
+	return carry;
+}
+
+/* Computes result = left + right, returning carry. Can modify in place. */
+static u64 vli_uadd(u64 *result, const u64 *left, u64 right,
+		    unsigned int ndigits)
+{
+	u64 carry;
+	int i;
+
+	if (ndigits == 0)
+		return right;
+
+	carry = __builtin_uaddll_overflow(left[0], right, &result[0]);
+
+	for (i = 1; i < ndigits; i++)
+		carry = __builtin_uaddll_overflow(left[i], carry, &result[i]);
+
+	return carry;
+}
+
+/* Computes result = left - right, returning borrow. Can modify in place. */
+u64 vli_sub(u64 *result, const u64 *left, const u64 *right,
+	    unsigned int ndigits)
+{
+	u64 borrow = 0;
+	int i;
+
+	for (i = 0; i < ndigits; i++)
+		sub_borrow(left[i], right[i], &result[i], borrow, &borrow);
+
+	return borrow;
+}
+EXPORT_SYMBOL(vli_sub);
+
+/* Computes result = left - right, returning borrow. Can modify in place. */
+static u64 vli_usub(u64 *result, const u64 *left, u64 right,
+		    unsigned int ndigits)
+{
+	u64 borrow;
+	int i;
+
+	if (ndigits == 0)
+		return right;
+
+	borrow = __builtin_usubll_overflow(left[0], right, &result[0]);
+
+	for (i = 1; i < ndigits; i++)
+		borrow = __builtin_usubll_overflow(left[i], borrow, &result[i]);
+
+	return borrow;
+}
+
+static uint128_t mul_64_64(u64 left, u64 right)
+{
+	uint128_t result;
+#if defined(CONFIG_ARCH_SUPPORTS_INT128)
+	unsigned __int128 m = (unsigned __int128)left * right;
+
+	result.m_low  = m;
+	result.m_high = m >> 64;
+#else
+	u64 a0 = left & 0xffffffffull;
+	u64 a1 = left >> 32;
+	u64 b0 = right & 0xffffffffull;
+	u64 b1 = right >> 32;
+	u64 m0 = a0 * b0;
+	u64 m1 = a0 * b1;
+	u64 m2 = a1 * b0;
+	u64 m3 = a1 * b1;
+
+	m2 += (m0 >> 32);
+	m2 += m1;
+
+	/* Overflow */
+	if (m2 < m1)
+		m3 += 0x100000000ull;
+
+	result.m_low = (m0 & 0xffffffffull) | (m2 << 32);
+	result.m_high = m3 + (m2 >> 32);
+#endif
+	return result;
+}
+
+/* Calculate addition with overflow checking. Returns true on wrap-around,
+ * false otherwise.
+ */
+static bool check_add_128_128_overflow(uint128_t *result, uint128_t a,
+				       uint128_t b)
+{
+	bool carry;
+
+	result->m_low = a.m_low + b.m_low;
+	carry = (result->m_low < a.m_low);
+
+	result->m_high = a.m_high + b.m_high + carry;
+
+	/* Using constant-time bitwise arithmetic to prevent timing
+	 * side-channels.
+	 */
+	carry = (result->m_high < a.m_high) |
+		((result->m_high == a.m_high) & carry);
+
+	return carry;
+}
+
+static void vli_mult(u64 *result, const u64 *left, const u64 *right,
+		     unsigned int ndigits)
+{
+	uint128_t r01 = { 0, 0 };
+	u64 r2 = 0;
+	unsigned int i, k;
+
+	/* Compute each digit of result in sequence, maintaining the
+	 * carries.
+	 */
+	for (k = 0; k < ndigits * 2 - 1; k++) {
+		unsigned int min;
+
+		if (k < ndigits)
+			min = 0;
+		else
+			min = (k + 1) - ndigits;
+
+		for (i = min; i <= k && i < ndigits; i++) {
+			uint128_t product;
+
+			product = mul_64_64(left[i], right[k - i]);
+			r2 += check_add_128_128_overflow(&r01, r01, product);
+		}
+
+		result[k] = r01.m_low;
+		r01.m_low = r01.m_high;
+		r01.m_high = r2;
+		r2 = 0;
+	}
+
+	result[ndigits * 2 - 1] = r01.m_low;
+}
+
+/* Compute product = left * right, for a small right value. */
+static void vli_umult(u64 *result, const u64 *left, u32 right,
+		      unsigned int ndigits)
+{
+	uint128_t r01 = { 0 };
+	unsigned int k;
+
+	for (k = 0; k < ndigits; k++) {
+		uint128_t product;
+
+		product = mul_64_64(left[k], right);
+		check_add_128_128_overflow(&r01, r01, product);
+		/* no carry */
+		result[k] = r01.m_low;
+		r01.m_low = r01.m_high;
+		r01.m_high = 0;
+	}
+	result[k] = r01.m_low;
+	for (++k; k < ndigits * 2; k++)
+		result[k] = 0;
+}
+
+static void vli_square(u64 *result, const u64 *left, unsigned int ndigits)
+{
+	uint128_t r01 = { 0, 0 };
+	u64 r2 = 0;
+	int i, k;
+
+	for (k = 0; k < ndigits * 2 - 1; k++) {
+		unsigned int min;
+
+		if (k < ndigits)
+			min = 0;
+		else
+			min = (k + 1) - ndigits;
+
+		for (i = min; i <= k && i <= k - i; i++) {
+			uint128_t product;
+
+			product = mul_64_64(left[i], left[k - i]);
+
+			if (i < k - i) {
+				r2 += product.m_high >> 63;
+				product.m_high = (product.m_high << 1) |
+						 (product.m_low >> 63);
+				product.m_low <<= 1;
+			}
+
+			r2 += check_add_128_128_overflow(&r01, r01, product);
+		}
+
+		result[k] = r01.m_low;
+		r01.m_low = r01.m_high;
+		r01.m_high = r2;
+		r2 = 0;
+	}
+
+	result[ndigits * 2 - 1] = r01.m_low;
+}
+
+/* Computes result = (left + right) % mod.
+ * Assumes that left < mod and right < mod, result != mod.
+ */
+static void vli_mod_add(u64 *result, const u64 *left, const u64 *right,
+			const u64 *mod, unsigned int ndigits)
+{
+	u64 carry;
+
+	carry = vli_add(result, left, right, ndigits);
+
+	/* result > mod (result = mod + remainder), so subtract mod to
+	 * get remainder.
+	 */
+	if (carry || vli_cmp(result, mod, ndigits) >= 0)
+		vli_sub(result, result, mod, ndigits);
+}
+
+/* Computes result = (left - right) % mod.
+ * Assumes that left < mod and right < mod, result != mod.
+ */
+static void vli_mod_sub(u64 *result, const u64 *left, const u64 *right,
+			const u64 *mod, unsigned int ndigits)
+{
+	u64 borrow = vli_sub(result, left, right, ndigits);
+
+	/* In this case, p_result == -diff == (max int) - diff.
+	 * Since -x % d == d - x, we can get the correct result from
+	 * result + mod (with overflow).
+	 */
+	if (borrow)
+		vli_add(result, result, mod, ndigits);
+}
+
+/*
+ * Computes result = product % mod
+ * for special form moduli: p = 2^k-c, for small c (note the minus sign)
+ *
+ * References:
+ * R. Crandall, C. Pomerance. Prime Numbers: A Computational Perspective.
+ * 9 Fast Algorithms for Large-Integer Arithmetic. 9.2.3 Moduli of special form
+ * Algorithm 9.2.13 (Fast mod operation for special-form moduli).
+ */
+static void vli_mmod_special(u64 *result, const u64 *product,
+			      const u64 *mod, unsigned int ndigits)
+{
+	u64 c = -mod[0];
+	u64 t[ECC_MAX_DIGITS * 2];
+	u64 r[ECC_MAX_DIGITS * 2];
+
+	vli_set(r, product, ndigits * 2);
+	while (!vli_is_zero(r + ndigits, ndigits)) {
+		vli_umult(t, r + ndigits, c, ndigits);
+		vli_clear(r + ndigits, ndigits);
+		vli_add(r, r, t, ndigits * 2);
+	}
+	vli_set(t, mod, ndigits);
+	vli_clear(t + ndigits, ndigits);
+	while (vli_cmp(r, t, ndigits * 2) >= 0)
+		vli_sub(r, r, t, ndigits * 2);
+	vli_set(result, r, ndigits);
+}
+
+/*
+ * Computes result = product % mod
+ * for special form moduli: p = 2^{k-1}+c, for small c (note the plus sign)
+ * where k-1 does not fit into qword boundary by -1 bit (such as 255).
+
+ * References (loosely based on):
+ * A. Menezes, P. van Oorschot, S. Vanstone. Handbook of Applied Cryptography.
+ * 14.3.4 Reduction methods for moduli of special form. Algorithm 14.47.
+ * URL: http://cacr.uwaterloo.ca/hac/about/chap14.pdf
+ *
+ * H. Cohen, G. Frey, R. Avanzi, C. Doche, T. Lange, K. Nguyen, F. Vercauteren.
+ * Handbook of Elliptic and Hyperelliptic Curve Cryptography.
+ * Algorithm 10.25 Fast reduction for special form moduli
+ */
+static void vli_mmod_special2(u64 *result, const u64 *product,
+			       const u64 *mod, unsigned int ndigits)
+{
+	u64 c2 = mod[0] * 2;
+	u64 q[ECC_MAX_DIGITS];
+	u64 r[ECC_MAX_DIGITS * 2];
+	u64 m[ECC_MAX_DIGITS * 2]; /* expanded mod */
+	int carry; /* last bit that doesn't fit into q */
+	int i;
+
+	vli_set(m, mod, ndigits);
+	vli_clear(m + ndigits, ndigits);
+
+	vli_set(r, product, ndigits);
+	/* q and carry are top bits */
+	vli_set(q, product + ndigits, ndigits);
+	vli_clear(r + ndigits, ndigits);
+	carry = vli_is_negative(r, ndigits);
+	if (carry)
+		r[ndigits - 1] &= (1ull << 63) - 1;
+	for (i = 1; carry || !vli_is_zero(q, ndigits); i++) {
+		u64 qc[ECC_MAX_DIGITS * 2];
+
+		vli_umult(qc, q, c2, ndigits);
+		if (carry)
+			vli_uadd(qc, qc, mod[0], ndigits * 2);
+		vli_set(q, qc + ndigits, ndigits);
+		vli_clear(qc + ndigits, ndigits);
+		carry = vli_is_negative(qc, ndigits);
+		if (carry)
+			qc[ndigits - 1] &= (1ull << 63) - 1;
+		if (i & 1)
+			vli_sub(r, r, qc, ndigits * 2);
+		else
+			vli_add(r, r, qc, ndigits * 2);
+	}
+	while (vli_is_negative(r, ndigits * 2))
+		vli_add(r, r, m, ndigits * 2);
+	while (vli_cmp(r, m, ndigits * 2) >= 0)
+		vli_sub(r, r, m, ndigits * 2);
+
+	vli_set(result, r, ndigits);
+}
+
+/*
+ * Computes result = product % mod, where product is 2N words long.
+ * Reference: Ken MacKay's micro-ecc.
+ * Currently only designed to work for curve_p or curve_n.
+ */
+static void vli_mmod_slow(u64 *result, u64 *product, const u64 *mod,
+			  unsigned int ndigits)
+{
+	u64 mod_m[2 * ECC_MAX_DIGITS];
+	u64 tmp[2 * ECC_MAX_DIGITS];
+	u64 *v[2] = { tmp, product };
+	u64 carry = 0;
+	unsigned int i;
+	/* Shift mod so its highest set bit is at the maximum position. */
+	int shift = (ndigits * 2 * 64) - vli_num_bits(mod, ndigits);
+	int word_shift = shift / 64;
+	int bit_shift = shift % 64;
+
+	vli_clear(mod_m, word_shift);
+	if (bit_shift > 0) {
+		for (i = 0; i < ndigits; ++i) {
+			mod_m[word_shift + i] = (mod[i] << bit_shift) | carry;
+			carry = mod[i] >> (64 - bit_shift);
+		}
+	} else
+		vli_set(mod_m + word_shift, mod, ndigits);
+
+	for (i = 1; shift >= 0; --shift) {
+		u64 borrow = 0;
+		unsigned int j;
+
+		for (j = 0; j < ndigits * 2; ++j) {
+			u64 diff = v[i][j] - mod_m[j] - borrow;
+
+			if (diff != v[i][j])
+				borrow = (diff > v[i][j]);
+			v[1 - i][j] = diff;
+		}
+		i = !(i ^ borrow); /* Swap the index if there was no borrow */
+		vli_rshift1(mod_m, ndigits);
+		mod_m[ndigits - 1] |= mod_m[ndigits] << (64 - 1);
+		vli_rshift1(mod_m + ndigits, ndigits);
+	}
+	vli_set(result, v[i], ndigits);
+}
+
+/* Computes result = product % mod using Barrett's reduction with precomputed
+ * value mu appended to the mod after ndigits, mu = (2^{2w} / mod) and have
+ * length ndigits + 1, where mu * (2^w - 1) should not overflow ndigits
+ * boundary.
+ *
+ * Reference:
+ * R. Brent, P. Zimmermann. Modern Computer Arithmetic. 2010.
+ * 2.4.1 Barrett's algorithm. Algorithm 2.5.
+ */
+static void vli_mmod_barrett(u64 *result, u64 *product, const u64 *mod,
+			     unsigned int ndigits)
+{
+	u64 q[ECC_MAX_DIGITS * 2];
+	u64 r[ECC_MAX_DIGITS * 2];
+	const u64 *mu = mod + ndigits;
+
+	vli_mult(q, product + ndigits, mu, ndigits);
+	if (mu[ndigits])
+		vli_add(q + ndigits, q + ndigits, product + ndigits, ndigits);
+	vli_mult(r, mod, q + ndigits, ndigits);
+	vli_sub(r, product, r, ndigits * 2);
+	while (!vli_is_zero(r + ndigits, ndigits) ||
+	       vli_cmp(r, mod, ndigits) != -1) {
+		u64 carry;
+
+		carry = vli_sub(r, r, mod, ndigits);
+		vli_usub(r + ndigits, r + ndigits, carry, ndigits);
+	}
+	vli_set(result, r, ndigits);
+}
+
+/* Computes p_result = p_product % curve_p.
+ * See algorithm 5 and 6 from
+ * http://www.isys.uni-klu.ac.at/PDF/2001-0126-MT.pdf
+ */
+static void vli_mmod_fast_192(u64 *result, const u64 *product,
+			      const u64 *curve_prime, u64 *tmp)
+{
+	const unsigned int ndigits = ECC_CURVE_NIST_P192_DIGITS;
+	int carry;
+
+	vli_set(result, product, ndigits);
+
+	vli_set(tmp, &product[3], ndigits);
+	carry = vli_add(result, result, tmp, ndigits);
+
+	tmp[0] = 0;
+	tmp[1] = product[3];
+	tmp[2] = product[4];
+	carry += vli_add(result, result, tmp, ndigits);
+
+	tmp[0] = tmp[1] = product[5];
+	tmp[2] = 0;
+	carry += vli_add(result, result, tmp, ndigits);
+
+	while (carry || vli_cmp(curve_prime, result, ndigits) != 1)
+		carry -= vli_sub(result, result, curve_prime, ndigits);
+}
+
+/* Computes result = product % curve_prime
+ * from http://www.nsa.gov/ia/_files/nist-routines.pdf
+ */
+static void vli_mmod_fast_256(u64 *result, const u64 *product,
+			      const u64 *curve_prime, u64 *tmp)
+{
+	int carry;
+	const unsigned int ndigits = ECC_CURVE_NIST_P256_DIGITS;
+
+	/* t */
+	vli_set(result, product, ndigits);
+
+	/* s1 */
+	tmp[0] = 0;
+	tmp[1] = product[5] & 0xffffffff00000000ull;
+	tmp[2] = product[6];
+	tmp[3] = product[7];
+	carry = vli_lshift(tmp, tmp, 1, ndigits);
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s2 */
+	tmp[1] = product[6] << 32;
+	tmp[2] = (product[6] >> 32) | (product[7] << 32);
+	tmp[3] = product[7] >> 32;
+	carry += vli_lshift(tmp, tmp, 1, ndigits);
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s3 */
+	tmp[0] = product[4];
+	tmp[1] = product[5] & 0xffffffff;
+	tmp[2] = 0;
+	tmp[3] = product[7];
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s4 */
+	tmp[0] = (product[4] >> 32) | (product[5] << 32);
+	tmp[1] = (product[5] >> 32) | (product[6] & 0xffffffff00000000ull);
+	tmp[2] = product[7];
+	tmp[3] = (product[6] >> 32) | (product[4] << 32);
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* d1 */
+	tmp[0] = (product[5] >> 32) | (product[6] << 32);
+	tmp[1] = (product[6] >> 32);
+	tmp[2] = 0;
+	tmp[3] = (product[4] & 0xffffffff) | (product[5] << 32);
+	carry -= vli_sub(result, result, tmp, ndigits);
+
+	/* d2 */
+	tmp[0] = product[6];
+	tmp[1] = product[7];
+	tmp[2] = 0;
+	tmp[3] = (product[4] >> 32) | (product[5] & 0xffffffff00000000ull);
+	carry -= vli_sub(result, result, tmp, ndigits);
+
+	/* d3 */
+	tmp[0] = (product[6] >> 32) | (product[7] << 32);
+	tmp[1] = (product[7] >> 32) | (product[4] << 32);
+	tmp[2] = (product[4] >> 32) | (product[5] << 32);
+	tmp[3] = (product[6] << 32);
+	carry -= vli_sub(result, result, tmp, ndigits);
+
+	/* d4 */
+	tmp[0] = product[7];
+	tmp[1] = product[4] & 0xffffffff00000000ull;
+	tmp[2] = product[5];
+	tmp[3] = product[6] & 0xffffffff00000000ull;
+	carry -= vli_sub(result, result, tmp, ndigits);
+
+	if (carry < 0) {
+		do {
+			carry += vli_add(result, result, curve_prime, ndigits);
+		} while (carry < 0);
+	} else {
+		while (carry || vli_cmp(curve_prime, result, ndigits) != 1)
+			carry -= vli_sub(result, result, curve_prime, ndigits);
+	}
+}
+
+#define SL32OR32(x32, y32) (((u64)x32 << 32) | y32)
+#define AND64H(x64)  (x64 & 0xffFFffFF00000000ull)
+#define AND64L(x64)  (x64 & 0x00000000ffFFffFFull)
+
+/* Computes result = product % curve_prime
+ * from "Mathematical routines for the NIST prime elliptic curves"
+ */
+static void vli_mmod_fast_384(u64 *result, const u64 *product,
+				const u64 *curve_prime, u64 *tmp)
+{
+	int carry;
+	const unsigned int ndigits = ECC_CURVE_NIST_P384_DIGITS;
+
+	/* t */
+	vli_set(result, product, ndigits);
+
+	/* s1 */
+	tmp[0] = 0;		// 0 || 0
+	tmp[1] = 0;		// 0 || 0
+	tmp[2] = SL32OR32(product[11], (product[10]>>32));	//a22||a21
+	tmp[3] = product[11]>>32;	// 0 ||a23
+	tmp[4] = 0;		// 0 || 0
+	tmp[5] = 0;		// 0 || 0
+	carry = vli_lshift(tmp, tmp, 1, ndigits);
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s2 */
+	tmp[0] = product[6];	//a13||a12
+	tmp[1] = product[7];	//a15||a14
+	tmp[2] = product[8];	//a17||a16
+	tmp[3] = product[9];	//a19||a18
+	tmp[4] = product[10];	//a21||a20
+	tmp[5] = product[11];	//a23||a22
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s3 */
+	tmp[0] = SL32OR32(product[11], (product[10]>>32));	//a22||a21
+	tmp[1] = SL32OR32(product[6], (product[11]>>32));	//a12||a23
+	tmp[2] = SL32OR32(product[7], (product[6])>>32);	//a14||a13
+	tmp[3] = SL32OR32(product[8], (product[7]>>32));	//a16||a15
+	tmp[4] = SL32OR32(product[9], (product[8]>>32));	//a18||a17
+	tmp[5] = SL32OR32(product[10], (product[9]>>32));	//a20||a19
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s4 */
+	tmp[0] = AND64H(product[11]);	//a23|| 0
+	tmp[1] = (product[10]<<32);	//a20|| 0
+	tmp[2] = product[6];	//a13||a12
+	tmp[3] = product[7];	//a15||a14
+	tmp[4] = product[8];	//a17||a16
+	tmp[5] = product[9];	//a19||a18
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s5 */
+	tmp[0] = 0;		//  0|| 0
+	tmp[1] = 0;		//  0|| 0
+	tmp[2] = product[10];	//a21||a20
+	tmp[3] = product[11];	//a23||a22
+	tmp[4] = 0;		//  0|| 0
+	tmp[5] = 0;		//  0|| 0
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* s6 */
+	tmp[0] = AND64L(product[10]);	// 0 ||a20
+	tmp[1] = AND64H(product[10]);	//a21|| 0
+	tmp[2] = product[11];	//a23||a22
+	tmp[3] = 0;		// 0 || 0
+	tmp[4] = 0;		// 0 || 0
+	tmp[5] = 0;		// 0 || 0
+	carry += vli_add(result, result, tmp, ndigits);
+
+	/* d1 */
+	tmp[0] = SL32OR32(product[6], (product[11]>>32));	//a12||a23
+	tmp[1] = SL32OR32(product[7], (product[6]>>32));	//a14||a13
+	tmp[2] = SL32OR32(product[8], (product[7]>>32));	//a16||a15
+	tmp[3] = SL32OR32(product[9], (product[8]>>32));	//a18||a17
+	tmp[4] = SL32OR32(product[10], (product[9]>>32));	//a20||a19
+	tmp[5] = SL32OR32(product[11], (product[10]>>32));	//a22||a21
+	carry -= vli_sub(result, result, tmp, ndigits);
+
+	/* d2 */
+	tmp[0] = (product[10]<<32);	//a20|| 0
+	tmp[1] = SL32OR32(product[11], (product[10]>>32));	//a22||a21
+	tmp[2] = (product[11]>>32);	// 0 ||a23
+	tmp[3] = 0;		// 0 || 0
+	tmp[4] = 0;		// 0 || 0
+	tmp[5] = 0;		// 0 || 0
+	carry -= vli_sub(result, result, tmp, ndigits);
+
+	/* d3 */
+	tmp[0] = 0;		// 0 || 0
+	tmp[1] = AND64H(product[11]);	//a23|| 0
+	tmp[2] = product[11]>>32;	// 0 ||a23
+	tmp[3] = 0;		// 0 || 0
+	tmp[4] = 0;		// 0 || 0
+	tmp[5] = 0;		// 0 || 0
+	carry -= vli_sub(result, result, tmp, ndigits);
+
+	if (carry < 0) {
+		do {
+			carry += vli_add(result, result, curve_prime, ndigits);
+		} while (carry < 0);
+	} else {
+		while (carry || vli_cmp(curve_prime, result, ndigits) != 1)
+			carry -= vli_sub(result, result, curve_prime, ndigits);
+	}
+
+}
+
+#undef SL32OR32
+#undef AND64H
+#undef AND64L
+
+/*
+ * Computes result = product % curve_prime
+ * from "Recommendations for Discrete Logarithm-Based Cryptography:
+ *       Elliptic Curve Domain Parameters" section G.1.4
+ */
+static void vli_mmod_fast_521(u64 *result, const u64 *product,
+			      const u64 *curve_prime, u64 *tmp)
+{
+	const unsigned int ndigits = ECC_CURVE_NIST_P521_DIGITS;
+	size_t i;
+
+	/* Initialize result with lowest 521 bits from product */
+	vli_set(result, product, ndigits);
+	result[8] &= 0x1ff;
+
+	for (i = 0; i < ndigits; i++)
+		tmp[i] = (product[8 + i] >> 9) | (product[9 + i] << 55);
+	tmp[8] &= 0x1ff;
+
+	vli_mod_add(result, result, tmp, curve_prime, ndigits);
+}
+
+/* Computes result = product % curve_prime for different curve_primes.
+ *
+ * Note that curve_primes are distinguished just by heuristic check and
+ * not by complete conformance check.
+ */
+static bool vli_mmod_fast(u64 *result, u64 *product,
+			  const struct ecc_curve *curve)
+{
+	u64 tmp[2 * ECC_MAX_DIGITS];
+	const u64 *curve_prime = curve->p;
+	const unsigned int ndigits = curve->g.ndigits;
+
+	/* All NIST curves have name prefix 'nist_' */
+	if (strncmp(curve->name, "nist_", 5) != 0) {
+		/* Try to handle Pseudo-Marsenne primes. */
+		if (curve_prime[ndigits - 1] == -1ull) {
+			vli_mmod_special(result, product, curve_prime,
+					 ndigits);
+			return true;
+		} else if (curve_prime[ndigits - 1] == 1ull << 63 &&
+			   curve_prime[ndigits - 2] == 0) {
+			vli_mmod_special2(result, product, curve_prime,
+					  ndigits);
+			return true;
+		}
+		vli_mmod_barrett(result, product, curve_prime, ndigits);
+		return true;
+	}
+
+	switch (ndigits) {
+	case ECC_CURVE_NIST_P192_DIGITS:
+		vli_mmod_fast_192(result, product, curve_prime, tmp);
+		break;
+	case ECC_CURVE_NIST_P256_DIGITS:
+		vli_mmod_fast_256(result, product, curve_prime, tmp);
+		break;
+	case ECC_CURVE_NIST_P384_DIGITS:
+		vli_mmod_fast_384(result, product, curve_prime, tmp);
+		break;
+	case ECC_CURVE_NIST_P521_DIGITS:
+		vli_mmod_fast_521(result, product, curve_prime, tmp);
+		break;
+	default:
+		pr_err_ratelimited("ecc: unsupported digits size!\n");
+		return false;
+	}
+
+	return true;
+}
+
+/* Computes result = (left * right) % mod.
+ * Assumes that mod is big enough curve order.
+ */
+void vli_mod_mult_slow(u64 *result, const u64 *left, const u64 *right,
+		       const u64 *mod, unsigned int ndigits)
+{
+	u64 product[ECC_MAX_DIGITS * 2];
+
+	vli_mult(product, left, right, ndigits);
+	vli_mmod_slow(result, product, mod, ndigits);
+}
+EXPORT_SYMBOL(vli_mod_mult_slow);
+
+/* Computes result = (left * right) % curve_prime. */
+static void vli_mod_mult_fast(u64 *result, const u64 *left, const u64 *right,
+			      const struct ecc_curve *curve)
+{
+	u64 product[2 * ECC_MAX_DIGITS];
+
+	vli_mult(product, left, right, curve->g.ndigits);
+	vli_mmod_fast(result, product, curve);
+}
+
+/* Computes result = left^2 % curve_prime. */
+static void vli_mod_square_fast(u64 *result, const u64 *left,
+				const struct ecc_curve *curve)
+{
+	u64 product[2 * ECC_MAX_DIGITS];
+
+	vli_square(product, left, curve->g.ndigits);
+	vli_mmod_fast(result, product, curve);
+}
+
+#define EVEN(vli) (!(vli[0] & 1))
+/* Computes result = (1 / p_input) % mod. All VLIs are the same size.
+ * See "From Euclid's GCD to Montgomery Multiplication to the Great Divide"
+ * https://labs.oracle.com/techrep/2001/smli_tr-2001-95.pdf
+ */
+void vli_mod_inv(u64 *result, const u64 *input, const u64 *mod,
+			unsigned int ndigits)
+{
+	u64 a[ECC_MAX_DIGITS], b[ECC_MAX_DIGITS];
+	u64 u[ECC_MAX_DIGITS], v[ECC_MAX_DIGITS];
+	u64 carry;
+	int cmp_result;
+
+	if (vli_is_zero(input, ndigits)) {
+		vli_clear(result, ndigits);
+		return;
+	}
+
+	vli_set(a, input, ndigits);
+	vli_set(b, mod, ndigits);
+	vli_clear(u, ndigits);
+	u[0] = 1;
+	vli_clear(v, ndigits);
+
+	while ((cmp_result = vli_cmp(a, b, ndigits)) != 0) {
+		carry = 0;
+
+		if (EVEN(a)) {
+			vli_rshift1(a, ndigits);
+
+			if (!EVEN(u))
+				carry = vli_add(u, u, mod, ndigits);
+
+			vli_rshift1(u, ndigits);
+			if (carry)
+				u[ndigits - 1] |= 0x8000000000000000ull;
+		} else if (EVEN(b)) {
+			vli_rshift1(b, ndigits);
+
+			if (!EVEN(v))
+				carry = vli_add(v, v, mod, ndigits);
+
+			vli_rshift1(v, ndigits);
+			if (carry)
+				v[ndigits - 1] |= 0x8000000000000000ull;
+		} else if (cmp_result > 0) {
+			vli_sub(a, a, b, ndigits);
+			vli_rshift1(a, ndigits);
+
+			if (vli_cmp(u, v, ndigits) < 0)
+				vli_add(u, u, mod, ndigits);
+
+			vli_sub(u, u, v, ndigits);
+			if (!EVEN(u))
+				carry = vli_add(u, u, mod, ndigits);
+
+			vli_rshift1(u, ndigits);
+			if (carry)
+				u[ndigits - 1] |= 0x8000000000000000ull;
+		} else {
+			vli_sub(b, b, a, ndigits);
+			vli_rshift1(b, ndigits);
+
+			if (vli_cmp(v, u, ndigits) < 0)
+				vli_add(v, v, mod, ndigits);
+
+			vli_sub(v, v, u, ndigits);
+			if (!EVEN(v))
+				carry = vli_add(v, v, mod, ndigits);
+
+			vli_rshift1(v, ndigits);
+			if (carry)
+				v[ndigits - 1] |= 0x8000000000000000ull;
+		}
+	}
+
+	vli_set(result, u, ndigits);
+}
+EXPORT_SYMBOL(vli_mod_inv);
+
+/* ------ Point operations ------ */
+
+/* Returns true if p_point is the point at infinity, false otherwise. */
+bool ecc_point_is_zero(const struct ecc_point *point)
+{
+	return (vli_is_zero(point->x, point->ndigits) &&
+		vli_is_zero(point->y, point->ndigits));
+}
+EXPORT_SYMBOL(ecc_point_is_zero);
+
+/* Point multiplication algorithm using Montgomery's ladder with co-Z
+ * coordinates. From https://eprint.iacr.org/2011/338.pdf
+ */
+
+/* Double in place */
+static void ecc_point_double_jacobian(u64 *x1, u64 *y1, u64 *z1,
+					const struct ecc_curve *curve)
+{
+	/* t1 = x, t2 = y, t3 = z */
+	u64 t4[ECC_MAX_DIGITS];
+	u64 t5[ECC_MAX_DIGITS];
+	const u64 *curve_prime = curve->p;
+	const unsigned int ndigits = curve->g.ndigits;
+
+	if (vli_is_zero(z1, ndigits))
+		return;
+
+	/* t4 = y1^2 */
+	vli_mod_square_fast(t4, y1, curve);
+	/* t5 = x1*y1^2 = A */
+	vli_mod_mult_fast(t5, x1, t4, curve);
+	/* t4 = y1^4 */
+	vli_mod_square_fast(t4, t4, curve);
+	/* t2 = y1*z1 = z3 */
+	vli_mod_mult_fast(y1, y1, z1, curve);
+	/* t3 = z1^2 */
+	vli_mod_square_fast(z1, z1, curve);
+
+	/* t1 = x1 + z1^2 */
+	vli_mod_add(x1, x1, z1, curve_prime, ndigits);
+	/* t3 = 2*z1^2 */
+	vli_mod_add(z1, z1, z1, curve_prime, ndigits);
+	/* t3 = x1 - z1^2 */
+	vli_mod_sub(z1, x1, z1, curve_prime, ndigits);
+	/* t1 = x1^2 - z1^4 */
+	vli_mod_mult_fast(x1, x1, z1, curve);
+
+	/* t3 = 2*(x1^2 - z1^4) */
+	vli_mod_add(z1, x1, x1, curve_prime, ndigits);
+	/* t1 = 3*(x1^2 - z1^4) */
+	vli_mod_add(x1, x1, z1, curve_prime, ndigits);
+	if (vli_test_bit(x1, 0)) {
+		u64 carry = vli_add(x1, x1, curve_prime, ndigits);
+
+		vli_rshift1(x1, ndigits);
+		x1[ndigits - 1] |= carry << 63;
+	} else {
+		vli_rshift1(x1, ndigits);
+	}
+	/* t1 = 3/2*(x1^2 - z1^4) = B */
+
+	/* t3 = B^2 */
+	vli_mod_square_fast(z1, x1, curve);
+	/* t3 = B^2 - A */
+	vli_mod_sub(z1, z1, t5, curve_prime, ndigits);
+	/* t3 = B^2 - 2A = x3 */
+	vli_mod_sub(z1, z1, t5, curve_prime, ndigits);
+	/* t5 = A - x3 */
+	vli_mod_sub(t5, t5, z1, curve_prime, ndigits);
+	/* t1 = B * (A - x3) */
+	vli_mod_mult_fast(x1, x1, t5, curve);
+	/* t4 = B * (A - x3) - y1^4 = y3 */
+	vli_mod_sub(t4, x1, t4, curve_prime, ndigits);
+
+	vli_set(x1, z1, ndigits);
+	vli_set(z1, y1, ndigits);
+	vli_set(y1, t4, ndigits);
+}
+
+/* Modify (x1, y1) => (x1 * z^2, y1 * z^3) */
+static void apply_z(u64 *x1, u64 *y1, u64 *z, const struct ecc_curve *curve)
+{
+	u64 t1[ECC_MAX_DIGITS];
+
+	vli_mod_square_fast(t1, z, curve);		/* z^2 */
+	vli_mod_mult_fast(x1, x1, t1, curve);	/* x1 * z^2 */
+	vli_mod_mult_fast(t1, t1, z, curve);	/* z^3 */
+	vli_mod_mult_fast(y1, y1, t1, curve);	/* y1 * z^3 */
+}
+
+/* P = (x1, y1) => 2P, (x2, y2) => P' */
+static void xycz_initial_double(u64 *x1, u64 *y1, u64 *x2, u64 *y2,
+				u64 *p_initial_z, const struct ecc_curve *curve)
+{
+	u64 z[ECC_MAX_DIGITS];
+	const unsigned int ndigits = curve->g.ndigits;
+
+	vli_set(x2, x1, ndigits);
+	vli_set(y2, y1, ndigits);
+
+	vli_clear(z, ndigits);
+	z[0] = 1;
+
+	if (p_initial_z)
+		vli_set(z, p_initial_z, ndigits);
+
+	apply_z(x1, y1, z, curve);
+
+	ecc_point_double_jacobian(x1, y1, z, curve);
+
+	apply_z(x2, y2, z, curve);
+}
+
+/* Input P = (x1, y1, Z), Q = (x2, y2, Z)
+ * Output P' = (x1', y1', Z3), P + Q = (x3, y3, Z3)
+ * or P => P', Q => P + Q
+ */
+static void xycz_add(u64 *x1, u64 *y1, u64 *x2, u64 *y2,
+			const struct ecc_curve *curve)
+{
+	/* t1 = X1, t2 = Y1, t3 = X2, t4 = Y2 */
+	u64 t5[ECC_MAX_DIGITS];
+	const u64 *curve_prime = curve->p;
+	const unsigned int ndigits = curve->g.ndigits;
+
+	/* t5 = x2 - x1 */
+	vli_mod_sub(t5, x2, x1, curve_prime, ndigits);
+	/* t5 = (x2 - x1)^2 = A */
+	vli_mod_square_fast(t5, t5, curve);
+	/* t1 = x1*A = B */
+	vli_mod_mult_fast(x1, x1, t5, curve);
+	/* t3 = x2*A = C */
+	vli_mod_mult_fast(x2, x2, t5, curve);
+	/* t4 = y2 - y1 */
+	vli_mod_sub(y2, y2, y1, curve_prime, ndigits);
+	/* t5 = (y2 - y1)^2 = D */
+	vli_mod_square_fast(t5, y2, curve);
+
+	/* t5 = D - B */
+	vli_mod_sub(t5, t5, x1, curve_prime, ndigits);
+	/* t5 = D - B - C = x3 */
+	vli_mod_sub(t5, t5, x2, curve_prime, ndigits);
+	/* t3 = C - B */
+	vli_mod_sub(x2, x2, x1, curve_prime, ndigits);
+	/* t2 = y1*(C - B) */
+	vli_mod_mult_fast(y1, y1, x2, curve);
+	/* t3 = B - x3 */
+	vli_mod_sub(x2, x1, t5, curve_prime, ndigits);
+	/* t4 = (y2 - y1)*(B - x3) */
+	vli_mod_mult_fast(y2, y2, x2, curve);
+	/* t4 = y3 */
+	vli_mod_sub(y2, y2, y1, curve_prime, ndigits);
+
+	vli_set(x2, t5, ndigits);
+}
+
+/* Input P = (x1, y1, Z), Q = (x2, y2, Z)
+ * Output P + Q = (x3, y3, Z3), P - Q = (x3', y3', Z3)
+ * or P => P - Q, Q => P + Q
+ */
+static void xycz_add_c(u64 *x1, u64 *y1, u64 *x2, u64 *y2,
+			const struct ecc_curve *curve)
+{
+	/* t1 = X1, t2 = Y1, t3 = X2, t4 = Y2 */
+	u64 t5[ECC_MAX_DIGITS];
+	u64 t6[ECC_MAX_DIGITS];
+	u64 t7[ECC_MAX_DIGITS];
+	const u64 *curve_prime = curve->p;
+	const unsigned int ndigits = curve->g.ndigits;
+
+	/* t5 = x2 - x1 */
+	vli_mod_sub(t5, x2, x1, curve_prime, ndigits);
+	/* t5 = (x2 - x1)^2 = A */
+	vli_mod_square_fast(t5, t5, curve);
+	/* t1 = x1*A = B */
+	vli_mod_mult_fast(x1, x1, t5, curve);
+	/* t3 = x2*A = C */
+	vli_mod_mult_fast(x2, x2, t5, curve);
+	/* t4 = y2 + y1 */
+	vli_mod_add(t5, y2, y1, curve_prime, ndigits);
+	/* t4 = y2 - y1 */
+	vli_mod_sub(y2, y2, y1, curve_prime, ndigits);
+
+	/* t6 = C - B */
+	vli_mod_sub(t6, x2, x1, curve_prime, ndigits);
+	/* t2 = y1 * (C - B) */
+	vli_mod_mult_fast(y1, y1, t6, curve);
+	/* t6 = B + C */
+	vli_mod_add(t6, x1, x2, curve_prime, ndigits);
+	/* t3 = (y2 - y1)^2 */
+	vli_mod_square_fast(x2, y2, curve);
+	/* t3 = x3 */
+	vli_mod_sub(x2, x2, t6, curve_prime, ndigits);
+
+	/* t7 = B - x3 */
+	vli_mod_sub(t7, x1, x2, curve_prime, ndigits);
+	/* t4 = (y2 - y1)*(B - x3) */
+	vli_mod_mult_fast(y2, y2, t7, curve);
+	/* t4 = y3 */
+	vli_mod_sub(y2, y2, y1, curve_prime, ndigits);
+
+	/* t7 = (y2 + y1)^2 = F */
+	vli_mod_square_fast(t7, t5, curve);
+	/* t7 = x3' */
+	vli_mod_sub(t7, t7, t6, curve_prime, ndigits);
+	/* t6 = x3' - B */
+	vli_mod_sub(t6, t7, x1, curve_prime, ndigits);
+	/* t6 = (y2 + y1)*(x3' - B) */
+	vli_mod_mult_fast(t6, t6, t5, curve);
+	/* t2 = y3' */
+	vli_mod_sub(y1, t6, y1, curve_prime, ndigits);
+
+	vli_set(x1, t7, ndigits);
+}
+
+static void ecc_point_mult(struct ecc_point *result,
+			   const struct ecc_point *point, const u64 *scalar,
+			   u64 *initial_z, const struct ecc_curve *curve,
+			   unsigned int ndigits)
+{
+	/* R0 and R1 */
+	u64 rx[2][ECC_MAX_DIGITS];
+	u64 ry[2][ECC_MAX_DIGITS];
+	u64 z[ECC_MAX_DIGITS];
+	u64 sk[2][ECC_MAX_DIGITS];
+	u64 *curve_prime = curve->p;
+	int i, nb;
+	int num_bits;
+	int carry;
+
+	carry = vli_add(sk[0], scalar, curve->n, ndigits);
+	vli_add(sk[1], sk[0], curve->n, ndigits);
+	scalar = sk[!carry];
+	if (curve->nbits == 521)	/* NIST P521 */
+		num_bits = curve->nbits + 2;
+	else
+		num_bits = sizeof(u64) * ndigits * 8 + 1;
+
+	vli_set(rx[1], point->x, ndigits);
+	vli_set(ry[1], point->y, ndigits);
+
+	xycz_initial_double(rx[1], ry[1], rx[0], ry[0], initial_z, curve);
+
+	for (i = num_bits - 2; i > 0; i--) {
+		nb = !vli_test_bit(scalar, i);
+		xycz_add_c(rx[1 - nb], ry[1 - nb], rx[nb], ry[nb], curve);
+		xycz_add(rx[nb], ry[nb], rx[1 - nb], ry[1 - nb], curve);
+	}
+
+	nb = !vli_test_bit(scalar, 0);
+	xycz_add_c(rx[1 - nb], ry[1 - nb], rx[nb], ry[nb], curve);
+
+	/* Find final 1/Z value. */
+	/* X1 - X0 */
+	vli_mod_sub(z, rx[1], rx[0], curve_prime, ndigits);
+	/* Yb * (X1 - X0) */
+	vli_mod_mult_fast(z, z, ry[1 - nb], curve);
+	/* xP * Yb * (X1 - X0) */
+	vli_mod_mult_fast(z, z, point->x, curve);
+
+	/* 1 / (xP * Yb * (X1 - X0)) */
+	vli_mod_inv(z, z, curve_prime, point->ndigits);
+
+	/* yP / (xP * Yb * (X1 - X0)) */
+	vli_mod_mult_fast(z, z, point->y, curve);
+	/* Xb * yP / (xP * Yb * (X1 - X0)) */
+	vli_mod_mult_fast(z, z, rx[1 - nb], curve);
+	/* End 1/Z calculation */
+
+	xycz_add(rx[nb], ry[nb], rx[1 - nb], ry[1 - nb], curve);
+
+	apply_z(rx[0], ry[0], z, curve);
+
+	vli_set(result->x, rx[0], ndigits);
+	vli_set(result->y, ry[0], ndigits);
+}
+
+/* Computes R = P + Q mod p */
+static void ecc_point_add(const struct ecc_point *result,
+		   const struct ecc_point *p, const struct ecc_point *q,
+		   const struct ecc_curve *curve)
+{
+	u64 z[ECC_MAX_DIGITS];
+	u64 px[ECC_MAX_DIGITS];
+	u64 py[ECC_MAX_DIGITS];
+	unsigned int ndigits = curve->g.ndigits;
+
+	vli_set(result->x, q->x, ndigits);
+	vli_set(result->y, q->y, ndigits);
+	vli_mod_sub(z, result->x, p->x, curve->p, ndigits);
+	vli_set(px, p->x, ndigits);
+	vli_set(py, p->y, ndigits);
+	xycz_add(px, py, result->x, result->y, curve);
+	vli_mod_inv(z, z, curve->p, ndigits);
+	apply_z(result->x, result->y, z, curve);
+}
+
+/* Computes R = u1P + u2Q mod p using Shamir's trick.
+ * Based on: Kenneth MacKay's micro-ecc (2014).
+ */
+void ecc_point_mult_shamir(const struct ecc_point *result,
+			   const u64 *u1, const struct ecc_point *p,
+			   const u64 *u2, const struct ecc_point *q,
+			   const struct ecc_curve *curve)
+{
+	u64 z[ECC_MAX_DIGITS];
+	u64 sump[2][ECC_MAX_DIGITS];
+	u64 *rx = result->x;
+	u64 *ry = result->y;
+	unsigned int ndigits = curve->g.ndigits;
+	unsigned int num_bits;
+	struct ecc_point sum = ECC_POINT_INIT(sump[0], sump[1], ndigits);
+	const struct ecc_point *points[4];
+	const struct ecc_point *point;
+	unsigned int idx;
+	int i;
+
+	ecc_point_add(&sum, p, q, curve);
+	points[0] = NULL;
+	points[1] = p;
+	points[2] = q;
+	points[3] = &sum;
+
+	num_bits = max(vli_num_bits(u1, ndigits), vli_num_bits(u2, ndigits));
+	i = num_bits - 1;
+	idx = !!vli_test_bit(u1, i);
+	idx |= (!!vli_test_bit(u2, i)) << 1;
+	point = points[idx];
+
+	vli_set(rx, point->x, ndigits);
+	vli_set(ry, point->y, ndigits);
+	vli_clear(z + 1, ndigits - 1);
+	z[0] = 1;
+
+	for (--i; i >= 0; i--) {
+		ecc_point_double_jacobian(rx, ry, z, curve);
+		idx = !!vli_test_bit(u1, i);
+		idx |= (!!vli_test_bit(u2, i)) << 1;
+		point = points[idx];
+		if (point) {
+			u64 tx[ECC_MAX_DIGITS];
+			u64 ty[ECC_MAX_DIGITS];
+			u64 tz[ECC_MAX_DIGITS];
+
+			vli_set(tx, point->x, ndigits);
+			vli_set(ty, point->y, ndigits);
+			apply_z(tx, ty, z, curve);
+			vli_mod_sub(tz, rx, tx, curve->p, ndigits);
+			xycz_add(tx, ty, rx, ry, curve);
+			vli_mod_mult_fast(z, z, tz, curve);
+		}
+	}
+	vli_mod_inv(z, z, curve->p, ndigits);
+	apply_z(rx, ry, z, curve);
+}
+EXPORT_SYMBOL(ecc_point_mult_shamir);
+
+/*
+ * This function performs checks equivalent to Appendix A.4.2 of FIPS 186-5.
+ * Whereas A.4.2 results in an integer in the interval [1, n-1], this function
+ * ensures that the integer is in the range of [2, n-3]. We are slightly
+ * stricter because of the currently used scalar multiplication algorithm.
+ */
+static int __ecc_is_key_valid(const struct ecc_curve *curve,
+			      const u64 *private_key, unsigned int ndigits)
+{
+	u64 one[ECC_MAX_DIGITS] = { 1, };
+	u64 res[ECC_MAX_DIGITS];
+
+	if (!private_key)
+		return -EINVAL;
+
+	if (curve->g.ndigits != ndigits)
+		return -EINVAL;
+
+	/* Make sure the private key is in the range [2, n-3]. */
+	if (vli_cmp(one, private_key, ndigits) != -1)
+		return -EINVAL;
+	vli_sub(res, curve->n, one, ndigits);
+	vli_sub(res, res, one, ndigits);
+	if (vli_cmp(res, private_key, ndigits) != 1)
+		return -EINVAL;
+
+	return 0;
+}
+
+int ecc_is_key_valid(unsigned int curve_id, unsigned int ndigits,
+		     const u64 *private_key, unsigned int private_key_len)
+{
+	int nbytes;
+	const struct ecc_curve *curve = ecc_get_curve(curve_id);
+
+	nbytes = ndigits << ECC_DIGITS_TO_BYTES_SHIFT;
+
+	if (private_key_len != nbytes)
+		return -EINVAL;
+
+	return __ecc_is_key_valid(curve, private_key, ndigits);
+}
+EXPORT_SYMBOL(ecc_is_key_valid);
+
+/*
+ * ECC private keys are generated using the method of rejection sampling,
+ * equivalent to that described in FIPS 186-5, Appendix A.2.2.
+ *
+ * This method generates a private key uniformly distributed in the range
+ * [2, n-3].
+ */
+int ecc_gen_privkey(unsigned int curve_id, unsigned int ndigits,
+		    u64 *private_key)
+{
+	const struct ecc_curve *curve = ecc_get_curve(curve_id);
+	unsigned int nbytes = ndigits << ECC_DIGITS_TO_BYTES_SHIFT;
+	unsigned int nbits = vli_num_bits(curve->n, ndigits);
+	int err;
+
+	/*
+	 * Step 1 & 2: check that N is included in Table 1 of FIPS 186-5,
+	 * section 6.1.1.
+	 */
+	if (nbits < 224)
+		return -EINVAL;
+
+	/*
+	 * FIPS 186-5 recommends that the private key should be obtained from a
+	 * RBG with a security strength equal to or greater than the security
+	 * strength associated with N.
+	 *
+	 * The maximum security strength identified by NIST SP800-57pt1r4 for
+	 * ECC is 256 (N >= 512).
+	 *
+	 * This condition is met by stdrng because it selects a favored DRBG
+	 * with a security strength of 256.
+	 */
+	/* Step 3: obtain N returned_bits from the DRBG. */
+	err = crypto_stdrng_get_bytes(private_key, nbytes);
+	if (err)
+		return err;
+
+	/* Step 4: make sure the private key is in the valid range. */
+	if (__ecc_is_key_valid(curve, private_key, ndigits))
+		return -EINVAL;
+
+	return 0;
+}
+EXPORT_SYMBOL(ecc_gen_privkey);
+
+int ecc_make_pub_key(unsigned int curve_id, unsigned int ndigits,
+		     const u64 *private_key, u64 *public_key)
+{
+	int ret = 0;
+	struct ecc_point *pk;
+	const struct ecc_curve *curve = ecc_get_curve(curve_id);
+
+	if (!private_key) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	pk = ecc_alloc_point(ndigits);
+	if (!pk) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ecc_point_mult(pk, &curve->g, private_key, NULL, curve, ndigits);
+
+	/* SP800-56A rev 3 5.6.2.1.3 key check */
+	if (ecc_is_pubkey_valid_full(curve, pk)) {
+		ret = -EAGAIN;
+		goto err_free_point;
+	}
+
+	ecc_swap_digits(pk->x, public_key, ndigits);
+	ecc_swap_digits(pk->y, &public_key[ndigits], ndigits);
+
+err_free_point:
+	ecc_free_point(pk);
+out:
+	return ret;
+}
+EXPORT_SYMBOL(ecc_make_pub_key);
+
+/* SP800-56A section 5.6.2.3.4 partial verification: ephemeral keys only */
+int ecc_is_pubkey_valid_partial(const struct ecc_curve *curve,
+				struct ecc_point *pk)
+{
+	u64 yy[ECC_MAX_DIGITS], xxx[ECC_MAX_DIGITS], w[ECC_MAX_DIGITS];
+
+	if (WARN_ON(pk->ndigits != curve->g.ndigits))
+		return -EINVAL;
+
+	/* Check 1: Verify key is not the zero point. */
+	if (ecc_point_is_zero(pk))
+		return -EINVAL;
+
+	/* Check 2: Verify key is in the range [1, p-1]. */
+	if (vli_cmp(curve->p, pk->x, pk->ndigits) != 1)
+		return -EINVAL;
+	if (vli_cmp(curve->p, pk->y, pk->ndigits) != 1)
+		return -EINVAL;
+
+	/* Check 3: Verify that y^2 == (x^3 + a·x + b) mod p */
+	vli_mod_square_fast(yy, pk->y, curve); /* y^2 */
+	vli_mod_square_fast(xxx, pk->x, curve); /* x^2 */
+	vli_mod_mult_fast(xxx, xxx, pk->x, curve); /* x^3 */
+	vli_mod_mult_fast(w, curve->a, pk->x, curve); /* a·x */
+	vli_mod_add(w, w, curve->b, curve->p, pk->ndigits); /* a·x + b */
+	vli_mod_add(w, w, xxx, curve->p, pk->ndigits); /* x^3 + a·x + b */
+	if (vli_cmp(yy, w, pk->ndigits) != 0) /* Equation */
+		return -EINVAL;
+
+	return 0;
+}
+EXPORT_SYMBOL(ecc_is_pubkey_valid_partial);
+
+/* SP800-56A section 5.6.2.3.3 full verification */
+int ecc_is_pubkey_valid_full(const struct ecc_curve *curve,
+			     struct ecc_point *pk)
+{
+	struct ecc_point *nQ;
+
+	/* Checks 1 through 3 */
+	int ret = ecc_is_pubkey_valid_partial(curve, pk);
+
+	if (ret)
+		return ret;
+
+	/* Check 4: Verify that nQ is the zero point. */
+	nQ = ecc_alloc_point(pk->ndigits);
+	if (!nQ)
+		return -ENOMEM;
+
+	ecc_point_mult(nQ, pk, curve->n, NULL, curve, pk->ndigits);
+	if (!ecc_point_is_zero(nQ))
+		ret = -EINVAL;
+
+	ecc_free_point(nQ);
+
+	return ret;
+}
+EXPORT_SYMBOL(ecc_is_pubkey_valid_full);
+
+int crypto_ecdh_shared_secret(unsigned int curve_id, unsigned int ndigits,
+			      const u64 *private_key, const u64 *public_key,
+			      u64 *secret)
+{
+	int ret = 0;
+	struct ecc_point *product, *pk;
+	u64 rand_z[ECC_MAX_DIGITS];
+	unsigned int nbytes;
+	const struct ecc_curve *curve = ecc_get_curve(curve_id);
+
+	if (!private_key || !public_key || ndigits > ARRAY_SIZE(rand_z)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	nbytes = ndigits << ECC_DIGITS_TO_BYTES_SHIFT;
+
+	get_random_bytes(rand_z, nbytes);
+
+	pk = ecc_alloc_point(ndigits);
+	if (!pk) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ecc_swap_digits(public_key, pk->x, ndigits);
+	ecc_swap_digits(&public_key[ndigits], pk->y, ndigits);
+	ret = ecc_is_pubkey_valid_partial(curve, pk);
+	if (ret)
+		goto err_alloc_product;
+
+	product = ecc_alloc_point(ndigits);
+	if (!product) {
+		ret = -ENOMEM;
+		goto err_alloc_product;
+	}
+
+	ecc_point_mult(product, pk, private_key, rand_z, curve, ndigits);
+
+	if (ecc_point_is_zero(product)) {
+		ret = -EFAULT;
+		goto err_validity;
+	}
+
+	ecc_swap_digits(product->x, secret, ndigits);
+
+err_validity:
+	memzero_explicit(rand_z, sizeof(rand_z));
+	ecc_free_point(product);
+err_alloc_product:
+	ecc_free_point(pk);
+out:
+	return ret;
+}
+EXPORT_SYMBOL(crypto_ecdh_shared_secret);
+
+MODULE_DESCRIPTION("core elliptic curve module");
+MODULE_LICENSE("Dual BSD/GPL");