[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 2/6] softfloat: fix float{32, 64}_maybe_silence_
From: |
Aurelien Jarno |
Subject: |
Re: [Qemu-devel] [PATCH 2/6] softfloat: fix float{32, 64}_maybe_silence_nan() for MIPS |
Date: |
Mon, 03 Jan 2011 16:24:50 +0100 |
User-agent: |
Mozilla-Thunderbird 2.0.0.24 (X11/20100329) |
Peter Maydell a écrit :
> On 3 January 2011 14:34, Aurelien Jarno <address@hidden> wrote:
>> On targets that define sNaN with the sNaN bit as one, simply clearing
>> this bit may correspond to an infinite value.
>>
>> Convert it to a default NaN if SNAN_BIT_IS_ONE, as it corresponds to
>> the MIPS implementation, the only emulated CPU with SNAN_BIT_IS_ONE.
>> When other CPU of this type are added, this might be updated to include
>> more cases.
>
> This patch doesn't apply to master:
>
>> Signed-off-by: Aurelien Jarno <address@hidden>
>> ---
>> fpu/softfloat-specialize.h | 12 ++++++------
>> 1 files changed, 6 insertions(+), 6 deletions(-)
>>
>> diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
>> index f23bd6a..31481e7 100644
>> --- a/fpu/softfloat-specialize.h
>> +++ b/fpu/softfloat-specialize.h
>> @@ -107,13 +107,13 @@ int float32_is_signaling_nan( float32 a_ )
>> float32 float32_maybe_silence_nan( float32 a_ )
>> {
>> if (float32_is_signaling_nan(a_)) {
>> - bits32 a = float32_val(a_);
>
> ...on master this line is
> uint32_t a = float32_val(a_);
>
> (different type) so the patch doesn't apply.
Oops, yes, my patch series should have started by a patch fixing types,
but i made a mistake selecting the commits to send. Will fix that in a v2.
> Other than that, looks OK. I think I'd like a comment somewhere
> along the lines of
> /* Rules for silencing a signaling NaN are target-specific. Typically
> * targets with !SNAN_BIT_IS_ONE use the rule that the NaN
> * is silenced by setting the bit. Targets where SNAN_BIT_IS_ONE
> * must do something more complicated, because clearing the
> * bit might turn a NaN into an infinity. This code is correct for
> * MIPS but new targets might need something different.
> */
>
> Or you could have the #ifdefs be on TARGET_whatever so
> that it's clear (because it won't compile) that adding a new
> TARGET_FOO means you have to check behaviour in this
> area. But I don't feel very strongly about that.
>
Ok, thanks for the review, will fix that.
--
Aurelien Jarno GPG: 1024D/F1BCDB73
address@hidden http://www.aurel32.net
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, (continued)
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Aurelien Jarno, 2011/01/06
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Peter Maydell, 2011/01/06
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Aurelien Jarno, 2011/01/06
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Peter Maydell, 2011/01/06
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Aurelien Jarno, 2011/01/06
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Nathan Froyd, 2011/01/06
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Andreas Färber, 2011/01/06
- Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code, Aurelien Jarno, 2011/01/06
[Qemu-devel] [PATCH 2/6] softfloat: fix float{32, 64}_maybe_silence_nan() for MIPS, Aurelien Jarno, 2011/01/03
[Qemu-devel] [PATCH 4/6] softfloat: use float{32, 64, x80, 128}_maybe_silence_nan(), Aurelien Jarno, 2011/01/03
[Qemu-devel] [PATCH 3/6] softfloat: add float{x80, 128}_maybe_silence_nan(), Aurelien Jarno, 2011/01/03
[Qemu-devel] [PATCH 5/6] target-mips: Implement correct NaN propagation rules, Aurelien Jarno, 2011/01/03
[Qemu-devel] [PATCH 6/6] target-ppc: Implement correct NaN propagation rules, Aurelien Jarno, 2011/01/03