From: caiyinyu <caiyinyu@loongson.cn>
To: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>,
libc-alpha@sourceware.org
Cc: joseph_myers@mentor.com
Subject: Re: [PATCH v6 07/13] LoongArch: Atomic and Locking Routines
Date: Fri, 15 Jul 2022 09:46:56 +0800 [thread overview]
Message-ID: <89885b7a-521d-2688-cd01-70b59dab9546@loongson.cn> (raw)
In-Reply-To: <9cc5ba1e-97ee-3ce6-7634-8fcfaec34821@linaro.org>
+
+/* Miscellaneous. */
+
+#define asm_amo(which, mem, value) \
+ ({ \
+ __atomic_check_size (mem); \
+ typeof (*mem) __tmp; \
+ if (sizeof (__tmp) == 4) \
+ asm volatile(which ".w" \
+ "\t%0, %z2, %1" \
+ : "=&r"(__tmp), "+ZB"(*(mem)) \
+ : "rJ"(value)); \
+ else if (sizeof (__tmp) == 8) \
It does not have the case of !__HAVE_64B_ATOMICS, but since only 64-bit
is support I don think this will be troublesome. Also, the idea is to
remove such atomic macros on next release.
*Removed.**
*
*>>>>>>>>>>>>>>>>>>>>>>>*
**
*diff --git a/sysdeps/unix/sysv/linux/loongarch/atomic-machine.h
b/sysdeps/unix/sysv/linux/loongarch/atomic-machine.h**
**index b51463915e..d1b8f1c11b 100644**
**--- a/sysdeps/unix/sysv/linux/loongarch/atomic-machine.h**
**+++ b/sysdeps/unix/sysv/linux/loongarch/atomic-machine.h**
**@@ -144,38 +144,4 @@**
** __atomic_val_bysize (__arch_exchange_and_add, int, mem, value, \**
** __ATOMIC_RELEASE)**
****
**-/* Miscellaneous. */**
**-**
**-#define asm_amo(which, mem, value) \**
**- ({ \**
**- __atomic_check_size (mem); \**
**- typeof (*mem) __tmp; \**
**- if (sizeof (__tmp) == 4) \**
**- asm volatile (which ".w" \**
**- "\t%0, %z2, %1" \**
**- : "=&r"(__tmp), "+ZB"(*(mem)) \**
**- : "rJ"(value)); \**
**- else if (sizeof (__tmp) == 8) \**
**- asm volatile (which ".d" \**
**- "\t%0, %z2, %1" \**
**- : "=&r"(__tmp), "+ZB"(*(mem)) \**
**- : "rJ"(value)); \**
**- else \**
**- abort (); \**
**- __tmp; \**
**- })**
**-**
**-#define atomic_max(mem, value) asm_amo ("ammax_db", mem, value)**
**-#define atomic_min(mem, value) asm_amo ("ammin_db", mem, value)**
**-**
**-#define atomic_bit_test_set(mem, bit) \**
**- ({ \**
**- typeof (*mem) __mask = (typeof (*mem)) 1 << (bit); \**
**- asm_amo ("amor_db", mem, __mask) & __mask; \**
**- })**
**-**
**-#define catomic_exchange_and_add(mem, value) \**
**- atomic_exchange_and_add (mem, value)**
**-#define catomic_max(mem, value) atomic_max (mem, value)**
**-**
** #endif /* bits/atomic.h */**
*
**
*<<<<<<<<<<<<<<<<<<<<<<<*
在 2022/7/13 下午9:53, Adhemerval Zanella Netto 写道:
> Although it would be to a more streamlined macros, it follows current practice.
> Only a small nit below.
>
> Reviewed-by: Adhemerval Zanella<adhemerval.zanella@linaro.org>
>
> On 08/07/22 03:52, caiyinyu wrote:
>> ---
>> .../sysv/linux/loongarch/atomic-machine.h | 181 ++++++++++++++++++
>> 1 file changed, 181 insertions(+)
>> create mode 100644 sysdeps/unix/sysv/linux/loongarch/atomic-machine.h
>>
>> diff --git a/sysdeps/unix/sysv/linux/loongarch/atomic-machine.h b/sysdeps/unix/sysv/linux/loongarch/atomic-machine.h
>> new file mode 100644
>> index 0000000000..60db25587e
>> --- /dev/null
>> +++ b/sysdeps/unix/sysv/linux/loongarch/atomic-machine.h
>> @@ -0,0 +1,181 @@
>> +/* Atomic operations.
>> + Copyright (C) 2022 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 _LINUX_LOONGARCH_BITS_ATOMIC_H
>> +#define _LINUX_LOONGARCH_BITS_ATOMIC_H 1
>> +
>> +#define atomic_full_barrier() __sync_synchronize ()
>> +
>> +#define __HAVE_64B_ATOMICS (__loongarch_grlen >=64)
> Missing space after >=
>
>> +#define USE_ATOMIC_COMPILER_BUILTINS 1
>> +#define ATOMIC_EXCHANGE_USES_CAS 0
>> +
>> +/* Compare and exchange.
>> + For all "bool" routines, we return FALSE if exchange succesful. */
>> +
>> +#define __arch_compare_and_exchange_bool_8_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + !__atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + })
>> +
>> +#define __arch_compare_and_exchange_bool_16_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + !__atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + })
>> +
>> +#define __arch_compare_and_exchange_bool_32_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + !__atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + })
>> +
>> +#define __arch_compare_and_exchange_bool_64_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + !__atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + })
>> +
>> +#define __arch_compare_and_exchange_val_8_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + __atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + __oldval; \
>> + })
>> +
>> +#define __arch_compare_and_exchange_val_16_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + __atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + __oldval; \
>> + })
>> +
>> +#define __arch_compare_and_exchange_val_32_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + __atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + __oldval; \
>> + })
>> +
>> +#define __arch_compare_and_exchange_val_64_int(mem, newval, oldval, model) \
>> + ({ \
>> + typeof (*mem) __oldval = (oldval); \
>> + __atomic_compare_exchange_n (mem, (void *) &__oldval, newval, 0, model, \
>> + __ATOMIC_RELAXED); \
>> + __oldval; \
>> + })
>> +
>> +/* Atomic compare and exchange. */
>> +
>> +#define atomic_compare_and_exchange_bool_acq(mem, new, old) \
>> + __atomic_bool_bysize (__arch_compare_and_exchange_bool, int, mem, new, old, \
>> + __ATOMIC_ACQUIRE)
>> +
>> +#define atomic_compare_and_exchange_val_acq(mem, new, old) \
>> + __atomic_val_bysize (__arch_compare_and_exchange_val, int, mem, new, old, \
>> + __ATOMIC_ACQUIRE)
>> +
>> +#define atomic_compare_and_exchange_val_rel(mem, new, old) \
>> + __atomic_val_bysize (__arch_compare_and_exchange_val, int, mem, new, old, \
>> + __ATOMIC_RELEASE)
>> +
>> +/* Atomic exchange (without compare). */
>> +
>> +#define __arch_exchange_8_int(mem, newval, model) \
>> + __atomic_exchange_n (mem, newval, model)
>> +
>> +#define __arch_exchange_16_int(mem, newval, model) \
>> + __atomic_exchange_n (mem, newval, model)
>> +
>> +#define __arch_exchange_32_int(mem, newval, model) \
>> + __atomic_exchange_n (mem, newval, model)
>> +
>> +#define __arch_exchange_64_int(mem, newval, model) \
>> + __atomic_exchange_n (mem, newval, model)
>> +
>> +#define atomic_exchange_acq(mem, value) \
>> + __atomic_val_bysize (__arch_exchange, int, mem, value, __ATOMIC_ACQUIRE)
>> +
>> +#define atomic_exchange_rel(mem, value) \
>> + __atomic_val_bysize (__arch_exchange, int, mem, value, __ATOMIC_RELEASE)
>> +
>> +/* Atomically add value and return the previous (unincremented) value. */
>> +
>> +#define __arch_exchange_and_add_8_int(mem, value, model) \
>> + __atomic_fetch_add (mem, value, model)
>> +
>> +#define __arch_exchange_and_add_16_int(mem, value, model) \
>> + __atomic_fetch_add (mem, value, model)
>> +
>> +#define __arch_exchange_and_add_32_int(mem, value, model) \
>> + __atomic_fetch_add (mem, value, model)
>> +
>> +#define __arch_exchange_and_add_64_int(mem, value, model) \
>> + __atomic_fetch_add (mem, value, model)
>> +
>> +#define atomic_exchange_and_add_acq(mem, value) \
>> + __atomic_val_bysize (__arch_exchange_and_add, int, mem, value, \
>> + __ATOMIC_ACQUIRE)
>> +
>> +#define atomic_exchange_and_add_rel(mem, value) \
>> + __atomic_val_bysize (__arch_exchange_and_add, int, mem, value, \
>> + __ATOMIC_RELEASE)
>> +
>> +/* Miscellaneous. */
>> +
>> +#define asm_amo(which, mem, value) \
>> + ({ \
>> + __atomic_check_size (mem); \
>> + typeof (*mem) __tmp; \
>> + if (sizeof (__tmp) == 4) \
>> + asm volatile(which ".w" \
>> + "\t%0, %z2, %1" \
>> + : "=&r"(__tmp), "+ZB"(*(mem)) \
>> + : "rJ"(value)); \
>> + else if (sizeof (__tmp) == 8) \
> It does not have the case of !__HAVE_64B_ATOMICS, but since only 64-bit
> is support I don think this will be troublesome. Also, the idea is to
> remove such atomic macros on next release.
>
>> + asm volatile(which ".d" \
>> + "\t%0, %z2, %1" \
>> + : "=&r"(__tmp), "+ZB"(*(mem)) \
>> + : "rJ"(value)); \
>> + else \
>> + abort (); \
>> + __tmp; \
>> + })
>> +
>> +#define atomic_max(mem, value) asm_amo ("ammax_db", mem, value)
>> +#define atomic_min(mem, value) asm_amo ("ammin_db", mem, value)
>> +
>> +#define atomic_bit_test_set(mem, bit) \
>> + ({ \
>> + typeof (*mem) __mask = (typeof (*mem)) 1 << (bit); \
>> + asm_amo ("amor_db", mem, __mask) & __mask; \
>> + })
>> +
>> +#define catomic_exchange_and_add(mem, value) \
>> + atomic_exchange_and_add (mem, value)
>> +#define catomic_max(mem, value) atomic_max (mem, value)
>> +
>> +#endif /* bits/atomic.h */
next prev parent reply other threads:[~2022-07-15 1:47 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-07-08 6:52 [PATCH v6 00/13] GLIBC LoongArch PATCHES caiyinyu
2022-07-08 6:52 ` [PATCH v6 01/13] LoongArch: Update NEWS and README for the LoongArch port caiyinyu
2022-07-12 11:06 ` Adhemerval Zanella Netto
2022-07-08 6:52 ` [PATCH v6 02/13] LoongArch: Add LoongArch entries to config.h.in caiyinyu
2022-07-12 18:51 ` Adhemerval Zanella Netto
2022-07-08 6:52 ` [PATCH v6 03/13] LoongArch: Add relocations and ELF flags to elf.h and scripts/glibcelf.py caiyinyu
2022-07-12 19:27 ` Adhemerval Zanella Netto
2022-07-08 6:52 ` [PATCH v6 04/13] LoongArch: ABI Implementation caiyinyu
2022-07-12 20:39 ` Adhemerval Zanella Netto
2022-07-15 1:46 ` caiyinyu
2022-07-08 6:52 ` [PATCH v6 05/13] LoongArch: Thread-Local Storage Support caiyinyu
2022-07-13 13:10 ` Adhemerval Zanella Netto
2022-07-08 6:52 ` [PATCH v6 06/13] LoongArch: Generic <math.h> and soft-fp Routines caiyinyu
2022-07-13 13:25 ` Adhemerval Zanella Netto
2022-07-08 6:52 ` [PATCH v6 07/13] LoongArch: Atomic and Locking Routines caiyinyu
2022-07-13 13:53 ` Adhemerval Zanella Netto
2022-07-15 1:46 ` caiyinyu [this message]
2022-07-08 6:52 ` [PATCH v6 08/13] LoongArch: Linux Syscall Interface caiyinyu
2022-07-13 16:19 ` Adhemerval Zanella Netto
2023-09-13 11:05 ` caiyinyu
2022-07-08 6:52 ` [PATCH v6 09/13] LoongArch: Linux ABI caiyinyu
2022-07-13 16:35 ` Adhemerval Zanella Netto
2022-07-15 1:48 ` caiyinyu
2022-07-08 6:52 ` [PATCH v6 10/13] LoongArch: Add ABI Lists caiyinyu
2022-07-13 17:12 ` Adhemerval Zanella Netto
2022-07-08 6:52 ` [PATCH v6 11/13] LoongArch: Build Infastructure caiyinyu
2022-07-13 3:07 ` Xi Ruoyao
2022-07-13 3:43 ` WANG Xuerui
2022-07-13 7:51 ` caiyinyu
2022-07-13 17:48 ` Adhemerval Zanella Netto
2022-07-15 1:49 ` caiyinyu
2022-07-08 6:52 ` [PATCH v6 12/13] LoongArch: Hard Float Support caiyinyu
2022-07-13 19:22 ` Adhemerval Zanella Netto
2022-07-14 13:00 ` caiyinyu
2022-07-08 7:13 ` [PATCH v6 00/13] GLIBC LoongArch PATCHES Xi Ruoyao
2022-07-08 7:15 ` Xi Ruoyao
2022-07-13 19:55 ` Adhemerval Zanella Netto
2022-07-14 11:33 ` Xi Ruoyao
2022-07-14 12:11 ` Adhemerval Zanella Netto
2022-07-18 13:54 ` Carlos O'Donell
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=89885b7a-521d-2688-cd01-70b59dab9546@loongson.cn \
--to=caiyinyu@loongson.cn \
--cc=adhemerval.zanella@linaro.org \
--cc=joseph_myers@mentor.com \
--cc=libc-alpha@sourceware.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 read-only IMAP folder(s) and NNTP newsgroup(s).