[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 0/2] ARM: fix VLD of one element to all lanes
From: |
Aurelien Jarno |
Subject: |
Re: [Qemu-devel] [PATCH 0/2] ARM: fix VLD of one element to all lanes |
Date: |
Fri, 1 Apr 2011 22:34:36 +0200 |
User-agent: |
Mutt/1.5.20 (2009-06-14) |
On Tue, Mar 15, 2011 at 04:26:50PM +0000, Peter Maydell wrote:
> This patchset fixes various bugs in the implementation of the Neon
> instructions loading a single element to all lanes.
>
> The primary bug is that VLD1's "single element to all lanes" form differs
> from those for VLD2, VLD3 and VLD4 in that bit 5 indicates whether the
> loaded element should be written to one or two Dregs, rather than being a
> register stride. (This is the issue addressed by Meego commit 6f3b4ee,
> although my choice of fix is different.)
>
> We were also incorrectly UNDEFfing VLD4.32 with 16 byte alignment
> specifier, failing to UNDEF for invalid size and alignment combinations,
> and leaking a TCG temporary on the UNDEF code paths.
>
> Tested via the usual random instruction set testing, and also with the
> valgrind 'neon64' test case.
>
> Peter Maydell (2):
> target-arm: Fix VLD of single element to all lanes
> target-arm: Don't leak TCG temp for UNDEFs in Neon load/store space
>
> target-arm/translate.c | 92
> ++++++++++++++++++++++++++++++++++--------------
> 1 files changed, 65 insertions(+), 27 deletions(-)
>
Thanks, both applied.
--
Aurelien Jarno GPG: 1024D/F1BCDB73
address@hidden http://www.aurel32.net
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- Re: [Qemu-devel] [PATCH 0/2] ARM: fix VLD of one element to all lanes,
Aurelien Jarno <=