Re: [RFC][PATCHSET] hopefully saner handling of pathnames in cifs
From: Al Viro <viro@zeniv.linux.org.uk>
Date: 2021-03-24 15:29:25
Also in:
linux-fsdevel
Subsystem:
filesystems (vfs and infrastructure), the rest · Maintainers:
Alexander Viro, Christian Brauner, Linus Torvalds
On Tue, Mar 23, 2021 at 12:04:34AM -0500, Steve French wrote:
reran with the updated patch 7 and it failed (although I didn't have time to dig much into it today) - see http://smb3-test-rhel-75.southcentralus.cloudapp.azure.com/#/builders/2/builds/534 but it seems to run ok without patch 7 (just the first six patches) http://smb3-test-rhel-75.southcentralus.cloudapp.azure.com/#/builders/2/builds/535
Hmm... Another bug, AFAICS, is that for root we end up with "/", not "". No idea if that's all there is, though ;-/ How does one set the things up for those tests? Anyway, incremental would be
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index ed16f75ac0fa..03afad8b24af 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c@@ -114,6 +114,8 @@ build_path_from_dentry_optional_prefix(struct dentry *direntry, void *page, s = dentry_path_raw(direntry, page, PAGE_SIZE); if (IS_ERR(s)) return s; + if (!s[1]) // for root we want "", not "/" + s++; if (s < (char *)page + pplen + dfsplen) return ERR_PTR(-ENAMETOOLONG); if (pplen) {
Folded and force-pushed. Could you throw the updated branch into testing?