summaryrefslogtreecommitdiffstats
path: root/fs/erofs/zdata.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/erofs/zdata.c')
-rw-r--r--fs/erofs/zdata.c287
1 files changed, 137 insertions, 150 deletions
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index ff0aa72b0d..3216b920d3 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -19,7 +19,10 @@
typedef void *z_erofs_next_pcluster_t;
struct z_erofs_bvec {
- struct page *page;
+ union {
+ struct page *page;
+ struct folio *folio;
+ };
int offset;
unsigned int end;
};
@@ -116,47 +119,46 @@ static inline unsigned int z_erofs_pclusterpages(struct z_erofs_pcluster *pcl)
return PAGE_ALIGN(pcl->pclustersize) >> PAGE_SHIFT;
}
+#define MNGD_MAPPING(sbi) ((sbi)->managed_cache->i_mapping)
+static bool erofs_folio_is_managed(struct erofs_sb_info *sbi, struct folio *fo)
+{
+ return fo->mapping == MNGD_MAPPING(sbi);
+}
+
/*
- * bit 30: I/O error occurred on this page
- * bit 0 - 29: remaining parts to complete this page
+ * bit 30: I/O error occurred on this folio
+ * bit 0 - 29: remaining parts to complete this folio
*/
-#define Z_EROFS_PAGE_EIO (1 << 30)
+#define Z_EROFS_FOLIO_EIO (1 << 30)
-static inline void z_erofs_onlinepage_init(struct page *page)
+static void z_erofs_onlinefolio_init(struct folio *folio)
{
union {
atomic_t o;
- unsigned long v;
+ void *v;
} u = { .o = ATOMIC_INIT(1) };
- set_page_private(page, u.v);
- smp_wmb();
- SetPagePrivate(page);
+ folio->private = u.v; /* valid only if file-backed folio is locked */
}
-static inline void z_erofs_onlinepage_split(struct page *page)
+static void z_erofs_onlinefolio_split(struct folio *folio)
{
- atomic_inc((atomic_t *)&page->private);
+ atomic_inc((atomic_t *)&folio->private);
}
-static void z_erofs_onlinepage_endio(struct page *page, int err)
+static void z_erofs_onlinefolio_end(struct folio *folio, int err)
{
int orig, v;
- DBG_BUGON(!PagePrivate(page));
-
do {
- orig = atomic_read((atomic_t *)&page->private);
- v = (orig - 1) | (err ? Z_EROFS_PAGE_EIO : 0);
- } while (atomic_cmpxchg((atomic_t *)&page->private, orig, v) != orig);
-
- if (!(v & ~Z_EROFS_PAGE_EIO)) {
- set_page_private(page, 0);
- ClearPagePrivate(page);
- if (!(v & Z_EROFS_PAGE_EIO))
- SetPageUptodate(page);
- unlock_page(page);
- }
+ orig = atomic_read((atomic_t *)&folio->private);
+ v = (orig - 1) | (err ? Z_EROFS_FOLIO_EIO : 0);
+ } while (atomic_cmpxchg((atomic_t *)&folio->private, orig, v) != orig);
+
+ if (v & ~Z_EROFS_FOLIO_EIO)
+ return;
+ folio->private = 0;
+ folio_end_read(folio, !(v & Z_EROFS_FOLIO_EIO));
}
#define Z_EROFS_ONSTACK_PAGES 32
@@ -572,17 +574,13 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
for (i = 0; i < pclusterpages; ++i) {
struct page *page, *newpage;
- void *t; /* mark pages just found for debugging */
/* Inaccurate check w/o locking to avoid unneeded lookups */
if (READ_ONCE(pcl->compressed_bvecs[i].page))
continue;
page = find_get_page(mc, pcl->obj.index + i);
- if (page) {
- t = (void *)((unsigned long)page | 1);
- newpage = NULL;
- } else {
+ if (!page) {
/* I/O is needed, no possible to decompress directly */
standalone = false;
if (!shouldalloc)
@@ -596,11 +594,10 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
if (!newpage)
continue;
set_page_private(newpage, Z_EROFS_PREALLOCATED_PAGE);
- t = (void *)((unsigned long)newpage | 1);
}
spin_lock(&pcl->obj.lockref.lock);
if (!pcl->compressed_bvecs[i].page) {
- pcl->compressed_bvecs[i].page = t;
+ pcl->compressed_bvecs[i].page = page ? page : newpage;
spin_unlock(&pcl->obj.lockref.lock);
continue;
}
@@ -620,9 +617,9 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
fe->mode = Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE;
}
-/* called by erofs_shrinker to get rid of all compressed_pages */
-int erofs_try_to_free_all_cached_pages(struct erofs_sb_info *sbi,
- struct erofs_workgroup *grp)
+/* called by erofs_shrinker to get rid of all cached compressed bvecs */
+int erofs_try_to_free_all_cached_folios(struct erofs_sb_info *sbi,
+ struct erofs_workgroup *grp)
{
struct z_erofs_pcluster *const pcl =
container_of(grp, struct z_erofs_pcluster, obj);
@@ -630,27 +627,22 @@ int erofs_try_to_free_all_cached_pages(struct erofs_sb_info *sbi,
int i;
DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
- /*
- * refcount of workgroup is now freezed as 0,
- * therefore no need to worry about available decompression users.
- */
+ /* There is no actice user since the pcluster is now freezed */
for (i = 0; i < pclusterpages; ++i) {
- struct page *page = pcl->compressed_bvecs[i].page;
+ struct folio *folio = pcl->compressed_bvecs[i].folio;
- if (!page)
+ if (!folio)
continue;
- /* block other users from reclaiming or migrating the page */
- if (!trylock_page(page))
+ /* Avoid reclaiming or migrating this folio */
+ if (!folio_trylock(folio))
return -EBUSY;
- if (!erofs_page_is_managed(sbi, page))
+ if (!erofs_folio_is_managed(sbi, folio))
continue;
-
- /* barrier is implied in the following 'unlock_page' */
- WRITE_ONCE(pcl->compressed_bvecs[i].page, NULL);
- detach_page_private(page);
- unlock_page(page);
+ pcl->compressed_bvecs[i].folio = NULL;
+ folio_detach_private(folio);
+ folio_unlock(folio);
}
return 0;
}
@@ -667,20 +659,17 @@ static bool z_erofs_cache_release_folio(struct folio *folio, gfp_t gfp)
ret = false;
spin_lock(&pcl->obj.lockref.lock);
- if (pcl->obj.lockref.count > 0)
- goto out;
-
- DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
- for (i = 0; i < pclusterpages; ++i) {
- if (pcl->compressed_bvecs[i].page == &folio->page) {
- WRITE_ONCE(pcl->compressed_bvecs[i].page, NULL);
- ret = true;
- break;
+ if (pcl->obj.lockref.count <= 0) {
+ DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
+ for (i = 0; i < pclusterpages; ++i) {
+ if (pcl->compressed_bvecs[i].folio == folio) {
+ pcl->compressed_bvecs[i].folio = NULL;
+ folio_detach_private(folio);
+ ret = true;
+ break;
+ }
}
}
- if (ret)
- folio_detach_private(folio);
-out:
spin_unlock(&pcl->obj.lockref.lock);
return ret;
}
@@ -962,20 +951,20 @@ static int z_erofs_read_fragment(struct super_block *sb, struct page *page,
return 0;
}
-static int z_erofs_do_read_page(struct z_erofs_decompress_frontend *fe,
- struct page *page, bool ra)
+static int z_erofs_scan_folio(struct z_erofs_decompress_frontend *fe,
+ struct folio *folio, bool ra)
{
struct inode *const inode = fe->inode;
struct erofs_map_blocks *const map = &fe->map;
- const loff_t offset = page_offset(page);
- const unsigned int bs = i_blocksize(inode);
+ const loff_t offset = folio_pos(folio);
+ const unsigned int bs = i_blocksize(inode), fs = folio_size(folio);
bool tight = true, exclusive;
unsigned int cur, end, len, split;
int err = 0;
- z_erofs_onlinepage_init(page);
+ z_erofs_onlinefolio_init(folio);
split = 0;
- end = PAGE_SIZE;
+ end = fs;
repeat:
if (offset + end - 1 < map->m_la ||
offset + end - 1 >= map->m_la + map->m_llen) {
@@ -992,7 +981,7 @@ repeat:
++split;
if (!(map->m_flags & EROFS_MAP_MAPPED)) {
- zero_user_segment(page, cur, end);
+ folio_zero_segment(folio, cur, end);
tight = false;
goto next_part;
}
@@ -1001,8 +990,8 @@ repeat:
erofs_off_t fpos = offset + cur - map->m_la;
len = min_t(unsigned int, map->m_llen - fpos, end - cur);
- err = z_erofs_read_fragment(inode->i_sb, page, cur, cur + len,
- EROFS_I(inode)->z_fragmentoff + fpos);
+ err = z_erofs_read_fragment(inode->i_sb, &folio->page, cur,
+ cur + len, EROFS_I(inode)->z_fragmentoff + fpos);
if (err)
goto out;
tight = false;
@@ -1017,25 +1006,25 @@ repeat:
}
/*
- * Ensure the current partial page belongs to this submit chain rather
+ * Ensure the current partial folio belongs to this submit chain rather
* than other concurrent submit chains or the noio(bypass) chain since
- * those chains are handled asynchronously thus the page cannot be used
+ * those chains are handled asynchronously thus the folio cannot be used
* for inplace I/O or bvpage (should be processed in a strict order.)
*/
tight &= (fe->mode > Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE);
- exclusive = (!cur && ((split <= 1) || (tight && bs == PAGE_SIZE)));
+ exclusive = (!cur && ((split <= 1) || (tight && bs == fs)));
if (cur)
tight &= (fe->mode >= Z_EROFS_PCLUSTER_FOLLOWED);
err = z_erofs_attach_page(fe, &((struct z_erofs_bvec) {
- .page = page,
+ .page = &folio->page,
.offset = offset - map->m_la,
.end = end,
}), exclusive);
if (err)
goto out;
- z_erofs_onlinepage_split(page);
+ z_erofs_onlinefolio_split(folio);
if (fe->pcl->pageofs_out != (map->m_la & ~PAGE_MASK))
fe->pcl->multibases = true;
if (fe->pcl->length < offset + end - map->m_la) {
@@ -1056,7 +1045,7 @@ next_part:
goto repeat;
out:
- z_erofs_onlinepage_endio(page, err);
+ z_erofs_onlinefolio_end(folio, err);
return err;
}
@@ -1159,7 +1148,7 @@ static void z_erofs_fill_other_copies(struct z_erofs_decompress_backend *be,
cur += len;
}
kunmap_local(dst);
- z_erofs_onlinepage_endio(bvi->bvec.page, err);
+ z_erofs_onlinefolio_end(page_folio(bvi->bvec.page), err);
list_del(p);
kfree(bvi);
}
@@ -1210,7 +1199,7 @@ static int z_erofs_parse_in_bvecs(struct z_erofs_decompress_backend *be,
be->compressed_pages[i] = page;
if (z_erofs_is_inline_pcluster(pcl) ||
- erofs_page_is_managed(EROFS_SB(be->sb), page)) {
+ erofs_folio_is_managed(EROFS_SB(be->sb), page_folio(page))) {
if (!PageUptodate(page))
err = -EIO;
continue;
@@ -1295,7 +1284,8 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
/* consider shortlived pages added when decompressing */
page = be->compressed_pages[i];
- if (!page || erofs_page_is_managed(sbi, page))
+ if (!page ||
+ erofs_folio_is_managed(sbi, page_folio(page)))
continue;
(void)z_erofs_put_shortlivedpage(be->pagepool, page);
WRITE_ONCE(pcl->compressed_bvecs[i].page, NULL);
@@ -1316,7 +1306,7 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
/* recycle all individual short-lived pages */
if (z_erofs_put_shortlivedpage(be->pagepool, page))
continue;
- z_erofs_onlinepage_endio(page, err);
+ z_erofs_onlinefolio_end(page_folio(page), err);
}
if (be->decompressed_pages != be->onstack_pages)
@@ -1430,38 +1420,34 @@ static void z_erofs_fill_bio_vec(struct bio_vec *bvec,
struct z_erofs_bvec zbv;
struct address_space *mapping;
struct page *page;
- int justfound, bs = i_blocksize(f->inode);
+ int bs = i_blocksize(f->inode);
- /* Except for inplace pages, the entire page can be used for I/Os */
+ /* Except for inplace folios, the entire folio can be used for I/Os */
bvec->bv_offset = 0;
bvec->bv_len = PAGE_SIZE;
repeat:
spin_lock(&pcl->obj.lockref.lock);
zbv = pcl->compressed_bvecs[nr];
- page = zbv.page;
- justfound = (unsigned long)page & 1UL;
- page = (struct page *)((unsigned long)page & ~1UL);
- pcl->compressed_bvecs[nr].page = page;
spin_unlock(&pcl->obj.lockref.lock);
- if (!page)
- goto out_allocpage;
+ if (!zbv.folio)
+ goto out_allocfolio;
- bvec->bv_page = page;
- DBG_BUGON(z_erofs_is_shortlived_page(page));
+ bvec->bv_page = &zbv.folio->page;
+ DBG_BUGON(z_erofs_is_shortlived_page(bvec->bv_page));
/*
- * Handle preallocated cached pages. We tried to allocate such pages
+ * Handle preallocated cached folios. We tried to allocate such folios
* without triggering direct reclaim. If allocation failed, inplace
- * file-backed pages will be used instead.
+ * file-backed folios will be used instead.
*/
- if (page->private == Z_EROFS_PREALLOCATED_PAGE) {
- set_page_private(page, 0);
+ if (zbv.folio->private == (void *)Z_EROFS_PREALLOCATED_PAGE) {
+ zbv.folio->private = 0;
tocache = true;
goto out_tocache;
}
- mapping = READ_ONCE(page->mapping);
+ mapping = READ_ONCE(zbv.folio->mapping);
/*
- * File-backed pages for inplace I/Os are all locked steady,
+ * File-backed folios for inplace I/Os are all locked steady,
* therefore it is impossible for `mapping` to be NULL.
*/
if (mapping && mapping != mc) {
@@ -1471,26 +1457,21 @@ repeat:
return;
}
- lock_page(page);
- /* only true if page reclaim goes wrong, should never happen */
- DBG_BUGON(justfound && PagePrivate(page));
-
- /* the cached page is still in managed cache */
- if (page->mapping == mc) {
+ folio_lock(zbv.folio);
+ if (zbv.folio->mapping == mc) {
/*
- * The cached page is still available but without a valid
- * `->private` pcluster hint. Let's reconnect them.
+ * The cached folio is still in managed cache but without
+ * a valid `->private` pcluster hint. Let's reconnect them.
*/
- if (!PagePrivate(page)) {
- DBG_BUGON(!justfound);
- /* compressed_bvecs[] already takes a ref */
- attach_page_private(page, pcl);
- put_page(page);
+ if (!folio_test_private(zbv.folio)) {
+ folio_attach_private(zbv.folio, pcl);
+ /* compressed_bvecs[] already takes a ref before */
+ folio_put(zbv.folio);
}
/* no need to submit if it is already up-to-date */
- if (PageUptodate(page)) {
- unlock_page(page);
+ if (folio_test_uptodate(zbv.folio)) {
+ folio_unlock(zbv.folio);
bvec->bv_page = NULL;
}
return;
@@ -1500,34 +1481,32 @@ repeat:
* It has been truncated, so it's unsafe to reuse this one. Let's
* allocate a new page for compressed data.
*/
- DBG_BUGON(page->mapping);
- DBG_BUGON(!justfound);
-
+ DBG_BUGON(zbv.folio->mapping);
tocache = true;
- unlock_page(page);
- put_page(page);
-out_allocpage:
+ folio_unlock(zbv.folio);
+ folio_put(zbv.folio);
+out_allocfolio:
page = erofs_allocpage(&f->pagepool, gfp | __GFP_NOFAIL);
spin_lock(&pcl->obj.lockref.lock);
- if (pcl->compressed_bvecs[nr].page) {
+ if (pcl->compressed_bvecs[nr].folio) {
erofs_pagepool_add(&f->pagepool, page);
spin_unlock(&pcl->obj.lockref.lock);
cond_resched();
goto repeat;
}
- pcl->compressed_bvecs[nr].page = page;
+ pcl->compressed_bvecs[nr].folio = zbv.folio = page_folio(page);
spin_unlock(&pcl->obj.lockref.lock);
bvec->bv_page = page;
out_tocache:
if (!tocache || bs != PAGE_SIZE ||
- add_to_page_cache_lru(page, mc, pcl->obj.index + nr, gfp)) {
- /* turn into a temporary shortlived page (1 ref) */
- set_page_private(page, Z_EROFS_SHORTLIVED_PAGE);
+ filemap_add_folio(mc, zbv.folio, pcl->obj.index + nr, gfp)) {
+ /* turn into a temporary shortlived folio (1 ref) */
+ zbv.folio->private = (void *)Z_EROFS_SHORTLIVED_PAGE;
return;
}
- attach_page_private(page, pcl);
+ folio_attach_private(zbv.folio, pcl);
/* drop a refcount added by allocpage (then 2 refs in total here) */
- put_page(page);
+ folio_put(zbv.folio);
}
static struct z_erofs_decompressqueue *jobqueue_init(struct super_block *sb,
@@ -1582,28 +1561,29 @@ static void move_to_bypass_jobqueue(struct z_erofs_pcluster *pcl,
qtail[JQ_BYPASS] = &pcl->next;
}
-static void z_erofs_submissionqueue_endio(struct bio *bio)
+static void z_erofs_endio(struct bio *bio)
{
struct z_erofs_decompressqueue *q = bio->bi_private;
blk_status_t err = bio->bi_status;
- struct bio_vec *bvec;
- struct bvec_iter_all iter_all;
+ struct folio_iter fi;
- bio_for_each_segment_all(bvec, bio, iter_all) {
- struct page *page = bvec->bv_page;
+ bio_for_each_folio_all(fi, bio) {
+ struct folio *folio = fi.folio;
- DBG_BUGON(PageUptodate(page));
- DBG_BUGON(z_erofs_page_is_invalidated(page));
- if (erofs_page_is_managed(EROFS_SB(q->sb), page)) {
- if (!err)
- SetPageUptodate(page);
- unlock_page(page);
- }
+ DBG_BUGON(folio_test_uptodate(folio));
+ DBG_BUGON(z_erofs_page_is_invalidated(&folio->page));
+ if (!erofs_folio_is_managed(EROFS_SB(q->sb), folio))
+ continue;
+
+ if (!err)
+ folio_mark_uptodate(folio);
+ folio_unlock(folio);
}
if (err)
q->eio = true;
z_erofs_decompress_kickoff(q, -1);
- bio_put(bio);
+ if (bio->bi_bdev)
+ bio_put(bio);
}
static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
@@ -1617,7 +1597,6 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
z_erofs_next_pcluster_t owned_head = f->owned_head;
/* bio is NULL initially, so no need to initialize last_{index,bdev} */
erofs_off_t last_pa;
- struct block_device *last_bdev;
unsigned int nr_bios = 0;
struct bio *bio = NULL;
unsigned long pflags;
@@ -1664,9 +1643,13 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
continue;
if (bio && (cur != last_pa ||
- last_bdev != mdev.m_bdev)) {
-submit_bio_retry:
- submit_bio(bio);
+ bio->bi_bdev != mdev.m_bdev)) {
+io_retry:
+ if (!erofs_is_fscache_mode(sb))
+ submit_bio(bio);
+ else
+ erofs_fscache_submit_bio(bio);
+
if (memstall) {
psi_memstall_leave(&pflags);
memstall = 0;
@@ -1681,15 +1664,16 @@ submit_bio_retry:
}
if (!bio) {
- bio = bio_alloc(mdev.m_bdev, BIO_MAX_VECS,
- REQ_OP_READ, GFP_NOIO);
- bio->bi_end_io = z_erofs_submissionqueue_endio;
+ bio = erofs_is_fscache_mode(sb) ?
+ erofs_fscache_bio_alloc(&mdev) :
+ bio_alloc(mdev.m_bdev, BIO_MAX_VECS,
+ REQ_OP_READ, GFP_NOIO);
+ bio->bi_end_io = z_erofs_endio;
bio->bi_iter.bi_sector = cur >> 9;
bio->bi_private = q[JQ_SUBMIT];
if (readahead)
bio->bi_opf |= REQ_RAHEAD;
++nr_bios;
- last_bdev = mdev.m_bdev;
}
if (cur + bvec.bv_len > end)
@@ -1697,7 +1681,7 @@ submit_bio_retry:
DBG_BUGON(bvec.bv_len < sb->s_blocksize);
if (!bio_add_page(bio, bvec.bv_page, bvec.bv_len,
bvec.bv_offset))
- goto submit_bio_retry;
+ goto io_retry;
last_pa = cur + bvec.bv_len;
bypass = false;
@@ -1710,7 +1694,10 @@ submit_bio_retry:
} while (owned_head != Z_EROFS_PCLUSTER_TAIL);
if (bio) {
- submit_bio(bio);
+ if (!erofs_is_fscache_mode(sb))
+ submit_bio(bio);
+ else
+ erofs_fscache_submit_bio(bio);
if (memstall)
psi_memstall_leave(&pflags);
}
@@ -1795,7 +1782,7 @@ static void z_erofs_pcluster_readmore(struct z_erofs_decompress_frontend *f,
if (PageUptodate(page))
unlock_page(page);
else
- (void)z_erofs_do_read_page(f, page, !!rac);
+ z_erofs_scan_folio(f, page_folio(page), !!rac);
put_page(page);
}
@@ -1816,7 +1803,7 @@ static int z_erofs_read_folio(struct file *file, struct folio *folio)
f.headoffset = (erofs_off_t)folio->index << PAGE_SHIFT;
z_erofs_pcluster_readmore(&f, NULL, true);
- err = z_erofs_do_read_page(&f, &folio->page, false);
+ err = z_erofs_scan_folio(&f, folio, false);
z_erofs_pcluster_readmore(&f, NULL, false);
z_erofs_pcluster_end(&f);
@@ -1857,7 +1844,7 @@ static void z_erofs_readahead(struct readahead_control *rac)
folio = head;
head = folio_get_private(folio);
- err = z_erofs_do_read_page(&f, &folio->page, true);
+ err = z_erofs_scan_folio(&f, folio, true);
if (err && err != -EINTR)
erofs_err(inode->i_sb, "readahead error at folio %lu @ nid %llu",
folio->index, EROFS_I(inode)->nid);