linux-kernel - Probable bug in renameat2 with RENAME

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <CA+jjjYTX80h7mRtxnYu2a=MNu5fuzxVRG4v8d=u0np0SD37qEg@mail.gmail.com>
Date: Wed, 2 Oct 2024 17:42:06 -0700
From: Joshua Hudson <joshudson@...il.com>
To: linux-kernel <linux-kernel@...r.kernel.org>
Subject: Probable bug in renameat2 with RENAME_NOREPLACE on vfat

System call:

renameat2(AT_FDCWD, "makefile", AT_FDCWD, "Makefile", RENAME_NOREPLACE);

where the current directory is on a vfat filesystem such as a USB stick

Expected return: 0

Actual return: -EEXIST

It's just not right. I really shouldn't have to remove the
RENAME_NOREPLACE to correct the case of a filename. Basically, if
oldname and newname are the same dnode (not just the same inode) the
call should proceed.

I'd rather not lose the race condition safeguard to correct the case
of a file name.