[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 02/17] target-openrisc: Streamline arithmetic an
From: |
Bastian Koppelmann |
Subject: |
Re: [Qemu-devel] [PATCH 02/17] target-openrisc: Streamline arithmetic and OVE |
Date: |
Fri, 4 Sep 2015 15:12:24 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.2.0 |
On 09/03/2015 04:44 PM, Richard Henderson wrote:
That's patch 7/17. ;-)
Woops , I just realized ;-)
R0 is not an *architectural* zero. It's a software convention. The
spec is fairly clear on this point.
I agree that there ought to be some special-casing of the software
convention, but that should require a TB flag to verify the convention
is followed. But even then I would not bother retaining the special
case here in multiply. I would only use it in the "move" and constant
formation types of pseudo instructions (or, ori, etc).
That sounds reasonable and is clearly for another patch, so
Reviewed-by: Bastian Koppelmann <address@hidden>
Cheers,
Bastian
- [Qemu-devel] [PATCH 00/17] target-openrisc improvements, Richard Henderson, 2015/09/02
- [Qemu-devel] [PATCH 05/17] target-openrisc: Use movcond where appropriate, Richard Henderson, 2015/09/02
- [Qemu-devel] [PATCH 06/17] target-openrisc: Put SR[OVE] in TB flags, Richard Henderson, 2015/09/02
- [Qemu-devel] [PATCH 04/17] target-openrisc: Keep SR_F in a separate variable, Richard Henderson, 2015/09/02
- [Qemu-devel] [PATCH 08/17] target-openrisc: Set flags on helpers, Richard Henderson, 2015/09/02
- [Qemu-devel] [PATCH 09/17] target-openrisc: Implement ff1 and fl1 for 64-bit, Richard Henderson, 2015/09/02