On Thu, Nov 14, 2019 at 4:12 PM Miklos Szeredi miklos@szeredi.hu wrote:
On Wed, Nov 13, 2019 at 5:34 PM Amir Goldstein amir73il@gmail.com wrote:
On Wed, Nov 13, 2019 at 6:02 PM Miklos Szeredi miklos@szeredi.hu wrote:
On Thu, Nov 7, 2019 at 11:50 AM Colin King colin.king@canonical.com wrote:
From: Colin Ian King colin.king@canonical.com
In the past, overlayfs required that lower fs have non null uuid in order to support nfs export and decode copy up origin file handles.
Commit 9df085f3c9a2 ("ovl: relax requirement for non null uuid of lower fs") relaxed this requirement for nfs export support, as long as uuid (even if null) is unique among all lower fs.
However, said commit unintentionally also relaxed the non null uuid requirement for decoding copy up origin file handles, regardless of the unique uuid requirement.
Amend this mistake by disabling decoding of copy up origin file handle from lower fs with a conflicting uuid.
We still encode copy up origin file handles from those fs, because file handles like those already exist in the wild and because they might provide useful information in the future.
Reported-by: Colin Ian King colin.king@canonical.com Link: https://lore.kernel.org/lkml/20191106234301.283006-1-colin.king@canonical.co... Fixes: 9df085f3c9a2 ("ovl: relax requirement for non null uuid ...") Cc: stable@vger.kernel.org # v4.20+ Signed-off-by: Amir Goldstein amir73il@gmail.com Signed-off-by: Colin Ian King colin.king@canonical.com
fs/overlayfs/namei.c | 8 ++++++++ fs/overlayfs/ovl_entry.h | 2 ++ fs/overlayfs/super.c | 16 ++++++++++------ 3 files changed, 20 insertions(+), 6 deletions(-)
diff --git a/fs/overlayfs/namei.c b/fs/overlayfs/namei.c index e9717c2f7d45..f47c591402d7 100644 --- a/fs/overlayfs/namei.c +++ b/fs/overlayfs/namei.c @@ -325,6 +325,14 @@ int ovl_check_origin_fh(struct ovl_fs *ofs, struct ovl_fh *fh, bool connected, int i;
for (i = 0; i < ofs->numlower; i++) {
/** If lower fs uuid is not unique among lower fs we cannot match* fh->uuid to layer.*/if (ofs->lower_layers[i].fsid &&ofs->lower_layers[i].fs->bad_uuid)continue;origin = ovl_decode_real_fh(fh, ofs->lower_layers[i].mnt, connected); if (origin)diff --git a/fs/overlayfs/ovl_entry.h b/fs/overlayfs/ovl_entry.h index a8279280e88d..28348c44ea5b 100644 --- a/fs/overlayfs/ovl_entry.h +++ b/fs/overlayfs/ovl_entry.h @@ -22,6 +22,8 @@ struct ovl_config { struct ovl_sb { struct super_block *sb; dev_t pseudo_dev;
/* Unusable (conflicting) uuid */bool bad_uuid;};
struct ovl_layer { diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c index afbcb116a7f1..5d4faab57ba0 100644 --- a/fs/overlayfs/super.c +++ b/fs/overlayfs/super.c @@ -1255,17 +1255,18 @@ static bool ovl_lower_uuid_ok(struct ovl_fs *ofs, const uuid_t *uuid) { unsigned int i;
if (!ofs->config.nfs_export && !(ofs->config.index && ofs->upper_mnt))return true;Colin, I mislead you, this should be (I think):
if (!ofs->config.nfs_export && !ofs->upper_mnt) return true;for (i = 0; i < ofs->numlowerfs; i++) { /* * We use uuid to associate an overlay lower file handle with a * lower layer, so we can accept lower fs with null uuid as long * as all lower layers with null uuid are on the same fs.
* if we detect multiple lower fs with the same uuid, we* disable lower file handle decoding on all of them. */
if (uuid_equal(&ofs->lower_fs[i].sb->s_uuid, uuid))
if (uuid_equal(&ofs->lower_fs[i].sb->s_uuid, uuid)) {ofs->lower_fs[i].bad_uuid = true; return false;} } return true;} @@ -1277,6 +1278,7 @@ static int ovl_get_fsid(struct ovl_fs *ofs, const struct path *path) unsigned int i; dev_t dev; int err;
bool bad_uuid = false; /* fsid 0 is reserved for upper fs even with non upper overlay */ if (ofs->upper_mnt && ofs->upper_mnt->mnt_sb == sb)@@ -1287,10 +1289,11 @@ static int ovl_get_fsid(struct ovl_fs *ofs, const struct path *path) return i + 1; }
if (!ovl_lower_uuid_ok(ofs, &sb->s_uuid)) {
if (ofs->upper_mnt && !ovl_lower_uuid_ok(ofs, &sb->s_uuid)) {This seems bogus: why only check conflicting lower layers if there's an upper layer?
It is bogus - it was my (wrong) suggestion. The thinking was that we only decode origin fh if we have an upper layer and index only valid with upper layer. I forgot the case of nfs_export and lower-only setup. Suggested fix above.
bad_uuid = true; ofs->config.index = false; ofs->config.nfs_export = false;
pr_warn("overlayfs: %s uuid detected in lower fs '%pd2', falling back to index=off,nfs_export=off.\n",
pr_warn("overlayfs: %s uuid detected in lower fs '%pd2', enforcing index=off,nfs_export=off.\n",And this while this makes sense, it doesn't really fit into this patch (no change of behavior regarding how index and nfs_export are handled).
Again, this was my (not wrong?) suggestion. What this patch changes is that ovl_lower_uuid_ok() can now return false and we get to this print although user did not ask for index nor nfs_export. So the "falling back" language no longer makes sense.
But does "enforcing" makes sense in this light? That's not what the detected bad_uuid condition is about, it's about failing to utilize origin markings to make inode numbers persistent for filesystems that have null uuid. Is that correct?
That is not exactly how I would describe bad_uuid. ovl_lower_uuid_ok() had already existed for a while it was requires for decoding lower file handles, which is a requirement of both index and nfs_export.
What Colin has now reported brings to light the fact that decoding lower file handles was also required for making inode numbers persistent.
So the bad_uuid condition is required for all of the above, not just for decoding origin.
Can we do a message that makes that somewhat more clearer?
What about the logs:
pr_warn("overlayfs: upper fs does not support xattr, falling back to index=off and metacopy=off.\n"); pr_warn("overlayfs: upper fs does not support file handles, falling back to index=off.\n"); pr_warn("overlayfs: fs on '%s' does not support file handles, falling back to index=off,nfs_export=off.\n",
Should we also change them to reflect the fact the decoding origin is not supported???
Seems like a lot of hassle that will end up writing too much information that most people won't understand.
IIRC, we also do not guaranty persistent inode numbers for hardlinks when index=off.
As for the change in question (falling back => enforcing), if that bothers you, we can get rid of this change by testing emitting the print only if (ofs->config.nfs_export || ofs->config.index).
Thanks, Amir.