From: DJ Delorie <dj@redhat.com>
To: bug-gnulib@gnu.org
Subject: [patch v2] glob: resolve DT_UNKNOWN via is_dir
Date: Fri, 11 Mar 2022 16:43:39 -0500 [thread overview]
Message-ID: <xnbkycjhbo.fsf@greed.delorie.com> (raw)
In-Reply-To: <xnsfvv30on.fsf@greed.delorie.com> (message from DJ Delorie on Tue, 16 Nov 2021 22:45:12 -0500)
[v2: changed malloc failure from ignore to error; added support for
alloca; tested by copying to glibc and testing there]
The DT_* values returned by getdents (readdir) are only hints and
not required. In fact, some Linux filesystems return DT_UNKNOWN
for most entries, regardless of actual type. This causes make
to mis-match patterns with a trailing slash (via GLOB_ONLYDIR)
(see make's functions/wildcard test case). Thus, this patch
detects that case and uses is_dir() to make the type known enough
for proper operation.
Performance in non-DT_UNKNOWN cases is not affected.
The lack of DT_* is a well known issue on older XFS installations
(for example, RHEL 7 and 8, Fedora 28) but can be recreated by
creating an XFS filesystem with flags that mimic older behavior:
$ fallocate -l 10G /xfs.fs
$ mkfs.xfs -n ftype=0 -m crc=0 -f /xfs.fs
$ mkdir /xfs
$ mount -o loop /xfs.fs /xfs
diff --git a/lib/glob.c b/lib/glob.c
index f8d8a306f2..c28c92a42c 100644
--- a/lib/glob.c
+++ b/lib/glob.c
@@ -1381,7 +1381,33 @@ glob_in_dir (const char *pattern, const char *directory, int flags,
if (flags & GLOB_ONLYDIR)
switch (readdir_result_type (d))
{
- case DT_DIR: case DT_LNK: case DT_UNKNOWN: break;
+ case DT_DIR: case DT_LNK: break;
+ case DT_UNKNOWN:
+ {
+ /* The filesystem was too lazy to give us a hint,
+ so we have to do it the hard way. */
+ char *fullpath, *p;
+ bool isdir;
+ int need = strlen (directory) + strlen (d.name) + 2;
+ int use_alloca = glob_use_alloca (alloca_used, need);
+ if (use_alloca)
+ fullpath = alloca_account (need, alloca_used);
+ else
+ {
+ fullpath = malloc (need);
+ if (fullpath == NULL)
+ goto memory_error;
+ }
+ p = stpcpy (fullpath, directory);
+ *p++ = '/';
+ strcpy (p, d.name);
+ isdir = is_dir (fullpath, flags, pglob);
+ if (!use_alloca)
+ free (fullpath);
+ if (isdir)
+ break;
+ continue;
+ }
default: continue;
}
next prev parent reply other threads:[~2022-03-11 21:43 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-17 3:45 [patch] glob: resolve DT_UNKNOWN via is_dir DJ Delorie
2021-11-17 11:12 ` Florian Weimer
2021-11-17 21:59 ` DJ Delorie
2021-11-17 22:14 ` Florian Weimer
2021-11-17 22:38 ` DJ Delorie
2022-03-11 21:43 ` DJ Delorie [this message]
2022-03-12 0:37 ` [patch v2] " Paul Eggert
2022-03-23 4:34 ` DJ Delorie
2022-03-23 17:35 ` Paul Eggert
2022-03-25 19:18 ` DJ Delorie
2022-05-14 20:34 ` Bruno Haible
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://lists.gnu.org/mailman/listinfo/bug-gnulib
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=xnbkycjhbo.fsf@greed.delorie.com \
--to=dj@redhat.com \
--cc=bug-gnulib@gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).