ext4: always initialize the crc32c checksum driver
Subsystems (2)
Theodore Ts'o
maintainerPATCH 1/2 ext4: always initialize the crc32c checksum driver
The extended attribute code now uses the crc32c checksum for hashing purposes, so we should just always always initialize it. We also want to prevent NULL pointer dereferences if one of the metadata checksum features is enabled after the file sytsem is originally mounted.
https://bugzilla.kernel.org/show_bug.cgi?id=199183
--- fs/ext4/super.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-)
fs/ext4/super.c +6 -9
@@ -3492,15 +3492,12 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) }/* Load the checksum driver */ -if (ext4_has_feature_metadata_csum (sb) || -ext4_has_feature_ea_inode (sb)) { - sbi->s_chksum_driver =crypto_alloc_shash ("crc32c" ,0 ,0 ); -if (IS_ERR (sbi->s_chksum_driver)) { -ext4_msg (sb,KERN_ERR ,"Cannot load crc32c driver." ); - ret =PTR_ERR (sbi->s_chksum_driver); - sbi->s_chksum_driver =NULL ; - goto failed_mount; - } + sbi->s_chksum_driver =crypto_alloc_shash ("crc32c" ,0 ,0 ); +if (IS_ERR (sbi->s_chksum_driver)) { +ext4_msg (sb,KERN_ERR ,"Cannot load crc32c driver." ); + ret =PTR_ERR (sbi->s_chksum_driver); + sbi->s_chksum_driver =NULL ; + goto failed_mount; }/* Check superblock checksum */
Signature · 1 line
2.16.1.72.g5be1f00a9a
Theodore Ts'o
maintainerPATCH 2/2 ext4: don't allow r/w mounts if metadata blocks overlap the superblock
If some metadata block, such as an allocation bitmap, overlaps the superblock, it's very likely that if the file system is mounted read/write, the results will not be pretty. So disallow r/w mounts for file systems corrupted in this particular way.
--- fs/ext4/super.c | 6 ++++++ 1 file changed, 6 insertions(+)
fs/ext4/super.c +6 -0
@@ -2335,6 +2335,8 @@ static int ext4_check_descriptors(struct super_block *sb,ext4_msg (sb,KERN_ERR ,"ext4_check_descriptors: " "Block bitmap for group %u overlaps " "superblock" , i); +if (!sb_rdonly (sb)) +return 0 ; }if (block_bitmap < first_block || block_bitmap > last_block) { ext4_msg(sb, KERN_ERR, "ext4_check_descriptors: "
@@ -2347,6 +2349,8 @@ static int ext4_check_descriptors(struct super_block *sb,ext4_msg (sb,KERN_ERR ,"ext4_check_descriptors: " "Inode bitmap for group %u overlaps " "superblock" , i); +if (!sb_rdonly (sb)) +return 0 ; }if (inode_bitmap < first_block || inode_bitmap > last_block) { ext4_msg(sb, KERN_ERR, "ext4_check_descriptors: "
@@ -2359,6 +2363,8 @@ static int ext4_check_descriptors(struct super_block *sb,ext4_msg (sb,KERN_ERR ,"ext4_check_descriptors: " "Inode table for group %u overlaps " "superblock" , i); +if (!sb_rdonly (sb)) +return 0 ; }if (inode_table < first_block || inode_table + sbi->s_itb_per_group -1 > last_block) {
Signature · 1 line
2.16.1.72.g5be1f00a9a
Re: [PATCH 2/2] ext4: don't allow r/w mounts if metadata blocks overlap the superblock
Hi EXT4 ML!
I'm working on a project to perform automatic patch selection for stable, and automate the testing being done for patches going in various stable tree.
I got some input from the XFS folks so far, and wanted to get some input from the EXT4 folks as well.
Below is a mail that would be sent as a response for ext4 patches tagged for stable and sent to the mailing list. Since this one is already tagged for stable, these results should be useful for people to understand if the patch properly applies to various stable trees and passes testing.
For EXT4, at this point testing is the xfstests tree at
https://github.com/tytso/xfstests , running with "./check -g auto". Complete build and test logs are available in the links embedded in the mail.
I'd be happy to get feedback for this. Thanks!
===
Hi Theodore Ts'o,
[This is an automated email]
This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: all
The bot has also determined it's probably a bug fixing patch. (score: 90.6059)
The bot has tested the following trees: v4.15.12, v4.14.29, v4.9.89, v4.4.123, v4.1.50, v3.18.101.
v4.15.12: Build OK!
v4.14.29: Build OK!
v4.9.89: Build failed! Errors:
fs/ext4/super.c:2263:9: error: implicit declaration of function ‘sb_rdonly’ [-Werror=implicit-function-declaration]
v4.4.123: Build failed! Errors:
fs/ext4/super.c:2134:9: error: implicit declaration of function ‘sb_rdonly’ [-Werror=implicit-function-declaration]
v4.1.50: Build failed! Errors:
fs/ext4/super.c:2129:9: error: implicit declaration of function ‘sb_rdonly’ [-Werror=implicit-function-declaration]
v3.18.101: Build failed! Errors:
fs/ext4/super.c:2096:9: error: implicit declaration of function ‘sb_rdonly’ [-Werror=implicit-function-declaration]
EXT4 Specific tests:
v4.15.12 (http://stable-bot.westus2.cloudapp.azure.com/ext4-patch/v4.15.12/tests/):
Tests complete.
v4.14.29 (http://stable-bot.westus2.cloudapp.azure.com/ext4-patch/v4.14.29/tests/):
Tests complete.
v4.9.89 (http://stable-bot.westus2.cloudapp.azure.com/ext4-patch/v4.9.89/tests/):
No tests completed!
v4.4.123 (http://stable-bot.westus2.cloudapp.azure.com/ext4-patch/v4.4.123/tests/):
No tests completed!
v4.1.50 (http://stable-bot.westus2.cloudapp.azure.com/ext4-patch/v4.1.50/tests/):
No tests completed!
v3.18.101 (http://stable-bot.westus2.cloudapp.azure.com/ext4-patch/v3.18.101/tests/):
No tests completed!
Please let us know if you'd like to have this patch included in a stable tree.
--
Thanks,
Sasha
Quoted 27 lines · Theodore Ts'o
The extended attribute code now uses the crc32c checksum for hashing purposes, so we should just always always initialize it. We also want to prevent NULL pointer dereferences if one of the metadata checksum features is enabled after the file sytsem is originally mounted.
https://bugzilla.kernel.org/show_bug.cgi?id=199183
Signed-off-by Theodore Ts'o <[email protected]>--- fs/ext4/super.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-)fs/ext4/super.c +0
-6@@ -3492,15 +3492,12 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) }/* Load the checksum driver */ -if (ext4_has_feature_metadata_csum (sb) || -ext4_has_feature_ea_inode (sb)) { - sbi->s_chksum_driver =crypto_alloc_shash ("crc32c" ,0 ,0 ); -if (IS_ERR (sbi->s_chksum_driver)) { -ext4_msg (sb,KERN_ERR ,"Cannot load crc32c driver." ); - ret =PTR_ERR (sbi->s_chksum_driver);
fs/ext4/super.c static int ext4_fill_super(struct super_block *sb, void *data, int silent) · line 3501
Quoted 10 lines
fs/ext4/super.c +6
-3- sbi->s_chksum_driver =NULL ; - goto failed_mount; - } + sbi->s_chksum_driver =crypto_alloc_shash ("crc32c" ,0 ,0 ); +if (IS_ERR (sbi->s_chksum_driver)) { +ext4_msg (sb,KERN_ERR ,"Cannot load crc32c driver." ); + ret =PTR_ERR (sbi->s_chksum_driver); + sbi->s_chksum_driver =NULL ; + goto failed_mount; }
Why do you need to do all the manual initialization of the crc32c lowlevel driver, when you can just make ext4 select LIBCRC32c and in the code just use crc32c? Refer to include/linux/crc32c.h for the interface and lib/libcrc32c.c for the low-level details.
Quoted 3 lines
fs/ext4/super.c +0
-0/* Check superblock checksum */