Linux SNPS ARC Archive on lore.kernel.org
 help / color / Atom feed
From: Vineet Gupta <Vineet.Gupta1@synopsys.com>
To: Adhemerval Zanella <adhemerval.zanella@linaro.org>,
	"libc-alpha@sourceware.org" <libc-alpha@sourceware.org>
Cc: "linux-snps-arc@lists.infradead.org"
	<linux-snps-arc@lists.infradead.org>
Subject: Re: [PATCH v6 06/13] ARC: hardware floating point support
Date: Fri, 29 May 2020 22:28:35 +0000
Message-ID: <91c2aba8-bff5-5ddc-7a95-c93c90883150@synopsys.com> (raw)
In-Reply-To: <de8975a0-bb03-8cec-be97-56f8c4e2a2f3@linaro.org>

On 5/29/20 7:12 AM, Adhemerval Zanella wrote:
> 
> 
> On 22/04/2020 22:41, Vineet Gupta via Libc-alpha wrote:
>> Signed-off-by: Vineet Gupta <vgupta@synopsys.com>

>> diff --git a/sysdeps/arc/fpu/e_sqrt.c b/sysdeps/arc/fpu/e_sqrt.c
>> +
>> +#include <math_private.h>
>> +#include <libm-alias-finite.h>
>> +
>> +double
>> +__ieee754_sqrt (double d)
>> +{
>> +  return __builtin_sqrt (d);
>> +}
>> +libm_alias_finite (__ieee754_sqrt, __sqrt)
> 
> I think it is better to extend to math-use-builtins.h for cover 
> e_sqrt{f} so we can also adapt to other architecutes.

Good idea.

Side question: what do s_ / e_ prefixes mean?
The suffix is clear s_roundf (float), s_round (double), s_roundl (128).

>> diff --git a/sysdeps/arc/fpu/e_sqrtf.c b/sysdeps/arc/fpu/e_sqrtf.c
>> new file mode 100644
>> index 000000000000..13008a4f45d6
>> --- /dev/null
>> +++ b/sysdeps/arc/fpu/e_sqrtf.c

>> +#include <math_private.h>
>> +#include <libm-alias-finite.h>
>> +
>> +float
>> +__ieee754_sqrtf (float s)
>> +{
>> +  return __builtin_sqrtf (s);
>> +}
>> +libm_alias_finite (__ieee754_sqrtf, __sqrtf)
> 
> Same as before.

OK.

>> diff --git a/sysdeps/arc/fpu/s_fma.c b/sysdeps/arc/fpu/s_fma.c

>> +
>> +#include <math.h>
>> +#include <libm-alias-double.h>
>> +
>> +double
>> +__fma (double x, double y, double z)
>> +{
>> +  return __builtin_fma (x, y, z);
>> +}
>> +
>> +libm_alias_double (__fma, fma)
> 
> Same as for sqrt{f}.

OK.


>> diff --git a/sysdeps/arc/fpu/s_fmaf.c b/sysdeps/arc/fpu/s_fmaf.c
>> new file mode 100644
>> index 000000000000..544f32e27aec
>> --- /dev/null
>> +++ b/sysdeps/arc/fpu/s_fmaf.c
>> @@ -0,0 +1,28 @@
>> +/* Copyright (C) 2011-2020 Free Software Foundation, Inc.
>> +
>> +   This file is part of the GNU C Library.
>> +
>> +   The GNU C Library is free software; you can redistribute it and/or
>> +   modify it under the terms of the GNU Lesser General Public License as
>> +   published by the Free Software Foundation; either version 2.1 of the
>> +   License, or (at your option) any later version.
>> +
>> +   The GNU C Library is distributed in the hope that it will be useful,
>> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
>> +   Lesser General Public License for more details.
>> +
>> +   You should have received a copy of the GNU Lesser General Public
>> +   License along with the GNU C Library; if not, see
>> +   <https://www.gnu.org/licenses/>.  */
>> +
>> +#include <math.h>
>> +#include <libm-alias-float.h>
>> +
>> +float
>> +__fmaf (float x, float y, float z)
>> +{
>> +  return __builtin_fmaf (x, y, z);
>> +}
>> +
>> +libm_alias_float (__fma, fma)
>> diff --git a/sysdeps/arc/fpu_control.h b/sysdeps/arc/fpu_control.h
>> new file mode 100644
>> index 000000000000..121656ff7d8c
>> --- /dev/null
>> +++ b/sysdeps/arc/fpu_control.h
>> @@ -0,0 +1,106 @@
>> +/* FPU control word bits.  ARC version.
>> +   Copyright (C) 2018-2020 Free Software Foundation, Inc.
>> +   This file is part of the GNU C Library.
>> +
>> +   The GNU C Library is free software; you can redistribute it and/or
>> +   modify it under the terms of the GNU Lesser General Public
>> +   License as published by the Free Software Foundation; either
>> +   version 2.1 of the License, or (at your option) any later version.
>> +
>> +   The GNU C Library is distributed in the hope that it will be useful,
>> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
>> +   Lesser General Public License for more details.
>> +
>> +   You should have received a copy of the GNU Lesser General Public
>> +   License along with the GNU C Library.  If not, see
>> +   <https://www.gnu.org/licenses/>.  */
>> +
>> +#ifndef _FPU_CONTROL_H
>> +#define _FPU_CONTROL_H
>> +
>> +/* ARC FPU control register bits.
>> +
>> +  [  0] -> IVE: Enable invalid operation exception.
>> +           if 0, soft exception: status register IV flag set.
>> +           if 1, hardware exception trap (not supported in Linux yet).
>> +  [  1] -> DZE: Enable division by zero exception.
>> +           if 0, soft exception: status register IV flag set.
>> +           if 1, hardware exception: (not supported in Linux yet).
>> +  [9:8] -> RM: Rounding Mode:
>> +           00 - Rounding toward zero.
>> +           01 - Rounding to nearest (default).
>> +           10 - Rounding (up) toward plus infinity.
>> +           11 - Rounding (down)toward minus infinity.
>> +
>> +   ARC FPU status register bits.
>> +
>> +   [ 0]  -> IV: flag invalid operation.
>> +   [ 1]  -> DZ: flag division by zero.
>> +   [ 2]  -> OV: flag Overflow operation.
>> +   [ 3]  -> UV: flag Underflow operation.
>> +   [ 4]  -> IX: flag Inexact operation.
>> +   [31]  -> FWE: Flag Write Enable.
>> +            If 1, above flags writable explicitly (clearing),
>> +            else IoW and only writable indirectly via bits [12:7].  */
>> +
>> +#include <features.h>
>> +
>> +#if !defined(__ARC_FPU_SP__) &&  !defined(__ARC_FPU_DP__)
>> +
>> +# define _FPU_RESERVED 0xffffffff
>> +# define _FPU_DEFAULT  0x00000000
>> +typedef unsigned int fpu_control_t;
>> +# define _FPU_GETCW(cw) (cw) = 0
>> +# define _FPU_SETCW(cw) (void) (cw)
>> +# define _FPU_GETS(cw) (cw) = 0
>> +# define _FPU_SETS(cw) (void) (cw)
>> +extern fpu_control_t __fpu_control;
>> +
>> +#else
>> +
>> +#define _FPU_RESERVED		0
>> +
>> +/* The fdlibm code requires strict IEEE double precision arithmetic,
>> +   and no interrupts for exceptions, rounding to nearest.
>> +   So only RM set to b'01.  */
>> +# define _FPU_DEFAULT		0x00000100
>> +
>> +/* Actually default needs to have FWE bit as 1 but that is already
>> +   ingrained into _FPU_SETS macro below.  */
>> +#define  _FPU_FPSR_DEFAULT	0x00000000
>> +
>> +#define __FPU_RND_SHIFT		8
>> +#define __FPU_RND_MASK		0x3
>> +
>> +/* Type of the control word.  */
>> +typedef unsigned int fpu_control_t;
>> +
>> +/* Macros for accessing the hardware control word.  */
>> +#  define _FPU_GETCW(cw) __asm__ volatile ("lr %0, [0x300]" : "=r" (cw))
>> +#  define _FPU_SETCW(cw) __asm__ volatile ("sr %0, [0x300]" : : "r" (cw))
>> +
>> +/*  Macros for accessing the hardware status word.
>> +    FWE bit is special as it controls if actual status bits could be wrritten
>> +    explicitly (other than FPU instructions). We handle it here to keep the
>> +    callers agnostic of it:
>> +      - clear it out when reporting status bits
>> +      - always set it when changing status bits.  */
>> +#  define _FPU_GETS(cw)				\
>> +    __asm__ volatile ("lr   %0, [0x301]	\r\n" 	\
>> +                      "bclr %0, %0, 31	\r\n" 	\
>> +                      : "=r" (cw))
>> +
>> +#  define _FPU_SETS(cw)				\
>> +    do {					\
>> +      unsigned int __tmp = 0x80000000 | (cw);	\
>> +      __asm__ volatile ("sr  %0, [0x301] \r\n" 	\
>> +                        : : "r" (__tmp));	\
>> +    } while (0)
>> +
> 
> Although this code follow other architectures, I think it woudl be better
> to move forward a macro that emulates function calls and use proper
> static inline function instead for _FPU_* (as for get-rounding-mode.h).

OK. do you have a preference for names, existing upper case names OK ?
_______________________________________________
linux-snps-arc mailing list
linux-snps-arc@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-snps-arc

  reply index

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-23  1:41 [PATCH v6 00/13] glibc port to ARC processors Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 01/13] ARC: ABI Implementation Vineet Gupta
     [not found]   ` <88508d10-2d29-026a-bb54-ad607154ab87@linaro.org>
2020-05-27 22:15     ` Vineet Gupta
2020-05-29 13:56       ` Adhemerval Zanella
     [not found]     ` <a56a35d4-3e9e-9a88-4be5-8553d5f11ad3@synopsys.com>
     [not found]       ` <87mu5jxkv7.fsf@oldenburg2.str.redhat.com>
2020-06-04 19:01         ` Vineet Gupta
2020-06-04 23:56           ` Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 02/13] ARC: startup and dynamic linking code Vineet Gupta
     [not found]   ` <17957ee6-2bc1-f43b-f184-f0703ba2765f@linaro.org>
2020-05-28  1:14     ` Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 03/13] ARC: Thread Local Storage support Vineet Gupta
     [not found]   ` <4f7a67fb-6f96-57e6-b827-d1ab5dd6733f@linaro.org>
2020-05-28  1:36     ` Vineet Gupta
2020-06-01 18:53       ` Adhemerval Zanella
2020-04-23  1:41 ` [PATCH v6 04/13] ARC: Atomics and Locking primitives Vineet Gupta
     [not found]   ` <03f4a9b3-b1ca-90fa-0b6a-609a3135267d@linaro.org>
2020-04-24  7:23     ` Vineet Gupta
     [not found]     ` <20200427215938.14136-1-vgupta@synopsys.com>
2020-04-27 22:13       ` [PATCH] semaphore: consolidate arch headers into a generic one Vineet Gupta
     [not found]       ` <ac93c301-36d3-b20a-d31c-50c1f3264c68@linaro.org>
2020-05-05 22:59         ` Vineet Gupta
2020-05-08 13:32           ` Adhemerval Zanella
2020-04-23  1:41 ` [PATCH v6 05/13] ARC: math soft float support Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 06/13] ARC: hardware floating point support Vineet Gupta
2020-05-29 14:12   ` Adhemerval Zanella
2020-05-29 22:28     ` Vineet Gupta [this message]
2020-05-29 23:50       ` Vineet Gupta
2020-06-02 17:51         ` Joseph Myers
     [not found]         ` <07887c48-7e07-9f89-035d-3f336a16f2da@synopsys.com>
2020-06-02 18:13           ` static inline math functions (was Re: [PATCH v6 06/13] ARC: hardware floating point support) Joseph Myers
2020-06-02 18:35             ` Adhemerval Zanella
2020-06-05  4:44         ` [PATCH v6 06/13] ARC: hardware floating point support Vineet Gupta
2020-06-05 17:22           ` Adhemerval Zanella
2020-06-02 17:48       ` Joseph Myers
2020-04-23  1:41 ` [PATCH v6 07/13] ARC: Linux Syscall Interface Vineet Gupta
2020-05-29 16:49   ` Adhemerval Zanella
2020-05-30  2:02     ` Vineet Gupta
2020-06-03 19:46     ` Vineet Gupta
2020-06-03 20:04       ` Adhemerval Zanella
2020-06-03 20:17         ` Vineet Gupta
2020-06-04 11:06           ` Adhemerval Zanella
2020-04-23  1:41 ` [PATCH v6 08/13] ARC: Linux ABI Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 09/13] ARC: Linux Startup and Dynamic Loading Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 10/13] ARC: ABI lists Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 11/13] ARC: Build Infrastructure Vineet Gupta
2020-06-03 19:58   ` Adhemerval Zanella
2020-06-04 15:25     ` Vineet Gupta
2020-06-04 17:05       ` Adhemerval Zanella
2020-06-08  4:18         ` Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 12/13] build-many-glibcs.py: Enable ARC builds Vineet Gupta
2020-04-23  1:41 ` [PATCH v6 13/13] Documentation for ARC port Vineet Gupta
2020-05-04 21:21 ` [PATCH v6 00/13] glibc port to ARC processors Vineet Gupta
2020-05-15  0:45   ` Vineet Gupta
2020-05-27  1:49     ` Vineet Gupta
     [not found]       ` <d7f1176c-87c6-90c6-161c-4705a47837ea@linaro.org>
2020-05-27 18:38         ` Vineet Gupta

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=91c2aba8-bff5-5ddc-7a95-c93c90883150@synopsys.com \
    --to=vineet.gupta1@synopsys.com \
    --cc=adhemerval.zanella@linaro.org \
    --cc=libc-alpha@sourceware.org \
    --cc=linux-snps-arc@lists.infradead.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

Linux SNPS ARC Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-snps-arc/0 linux-snps-arc/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-snps-arc linux-snps-arc/ https://lore.kernel.org/linux-snps-arc \
		linux-snps-arc@lists.infradead.org
	public-inbox-index linux-snps-arc

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.infradead.lists.linux-snps-arc


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git