qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: Paul Brook <paul@nowt.org>,
	Richard Henderson <richard.henderson@linaro.org>,
	Eduardo Habkost <eduardo@habkost.net>
Cc: "open list:All patches CC here" <qemu-devel@nongnu.org>
Subject: Re: [PATCH v2 01/42] i386: pcmpestr 64-bit sign extension bug
Date: Wed, 27 Apr 2022 09:00:31 +0200	[thread overview]
Message-ID: <860ea92c-355a-07df-23b7-a46c7e84d9b2@redhat.com> (raw)
In-Reply-To: <20220424220204.2493824-2-paul@nowt.org>

On 4/25/22 00:01, Paul Brook wrote:
> The abs1 function in ops_sse.h only works sorrectly when the result fits
> in a signed int. This is fine most of the time because we're only dealing
> with byte sized values.
> 
> However pcmp_elen helper function uses abs1 to calculate the absolute value
> of a cpu register. This incorrectly truncates to 32 bits, and will give
> the wrong anser for the most negative value.
> 
> Fix by open coding the saturation check before taking the absolute value.
> 
> Signed-off-by: Paul Brook <paul@nowt.org>

Queued, thanks.

Paolo

> ---
>   target/i386/ops_sse.h | 20 +++++++++-----------
>   1 file changed, 9 insertions(+), 11 deletions(-)
> 
> diff --git a/target/i386/ops_sse.h b/target/i386/ops_sse.h
> index e4d74b814a..535440f882 100644
> --- a/target/i386/ops_sse.h
> +++ b/target/i386/ops_sse.h
> @@ -2011,25 +2011,23 @@ SSE_HELPER_Q(helper_pcmpgtq, FCMPGTQ)
>   
>   static inline int pcmp_elen(CPUX86State *env, int reg, uint32_t ctrl)
>   {
> -    int val;
> +    target_long val, limit;
>   
>       /* Presence of REX.W is indicated by a bit higher than 7 set */
>       if (ctrl >> 8) {
> -        val = abs1((int64_t)env->regs[reg]);
> +        val = (target_long)env->regs[reg];
>       } else {
> -        val = abs1((int32_t)env->regs[reg]);
> +        val = (int32_t)env->regs[reg];
>       }
> -
>       if (ctrl & 1) {
> -        if (val > 8) {
> -            return 8;
> -        }
> +        limit = 8;
>       } else {
> -        if (val > 16) {
> -            return 16;
> -        }
> +        limit = 16;
>       }
> -    return val;
> +    if ((val > limit) || (val < -limit)) {
> +        return limit;
> +    }
> +    return abs1(val);
>   }
>   
>   static inline int pcmp_ilen(Reg *r, uint8_t ctrl)



  parent reply	other threads:[~2022-04-27  7:03 UTC|newest]

Thread overview: 67+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-04-18 17:39 [PATCH 0/3] AVX guest implementation Paul Brook
2022-04-18 17:39 ` [PATCH 1/4] Add AVX_EN hflag Paul Brook
2022-04-18 17:39 ` [PATCH 2/4] TCG support for AVX Paul Brook
2022-04-18 19:33   ` Peter Maydell
2022-04-18 19:45     ` Paul Brook
2022-04-18 19:50       ` Peter Maydell
2022-04-18 23:14       ` Richard Henderson
2022-04-20 14:19       ` Paolo Bonzini
2022-04-20 18:59         ` Paul Brook
2022-04-18 17:39 ` [PATCH 3/4] Enable all x86-64 cpu features in user mode Paul Brook
2022-04-18 17:39 ` [PATCH 4/4] AVX tests Paul Brook
2022-04-19 10:34   ` Alex Bennée
2022-04-24 22:01 ` [PATCH v2 01/42] i386: pcmpestr 64-bit sign extension bug Paul Brook
2022-04-25 15:50   ` Richard Henderson
2022-04-27  7:00   ` Paolo Bonzini [this message]
2022-04-24 22:01 ` [PATCH v2 02/42] i386: DPPS rounding fix Paul Brook
2022-04-25 16:09   ` Richard Henderson
2022-04-24 22:01 ` [PATCH v2 03/42] Add AVX_EN hflag Paul Brook
2022-04-25 17:27   ` Richard Henderson
2022-04-24 22:01 ` [PATCH v2 04/42] i386: Rework sse_op_table1 Paul Brook
2022-04-24 22:01 ` [PATCH v2 05/42] i386: Rework sse_op_table6/7 Paul Brook
2022-04-24 22:01 ` [PATCH v2 06/42] i386: Add CHECK_NO_VEX Paul Brook
2022-04-25 20:39   ` Richard Henderson
2022-04-25 20:41   ` Richard Henderson
2022-04-24 22:01 ` [PATCH v2 07/42] Enforce VEX encoding restrictions Paul Brook
2022-04-25 20:42   ` Richard Henderson
2022-04-25 21:00   ` Richard Henderson
2022-04-27  9:08   ` Paolo Bonzini
2022-04-24 22:01 ` [PATCH v2 08/42] i386: Add ZMM_OFFSET macro Paul Brook
2022-04-25 21:03   ` Richard Henderson
2022-04-24 22:01 ` [PATCH v2 09/42] i386: Helper macro for 256 bit AVX helpers Paul Brook
2022-04-24 22:01 ` [PATCH v2 10/42] i386: Rewrite vector shift helper Paul Brook
2022-04-25 21:33   ` Richard Henderson
2022-04-27  6:51     ` Paolo Bonzini
2022-04-24 22:01 ` [PATCH v2 11/42] i386: Rewrite simple integer vector helpers Paul Brook
2022-04-24 22:01 ` [PATCH v2 12/42] i386: Misc integer AVX helper prep Paul Brook
2022-04-24 22:01 ` [PATCH v2 13/42] i386: Destructive vector helpers for AVX Paul Brook
2022-04-27  6:53   ` Paolo Bonzini
2022-04-24 22:01 ` [PATCH v2 14/42] i386: Add size suffix to vector FP helpers Paul Brook
2022-04-24 22:01 ` [PATCH v2 15/42] i386: Floating point atithmetic helper AVX prep Paul Brook
2022-04-24 22:01 ` [PATCH v2 16/42] i386: Dot product AVX helper prep Paul Brook
2022-04-24 22:01 ` [PATCH v2 17/42] i386: Destructive FP helpers for AVX Paul Brook
2022-04-24 22:01 ` [PATCH v2 18/42] i386: Misc AVX helper prep Paul Brook
2022-04-24 22:01 ` [PATCH v2 19/42] i386: Rewrite blendv helpers Paul Brook
2022-04-24 22:01 ` [PATCH v2 20/42] i386: AVX pclmulqdq Paul Brook
2022-04-24 22:01 ` [PATCH v2 21/42] i386: AVX+AES helpers Paul Brook
2022-04-24 22:01 ` [PATCH v2 22/42] i386: Update ops_sse_helper.h ready for 256 bit AVX Paul Brook
2022-04-24 22:01 ` [PATCH v2 23/42] i386: AVX comparison helpers Paul Brook
2022-04-24 22:01 ` [PATCH v2 24/42] i386: Move 3DNOW decoder Paul Brook
2022-04-24 22:01 ` [PATCH v2 25/42] i386: VEX.V encodings (3 operand) Paul Brook
2022-04-24 22:01 ` [PATCH v2 26/42] i386: Utility function for 128 bit AVX Paul Brook
2022-04-24 22:01 ` [PATCH v2 27/42] i386: Translate 256 bit AVX instructions Paul Brook
2022-04-24 22:01 ` [PATCH v2 28/42] i386: Implement VZEROALL and VZEROUPPER Paul Brook
2022-04-24 22:01 ` [PATCH v2 29/42] i386: Implement VBROADCAST Paul Brook
2022-04-24 22:01 ` [PATCH v2 30/42] i386: Implement VPERMIL Paul Brook
2022-04-24 22:01 ` [PATCH v2 31/42] i386: Implement AVX variable shifts Paul Brook
2022-04-24 22:01 ` [PATCH v2 32/42] i386: Implement VTEST Paul Brook
2022-04-24 22:01 ` [PATCH v2 33/42] i386: Implement VMASKMOV Paul Brook
2022-04-24 22:01 ` [PATCH v2 34/42] i386: Implement VGATHER Paul Brook
2022-04-24 22:01 ` [PATCH v2 35/42] i386: Implement VPERM Paul Brook
2022-04-24 22:01 ` [PATCH v2 36/42] i386: Implement VINSERT128/VEXTRACT128 Paul Brook
2022-04-24 22:01 ` [PATCH v2 37/42] i386: Implement VBLENDV Paul Brook
2022-04-24 22:02 ` [PATCH v2 38/42] i386: Implement VPBLENDD Paul Brook
2022-04-24 22:02 ` [PATCH v2 39/42] i386: Enable AVX cpuid bits when using TCG Paul Brook
2022-04-24 22:02 ` [PATCH v2 40/42] Enable all x86-64 cpu features in user mode Paul Brook
2022-04-24 22:02 ` [PATCH v2 41/42] AVX tests Paul Brook
2022-04-24 22:02 ` [PATCH v2 42/42] i386: Add sha512-avx test Paul Brook

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=860ea92c-355a-07df-23b7-a46c7e84d9b2@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=eduardo@habkost.net \
    --cc=paul@nowt.org \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.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).