Message ID | 20170909002556.23382-1-cascardo@canonical.com |
---|---|
Headers | show |
Series | replace mbcache with mbcache2 for ext2/ext4 | expand |
On 09.09.2017 02:25, Thadeu Lima de Souza Cascardo wrote: > Third patch needed backport because of context due to commit > c6d5612f80923170dad06c04736984cc9bcc8568 ("ext4: lock the xattr block before checksuming it"). > > Removal of old mbcache was not included to minimize number of patches. > > Jan Kara (3): > mbcache2: reimplement mbcache > ext2: convert to mbcache2 > ext4: convert to mbcache2 > > fs/Makefile | 2 +- > fs/ext2/ext2.h | 3 + > fs/ext2/super.c | 25 ++-- > fs/ext2/xattr.c | 143 +++++++++---------- > fs/ext2/xattr.h | 21 +-- > fs/ext4/ext4.h | 2 +- > fs/ext4/super.c | 7 +- > fs/ext4/xattr.c | 136 +++++++++--------- > fs/ext4/xattr.h | 5 +- > fs/mbcache2.c | 359 +++++++++++++++++++++++++++++++++++++++++++++++ > include/linux/mbcache2.h | 50 +++++++ > 11 files changed, 577 insertions(+), 176 deletions(-) > create mode 100644 fs/mbcache2.c > create mode 100644 include/linux/mbcache2.h > It feels a bit excessive to fix that CVE by replacing one cache with another. Also I noticed that in the bugzilla which gets referred to from the CVE data, there are a couple of other changes around the drop of the old code. Those sound at least like improving performance. For that reason I would at least discuss this with Thadeu next week. And then decide how to go on with this. -Stefan