linux-kernel - [RFC PATCH 0/3] Prohibit irq disable when disabled or enable when enabled

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: <20230509110739.241735-1-npiggin@gmail.com>
Date:   Tue,  9 May 2023 21:07:36 +1000
From:   Nicholas Piggin <npiggin@...il.com>
To:     linux-kernel@...r.kernel.org
Cc:     Nicholas Piggin <npiggin@...il.com>, linux-arch@...r.kernel.org,
        Thomas Gleixner <tglx@...utronix.de>,
        Peter Zijlstra <peterz@...radead.org>
Subject: [RFC PATCH 0/3] Prohibit irq disable when disabled or enable when enabled

Hi,

Any thoughts on tightening this up and adding some warnings for this?
Motivated by a bug I added to powerpc which broke disabled-disable
callers.

Not that I necessarily want archs to be able to depend on this, it
would always be better to be tolerant. But it seems a risky pattern
for random code to be using.

This still fires off quite a few warnings on powerpc, and I haven't
test booted any other arch so probably wouldn't ask to upstream patch
3 just yet, but if there are no objections to the idea I might do a
bit more work on it.

Thanks,
Nick

Nicholas Piggin (3):
  hrtimer: balance irq save/restore
  init: Require archs call start_kernel with arch irqs disabled
  irqflags: Warn on irq disable when disabled and enable when enabled

 include/linux/irqflags.h       | 26 ++++++++++++++++++++++++--
 init/main.c                    |  6 +++++-
 kernel/locking/irqflag-debug.c | 14 ++++++++++++++
 kernel/time/hrtimer.c          | 22 +++++++++++-----------
 4 files changed, 54 insertions(+), 14 deletions(-)

-- 
2.40.1