test_dummy_encryption fixes and cleanups
Subsystems (4)
Eric Biggers
maintainerv3 cover test_dummy_encryption fixes and cleanups
This series cleans up and fixes the way that ext4 and f2fs handle the test_dummy_encryption mount option.
Some patches from v2 were already applied to the fscrypt and f2fs trees. This series just includes the remaining patches. Patches 1-2 can be applied to the ext4 tree now. The remaining patches will need to wait until their prerequisites get merged via the fscrypt tree.
Changed from v2 (besides the omitted patches as mentioned above):
- Split the parse_apply_sb_mount_options() fix into its own patch
- Fixed a couple bugs in
"ext4: fix up test_dummy_encryption handling for new mount API"
Eric Biggers (5):
ext4: fix memory leak in parse_apply_sb_mount_options()
ext4: only allow test_dummy_encryption when supported
ext4: fix up test_dummy_encryption handling for new mount API
f2fs: use the updated test_dummy_encryption helper functions
fscrypt: remove fscrypt_set_test_dummy_encryption()
fs/crypto/policy.c | 13 --- fs/ext4/ext4.h | 6 -- fs/ext4/super.c | 180 +++++++++++++++++++++++----------------- fs/f2fs/super.c | 29 +++++-- include/linux/fscrypt.h | 2 - 5 files changed, 124 insertions(+), 106 deletions(-)
Signature · 8 lines
2.36.1 _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Eric Biggers
maintainerPATCH v3 1/5 ext4: fix memory leak in parse_apply_sb_mount_options()
From: Eric Biggers <[email protected]>
If processing the on-disk mount options fails after any memory was allocated in the ext4_fs_context, e.g. s_qf_names, then this memory is leaked. Fix this by calling ext4_fc_free() instead of kfree() directly.
Reproducer:
mkfs.ext4 -F /dev/vdc
tune2fs /dev/vdc -E mount_opts=usrjquota=file
echo clear > /sys/kernel/debug/kmemleak
mount /dev/vdc /vdc
echo scan > /sys/kernel/debug/kmemleak
sleep 5
echo scan > /sys/kernel/debug/kmemleak
cat /sys/kernel/debug/kmemleak
--- fs/ext4/super.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
fs/ext4/super.c +4 -2
@@ -2625,8 +2625,10 @@ static int parse_apply_sb_mount_options(struct super_block *sb, ret =ext4_apply_options (fc, sb); out_free: -kfree (s_ctx); -kfree (fc); +if (fc) { +ext4_fc_free (fc); +kfree (fc); + }kfree (s_mount_opts);return ret; }
Signature · 8 lines
2.36.1 _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Re: [f2fs-dev] [PATCH v3 1/5] ext4: fix memory leak in parse_apply_sb_mount_options()
Quoted 5 lines · Eric Biggers
From: Eric Biggers <[email protected]>
If processing the on-disk mount options fails after any memory was allocated in the
ext4_fs_context, e.g.s_qf_names, then this memory is leaked. Fix this by callingext4_fc_free()instead ofkfree()directly.
Thanks for splitting the patch. It becomes an easy backport.
Quoted 11 lines
Reproducer:
mkfs.ext4 -F
/dev/vdc
tune2fs/dev/vdc-E mount_opts=usrjquota=file
echo clear >/sys/kernel/debug/kmemleak
mount/dev/vdc/vdc
echo scan >/sys/kernel/debug/kmemleak
sleep 5
echo scan >/sys/kernel/debug/kmemleak
cat/sys/kernel/debug/kmemleak
Tested this and as you mentioned this patch fixes the memory leak with s_qf_names in note_qf_name().
tune2fs 1.46.5 (30-Dec-2021)
Setting extended default mount options to 'usrjquota=file' unreferenced object 0xffff8881126b9a50 (size 8):
comm "mount", pid 1475, jiffies 4294829180 (age 48.670s) hex dump (first 8 bytes):
66 69 6c 65 00 6b 6b a5 file.kk.
backtrace:
[<ffffffff8153b09d>] __kmalloc_track_caller+0x17d/0x2f0
[<ffffffff8149b7e8>] kmemdup_nul+0x28/0x70
[<ffffffff81753a75>] note_qf_name.isra.0+0x95/0x180
[<ffffffff817548a8>] ext4_parse_param+0xd48/0x11c0
[<ffffffff8175a131>] ext4_fill_super+0x1cc1/0x6260
[<ffffffff8155edce>] get_tree_bdev+0x24e/0x3a0
[<ffffffff81740355>] ext4_get_tree+0x15/0x20
[<ffffffff8155d3a2>] vfs_get_tree+0x52/0x140
[<ffffffff815a2048>] path_mount+0x3f8/0xf30
[<ffffffff815a2c52>] do_mount+0xd2/0xf0
[<ffffffff815a2e4a>] __x64_sys_mount+0xca/0x110
[<ffffffff82e6674b>] do_syscall_64+0x3b/0x90
[<ffffffff8300007c>] entry_SYSCALL_64_after_hwframe+0x44/0xaeFeel free to add by -
-ritesh
Quoted 34 lines
Fixes 7edfd85b1ffd ("ext4: Completely separate options parsing and sb setup")Signed-off-by Eric Biggers <[email protected]>--- fs/ext4/super.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)fs/ext4/super.c +4
-3@@ -2625,8 +2625,10 @@ static int parse_apply_sb_mount_options(struct super_block *sb, ret =ext4_apply_options (fc, sb); out_free: -kfree (s_ctx); -kfree (fc); +if (fc) { +ext4_fc_free (fc); +kfree (fc); + }kfree (s_mount_opts);return ret; } --2.36.1
_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Re: [PATCH v3 1/5] ext4: fix memory leak in parse_apply_sb_mount_options()
In reply to Eric BiggersQuoted 17 lines · Eric Biggers
From: Eric Biggers <[email protected]>
If processing the on-disk mount options fails after any memory was allocated in the
ext4_fs_context, e.g.s_qf_names, then this memory is leaked. Fix this by callingext4_fc_free()instead ofkfree()directly.Reproducer:
mkfs.ext4 -F
/dev/vdc
tune2fs/dev/vdc-E mount_opts=usrjquota=file
echo clear >/sys/kernel/debug/kmemleak
mount/dev/vdc/vdc
echo scan >/sys/kernel/debug/kmemleak
sleep 5
echo scan >/sys/kernel/debug/kmemleak
cat/sys/kernel/debug/kmemleak
Quoted 3 lines
Fixes 7edfd85b1ffd ("ext4: Completely separate options parsing and sb setup")Signed-off-by Eric Biggers <[email protected]>
Thanks, applied.
- Ted
_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Eric Biggers
maintainerPATCH v3 2/5 ext4: only allow test_dummy_encryption when supported
In reply to Eric BiggersFrom: Eric Biggers <[email protected]>
Make the test_dummy_encryption mount option require that the encrypt feature flag be already enabled on the filesystem, rather than automatically enabling it. Practically, this means that "-O encrypt" will need to be included in MKFS_OPTIONS when running xfstests with the test_dummy_encryption mount option. (ext4/053 also needs an update.)
Moreover, as long as the preconditions for test_dummy_encryption are being tightened anyway, take the opportunity to start rejecting it when !CONFIG_FS_ENCRYPTION rather than ignoring it.
The motivation for requiring the encrypt feature flag is that:
- Having the filesystem auto-enable feature flags is problematic, as it bypasses the usual sanity checks. The specific issue which came up recently is that in kernel versions where ext4 supports casefold but not encrypt+casefold (v5.1 through v5.10), the kernel will happily add the encrypt flag to a filesystem that has the casefold flag, making it unmountable -- but only for subsequent mounts, not the initial one. This confused the casefold support detection in xfstests, causing generic/556 to fail rather than be skipped.
- The xfstests-bld test runners (kvm-xfstests et al.) already use the required mkfs flag, so they will not be affected by this change. Only users of test_dummy_encryption alone will be affected. But, this option has always been for testing only, so it should be fine to require that the few users of this option update their test scripts.
- f2fs already requires it (for its equivalent feature flag).
--- fs/ext4/super.c | 59 +++++++++++++++++++++++++++++++------------------ 1 file changed, 37 insertions(+), 22 deletions(-)
fs/ext4/super.c +37 -22
@@ -2427,11 +2427,12 @@ static int ext4_parse_param(struct fs_context *fc, struct fs_parameter *param) ctx->spec |=EXT4_SPEC_DUMMY_ENCRYPTION ; ctx->test_dummy_enc_arg =kmemdup_nul (param->string, param->size,GFP_KERNEL ); +return 0 ;#else ext4_msg (NULL ,KERN_WARNING , -"Test dummy encryption mount option ignored" ); +"test_dummy_encryption option not supported" ); +return -EINVAL ;#endif -return 0 ;case Opt_dax:case Opt_dax_type:#ifdef CONFIG_FS_DAX
@@ -2788,12 +2789,43 @@ static int ext4_check_quota_consistency(struct fs_context *fc,#endif } +static int ext4_check_test_dummy_encryption (const struct fs_context *fc, +struct super_block *sb) +{ +const struct ext4_fs_context *ctx = fc->fs_private; +const struct ext4_sb_info *sbi =EXT4_SB (sb); + +if (!IS_ENABLED (CONFIG_FS_ENCRYPTION ) || + !(ctx->spec &EXT4_SPEC_DUMMY_ENCRYPTION )) +return 0 ; + +if (!ext4_has_feature_encrypt (sb)) { +ext4_msg (NULL ,KERN_WARNING , +"test_dummy_encryption requires encrypt feature" ); +return -EINVAL ; + } +/* +* This mount option is just for testing, and it's not worthwhile to +* implement the extra complexity (e.g. RCU protection) that would be +* needed to allow it to be set or changed during remount. We do allow +* it to be specified during remount, but only if there is no change. +*/ +if (fc->purpose ==FS_CONTEXT_FOR_RECONFIGURE && + !DUMMY_ENCRYPTION_ENABLED (sbi)) { +ext4_msg (NULL ,KERN_WARNING , +"Can't set test_dummy_encryption on remount" ); +return -EINVAL ; + } +return 0 ; +} +static int ext4_check_opt_consistency (struct fs_context *fc,struct super_block *sb) {struct ext4_fs_context *ctx = fc->fs_private;struct ext4_sb_info *sbi = fc->s_fs_info;int is_remount = fc->purpose ==FS_CONTEXT_FOR_RECONFIGURE ; +int err;if ((ctx->opt_flags &MOPT_NO_EXT2 ) &&IS_EXT2_SB (sb)) { ext4_msg(NULL, KERN_ERR,
@@ -2823,20 +2855,9 @@ static int ext4_check_opt_consistency(struct fs_context *fc, "for blocksize < PAGE_SIZE"); } -#ifdef CONFIG_FS_ENCRYPTION -/* -* This mount option is just for testing, and it's not worthwhile to -* implement the extra complexity (e.g. RCU protection) that would be -* needed to allow it to be set or changed during remount. We do allow -* it to be specified during remount, but only if there is no change. -*/ -if ((ctx->spec &EXT4_SPEC_DUMMY_ENCRYPTION ) && - is_remount && !sbi->s_dummy_enc_policy.policy) { -ext4_msg (NULL ,KERN_WARNING , -"Can't set test_dummy_encryption on remount" ); -return -1 ; - } -#endif + err =ext4_check_test_dummy_encryption (fc, sb); +if (err) +return err;if ((ctx->spec &EXT4_SPEC_DATAJ ) && is_remount) {if (!sbi->s_journal) {
@@ -5281,12 +5302,6 @@ static int __ext4_fill_super(struct fs_context *fc, struct super_block *sb) goto failed_mount_wq; } -if (DUMMY_ENCRYPTION_ENABLED (sbi) && !sb_rdonly (sb) && - !ext4_has_feature_encrypt (sb)) { -ext4_set_feature_encrypt (sb); -ext4_commit_super (sb); - } -/* * Get the # of file system overhead blocks from the * superblock if present.
Signature · 8 lines
2.36.1 _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Re: [PATCH v3 2/5] ext4: only allow test_dummy_encryption when supported
Quoted 29 lines · Eric Biggers
From: Eric Biggers <[email protected]>
Make the
test_dummy_encryptionmount option require that the encrypt feature flag be already enabled on the filesystem, rather than automatically enabling it. Practically, this means that "-O encrypt" will need to be included inMKFS_OPTIONSwhen running xfstests with thetest_dummy_encryptionmount option. (ext4/053 also needs an update.)Moreover, as long as the preconditions for
test_dummy_encryptionare being tightened anyway, take the opportunity to start rejecting it when !CONFIG_FS_ENCRYPTION rather than ignoring it.The motivation for requiring the encrypt feature flag is that:
- Having the filesystem auto-enable feature flags is problematic, as it bypasses the usual sanity checks. The specific issue which came up recently is that in kernel versions where ext4 supports casefold but not encrypt+casefold (v5.1 through v5.10), the kernel will happily add the encrypt flag to a filesystem that has the casefold flag, making it unmountable -- but only for subsequent mounts, not the initial one. This confused the casefold support detection in xfstests, causing generic/556 to fail rather than be skipped.
- The xfstests-bld test runners (kvm-xfstests et al.) already use the required mkfs flag, so they will not be affected by this change. Only users of
test_dummy_encryptionalone will be affected. But, this option has always been for testing only, so it should be fine to require that the few users of this option update their test scripts.
Quoted 3 lines
- f2fs already requires it (for its equivalent feature flag).
Signed-off-by Eric Biggers <[email protected]>
Thanks, applied.
- Ted
Eric Biggers
maintainerPATCH v3 3/5 ext4: fix up test_dummy_encryption handling for new mount API
In reply to Eric BiggersFrom: Eric Biggers <[email protected]>
Since ext4 was converted to the new mount API, the test_dummy_encryption mount option isn't being handled entirely correctly, because the needed fscrypt_set_test_dummy_encryption() helper function combines parsing/checking/applying into one function. That doesn't work well with the new mount API, which split these into separate steps.
This was sort of okay anyway, due to the parsing logic that was copied from fscrypt_set_test_dummy_encryption() into ext4_parse_param(), combined with an additional check in ext4_check_test_dummy_encryption(). However, these overlooked the case of changing the value of test_dummy_encryption on remount, which isn't allowed but ext4 wasn't detecting until ext4_apply_options() when it's too late to fail. Another bug is that if test_dummy_encryption was specified multiple times with an argument, memory was leaked.
Fix this up properly by using the new helper functions that allow splitting up the parse/check/apply steps for test_dummy_encryption.
--- fs/ext4/ext4.h | 6 --- fs/ext4/super.c | 133 ++++++++++++++++++++++++++---------------------- 2 files changed, 71 insertions(+), 68 deletions(-)
fs/ext4/ext4.h +0 -6
@@ -1440,12 +1440,6 @@ struct ext4_super_block {#ifdef __KERNEL__ -#ifdef CONFIG_FS_ENCRYPTION -#define DUMMY_ENCRYPTION_ENABLED (sbi) ((sbi)->s_dummy_enc_policy.policy != NULL) -#else -#define DUMMY_ENCRYPTION_ENABLED (sbi) (0) -#endif -/* Number of quota types we support */ #define EXT4_MAXQUOTAS 3
fs/ext4/super.c +71 -62
218 lines not on this page · raw message
Signature · 8 lines
2.36.1 _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Eric Biggers
maintainerPATCH v3 4/5 f2fs: use the updated test_dummy_encryption helper functions
In reply to Eric BiggersFrom: Eric Biggers <[email protected]>
Switch f2fs over to the functions that are replacing fscrypt_set_test_dummy_encryption(). Since f2fs hasn't been converted to the new mount API yet, this doesn't really provide a benefit for f2fs. But it allows fscrypt_set_test_dummy_encryption() to be removed.
Also take the opportunity to eliminate an #ifdef.
--- fs/f2fs/super.c | 29 +++++++++++++++++++++-------- 1 file changed, 21 insertions(+), 8 deletions(-)
fs/f2fs/super.c +21 -8
@@ -8,6 +8,7 @@#include <linux/module.h> #include <linux/init.h> #include <linux/fs.h> +#include <linux/fs_context.h> #include <linux/sched/mm.h> #include <linux/statfs.h> #include <linux/buffer_head.h>
@@ -492,9 +493,19 @@ static int f2fs_set_test_dummy_encryption(struct super_block *sb, bool is_remount) {struct f2fs_sb_info *sbi =F2FS_SB (sb); -#ifdef CONFIG_FS_ENCRYPTION +struct fs_parameter param = { + .type = fs_value_is_string, + .string = arg->from ? arg->from :"" , + }; +struct fscrypt_dummy_policy *policy = + &F2FS_OPTION (sbi).dummy_enc_policy;int err; +if (!IS_ENABLED (CONFIG_FS_ENCRYPTION )) { +f2fs_warn (sbi,"test_dummy_encryption option not supported" ); +return -EINVAL ; + } +if (!f2fs_sb_has_encrypt (sbi)) {f2fs_err (sbi,"Encrypt feature is off" );return -EINVAL ;
@@ -506,12 +517,12 @@ static int f2fs_set_test_dummy_encryption(struct super_block *sb,* needed to allow it to be set or changed during remount. We do allow * it to be specified during remount, but only if there is no change. */ -if (is_remount && !F2FS_OPTION (sbi).dummy_enc_policy.policy) { +if (is_remount && !fscrypt_is_dummy_policy_set (policy)) {f2fs_warn (sbi,"Can't set test_dummy_encryption on remount" );return -EINVAL ; } - err =fscrypt_set_test_dummy_encryption ( - sb, arg->from, &F2FS_OPTION (sbi).dummy_enc_policy); + + err =fscrypt_parse_test_dummy_encryption (¶m, policy);if (err) {if (err == -EEXIST ) f2fs_warn(sbi,
@@ -524,12 +535,14 @@ static int f2fs_set_test_dummy_encryption(struct super_block *sb, opt, err);return -EINVAL ; } + err =fscrypt_add_test_dummy_key (sb, policy); +if (err) { +f2fs_warn (sbi,"Error adding test dummy encryption key [%d]" , + err); +return err; + }f2fs_warn (sbi,"Test dummy encryption mode enabled" );return 0 ; -#else -f2fs_warn (sbi,"test_dummy_encryption option not supported" ); -return -EINVAL ; -#endif }#ifdef CONFIG_F2FS_FS_COMPRESSION
Signature · 8 lines
2.36.1 _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Re: [f2fs-dev] [PATCH v3 4/5] f2fs: use the updated test_dummy_encryption helper functions
Quoted 10 lines · Eric Biggers
From: Eric Biggers <[email protected]>
Switch f2fs over to the functions that are replacing
fscrypt_set_test_dummy_encryption(). Since f2fs hasn't been converted to the new mount API yet, this doesn't really provide a benefit for f2fs. But it allowsfscrypt_set_test_dummy_encryption()to be removed.Also take the opportunity to eliminate an #ifdef.
Signed-off-by Eric Biggers <[email protected]>
Thanks,
Eric Biggers
maintainerPATCH v3 5/5 fscrypt: remove fscrypt_set_test_dummy_encryption()
In reply to Eric BiggersFrom: Eric Biggers <[email protected]>
Now that all its callers have been converted to fscrypt_parse_test_dummy_encryption() and fscrypt_add_test_dummy_key() instead, fscrypt_set_test_dummy_encryption() can be removed.
--- fs/crypto/policy.c | 13 ------------- include/linux/fscrypt.h | 2 -- 2 files changed, 15 deletions(-)
fs/crypto/policy.c +0 -13
@@ -802,19 +802,6 @@ bool fscrypt_dummy_policies_equal(const struct fscrypt_dummy_policy *p1, }EXPORT_SYMBOL_GPL (fscrypt_dummy_policies_equal); -/* Deprecated, do not use */ -int fscrypt_set_test_dummy_encryption (struct super_block *sb,const char *arg, -struct fscrypt_dummy_policy *dummy_policy) -{ -struct fs_parameter param = { - .type = fs_value_is_string, - .string = arg ? (char *)arg :"" , - }; -return fscrypt_parse_test_dummy_encryption (¶m, dummy_policy) ?: -fscrypt_add_test_dummy_key (sb, dummy_policy); -} -EXPORT_SYMBOL_GPL (fscrypt_set_test_dummy_encryption); -/** * fscrypt_show_test_dummy_encryption() - show '-o test_dummy_encryption' * @seq: the seq_file to print the option to
include/linux/fscrypt.h +0 -2
@@ -284,8 +284,6 @@ int fscrypt_parse_test_dummy_encryption(const struct fs_parameter *param, struct fscrypt_dummy_policy *dummy_policy);bool fscrypt_dummy_policies_equal (const struct fscrypt_dummy_policy *p1,const struct fscrypt_dummy_policy *p2); -int fscrypt_set_test_dummy_encryption (struct super_block *sb,const char *arg, -struct fscrypt_dummy_policy *dummy_policy);void fscrypt_show_test_dummy_encryption (struct seq_file *seq,char sep,struct super_block *sb); static inline bool
Signature · 8 lines
2.36.1 _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
Eric Biggers
maintainerRe: [PATCH v3 5/5] fscrypt: remove fscrypt_set_test_dummy_encryption()
Quoted 11 lines · Eric Biggers
From: Eric Biggers <[email protected]>
Now that all its callers have been converted to
fscrypt_parse_test_dummy_encryption()andfscrypt_add_test_dummy_key()instead,fscrypt_set_test_dummy_encryption()can be removed.Signed-off-by Eric Biggers <[email protected]>--- fs/crypto/policy.c | 13 ------------- include/linux/fscrypt.h | 2 -- 2 files changed, 15 deletions(-)
Now that all its prerequisites are upstream, I've applied this patch to fscrypt.git#master for 6.1.
- Eric
_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel