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 [thread overview]
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
next prev parent reply other threads:[~2020-05-29 22:29 UTC|newest]
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
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).