* [PATCH v2 1/2] erofs: simplify tail inline pcluster handling
@ 2025-02-24 12:37 Gao Xiang
2025-02-24 12:37 ` [PATCH v2 2/2] erofs: clean up header parsing for ztailpacking and fragments Gao Xiang
2025-02-24 12:40 ` [PATCH v2 1/2] erofs: simplify tail inline pcluster handling Gao Xiang
0 siblings, 2 replies; 6+ messages in thread
From: Gao Xiang @ 2025-02-24 12:37 UTC (permalink / raw)
To: linux-erofs; +Cc: LKML, Gao Xiang
Commit ab92184ff8f1 ("erofs: add on-disk compressed tail-packing inline
support") introduced the flag `Z_EROFS_ADVISE_INLINE_PCLUSTER`, which is
redundant because `h_idata_size != 0` serves the same purpose.
Additionally, merge `z_idataoff` and `z_fragmentoff` since these two
features are mutually exclusive for a given inode.
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
new patch.
fs/erofs/internal.h | 9 ++-------
fs/erofs/zmap.c | 19 ++++++++++---------
2 files changed, 12 insertions(+), 16 deletions(-)
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index f955793146f4..afce4b03f586 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -265,13 +265,8 @@ struct erofs_inode {
unsigned char z_algorithmtype[2];
unsigned char z_logical_clusterbits;
unsigned long z_tailextent_headlcn;
- union {
- struct {
- erofs_off_t z_idataoff;
- unsigned short z_idata_size;
- };
- erofs_off_t z_fragmentoff;
- };
+ erofs_off_t z_fragmentoff;
+ unsigned short z_idata_size;
};
#endif /* CONFIG_EROFS_FS_ZIP */
};
diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index a0c1c6450a55..9518d341cd82 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -395,8 +395,8 @@ static int z_erofs_do_map_blocks(struct inode *inode,
struct erofs_map_blocks *map, int flags)
{
struct erofs_inode *const vi = EROFS_I(inode);
- bool ztailpacking = vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER;
bool fragment = vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER;
+ bool ztailpacking = vi->z_idata_size;
struct z_erofs_maprecorder m = {
.inode = inode,
.map = map,
@@ -415,9 +415,8 @@ static int z_erofs_do_map_blocks(struct inode *inode,
if (err)
goto unmap_out;
- if (ztailpacking && (flags & EROFS_GET_BLOCKS_FINDTAIL))
- vi->z_idataoff = m.nextpackoff;
-
+ if ((flags & EROFS_GET_BLOCKS_FINDTAIL) && ztailpacking)
+ vi->z_fragmentoff = m.nextpackoff;
map->m_flags = EROFS_MAP_MAPPED | EROFS_MAP_ENCODED;
end = (m.lcn + 1ULL) << lclusterbits;
@@ -472,7 +471,7 @@ static int z_erofs_do_map_blocks(struct inode *inode,
}
if (ztailpacking && m.lcn == vi->z_tailextent_headlcn) {
map->m_flags |= EROFS_MAP_META;
- map->m_pa = vi->z_idataoff;
+ map->m_pa = vi->z_fragmentoff;
map->m_plen = vi->z_idata_size;
} else if (fragment && m.lcn == vi->z_tailextent_headlcn) {
map->m_flags |= EROFS_MAP_FRAGMENT;
@@ -565,6 +564,10 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
vi->z_advise = le16_to_cpu(h->h_advise);
vi->z_algorithmtype[0] = h->h_algorithmtype & 15;
vi->z_algorithmtype[1] = h->h_algorithmtype >> 4;
+ if (vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER)
+ vi->z_fragmentoff = le32_to_cpu(h->h_fragmentoff);
+ else
+ vi->z_idata_size = le16_to_cpu(h->h_idata_size);
headnr = 0;
if (vi->z_algorithmtype[0] >= Z_EROFS_COMPRESSION_MAX ||
@@ -593,18 +596,16 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
goto out_put_metabuf;
}
- if (vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER) {
+ if (vi->z_idata_size) {
struct erofs_map_blocks map = {
.buf = __EROFS_BUF_INITIALIZER
};
- vi->z_idata_size = le16_to_cpu(h->h_idata_size);
err = z_erofs_do_map_blocks(inode, &map,
EROFS_GET_BLOCKS_FINDTAIL);
erofs_put_metabuf(&map.buf);
- if (!map.m_plen ||
- erofs_blkoff(sb, map.m_pa) + map.m_plen > sb->s_blocksize) {
+ if (erofs_blkoff(sb, map.m_pa) + map.m_plen > sb->s_blocksize) {
erofs_err(sb, "invalid tail-packing pclustersize %llu",
map.m_plen);
err = -EFSCORRUPTED;
--
2.43.5
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v2 2/2] erofs: clean up header parsing for ztailpacking and fragments
2025-02-24 12:37 [PATCH v2 1/2] erofs: simplify tail inline pcluster handling Gao Xiang
@ 2025-02-24 12:37 ` Gao Xiang
2025-03-16 1:50 ` Chao Yu
2025-02-24 12:40 ` [PATCH v2 1/2] erofs: simplify tail inline pcluster handling Gao Xiang
1 sibling, 1 reply; 6+ messages in thread
From: Gao Xiang @ 2025-02-24 12:37 UTC (permalink / raw)
To: linux-erofs; +Cc: LKML, Gao Xiang
Simplify the logic in z_erofs_fill_inode_lazy() by combining the
handling of ztailpacking and fragments, as they are mutually exclusive.
Note that `h->h_clusterbits >> Z_EROFS_FRAGMENT_INODE_BIT` is handled
above, so no need to duplicate the check.
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
v2:
- rebase on the previous one;
fs/erofs/zmap.c | 40 ++++++++++++++--------------------------
1 file changed, 14 insertions(+), 26 deletions(-)
diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index 9518d341cd82..dd120306a272 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -394,7 +394,8 @@ static int z_erofs_get_extent_decompressedlen(struct z_erofs_maprecorder *m)
static int z_erofs_do_map_blocks(struct inode *inode,
struct erofs_map_blocks *map, int flags)
{
- struct erofs_inode *const vi = EROFS_I(inode);
+ struct erofs_inode *vi = EROFS_I(inode);
+ struct super_block *sb = inode->i_sb;
bool fragment = vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER;
bool ztailpacking = vi->z_idata_size;
struct z_erofs_maprecorder m = {
@@ -438,7 +439,7 @@ static int z_erofs_do_map_blocks(struct inode *inode,
}
/* m.lcn should be >= 1 if endoff < m.clusterofs */
if (!m.lcn) {
- erofs_err(inode->i_sb, "invalid logical cluster 0 at nid %llu",
+ erofs_err(sb, "invalid logical cluster 0 at nid %llu",
vi->nid);
err = -EFSCORRUPTED;
goto unmap_out;
@@ -454,7 +455,7 @@ static int z_erofs_do_map_blocks(struct inode *inode,
goto unmap_out;
break;
default:
- erofs_err(inode->i_sb, "unknown type %u @ offset %llu of nid %llu",
+ erofs_err(sb, "unknown type %u @ offset %llu of nid %llu",
m.type, ofs, vi->nid);
err = -EOPNOTSUPP;
goto unmap_out;
@@ -473,10 +474,16 @@ static int z_erofs_do_map_blocks(struct inode *inode,
map->m_flags |= EROFS_MAP_META;
map->m_pa = vi->z_fragmentoff;
map->m_plen = vi->z_idata_size;
+ if (erofs_blkoff(sb, map->m_pa) + map->m_plen > sb->s_blocksize) {
+ erofs_err(sb, "invalid tail-packing pclustersize %llu",
+ map->m_plen);
+ err = -EFSCORRUPTED;
+ goto unmap_out;
+ }
} else if (fragment && m.lcn == vi->z_tailextent_headlcn) {
map->m_flags |= EROFS_MAP_FRAGMENT;
} else {
- map->m_pa = erofs_pos(inode->i_sb, m.pblk);
+ map->m_pa = erofs_pos(sb, m.pblk);
err = z_erofs_get_extent_compressedlen(&m, initial_lcn);
if (err)
goto unmap_out;
@@ -495,7 +502,7 @@ static int z_erofs_do_map_blocks(struct inode *inode,
afmt = m.headtype == Z_EROFS_LCLUSTER_TYPE_HEAD2 ?
vi->z_algorithmtype[1] : vi->z_algorithmtype[0];
if (!(EROFS_I_SB(inode)->available_compr_algs & (1 << afmt))) {
- erofs_err(inode->i_sb, "inconsistent algorithmtype %u for nid %llu",
+ erofs_err(sb, "inconsistent algorithmtype %u for nid %llu",
afmt, vi->nid);
err = -EFSCORRUPTED;
goto unmap_out;
@@ -596,31 +603,12 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
goto out_put_metabuf;
}
- if (vi->z_idata_size) {
+ if (vi->z_idata_size ||
+ (vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER)) {
struct erofs_map_blocks map = {
.buf = __EROFS_BUF_INITIALIZER
};
- err = z_erofs_do_map_blocks(inode, &map,
- EROFS_GET_BLOCKS_FINDTAIL);
- erofs_put_metabuf(&map.buf);
-
- if (erofs_blkoff(sb, map.m_pa) + map.m_plen > sb->s_blocksize) {
- erofs_err(sb, "invalid tail-packing pclustersize %llu",
- map.m_plen);
- err = -EFSCORRUPTED;
- }
- if (err < 0)
- goto out_put_metabuf;
- }
-
- if (vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER &&
- !(h->h_clusterbits >> Z_EROFS_FRAGMENT_INODE_BIT)) {
- struct erofs_map_blocks map = {
- .buf = __EROFS_BUF_INITIALIZER
- };
-
- vi->z_fragmentoff = le32_to_cpu(h->h_fragmentoff);
err = z_erofs_do_map_blocks(inode, &map,
EROFS_GET_BLOCKS_FINDTAIL);
erofs_put_metabuf(&map.buf);
--
2.43.5
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v2 1/2] erofs: simplify tail inline pcluster handling
2025-02-24 12:37 [PATCH v2 1/2] erofs: simplify tail inline pcluster handling Gao Xiang
2025-02-24 12:37 ` [PATCH v2 2/2] erofs: clean up header parsing for ztailpacking and fragments Gao Xiang
@ 2025-02-24 12:40 ` Gao Xiang
2025-02-25 11:40 ` [PATCH v3 " Gao Xiang
1 sibling, 1 reply; 6+ messages in thread
From: Gao Xiang @ 2025-02-24 12:40 UTC (permalink / raw)
To: linux-erofs; +Cc: LKML
On 2025/2/24 20:37, Gao Xiang wrote:
> Commit ab92184ff8f1 ("erofs: add on-disk compressed tail-packing inline
> support") introduced the flag `Z_EROFS_ADVISE_INLINE_PCLUSTER`, which is
> redundant because `h_idata_size != 0` serves the same purpose.
>
> Additionally, merge `z_idataoff` and `z_fragmentoff` since these two
> features are mutually exclusive for a given inode.
>
> Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
> ---
> new patch.
Note that `Z_EROFS_ADVISE_INLINE_PCLUSTER` now is left for
the compatibility purposes, only `mkfs.erofs` uses it in
the future.
Thanks,
Gao Xiang
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v3 1/2] erofs: simplify tail inline pcluster handling
2025-02-24 12:40 ` [PATCH v2 1/2] erofs: simplify tail inline pcluster handling Gao Xiang
@ 2025-02-25 11:40 ` Gao Xiang
2025-03-16 1:39 ` Chao Yu
0 siblings, 1 reply; 6+ messages in thread
From: Gao Xiang @ 2025-02-25 11:40 UTC (permalink / raw)
To: linux-erofs; +Cc: LKML, Gao Xiang
Use `z_idata_size != 0` to indicate that ztailpacking is enabled.
`Z_EROFS_ADVISE_INLINE_PCLUSTER` cannot be ignored, as `h_idata_size`
could be non-zero prior to erofs-utils 1.6 [1].
Additionally, merge `z_idataoff` and `z_fragmentoff` since these two
features are mutually exclusive for a given inode.
[1] https://git.kernel.org/xiang/erofs-utils/c/547bea3cb71a
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
---
change since v2:
- should not ignore `Z_EROFS_ADVISE_INLINE_PCLUSTER` as mentioned in
the new commit message.
fs/erofs/internal.h | 9 ++-------
fs/erofs/zmap.c | 20 ++++++++++----------
2 files changed, 12 insertions(+), 17 deletions(-)
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index b452b6557aa5..b357cbbce764 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -263,13 +263,8 @@ struct erofs_inode {
unsigned char z_algorithmtype[2];
unsigned char z_logical_clusterbits;
unsigned long z_tailextent_headlcn;
- union {
- struct {
- erofs_off_t z_idataoff;
- unsigned short z_idata_size;
- };
- erofs_off_t z_fragmentoff;
- };
+ erofs_off_t z_fragmentoff;
+ unsigned short z_idata_size;
};
#endif /* CONFIG_EROFS_FS_ZIP */
};
diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index a0c1c6450a55..ba4533eabc6a 100644
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -395,8 +395,8 @@ static int z_erofs_do_map_blocks(struct inode *inode,
struct erofs_map_blocks *map, int flags)
{
struct erofs_inode *const vi = EROFS_I(inode);
- bool ztailpacking = vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER;
bool fragment = vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER;
+ bool ztailpacking = vi->z_idata_size;
struct z_erofs_maprecorder m = {
.inode = inode,
.map = map,
@@ -415,9 +415,8 @@ static int z_erofs_do_map_blocks(struct inode *inode,
if (err)
goto unmap_out;
- if (ztailpacking && (flags & EROFS_GET_BLOCKS_FINDTAIL))
- vi->z_idataoff = m.nextpackoff;
-
+ if ((flags & EROFS_GET_BLOCKS_FINDTAIL) && ztailpacking)
+ vi->z_fragmentoff = m.nextpackoff;
map->m_flags = EROFS_MAP_MAPPED | EROFS_MAP_ENCODED;
end = (m.lcn + 1ULL) << lclusterbits;
@@ -472,7 +471,7 @@ static int z_erofs_do_map_blocks(struct inode *inode,
}
if (ztailpacking && m.lcn == vi->z_tailextent_headlcn) {
map->m_flags |= EROFS_MAP_META;
- map->m_pa = vi->z_idataoff;
+ map->m_pa = vi->z_fragmentoff;
map->m_plen = vi->z_idata_size;
} else if (fragment && m.lcn == vi->z_tailextent_headlcn) {
map->m_flags |= EROFS_MAP_FRAGMENT;
@@ -565,6 +564,10 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
vi->z_advise = le16_to_cpu(h->h_advise);
vi->z_algorithmtype[0] = h->h_algorithmtype & 15;
vi->z_algorithmtype[1] = h->h_algorithmtype >> 4;
+ if (vi->z_advise & Z_EROFS_ADVISE_FRAGMENT_PCLUSTER)
+ vi->z_fragmentoff = le32_to_cpu(h->h_fragmentoff);
+ else if (vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER)
+ vi->z_idata_size = le16_to_cpu(h->h_idata_size);
headnr = 0;
if (vi->z_algorithmtype[0] >= Z_EROFS_COMPRESSION_MAX ||
@@ -593,18 +596,16 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
goto out_put_metabuf;
}
- if (vi->z_advise & Z_EROFS_ADVISE_INLINE_PCLUSTER) {
+ if (vi->z_idata_size) {
struct erofs_map_blocks map = {
.buf = __EROFS_BUF_INITIALIZER
};
- vi->z_idata_size = le16_to_cpu(h->h_idata_size);
err = z_erofs_do_map_blocks(inode, &map,
EROFS_GET_BLOCKS_FINDTAIL);
erofs_put_metabuf(&map.buf);
- if (!map.m_plen ||
- erofs_blkoff(sb, map.m_pa) + map.m_plen > sb->s_blocksize) {
+ if (erofs_blkoff(sb, map.m_pa) + map.m_plen > sb->s_blocksize) {
erofs_err(sb, "invalid tail-packing pclustersize %llu",
map.m_plen);
err = -EFSCORRUPTED;
@@ -619,7 +620,6 @@ static int z_erofs_fill_inode_lazy(struct inode *inode)
.buf = __EROFS_BUF_INITIALIZER
};
- vi->z_fragmentoff = le32_to_cpu(h->h_fragmentoff);
err = z_erofs_do_map_blocks(inode, &map,
EROFS_GET_BLOCKS_FINDTAIL);
erofs_put_metabuf(&map.buf);
--
2.43.5
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v3 1/2] erofs: simplify tail inline pcluster handling
2025-02-25 11:40 ` [PATCH v3 " Gao Xiang
@ 2025-03-16 1:39 ` Chao Yu
0 siblings, 0 replies; 6+ messages in thread
From: Chao Yu @ 2025-03-16 1:39 UTC (permalink / raw)
To: Gao Xiang, linux-erofs; +Cc: chao, LKML
On 2025/2/25 19:40, Gao Xiang wrote:
> Use `z_idata_size != 0` to indicate that ztailpacking is enabled.
> `Z_EROFS_ADVISE_INLINE_PCLUSTER` cannot be ignored, as `h_idata_size`
> could be non-zero prior to erofs-utils 1.6 [1].
>
> Additionally, merge `z_idataoff` and `z_fragmentoff` since these two
> features are mutually exclusive for a given inode.
>
> [1] https://git.kernel.org/xiang/erofs-utils/c/547bea3cb71a
> Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Thanks,
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v2 2/2] erofs: clean up header parsing for ztailpacking and fragments
2025-02-24 12:37 ` [PATCH v2 2/2] erofs: clean up header parsing for ztailpacking and fragments Gao Xiang
@ 2025-03-16 1:50 ` Chao Yu
0 siblings, 0 replies; 6+ messages in thread
From: Chao Yu @ 2025-03-16 1:50 UTC (permalink / raw)
To: Gao Xiang, linux-erofs; +Cc: chao, LKML
On 2025/2/24 20:37, Gao Xiang wrote:
> Simplify the logic in z_erofs_fill_inode_lazy() by combining the
> handling of ztailpacking and fragments, as they are mutually exclusive.
>
> Note that `h->h_clusterbits >> Z_EROFS_FRAGMENT_INODE_BIT` is handled
> above, so no need to duplicate the check.
>
> Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Thanks,
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-03-16 1:50 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-02-24 12:37 [PATCH v2 1/2] erofs: simplify tail inline pcluster handling Gao Xiang
2025-02-24 12:37 ` [PATCH v2 2/2] erofs: clean up header parsing for ztailpacking and fragments Gao Xiang
2025-03-16 1:50 ` Chao Yu
2025-02-24 12:40 ` [PATCH v2 1/2] erofs: simplify tail inline pcluster handling Gao Xiang
2025-02-25 11:40 ` [PATCH v3 " Gao Xiang
2025-03-16 1:39 ` Chao Yu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox