linux-kernel - Re: [RFC][PATCH 00/10] pstore: directly mapped regions

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

Message-ID: <b304d582-9328-4e1b-9e34-5604125b0c06@linaro.org>
Date: Mon, 17 Feb 2025 12:44:29 +0200
From: Eugen Hristev <eugen.hristev@...aro.org>
To: Johannes Berg <johannes@...solutions.net>, linux-arm-msm@...r.kernel.org,
 linux-hardening@...r.kernel.org, kees@...nel.org
Cc: linux-kernel@...r.kernel.org, gregkh@...uxfoundation.org,
 rafael@...nel.org, dakr@...nel.org, andersson@...nel.org,
 konradybcio@...nel.org, tony.luck@...el.com, gpiccoli@...lia.com,
 pmladek@...e.com, rostedt@...dmis.org, john.ogness@...utronix.de,
 senozhatsky@...omium.org, quic_mojha@...cinc.com,
 linux-arm-kernel@...ts.infradead.org, kernel@...cinc.com
Subject: Re: [RFC][PATCH 00/10] pstore: directly mapped regions

On 2/17/25 12:23, Johannes Berg wrote:
> On Mon, 2025-02-17 at 12:16 +0200, Eugen Hristev wrote:
> 
>> This series comes as an RFC proposed solution to enhance pstore and
>> devcoredump with the following functionality:
> 
> ...
> 
>> This patch series attempts to solve this by reusing existing
>> infrastructure in pstore and devcoredump, and provide a copy-free
> 
> ...
> 
> You mention devcoredump multiple times, but it almost seems like you
> don't even know what devcoredump does? I don't see how there's any
> relation at all, and the code added to it seems to have no relation to
> the actual functionality of devcoredump either?

At this moment going through devcoredump is not something that impacts
the idea of the implementation.
The whole reason of going through it (because things work without it as
well), is to see whether this has any kind of impact or not, and if
there is any kind of fit/reason of going through it.
Devcoredump is involved because the whole core registration is similar
to a core area that devcoredump could use.
For example, would it be interesting to have a handler going through all
devices, and have the dump areas already registered ?
Meaning, when there is a request to generate a core dump, one could
directly dump this area instead of calling back the driver, and provide
that to the userspace instead of the driver calling the dev_coredumpv by
its own.

> 
> johannes