zram: add dictionary support to zstd backend

This adds support for pre-trained zstd dictionaries [1] Dictionary is
setup in params once (per-comp) and loaded to Cctx and Dctx by reference,
so we don't allocate extra memory.

TEST
====

*** zstd
/sys/block/zram0/mm_stat
1750654976 504565092 514203648 0 514203648 1 0 34204 34204

*** zstd dict=/etc/zstd-dict-amd64
/sys/block/zram0/mm_stat
1750638592 465851259 475373568 0 475373568 1 0 34185 34185

*** zstd level=8 dict=/etc/zstd-dict-amd64
/sys/block/zram0/mm_stat
1750642688 430765171 439955456 0 439955456 1 0 34185 34185

[1] https://github.com/facebook/zstd/blob/dev/programs/zstd.1.md#dictionary-builder

Link: https://lkml.kernel.org/r/[e-mail]
Signed-off-by: Sergey Senozhatsky <[e-mail]>
Cc: Minchan Kim <[e-mail]>
Cc: Nick Terrell <[e-mail]>
Signed-off-by: Andrew Morton <[e-mail]>

drivers/block/zram/backend_zstd.c

···
 };
 
 struct zstd_params {
+	zstd_custom_mem custom_mem;
+	zstd_cdict *cdict;
+	zstd_ddict *ddict;
 	zstd_parameters cprm;
 };
 
+/*
+ * For C/D dictionaries we need to provide zstd with zstd_custom_mem,
+ * which zstd uses internally to allocate/free memory when needed.
+ *
+ * This means that allocator.customAlloc() can be called from zcomp_compress()
+ * under local-lock (per-CPU compression stream), in which case we must use
+ * GFP_ATOMIC.
+ *
+ * Another complication here is that we can be configured as a swap device.
+ */
+static void *zstd_custom_alloc(void *opaque, size_t size)
+{
+	if (!preemptible())
+		return kvzalloc(size, GFP_ATOMIC);
+
+	return kvzalloc(size, __GFP_KSWAPD_RECLAIM | __GFP_NOWARN);
+}
+
+static void zstd_custom_free(void *opaque, void *address)
+{
+	kvfree(address);
+}
+
 static void zstd_release_params(struct zcomp_params *params)
 {
-	kfree(params->drv_data);
+	struct zstd_params *zp = params->drv_data;
+
+	params->drv_data = NULL;
+	if (!zp)
+		return;
+
+	zstd_free_cdict(zp->cdict);
+	zstd_free_ddict(zp->ddict);
+	kfree(zp);
 }
 
 static int zstd_setup_params(struct zcomp_params *params)
 {
+	zstd_compression_parameters prm;
 	struct zstd_params *zp;
 
 	zp = kzalloc(sizeof(*zp), GFP_KERNEL);
 	if (!zp)
 		return -ENOMEM;
 
+	params->drv_data = zp;
 	if (params->level == ZCOMP_PARAM_NO_LEVEL)
 		params->level = zstd_default_clevel();
 
 	zp->cprm = zstd_get_params(params->level, PAGE_SIZE);
-	params->drv_data = zp;
+
+	zp->custom_mem.customAlloc = zstd_custom_alloc;
+	zp->custom_mem.customFree = zstd_custom_free;
+
+	prm = zstd_get_cparams(params->level, PAGE_SIZE,
+			       params->dict_sz);
+
+	zp->cdict = zstd_create_cdict_byreference(params->dict,
+						  params->dict_sz,
+						  prm,
+						  zp->custom_mem);
+	if (!zp->cdict)
+		goto error;
+
+	zp->ddict = zstd_create_ddict_byreference(params->dict,
+						  params->dict_sz,
+						  zp->custom_mem);
+	if (!zp->ddict)
+		goto error;
 
 	return 0;
+
+error:
+	zstd_release_params(params);
+	return -EINVAL;
 }
 
 static void zstd_destroy(struct zcomp_ctx *ctx)
···
 	if (!zctx)
 		return;
 
-	vfree(zctx->cctx_mem);
-	vfree(zctx->dctx_mem);
+	/*
+	 * If ->cctx_mem and ->dctx_mem were allocated then we didn't use
+	 * C/D dictionary and ->cctx / ->dctx were "embedded" into these
+	 * buffers.
+	 *
+	 * If otherwise then we need to explicitly release ->cctx / ->dctx.
+	 */
+	if (zctx->cctx_mem)
+		vfree(zctx->cctx_mem);
+	else
+		zstd_free_cctx(zctx->cctx);
+
+	if (zctx->dctx_mem)
+		vfree(zctx->dctx_mem);
+	else
+		zstd_free_dctx(zctx->dctx);
+
 	kfree(zctx);
 }
 
···
 		return -ENOMEM;
 
 	ctx->context = zctx;
-	prm = zstd_get_params(params->level, PAGE_SIZE);
-	sz = zstd_cctx_workspace_bound(&prm.cParams);
-	zctx->cctx_mem = vzalloc(sz);
-	if (!zctx->cctx_mem)
-		goto error;
+	if (params->dict_sz == 0) {
+		prm = zstd_get_params(params->level, PAGE_SIZE);
+		sz = zstd_cctx_workspace_bound(&prm.cParams);
+		zctx->cctx_mem = vzalloc(sz);
+		if (!zctx->cctx_mem)
+			goto error;
 
-	zctx->cctx = zstd_init_cctx(zctx->cctx_mem, sz);
-	if (!zctx->cctx)
-		goto error;
+		zctx->cctx = zstd_init_cctx(zctx->cctx_mem, sz);
+		if (!zctx->cctx)
+			goto error;
 
-	sz = zstd_dctx_workspace_bound();
-	zctx->dctx_mem = vzalloc(sz);
-	if (!zctx->dctx_mem)
-		goto error;
+		sz = zstd_dctx_workspace_bound();
+		zctx->dctx_mem = vzalloc(sz);
+		if (!zctx->dctx_mem)
+			goto error;
 
-	zctx->dctx = zstd_init_dctx(zctx->dctx_mem, sz);
-	if (!zctx->dctx)
-		goto error;
+		zctx->dctx = zstd_init_dctx(zctx->dctx_mem, sz);
+		if (!zctx->dctx)
+			goto error;
+	} else {
+		struct zstd_params *zp = params->drv_data;
+
+		zctx->cctx = zstd_create_cctx_advanced(zp->custom_mem);
+		if (!zctx->cctx)
+			goto error;
+
+		zctx->dctx = zstd_create_dctx_advanced(zp->custom_mem);
+		if (!zctx->dctx)
+			goto error;
+	}
 
 	return 0;
 
 error:
+	zstd_release_params(params);
 	zstd_destroy(ctx);
 	return -EINVAL;
 }
···
 	struct zstd_ctx *zctx = ctx->context;
 	size_t ret;
 
-	ret = zstd_compress_cctx(zctx->cctx, req->dst, req->dst_len,
-				 req->src, req->src_len, &zp->cprm);
+	if (params->dict_sz == 0)
+		ret = zstd_compress_cctx(zctx->cctx, req->dst, req->dst_len,
+					 req->src, req->src_len, &zp->cprm);
+	else
+		ret = zstd_compress_using_cdict(zctx->cctx, req->dst,
+						req->dst_len, req->src,
+						req->src_len,
+						zp->cdict);
 	if (zstd_is_error(ret))
 		return -EINVAL;
 	req->dst_len = ret;
···
 static int zstd_decompress(struct zcomp_params *params, struct zcomp_ctx *ctx,
 			   struct zcomp_req *req)
 {
+	struct zstd_params *zp = params->drv_data;
 	struct zstd_ctx *zctx = ctx->context;
 	size_t ret;
 
-	ret = zstd_decompress_dctx(zctx->dctx, req->dst, req->dst_len,
-				   req->src, req->src_len);
+	if (params->dict_sz == 0)
+		ret = zstd_decompress_dctx(zctx->dctx, req->dst, req->dst_len,
+					   req->src, req->src_len);
+	else
+		ret = zstd_decompress_using_ddict(zctx->dctx, req->dst,
+						  req->dst_len, req->src,
+						  req->src_len, zp->ddict);
 	if (zstd_is_error(ret))
 		return -EINVAL;
 	return 0;