[f2fs-dev] [PATCH 0/8] Revert setting casefolding dentry operations through s_d_op

[f2fs-dev] [PATCH 0/8] Revert setting casefolding dentry operations through s_d_op

[Gabriel Krisman Bertazi]

When case-insensitive and fscrypt were adapted to work together, we moved the
code that sets the dentry operations for case-insensitive dentries(d_hash and
d_compare) to happen from a helper inside ->lookup.  This is because fscrypt
wants to set d_revalidate only on some dentries, so it does it only for them in
d_revalidate.

But, case-insensitive hooks are actually set on all dentries in the filesystem,
so the natural place to do it is through s_d_op and let d_alloc handle it [1].
In addition, doing it inside the ->lookup is a problem for case-insensitive
dentries that are not created through ->lookup, like those coming
open-by-fhandle[2], which will not see the required d_ops.

This patchset therefore reverts to using sb->s_d_op to set the dentry operations
for case-insensitive filesystems.  In order to set case-insensitive hooks early
and not require every dentry to have d_revalidate in case-insensitive
filesystems, it introduces a patch suggested by Al Viro to disable d_revalidate
on some dentries on the fly.

It survives fstests encrypt and quick groups without regressions.  Based on v6.7-rc1.

[1] https://lore.kernel.org/linux-fsdevel/20231123195327.GP38156@ZenIV/
[2] https://lore.kernel.org/linux-fsdevel/20231123171255.GN38156@ZenIV/

Gabriel Krisman Bertazi (8):
  dcache: Add helper to disable d_revalidate for a specific dentry
  fscrypt: Drop d_revalidate if key is available
  libfs: Merge encrypted_ci_dentry_ops and ci_dentry_ops
  libfs: Expose generic_ci_dentry_ops outside of libfs
  ext4: Set the case-insensitive dentry operations through ->s_d_op
  f2fs: Set the case-insensitive dentry operations through ->s_d_op
  libfs: Don't support setting casefold operations during lookup
  fscrypt: Move d_revalidate configuration back into fscrypt

 fs/crypto/fname.c       |  9 +++++-
 fs/crypto/hooks.c       |  8 ++++++
 fs/dcache.c             | 10 +++++++
 fs/ext4/namei.c         |  1 -
 fs/ext4/super.c         |  3 ++
 fs/f2fs/namei.c         |  1 -
 fs/f2fs/super.c         |  3 ++
 fs/libfs.c              | 64 +++--------------------------------------
 fs/ubifs/dir.c          |  1 -
 include/linux/dcache.h  |  1 +
 include/linux/fs.h      |  2 +-
 include/linux/fscrypt.h | 10 +++----
 12 files changed, 43 insertions(+), 70 deletions(-)

-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 1/8] dcache: Add helper to disable d_revalidate for a specific dentry

[Gabriel Krisman Bertazi]

Case-insensitive wants d_compare/d_hash for every dentry in the
filesystem, while fscrypt needs d_revalidate only for DCACHE_NOKEY_NAME.
This means we currently can't use sb->s_d_op to set case-insensitive
hooks in fscrypt+case-insensitive filesystems without paying the cost to
call d_revalidate for every dentry in the filesystem.

In preparation to doing exactly that, add a way to disable d_revalidate
later.

Suggested-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/dcache.c            | 10 ++++++++++
 include/linux/dcache.h |  1 +
 2 files changed, 11 insertions(+)

diff --git a/fs/dcache.c b/fs/dcache.c
index c82ae731df9a..1f5464cd3bd1 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -1911,6 +1911,16 @@ struct dentry *d_alloc_name(struct dentry *parent, const char *name)
 }
 EXPORT_SYMBOL(d_alloc_name);
 
+void d_set_always_valid(struct dentry *dentry)
+{
+	if (!(dentry->d_flags & DCACHE_OP_REVALIDATE))
+		return;
+
+	spin_lock(&dentry->d_lock);
+	dentry->d_flags &= ~DCACHE_OP_REVALIDATE;
+	spin_unlock(&dentry->d_lock);
+}
+
 void d_set_d_op(struct dentry *dentry, const struct dentry_operations *op)
 {
 	WARN_ON_ONCE(dentry->d_op);
diff --git a/include/linux/dcache.h b/include/linux/dcache.h
index 3da2f0545d5d..d2ce151b2d8e 100644
--- a/include/linux/dcache.h
+++ b/include/linux/dcache.h
@@ -225,6 +225,7 @@ extern struct dentry * d_instantiate_anon(struct dentry *, struct inode *);
 extern void __d_drop(struct dentry *dentry);
 extern void d_drop(struct dentry *dentry);
 extern void d_delete(struct dentry *);
+extern void d_set_always_valid(struct dentry *dentry);
 extern void d_set_d_op(struct dentry *dentry, const struct dentry_operations *op);
 
 /* allocate/de-allocate */
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 2/8] fscrypt: Drop d_revalidate if key is available

[Gabriel Krisman Bertazi]

fscrypt dentries are always valid once the key is available.  Since the
key cannot be removed without evicting the dentry, we don't need to keep
retrying to revalidate it.

Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/crypto/fname.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/fs/crypto/fname.c b/fs/crypto/fname.c
index 7b3fc189593a..0457ba2d7d76 100644
--- a/fs/crypto/fname.c
+++ b/fs/crypto/fname.c
@@ -591,8 +591,15 @@ int fscrypt_d_revalidate(struct dentry *dentry, unsigned int flags)
 	 * reverting to no-key names without evicting the directory's inode
 	 * -- which implies eviction of the dentries in the directory.
 	 */
-	if (!(dentry->d_flags & DCACHE_NOKEY_NAME))
+	if (!(dentry->d_flags & DCACHE_NOKEY_NAME)) {
+		/*
+		 * If fscrypt is the only feature requiring
+		 * revalidation for this dentry, we can just disable it.
+		 */
+		if (dentry->d_op->d_revalidate == &fscrypt_d_revalidate)
+			d_set_always_valid(dentry);
 		return 1;
+	}
 
 	/*
 	 * No-key name; valid if the directory's key is still unavailable.
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 3/8] libfs: Merge encrypted_ci_dentry_ops and ci_dentry_ops

[Gabriel Krisman Bertazi]

In preparation to get case-insensitive dentry operations from
sb->s_d_op again, use the same structure for case-insensitive
filesystems with and without fscrypt.

This means that on a casefolded filesystem without fscrypt, we end up
having to call fscrypt_d_revalidate once per dentry, which does the
function call extra cost.  We could avoid it by calling
d_set_always_valid in generic_set_encrypted_ci_d_ops, but this entire
function will go away in the following patches, and it is not worth the
extra complexity. Also, since the first fscrypt_d_revalidate will call
d_set_always_valid for us, we'll only have only pay the cost once, and
not per-lookup.

Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/libfs.c | 34 ++++++----------------------------
 1 file changed, 6 insertions(+), 28 deletions(-)

diff --git a/fs/libfs.c b/fs/libfs.c
index e9440d55073c..52c944173e57 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -1768,6 +1768,9 @@ static int generic_ci_d_hash(const struct dentry *dentry, struct qstr *str)
 static const struct dentry_operations generic_ci_dentry_ops = {
 	.d_hash = generic_ci_d_hash,
 	.d_compare = generic_ci_d_compare,
+#if defined(CONFIG_FS_ENCRYPTION)
+	.d_revalidate = fscrypt_d_revalidate,
+#endif
 };
 #endif
 
@@ -1777,14 +1780,6 @@ static const struct dentry_operations generic_encrypted_dentry_ops = {
 };
 #endif
 
-#if defined(CONFIG_FS_ENCRYPTION) && IS_ENABLED(CONFIG_UNICODE)
-static const struct dentry_operations generic_encrypted_ci_dentry_ops = {
-	.d_hash = generic_ci_d_hash,
-	.d_compare = generic_ci_d_compare,
-	.d_revalidate = fscrypt_d_revalidate,
-};
-#endif
-
 /**
  * generic_set_encrypted_ci_d_ops - helper for setting d_ops for given dentry
  * @dentry:	dentry to set ops on
@@ -1801,38 +1796,21 @@ static const struct dentry_operations generic_encrypted_ci_dentry_ops = {
  * Encryption works differently in that the only dentry operation it needs is
  * d_revalidate, which it only needs on dentries that have the no-key name flag.
  * The no-key flag can't be set "later", so we don't have to worry about that.
- *
- * Finally, to maximize compatibility with overlayfs (which isn't compatible
- * with certain dentry operations) and to avoid taking an unnecessary
- * performance hit, we use custom dentry_operations for each possible
- * combination rather than always installing all operations.
  */
 void generic_set_encrypted_ci_d_ops(struct dentry *dentry)
 {
-#ifdef CONFIG_FS_ENCRYPTION
-	bool needs_encrypt_ops = dentry->d_flags & DCACHE_NOKEY_NAME;
-#endif
 #if IS_ENABLED(CONFIG_UNICODE)
-	bool needs_ci_ops = dentry->d_sb->s_encoding;
-#endif
-#if defined(CONFIG_FS_ENCRYPTION) && IS_ENABLED(CONFIG_UNICODE)
-	if (needs_encrypt_ops && needs_ci_ops) {
-		d_set_d_op(dentry, &generic_encrypted_ci_dentry_ops);
+	if (dentry->d_sb->s_encoding) {
+		d_set_d_op(dentry, &generic_ci_dentry_ops);
 		return;
 	}
 #endif
 #ifdef CONFIG_FS_ENCRYPTION
-	if (needs_encrypt_ops) {
+	if (dentry->d_flags & DCACHE_NOKEY_NAME) {
 		d_set_d_op(dentry, &generic_encrypted_dentry_ops);
 		return;
 	}
 #endif
-#if IS_ENABLED(CONFIG_UNICODE)
-	if (needs_ci_ops) {
-		d_set_d_op(dentry, &generic_ci_dentry_ops);
-		return;
-	}
-#endif
 }
 EXPORT_SYMBOL(generic_set_encrypted_ci_d_ops);
 
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 4/8] libfs: Expose generic_ci_dentry_ops outside of libfs

[Gabriel Krisman Bertazi]

In preparation to allow filesystems to set this through sb->s_d_op,
expose the symbol directly to the filesystems.

Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/libfs.c         | 2 +-
 include/linux/fs.h | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/libfs.c b/fs/libfs.c
index 52c944173e57..b8ecada3a5b2 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -1765,7 +1765,7 @@ static int generic_ci_d_hash(const struct dentry *dentry, struct qstr *str)
 	return 0;
 }
 
-static const struct dentry_operations generic_ci_dentry_ops = {
+const struct dentry_operations generic_ci_dentry_ops = {
 	.d_hash = generic_ci_d_hash,
 	.d_compare = generic_ci_d_compare,
 #if defined(CONFIG_FS_ENCRYPTION)
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 98b7a7a8c42e..887a27d07f96 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -3202,6 +3202,7 @@ extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
 extern int generic_check_addressable(unsigned, u64);
 
 extern void generic_set_encrypted_ci_d_ops(struct dentry *dentry);
+extern const struct dentry_operations generic_ci_dentry_ops;
 
 int may_setattr(struct mnt_idmap *idmap, struct inode *inode,
 		unsigned int ia_valid);
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 5/8] ext4: Set the case-insensitive dentry operations through ->s_d_op

[Gabriel Krisman Bertazi]

All dentries in a case-insensitive filesystem have the same set of
dentry operations.  Therefore, we should let VFS propagate them from
sb->s_d_op d_alloc instead of setting at lookup time.

This was already the case before commit bb9cd9106b22 ("fscrypt: Have
filesystems handle their d_ops"), but it was changed to set at
lookup-time to facilitate the integration with fscrypt.  But it's a
problem because dentries that don't get created through ->lookup() won't
have any visibility of the operations.  Let's revert to the previous
implementation.

Suggested-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/ext4/namei.c | 6 +++++-
 fs/ext4/super.c | 3 +++
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index d252935f9c8a..3c1208d5d85b 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1762,7 +1762,11 @@ static struct buffer_head *ext4_lookup_entry(struct inode *dir,
 	struct buffer_head *bh;
 
 	err = ext4_fname_prepare_lookup(dir, dentry, &fname);
-	generic_set_encrypted_ci_d_ops(dentry);
+
+	/* Case-insensitive volumes set dentry ops through sb->s_d_op. */
+	if (!dir->i_sb->s_encoding)
+		generic_set_encrypted_ci_d_ops(dentry);
+
 	if (err == -ENOENT)
 		return NULL;
 	if (err)
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index c5fcf377ab1f..5ac1c9df9956 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -5493,6 +5493,9 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
 		goto failed_mount4;
 	}
 
+	if (sb->s_encoding)
+		sb->s_d_op = &generic_ci_dentry_ops;
+
 	sb->s_root = d_make_root(root);
 	if (!sb->s_root) {
 		ext4_msg(sb, KERN_ERR, "get root dentry failed");
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Re: [f2fs-dev] [PATCH 5/8] ext4: Set the case-insensitive dentry operations through ->s_d_op

[kernel test robot]

Hi Gabriel,

kernel test robot noticed the following build errors:

[auto build test ERROR on jaegeuk-f2fs/dev-test]
[also build test ERROR on jaegeuk-f2fs/dev tytso-ext4/dev linus/master v6.7-rc5 next-20231214]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Gabriel-Krisman-Bertazi/dcache-Add-helper-to-disable-d_revalidate-for-a-specific-dentry/20231214-074322
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git dev-test
patch link:    https://lore.kernel.org/r/20231213234031.1081-6-krisman%40suse.de
patch subject: [PATCH 5/8] ext4: Set the case-insensitive dentry operations through ->s_d_op
config: arc-vdk_hs38_defconfig (https://download.01.org/0day-ci/archive/20231214/202312142039.u2475qPS-lkp@intel.com/config)
compiler: arc-elf-gcc (GCC) 13.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231214/202312142039.u2475qPS-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202312142039.u2475qPS-lkp@intel.com/

All errors (new ones prefixed by >>):

   fs/ext4/super.c: In function '__ext4_fill_super':
>> fs/ext4/super.c:5496:15: error: 'struct super_block' has no member named 's_encoding'
    5496 |         if (sb->s_encoding)
         |               ^~


vim +5496 fs/ext4/super.c

  5215	
  5216	static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb)
  5217	{
  5218		struct ext4_super_block *es = NULL;
  5219		struct ext4_sb_info *sbi = EXT4_SB(sb);
  5220		ext4_fsblk_t logical_sb_block;
  5221		struct inode *root;
  5222		int needs_recovery;
  5223		int err;
  5224		ext4_group_t first_not_zeroed;
  5225		struct ext4_fs_context *ctx = fc->fs_private;
  5226		int silent = fc->sb_flags & SB_SILENT;
  5227	
  5228		/* Set defaults for the variables that will be set during parsing */
  5229		if (!(ctx->spec & EXT4_SPEC_JOURNAL_IOPRIO))
  5230			ctx->journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
  5231	
  5232		sbi->s_inode_readahead_blks = EXT4_DEF_INODE_READAHEAD_BLKS;
  5233		sbi->s_sectors_written_start =
  5234			part_stat_read(sb->s_bdev, sectors[STAT_WRITE]);
  5235	
  5236		err = ext4_load_super(sb, &logical_sb_block, silent);
  5237		if (err)
  5238			goto out_fail;
  5239	
  5240		es = sbi->s_es;
  5241		sbi->s_kbytes_written = le64_to_cpu(es->s_kbytes_written);
  5242	
  5243		err = ext4_init_metadata_csum(sb, es);
  5244		if (err)
  5245			goto failed_mount;
  5246	
  5247		ext4_set_def_opts(sb, es);
  5248	
  5249		sbi->s_resuid = make_kuid(&init_user_ns, le16_to_cpu(es->s_def_resuid));
  5250		sbi->s_resgid = make_kgid(&init_user_ns, le16_to_cpu(es->s_def_resgid));
  5251		sbi->s_commit_interval = JBD2_DEFAULT_MAX_COMMIT_AGE * HZ;
  5252		sbi->s_min_batch_time = EXT4_DEF_MIN_BATCH_TIME;
  5253		sbi->s_max_batch_time = EXT4_DEF_MAX_BATCH_TIME;
  5254	
  5255		/*
  5256		 * set default s_li_wait_mult for lazyinit, for the case there is
  5257		 * no mount option specified.
  5258		 */
  5259		sbi->s_li_wait_mult = EXT4_DEF_LI_WAIT_MULT;
  5260	
  5261		err = ext4_inode_info_init(sb, es);
  5262		if (err)
  5263			goto failed_mount;
  5264	
  5265		err = parse_apply_sb_mount_options(sb, ctx);
  5266		if (err < 0)
  5267			goto failed_mount;
  5268	
  5269		sbi->s_def_mount_opt = sbi->s_mount_opt;
  5270		sbi->s_def_mount_opt2 = sbi->s_mount_opt2;
  5271	
  5272		err = ext4_check_opt_consistency(fc, sb);
  5273		if (err < 0)
  5274			goto failed_mount;
  5275	
  5276		ext4_apply_options(fc, sb);
  5277	
  5278		err = ext4_encoding_init(sb, es);
  5279		if (err)
  5280			goto failed_mount;
  5281	
  5282		err = ext4_check_journal_data_mode(sb);
  5283		if (err)
  5284			goto failed_mount;
  5285	
  5286		sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
  5287			(test_opt(sb, POSIX_ACL) ? SB_POSIXACL : 0);
  5288	
  5289		/* i_version is always enabled now */
  5290		sb->s_flags |= SB_I_VERSION;
  5291	
  5292		err = ext4_check_feature_compatibility(sb, es, silent);
  5293		if (err)
  5294			goto failed_mount;
  5295	
  5296		err = ext4_block_group_meta_init(sb, silent);
  5297		if (err)
  5298			goto failed_mount;
  5299	
  5300		ext4_hash_info_init(sb);
  5301	
  5302		err = ext4_handle_clustersize(sb);
  5303		if (err)
  5304			goto failed_mount;
  5305	
  5306		err = ext4_check_geometry(sb, es);
  5307		if (err)
  5308			goto failed_mount;
  5309	
  5310		timer_setup(&sbi->s_err_report, print_daily_error_info, 0);
  5311		spin_lock_init(&sbi->s_error_lock);
  5312		INIT_WORK(&sbi->s_sb_upd_work, update_super_work);
  5313	
  5314		err = ext4_group_desc_init(sb, es, logical_sb_block, &first_not_zeroed);
  5315		if (err)
  5316			goto failed_mount3;
  5317	
  5318		err = ext4_es_register_shrinker(sbi);
  5319		if (err)
  5320			goto failed_mount3;
  5321	
  5322		sbi->s_stripe = ext4_get_stripe_size(sbi);
  5323		/*
  5324		 * It's hard to get stripe aligned blocks if stripe is not aligned with
  5325		 * cluster, just disable stripe and alert user to simpfy code and avoid
  5326		 * stripe aligned allocation which will rarely successes.
  5327		 */
  5328		if (sbi->s_stripe > 0 && sbi->s_cluster_ratio > 1 &&
  5329		    sbi->s_stripe % sbi->s_cluster_ratio != 0) {
  5330			ext4_msg(sb, KERN_WARNING,
  5331				 "stripe (%lu) is not aligned with cluster size (%u), "
  5332				 "stripe is disabled",
  5333				 sbi->s_stripe, sbi->s_cluster_ratio);
  5334			sbi->s_stripe = 0;
  5335		}
  5336		sbi->s_extent_max_zeroout_kb = 32;
  5337	
  5338		/*
  5339		 * set up enough so that it can read an inode
  5340		 */
  5341		sb->s_op = &ext4_sops;
  5342		sb->s_export_op = &ext4_export_ops;
  5343		sb->s_xattr = ext4_xattr_handlers;
  5344	#ifdef CONFIG_FS_ENCRYPTION
  5345		sb->s_cop = &ext4_cryptops;
  5346	#endif
  5347	#ifdef CONFIG_FS_VERITY
  5348		sb->s_vop = &ext4_verityops;
  5349	#endif
  5350	#ifdef CONFIG_QUOTA
  5351		sb->dq_op = &ext4_quota_operations;
  5352		if (ext4_has_feature_quota(sb))
  5353			sb->s_qcop = &dquot_quotactl_sysfile_ops;
  5354		else
  5355			sb->s_qcop = &ext4_qctl_operations;
  5356		sb->s_quota_types = QTYPE_MASK_USR | QTYPE_MASK_GRP | QTYPE_MASK_PRJ;
  5357	#endif
  5358		memcpy(&sb->s_uuid, es->s_uuid, sizeof(es->s_uuid));
  5359	
  5360		INIT_LIST_HEAD(&sbi->s_orphan); /* unlinked but open files */
  5361		mutex_init(&sbi->s_orphan_lock);
  5362	
  5363		ext4_fast_commit_init(sb);
  5364	
  5365		sb->s_root = NULL;
  5366	
  5367		needs_recovery = (es->s_last_orphan != 0 ||
  5368				  ext4_has_feature_orphan_present(sb) ||
  5369				  ext4_has_feature_journal_needs_recovery(sb));
  5370	
  5371		if (ext4_has_feature_mmp(sb) && !sb_rdonly(sb)) {
  5372			err = ext4_multi_mount_protect(sb, le64_to_cpu(es->s_mmp_block));
  5373			if (err)
  5374				goto failed_mount3a;
  5375		}
  5376	
  5377		err = -EINVAL;
  5378		/*
  5379		 * The first inode we look at is the journal inode.  Don't try
  5380		 * root first: it may be modified in the journal!
  5381		 */
  5382		if (!test_opt(sb, NOLOAD) && ext4_has_feature_journal(sb)) {
  5383			err = ext4_load_and_init_journal(sb, es, ctx);
  5384			if (err)
  5385				goto failed_mount3a;
  5386		} else if (test_opt(sb, NOLOAD) && !sb_rdonly(sb) &&
  5387			   ext4_has_feature_journal_needs_recovery(sb)) {
  5388			ext4_msg(sb, KERN_ERR, "required journal recovery "
  5389			       "suppressed and not mounted read-only");
  5390			goto failed_mount3a;
  5391		} else {
  5392			/* Nojournal mode, all journal mount options are illegal */
  5393			if (test_opt(sb, JOURNAL_ASYNC_COMMIT)) {
  5394				ext4_msg(sb, KERN_ERR, "can't mount with "
  5395					 "journal_async_commit, fs mounted w/o journal");
  5396				goto failed_mount3a;
  5397			}
  5398	
  5399			if (test_opt2(sb, EXPLICIT_JOURNAL_CHECKSUM)) {
  5400				ext4_msg(sb, KERN_ERR, "can't mount with "
  5401					 "journal_checksum, fs mounted w/o journal");
  5402				goto failed_mount3a;
  5403			}
  5404			if (sbi->s_commit_interval != JBD2_DEFAULT_MAX_COMMIT_AGE*HZ) {
  5405				ext4_msg(sb, KERN_ERR, "can't mount with "
  5406					 "commit=%lu, fs mounted w/o journal",
  5407					 sbi->s_commit_interval / HZ);
  5408				goto failed_mount3a;
  5409			}
  5410			if (EXT4_MOUNT_DATA_FLAGS &
  5411			    (sbi->s_mount_opt ^ sbi->s_def_mount_opt)) {
  5412				ext4_msg(sb, KERN_ERR, "can't mount with "
  5413					 "data=, fs mounted w/o journal");
  5414				goto failed_mount3a;
  5415			}
  5416			sbi->s_def_mount_opt &= ~EXT4_MOUNT_JOURNAL_CHECKSUM;
  5417			clear_opt(sb, JOURNAL_CHECKSUM);
  5418			clear_opt(sb, DATA_FLAGS);
  5419			clear_opt2(sb, JOURNAL_FAST_COMMIT);
  5420			sbi->s_journal = NULL;
  5421			needs_recovery = 0;
  5422		}
  5423	
  5424		if (!test_opt(sb, NO_MBCACHE)) {
  5425			sbi->s_ea_block_cache = ext4_xattr_create_cache();
  5426			if (!sbi->s_ea_block_cache) {
  5427				ext4_msg(sb, KERN_ERR,
  5428					 "Failed to create ea_block_cache");
  5429				err = -EINVAL;
  5430				goto failed_mount_wq;
  5431			}
  5432	
  5433			if (ext4_has_feature_ea_inode(sb)) {
  5434				sbi->s_ea_inode_cache = ext4_xattr_create_cache();
  5435				if (!sbi->s_ea_inode_cache) {
  5436					ext4_msg(sb, KERN_ERR,
  5437						 "Failed to create ea_inode_cache");
  5438					err = -EINVAL;
  5439					goto failed_mount_wq;
  5440				}
  5441			}
  5442		}
  5443	
  5444		/*
  5445		 * Get the # of file system overhead blocks from the
  5446		 * superblock if present.
  5447		 */
  5448		sbi->s_overhead = le32_to_cpu(es->s_overhead_clusters);
  5449		/* ignore the precalculated value if it is ridiculous */
  5450		if (sbi->s_overhead > ext4_blocks_count(es))
  5451			sbi->s_overhead = 0;
  5452		/*
  5453		 * If the bigalloc feature is not enabled recalculating the
  5454		 * overhead doesn't take long, so we might as well just redo
  5455		 * it to make sure we are using the correct value.
  5456		 */
  5457		if (!ext4_has_feature_bigalloc(sb))
  5458			sbi->s_overhead = 0;
  5459		if (sbi->s_overhead == 0) {
  5460			err = ext4_calculate_overhead(sb);
  5461			if (err)
  5462				goto failed_mount_wq;
  5463		}
  5464	
  5465		/*
  5466		 * The maximum number of concurrent works can be high and
  5467		 * concurrency isn't really necessary.  Limit it to 1.
  5468		 */
  5469		EXT4_SB(sb)->rsv_conversion_wq =
  5470			alloc_workqueue("ext4-rsv-conversion", WQ_MEM_RECLAIM | WQ_UNBOUND, 1);
  5471		if (!EXT4_SB(sb)->rsv_conversion_wq) {
  5472			printk(KERN_ERR "EXT4-fs: failed to create workqueue\n");
  5473			err = -ENOMEM;
  5474			goto failed_mount4;
  5475		}
  5476	
  5477		/*
  5478		 * The jbd2_journal_load will have done any necessary log recovery,
  5479		 * so we can safely mount the rest of the filesystem now.
  5480		 */
  5481	
  5482		root = ext4_iget(sb, EXT4_ROOT_INO, EXT4_IGET_SPECIAL);
  5483		if (IS_ERR(root)) {
  5484			ext4_msg(sb, KERN_ERR, "get root inode failed");
  5485			err = PTR_ERR(root);
  5486			root = NULL;
  5487			goto failed_mount4;
  5488		}
  5489		if (!S_ISDIR(root->i_mode) || !root->i_blocks || !root->i_size) {
  5490			ext4_msg(sb, KERN_ERR, "corrupt root inode, run e2fsck");
  5491			iput(root);
  5492			err = -EFSCORRUPTED;
  5493			goto failed_mount4;
  5494		}
  5495	
> 5496		if (sb->s_encoding)
  5497			sb->s_d_op = &generic_ci_dentry_ops;
  5498	
  5499		sb->s_root = d_make_root(root);
  5500		if (!sb->s_root) {
  5501			ext4_msg(sb, KERN_ERR, "get root dentry failed");
  5502			err = -ENOMEM;
  5503			goto failed_mount4;
  5504		}
  5505	
  5506		err = ext4_setup_super(sb, es, sb_rdonly(sb));
  5507		if (err == -EROFS) {
  5508			sb->s_flags |= SB_RDONLY;
  5509		} else if (err)
  5510			goto failed_mount4a;
  5511	
  5512		ext4_set_resv_clusters(sb);
  5513	
  5514		if (test_opt(sb, BLOCK_VALIDITY)) {
  5515			err = ext4_setup_system_zone(sb);
  5516			if (err) {
  5517				ext4_msg(sb, KERN_ERR, "failed to initialize system "
  5518					 "zone (%d)", err);
  5519				goto failed_mount4a;
  5520			}
  5521		}
  5522		ext4_fc_replay_cleanup(sb);
  5523	
  5524		ext4_ext_init(sb);
  5525	
  5526		/*
  5527		 * Enable optimize_scan if number of groups is > threshold. This can be
  5528		 * turned off by passing "mb_optimize_scan=0". This can also be
  5529		 * turned on forcefully by passing "mb_optimize_scan=1".
  5530		 */
  5531		if (!(ctx->spec & EXT4_SPEC_mb_optimize_scan)) {
  5532			if (sbi->s_groups_count >= MB_DEFAULT_LINEAR_SCAN_THRESHOLD)
  5533				set_opt2(sb, MB_OPTIMIZE_SCAN);
  5534			else
  5535				clear_opt2(sb, MB_OPTIMIZE_SCAN);
  5536		}
  5537	
  5538		err = ext4_mb_init(sb);
  5539		if (err) {
  5540			ext4_msg(sb, KERN_ERR, "failed to initialize mballoc (%d)",
  5541				 err);
  5542			goto failed_mount5;
  5543		}
  5544	
  5545		/*
  5546		 * We can only set up the journal commit callback once
  5547		 * mballoc is initialized
  5548		 */
  5549		if (sbi->s_journal)
  5550			sbi->s_journal->j_commit_callback =
  5551				ext4_journal_commit_callback;
  5552	
  5553		err = ext4_percpu_param_init(sbi);
  5554		if (err)
  5555			goto failed_mount6;
  5556	
  5557		if (ext4_has_feature_flex_bg(sb))
  5558			if (!ext4_fill_flex_info(sb)) {
  5559				ext4_msg(sb, KERN_ERR,
  5560				       "unable to initialize "
  5561				       "flex_bg meta info!");
  5562				err = -ENOMEM;
  5563				goto failed_mount6;
  5564			}
  5565	
  5566		err = ext4_register_li_request(sb, first_not_zeroed);
  5567		if (err)
  5568			goto failed_mount6;
  5569	
  5570		err = ext4_register_sysfs(sb);
  5571		if (err)
  5572			goto failed_mount7;
  5573	
  5574		err = ext4_init_orphan_info(sb);
  5575		if (err)
  5576			goto failed_mount8;
  5577	#ifdef CONFIG_QUOTA
  5578		/* Enable quota usage during mount. */
  5579		if (ext4_has_feature_quota(sb) && !sb_rdonly(sb)) {
  5580			err = ext4_enable_quotas(sb);
  5581			if (err)
  5582				goto failed_mount9;
  5583		}
  5584	#endif  /* CONFIG_QUOTA */
  5585	
  5586		/*
  5587		 * Save the original bdev mapping's wb_err value which could be
  5588		 * used to detect the metadata async write error.
  5589		 */
  5590		spin_lock_init(&sbi->s_bdev_wb_lock);
  5591		errseq_check_and_advance(&sb->s_bdev->bd_inode->i_mapping->wb_err,
  5592					 &sbi->s_bdev_wb_err);
  5593		EXT4_SB(sb)->s_mount_state |= EXT4_ORPHAN_FS;
  5594		ext4_orphan_cleanup(sb, es);
  5595		EXT4_SB(sb)->s_mount_state &= ~EXT4_ORPHAN_FS;
  5596		/*
  5597		 * Update the checksum after updating free space/inode counters and
  5598		 * ext4_orphan_cleanup. Otherwise the superblock can have an incorrect
  5599		 * checksum in the buffer cache until it is written out and
  5600		 * e2fsprogs programs trying to open a file system immediately
  5601		 * after it is mounted can fail.
  5602		 */
  5603		ext4_superblock_csum_set(sb);
  5604		if (needs_recovery) {
  5605			ext4_msg(sb, KERN_INFO, "recovery complete");
  5606			err = ext4_mark_recovery_complete(sb, es);
  5607			if (err)
  5608				goto failed_mount10;
  5609		}
  5610	
  5611		if (test_opt(sb, DISCARD) && !bdev_max_discard_sectors(sb->s_bdev))
  5612			ext4_msg(sb, KERN_WARNING,
  5613				 "mounting with \"discard\" option, but the device does not support discard");
  5614	
  5615		if (es->s_error_count)
  5616			mod_timer(&sbi->s_err_report, jiffies + 300*HZ); /* 5 minutes */
  5617	
  5618		/* Enable message ratelimiting. Default is 10 messages per 5 secs. */
  5619		ratelimit_state_init(&sbi->s_err_ratelimit_state, 5 * HZ, 10);
  5620		ratelimit_state_init(&sbi->s_warning_ratelimit_state, 5 * HZ, 10);
  5621		ratelimit_state_init(&sbi->s_msg_ratelimit_state, 5 * HZ, 10);
  5622		atomic_set(&sbi->s_warning_count, 0);
  5623		atomic_set(&sbi->s_msg_count, 0);
  5624	
  5625		return 0;
  5626	
  5627	failed_mount10:
  5628		ext4_quotas_off(sb, EXT4_MAXQUOTAS);
  5629	failed_mount9: __maybe_unused
  5630		ext4_release_orphan_info(sb);
  5631	failed_mount8:
  5632		ext4_unregister_sysfs(sb);
  5633		kobject_put(&sbi->s_kobj);
  5634	failed_mount7:
  5635		ext4_unregister_li_request(sb);
  5636	failed_mount6:
  5637		ext4_mb_release(sb);
  5638		ext4_flex_groups_free(sbi);
  5639		ext4_percpu_param_destroy(sbi);
  5640	failed_mount5:
  5641		ext4_ext_release(sb);
  5642		ext4_release_system_zone(sb);
  5643	failed_mount4a:
  5644		dput(sb->s_root);
  5645		sb->s_root = NULL;
  5646	failed_mount4:
  5647		ext4_msg(sb, KERN_ERR, "mount failed");
  5648		if (EXT4_SB(sb)->rsv_conversion_wq)
  5649			destroy_workqueue(EXT4_SB(sb)->rsv_conversion_wq);
  5650	failed_mount_wq:
  5651		ext4_xattr_destroy_cache(sbi->s_ea_inode_cache);
  5652		sbi->s_ea_inode_cache = NULL;
  5653	
  5654		ext4_xattr_destroy_cache(sbi->s_ea_block_cache);
  5655		sbi->s_ea_block_cache = NULL;
  5656	
  5657		if (sbi->s_journal) {
  5658			/* flush s_sb_upd_work before journal destroy. */
  5659			flush_work(&sbi->s_sb_upd_work);
  5660			jbd2_journal_destroy(sbi->s_journal);
  5661			sbi->s_journal = NULL;
  5662		}
  5663	failed_mount3a:
  5664		ext4_es_unregister_shrinker(sbi);
  5665	failed_mount3:
  5666		/* flush s_sb_upd_work before sbi destroy */
  5667		flush_work(&sbi->s_sb_upd_work);
  5668		del_timer_sync(&sbi->s_err_report);
  5669		ext4_stop_mmpd(sbi);
  5670		ext4_group_desc_free(sbi);
  5671	failed_mount:
  5672		if (sbi->s_chksum_driver)
  5673			crypto_free_shash(sbi->s_chksum_driver);
  5674	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 6/8] f2fs: Set the case-insensitive dentry operations through ->s_d_op

[Gabriel Krisman Bertazi]

All dentries in a case-insensitive filesystem have the same set of
dentry operations.  Therefore, we should let VFS propagate them from
sb->s_d_op d_alloc instead of setting at lookup time.

This was already the case before commit bb9cd9106b22 ("fscrypt: Have
filesystems handle their d_ops"), but it was changed to set at
lookup-time to facilitate the integration with fscrypt.  But it's a
problem because dentries that don't get created through ->lookup() won't
have any visibility of the operations.  Let's revert to the previous
implementation.

Suggested-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/f2fs/namei.c | 6 +++++-
 fs/f2fs/super.c | 3 +++
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c
index d0053b0284d8..4053846e2cd3 100644
--- a/fs/f2fs/namei.c
+++ b/fs/f2fs/namei.c
@@ -532,7 +532,11 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry,
 	}
 
 	err = f2fs_prepare_lookup(dir, dentry, &fname);
-	generic_set_encrypted_ci_d_ops(dentry);
+
+	/* Case-insensitive volumes set dentry ops through sb->s_d_op. */
+	if (!dir->i_sb->s_encoding)
+		generic_set_encrypted_ci_d_ops(dentry);
+
 	if (err == -ENOENT)
 		goto out_splice;
 	if (err)
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 033af907c3b1..0a199eb5b01e 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -4663,6 +4663,9 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 		goto free_node_inode;
 	}
 
+	if (sb->s_encoding)
+		sb->s_d_op = &generic_ci_dentry_ops;
+
 	sb->s_root = d_make_root(root); /* allocate root dentry */
 	if (!sb->s_root) {
 		err = -ENOMEM;
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Re: [f2fs-dev] [PATCH 6/8] f2fs: Set the case-insensitive dentry operations through ->s_d_op

[kernel test robot]

Hi Gabriel,

kernel test robot noticed the following build errors:

[auto build test ERROR on jaegeuk-f2fs/dev-test]
[also build test ERROR on jaegeuk-f2fs/dev tytso-ext4/dev linus/master v6.7-rc5 next-20231214]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Gabriel-Krisman-Bertazi/dcache-Add-helper-to-disable-d_revalidate-for-a-specific-dentry/20231214-074322
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git dev-test
patch link:    https://lore.kernel.org/r/20231213234031.1081-7-krisman%40suse.de
patch subject: [PATCH 6/8] f2fs: Set the case-insensitive dentry operations through ->s_d_op
config: x86_64-buildonly-randconfig-002-20231214 (https://download.01.org/0day-ci/archive/20231214/202312142238.71N8cNUB-lkp@intel.com/config)
compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231214/202312142238.71N8cNUB-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202312142238.71N8cNUB-lkp@intel.com/

All errors (new ones prefixed by >>):

   fs/f2fs/super.c: In function 'f2fs_fill_super':
>> fs/f2fs/super.c:4669:15: error: 'struct super_block' has no member named 's_encoding'
    4669 |         if (sb->s_encoding)
         |               ^~


vim +4669 fs/f2fs/super.c

  4491	
  4492		sb->s_op = &f2fs_sops;
  4493	#ifdef CONFIG_FS_ENCRYPTION
  4494		sb->s_cop = &f2fs_cryptops;
  4495	#endif
  4496	#ifdef CONFIG_FS_VERITY
  4497		sb->s_vop = &f2fs_verityops;
  4498	#endif
  4499		sb->s_xattr = f2fs_xattr_handlers;
  4500		sb->s_export_op = &f2fs_export_ops;
  4501		sb->s_magic = F2FS_SUPER_MAGIC;
  4502		sb->s_time_gran = 1;
  4503		sb->s_flags = (sb->s_flags & ~SB_POSIXACL) |
  4504			(test_opt(sbi, POSIX_ACL) ? SB_POSIXACL : 0);
  4505		memcpy(&sb->s_uuid, raw_super->uuid, sizeof(raw_super->uuid));
  4506		sb->s_iflags |= SB_I_CGROUPWB;
  4507	
  4508		/* init f2fs-specific super block info */
  4509		sbi->valid_super_block = valid_super_block;
  4510	
  4511		/* disallow all the data/node/meta page writes */
  4512		set_sbi_flag(sbi, SBI_POR_DOING);
  4513	
  4514		err = f2fs_init_write_merge_io(sbi);
  4515		if (err)
  4516			goto free_bio_info;
  4517	
  4518		init_sb_info(sbi);
  4519	
  4520		err = f2fs_init_iostat(sbi);
  4521		if (err)
  4522			goto free_bio_info;
  4523	
  4524		err = init_percpu_info(sbi);
  4525		if (err)
  4526			goto free_iostat;
  4527	
  4528		if (F2FS_IO_ALIGNED(sbi)) {
  4529			sbi->write_io_dummy =
  4530				mempool_create_page_pool(2 * (F2FS_IO_SIZE(sbi) - 1), 0);
  4531			if (!sbi->write_io_dummy) {
  4532				err = -ENOMEM;
  4533				goto free_percpu;
  4534			}
  4535		}
  4536	
  4537		/* init per sbi slab cache */
  4538		err = f2fs_init_xattr_caches(sbi);
  4539		if (err)
  4540			goto free_io_dummy;
  4541		err = f2fs_init_page_array_cache(sbi);
  4542		if (err)
  4543			goto free_xattr_cache;
  4544	
  4545		/* get an inode for meta space */
  4546		sbi->meta_inode = f2fs_iget(sb, F2FS_META_INO(sbi));
  4547		if (IS_ERR(sbi->meta_inode)) {
  4548			f2fs_err(sbi, "Failed to read F2FS meta data inode");
  4549			err = PTR_ERR(sbi->meta_inode);
  4550			goto free_page_array_cache;
  4551		}
  4552	
  4553		err = f2fs_get_valid_checkpoint(sbi);
  4554		if (err) {
  4555			f2fs_err(sbi, "Failed to get valid F2FS checkpoint");
  4556			goto free_meta_inode;
  4557		}
  4558	
  4559		if (__is_set_ckpt_flags(F2FS_CKPT(sbi), CP_QUOTA_NEED_FSCK_FLAG))
  4560			set_sbi_flag(sbi, SBI_QUOTA_NEED_REPAIR);
  4561		if (__is_set_ckpt_flags(F2FS_CKPT(sbi), CP_DISABLED_QUICK_FLAG)) {
  4562			set_sbi_flag(sbi, SBI_CP_DISABLED_QUICK);
  4563			sbi->interval_time[DISABLE_TIME] = DEF_DISABLE_QUICK_INTERVAL;
  4564		}
  4565	
  4566		if (__is_set_ckpt_flags(F2FS_CKPT(sbi), CP_FSCK_FLAG))
  4567			set_sbi_flag(sbi, SBI_NEED_FSCK);
  4568	
  4569		/* Initialize device list */
  4570		err = f2fs_scan_devices(sbi);
  4571		if (err) {
  4572			f2fs_err(sbi, "Failed to find devices");
  4573			goto free_devices;
  4574		}
  4575	
  4576		err = f2fs_init_post_read_wq(sbi);
  4577		if (err) {
  4578			f2fs_err(sbi, "Failed to initialize post read workqueue");
  4579			goto free_devices;
  4580		}
  4581	
  4582		sbi->total_valid_node_count =
  4583					le32_to_cpu(sbi->ckpt->valid_node_count);
  4584		percpu_counter_set(&sbi->total_valid_inode_count,
  4585					le32_to_cpu(sbi->ckpt->valid_inode_count));
  4586		sbi->user_block_count = le64_to_cpu(sbi->ckpt->user_block_count);
  4587		sbi->total_valid_block_count =
  4588					le64_to_cpu(sbi->ckpt->valid_block_count);
  4589		sbi->last_valid_block_count = sbi->total_valid_block_count;
  4590		sbi->reserved_blocks = 0;
  4591		sbi->current_reserved_blocks = 0;
  4592		limit_reserve_root(sbi);
  4593		adjust_unusable_cap_perc(sbi);
  4594	
  4595		f2fs_init_extent_cache_info(sbi);
  4596	
  4597		f2fs_init_ino_entry_info(sbi);
  4598	
  4599		f2fs_init_fsync_node_info(sbi);
  4600	
  4601		/* setup checkpoint request control and start checkpoint issue thread */
  4602		f2fs_init_ckpt_req_control(sbi);
  4603		if (!f2fs_readonly(sb) && !test_opt(sbi, DISABLE_CHECKPOINT) &&
  4604				test_opt(sbi, MERGE_CHECKPOINT)) {
  4605			err = f2fs_start_ckpt_thread(sbi);
  4606			if (err) {
  4607				f2fs_err(sbi,
  4608				    "Failed to start F2FS issue_checkpoint_thread (%d)",
  4609				    err);
  4610				goto stop_ckpt_thread;
  4611			}
  4612		}
  4613	
  4614		/* setup f2fs internal modules */
  4615		err = f2fs_build_segment_manager(sbi);
  4616		if (err) {
  4617			f2fs_err(sbi, "Failed to initialize F2FS segment manager (%d)",
  4618				 err);
  4619			goto free_sm;
  4620		}
  4621		err = f2fs_build_node_manager(sbi);
  4622		if (err) {
  4623			f2fs_err(sbi, "Failed to initialize F2FS node manager (%d)",
  4624				 err);
  4625			goto free_nm;
  4626		}
  4627	
  4628		err = adjust_reserved_segment(sbi);
  4629		if (err)
  4630			goto free_nm;
  4631	
  4632		/* For write statistics */
  4633		sbi->sectors_written_start = f2fs_get_sectors_written(sbi);
  4634	
  4635		/* Read accumulated write IO statistics if exists */
  4636		seg_i = CURSEG_I(sbi, CURSEG_HOT_NODE);
  4637		if (__exist_node_summaries(sbi))
  4638			sbi->kbytes_written =
  4639				le64_to_cpu(seg_i->journal->info.kbytes_written);
  4640	
  4641		f2fs_build_gc_manager(sbi);
  4642	
  4643		err = f2fs_build_stats(sbi);
  4644		if (err)
  4645			goto free_nm;
  4646	
  4647		/* get an inode for node space */
  4648		sbi->node_inode = f2fs_iget(sb, F2FS_NODE_INO(sbi));
  4649		if (IS_ERR(sbi->node_inode)) {
  4650			f2fs_err(sbi, "Failed to read node inode");
  4651			err = PTR_ERR(sbi->node_inode);
  4652			goto free_stats;
  4653		}
  4654	
  4655		/* read root inode and dentry */
  4656		root = f2fs_iget(sb, F2FS_ROOT_INO(sbi));
  4657		if (IS_ERR(root)) {
  4658			f2fs_err(sbi, "Failed to read root inode");
  4659			err = PTR_ERR(root);
  4660			goto free_node_inode;
  4661		}
  4662		if (!S_ISDIR(root->i_mode) || !root->i_blocks ||
  4663				!root->i_size || !root->i_nlink) {
  4664			iput(root);
  4665			err = -EINVAL;
  4666			goto free_node_inode;
  4667		}
  4668	
> 4669		if (sb->s_encoding)
  4670			sb->s_d_op = &generic_ci_dentry_ops;
  4671	
  4672		sb->s_root = d_make_root(root); /* allocate root dentry */
  4673		if (!sb->s_root) {
  4674			err = -ENOMEM;
  4675			goto free_node_inode;
  4676		}
  4677	
  4678		err = f2fs_init_compress_inode(sbi);
  4679		if (err)
  4680			goto free_root_inode;
  4681	
  4682		err = f2fs_register_sysfs(sbi);
  4683		if (err)
  4684			goto free_compress_inode;
  4685	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 7/8] libfs: Don't support setting casefold operations during lookup

[Gabriel Krisman Bertazi]

No filesystems depend on it anymore, and it is generally a bad idea.
Since all dentries should have the same set of dentry operations in
case-insensitive filesystems, it should be configured through ->s_d_op.

Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/libfs.c | 15 ---------------
 1 file changed, 15 deletions(-)

diff --git a/fs/libfs.c b/fs/libfs.c
index b8ecada3a5b2..41c02c003265 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -1784,27 +1784,12 @@ static const struct dentry_operations generic_encrypted_dentry_ops = {
  * generic_set_encrypted_ci_d_ops - helper for setting d_ops for given dentry
  * @dentry:	dentry to set ops on
  *
- * Casefolded directories need d_hash and d_compare set, so that the dentries
- * contained in them are handled case-insensitively.  Note that these operations
- * are needed on the parent directory rather than on the dentries in it, and
- * while the casefolding flag can be toggled on and off on an empty directory,
- * dentry_operations can't be changed later.  As a result, if the filesystem has
- * casefolding support enabled at all, we have to give all dentries the
- * casefolding operations even if their inode doesn't have the casefolding flag
- * currently (and thus the casefolding ops would be no-ops for now).
- *
  * Encryption works differently in that the only dentry operation it needs is
  * d_revalidate, which it only needs on dentries that have the no-key name flag.
  * The no-key flag can't be set "later", so we don't have to worry about that.
  */
 void generic_set_encrypted_ci_d_ops(struct dentry *dentry)
 {
-#if IS_ENABLED(CONFIG_UNICODE)
-	if (dentry->d_sb->s_encoding) {
-		d_set_d_op(dentry, &generic_ci_dentry_ops);
-		return;
-	}
-#endif
 #ifdef CONFIG_FS_ENCRYPTION
 	if (dentry->d_flags & DCACHE_NOKEY_NAME) {
 		d_set_d_op(dentry, &generic_encrypted_dentry_ops);
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[f2fs-dev] [PATCH 8/8] fscrypt: Move d_revalidate configuration back into fscrypt

[Gabriel Krisman Bertazi]

This partially reverts commit bb9cd9106b22 ("fscrypt: Have filesystems
handle their d_ops"), which moved this handler out of fscrypt and into
the filesystems, in preparation to support casefold and fscrypt
combinations.  Now that we set casefolding operations through
->s_d_op, move this back into fscrypt, where it belongs, but take care
to handle filesystems that set their own sb->s_d_op.

Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/crypto/hooks.c       |  8 ++++++++
 fs/ext4/namei.c         |  5 -----
 fs/f2fs/namei.c         |  5 -----
 fs/libfs.c              | 19 -------------------
 fs/ubifs/dir.c          |  1 -
 include/linux/fs.h      |  1 -
 include/linux/fscrypt.h | 10 +++++-----
 7 files changed, 13 insertions(+), 36 deletions(-)

diff --git a/fs/crypto/hooks.c b/fs/crypto/hooks.c
index 52504dd478d3..166837d5af29 100644
--- a/fs/crypto/hooks.c
+++ b/fs/crypto/hooks.c
@@ -94,6 +94,10 @@ int __fscrypt_prepare_rename(struct inode *old_dir, struct dentry *old_dentry,
 }
 EXPORT_SYMBOL_GPL(__fscrypt_prepare_rename);
 
+static const struct dentry_operations fscrypt_dentry_ops = {
+	.d_revalidate = fscrypt_d_revalidate,
+};
+
 int __fscrypt_prepare_lookup(struct inode *dir, struct dentry *dentry,
 			     struct fscrypt_name *fname)
 {
@@ -106,6 +110,10 @@ int __fscrypt_prepare_lookup(struct inode *dir, struct dentry *dentry,
 		spin_lock(&dentry->d_lock);
 		dentry->d_flags |= DCACHE_NOKEY_NAME;
 		spin_unlock(&dentry->d_lock);
+
+		/* Give preference to the filesystem hooks, if any. */
+		if (!dentry->d_op)
+			d_set_d_op(dentry, &fscrypt_dentry_ops);
 	}
 	return err;
 }
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 3c1208d5d85b..3f0b853a371e 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1762,11 +1762,6 @@ static struct buffer_head *ext4_lookup_entry(struct inode *dir,
 	struct buffer_head *bh;
 
 	err = ext4_fname_prepare_lookup(dir, dentry, &fname);
-
-	/* Case-insensitive volumes set dentry ops through sb->s_d_op. */
-	if (!dir->i_sb->s_encoding)
-		generic_set_encrypted_ci_d_ops(dentry);
-
 	if (err == -ENOENT)
 		return NULL;
 	if (err)
diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c
index 4053846e2cd3..b40c6c393bd6 100644
--- a/fs/f2fs/namei.c
+++ b/fs/f2fs/namei.c
@@ -532,11 +532,6 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry,
 	}
 
 	err = f2fs_prepare_lookup(dir, dentry, &fname);
-
-	/* Case-insensitive volumes set dentry ops through sb->s_d_op. */
-	if (!dir->i_sb->s_encoding)
-		generic_set_encrypted_ci_d_ops(dentry);
-
 	if (err == -ENOENT)
 		goto out_splice;
 	if (err)
diff --git a/fs/libfs.c b/fs/libfs.c
index 41c02c003265..4fed170dfe49 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -1780,25 +1780,6 @@ static const struct dentry_operations generic_encrypted_dentry_ops = {
 };
 #endif
 
-/**
- * generic_set_encrypted_ci_d_ops - helper for setting d_ops for given dentry
- * @dentry:	dentry to set ops on
- *
- * Encryption works differently in that the only dentry operation it needs is
- * d_revalidate, which it only needs on dentries that have the no-key name flag.
- * The no-key flag can't be set "later", so we don't have to worry about that.
- */
-void generic_set_encrypted_ci_d_ops(struct dentry *dentry)
-{
-#ifdef CONFIG_FS_ENCRYPTION
-	if (dentry->d_flags & DCACHE_NOKEY_NAME) {
-		d_set_d_op(dentry, &generic_encrypted_dentry_ops);
-		return;
-	}
-#endif
-}
-EXPORT_SYMBOL(generic_set_encrypted_ci_d_ops);
-
 /**
  * inode_maybe_inc_iversion - increments i_version
  * @inode: inode with the i_version that should be updated
diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c
index 3b13c648d490..51b9a10a9851 100644
--- a/fs/ubifs/dir.c
+++ b/fs/ubifs/dir.c
@@ -205,7 +205,6 @@ static struct dentry *ubifs_lookup(struct inode *dir, struct dentry *dentry,
 	dbg_gen("'%pd' in dir ino %lu", dentry, dir->i_ino);
 
 	err = fscrypt_prepare_lookup(dir, dentry, &nm);
-	generic_set_encrypted_ci_d_ops(dentry);
 	if (err == -ENOENT)
 		return d_splice_alias(NULL, dentry);
 	if (err)
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 887a27d07f96..e5ae21f9f637 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -3201,7 +3201,6 @@ extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
 
 extern int generic_check_addressable(unsigned, u64);
 
-extern void generic_set_encrypted_ci_d_ops(struct dentry *dentry);
 extern const struct dentry_operations generic_ci_dentry_ops;
 
 int may_setattr(struct mnt_idmap *idmap, struct inode *inode,
diff --git a/include/linux/fscrypt.h b/include/linux/fscrypt.h
index 12f9e455d569..97a11280c2bd 100644
--- a/include/linux/fscrypt.h
+++ b/include/linux/fscrypt.h
@@ -961,11 +961,11 @@ static inline int fscrypt_prepare_rename(struct inode *old_dir,
  * key is available, then the lookup is assumed to be by plaintext name;
  * otherwise, it is assumed to be by no-key name.
  *
- * This will set DCACHE_NOKEY_NAME on the dentry if the lookup is by no-key
- * name.  In this case the filesystem must assign the dentry a dentry_operations
- * which contains fscrypt_d_revalidate (or contains a d_revalidate method that
- * calls fscrypt_d_revalidate), so that the dentry will be invalidated if the
- * directory's encryption key is later added.
+ * This also optionally installs a custom ->d_revalidate() method which will
+ * invalidate the dentry if it was created without the key and the key is later
+ * added.  If the filesystem provides its own ->d_op hooks, they will be used
+ * instead, but then the filesystem must make sure to call fscrypt_d_revalidate
+ * in its d_revalidate hook, to check if fscrypt considers the dentry stale.
  *
  * Return: 0 on success; -ENOENT if the directory's key is unavailable but the
  * filename isn't a valid no-key name, so a negative dentry should be created;
-- 
2.43.0



_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Re: [f2fs-dev] [PATCH 8/8] fscrypt: Move d_revalidate configuration back into fscrypt

[kernel test robot]

Hi Gabriel,

kernel test robot noticed the following build warnings:

[auto build test WARNING on jaegeuk-f2fs/dev-test]
[also build test WARNING on jaegeuk-f2fs/dev tytso-ext4/dev linus/master v6.7-rc5 next-20231214]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Gabriel-Krisman-Bertazi/dcache-Add-helper-to-disable-d_revalidate-for-a-specific-dentry/20231214-074322
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git dev-test
patch link:    https://lore.kernel.org/r/20231213234031.1081-9-krisman%40suse.de
patch subject: [PATCH 8/8] fscrypt: Move d_revalidate configuration back into fscrypt
config: x86_64-randconfig-123-20231214 (https://download.01.org/0day-ci/archive/20231214/202312142213.uyrNJniX-lkp@intel.com/config)
compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231214/202312142213.uyrNJniX-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202312142213.uyrNJniX-lkp@intel.com/

All warnings (new ones prefixed by >>):

>> fs/libfs.c:1778:39: warning: unused variable 'generic_encrypted_dentry_ops' [-Wunused-const-variable]
   static const struct dentry_operations generic_encrypted_dentry_ops = {
                                         ^
   1 warning generated.


vim +/generic_encrypted_dentry_ops +1778 fs/libfs.c

608af703519a58 Daniel Rosenberg 2020-11-19  1776  
608af703519a58 Daniel Rosenberg 2020-11-19  1777  #ifdef CONFIG_FS_ENCRYPTION
608af703519a58 Daniel Rosenberg 2020-11-19 @1778  static const struct dentry_operations generic_encrypted_dentry_ops = {
608af703519a58 Daniel Rosenberg 2020-11-19  1779  	.d_revalidate = fscrypt_d_revalidate,
608af703519a58 Daniel Rosenberg 2020-11-19  1780  };
608af703519a58 Daniel Rosenberg 2020-11-19  1781  #endif
608af703519a58 Daniel Rosenberg 2020-11-19  1782  

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel