Merge tag 'for-5.15-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba:
 "The highlights of this round are integrations with fs-verity and
  idmapped mounts, the rest is usual mix of minor improvements, speedups
  and cleanups.

  There are some patches outside of btrfs, namely updating some VFS
  interfaces, all straightforward and acked.

  Features:

   - fs-verity support, using standard ioctls, backward compatible with
     read-only limitation on inodes with previously enabled fs-verity

   - idmapped mount support

   - make mount with rescue=ibadroots more tolerant to partially damaged
     trees

   - allow raid0 on a single device and raid10 on two devices,
     degenerate cases but might be useful as an intermediate step during
     conversion to other profiles

   - zoned mode block group auto reclaim can be disabled via sysfs knob

  Performance improvements:

   - continue readahead of node siblings even if target node is in
     memory, could speed up full send (on sample test +11%)

   - batching of delayed items can speed up creating many files

   - fsync/tree-log speedups
       - avoid unnecessary work (gains +2% throughput, -2% run time on
         sample load)
       - reduced lock contention on renames (on dbench +4% throughput,
         up to -30% latency)

  Fixes:

   - various zoned mode fixes

   - preemptive flushing threshold tuning, avoid excessive work on
     almost full filesystems

  Core:

   - continued subpage support, preparation for implementing remaining
     features like compression and defragmentation; with some
     limitations, write is now enabled on 64K page systems with 4K
     sectors, still considered experimental
       - no readahead on compressed reads
       - inline extents disabled
       - disabled raid56 profile conversion and mount

   - improved flushing logic, fixing early ENOSPC on some workloads

   - inode flags have been internally split to read-only and read-write
     incompat bit parts, used by fs-verity

   - new tree items for fs-verity
       - descriptor item
       - Merkle tree item

   - inode operations extended to be namespace-aware

   - cleanups and refactoring

  Generic code changes:

   - fs: new export filemap_fdatawrite_wbc

   - fs: removed sync_inode

   - block: bio_trim argument type fixups

   - vfs: add namespace-aware lookup"

* tag 'for-5.15-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (114 commits)
  btrfs: reset replace target device to allocation state on close
  btrfs: zoned: fix ordered extent boundary calculation
  btrfs: do not do preemptive flushing if the majority is global rsv
  btrfs: reduce the preemptive flushing threshold to 90%
  btrfs: tree-log: check btrfs_lookup_data_extent return value
  btrfs: avoid unnecessarily logging directories that had no changes
  btrfs: allow idmapped mount
  btrfs: handle ACLs on idmapped mounts
  btrfs: allow idmapped INO_LOOKUP_USER ioctl
  btrfs: allow idmapped SUBVOL_SETFLAGS ioctl
  btrfs: allow idmapped SET_RECEIVED_SUBVOL ioctls
  btrfs: relax restrictions for SNAP_DESTROY_V2 with subvolids
  btrfs: allow idmapped SNAP_DESTROY ioctls
  btrfs: allow idmapped SNAP_CREATE/SUBVOL_CREATE ioctls
  btrfs: check whether fsgid/fsuid are mapped during subvolume creation
  btrfs: allow idmapped permission inode op
  btrfs: allow idmapped setattr inode op
  btrfs: allow idmapped tmpfile inode op
  btrfs: allow idmapped symlink inode op
  btrfs: allow idmapped mkdir inode op
  ...

1 files changed, 100 insertions, 136 deletions

diff --git a/fs/btrfs/lzo.c b/fs/btrfs/lzo.c
index cd042c7567a4..c25dfd1a8a54 100644
--- a/fs/btrfs/lzo.c
+++ b/fs/btrfs/lzo.c
@@ -14,6 +14,7 @@
 #include <linux/lzo.h>
 #include <linux/refcount.h>
 #include "compression.h"
+#include "ctree.h"
 
 #define LZO_LEN	4
 
@@ -140,18 +141,18 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 	*total_in = 0;
 
 	in_page = find_get_page(mapping, start >> PAGE_SHIFT);
-	data_in = kmap(in_page);
+	data_in = page_address(in_page);
 
 	/*
 	 * store the size of all chunks of compressed data in
 	 * the first 4 bytes
 	 */
-	out_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
+	out_page = alloc_page(GFP_NOFS);
 	if (out_page == NULL) {
 		ret = -ENOMEM;
 		goto out;
 	}
-	cpage_out = kmap(out_page);
+	cpage_out = page_address(out_page);
 	out_offset = LZO_LEN;
 	tot_out = LZO_LEN;
 	pages[0] = out_page;
@@ -209,19 +210,18 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 				if (out_len == 0 && tot_in >= len)
 					break;
 
-				kunmap(out_page);
 				if (nr_pages == nr_dest_pages) {
 					out_page = NULL;
 					ret = -E2BIG;
 					goto out;
 				}
 
-				out_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
+				out_page = alloc_page(GFP_NOFS);
 				if (out_page == NULL) {
 					ret = -ENOMEM;
 					goto out;
 				}
-				cpage_out = kmap(out_page);
+				cpage_out = page_address(out_page);
 				pages[nr_pages++] = out_page;
 
 				pg_bytes_left = PAGE_SIZE;
@@ -243,12 +243,11 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 			break;
 
 		bytes_left = len - tot_in;
-		kunmap(in_page);
 		put_page(in_page);
 
 		start += PAGE_SIZE;
 		in_page = find_get_page(mapping, start >> PAGE_SHIFT);
-		data_in = kmap(in_page);
+		data_in = page_address(in_page);
 		in_len = min(bytes_left, PAGE_SIZE);
 	}
 
@@ -258,164 +257,130 @@ int lzo_compress_pages(struct list_head *ws, struct address_space *mapping,
 	}
 
 	/* store the size of all chunks of compressed data */
-	sizes_ptr = kmap_local_page(pages[0]);
+	sizes_ptr = page_address(pages[0]);
 	write_compress_length(sizes_ptr, tot_out);
-	kunmap_local(sizes_ptr);
 
 	ret = 0;
 	*total_out = tot_out;
 	*total_in = tot_in;
 out:
 	*out_pages = nr_pages;
-	if (out_page)
-		kunmap(out_page);
 
-	if (in_page) {
-		kunmap(in_page);
+	if (in_page)
 		put_page(in_page);
-	}
 
 	return ret;
 }
 
+/*
+ * Copy the compressed segment payload into @dest.
+ *
+ * For the payload there will be no padding, just need to do page switching.
+ */
+static void copy_compressed_segment(struct compressed_bio *cb,
+				    char *dest, u32 len, u32 *cur_in)
+{
+	u32 orig_in = *cur_in;
+
+	while (*cur_in < orig_in + len) {
+		struct page *cur_page;
+		u32 copy_len = min_t(u32, PAGE_SIZE - offset_in_page(*cur_in),
+					  orig_in + len - *cur_in);
+
+		ASSERT(copy_len);
+		cur_page = cb->compressed_pages[*cur_in / PAGE_SIZE];
+
+		memcpy(dest + *cur_in - orig_in,
+			page_address(cur_page) + offset_in_page(*cur_in),
+			copy_len);
+
+		*cur_in += copy_len;
+	}
+}
+
 int lzo_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
 {
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
-	int ret = 0, ret2;
-	char *data_in;
-	unsigned long page_in_index = 0;
-	size_t srclen = cb->compressed_len;
-	unsigned long total_pages_in = DIV_ROUND_UP(srclen, PAGE_SIZE);
-	unsigned long buf_start;
-	unsigned long buf_offset = 0;
-	unsigned long bytes;
-	unsigned long working_bytes;
-	size_t in_len;
-	size_t out_len;
-	const size_t max_segment_len = lzo1x_worst_compress(PAGE_SIZE);
-	unsigned long in_offset;
-	unsigned long in_page_bytes_left;
-	unsigned long tot_in;
-	unsigned long tot_out;
-	unsigned long tot_len;
-	char *buf;
-	bool may_late_unmap, need_unmap;
-	struct page **pages_in = cb->compressed_pages;
-	u64 disk_start = cb->start;
-	struct bio *orig_bio = cb->orig_bio;
+	const struct btrfs_fs_info *fs_info = btrfs_sb(cb->inode->i_sb);
+	const u32 sectorsize = fs_info->sectorsize;
+	int ret;
+	/* Compressed data length, can be unaligned */
+	u32 len_in;
+	/* Offset inside the compressed data */
+	u32 cur_in = 0;
+	/* Bytes decompressed so far */
+	u32 cur_out = 0;
+
+	len_in = read_compress_length(page_address(cb->compressed_pages[0]));
+	cur_in += LZO_LEN;
 
-	data_in = kmap(pages_in[0]);
-	tot_len = read_compress_length(data_in);
 	/*
-	 * Compressed data header check.
+	 * LZO header length check
 	 *
-	 * The real compressed size can't exceed the maximum extent length, and
-	 * all pages should be used (whole unused page with just the segment
-	 * header is not possible).  If this happens it means the compressed
-	 * extent is corrupted.
+	 * The total length should not exceed the maximum extent length,
+	 * and all sectors should be used.
+	 * If this happens, it means the compressed extent is corrupted.
 	 */
-	if (tot_len > min_t(size_t, BTRFS_MAX_COMPRESSED, srclen) ||
-	    tot_len < srclen - PAGE_SIZE) {
-		ret = -EUCLEAN;
-		goto done;
+	if (len_in > min_t(size_t, BTRFS_MAX_COMPRESSED, cb->compressed_len) ||
+	    round_up(len_in, sectorsize) < cb->compressed_len) {
+		btrfs_err(fs_info,
+			"invalid lzo header, lzo len %u compressed len %u",
+			len_in, cb->compressed_len);
+		return -EUCLEAN;
 	}
 
-	tot_in = LZO_LEN;
-	in_offset = LZO_LEN;
-	in_page_bytes_left = PAGE_SIZE - LZO_LEN;
-
-	tot_out = 0;
-
-	while (tot_in < tot_len) {
-		in_len = read_compress_length(data_in + in_offset);
-		in_page_bytes_left -= LZO_LEN;
-		in_offset += LZO_LEN;
-		tot_in += LZO_LEN;
+	/* Go through each lzo segment */
+	while (cur_in < len_in) {
+		struct page *cur_page;
+		/* Length of the compressed segment */
+		u32 seg_len;
+		u32 sector_bytes_left;
+		size_t out_len = lzo1x_worst_compress(sectorsize);
 
 		/*
-		 * Segment header check.
-		 *
-		 * The segment length must not exceed the maximum LZO
-		 * compression size, nor the total compressed size.
+		 * We should always have enough space for one segment header
+		 * inside current sector.
 		 */
-		if (in_len > max_segment_len || tot_in + in_len > tot_len) {
-			ret = -EUCLEAN;
-			goto done;
-		}
-
-		tot_in += in_len;
-		working_bytes = in_len;
-		may_late_unmap = need_unmap = false;
-
-		/* fast path: avoid using the working buffer */
-		if (in_page_bytes_left >= in_len) {
-			buf = data_in + in_offset;
-			bytes = in_len;
-			may_late_unmap = true;
-			goto cont;
-		}
-
-		/* copy bytes from the pages into the working buffer */
-		buf = workspace->cbuf;
-		buf_offset = 0;
-		while (working_bytes) {
-			bytes = min(working_bytes, in_page_bytes_left);
-
-			memcpy(buf + buf_offset, data_in + in_offset, bytes);
-			buf_offset += bytes;
-cont:
-			working_bytes -= bytes;
-			in_page_bytes_left -= bytes;
-			in_offset += bytes;
-
-			/* check if we need to pick another page */
-			if ((working_bytes == 0 && in_page_bytes_left < LZO_LEN)
-			    || in_page_bytes_left == 0) {
-				tot_in += in_page_bytes_left;
-
-				if (working_bytes == 0 && tot_in >= tot_len)
-					break;
-
-				if (page_in_index + 1 >= total_pages_in) {
-					ret = -EIO;
-					goto done;
-				}
-
-				if (may_late_unmap)
-					need_unmap = true;
-				else
-					kunmap(pages_in[page_in_index]);
-
-				data_in = kmap(pages_in[++page_in_index]);
-
-				in_page_bytes_left = PAGE_SIZE;
-				in_offset = 0;
-			}
-		}
-
-		out_len = max_segment_len;
-		ret = lzo1x_decompress_safe(buf, in_len, workspace->buf,
-					    &out_len);
-		if (need_unmap)
-			kunmap(pages_in[page_in_index - 1]);
+		ASSERT(cur_in / sectorsize ==
+		       (cur_in + LZO_LEN - 1) / sectorsize);
+		cur_page = cb->compressed_pages[cur_in / PAGE_SIZE];
+		ASSERT(cur_page);
+		seg_len = read_compress_length(page_address(cur_page) +
+					       offset_in_page(cur_in));
+		cur_in += LZO_LEN;
+
+		/* Copy the compressed segment payload into workspace */
+		copy_compressed_segment(cb, workspace->cbuf, seg_len, &cur_in);
+
+		/* Decompress the data */
+		ret = lzo1x_decompress_safe(workspace->cbuf, seg_len,
+					    workspace->buf, &out_len);
 		if (ret != LZO_E_OK) {
-			pr_warn("BTRFS: decompress failed\n");
+			btrfs_err(fs_info, "failed to decompress");
 			ret = -EIO;
-			break;
+			goto out;
 		}
 
-		buf_start = tot_out;
-		tot_out += out_len;
+		/* Copy the data into inode pages */
+		ret = btrfs_decompress_buf2page(workspace->buf, out_len, cb, cur_out);
+		cur_out += out_len;
 
-		ret2 = btrfs_decompress_buf2page(workspace->buf, buf_start,
-						 tot_out, disk_start, orig_bio);
-		if (ret2 == 0)
-			break;
+		/* All data read, exit */
+		if (ret == 0)
+			goto out;
+		ret = 0;
+
+		/* Check if the sector has enough space for a segment header */
+		sector_bytes_left = sectorsize - (cur_in % sectorsize);
+		if (sector_bytes_left >= LZO_LEN)
+			continue;
+
+		/* Skip the padding zeros */
+		cur_in += sector_bytes_left;
 	}
-done:
-	kunmap(pages_in[page_in_index]);
+out:
 	if (!ret)
-		zero_fill_bio(orig_bio);
+		zero_fill_bio(cb->orig_bio);
 	return ret;
 }
 
@@ -466,7 +431,7 @@ int lzo_decompress(struct list_head *ws, unsigned char *data_in,
 	destlen = min_t(unsigned long, destlen, PAGE_SIZE);
 	bytes = min_t(unsigned long, destlen, out_len - start_byte);
 
-	kaddr = kmap_local_page(dest_page);
+	kaddr = page_address(dest_page);
 	memcpy(kaddr, workspace->buf + start_byte, bytes);
 
 	/*
@@ -476,7 +441,6 @@ int lzo_decompress(struct list_head *ws, unsigned char *data_in,
 	 */
 	if (bytes < destlen)
 		memset(kaddr+bytes, 0, destlen-bytes);
-	kunmap_local(kaddr);
 out:
 	return ret;
 }