linux-ext4 - stop overriding SCRATCH

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>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20251218073023.1547648-1-hch@lst.de>
Date: Thu, 18 Dec 2025 08:29:58 +0100
From: Christoph Hellwig <hch@....de>
To: Zorro Lang <zlang@...nel.org>
Cc: Anand Jain <asj@...nel.org>,
	Filipe Manana <fdmanana@...e.com>,
	"Darrick J. Wong" <djwong@...nel.org>,
	fstests@...r.kernel.org,
	linux-ext4@...r.kernel.org,
	linux-xfs@...r.kernel.org
Subject: stop overriding SCRATCH_{,LOG,RT}DEV v3

Hi all,

this series cleans up various tests to get out of the business of
overriding the scratch devices setup.  This is not only generally a
little ugly, but breaks when mkfs options are incompatible with the
synthesized setups.

My prime example right now are two zoned XFS options:  -r rtreserved,
which only works for zoned RT devices, and breaks as soon as we're
running on non-RT or non-Zoned setups, and -r rtstart which only works
for internal RT zoned RT devices, and breaks as soon as an actual
SCRATCH_RTDEV was used.  There's probably more that we've been
papering over with the try_ options and by scratch_mkfs dropping
options when they conflict.  I plan to remove the need for the latter
in a follow-on series as it leads to hard to debug bugs.

Changes since v2:
 - fix the error table override in xfs/438
 - fix spelling in a commit log

Changes since v1:
 - pass a device to e2fsck in ext4/006
 - add a _require_scratch_size to xfs/521
 - ensure the file system is still mounted when checking it in xfs/528