linux-kernel - [RFC 0/2] printk: Add more metadata for each record

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]

Date:   Wed, 23 Sep 2020 15:56:15 +0200
From:   Petr Mladek <pmladek@...e.com>
To:     Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        John Ogness <john.ogness@...utronix.de>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Prarit Bhargava <prarit@...hat.com>,
        Mark Salyzyn <salyzyn@...roid.com>,
        Chunyan Zhang <zhang.lyra@...il.com>,
        Orson Zhai <orsonzhai@...il.com>,
        Changki Kim <changki.kim@...sung.com>,
        Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
        linux-kernel@...r.kernel.org, Petr Mladek <pmladek@...e.com>
Subject: [RFC 0/2] printk: Add more metadata for each record

There have been repeated requests to store and print more metadata
for each line, see the two patches for more details.

This patchset handles the storing part so that we could do it together
with the other changes of the ring buffer structure. It would be nice
to get it into the same kernel release to reduce pain on the crash
tools side.

My proposal is to make this information available via /dev/kmsg as
dictionary values. The console output would probably need to be configurable.

I would ignore syslog syscall that is hard to extend and is basically
obsoleted by /dev/kmsg.

The patchset applies on top of the current printk/linux.git,
branch printk-rework (after the dict ring removal).

You also need to merge "timekeeping-for-printk-2020-08-23" tag
from tip.git. It provides "struct ktime_timestamps" and
ktime_get_fast_timestamps().

I still need to add the gdb-related changes if there is not a strong
opposition to these changes. I just wanted to send what I had before
starting review of the logbuf_lock removal patchset.

Petr Mladek (2):
  printk: Store all three timestamps
  printk: Add more information about the printk caller

 kernel/printk/printk.c            | 129 ++++++++++++++++++++----------
 kernel/printk/printk_ringbuffer.c |  43 +++++-----
 kernel/printk/printk_ringbuffer.h |  42 +++++++++-
 3 files changed, 148 insertions(+), 66 deletions(-)

-- 
2.26.2