linux-kernel - Re: [PATCH 1/4] LoongArch: Use trampoline for exception handlers and kill la.abs

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] [day] [month] [year] [list]

Message-ID: <CAAhV-H7KpR3wKUYVR2mSfbCuwbONAmPwSB9oaLAE-4zGhx1r4g@mail.gmail.com>
Date:   Tue, 17 Jan 2023 09:38:11 +0800
From:   Huacai Chen <chenhuacai@...nel.org>
To:     Xi Ruoyao <xry111@...111.site>
Cc:     Jinyang He <hejinyang@...ngson.cn>,
        Youling Tang <tangyouling@...ngson.cn>,
        Xuerui Wang <kernel@...0n.name>, loongarch@...ts.linux.dev,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/4] LoongArch: Use trampoline for exception handlers and
 kill la.abs

On Mon, Jan 16, 2023 at 1:41 PM Xi Ruoyao <xry111@...111.site> wrote:
>
> On Mon, 2023-01-16 at 09:30 +0800, Huacai Chen wrote:
> > On Sat, Jan 14, 2023 at 2:38 PM Jinyang He <hejinyang@...ngson.cn>
> > wrote:
> > >
> > > Hi, Ruoyao and Youling,
> > >
> > > I care about the performance when NUMA enabled. We set CSR.EENTRY
> > > for each possible cpus where is NUMA-relative. So, I guess the more
> > > codes in NUMA-relative memory makes more performance. If we just set
> > > handler_trampoline as exception handler, the performance may be
> > > influenced.
> > So copying both the handlers and handler_trampoline can solve the
> > problem? If that is possible, please do that on top of the latest code
> > in
> > https://github.com/loongson/linux/commits/loongarch-next
>
> Hi folks,
>
> I just wrote the trampoline code as a PoC to show "relocatable kernel
> can work" and there must be some better way.  But I'm too sad to write
> any serious code in this month, and I don't have access to a LoongArch
> NUMA system.  So I think it's better to leave the job for you guys now
> :).
Hmm, I hope this series can be merged in 6.3. :)

Huacai
>
> Best regards
>
> --
> Xi Ruoyao <xry111@...111.site>
> School of Aerospace Science and Technology, Xidian University