[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <da955b87-e2d4-6019-72a2-e00da0c8b714@zytor.com>
Date: Fri, 10 Sep 2021 00:14:26 -0700
From: "H. Peter Anvin" <hpa@...or.com>
To: Lai Jiangshan <laijs@...ux.alibaba.com>,
Lai Jiangshan <jiangshanlai@...il.com>,
linux-kernel@...r.kernel.org
Cc: Andy Lutomirski <luto@...nel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
x86@...nel.org, Joerg Roedel <jroedel@...e.de>,
Youquan Song <youquan.song@...el.com>,
Tony Luck <tony.luck@...el.com>,
Juergen Gross <jgross@...e.com>,
"Peter Zijlstra (Intel)" <peterz@...radead.org>
Subject: Re: [PATCH 17/24] x86/entry: Introduce struct ist_regs
On 9/10/21 12:13 AM, Lai Jiangshan wrote:
>
>
> On 2021/9/10 12:31, H. Peter Anvin wrote:
>> Note: the examples in this email all compiled with:
>>
>> gcc -O2 -mpreferred-stack-boundary=3 -mcmodel=kernel
>>
>> The disassembly has been slightly simplified.
>
> Saving results in registers is non-flexible no matter in ASM or hack in
> C like this.
>
> Saving CR3 in ist_regs is not different than saving rax in pt_regs, and
> both of
> ist_regs and embedded pt_regs can be moved when stack is required to
> switch.
>
> I prefer to my original design.
>
I agree. I was having some issues following your arguments.
-hpa
Powered by blists - more mailing lists