qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Richard Henderson <richard.henderson@linaro.org>
To: "Philippe Mathieu-Daudé" <f4bug@amsat.org>, qemu-devel@nongnu.org
Cc: alex.bennee@linaro.org
Subject: Re: [Qemu-devel] [PATCH v2 07/16] target/arm: Align vector registers
Date: Wed, 13 Sep 2017 08:28:28 -0700	[thread overview]
Message-ID: <4262b16c-fd18-b551-a510-30833df414ad@linaro.org> (raw)
In-Reply-To: <d1f97f10-921e-1c50-9407-2007b1cbe169@amsat.org>

On 09/12/2017 11:50 AM, Philippe Mathieu-Daudé wrote:
>>
>> -        float64 regs[64];
>> +        float64 regs[64] __attribute__((aligned(16)));
> 
> I understand this should be aligned to the biggest vector register the host
> support, i.e. for AVX-512 this would be QEMU_ALIGNED(64), is it correct?

No.

Alignment of 16 is sufficient for "older" vector extensions, like altivec,
which require alignment in load/store insns.  But (so far at least) newer
vector extensions with larger vector sizes (AVX2, AVX512, ARM SVE) handle
unaligned load/store operations just fine.

Which means we need not require excessive alignment within the cpu struct.

The rule for this is documented in tcg/tcg-op-gvec.h, iirc.


r~

  parent reply	other threads:[~2017-09-13 15:28 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-12 16:24 [Qemu-devel] [PATCH v2 00/16] TCG vectorization and example conversion Richard Henderson
2017-09-12 16:24 ` [Qemu-devel] [PATCH v2 01/16] tcg: Add expanders for out-of-line vector helpers Richard Henderson
2017-09-12 16:24 ` [Qemu-devel] [PATCH v2 02/16] tcg: Add types for host vectors Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 03/16] tcg: Add operations " Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 04/16] tcg: Add tcg_op_supported Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 05/16] tcg: Add INDEX_op_invalid Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 06/16] tcg: Add vector infrastructure and ops for add/sub/logic Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 07/16] target/arm: Align vector registers Richard Henderson
2017-09-12 18:50   ` Philippe Mathieu-Daudé
2017-09-12 20:44     ` Philippe Mathieu-Daudé
2017-09-13 15:28     ` Richard Henderson [this message]
2017-09-12 18:55   ` Peter Maydell
2017-09-12 20:17     ` Philippe Mathieu-Daudé
2017-09-12 20:20       ` Peter Maydell
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 08/16] target/arm: Use vector infrastructure for aa64 add/sub/logic Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 09/16] tcg/i386: Add vector operations Richard Henderson
2017-09-14 16:20   ` Alex Bennée
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 10/16] tcg/aarch64: Fully convert tcg_target_op_def Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 11/16] tcg: Remove tcg_regset_clear Richard Henderson
2017-09-12 18:52   ` Philippe Mathieu-Daudé
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 12/16] tcg: Remove tcg_regset_set Richard Henderson
2017-09-12 18:52   ` Philippe Mathieu-Daudé
2017-09-15 10:21   ` Alex Bennée
2017-09-15 17:03     ` Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 13/16] tcg: Remove tcg_regset_{or, and, andnot, not} Richard Henderson
2017-09-12 18:52   ` Philippe Mathieu-Daudé
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 14/16] tcg: Remove tcg_regset_set32 Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 15/16] tcg: Fix types in tcg_regset_{set, reset}_reg Richard Henderson
2017-09-12 16:25 ` [Qemu-devel] [PATCH v2 16/16] tcg/aarch64: Add vector operations Richard Henderson
2017-09-12 16:40 ` [Qemu-devel] [PATCH v2 00/16] TCG vectorization and example conversion no-reply
2017-09-12 16:52 ` no-reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4262b16c-fd18-b551-a510-30833df414ad@linaro.org \
    --to=richard.henderson@linaro.org \
    --cc=alex.bennee@linaro.org \
    --cc=f4bug@amsat.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).