From d53238b614e01266a3d36b417b60a502e0698504 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Mon, 7 Jul 2025 16:48:32 +0800 Subject: [PATCH 1/6] erofs: fix to add missing tracepoint in erofs_readahead() Commit 771c994ea51f ("erofs: convert all uncompressed cases to iomap") converts to use iomap interface, it removed trace_erofs_readahead() tracepoint in the meantime, let's add it back. Fixes: 771c994ea51f ("erofs: convert all uncompressed cases to iomap") Signed-off-by: Chao Yu Reviewed-by: Gao Xiang Link: https://lore.kernel.org/r/20250707084832.2725677-1-chao@kernel.org Signed-off-by: Gao Xiang --- fs/erofs/data.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fs/erofs/data.c b/fs/erofs/data.c index 6a329c329f43..534ac359976e 100644 --- a/fs/erofs/data.c +++ b/fs/erofs/data.c @@ -356,6 +356,9 @@ static int erofs_read_folio(struct file *file, struct folio *folio) static void erofs_readahead(struct readahead_control *rac) { + trace_erofs_readahead(rac->mapping->host, readahead_index(rac), + readahead_count(rac), true); + return iomap_readahead(rac, &erofs_iomap_ops); } From 99f7619a77a0a2e3e2bcae676d0f301769167754 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 8 Jul 2025 19:19:42 +0800 Subject: [PATCH 2/6] erofs: fix to add missing tracepoint in erofs_read_folio() Commit 771c994ea51f ("erofs: convert all uncompressed cases to iomap") converts to use iomap interface, it removed trace_erofs_readpage() tracepoint in the meantime, let's add it back. Fixes: 771c994ea51f ("erofs: convert all uncompressed cases to iomap") Signed-off-by: Chao Yu Reviewed-by: Gao Xiang Link: https://lore.kernel.org/r/20250708111942.3120926-1-chao@kernel.org Signed-off-by: Gao Xiang --- fs/erofs/data.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/erofs/data.c b/fs/erofs/data.c index 534ac359976e..221e0ff1ed0d 100644 --- a/fs/erofs/data.c +++ b/fs/erofs/data.c @@ -351,6 +351,8 @@ int erofs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, */ static int erofs_read_folio(struct file *file, struct folio *folio) { + trace_erofs_read_folio(folio, true); + return iomap_read_folio(folio, &erofs_iomap_ops); } From f5443d0d1ad775927f1473b1c256ef68f2515b9c Mon Sep 17 00:00:00 2001 From: Gao Xiang Date: Wed, 9 Jul 2025 11:46:13 +0800 Subject: [PATCH 3/6] erofs: use memcpy_to_folio() to replace copy_to_iter() Using copy_to_iter() here is overkill and even messy. Signed-off-by: Gao Xiang Link: https://lore.kernel.org/r/20250709034614.2780117-1-hsiangkao@linux.alibaba.com --- fs/erofs/fileio.c | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/fs/erofs/fileio.c b/fs/erofs/fileio.c index df5cc63f2c01..fe2cd2982b4b 100644 --- a/fs/erofs/fileio.c +++ b/fs/erofs/fileio.c @@ -96,8 +96,6 @@ static int erofs_fileio_scan_folio(struct erofs_fileio *io, struct folio *folio) struct erofs_map_blocks *map = &io->map; unsigned int cur = 0, end = folio_size(folio), len, attached = 0; loff_t pos = folio_pos(folio), ofs; - struct iov_iter iter; - struct bio_vec bv; int err = 0; erofs_onlinefolio_init(folio); @@ -122,13 +120,7 @@ static int erofs_fileio_scan_folio(struct erofs_fileio *io, struct folio *folio) err = PTR_ERR(src); break; } - bvec_set_folio(&bv, folio, len, cur); - iov_iter_bvec(&iter, ITER_DEST, &bv, 1, len); - if (copy_to_iter(src, len, &iter) != len) { - erofs_put_metabuf(&buf); - err = -EIO; - break; - } + memcpy_to_folio(folio, cur, src, len); erofs_put_metabuf(&buf); } else if (!(map->m_flags & EROFS_MAP_MAPPED)) { folio_zero_segment(folio, cur, cur + len); From 27917e8194f91dffd8b4825350c63cb68e98ce58 Mon Sep 17 00:00:00 2001 From: Gao Xiang Date: Wed, 9 Jul 2025 11:46:14 +0800 Subject: [PATCH 4/6] erofs: address D-cache aliasing Flush the D-cache before unlocking folios for compressed inodes, as they are dirtied during decompression. Avoid calling flush_dcache_folio() on every CPU write, since it's more like playing whack-a-mole without real benefit. It has no impact on x86 and arm64/risc-v: on x86, flush_dcache_folio() is a no-op, and on arm64/risc-v, PG_dcache_clean (PG_arch_1) is clear for new page cache folios. However, certain ARM boards are affected, as reported. Fixes: 3883a79abd02 ("staging: erofs: introduce VLE decompression support") Closes: https://lore.kernel.org/r/c1e51e16-6cc6-49d0-a63e-4e9ff6c4dd53@pengutronix.de Closes: https://lore.kernel.org/r/38d43fae-1182-4155-9c5b-ffc7382d9917@siemens.com Tested-by: Jan Kiszka Tested-by: Stefan Kerkmann Signed-off-by: Gao Xiang Link: https://lore.kernel.org/r/20250709034614.2780117-2-hsiangkao@linux.alibaba.com --- fs/erofs/data.c | 16 +++++++++++----- fs/erofs/decompressor.c | 12 ++++-------- fs/erofs/fileio.c | 4 ++-- fs/erofs/internal.h | 2 +- fs/erofs/zdata.c | 6 +++--- 5 files changed, 21 insertions(+), 19 deletions(-) diff --git a/fs/erofs/data.c b/fs/erofs/data.c index 221e0ff1ed0d..16e4a6bd9b97 100644 --- a/fs/erofs/data.c +++ b/fs/erofs/data.c @@ -214,9 +214,11 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map) /* * bit 30: I/O error occurred on this folio + * bit 29: CPU has dirty data in D-cache (needs aliasing handling); * bit 0 - 29: remaining parts to complete this folio */ -#define EROFS_ONLINEFOLIO_EIO (1 << 30) +#define EROFS_ONLINEFOLIO_EIO 30 +#define EROFS_ONLINEFOLIO_DIRTY 29 void erofs_onlinefolio_init(struct folio *folio) { @@ -233,19 +235,23 @@ void erofs_onlinefolio_split(struct folio *folio) atomic_inc((atomic_t *)&folio->private); } -void erofs_onlinefolio_end(struct folio *folio, int err) +void erofs_onlinefolio_end(struct folio *folio, int err, bool dirty) { int orig, v; do { orig = atomic_read((atomic_t *)&folio->private); - v = (orig - 1) | (err ? EROFS_ONLINEFOLIO_EIO : 0); + DBG_BUGON(orig <= 0); + v = dirty << EROFS_ONLINEFOLIO_DIRTY; + v |= (orig - 1) | (!!err << EROFS_ONLINEFOLIO_EIO); } while (atomic_cmpxchg((atomic_t *)&folio->private, orig, v) != orig); - if (v & ~EROFS_ONLINEFOLIO_EIO) + if (v & (BIT(EROFS_ONLINEFOLIO_DIRTY) - 1)) return; folio->private = 0; - folio_end_read(folio, !(v & EROFS_ONLINEFOLIO_EIO)); + if (v & BIT(EROFS_ONLINEFOLIO_DIRTY)) + flush_dcache_folio(folio); + folio_end_read(folio, !(v & BIT(EROFS_ONLINEFOLIO_EIO))); } static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length, diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c index bf62e2836b60..358061d7b660 100644 --- a/fs/erofs/decompressor.c +++ b/fs/erofs/decompressor.c @@ -301,13 +301,11 @@ static int z_erofs_transform_plain(struct z_erofs_decompress_req *rq, 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]) { + if (rq->out[0] == rq->in[nrpages_in - 1]) memmove(kin + rq->pageofs_out, kin + pi, cur); - flush_dcache_page(rq->out[0]); - } else { + else memcpy_to_page(rq->out[0], rq->pageofs_out, kin + pi, cur); - } kunmap_local(kin); } rq->outputsize -= cur; @@ -325,14 +323,12 @@ static int z_erofs_transform_plain(struct z_erofs_decompress_req *rq, 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]) { + if (rq->out[no] == rq->in[ni]) memmove(kin + po, kin + rq->pageofs_in + pi, cnt); - flush_dcache_page(rq->out[no]); - } else if (rq->out[no]) { + 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); diff --git a/fs/erofs/fileio.c b/fs/erofs/fileio.c index fe2cd2982b4b..91781718199e 100644 --- a/fs/erofs/fileio.c +++ b/fs/erofs/fileio.c @@ -38,7 +38,7 @@ static void erofs_fileio_ki_complete(struct kiocb *iocb, long ret) } else { bio_for_each_folio_all(fi, &rq->bio) { DBG_BUGON(folio_test_uptodate(fi.folio)); - erofs_onlinefolio_end(fi.folio, ret); + erofs_onlinefolio_end(fi.folio, ret, false); } } bio_uninit(&rq->bio); @@ -154,7 +154,7 @@ static int erofs_fileio_scan_folio(struct erofs_fileio *io, struct folio *folio) } cur += len; } - erofs_onlinefolio_end(folio, err); + erofs_onlinefolio_end(folio, err, false); return err; } diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h index a32c03a80c70..0d19bde8c094 100644 --- a/fs/erofs/internal.h +++ b/fs/erofs/internal.h @@ -390,7 +390,7 @@ int erofs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, int erofs_map_blocks(struct inode *inode, struct erofs_map_blocks *map); void erofs_onlinefolio_init(struct folio *folio); void erofs_onlinefolio_split(struct folio *folio); -void erofs_onlinefolio_end(struct folio *folio, int err); +void erofs_onlinefolio_end(struct folio *folio, int err, bool dirty); struct inode *erofs_iget(struct super_block *sb, erofs_nid_t nid); int erofs_getattr(struct mnt_idmap *idmap, const struct path *path, struct kstat *stat, u32 request_mask, diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c index fe8071844724..32e3905e75fe 100644 --- a/fs/erofs/zdata.c +++ b/fs/erofs/zdata.c @@ -1091,7 +1091,7 @@ static int z_erofs_scan_folio(struct z_erofs_frontend *f, tight = (bs == PAGE_SIZE); } } while ((end = cur) > 0); - erofs_onlinefolio_end(folio, err); + erofs_onlinefolio_end(folio, err, false); return err; } @@ -1196,7 +1196,7 @@ static void z_erofs_fill_other_copies(struct z_erofs_backend *be, int err) cur += len; } kunmap_local(dst); - erofs_onlinefolio_end(page_folio(bvi->bvec.page), err); + erofs_onlinefolio_end(page_folio(bvi->bvec.page), err, true); list_del(p); kfree(bvi); } @@ -1355,7 +1355,7 @@ static int z_erofs_decompress_pcluster(struct z_erofs_backend *be, int err) DBG_BUGON(z_erofs_page_is_invalidated(page)); if (!z_erofs_is_shortlived_page(page)) { - erofs_onlinefolio_end(page_folio(page), err); + erofs_onlinefolio_end(page_folio(page), err, true); continue; } if (pcl->algorithmformat != Z_EROFS_COMPRESSION_LZ4) { From d31fbdc4c7252846ea80235db8c1a8c932da9d39 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Thu, 10 Jul 2025 15:36:18 +0800 Subject: [PATCH 5/6] erofs: allow readdir() to be interrupted In a quick slow device, readdir() may loop for long time in large directory, let's give a chance to allow it to be interrupted by userspace. Signed-off-by: Chao Yu Reviewed-by: Gao Xiang Link: https://lore.kernel.org/r/20250710073619.4083422-1-chao@kernel.org [ Gao Xiang: move cond_resched() to the end of the while loop. ] Signed-off-by: Gao Xiang --- fs/erofs/dir.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/erofs/dir.c b/fs/erofs/dir.c index 2fae209d0274..3e4b38bec0aa 100644 --- a/fs/erofs/dir.c +++ b/fs/erofs/dir.c @@ -58,6 +58,11 @@ static int erofs_readdir(struct file *f, struct dir_context *ctx) struct erofs_dirent *de; unsigned int nameoff, maxsize; + if (fatal_signal_pending(current)) { + err = -ERESTARTSYS; + break; + } + de = erofs_bread(&buf, dbstart, true); if (IS_ERR(de)) { erofs_err(sb, "failed to readdir of logical block %llu of nid %llu", @@ -88,6 +93,7 @@ static int erofs_readdir(struct file *f, struct dir_context *ctx) break; ctx->pos = dbstart + maxsize; ofs = 0; + cond_resched(); } erofs_put_metabuf(&buf); if (EROFS_I(dir)->dot_omitted && ctx->pos == dir->i_size) { From b44686c8391b427fb1c85a31c35077e6947c6d90 Mon Sep 17 00:00:00 2001 From: Gao Xiang Date: Sat, 12 Jul 2025 03:58:26 +0800 Subject: [PATCH 6/6] erofs: fix large fragment handling Fragments aren't limited by Z_EROFS_PCLUSTER_MAX_DSIZE. However, if a fragment's logical length is larger than Z_EROFS_PCLUSTER_MAX_DSIZE but the fragment is not the whole inode, it currently returns -EOPNOTSUPP because m_flags has the wrong EROFS_MAP_ENCODED flag set. It is not intended by design but should be rare, as it can only be reproduced by mkfs with `-Eall-fragments` in a specific case. Let's normalize fragment m_flags using the new EROFS_MAP_FRAGMENT. Reported-by: Axel Fontaine Closes: https://github.com/erofs/erofs-utils/issues/23 Fixes: 7c3ca1838a78 ("erofs: restrict pcluster size limitations") Signed-off-by: Gao Xiang Link: https://lore.kernel.org/r/20250711195826.3601157-1-hsiangkao@linux.alibaba.com --- fs/erofs/internal.h | 4 +++- fs/erofs/zdata.c | 2 +- fs/erofs/zmap.c | 9 ++++----- 3 files changed, 8 insertions(+), 7 deletions(-) diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h index 0d19bde8c094..06b867d2fc3b 100644 --- a/fs/erofs/internal.h +++ b/fs/erofs/internal.h @@ -315,10 +315,12 @@ static inline struct folio *erofs_grab_folio_nowait(struct address_space *as, /* The length of extent is full */ #define EROFS_MAP_FULL_MAPPED 0x0008 /* Located in the special packed inode */ -#define EROFS_MAP_FRAGMENT 0x0010 +#define __EROFS_MAP_FRAGMENT 0x0010 /* The extent refers to partial decompressed data */ #define EROFS_MAP_PARTIAL_REF 0x0020 +#define EROFS_MAP_FRAGMENT (EROFS_MAP_MAPPED | __EROFS_MAP_FRAGMENT) + struct erofs_map_blocks { struct erofs_buf buf; diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c index 32e3905e75fe..e3f28a1bb945 100644 --- a/fs/erofs/zdata.c +++ b/fs/erofs/zdata.c @@ -1034,7 +1034,7 @@ static int z_erofs_scan_folio(struct z_erofs_frontend *f, if (!(map->m_flags & EROFS_MAP_MAPPED)) { folio_zero_segment(folio, cur, end); tight = false; - } else if (map->m_flags & EROFS_MAP_FRAGMENT) { + } else if (map->m_flags & __EROFS_MAP_FRAGMENT) { erofs_off_t fpos = offset + cur - map->m_la; err = z_erofs_read_fragment(inode->i_sb, folio, cur, diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c index 0bebc6e3a4d7..f1a15ff22147 100644 --- a/fs/erofs/zmap.c +++ b/fs/erofs/zmap.c @@ -413,8 +413,7 @@ static int z_erofs_map_blocks_fo(struct inode *inode, !vi->z_tailextent_headlcn) { map->m_la = 0; map->m_llen = inode->i_size; - map->m_flags = EROFS_MAP_MAPPED | - EROFS_MAP_FULL_MAPPED | EROFS_MAP_FRAGMENT; + map->m_flags = EROFS_MAP_FRAGMENT; return 0; } initial_lcn = ofs >> lclusterbits; @@ -489,7 +488,7 @@ static int z_erofs_map_blocks_fo(struct inode *inode, goto unmap_out; } } else if (fragment && m.lcn == vi->z_tailextent_headlcn) { - map->m_flags |= EROFS_MAP_FRAGMENT; + map->m_flags = EROFS_MAP_FRAGMENT; } else { map->m_pa = erofs_pos(sb, m.pblk); err = z_erofs_get_extent_compressedlen(&m, initial_lcn); @@ -617,7 +616,7 @@ static int z_erofs_map_blocks_ext(struct inode *inode, if (lstart < lend) { map->m_la = lstart; if (last && (vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER)) { - map->m_flags |= EROFS_MAP_MAPPED | EROFS_MAP_FRAGMENT; + map->m_flags = EROFS_MAP_FRAGMENT; vi->z_fragmentoff = map->m_plen; if (recsz > offsetof(struct z_erofs_extent, pstart_lo)) vi->z_fragmentoff |= map->m_pa << 32; @@ -797,7 +796,7 @@ static int z_erofs_iomap_begin_report(struct inode *inode, loff_t offset, iomap->length = map.m_llen; if (map.m_flags & EROFS_MAP_MAPPED) { iomap->type = IOMAP_MAPPED; - iomap->addr = map.m_flags & EROFS_MAP_FRAGMENT ? + iomap->addr = map.m_flags & __EROFS_MAP_FRAGMENT ? IOMAP_NULL_ADDR : map.m_pa; } else { iomap->type = IOMAP_HOLE;