linux-kernel - [patch part-II V2 00/13] x86/entry: Consolidation

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: <20200308222359.370649591@linutronix.de>
Date:   Sun, 08 Mar 2020 23:23:59 +0100
From:   Thomas Gleixner <tglx@...utronix.de>
To:     LKML <linux-kernel@...r.kernel.org>
Cc:     x86@...nel.org, Steven Rostedt <rostedt@...dmis.org>,
        Brian Gerst <brgerst@...il.com>,
        Juergen Gross <jgross@...e.com>,
        Frederic Weisbecker <frederic@...nel.org>,
        Alexandre Chartre <alexandre.chartre@...cle.com>
Subject: [patch part-II V2 00/13]  x86/entry: Consolidation - Part II (syscalls)

Hi!

This is the second version of the syscall entry code consolidation
series. V1 can be found here:

  https://lore.kernel.org/r/20200225220801.571835584@linutronix.de

It applies on top of

  git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/entry

and is also available from git:

    git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel entry-v2-part2

The changes vs. V1:

  - Addressed the review comments. The main change is the rework of
    preparing the move of interrupt tracing to C-code. The new version
    creates less churn which needs to be mopped up afterwards again.

  - Provide new variants for trace_hardirqs_on() which are not using
    the rcuidle tracepoint mechanism and do not call into lockdep.

    This allows to split up lockdep and tracing in the enter from user and
    exit to user implementation.

    Aside of addressing the rcuidle issue of functions attached to
    tracepoints which are not rcuidle safe, e.g. BPF, this also has a
    performance advantage as it spares the srcu/rcu_irq dance around the
    tracepoint before enter_from_user_mode() which turns RCU back on
    anyway. Same on the way out.

  - Picked up Reviewed-by and Acked-by tags

Thanks,

	tglx
---
 arch/x86/entry/common.c                |  106 +++++++++++++++++++++++++--------
 arch/x86/entry/entry_32.S              |   26 +-------
 arch/x86/entry/entry_64.S              |    6 -
 arch/x86/entry/entry_64_compat.S       |   32 +--------
 arch/x86/include/asm/nospec-branch.h   |    4 -
 include/linux/context_tracking.h       |   14 ++--
 include/linux/context_tracking_state.h |    6 -
 include/linux/irqflags.h               |    4 +
 kernel/context_tracking.c              |    9 +-
 kernel/trace/trace_preemptirq.c        |   23 +++++++
 10 files changed, 136 insertions(+), 94 deletions(-)