qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v5 56/60] target/riscv: floating-point scalar move instructio


From: Richard Henderson
Subject: Re: [PATCH v5 56/60] target/riscv: floating-point scalar move instructions
Date: Tue, 17 Mar 2020 08:11:29 -0700
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.4.1

On 3/16/20 11:01 PM, LIU Zhiwei wrote:
> Two questions here. I don't find the answer in the specification.
> 
> 1. Should  I check RVF if the instruction uses float register,  such as all
> float point instructions and some other instructions?

I would think so, but even the 0.8 spec isn't clear.


> 2. Should I check mstatus_fs if the instruction uses float registers, or just
> for instructions that write float point register?

Definitely, just like the regular fp instructions.

This trap is how the kernel implements lazy fp context switching, so if you
allow access to fp when disabled you may be accessing values from a different
process.


r~



reply via email to

[Prev in Thread] Current Thread [Next in Thread]