1 files changed, 308 insertions, 171 deletions

diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c
index 332ec5f74002..d5d090276391 100644
--- a/fs/erofs/decompressor.c
+++ b/fs/erofs/decompressor.c
@@ -2,33 +2,18 @@
 /*
  * Copyright (C) 2019 HUAWEI, Inc.
  *             https://www.huawei.com/
+ * Copyright (C) 2024 Alibaba Cloud
  */
 #include "compress.h"
-#include <linux/module.h>
 #include <linux/lz4.h>
 
-#ifndef LZ4_DISTANCE_MAX	/* history window size */
-#define LZ4_DISTANCE_MAX 65535	/* set to maximum value by default */
-#endif
-
 #define LZ4_MAX_DISTANCE_PAGES	(DIV_ROUND_UP(LZ4_DISTANCE_MAX, PAGE_SIZE) + 1)
-#ifndef LZ4_DECOMPRESS_INPLACE_MARGIN
-#define LZ4_DECOMPRESS_INPLACE_MARGIN(srcsize)  (((srcsize) >> 8) + 32)
-#endif
-
-struct z_erofs_lz4_decompress_ctx {
-	struct z_erofs_decompress_req *rq;
-	/* # of encoded, decoded pages */
-	unsigned int inpages, outpages;
-	/* decoded block total length (used for in-place decompression) */
-	unsigned int oend;
-};
 
-int z_erofs_load_lz4_config(struct super_block *sb,
-			    struct erofs_super_block *dsb,
-			    struct z_erofs_lz4_cfgs *lz4, int size)
+static int z_erofs_load_lz4_config(struct super_block *sb,
+			    struct erofs_super_block *dsb, void *data, int size)
 {
 	struct erofs_sb_info *sbi = EROFS_SB(sb);
+	struct z_erofs_lz4_cfgs *lz4 = data;
 	u16 distance;
 
 	if (lz4) {
@@ -55,17 +40,16 @@ int z_erofs_load_lz4_config(struct super_block *sb,
 	sbi->lz4.max_distance_pages = distance ?
 					DIV_ROUND_UP(distance, PAGE_SIZE) + 1 :
 					LZ4_MAX_DISTANCE_PAGES;
-	return erofs_pcpubuf_growsize(sbi->lz4.max_pclusterblks);
+	return z_erofs_gbuf_growsize(sbi->lz4.max_pclusterblks);
 }
 
 /*
  * Fill all gaps with bounce pages if it's a sparse page list. Also check if
  * all physical pages are consecutive, which can be seen for moderate CR.
  */
-static int z_erofs_lz4_prepare_dstpages(struct z_erofs_lz4_decompress_ctx *ctx,
+static int z_erofs_lz4_prepare_dstpages(struct z_erofs_decompress_req *rq,
 					struct page **pagepool)
 {
-	struct z_erofs_decompress_req *rq = ctx->rq;
 	struct page *availables[LZ4_MAX_DISTANCE_PAGES] = { NULL };
 	unsigned long bounced[DIV_ROUND_UP(LZ4_MAX_DISTANCE_PAGES,
 					   BITS_PER_LONG)] = { 0 };
@@ -75,7 +59,7 @@ static int z_erofs_lz4_prepare_dstpages(struct z_erofs_lz4_decompress_ctx *ctx,
 	unsigned int i, j, top;
 
 	top = 0;
-	for (i = j = 0; i < ctx->outpages; ++i, ++j) {
+	for (i = j = 0; i < rq->outpages; ++i, ++j) {
 		struct page *const page = rq->out[i];
 		struct page *victim;
 
@@ -110,10 +94,10 @@ static int z_erofs_lz4_prepare_dstpages(struct z_erofs_lz4_decompress_ctx *ctx,
 
 		if (top) {
 			victim = availables[--top];
-			get_page(victim);
 		} else {
-			victim = erofs_allocpage(pagepool,
-						 GFP_KERNEL | __GFP_NOFAIL);
+			victim = __erofs_allocpage(pagepool, rq->gfp, true);
+			if (!victim)
+				return -ENOMEM;
 			set_page_private(victim, Z_EROFS_SHORTLIVED_PAGE);
 		}
 		rq->out[i] = victim;
@@ -121,65 +105,72 @@ static int z_erofs_lz4_prepare_dstpages(struct z_erofs_lz4_decompress_ctx *ctx,
 	return kaddr ? 1 : 0;
 }
 
-static void *z_erofs_lz4_handle_overlap(struct z_erofs_lz4_decompress_ctx *ctx,
-			void *inpage, unsigned int *inputmargin, int *maptype,
-			bool may_inplace)
+static void *z_erofs_lz4_handle_overlap(const struct z_erofs_decompress_req *rq,
+			void *inpage, void *out, unsigned int *inputmargin,
+			int *maptype, bool may_inplace)
 {
-	struct z_erofs_decompress_req *rq = ctx->rq;
-	unsigned int omargin, total, i, j;
+	unsigned int oend, omargin, cnt, i;
 	struct page **in;
-	void *src, *tmp;
-
-	if (rq->inplace_io) {
-		omargin = PAGE_ALIGN(ctx->oend) - ctx->oend;
-		if (rq->partial_decoding || !may_inplace ||
-		    omargin < LZ4_DECOMPRESS_INPLACE_MARGIN(rq->inputsize))
-			goto docopy;
-
-		for (i = 0; i < ctx->inpages; ++i) {
-			DBG_BUGON(rq->in[i] == NULL);
-			for (j = 0; j < ctx->outpages - ctx->inpages + i; ++j)
-				if (rq->out[j] == rq->in[i])
-					goto docopy;
+	void *src;
+
+	/*
+	 * If in-place I/O isn't used, for example, the bounce compressed cache
+	 * can hold data for incomplete read requests. Just map the compressed
+	 * buffer as well and decompress directly.
+	 */
+	if (!rq->inplace_io) {
+		if (rq->inpages <= 1) {
+			*maptype = 0;
+			return inpage;
 		}
+		kunmap_local(inpage);
+		src = erofs_vm_map_ram(rq->in, rq->inpages);
+		if (!src)
+			return ERR_PTR(-ENOMEM);
+		*maptype = 1;
+		return src;
 	}
-
-	if (ctx->inpages <= 1) {
-		*maptype = 0;
-		return inpage;
+	/*
+	 * Then, deal with in-place I/Os. The reasons why in-place I/O is useful
+	 * are: (1) It minimizes memory footprint during the I/O submission,
+	 * which is useful for slow storage (including network devices and
+	 * low-end HDDs/eMMCs) but with a lot inflight I/Os; (2) If in-place
+	 * decompression can also be applied, it will reuse the unique buffer so
+	 * that no extra CPU D-cache is polluted with temporary compressed data
+	 * for extreme performance.
+	 */
+	oend = rq->pageofs_out + rq->outputsize;
+	omargin = PAGE_ALIGN(oend) - oend;
+	if (!rq->partial_decoding && may_inplace &&
+	    omargin >= LZ4_DECOMPRESS_INPLACE_MARGIN(rq->inputsize)) {
+		for (i = 0; i < rq->inpages; ++i)
+			if (rq->out[rq->outpages - rq->inpages + i] !=
+			    rq->in[i])
+				break;
+		if (i >= rq->inpages) {
+			kunmap_local(inpage);
+			*maptype = 3;
+			return out + ((rq->outpages - rq->inpages) << PAGE_SHIFT);
+		}
 	}
-	kunmap_local(inpage);
-	might_sleep();
-	src = erofs_vm_map_ram(rq->in, ctx->inpages);
-	if (!src)
-		return ERR_PTR(-ENOMEM);
-	*maptype = 1;
-	return src;
-
-docopy:
-	/* Or copy compressed data which can be overlapped to per-CPU buffer */
-	in = rq->in;
-	src = erofs_get_pcpubuf(ctx->inpages);
+	/*
+	 * If in-place decompression can't be applied, copy compressed data that
+	 * may potentially overlap during decompression to a per-CPU buffer.
+	 */
+	src = z_erofs_get_gbuf(rq->inpages);
 	if (!src) {
 		DBG_BUGON(1);
 		kunmap_local(inpage);
 		return ERR_PTR(-EFAULT);
 	}
 
-	tmp = src;
-	total = rq->inputsize;
-	while (total) {
-		unsigned int page_copycnt =
-			min_t(unsigned int, total, PAGE_SIZE - *inputmargin);
-
+	for (i = 0, in = rq->in; i < rq->inputsize; i += cnt, ++in) {
+		cnt = min_t(u32, rq->inputsize - i, PAGE_SIZE - *inputmargin);
 		if (!inpage)
 			inpage = kmap_local_page(*in);
-		memcpy(tmp, inpage + *inputmargin, page_copycnt);
+		memcpy(src + i, inpage + *inputmargin, cnt);
 		kunmap_local(inpage);
 		inpage = NULL;
-		tmp += page_copycnt;
-		total -= page_copycnt;
-		++in;
 		*inputmargin = 0;
 	}
 	*maptype = 2;
@@ -187,30 +178,29 @@ docopy:
 }
 
 /*
- * Get the exact inputsize with zero_padding feature.
- *  - For LZ4, it should work if zero_padding feature is on (5.3+);
- *  - For MicroLZMA, it'd be enabled all the time.
+ * Get the exact on-disk size of the compressed data:
+ *  - For LZ4, it should apply if the zero_padding feature is on (5.3+);
+ *  - For others, zero_padding is enabled all the time.
  */
-int z_erofs_fixup_insize(struct z_erofs_decompress_req *rq, const char *padbuf,
-			 unsigned int padbufsize)
+const char *z_erofs_fixup_insize(struct z_erofs_decompress_req *rq,
+				 const char *padbuf, unsigned int padbufsize)
 {
 	const char *padend;
 
 	padend = memchr_inv(padbuf, 0, padbufsize);
 	if (!padend)
-		return -EFSCORRUPTED;
+		return "compressed data start not found";
 	rq->inputsize -= padend - padbuf;
 	rq->pageofs_in += padend - padbuf;
-	return 0;
+	return NULL;
 }
 
-static int z_erofs_lz4_decompress_mem(struct z_erofs_lz4_decompress_ctx *ctx,
-				      u8 *out)
+static int z_erofs_lz4_decompress_mem(struct z_erofs_decompress_req *rq, u8 *dst)
 {
-	struct z_erofs_decompress_req *rq = ctx->rq;
 	bool support_0padding = false, may_inplace = false;
 	unsigned int inputmargin;
-	u8 *headpage, *src;
+	u8 *out, *headpage, *src;
+	const char *reason;
 	int ret, maptype;
 
 	DBG_BUGON(*rq->in == NULL);
@@ -219,23 +209,24 @@ static int z_erofs_lz4_decompress_mem(struct z_erofs_lz4_decompress_ctx *ctx,
 	/* LZ4 decompression inplace is only safe if zero_padding is enabled */
 	if (erofs_sb_has_zero_padding(EROFS_SB(rq->sb))) {
 		support_0padding = true;
-		ret = z_erofs_fixup_insize(rq, headpage + rq->pageofs_in,
+		reason = z_erofs_fixup_insize(rq, headpage + rq->pageofs_in,
 				min_t(unsigned int, rq->inputsize,
 				      rq->sb->s_blocksize - rq->pageofs_in));
-		if (ret) {
+		if (reason) {
 			kunmap_local(headpage);
-			return ret;
+			return IS_ERR(reason) ? PTR_ERR(reason) : -EFSCORRUPTED;
 		}
 		may_inplace = !((rq->pageofs_in + rq->inputsize) &
 				(rq->sb->s_blocksize - 1));
 	}
 
 	inputmargin = rq->pageofs_in;
-	src = z_erofs_lz4_handle_overlap(ctx, headpage, &inputmargin,
+	src = z_erofs_lz4_handle_overlap(rq, headpage, dst, &inputmargin,
 					 &maptype, may_inplace);
 	if (IS_ERR(src))
 		return PTR_ERR(src);
 
+	out = dst + rq->pageofs_out;
 	/* legacy format could compress extra data in a pcluster. */
 	if (rq->partial_decoding || !support_0padding)
 		ret = LZ4_decompress_safe_partial(src + inputmargin, out,
@@ -245,17 +236,9 @@ static int z_erofs_lz4_decompress_mem(struct z_erofs_lz4_decompress_ctx *ctx,
 					  rq->inputsize, rq->outputsize);
 
 	if (ret != rq->outputsize) {
-		erofs_err(rq->sb, "failed to decompress %d in[%u, %u] out[%u]",
-			  ret, rq->inputsize, inputmargin, rq->outputsize);
-
-		print_hex_dump(KERN_DEBUG, "[ in]: ", DUMP_PREFIX_OFFSET,
-			       16, 1, src + inputmargin, rq->inputsize, true);
-		print_hex_dump(KERN_DEBUG, "[out]: ", DUMP_PREFIX_OFFSET,
-			       16, 1, out, rq->outputsize, true);
-
 		if (ret >= 0)
 			memset(out + ret, 0, rq->outputsize - ret);
-		ret = -EIO;
+		ret = -EFSCORRUPTED;
 	} else {
 		ret = 0;
 	}
@@ -263,126 +246,280 @@ static int z_erofs_lz4_decompress_mem(struct z_erofs_lz4_decompress_ctx *ctx,
 	if (maptype == 0) {
 		kunmap_local(headpage);
 	} else if (maptype == 1) {
-		vm_unmap_ram(src, ctx->inpages);
+		vm_unmap_ram(src, rq->inpages);
 	} else if (maptype == 2) {
-		erofs_put_pcpubuf(src);
-	} else {
+		z_erofs_put_gbuf(src);
+	} else if (maptype != 3) {
 		DBG_BUGON(1);
 		return -EFAULT;
 	}
 	return ret;
 }
 
-static int z_erofs_lz4_decompress(struct z_erofs_decompress_req *rq,
-				  struct page **pagepool)
+static const char *z_erofs_lz4_decompress(struct z_erofs_decompress_req *rq,
+					  struct page **pagepool)
 {
-	struct z_erofs_lz4_decompress_ctx ctx;
 	unsigned int dst_maptype;
 	void *dst;
 	int ret;
 
-	ctx.rq = rq;
-	ctx.oend = rq->pageofs_out + rq->outputsize;
-	ctx.outpages = PAGE_ALIGN(ctx.oend) >> PAGE_SHIFT;
-	ctx.inpages = PAGE_ALIGN(rq->inputsize) >> PAGE_SHIFT;
-
 	/* one optimized fast path only for non bigpcluster cases yet */
-	if (ctx.inpages == 1 && ctx.outpages == 1 && !rq->inplace_io) {
+	if (rq->inpages == 1 && rq->outpages == 1 && !rq->inplace_io) {
 		DBG_BUGON(!*rq->out);
 		dst = kmap_local_page(*rq->out);
 		dst_maptype = 0;
-		goto dstmap_out;
-	}
-
-	/* general decoding path which can be used for all cases */
-	ret = z_erofs_lz4_prepare_dstpages(&ctx, pagepool);
-	if (ret < 0) {
-		return ret;
-	} else if (ret > 0) {
-		dst = page_address(*rq->out);
-		dst_maptype = 1;
 	} else {
-		dst = erofs_vm_map_ram(rq->out, ctx.outpages);
-		if (!dst)
-			return -ENOMEM;
-		dst_maptype = 2;
+		/* general decoding path which can be used for all cases */
+		ret = z_erofs_lz4_prepare_dstpages(rq, pagepool);
+		if (ret < 0)
+			return ERR_PTR(ret);
+		if (ret > 0) {
+			dst = page_address(*rq->out);
+			dst_maptype = 1;
+		} else {
+			dst = erofs_vm_map_ram(rq->out, rq->outpages);
+			if (!dst)
+				return ERR_PTR(-ENOMEM);
+			dst_maptype = 2;
+		}
 	}
-
-dstmap_out:
-	ret = z_erofs_lz4_decompress_mem(&ctx, dst + rq->pageofs_out);
+	ret = z_erofs_lz4_decompress_mem(rq, dst);
 	if (!dst_maptype)
 		kunmap_local(dst);
 	else if (dst_maptype == 2)
-		vm_unmap_ram(dst, ctx.outpages);
-	return ret;
+		vm_unmap_ram(dst, rq->outpages);
+	return ERR_PTR(ret);
 }
 
-static int z_erofs_transform_plain(struct z_erofs_decompress_req *rq,
-				   struct page **pagepool)
+static const char *z_erofs_transform_plain(struct z_erofs_decompress_req *rq,
+					   struct page **pagepool)
 {
-	const unsigned int inpages = PAGE_ALIGN(rq->inputsize) >> PAGE_SHIFT;
-	const unsigned int outpages =
-		PAGE_ALIGN(rq->pageofs_out + rq->outputsize) >> PAGE_SHIFT;
-	const unsigned int righthalf = min_t(unsigned int, rq->outputsize,
-					     PAGE_SIZE - rq->pageofs_out);
-	const unsigned int lefthalf = rq->outputsize - righthalf;
-	const unsigned int interlaced_offset =
-		rq->alg == Z_EROFS_COMPRESSION_SHIFTED ? 0 : rq->pageofs_out;
-	u8 *src;
-
-	if (outpages > 2 && rq->alg == Z_EROFS_COMPRESSION_SHIFTED) {
-		DBG_BUGON(1);
-		return -EFSCORRUPTED;
+	const unsigned int nrpages_in = rq->inpages, nrpages_out = rq->outpages;
+	const unsigned int bs = rq->sb->s_blocksize;
+	unsigned int cur = 0, ni = 0, no, pi, po, insz, cnt;
+	u8 *kin;
+
+	if (rq->outputsize > rq->inputsize)
+		return ERR_PTR(-EOPNOTSUPP);
+	if (rq->alg == Z_EROFS_COMPRESSION_INTERLACED) {
+		cur = bs - (rq->pageofs_out & (bs - 1));
+		pi = (rq->pageofs_in + rq->inputsize - cur) & ~PAGE_MASK;
+		cur = min(cur, rq->outputsize);
+		if (cur && rq->out[0]) {
+			kin = kmap_local_page(rq->in[nrpages_in - 1]);
+			if (rq->out[0] == rq->in[nrpages_in - 1])
+				memmove(kin + rq->pageofs_out, kin + pi, cur);
+			else
+				memcpy_to_page(rq->out[0], rq->pageofs_out,
+					       kin + pi, cur);
+			kunmap_local(kin);
+		}
+		rq->outputsize -= cur;
 	}
 
-	if (rq->out[0] == *rq->in) {
-		DBG_BUGON(rq->pageofs_out);
-		return 0;
+	for (; rq->outputsize; rq->pageofs_in = 0, cur += insz, ni++) {
+		insz = min(PAGE_SIZE - rq->pageofs_in, rq->outputsize);
+		rq->outputsize -= insz;
+		if (!rq->in[ni])
+			continue;
+		kin = kmap_local_page(rq->in[ni]);
+		pi = 0;
+		do {
+			no = (rq->pageofs_out + cur + pi) >> PAGE_SHIFT;
+			po = (rq->pageofs_out + cur + pi) & ~PAGE_MASK;
+			DBG_BUGON(no >= nrpages_out);
+			cnt = min(insz - pi, PAGE_SIZE - po);
+			if (rq->out[no] == rq->in[ni])
+				memmove(kin + po,
+					kin + rq->pageofs_in + pi, cnt);
+			else if (rq->out[no])
+				memcpy_to_page(rq->out[no], po,
+					       kin + rq->pageofs_in + pi, cnt);
+			pi += cnt;
+		} while (pi < insz);
+		kunmap_local(kin);
 	}
+	DBG_BUGON(ni > nrpages_in);
+	return NULL;
+}
 
-	src = kmap_local_page(rq->in[inpages - 1]) + rq->pageofs_in;
-	if (rq->out[0])
-		memcpy_to_page(rq->out[0], rq->pageofs_out,
-			       src + interlaced_offset, righthalf);
-
-	if (outpages > inpages) {
-		DBG_BUGON(!rq->out[outpages - 1]);
-		if (rq->out[outpages - 1] != rq->in[inpages - 1]) {
-			memcpy_to_page(rq->out[outpages - 1], 0, src +
-					(interlaced_offset ? 0 : righthalf),
-				       lefthalf);
-		} else if (!interlaced_offset) {
-			memmove(src, src + righthalf, lefthalf);
-			flush_dcache_page(rq->in[inpages - 1]);
+const char *z_erofs_stream_switch_bufs(struct z_erofs_stream_dctx *dctx,
+				void **dst, void **src, struct page **pgpl)
+{
+	struct z_erofs_decompress_req *rq = dctx->rq;
+	struct page **pgo, *tmppage;
+	unsigned int j;
+
+	if (!dctx->avail_out) {
+		if (++dctx->no >= rq->outpages || !rq->outputsize)
+			return "insufficient space for decompressed data";
+
+		if (dctx->kout)
+			kunmap_local(dctx->kout);
+		dctx->avail_out = min(rq->outputsize, PAGE_SIZE - rq->pageofs_out);
+		rq->outputsize -= dctx->avail_out;
+		pgo = &rq->out[dctx->no];
+		if (!*pgo && rq->fillgaps) {		/* deduped */
+			*pgo = erofs_allocpage(pgpl, rq->gfp);
+			if (!*pgo) {
+				dctx->kout = NULL;
+				return ERR_PTR(-ENOMEM);
+			}
+			set_page_private(*pgo, Z_EROFS_SHORTLIVED_PAGE);
+		}
+		if (*pgo) {
+			dctx->kout = kmap_local_page(*pgo);
+			*dst = dctx->kout + rq->pageofs_out;
+		} else {
+			*dst = dctx->kout = NULL;
 		}
+		rq->pageofs_out = 0;
 	}
-	kunmap_local(src);
-	return 0;
+
+	if (dctx->inbuf_pos == dctx->inbuf_sz && rq->inputsize) {
+		if (++dctx->ni >= rq->inpages)
+			return "invalid compressed data";
+		if (dctx->kout) /* unlike kmap(), take care of the orders */
+			kunmap_local(dctx->kout);
+		kunmap_local(dctx->kin);
+
+		dctx->inbuf_sz = min_t(u32, rq->inputsize, PAGE_SIZE);
+		rq->inputsize -= dctx->inbuf_sz;
+		dctx->kin = kmap_local_page(rq->in[dctx->ni]);
+		*src = dctx->kin;
+		dctx->bounced = false;
+		if (dctx->kout) {
+			j = (u8 *)*dst - dctx->kout;
+			dctx->kout = kmap_local_page(rq->out[dctx->no]);
+			*dst = dctx->kout + j;
+		}
+		dctx->inbuf_pos = 0;
+	}
+
+	/*
+	 * Handle overlapping: Use the given bounce buffer if the input data is
+	 * under processing; Or utilize short-lived pages from the on-stack page
+	 * pool, where pages are shared among the same request.  Note that only
+	 * a few inplace I/O pages need to be doubled.
+	 */
+	if (!dctx->bounced && rq->out[dctx->no] == rq->in[dctx->ni]) {
+		memcpy(dctx->bounce, *src, dctx->inbuf_sz);
+		*src = dctx->bounce;
+		dctx->bounced = true;
+	}
+
+	for (j = dctx->ni + 1; j < rq->inpages; ++j) {
+		if (rq->out[dctx->no] != rq->in[j])
+			continue;
+		tmppage = erofs_allocpage(pgpl, rq->gfp);
+		if (!tmppage)
+			return ERR_PTR(-ENOMEM);
+		set_page_private(tmppage, Z_EROFS_SHORTLIVED_PAGE);
+		copy_highpage(tmppage, rq->in[j]);
+		rq->in[j] = tmppage;
+	}
+	return NULL;
 }
 
-const struct z_erofs_decompressor erofs_decompressors[] = {
-	[Z_EROFS_COMPRESSION_SHIFTED] = {
+const struct z_erofs_decompressor *z_erofs_decomp[] = {
+	[Z_EROFS_COMPRESSION_SHIFTED] = &(const struct z_erofs_decompressor) {
 		.decompress = z_erofs_transform_plain,
 		.name = "shifted"
 	},
-	[Z_EROFS_COMPRESSION_INTERLACED] = {
+	[Z_EROFS_COMPRESSION_INTERLACED] = &(const struct z_erofs_decompressor) {
 		.decompress = z_erofs_transform_plain,
 		.name = "interlaced"
 	},
-	[Z_EROFS_COMPRESSION_LZ4] = {
+	[Z_EROFS_COMPRESSION_LZ4] = &(const struct z_erofs_decompressor) {
+		.config = z_erofs_load_lz4_config,
 		.decompress = z_erofs_lz4_decompress,
+		.init = z_erofs_gbuf_init,
+		.exit = z_erofs_gbuf_exit,
 		.name = "lz4"
 	},
 #ifdef CONFIG_EROFS_FS_ZIP_LZMA
-	[Z_EROFS_COMPRESSION_LZMA] = {
-		.decompress = z_erofs_lzma_decompress,
-		.name = "lzma"
-	},
+	[Z_EROFS_COMPRESSION_LZMA] = &z_erofs_lzma_decomp,
 #endif
 #ifdef CONFIG_EROFS_FS_ZIP_DEFLATE
-	[Z_EROFS_COMPRESSION_DEFLATE] = {
-		.decompress = z_erofs_deflate_decompress,
-		.name = "deflate"
-	},
+	[Z_EROFS_COMPRESSION_DEFLATE] = &z_erofs_deflate_decomp,
+#endif
+#ifdef CONFIG_EROFS_FS_ZIP_ZSTD
+	[Z_EROFS_COMPRESSION_ZSTD] = &z_erofs_zstd_decomp,
 #endif
 };
+
+int z_erofs_parse_cfgs(struct super_block *sb, struct erofs_super_block *dsb)
+{
+	struct erofs_sb_info *sbi = EROFS_SB(sb);
+	struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
+	unsigned int algs, alg;
+	erofs_off_t offset;
+	int size, ret = 0;
+
+	if (!erofs_sb_has_compr_cfgs(sbi)) {
+		sbi->available_compr_algs = 1 << Z_EROFS_COMPRESSION_LZ4;
+		return z_erofs_load_lz4_config(sb, dsb, NULL, 0);
+	}
+
+	sbi->available_compr_algs = le16_to_cpu(dsb->u1.available_compr_algs);
+	if (sbi->available_compr_algs & ~Z_EROFS_ALL_COMPR_ALGS) {
+		erofs_err(sb, "unidentified algorithms %x, please upgrade kernel",
+			  sbi->available_compr_algs & ~Z_EROFS_ALL_COMPR_ALGS);
+		return -EOPNOTSUPP;
+	}
+
+	(void)erofs_init_metabuf(&buf, sb, false);
+	offset = EROFS_SUPER_OFFSET + sbi->sb_size;
+	alg = 0;
+	for (algs = sbi->available_compr_algs; algs; algs >>= 1, ++alg) {
+		const struct z_erofs_decompressor *dec = z_erofs_decomp[alg];
+		void *data;
+
+		if (!(algs & 1))
+			continue;
+
+		data = erofs_read_metadata(sb, &buf, &offset, &size);
+		if (IS_ERR(data)) {
+			ret = PTR_ERR(data);
+			break;
+		}
+
+		if (alg < Z_EROFS_COMPRESSION_MAX && dec && dec->config) {
+			ret = dec->config(sb, dsb, data, size);
+		} else {
+			erofs_err(sb, "algorithm %d isn't enabled on this kernel",
+				  alg);
+			ret = -EOPNOTSUPP;
+		}
+		kfree(data);
+		if (ret)
+			break;
+	}
+	erofs_put_metabuf(&buf);
+	return ret;
+}
+
+int __init z_erofs_init_decompressor(void)
+{
+	int i, err;
+
+	for (i = 0; i < Z_EROFS_COMPRESSION_MAX; ++i) {
+		err = z_erofs_decomp[i] ? z_erofs_decomp[i]->init() : 0;
+		if (err) {
+			while (i--)
+				if (z_erofs_decomp[i])
+					z_erofs_decomp[i]->exit();
+			return err;
+		}
+	}
+	return 0;
+}
+
+void z_erofs_exit_decompressor(void)
+{
+	int i;
+
+	for (i = 0; i < Z_EROFS_COMPRESSION_MAX; ++i)
+		if (z_erofs_decomp[i])
+			z_erofs_decomp[i]->exit();
+}