Skip to content

Commit

Permalink
Merge bitcoin#600: scratch space: use single allocation
Browse files Browse the repository at this point in the history 
98836b1 scratch: replace frames with "checkpoint" system (Andrew Poelstra)
7623cf2 scratch: save a couple bytes of unnecessarily-allocated memory (Andrew Poelstra)
a7a164f scratch: rename `max_size` to `size`, document that extra will actually be allocated (Andrew Poelstra)
5a4bc0b scratch: unify allocations (Andrew Poelstra)
c2b028a scratch space: thread `error_callback` into all scratch space functions (Andrew Poelstra)
0be1a4a scratch: add magic bytes to beginning of structure (Andrew Poelstra)
92a48a7 scratch space: use single allocation (Andrew Poelstra)

Pull request description:

ACKs for commit 98836b:

Tree-SHA512: 6e251f704644a5f61b24aa05c6f7a31ad8c58d147195079d52fe45daacd28a9fd2f4aaf71273183b99b3795a01a88f8389170d4280489b2a28a14a56e03153d7
  • Loading branch information
@gmaxwell
gmaxwell committed May 26, 2019
2 parents 40839e2 + 98836b1 commit 6c36de7
Show file tree
Hide file tree
Showing 9 changed files with 220 additions and 153 deletions.
11 changes: 7 additions & 4 deletions include/secp256k1.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -285,21 +285,24 @@ SECP256K1_API void secp256k1_context_set_error_callback(
*
* Returns: a newly created scratch space.
* Args: ctx: an existing context object (cannot be NULL)
* In: max_size: maximum amount of memory to allocate
* In: size: amount of memory to be available as scratch space. Some extra
* (<100 bytes) will be allocated for extra accounting.
*/
SECP256K1_API SECP256K1_WARN_UNUSED_RESULT secp256k1_scratch_space* secp256k1_scratch_space_create(
const secp256k1_context* ctx,
size_t max_size
size_t size
) SECP256K1_ARG_NONNULL(1);

/** Destroy a secp256k1 scratch space.
*
* The pointer may not be used afterwards.
* Args: scratch: space to destroy
* Args: ctx: a secp256k1 context object.
* scratch: space to destroy
*/
SECP256K1_API void secp256k1_scratch_space_destroy(
const secp256k1_context* ctx,
secp256k1_scratch_space* scratch
);
) SECP256K1_ARG_NONNULL(1);

/** Parse a variable-length public key into the pubkey object.
*
Expand Down
8 changes: 4 additions & 4 deletions src/bench_ecmult.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -64,7 +64,7 @@ static void bench_ecmult(void* arg) {
size_t iter;

for (iter = 0; iter < iters; ++iter) {
data->ecmult_multi(&data->ctx->ecmult_ctx, data->scratch, &data->output[iter], data->includes_g ? &data->scalars[data->offset1] : NULL, bench_callback, arg, count - includes_g);
data->ecmult_multi(&data->ctx->error_callback, &data->ctx->ecmult_ctx, data->scratch, &data->output[iter], data->includes_g ? &data->scalars[data->offset1] : NULL, bench_callback, arg, count - includes_g);
data->offset1 = (data->offset1 + count) % POINTS;
data->offset2 = (data->offset2 + count - 1) % POINTS;
}
Expand Down Expand Up @@ -154,7 +154,7 @@ int main(int argc, char **argv) {
} else if(have_flag(argc, argv, "simple")) {
printf("Using simple algorithm:\n");
data.ecmult_multi = secp256k1_ecmult_multi_var;
secp256k1_scratch_space_destroy(data.scratch);
secp256k1_scratch_space_destroy(data.ctx, data.scratch);
data.scratch = NULL;
} else {
fprintf(stderr, "%s: unrecognized argument '%s'.\n", argv[0], argv[1]);
Expand Down Expand Up @@ -193,10 +193,10 @@ int main(int argc, char **argv) {
run_test(&data, i << p, 1);
}
}
secp256k1_context_destroy(data.ctx);
if (data.scratch != NULL) {
secp256k1_scratch_space_destroy(data.scratch);
secp256k1_scratch_space_destroy(data.ctx, data.scratch);
}
secp256k1_context_destroy(data.ctx);
free(data.scalars);
free(data.pubkeys);
free(data.seckeys);
Expand Down
2 changes: 1 addition & 1 deletion src/ecmult.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -43,6 +43,6 @@ typedef int (secp256k1_ecmult_multi_callback)(secp256k1_scalar *sc, secp256k1_ge
* 0 if there is not enough scratch space for a single point or
* callback returns 0
*/
static int secp256k1_ecmult_multi_var(const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n);
static int secp256k1_ecmult_multi_var(const secp256k1_callback* error_callback, const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n);

#endif /* SECP256K1_ECMULT_H */
84 changes: 47 additions & 37 deletions src/ecmult_impl.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -648,52 +648,55 @@ static size_t secp256k1_strauss_scratch_size(size_t n_points) {
return n_points*point_size;
}

static int secp256k1_ecmult_strauss_batch(const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n_points, size_t cb_offset) {
static int secp256k1_ecmult_strauss_batch(const secp256k1_callback* error_callback, const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n_points, size_t cb_offset) {
secp256k1_gej* points;
secp256k1_scalar* scalars;
struct secp256k1_strauss_state state;
size_t i;
const size_t scratch_checkpoint = secp256k1_scratch_checkpoint(error_callback, scratch);

secp256k1_gej_set_infinity(r);
if (inp_g_sc == NULL && n_points == 0) {
return 1;
}

if (!secp256k1_scratch_allocate_frame(scratch, secp256k1_strauss_scratch_size(n_points), STRAUSS_SCRATCH_OBJECTS)) {
return 0;
}
points = (secp256k1_gej*)secp256k1_scratch_alloc(scratch, n_points * sizeof(secp256k1_gej));
scalars = (secp256k1_scalar*)secp256k1_scratch_alloc(scratch, n_points * sizeof(secp256k1_scalar));
state.prej = (secp256k1_gej*)secp256k1_scratch_alloc(scratch, n_points * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_gej));
state.zr = (secp256k1_fe*)secp256k1_scratch_alloc(scratch, n_points * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_fe));
points = (secp256k1_gej*)secp256k1_scratch_alloc(error_callback, scratch, n_points * sizeof(secp256k1_gej));
scalars = (secp256k1_scalar*)secp256k1_scratch_alloc(error_callback, scratch, n_points * sizeof(secp256k1_scalar));
state.prej = (secp256k1_gej*)secp256k1_scratch_alloc(error_callback, scratch, n_points * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_gej));
state.zr = (secp256k1_fe*)secp256k1_scratch_alloc(error_callback, scratch, n_points * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_fe));
#ifdef USE_ENDOMORPHISM
state.pre_a = (secp256k1_ge*)secp256k1_scratch_alloc(scratch, n_points * 2 * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_ge));
state.pre_a = (secp256k1_ge*)secp256k1_scratch_alloc(error_callback, scratch, n_points * 2 * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_ge));
state.pre_a_lam = state.pre_a + n_points * ECMULT_TABLE_SIZE(WINDOW_A);
#else
state.pre_a = (secp256k1_ge*)secp256k1_scratch_alloc(scratch, n_points * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_ge));
state.pre_a = (secp256k1_ge*)secp256k1_scratch_alloc(error_callback, scratch, n_points * ECMULT_TABLE_SIZE(WINDOW_A) * sizeof(secp256k1_ge));
#endif
state.ps = (struct secp256k1_strauss_point_state*)secp256k1_scratch_alloc(scratch, n_points * sizeof(struct secp256k1_strauss_point_state));
state.ps = (struct secp256k1_strauss_point_state*)secp256k1_scratch_alloc(error_callback, scratch, n_points * sizeof(struct secp256k1_strauss_point_state));

if (points == NULL || scalars == NULL || state.prej == NULL || state.zr == NULL || state.pre_a == NULL) {
secp256k1_scratch_apply_checkpoint(error_callback, scratch, scratch_checkpoint);
return 0;
}

for (i = 0; i < n_points; i++) {
secp256k1_ge point;
if (!cb(&scalars[i], &point, i+cb_offset, cbdata)) {
secp256k1_scratch_deallocate_frame(scratch);
secp256k1_scratch_apply_checkpoint(error_callback, scratch, scratch_checkpoint);
return 0;
}
secp256k1_gej_set_ge(&points[i], &point);
}
secp256k1_ecmult_strauss_wnaf(ctx, &state, r, n_points, points, scalars, inp_g_sc);
secp256k1_scratch_deallocate_frame(scratch);
secp256k1_scratch_apply_checkpoint(error_callback, scratch, scratch_checkpoint);
return 1;
}

/* Wrapper for secp256k1_ecmult_multi_func interface */
static int secp256k1_ecmult_strauss_batch_single(const secp256k1_ecmult_context *actx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n) {
return secp256k1_ecmult_strauss_batch(actx, scratch, r, inp_g_sc, cb, cbdata, n, 0);
static int secp256k1_ecmult_strauss_batch_single(const secp256k1_callback* error_callback, const secp256k1_ecmult_context *actx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n) {
return secp256k1_ecmult_strauss_batch(error_callback, actx, scratch, r, inp_g_sc, cb, cbdata, n, 0);
}

static size_t secp256k1_strauss_max_points(secp256k1_scratch *scratch) {
return secp256k1_scratch_max_allocation(scratch, STRAUSS_SCRATCH_OBJECTS) / secp256k1_strauss_scratch_size(1);
static size_t secp256k1_strauss_max_points(const secp256k1_callback* error_callback, secp256k1_scratch *scratch) {
return secp256k1_scratch_max_allocation(error_callback, scratch, STRAUSS_SCRATCH_OBJECTS) / secp256k1_strauss_scratch_size(1);
}

/** Convert a number to WNAF notation.
Expand Down Expand Up @@ -985,7 +988,8 @@ static size_t secp256k1_pippenger_scratch_size(size_t n_points, int bucket_windo
return (sizeof(secp256k1_gej) << bucket_window) + sizeof(struct secp256k1_pippenger_state) + entries * entry_size;
}

static int secp256k1_ecmult_pippenger_batch(const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n_points, size_t cb_offset) {
static int secp256k1_ecmult_pippenger_batch(const secp256k1_callback* error_callback, const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n_points, size_t cb_offset) {
const size_t scratch_checkpoint = secp256k1_scratch_checkpoint(error_callback, scratch);
/* Use 2(n+1) with the endomorphism, n+1 without, when calculating batch
* sizes. The reason for +1 is that we add the G scalar to the list of
* other scalars. */
Expand All @@ -1010,15 +1014,21 @@ static int secp256k1_ecmult_pippenger_batch(const secp256k1_ecmult_context *ctx,
}

bucket_window = secp256k1_pippenger_bucket_window(n_points);
if (!secp256k1_scratch_allocate_frame(scratch, secp256k1_pippenger_scratch_size(n_points, bucket_window), PIPPENGER_SCRATCH_OBJECTS)) {
points = (secp256k1_ge *) secp256k1_scratch_alloc(error_callback, scratch, entries * sizeof(*points));
scalars = (secp256k1_scalar *) secp256k1_scratch_alloc(error_callback, scratch, entries * sizeof(*scalars));
state_space = (struct secp256k1_pippenger_state *) secp256k1_scratch_alloc(error_callback, scratch, sizeof(*state_space));
if (points == NULL || scalars == NULL || state_space == NULL) {
secp256k1_scratch_apply_checkpoint(error_callback, scratch, scratch_checkpoint);
return 0;
}

state_space->ps = (struct secp256k1_pippenger_point_state *) secp256k1_scratch_alloc(error_callback, scratch, entries * sizeof(*state_space->ps));
state_space->wnaf_na = (int *) secp256k1_scratch_alloc(error_callback, scratch, entries*(WNAF_SIZE(bucket_window+1)) * sizeof(int));
buckets = (secp256k1_gej *) secp256k1_scratch_alloc(error_callback, scratch, (1<<bucket_window) * sizeof(*buckets));
if (state_space->ps == NULL || state_space->wnaf_na == NULL || buckets == NULL) {
secp256k1_scratch_apply_checkpoint(error_callback, scratch, scratch_checkpoint);
return 0;
}
points = (secp256k1_ge *) secp256k1_scratch_alloc(scratch, entries * sizeof(*points));
scalars = (secp256k1_scalar *) secp256k1_scratch_alloc(scratch, entries * sizeof(*scalars));
state_space = (struct secp256k1_pippenger_state *) secp256k1_scratch_alloc(scratch, sizeof(*state_space));
state_space->ps = (struct secp256k1_pippenger_point_state *) secp256k1_scratch_alloc(scratch, entries * sizeof(*state_space->ps));
state_space->wnaf_na = (int *) secp256k1_scratch_alloc(scratch, entries*(WNAF_SIZE(bucket_window+1)) * sizeof(int));
buckets = (secp256k1_gej *) secp256k1_scratch_alloc(scratch, sizeof(*buckets) << bucket_window);

if (inp_g_sc != NULL) {
scalars[0] = *inp_g_sc;
Expand All @@ -1032,7 +1042,7 @@ static int secp256k1_ecmult_pippenger_batch(const secp256k1_ecmult_context *ctx,

while (point_idx < n_points) {
if (!cb(&scalars[idx], &points[idx], point_idx + cb_offset, cbdata)) {
secp256k1_scratch_deallocate_frame(scratch);
secp256k1_scratch_apply_checkpoint(error_callback, scratch, scratch_checkpoint);
return 0;
}
idx++;
Expand All @@ -1056,22 +1066,22 @@ static int secp256k1_ecmult_pippenger_batch(const secp256k1_ecmult_context *ctx,
for(i = 0; i < 1<<bucket_window; i++) {
secp256k1_gej_clear(&buckets[i]);
}
secp256k1_scratch_deallocate_frame(scratch);
secp256k1_scratch_apply_checkpoint(error_callback, scratch, scratch_checkpoint);
return 1;
}

/* Wrapper for secp256k1_ecmult_multi_func interface */
static int secp256k1_ecmult_pippenger_batch_single(const secp256k1_ecmult_context *actx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n) {
return secp256k1_ecmult_pippenger_batch(actx, scratch, r, inp_g_sc, cb, cbdata, n, 0);
static int secp256k1_ecmult_pippenger_batch_single(const secp256k1_callback* error_callback, const secp256k1_ecmult_context *actx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n) {
return secp256k1_ecmult_pippenger_batch(error_callback, actx, scratch, r, inp_g_sc, cb, cbdata, n, 0);
}

/**
* Returns the maximum number of points in addition to G that can be used with
* a given scratch space. The function ensures that fewer points may also be
* used.
*/
static size_t secp256k1_pippenger_max_points(secp256k1_scratch *scratch) {
size_t max_alloc = secp256k1_scratch_max_allocation(scratch, PIPPENGER_SCRATCH_OBJECTS);
static size_t secp256k1_pippenger_max_points(const secp256k1_callback* error_callback, secp256k1_scratch *scratch) {
size_t max_alloc = secp256k1_scratch_max_allocation(error_callback, scratch, PIPPENGER_SCRATCH_OBJECTS);
int bucket_window;
size_t res = 0;

Expand Down Expand Up @@ -1153,11 +1163,11 @@ static int secp256k1_ecmult_multi_batch_size_helper(size_t *n_batches, size_t *n
return 1;
}

typedef int (*secp256k1_ecmult_multi_func)(const secp256k1_ecmult_context*, secp256k1_scratch*, secp256k1_gej*, const secp256k1_scalar*, secp256k1_ecmult_multi_callback cb, void*, size_t);
static int secp256k1_ecmult_multi_var(const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n) {
typedef int (*secp256k1_ecmult_multi_func)(const secp256k1_callback* error_callback, const secp256k1_ecmult_context*, secp256k1_scratch*, secp256k1_gej*, const secp256k1_scalar*, secp256k1_ecmult_multi_callback cb, void*, size_t);
static int secp256k1_ecmult_multi_var(const secp256k1_callback* error_callback, const secp256k1_ecmult_context *ctx, secp256k1_scratch *scratch, secp256k1_gej *r, const secp256k1_scalar *inp_g_sc, secp256k1_ecmult_multi_callback cb, void *cbdata, size_t n) {
size_t i;

int (*f)(const secp256k1_ecmult_context*, secp256k1_scratch*, secp256k1_gej*, const secp256k1_scalar*, secp256k1_ecmult_multi_callback cb, void*, size_t, size_t);
int (*f)(const secp256k1_callback* error_callback, const secp256k1_ecmult_context*, secp256k1_scratch*, secp256k1_gej*, const secp256k1_scalar*, secp256k1_ecmult_multi_callback cb, void*, size_t, size_t);
size_t n_batches;
size_t n_batch_points;

Expand All @@ -1178,13 +1188,13 @@ static int secp256k1_ecmult_multi_var(const secp256k1_ecmult_context *ctx, secp2
* a threshold use Pippenger's algorithm. Otherwise use Strauss' algorithm.
* As a first step check if there's enough space for Pippenger's algo (which requires less space
* than Strauss' algo) and if not, use the simple algorithm. */
if (!secp256k1_ecmult_multi_batch_size_helper(&n_batches, &n_batch_points, secp256k1_pippenger_max_points(scratch), n)) {
if (!secp256k1_ecmult_multi_batch_size_helper(&n_batches, &n_batch_points, secp256k1_pippenger_max_points(error_callback, scratch), n)) {
return secp256k1_ecmult_multi_simple_var(ctx, r, inp_g_sc, cb, cbdata, n);
}
if (n_batch_points >= ECMULT_PIPPENGER_THRESHOLD) {
f = secp256k1_ecmult_pippenger_batch;
} else {
if (!secp256k1_ecmult_multi_batch_size_helper(&n_batches, &n_batch_points, secp256k1_strauss_max_points(scratch), n)) {
if (!secp256k1_ecmult_multi_batch_size_helper(&n_batches, &n_batch_points, secp256k1_strauss_max_points(error_callback, scratch), n)) {
return secp256k1_ecmult_multi_simple_var(ctx, r, inp_g_sc, cb, cbdata, n);
}
f = secp256k1_ecmult_strauss_batch;
Expand All @@ -1193,7 +1203,7 @@ static int secp256k1_ecmult_multi_var(const secp256k1_ecmult_context *ctx, secp2
size_t nbp = n < n_batch_points ? n : n_batch_points;
size_t offset = n_batch_points*i;
secp256k1_gej tmp;
if (!f(ctx, scratch, &tmp, i == 0 ? inp_g_sc : NULL, cb, cbdata, nbp, offset)) {
if (!f(error_callback, ctx, scratch, &tmp, i == 0 ? inp_g_sc : NULL, cb, cbdata, nbp, offset)) {
return 0;
}
secp256k1_gej_add_var(r, r, &tmp, NULL);
Expand Down
31 changes: 17 additions & 14 deletions src/scratch.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,33 +7,36 @@
#ifndef _SECP256K1_SCRATCH_
#define _SECP256K1_SCRATCH_

#define SECP256K1_SCRATCH_MAX_FRAMES 5

/* The typedef is used internally; the struct name is used in the public API
* (where it is exposed as a different typedef) */
typedef struct secp256k1_scratch_space_struct {
void *data[SECP256K1_SCRATCH_MAX_FRAMES];
size_t offset[SECP256K1_SCRATCH_MAX_FRAMES];
size_t frame_size[SECP256K1_SCRATCH_MAX_FRAMES];
size_t frame;
/** guard against interpreting this object as other types */
unsigned char magic[8];
/** actual allocated data */
void *data;
/** amount that has been allocated (i.e. `data + offset` is the next
* available pointer) */
size_t alloc_size;
/** maximum size available to allocate */
size_t max_size;
const secp256k1_callback* error_callback;
} secp256k1_scratch;

static secp256k1_scratch* secp256k1_scratch_create(const secp256k1_callback* error_callback, size_t max_size);

static void secp256k1_scratch_destroy(secp256k1_scratch* scratch);
static void secp256k1_scratch_destroy(const secp256k1_callback* error_callback, secp256k1_scratch* scratch);

/** Attempts to allocate a new stack frame with `n` available bytes. Returns 1 on success, 0 on failure */
static int secp256k1_scratch_allocate_frame(secp256k1_scratch* scratch, size_t n, size_t objects);
/** Returns an opaque object used to "checkpoint" a scratch space. Used
* with `secp256k1_scratch_apply_checkpoint` to undo allocations. */
static size_t secp256k1_scratch_checkpoint(const secp256k1_callback* error_callback, const secp256k1_scratch* scratch);

/** Deallocates a stack frame */
static void secp256k1_scratch_deallocate_frame(secp256k1_scratch* scratch);
/** Applies a check point received from `secp256k1_scratch_checkpoint`,
* undoing all allocations since that point. */
static void secp256k1_scratch_apply_checkpoint(const secp256k1_callback* error_callback, secp256k1_scratch* scratch, size_t checkpoint);

/** Returns the maximum allocation the scratch space will allow */
static size_t secp256k1_scratch_max_allocation(const secp256k1_scratch* scratch, size_t n_objects);
static size_t secp256k1_scratch_max_allocation(const secp256k1_callback* error_callback, const secp256k1_scratch* scratch, size_t n_objects);

/** Returns a pointer into the most recently allocated frame, or NULL if there is insufficient available space */
static void *secp256k1_scratch_alloc(secp256k1_scratch* scratch, size_t n);
static void *secp256k1_scratch_alloc(const secp256k1_callback* error_callback, secp256k1_scratch* scratch, size_t n);

#endif
Loading

0 comments on commit 6c36de7

Please sign in to comment.