On Thu, Aug 14, 2025 at 7:22 PM André Almeida <andrealmeid@xxxxxxxxxx> wrote:
>
> To add overlayfs support casefold layers, create a new function
> ovl_casefold(), to be able to do case-insensitive strncmp().
>
> ovl_casefold() allocates a new buffer and stores the casefolded version
> of the string on it. If the allocation or the casefold operation fails,
> fallback to use the original string.
>
> The case-insentive name is then used in the rb-tree search/insertion
> operation. If the name is found in the rb-tree, the name can be
> discarded and the buffer is freed. If the name isn't found, it's then
> stored at struct ovl_cache_entry to be used later.
>
> Signed-off-by: André Almeida <andrealmeid@xxxxxxxxxx>
> ---
> Changes from v4:
> - Move the consumer/free buffer logic out to the caller
> - s/aux/c_name
>
> Changes from v3:
> - Improve commit message text
> - s/OVL_NAME_LEN/NAME_MAX
> - drop #ifdef in favor of if(IS_ENABLED)
> - use new helper sb_encoding
> - merged patch "Store casefold name..." and "Create ovl_casefold()..."
> - Guard all the casefolding inside of IS_ENABLED(UNICODE)
>
> Changes from v2:
> - Refactor the patch to do a single kmalloc() per rb_tree operation
> - Instead of casefolding the cache entry name everytime per strncmp(),
> casefold it once and reuse it for every strncmp().
> ---
> fs/overlayfs/readdir.c | 115 +++++++++++++++++++++++++++++++++++++++++--------
> 1 file changed, 97 insertions(+), 18 deletions(-)
>
> diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
> index b65cdfce31ce27172d28d879559f1008b9c87320..803ac6a7516d0156ae7793ee1ff884dbbf2e20b0 100644
> --- a/fs/overlayfs/readdir.c
> +++ b/fs/overlayfs/readdir.c
> @@ -27,6 +27,8 @@ struct ovl_cache_entry {
> bool is_upper;
> bool is_whiteout;
> bool check_xwhiteout;
> + const char *cf_name;
> + int cf_len;
We should also change these member names to c_name
Because they are the "compare/canonicalized" name, which
may or may not be casefolded.
> char name[];
> };
>
> @@ -45,6 +47,7 @@ struct ovl_readdir_data {
> struct list_head *list;
> struct list_head middle;
> struct ovl_cache_entry *first_maybe_whiteout;
> + struct unicode_map *map;
> int count;
> int err;
> bool is_upper;
> @@ -66,6 +69,27 @@ static struct ovl_cache_entry *ovl_cache_entry_from_node(struct rb_node *n)
> return rb_entry(n, struct ovl_cache_entry, node);
> }
>
> +static int ovl_casefold(struct unicode_map *map, const char *str, int len, char **dst)
> +{
> + const struct qstr qstr = { .name = str, .len = len };
> + int cf_len;
> +
> + if (!IS_ENABLED(CONFIG_UNICODE) || !map || is_dot_dotdot(str, len))
> + return 0;
> +
> + *dst = kmalloc(NAME_MAX, GFP_KERNEL);
> +
> + if (dst) {
> + cf_len = utf8_casefold(map, &qstr, *dst, NAME_MAX);
> +
> + if (cf_len > 0)
> + return cf_len;
> + }
> +
> + kfree(*dst);
> + return 0;
> +}
> +
> static bool ovl_cache_entry_find_link(const char *name, int len,
> struct rb_node ***link,
> struct rb_node **parent)
> @@ -79,7 +103,7 @@ static bool ovl_cache_entry_find_link(const char *name, int len,
>
> *parent = *newp;
> tmp = ovl_cache_entry_from_node(*newp);
> - cmp = strncmp(name, tmp->name, len);
> + cmp = strncmp(name, tmp->cf_name, tmp->cf_len);
> if (cmp > 0)
> newp = &tmp->node.rb_right;
> else if (cmp < 0 || len < tmp->len)
This looks like a bug - should be len < tmp->c_len
> @@ -101,7 +125,7 @@ static struct ovl_cache_entry *ovl_cache_entry_find(struct rb_root *root,
> while (node) {
> struct ovl_cache_entry *p = ovl_cache_entry_from_node(node);
>
> - cmp = strncmp(name, p->name, len);
> + cmp = strncmp(name, p->cf_name, p->cf_len);
> if (cmp > 0)
> node = p->node.rb_right;
> else if (cmp < 0 || len < p->len)
Same here.
But it's not the only bug, because this patch regresses 3 fstests without
enabling any casefolding:
overlay/038 12s ... [14:16:39] [14:16:50]- output mismatch (see
/results/overlay/results-large/overlay/038.out.bad)
--- tests/overlay/038.out 2025-05-25 08:52:54.000000000 +0000
+++ /results/overlay/results-large/overlay/038.out.bad 2025-08-17
14:16:50.549367654 +0000
@@ -1,2 +1,3 @@
QA output created by 038
+Merged dir: Invalid d_ino reported for ..
Silence is golden
overlay/041 11s ... [14:16:54] [14:17:05]- output mismatch (see
/results/overlay/results-large/overlay/041.out.bad)
--- tests/overlay/041.out 2025-05-25 08:52:54.000000000 +0000
+++ /results/overlay/results-large/overlay/041.out.bad 2025-08-17
14:17:05.275206922 +0000
@@ -1,2 +1,3 @@
QA output created by 041
+Merged dir: Invalid d_ino reported for ..
Silence is golden
overlay/077 19s ... [14:17:08][ 107.348626] WARNING: CPU: 3 PID:
5414 at fs/overlayfs/readdir.c:677 ovl_dir_read_impure+0x178/0x1c0
[ 107.354647] ---[ end trace 0000000000000000 ]---
[ 107.399525] WARNING: CPU: 2 PID: 5415 at fs/overlayfs/readdir.c:677
ovl_dir_read_impure+0x178/0x1c0
[ 107.406826] ---[ end trace 0000000000000000 ]---
_check_dmesg: something found in dmesg (see
/results/overlay/results-large/overlay/077.dmesg)
[14:17:28]- output mismatch (see
/results/overlay/results-large/overlay/077.out.bad)
--- tests/overlay/077.out 2025-05-25 08:52:54.000000000 +0000
+++ /results/overlay/results-large/overlay/077.out.bad 2025-08-17
14:17:28.762250671 +0000
@@ -1,2 +1,6 @@
QA output created by 077
+getdents: Input/output error
+Missing created file in impure upper dir (see
/results/overlay/results-large/overlay/077.full for details)
+getdents: Input/output error
+Found unlinked file in impure upper dir (see
/results/overlay/results-large/overlay/077.full for details)
Silence is golden
Thanks,
Amir.
