]> git.ipfire.org Git - thirdparty/kernel/linux.git/commit
dcache: permit dynamic_dname()s up to NAME_MAX
authorAleksa Sarai <aleksa@amutable.com>
Tue, 31 Mar 2026 14:46:21 +0000 (01:46 +1100)
committerChristian Brauner <brauner@kernel.org>
Tue, 7 Apr 2026 10:32:22 +0000 (12:32 +0200)
commit97b67e64affb0e709eeecc50f6a9222fc20bd14b
tree1de5b0c9bfb5b3c83721595b091d881540d893f9
parent4639f1cfba03e77e4fa2062cf904de63021fb746
dcache: permit dynamic_dname()s up to NAME_MAX

dynamic_dname() has had an implicit limit of 64 characters since it was
introduced in commit c23fbb6bcb3e ("VFS: delay the dentry name
generation on sockets and pipes"), however it seems that this was a
fairly arbitrary number (suspiciously it was double the previously
hardcoded buffer size).

NAME_MAX seems like a more reasonable and consistent limit for d_name
lengths. While we're at it, we can also remove the unnecessary
stack-allocated array and just memmove() the formatted string to the end
of the buffer.

It should also be noted that at least one driver (in particular,
liveupdate's usage of anon_inode for session files) already exceeded
this limit without noticing that readlink(/proc/self/fd/$n) always
returns -ENAMETOOLONG, so this fixes those drivers as well.

Fixes: 0153094d03df ("liveupdate: luo_session: add sessions support")
Fixes: c23fbb6bcb3e ("VFS: delay the dentry name generation on sockets and pipes")
Signed-off-by: Aleksa Sarai <aleksa@amutable.com>
Link: https://patch.msgid.link/20260401-dynamic-dname-name_max-v1-1-8ca20ab2642e@amutable.com
Tested-by: Luca Boccassi <luca.boccassi@gmail.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/d_path.c