Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Tue, 15 Nov 2022 15:17:54 +0800
From: 王洪亮 <wanghongliang@...ngson.cn>
To: musl@...ts.openwall.com
Cc: WANG Xuerui <i@...0n.name>
Subject: add loongarch64 port v6.

Hi,

I published loongarch64 port v6 in
https://github.com/loongson/musl/tree/loongarch-v3.0

Fixed the following problems:
1.Fixed some inappropriate comments in elf.h.
2.Fixed spaces and tabs used.
3.Removed crt/loongarch64/crti.s and crt/loongarch64/crtn.s.
4.Fixed the problems mentioned below email.

There are also some unmodified problems as follows:
1.Whether we still implement __clone() within __NR_clone3
syscall in LoongArch?
In latest kernel,function clone3_stack_valid(),the condition
of (kargs->stack_size == 0) is still valid, call __clone()
(within __NR_clone3 syscall) is fail.
I restore to the __NR_clone for libc-test OK temporary.

2.I don't remove __syscall7,I understand keep it without negative effects.
maybe used in the future.glibc also realize __syscall7 in LoongArch.

3.I don't remove arch/loongarch64/bits/ptrace.h,I'm not sure if it will be
used in the future.the most architectures support it.

4.The definition of extcontext[] in mcontext_t in musl is not consistent
with kernel,Because [0] is not valid C. it is not standard in musl,so use
extcontext[] instead of extcontext[0].This question has been discussed in
previous email named [musl] add loongarch64 port v3 in May.

Please check first,I will prepare a physical execution environment for 
run checks.

I'm Looking forward to your reply,thanks.

Hongliang Wang.


在 2022/10/12 下午4:22, WANG Xuerui 写道:
> Hi,
>
> Sorry for the delay in reviewing this. I haven't test-run the changes 
> yet but I've gone through most of it and left some comments below.
>
> Besides, MaskRay has hinted in a related LLVM patch [1] that the musl 
> maintainers may want to review the current LoongArch spec where target 
> triples are (more-or-less tentatively) defined and including musl ones 
> ([2], see Table 12). Maintainers, opinions?
>
> [1]: https://reviews.llvm.org/D135751
> [2]: 
> https://loongson.github.io/LoongArch-Documentation/LoongArch-toolchain-conventions-EN.html
>
> On 2022/8/3 16:15, 王洪亮 wrote:
>> From 7c08ecde9237a0e5de182829a947f6b02cb52cf5 Mon Sep 17 00:00:00 2001
>> From: wanghongliang <wanghongliang@...ngson.cn>
>> Date: Wed, 3 Aug 2022 02:01:23 +0800
>> Subject: [PATCH] add loongarch64 port v5.
>
> Patch revision should be inside the PATCH label, because it's not 
> supposed to be
> a part of the commit message when merged.
>
> You should generate the patch like `git format-patch -v5 master` and 
> ideally
> don't send it as an attachment, it makes inline replies cumbersome 
> (e.g. I can't
> review and compose the mail on my phone, which I often do).
>
>>
>> Author: xiaojuanZhai <zhaixiaojuan@...ngson.cn>
>> Author: meidanLi <limeidan@...ngson.cn>
>> Author: guoqiChen <chenguoqi@...ngson.cn>
>> Author: xiaolinZhao <zhaoxiaolin@...ngson.cn>
>> Author: Fanpeng <fanpeng@...ngson.cn>
>> Author: jiantaoShan <shanjiantao@...ngson.cn>
>> Author: xuhuiQiang <qiangxuhui@...ngson.cn>
>> Author: jingyunHua <huajingyun@...ngson.cn>
>> Author: Liuxue <liuxue@...ngson.cn>
>> Author: HongliangWang <wanghongliang@...ngson.cn>
>>
>> Signed-off-by: wanghongliang <wanghongliang@...ngson.cn>
>
> As I have pointed out earlier, Chinese names are NOT camelCase, please 
> use the
> native form of "Lastname Firstname" (or reversing the two) but never 
> the current
> form.
>
>> ---
>>  arch/loongarch64/atomic_arch.h             |  53 ++++
>>  arch/loongarch64/bits/alltypes.h.in        |  18 ++
>>  arch/loongarch64/bits/fenv.h               |  20 ++
>>  arch/loongarch64/bits/float.h              |  16 ++
>>  arch/loongarch64/bits/posix.h              |   2 +
>>  arch/loongarch64/bits/ptrace.h             |   4 +
>>  arch/loongarch64/bits/reg.h                |   2 +
>>  arch/loongarch64/bits/setjmp.h             |   1 +
>>  arch/loongarch64/bits/signal.h             |  80 ++++++
>>  arch/loongarch64/bits/stat.h               |  18 ++
>>  arch/loongarch64/bits/stdint.h             |  20 ++
>>  arch/loongarch64/bits/syscall.h.in         | 303 +++++++++++++++++++++
>>  arch/loongarch64/bits/user.h               |   5 +
>>  arch/loongarch64/crt_arch.h                |  13 +
>>  arch/loongarch64/pthread_arch.h            |  13 +
>>  arch/loongarch64/reloc.h                   |  27 ++
>>  arch/loongarch64/syscall_arch.h            | 137 ++++++++++
>>  configure                                  |   1 +
>>  crt/loongarch64/crti.s                     |  15 +
>>  crt/loongarch64/crtn.s                     |  12 +
>>  include/elf.h                              |  65 +++++
>>  src/fenv/loongarch64/fenv.S                |  72 +++++
>>  src/ldso/loongarch64/dlsym.s               |   7 +
>>  src/setjmp/loongarch64/longjmp.S           |  37 +++
>>  src/setjmp/loongarch64/setjmp.S            |  34 +++
>>  src/signal/loongarch64/restore.s           |  10 +
>>  src/signal/loongarch64/sigsetjmp.s         |  29 ++
>>  src/thread/loongarch64/__set_thread_area.s |   7 +
>>  src/thread/loongarch64/__unmapself.s       |   7 +
>>  src/thread/loongarch64/clone.s             |  47 ++++
>>  src/thread/loongarch64/syscall_cp.s        |  29 ++
>>  31 files changed, 1104 insertions(+)
>>  create mode 100644 arch/loongarch64/atomic_arch.h
>>  create mode 100644 arch/loongarch64/bits/alltypes.h.in
>>  create mode 100644 arch/loongarch64/bits/fenv.h
>>  create mode 100644 arch/loongarch64/bits/float.h
>>  create mode 100644 arch/loongarch64/bits/posix.h
>>  create mode 100644 arch/loongarch64/bits/ptrace.h
>>  create mode 100644 arch/loongarch64/bits/reg.h
>>  create mode 100644 arch/loongarch64/bits/setjmp.h
>>  create mode 100644 arch/loongarch64/bits/signal.h
>>  create mode 100644 arch/loongarch64/bits/stat.h
>>  create mode 100644 arch/loongarch64/bits/stdint.h
>>  create mode 100644 arch/loongarch64/bits/syscall.h.in
>>  create mode 100644 arch/loongarch64/bits/user.h
>>  create mode 100644 arch/loongarch64/crt_arch.h
>>  create mode 100644 arch/loongarch64/pthread_arch.h
>>  create mode 100644 arch/loongarch64/reloc.h
>>  create mode 100644 arch/loongarch64/syscall_arch.h
>>  create mode 100644 crt/loongarch64/crti.s
>>  create mode 100644 crt/loongarch64/crtn.s
>>  create mode 100644 src/fenv/loongarch64/fenv.S
>>  create mode 100644 src/ldso/loongarch64/dlsym.s
>>  create mode 100644 src/setjmp/loongarch64/longjmp.S
>>  create mode 100644 src/setjmp/loongarch64/setjmp.S
>>  create mode 100644 src/signal/loongarch64/restore.s
>>  create mode 100644 src/signal/loongarch64/sigsetjmp.s
>>  create mode 100644 src/thread/loongarch64/__set_thread_area.s
>>  create mode 100644 src/thread/loongarch64/__unmapself.s
>>  create mode 100644 src/thread/loongarch64/clone.s
>>  create mode 100644 src/thread/loongarch64/syscall_cp.s
>>
>> diff --git a/arch/loongarch64/atomic_arch.h 
>> b/arch/loongarch64/atomic_arch.h
>> new file mode 100644
>> index 00000000..bf4805c9
>> --- /dev/null
>> +++ b/arch/loongarch64/atomic_arch.h
>> @@ -0,0 +1,53 @@
>> +#define a_ll a_ll
>> +static inline int a_ll(volatile int *p)
>> +{
>> +    int v;
>> +    __asm__ __volatile__ (
>> +        "ll.w %0, %1"
>> +        : "=r"(v)
>> +        : "ZC"(*p));
>> +    return v;
>> +}
>> +
>> +#define a_sc a_sc
>> +static inline int a_sc(volatile int *p, int v)
>> +{
>> +    int r;
>> +    __asm__ __volatile__ (
>> +        "sc.w %0, %1"
>> +        : "=r"(r), "=ZC"(*p)
>> +        : "0"(v) : "memory");
>> +    return r;
>> +}
>> +
>> +#define a_ll_p a_ll_p
>> +static inline void *a_ll_p(volatile void *p)
>> +{
>> +    void *v;
>> +    __asm__ __volatile__ (
>> +        "ll.d %0, %1"
>> +        : "=r"(v)
>> +        : "ZC"(*(void *volatile *)p));
>> +    return v;
>> +}
>> +
>> +#define a_sc_p a_sc_p
>> +static inline int a_sc_p(volatile void *p, void *v)
>> +{
>> +    long r;
>> +    __asm__ __volatile__ (
>> +        "sc.d %0, %1"
>> +        : "=r"(r), "=ZC"(*(void *volatile *)p)
>> +        : "0"(v)
>> +        : "memory");
>> +    return r;
>> +}
>> +
>> +#define a_barrier a_barrier
>> +static inline void a_barrier()
>> +{
>> +    __asm__ __volatile__ ("dbar 0" : : : "memory");
>> +}
>> +
>> +#define a_pre_llsc a_barrier
>> +#define a_post_llsc a_barrier
>> diff --git a/arch/loongarch64/bits/alltypes.h.in 
>> b/arch/loongarch64/bits/alltypes.h.in
>> new file mode 100644
>> index 00000000..d1807aca
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/alltypes.h.in
>> @@ -0,0 +1,18 @@
>> +#define _Addr long
>> +#define _Int64 long
>> +#define _Reg long
>> +
>> +#define __BYTE_ORDER 1234
>
> `#define __BYTE_ORDER __LITTLE_ENDIAN` could be more consistent with 
> other
> arches.
>
>> +#define __LONG_MAX 0x7fffffffffffffffL
>> +
>> +#ifndef __cplusplus
>> +TYPEDEF int wchar_t;
>> +#endif
>> +
>> +TYPEDEF float float_t;
>> +TYPEDEF double double_t;
>> +
>> +TYPEDEF struct { long long __ll; long double __ld; } max_align_t;
>> +
>> +TYPEDEF unsigned nlink_t;
>
> `unsigned int`? Same for other bare `unsigned` usages.
>
>> +TYPEDEF int blksize_t;
>> diff --git a/arch/loongarch64/bits/fenv.h b/arch/loongarch64/bits/fenv.h
>> new file mode 100644
>> index 00000000..264cafb5
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/fenv.h
>> @@ -0,0 +1,20 @@
>> +#define FE_INEXACT    0x010000
>> +#define FE_UNDERFLOW  0x020000
>> +#define FE_OVERFLOW   0x040000
>> +#define FE_DIVBYZERO  0x080000
>> +#define FE_INVALID    0x100000
>> +
>> +#define FE_ALL_EXCEPT 0x1F0000
>> +
>> +#define FE_TONEAREST  0x000
>> +#define FE_TOWARDZERO 0x100
>> +#define FE_UPWARD     0x200
>> +#define FE_DOWNWARD   0x300
>> +
>> +typedef unsigned fexcept_t;
>> +
>> +typedef struct {
>> +    unsigned __cw;
>> +} fenv_t;
>> +
>> +#define FE_DFL_ENV ((const fenv_t *) -1)
>> diff --git a/arch/loongarch64/bits/float.h 
>> b/arch/loongarch64/bits/float.h
>> new file mode 100644
>> index 00000000..719c7908
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/float.h
>> @@ -0,0 +1,16 @@
>> +#define FLT_EVAL_METHOD 0
>> +
>> +#define LDBL_TRUE_MIN 6.47517511943802511092443895822764655e-4966L
>> +#define LDBL_MIN 3.36210314311209350626267781732175260e-4932L
>> +#define LDBL_MAX 1.18973149535723176508575932662800702e+4932L
>> +#define LDBL_EPSILON 1.92592994438723585305597794258492732e-34L
>> +
>> +#define LDBL_MANT_DIG 113
>> +#define LDBL_MIN_EXP (-16381)
>> +#define LDBL_MAX_EXP 16384
>> +
>> +#define LDBL_DIG 33
>> +#define LDBL_MIN_10_EXP (-4931)
>> +#define LDBL_MAX_10_EXP 4932
>> +
>> +#define DECIMAL_DIG 36
>> diff --git a/arch/loongarch64/bits/posix.h 
>> b/arch/loongarch64/bits/posix.h
>> new file mode 100644
>> index 00000000..8068ce98
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/posix.h
>> @@ -0,0 +1,2 @@
>> +#define _POSIX_V6_LP64_OFF64 1
>> +#define _POSIX_V7_LP64_OFF64 1
>> diff --git a/arch/loongarch64/bits/ptrace.h 
>> b/arch/loongarch64/bits/ptrace.h
>> new file mode 100644
>> index 00000000..741fc668
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/ptrace.h
>> @@ -0,0 +1,4 @@
>> +#define PTRACE_GET_THREAD_AREA    25
>> +#define PTRACE_SET_THREAD_AREA    26
>> +#define PTRACE_GET_WATCH_REGS    0xd0
>> +#define PTRACE_SET_WATCH_REGS    0xd1
>
> This file doesn't seem necessary.
>
>> diff --git a/arch/loongarch64/bits/reg.h b/arch/loongarch64/bits/reg.h
>> new file mode 100644
>> index 00000000..2633f39d
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/reg.h
>> @@ -0,0 +1,2 @@
>> +#undef __WORDSIZE
>> +#define __WORDSIZE 64
>> diff --git a/arch/loongarch64/bits/setjmp.h 
>> b/arch/loongarch64/bits/setjmp.h
>> new file mode 100644
>> index 00000000..f4a7f8a3
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/setjmp.h
>> @@ -0,0 +1 @@
>> +typedef unsigned long long __jmp_buf[22];
>
> `unsigned long` would suffice?
>
>> diff --git a/arch/loongarch64/bits/signal.h 
>> b/arch/loongarch64/bits/signal.h
>> new file mode 100644
>> index 00000000..16f56f21
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/signal.h
>> @@ -0,0 +1,80 @@
>> +#if defined(_POSIX_SOURCE) || defined(_POSIX_C_SOURCE) \
>> + || defined(_XOPEN_SOURCE) || defined(_GNU_SOURCE) || 
>> defined(_BSD_SOURCE)
>> +
>> +#if defined(_XOPEN_SOURCE) || defined(_GNU_SOURCE) || 
>> defined(_BSD_SOURCE)
>> +#define MINSIGSTKSZ 4096
>> +#define SIGSTKSZ 16384
>> +#endif
>> +
>> +typedef unsigned long greg_t, gregset_t[32];
>> +
>> +typedef struct sigcontext {
>> +    unsigned long pc;
>> +    gregset_t gregs;
>> +    unsigned int flags;
>> +    unsigned long extcontext[];
>> +} mcontext_t;
>
> This is different from the glibc definition:
>
> ```c
> // Linux arch/loongarch/include/uapi/asm/sigcontext.h
> struct sigcontext {
>         __u64   sc_pc;
>         __u64   sc_regs[32];
>         __u32   sc_flags;
>         __u64   sc_extcontext[0] __attribute__((__aligned__(16)));
> };
>
> #define CONTEXT_INFO_ALIGN      16
> struct sctx_info {
>         __u32   magic;
>         __u32   size;
>         __u64   padding;        /* padding to 16 bytes */
> };
>
> /* FPU context */
> #define FPU_CTX_MAGIC           0x46505501
> #define FPU_CTX_ALIGN           8p
> struct fpu_context {
>         __u64   regs[32];
>         __u64   fcc;
>         __u32   fcsr;
> };
>
> // glibc sysdeps/unix/sysv/linux/loongarch/sys/ucontext.h
> typedef struct mcontext_t
> {
>   unsigned long long __pc;
>   unsigned long long __gregs[32];
>   unsigned int __flags;
>   unsigned long long __extcontext[0] __attribute__((__aligned__(16)));
> } mcontext_t;
> ```
>
> Which you may want to keep consistent with.
>
>> +
>> +struct sigaltstack {
>> +    void *ss_sp;
>> +    int ss_flags;
>> +    size_t ss_size;
>> +};
>> +
>> +typedef struct __ucontext
>> +{
>> +    unsigned long uc_flags;
>> +    struct __ucontext *uc_link;
>> +    stack_t uc_stack;
>> +    sigset_t uc_sigmask;
>> +    long __uc_pad;
>> +    mcontext_t uc_mcontext;
>> +} ucontext_t;
>> +
>> +#define SA_NOCLDSTOP  1
>> +#define SA_NOCLDWAIT  2
>> +#define SA_SIGINFO    4
>> +#define SA_ONSTACK    0x08000000
>> +#define SA_RESTART    0x10000000
>> +#define SA_NODEFER    0x40000000
>> +#define SA_RESETHAND  0x80000000
>> +#define SA_RESTORER   0x0
>> +
>> +#endif
>> +
>> +#define SIGHUP           1
>> +#define SIGINT           2
>> +#define SIGQUIT          3
>> +#define SIGILL           4
>> +#define SIGTRAP          5
>> +#define SIGABRT          6
>> +#define SIGIOT           SIGABRT
>> +#define SIGBUS           7
>> +#define SIGFPE           8
>> +#define SIGKILL          9
>> +#define SIGUSR1         10
>> +#define SIGSEGV         11
>> +#define SIGUSR2         12
>> +#define SIGPIPE         13
>> +#define SIGALRM         14
>> +#define SIGTERM         15
>> +#define SIGSTKFLT       16
>> +#define SIGCHLD         17
>> +#define SIGCONT         18
>> +#define SIGSTOP         19
>> +#define SIGTSTP         20
>> +#define SIGTTIN         21
>> +#define SIGTTOU         22
>> +#define SIGURG          23
>> +#define SIGXCPU         24
>> +#define SIGXFSZ         25
>> +#define SIGVTALRM       26
>> +#define SIGPROF         27
>> +#define SIGWINCH        28
>> +#define SIGIO           29
>> +#define SIGPOLL         SIGIO
>> +#define SIGPWR          30
>> +#define SIGSYS          31
>> +#define SIGUNUSED       SIGSYS
>> +
>> +#define _NSIG 65
>> diff --git a/arch/loongarch64/bits/stat.h b/arch/loongarch64/bits/stat.h
>> new file mode 100644
>> index 00000000..b7f4221b
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/stat.h
>> @@ -0,0 +1,18 @@
>> +struct stat {
>> +    dev_t st_dev;
>> +    ino_t st_ino;
>> +    mode_t st_mode;
>> +    nlink_t st_nlink;
>> +    uid_t st_uid;
>> +    gid_t st_gid;
>> +    dev_t st_rdev;
>> +    unsigned long __pad;
>> +    off_t st_size;
>> +    blksize_t st_blksize;
>> +    int __pad2;
>> +    blkcnt_t st_blocks;
>> +    struct timespec st_atim;
>> +    struct timespec st_mtim;
>> +    struct timespec st_ctim;
>> +    unsigned __unused[2];
>> +};
>> diff --git a/arch/loongarch64/bits/stdint.h 
>> b/arch/loongarch64/bits/stdint.h
>> new file mode 100644
>> index 00000000..1bb147f2
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/stdint.h
>> @@ -0,0 +1,20 @@
>> +typedef int32_t int_fast16_t;
>> +typedef int32_t int_fast32_t;
>> +typedef uint32_t uint_fast16_t;
>> +typedef uint32_t uint_fast32_t;
>> +
>> +#define INT_FAST16_MIN  INT32_MIN
>> +#define INT_FAST32_MIN  INT32_MIN
>> +
>> +#define INT_FAST16_MAX  INT32_MAX
>> +#define INT_FAST32_MAX  INT32_MAX
>> +
>> +#define UINT_FAST16_MAX UINT32_MAX
>> +#define UINT_FAST32_MAX UINT32_MAX
>> +
>> +#define INTPTR_MIN      INT64_MIN
>> +#define INTPTR_MAX      INT64_MAX
>> +#define UINTPTR_MAX     UINT64_MAX
>> +#define PTRDIFF_MIN     INT64_MIN
>> +#define PTRDIFF_MAX     INT64_MAX
>> +#define SIZE_MAX        UINT64_MAX
>> diff --git a/arch/loongarch64/bits/syscall.h.in 
>> b/arch/loongarch64/bits/syscall.h.in
>> new file mode 100644
>> index 00000000..689ff36b
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/syscall.h.in
>> @@ -0,0 +1,303 @@
>> +#define __NR_io_setup                   0
>> +#define __NR_io_destroy                 1
>> +#define __NR_io_submit                  2
>> +#define __NR_io_cancel                  3
>> +#define __NR_io_getevents               4
>> +#define __NR_setxattr                   5
>> +#define __NR_lsetxattr                  6
>> +#define __NR_fsetxattr                  7
>> +#define __NR_getxattr                   8
>> +#define __NR_lgetxattr                  9
>> +#define __NR_fgetxattr                  10
>> +#define __NR_listxattr                  11
>> +#define __NR_llistxattr                 12
>> +#define __NR_flistxattr                 13
>> +#define __NR_removexattr                14
>> +#define __NR_lremovexattr               15
>> +#define __NR_fremovexattr               16
>> +#define __NR_getcwd                     17
>> +#define __NR_lookup_dcookie             18
>> +#define __NR_eventfd2                   19
>> +#define __NR_epoll_create1              20
>> +#define __NR_epoll_ctl                  21
>> +#define __NR_epoll_pwait                22
>> +#define __NR_dup                        23
>> +#define __NR_dup3                       24
>> +#define __NR3264_fcntl                  25
>> +#define __NR_inotify_init1              26
>> +#define __NR_inotify_add_watch          27
>> +#define __NR_inotify_rm_watch           28
>> +#define __NR_ioctl                      29
>> +#define __NR_ioprio_set                 30
>> +#define __NR_ioprio_get                 31
>> +#define __NR_flock                      32
>> +#define __NR_mknodat                    33
>> +#define __NR_mkdirat                    34
>> +#define __NR_unlinkat                   35
>> +#define __NR_symlinkat                  36
>> +#define __NR_linkat                     37
>> +#define __NR_umount2                    39
>> +#define __NR_mount                      40
>> +#define __NR_pivot_root                 41
>> +#define __NR_nfsservctl                 42
>> +#define __NR3264_statfs                 43
>> +#define __NR3264_fstatfs                44
>> +#define __NR3264_truncate               45
>> +#define __NR3264_ftruncate              46
>> +#define __NR_fallocate                  47
>> +#define __NR_faccessat                  48
>> +#define __NR_chdir                      49
>> +#define __NR_fchdir                     50
>> +#define __NR_chroot                     51
>> +#define __NR_fchmod                     52
>> +#define __NR_fchmodat                   53
>> +#define __NR_fchownat                   54
>> +#define __NR_fchown                     55
>> +#define __NR_openat                     56
>> +#define __NR_close                      57
>> +#define __NR_vhangup                    58
>> +#define __NR_pipe2                      59
>> +#define __NR_quotactl                   60
>> +#define __NR_getdents64                 61
>> +#define __NR3264_lseek                  62
>> +#define __NR_read                       63
>> +#define __NR_write                      64
>> +#define __NR_readv                      65
>> +#define __NR_writev                     66
>> +#define __NR_pread64                    67
>> +#define __NR_pwrite64                   68
>> +#define __NR_preadv                     69
>> +#define __NR_pwritev                    70
>> +#define __NR3264_sendfile               71
>> +#define __NR_pselect6                   72
>> +#define __NR_ppoll                      73
>> +#define __NR_signalfd4                  74
>> +#define __NR_vmsplice                   75
>> +#define __NR_splice                     76
>> +#define __NR_tee                        77
>> +#define __NR_readlinkat                 78
>> +#define __NR_sync                       81
>> +#define __NR_fsync                      82
>> +#define __NR_fdatasync                  83
>> +#define __NR_sync_file_range            84
>> +#define __NR_timerfd_create             85
>> +#define __NR_timerfd_settime            86
>> +#define __NR_timerfd_gettime            87
>> +#define __NR_utimensat                  88
>> +#define __NR_acct                       89
>> +#define __NR_capget                     90
>> +#define __NR_capset                     91
>> +#define __NR_personality                92
>> +#define __NR_exit                       93
>> +#define __NR_exit_group                 94
>> +#define __NR_waitid                     95
>> +#define __NR_set_tid_address            96
>> +#define __NR_unshare                    97
>> +#define __NR_futex                      98
>> +#define __NR_set_robust_list            99
>> +#define __NR_get_robust_list            100
>> +#define __NR_nanosleep                  101
>> +#define __NR_getitimer                  102
>> +#define __NR_setitimer                  103
>> +#define __NR_kexec_load                 104
>> +#define __NR_init_module                105
>> +#define __NR_delete_module              106
>> +#define __NR_timer_create               107
>> +#define __NR_timer_gettime              108
>> +#define __NR_timer_getoverrun           109
>> +#define __NR_timer_settime              110
>> +#define __NR_timer_delete               111
>> +#define __NR_clock_settime              112
>> +#define __NR_clock_gettime              113
>> +#define __NR_clock_getres               114
>> +#define __NR_clock_nanosleep            115
>> +#define __NR_syslog                     116
>> +#define __NR_ptrace                     117
>> +#define __NR_sched_setparam             118
>> +#define __NR_sched_setscheduler         119
>> +#define __NR_sched_getscheduler         120
>> +#define __NR_sched_getparam             121
>> +#define __NR_sched_setaffinity          122
>> +#define __NR_sched_getaffinity          123
>> +#define __NR_sched_yield                124
>> +#define __NR_sched_get_priority_max     125
>> +#define __NR_sched_get_priority_min     126
>> +#define __NR_sched_rr_get_interval      127
>> +#define __NR_restart_syscall            128
>> +#define __NR_kill                       129
>> +#define __NR_tkill                      130
>> +#define __NR_tgkill                     131
>> +#define __NR_sigaltstack                132
>> +#define __NR_rt_sigsuspend              133
>> +#define __NR_rt_sigaction               134
>> +#define __NR_rt_sigprocmask             135
>> +#define __NR_rt_sigpending              136
>> +#define __NR_rt_sigtimedwait            137
>> +#define __NR_rt_sigqueueinfo            138
>> +#define __NR_rt_sigreturn               139
>> +#define __NR_setpriority                140
>> +#define __NR_getpriority                141
>> +#define __NR_reboot                     142
>> +#define __NR_setregid                   143
>> +#define __NR_setgid                     144
>> +#define __NR_setreuid                   145
>> +#define __NR_setuid                     146
>> +#define __NR_setresuid                  147
>> +#define __NR_getresuid                  148
>> +#define __NR_setresgid                  149
>> +#define __NR_getresgid                  150
>> +#define __NR_setfsuid                   151
>> +#define __NR_setfsgid                   152
>> +#define __NR_times                      153
>> +#define __NR_setpgid                    154
>> +#define __NR_getpgid                    155
>> +#define __NR_getsid                     156
>> +#define __NR_setsid                     157
>> +#define __NR_getgroups                  158
>> +#define __NR_setgroups                  159
>> +#define __NR_uname                      160
>> +#define __NR_sethostname                161
>> +#define __NR_setdomainname              162
>> +#define __NR_getrlimit                  163
>> +#define __NR_setrlimit                  164
>> +#define __NR_getrusage                  165
>> +#define __NR_umask                      166
>> +#define __NR_prctl                      167
>> +#define __NR_getcpu                     168
>> +#define __NR_gettimeofday               169
>> +#define __NR_settimeofday               170
>> +#define __NR_adjtimex                   171
>> +#define __NR_getpid                     172
>> +#define __NR_getppid                    173
>> +#define __NR_getuid                     174
>> +#define __NR_geteuid                    175
>> +#define __NR_getgid                     176
>> +#define __NR_getegid                    177
>> +#define __NR_gettid                     178
>> +#define __NR_sysinfo                    179
>> +#define __NR_mq_open                    180
>> +#define __NR_mq_unlink                  181
>> +#define __NR_mq_timedsend               182
>> +#define __NR_mq_timedreceive            183
>> +#define __NR_mq_notify                  184
>> +#define __NR_mq_getsetattr              185
>> +#define __NR_msgget                     186
>> +#define __NR_msgctl                     187
>> +#define __NR_msgrcv                     188
>> +#define __NR_msgsnd                     189
>> +#define __NR_semget                     190
>> +#define __NR_semctl                     191
>> +#define __NR_semtimedop                 192
>> +#define __NR_semop                      193
>> +#define __NR_shmget                     194
>> +#define __NR_shmctl                     195
>> +#define __NR_shmat                      196
>> +#define __NR_shmdt                      197
>> +#define __NR_socket                     198
>> +#define __NR_socketpair                 199
>> +#define __NR_bind                       200
>> +#define __NR_listen                     201
>> +#define __NR_accept                     202
>> +#define __NR_connect                    203
>> +#define __NR_getsockname                204
>> +#define __NR_getpeername                205
>> +#define __NR_sendto                     206
>> +#define __NR_recvfrom                   207
>> +#define __NR_setsockopt                 208
>> +#define __NR_getsockopt                 209
>> +#define __NR_shutdown                   210
>> +#define __NR_sendmsg                    211
>> +#define __NR_recvmsg                    212
>> +#define __NR_readahead                  213
>> +#define __NR_brk                        214
>> +#define __NR_munmap                     215
>> +#define __NR_mremap                     216
>> +#define __NR_add_key                    217
>> +#define __NR_request_key                218
>> +#define __NR_keyctl                     219
>> +#define __NR_clone                      220
>> +#define __NR_execve                     221
>> +#define __NR3264_mmap                   222
>> +#define __NR3264_fadvise64              223
>> +#define __NR_swapon                     224
>> +#define __NR_swapoff                    225
>> +#define __NR_mprotect                   226
>> +#define __NR_msync                      227
>> +#define __NR_mlock                      228
>> +#define __NR_munlock                    229
>> +#define __NR_mlockall                   230
>> +#define __NR_munlockall                 231
>> +#define __NR_mincore                    232
>> +#define __NR_madvise                    233
>> +#define __NR_remap_file_pages           234
>> +#define __NR_mbind                      235
>> +#define __NR_get_mempolicy              236
>> +#define __NR_set_mempolicy              237
>> +#define __NR_migrate_pages              238
>> +#define __NR_move_pages                 239
>> +#define __NR_rt_tgsigqueueinfo          240
>> +#define __NR_perf_event_open            241
>> +#define __NR_accept4                    242
>> +#define __NR_recvmmsg                   243
>> +#define __NR_arch_specific_syscall      244
>> +#define __NR_wait4                      260
>> +#define __NR_prlimit64                  261
>> +#define __NR_fanotify_init              262
>> +#define __NR_fanotify_mark              263
>> +#define __NR_name_to_handle_at          264
>> +#define __NR_open_by_handle_at          265
>> +#define __NR_clock_adjtime              266
>> +#define __NR_syncfs                     267
>> +#define __NR_setns                      268
>> +#define __NR_sendmmsg                   269
>> +#define __NR_process_vm_readv           270
>> +#define __NR_process_vm_writev          271
>> +#define __NR_kcmp                       272
>> +#define __NR_finit_module               273
>> +#define __NR_sched_setattr              274
>> +#define __NR_sched_getattr              275
>> +#define __NR_renameat2                  276
>> +#define __NR_seccomp                    277
>> +#define __NR_getrandom                  278
>> +#define __NR_memfd_create               279
>> +#define __NR_bpf                        280
>> +#define __NR_execveat                   281
>> +#define __NR_userfaultfd                282
>> +#define __NR_membarrier                 283
>> +#define __NR_mlock2                     284
>> +#define __NR_copy_file_range            285
>> +#define __NR_preadv2                    286
>> +#define __NR_pwritev2                   287
>> +#define __NR_pkey_mprotect              288
>> +#define __NR_pkey_alloc                 289
>> +#define __NR_pkey_free                  290
>> +#define __NR_statx                      291
>> +#define __NR_io_pgetevents              292
>> +#define __NR_rseq                       293
>> +#define __NR_kexec_file_load            294
>> +#define __NR_pidfd_send_signal          424
>> +#define __NR_io_uring_setup             425
>> +#define __NR_io_uring_enter             426
>> +#define __NR_io_uring_register          427
>> +#define __NR_open_tree                428
>> +#define __NR_move_mount                429
>> +#define __NR_fsopen                430
>> +#define __NR_fsconfig                431
>> +#define __NR_fsmount                432
>> +#define __NR_fspick                433
>> +#define __NR_pidfd_open                434
>> +#define __NR_clone3                435
>> +#define __NR_close_range            436
>> +#define __NR_openat2                437
>> +#define __NR_pidfd_getfd            438
>> +#define __NR_faccessat2                439
>
> Mixed tab and spaces for some lines, please use spaces only for 
> consistency
> with bulk of surrounding code.
>
>> +#define __NR_process_madvise 440
>> +#define __NR_fcntl                      __NR3264_fcntl
>> +#define __NR_statfs                     __NR3264_statfs
>> +#define __NR_fstatfs                    __NR3264_fstatfs
>> +#define __NR_truncate                   __NR3264_truncate
>> +#define __NR_ftruncate                  __NR3264_ftruncate
>> +#define __NR_lseek                      __NR3264_lseek
>> +#define __NR_sendfile                   __NR3264_sendfile
>> +#define __NR_mmap                       __NR3264_mmap
>> +#define __NR_fadvise64                  __NR3264_fadvise64
>> diff --git a/arch/loongarch64/bits/user.h b/arch/loongarch64/bits/user.h
>> new file mode 100644
>> index 00000000..4d4cd534
>> --- /dev/null
>> +++ b/arch/loongarch64/bits/user.h
>> @@ -0,0 +1,5 @@
>> +#define ELF_NGREG    45
>> +#define ELF_NFPREG    33
>> +
>> +typedef unsigned long elf_greg_t, elf_gregset_t[ELF_NGREG];
>> +typedef double elf_fpreg_t, elf_fpregset_t[ELF_NFPREG];
>> diff --git a/arch/loongarch64/crt_arch.h b/arch/loongarch64/crt_arch.h
>> new file mode 100644
>> index 00000000..e0760d9e
>> --- /dev/null
>> +++ b/arch/loongarch64/crt_arch.h
>> @@ -0,0 +1,13 @@
>> +__asm__(
>> +".text \n"
>> +".global " START "\n"
>> +".type   " START ", @function\n"
>> +START ":\n"
>> +"    move $fp, $zero\n"
>> +"    move $a0, $sp\n"
>> +".weak _DYNAMIC\n"
>> +".hidden _DYNAMIC\n"
>> +"    la.local $a1, _DYNAMIC\n"
>> +"    bstrins.d $sp, $zero, 3, 0\n"
>> +"    b " START "_c\n"
>> +);
>> diff --git a/arch/loongarch64/pthread_arch.h 
>> b/arch/loongarch64/pthread_arch.h
>> new file mode 100644
>> index 00000000..95ee4c7a
>> --- /dev/null
>> +++ b/arch/loongarch64/pthread_arch.h
>> @@ -0,0 +1,13 @@
>> +static inline uintptr_t __get_tp()
>> +{
>> +    register uintptr_t tp __asm__("tp");
>> +    __asm__ ("" : "=r" (tp) );
>
> While the current approach works, it's a bit fragile [1], and the 
> simple and
> plain riscv version works too:
>
> uintptr_t tp;
> __asm__ ("move %0, $tp" : "=r"(tp));
>
> [1]:https://gcc.gnu.org/onlinedocs/gcc/Local-Register-Variables.html#Local-Register-Variables 
>
>
>> +    return tp;
>> +}
>> +
>> +#define TLS_ABOVE_TP
>> +#define GAP_ABOVE_TP 0
>> +
>> +#define DTP_OFFSET 0
>> +
>> +#define MC_PC pc
>> diff --git a/arch/loongarch64/reloc.h b/arch/loongarch64/reloc.h
>> new file mode 100644
>> index 00000000..865a648d
>> --- /dev/null
>> +++ b/arch/loongarch64/reloc.h
>> @@ -0,0 +1,27 @@
>> +#ifdef __loongarch64_soft_float
>
> Should be `__loongarch_soft_float`.
>
>> +#define FP_SUFFIX "-sf"
>> +#else
>> +#define FP_SUFFIX ""
>> +#endif
>> +
>> +#define LDSO_ARCH "loongarch64"  FP_SUFFIX
>> +
>> +#define TPOFF_K (0x0)
>
> a simple "0" would be enough...
>
>> +
>> +#define REL_PLT         R_LARCH_JUMP_SLOT
>> +#define REL_COPY        R_LARCH_COPY
>> +#define REL_DTPMOD      R_LARCH_TLS_DTPMOD64
>> +#define REL_DTPOFF      R_LARCH_TLS_DTPREL64
>> +#define REL_TPOFF       R_LARCH_TLS_TPREL64
>> +#define REL_RELATIVE    R_LARCH_RELATIVE
>> +#define REL_SYMBOLIC    R_LARCH_64
>> +
>> +#define CRTJMP(pc,sp) __asm__ __volatile__( \
>> +    "move $sp,%1 ; jr %0" : : "r"(pc), "r"(sp) : "memory" )
>
> One space after "$sp,".
>
>> +
>> +#define GETFUNCSYM(fp, sym, got) __asm__ ( \
>> +    ".hidden " #sym "\n" \
>> +    ".align 8 \n" \
>> +    "    la.local $t1, "#sym" \n" \
>> +    "    move %0, $t1 \n" \
>> +    : "=r"(*(fp)) : : "memory" )
>
> Does the generic version residing in ldso/dlstart.c not work?
>
>> diff --git a/arch/loongarch64/syscall_arch.h 
>> b/arch/loongarch64/syscall_arch.h
>> new file mode 100644
>> index 00000000..4d5e1885
>> --- /dev/null
>> +++ b/arch/loongarch64/syscall_arch.h
>> @@ -0,0 +1,137 @@
>> +#define __SYSCALL_LL_E(x) (x)
>> +#define __SYSCALL_LL_O(x) (x)
>> +
>> +#define SYSCALL_CLOBBERLIST \
>> +    "$t0", "$t1", "$t2", "$t3", \
>> +    "$t4", "$t5", "$t6", "$t7", "$t8", "memory"
>> +
>> +static inline long __syscall0(long n)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0");
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "=r"(a0)
>> +        : "r"(a7)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>> +
>> +static inline long __syscall1(long n, long a)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0") = a;
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "+r"(a0)
>> +        : "r"(a7)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>> +
>> +static inline long __syscall2(long n, long a, long b)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0") = a;
>> +    register long a1 __asm__("$a1") = b;
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "+r"(a0)
>> +            : "r"(a7), "r"(a1)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>> +
>> +static inline long __syscall3(long n, long a, long b, long c)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0") = a;
>> +    register long a1 __asm__("$a1") = b;
>> +    register long a2 __asm__("$a2") = c;
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "+r"(a0)
>> +            : "r"(a7), "r"(a1), "r"(a2)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>> +
>> +static inline long __syscall4(long n, long a, long b, long c, long d)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0") = a;
>> +    register long a1 __asm__("$a1") = b;
>> +    register long a2 __asm__("$a2") = c;
>> +    register long a3 __asm__("$a3") = d;
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "+r"(a0)
>> +            : "r"(a7), "r"(a1), "r"(a2), "r"(a3)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>> +
>> +static inline long __syscall5(long n, long a, long b, long c, long 
>> d, long e)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0") = a;
>> +    register long a1 __asm__("$a1") = b;
>> +    register long a2 __asm__("$a2") = c;
>> +    register long a3 __asm__("$a3") = d;
>> +    register long a4 __asm__("$a4") = e;
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "+r"(a0)
>> +            : "r"(a7), "r"(a1), "r"(a2), "r"(a3), "r"(a4)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>> +
>> +static inline long __syscall6(long n, long a, long b, long c, long 
>> d, long e, long f)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0") = a;
>> +    register long a1 __asm__("$a1") = b;
>> +    register long a2 __asm__("$a2") = c;
>> +    register long a3 __asm__("$a3") = d;
>> +    register long a4 __asm__("$a4") = e;
>> +    register long a5 __asm__("$a5") = f;
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "+r"(a0)
>> +            : "r"(a7), "r"(a1), "r"(a2), "r"(a3), "r"(a4), "r"(a5)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>> +
>> +static inline long __syscall7(long n, long a, long b, long c, long 
>> d, long e, long f, long g)
>> +{
>> +    register long a7 __asm__("$a7") = n;
>> +    register long a0 __asm__("$a0") = a;
>> +    register long a1 __asm__("$a1") = b;
>> +    register long a2 __asm__("$a2") = c;
>> +    register long a3 __asm__("$a3") = d;
>> +    register long a4 __asm__("$a4") = e;
>> +    register long a5 __asm__("$a5") = f;
>> +    register long a6 __asm__("$a6") = g;
>> +
>> +    __asm__ __volatile__ (
>> +        "syscall 0"
>> +        : "+r"(a0)
>> +            : "r"(a7), "r"(a1), "r"(a2), "r"(a3), "r"(a4), "r"(a5), 
>> "r"(a6)
>> +        : SYSCALL_CLOBBERLIST);
>> +    return a0;
>> +}
>
> __syscall7 seems useless, you could remove this function altogether. 
> Other syscall helpers look good though.
>
>> +
>> +#define VDSO_USEFUL
>> +#define VDSO_CGT_SYM "__vdso_clock_gettime"
>> +#define VDSO_CGT_VER "LINUX_5.10"
>
> Correct.
>
>> +
>> +#define IPC_64  0
>> diff --git a/configure b/configure
>> index 6f5453f5..68062071 100755
>> --- a/configure
>> +++ b/configure
>> @@ -328,6 +328,7 @@ i?86*) ARCH=i386 ;;
>>  x86_64-x32*|x32*|x86_64*x32) ARCH=x32 ;;
>>  x86_64-nt64*) ARCH=nt64 ;;
>>  x86_64*) ARCH=x86_64 ;;
>> +loongarch64*) ARCH=loongarch64 ;;
>>  m68k*) ARCH=m68k ;;
>>  mips64*|mipsisa64*) ARCH=mips64 ;;
>>  mips*) ARCH=mips ;;
>
> I think you're missing one change w.r.t. soft-float subarch detection, 
> somewhere
> below (should be between similar blocks for aarch64 and m68k):
>
> if test "$ARCH" = "loongarch64" ; then
> trycppif __loongarch_soft_float "$t" && SUBARCH=${SUBARCH}-sf
> fi
>
>> diff --git a/crt/loongarch64/crti.s b/crt/loongarch64/crti.s
>> new file mode 100644
>> index 00000000..81c43e6e
>> --- /dev/null
>> +++ b/crt/loongarch64/crti.s
>> @@ -0,0 +1,15 @@
>> +.section .init
>> +.global _init
>> +_init:
>> +    addi.d $sp,$sp,-16
>> +    st.d $fp,$sp,0
>> +    st.d $ra,$sp,8
>> +    addi.d $fp,$sp,16
>> +
>> +.section .fini
>> +.global _fini
>> +_fini:
>> +    addi.d $sp,$sp,-16
>> +    st.d $fp,$sp,0
>> +    st.d $ra,$sp,8
>> +    addi.d $fp,$sp,16
>> diff --git a/crt/loongarch64/crtn.s b/crt/loongarch64/crtn.s
>> new file mode 100644
>> index 00000000..ca3fe80e
>> --- /dev/null
>> +++ b/crt/loongarch64/crtn.s
>> @@ -0,0 +1,12 @@
>> +.section .init
>> +    ld.d $fp,$sp,0
>> +    ld.d $ra,$sp,8
>> +    addi.d $sp,$sp,16
>> +    jr $ra
>> +
>> +
>> +.section .fini
>> +    ld.d $fp,$sp,0
>> +    ld.d $ra,$sp,8
>> +    addi.d $sp,$sp,16
>> +    jr $ra
>
> The crt changes shouldn't be necessary at all huh? I didn't see any 
> custom asm
> for riscv, for example.
>
>> diff --git a/include/elf.h b/include/elf.h
>> index 86e2f0bb..1b0e9e71 100644
>> --- a/include/elf.h
>> +++ b/include/elf.h
>> @@ -697,6 +697,11 @@ typedef struct {
>>  #define NT_MIPS_FP_MODE    0x801
>>  #define NT_MIPS_MSA    0x802
>>  #define NT_VERSION    1
>> +#define NT_LOONGARCH_CPUCFG    0xa00
>> +#define NT_LOONGARCH_CSR    0xa01
>> +#define NT_LOONGARCH_LSX    0xa02
>> +#define NT_LOONGARCH_LASX    0xa03
>> +#define NT_LOONGARCH_LBT    0xa04
>>
>>
>>
>> @@ -3288,6 +3293,66 @@ enum
>>  #define R_RISCV_SET32           56
>>  #define R_RISCV_32_PCREL        57
>>
>> +/* LoongArch ELF Flags */
>> +#define EM_LOONGARCH  258
>
> Please move this definition to the EM_* group above.
>
>> +
>> +#define EF_LARCH_ABI             0x07
>> +#define EF_LARCH_ABI_LP64D       0x03
>> +
>> +/* LoongArch specific dynamic relocations. */
>> +#define R_LARCH_NONE                        0
>> +#define R_LARCH_32                          1
>> +#define R_LARCH_64                          2
>> +#define R_LARCH_RELATIVE                    3
>> +#define R_LARCH_COPY                        4
>> +#define R_LARCH_JUMP_SLOT                   5
>> +#define R_LARCH_TLS_DTPMOD32                6
>> +#define R_LARCH_TLS_DTPMOD64                7
>> +#define R_LARCH_TLS_DTPREL32                8
>> +#define R_LARCH_TLS_DTPREL64                9
>> +#define R_LARCH_TLS_TPREL32                 10
>> +#define R_LARCH_TLS_TPREL64                 11
>> +#define R_LARCH_IRELATIVE                   12
>> +#define R_LARCH_MARK_LA                     20
>> +#define R_LARCH_MARK_PCREL                  21
>> +#define R_LARCH_SOP_PUSH_PCREL              22
>> +#define R_LARCH_SOP_PUSH_ABSOLUTE           23
>> +#define R_LARCH_SOP_PUSH_DUP                24
>> +#define R_LARCH_SOP_PUSH_GPREL              25
>> +#define R_LARCH_SOP_PUSH_TLS_TPREL          26
>> +#define R_LARCH_SOP_PUSH_TLS_GOT            27
>> +#define R_LARCH_SOP_PUSH_TLS_GD             28
>> +#define R_LARCH_SOP_PUSH_PLT_PCREL          29
>> +#define R_LARCH_SOP_ASSERT                  30
>> +#define R_LARCH_SOP_NOT                     31
>> +#define R_LARCH_SOP_SUB                     32
>> +#define R_LARCH_SOP_SL                      33
>> +#define R_LARCH_SOP_SR                      34
>> +#define R_LARCH_SOP_ADD                     35
>> +#define R_LARCH_SOP_AND                     36
>> +#define R_LARCH_SOP_IF_ELSE                 37
>> +#define R_LARCH_SOP_POP_32_S_10_5           38
>> +#define R_LARCH_SOP_POP_32_U_10_12          39
>> +#define R_LARCH_SOP_POP_32_S_10_12          40
>> +#define R_LARCH_SOP_POP_32_S_10_16          41
>> +#define R_LARCH_SOP_POP_32_S_10_16_S2       42
>> +#define R_LARCH_SOP_POP_32_S_5_20           43
>> +#define R_LARCH_SOP_POP_32_S_0_5_10_16_S2   44
>> +#define R_LARCH_SOP_POP_32_S_0_10_10_16_S2  45
>> +#define R_LARCH_SOP_POP_32_U                46
>> +#define R_LARCH_ADD8                        47
>> +#define R_LARCH_ADD16                       48
>> +#define R_LARCH_ADD24                       49
>> +#define R_LARCH_ADD32                       50
>> +#define R_LARCH_ADD64                       51
>> +#define R_LARCH_SUB8                        52
>> +#define R_LARCH_SUB16                       53
>> +#define R_LARCH_SUB24                       54
>> +#define R_LARCH_SUB32                       55
>> +#define R_LARCH_SUB64                       56
>> +#define R_LARCH_GNU_VTINHERIT               57
>> +#define R_LARCH_GNU_VTENTRY                 58
>
> You may want to update the list here to include new reloc types in the 
> LoongArch
> ELF psABI v2.00. Check the latest binutils code for details. The e_flags
> constants need updating too.
>
>> +
>>  #ifdef __cplusplus
>>  }
>>  #endif
>> diff --git a/src/fenv/loongarch64/fenv.S b/src/fenv/loongarch64/fenv.S
>> new file mode 100644
>> index 00000000..aa012c97
>> --- /dev/null
>> +++ b/src/fenv/loongarch64/fenv.S
>> @@ -0,0 +1,72 @@
>> +#ifndef __loongarch_soft_float
>> +
>> +.global    feclearexcept
>> +.type    feclearexcept,@function
>
> Convention is to use a space after the directive, not a tab (it seems 
> only the
> mips port is using tabs here).
>
>> +feclearexcept:
>> +    li.w    $a1, 0x1f0000
>> +    and     $a0, $a0, $a1
>> +    movfcsr2gr $a1, $r0
>> +    or    $a1, $a1, $a0
>> +    xor    $a1, $a1, $a0
>> +    movgr2fcsr $r0, $a1
>> +    li.w    $a0, 0
>> +    jr      $ra
>
> Mixed space and tabs after the mnemonics; there are similar mistakes all
> around so please fix all of them.
>
> Other than that, the implementation isn't optimal, you should be able 
> to use
> `orn` to quickly clear the bits specified by the mask. Or you could 
> consider
> using the other FCSR selectors to hopefully avoid masking altogether.
>
>> +
>> +.global    feraiseexcept
>> +.type    feraiseexcept,@function
>> +feraiseexcept:
>> +    li.w    $a1, 0x1f0000
>> +    and     $a0, $a0, $a1
>> +    movfcsr2gr $a1, $r0
>> +    or    $a1, $a1, $a0
>> +    movgr2fcsr $r0, $a1
>> +        li.w    $a0, 0
>> +    jr      $ra
>> +
>> +.global    fetestexcept
>> +.type    fetestexcept,@function
>> +fetestexcept:
>> +        li.w    $a1, 0x1f0000
>> +        and     $a0, $a0, $a1
>> +    movfcsr2gr $a1, $r0
>> +        and     $a0, $a1, $a0
>> +    jr      $ra
>> +
>> +.global    fegetround
>> +.type    fegetround,@function
>> +fegetround:
>> +    movfcsr2gr $a0, $r0
>> +        andi    $a0, $a0, 0x300  //fcsr0.RM
>> +    jr      $ra
>> +
>> +.global    __fesetround
>> +.hidden __fesetround
>> +.type    __fesetround,@function
>> +__fesetround:
>> +    movfcsr2gr $a1, $r0
>> +    li.w    $a2, -769  //0xfffffcff
>> +    and    $a1, $a1, $a2
>> +    or    $a1, $a1, $a0
>> +    movgr2fcsr $r0, $a1
>> +        li.w    $a0, 0
>> +    jr      $ra
>> +
>> +.global    fegetenv
>> +.type    fegetenv,@function
>> +fegetenv:
>> +    movfcsr2gr $a1, $r0
>> +    st.w    $a1, $a0, 0
>> +        li.w    $a0, 0
>> +    jr      $ra
>> +
>> +.global    fesetenv
>> +.type    fesetenv,@function
>> +fesetenv:
>> +    addi.d    $a1, $a0, 1
>> +    beq    $a1, $r0, 1f
>> +    ld.w    $a1, $a0, 0
>> +1:    movgr2fcsr $r0, $a1
>> +        li.w    $a0, 0
>> +    jr      $ra
>> +
>> +#endif
>> diff --git a/src/ldso/loongarch64/dlsym.s b/src/ldso/loongarch64/dlsym.s
>> new file mode 100644
>> index 00000000..edb8214c
>> --- /dev/null
>> +++ b/src/ldso/loongarch64/dlsym.s
>> @@ -0,0 +1,7 @@
>> +.global dlsym
>> +.hidden __dlsym
>> +.type   dlsym,@function
>> +dlsym:
>> +    move    $a2, $ra
>> +    la.global    $t0, __dlsym
>> +    jr    $t0
>> diff --git a/src/setjmp/loongarch64/longjmp.S 
>> b/src/setjmp/loongarch64/longjmp.S
>> new file mode 100644
>> index 00000000..4186974f
>> --- /dev/null
>> +++ b/src/setjmp/loongarch64/longjmp.S
>> @@ -0,0 +1,37 @@
>> +.global    _longjmp
>> +.global    longjmp
>> +.type    _longjmp,@function
>> +.type    longjmp,@function
>> +_longjmp:
>> +longjmp:
>> +    move   $t5, $a0
>> +    move   $a0, $a1
>> +
>> +    bne     $a0, $zero, 1f
>
> bnez
>
>> +    addi.d  $a0, $a0, 1
>> +
>> +1:
>> +    ld.d    $ra, $t5, 0
>> +    ld.d    $sp, $t5, 8
>> +    ld.d    $r21,$t5, 16
>> +    ld.d    $fp, $t5, 24
>> +    ld.d    $s0, $t5, 32
>> +    ld.d    $s1, $t5, 40
>> +    ld.d    $s2, $t5, 48
>> +    ld.d    $s3, $t5, 56
>> +    ld.d    $s4, $t5, 64
>> +    ld.d    $s5, $t5, 72
>> +    ld.d    $s6, $t5, 80
>> +    ld.d    $s7, $t5, 88
>> +    ld.d    $s8, $t5, 96
>> +#ifndef __loongarch64_soft_float
>
> `__loongarch_soft_float`.
>
>> +    fld.d   $fs0, $t5, 104
>> +    fld.d   $fs1, $t5, 112
>> +    fld.d   $fs2, $t5, 120
>> +    fld.d   $fs3, $t5, 128
>> +    fld.d   $fs4, $t5, 136
>> +    fld.d   $fs5, $t5, 144
>> +    fld.d   $fs6, $t5, 152
>> +    fld.d   $fs7, $t5, 160
>> +#endif
>> +    jr      $ra
>> diff --git a/src/setjmp/loongarch64/setjmp.S 
>> b/src/setjmp/loongarch64/setjmp.S
>> new file mode 100644
>> index 00000000..f3bb7c70
>> --- /dev/null
>> +++ b/src/setjmp/loongarch64/setjmp.S
>> @@ -0,0 +1,34 @@
>> +.global    __setjmp
>> +.global    _setjmp
>> +.global    setjmp
>> +.type    __setjmp,@function
>> +.type    _setjmp,@function
>> +.type    setjmp,@function
>> +__setjmp:
>> +_setjmp:
>> +setjmp:
>> +    st.d    $ra, $a0, 0
>> +    st.d    $sp, $a0, 8
>> +    st.d    $r21,$a0, 16
>> +    st.d    $fp, $a0, 24
>> +    st.d    $s0, $a0, 32
>> +    st.d    $s1, $a0, 40
>> +    st.d    $s2, $a0, 48
>> +    st.d    $s3, $a0, 56
>> +    st.d    $s4, $a0, 64
>> +    st.d    $s5, $a0, 72
>> +    st.d    $s6, $a0, 80
>> +    st.d    $s7, $a0, 88
>> +    st.d    $s8, $a0, 96
>> +#ifndef __loongarch64_soft_float
>
> `__loongarch_soft_float`.
>
>> +    fst.d   $fs0, $a0, 104
>> +    fst.d   $fs1, $a0, 112
>> +    fst.d   $fs2, $a0, 120
>> +    fst.d   $fs3, $a0, 128
>> +    fst.d   $fs4, $a0, 136
>> +    fst.d   $fs5, $a0, 144
>> +    fst.d   $fs6, $a0, 152
>> +    fst.d   $fs7, $a0, 160
>> +#endif
>> +    xor     $a0, $a0, $a0
>
> move $a0, $zero
>
>> +    jr      $ra
>> diff --git a/src/signal/loongarch64/restore.s 
>> b/src/signal/loongarch64/restore.s
>> new file mode 100644
>> index 00000000..bca17eb7
>> --- /dev/null
>> +++ b/src/signal/loongarch64/restore.s
>> @@ -0,0 +1,10 @@
>> +.global    __restore_rt
>> +.global    __restore
>> +.hidden __restore_rt
>> +.hidden __restore
>> +.type    __restore_rt,@function
>> +.type    __restore,@function
>> +__restore_rt:
>> +__restore:
>> +    li.w    $a7, 139
>> +    syscall    0
>> diff --git a/src/signal/loongarch64/sigsetjmp.s 
>> b/src/signal/loongarch64/sigsetjmp.s
>> new file mode 100644
>> index 00000000..abd96c62
>> --- /dev/null
>> +++ b/src/signal/loongarch64/sigsetjmp.s
>> @@ -0,0 +1,29 @@
>> +.global sigsetjmp
>> +.global __sigsetjmp
>> +.type   sigsetjmp,@function
>> +.type   __sigsetjmp,@function
>> +sigsetjmp:
>> +__sigsetjmp:
>> +        move $t5, $a0
>> +        move $t6, $a1
>> +
>> +        # comparing save mask with 0, if equals to 0 then
>> +        # sigsetjmp is equal to setjmp.
>> +        beq     $t6, $zero, 1f
>> +        st.d    $ra, $t5, 168
>> +
>> +        # save base of got so that we can use it later
>> +        # once we return from 'longjmp'
>> +        la.global  $t8, setjmp
>> +        jirl    $ra, $t8, 0
>> +
>> +        move    $a1, $a0        # Return from 'setjmp' or 'longjmp'
>> +        ld.d    $ra, $t5, 168   # Restore ra of sigsetjmp
>> +        move    $a0, $t5
>> +
>> +.hidden __sigsetjmp_tail
>> +        la.global  $t8, __sigsetjmp_tail
>> +        jr         $t8
>> +1:
>> +        la.global  $t8, setjmp
>> +        jr         $t8
>
> This is crazy complicated compared to the riscv port, why is the 
> juggling between a0/a1 and t5/t6 necessary?
>
>> diff --git a/src/thread/loongarch64/__set_thread_area.s 
>> b/src/thread/loongarch64/__set_thread_area.s
>> new file mode 100644
>> index 00000000..6fd09a92
>> --- /dev/null
>> +++ b/src/thread/loongarch64/__set_thread_area.s
>> @@ -0,0 +1,7 @@
>> +.global    __set_thread_area
>> +.hidden __set_thread_area
>> +.type    __set_thread_area,@function
>> +__set_thread_area:
>> +    move    $tp, $a0
>> +    ori    $a0, $zero, 0
>
> `move $a0, $zero` is enough.
>
>> +    jr    $ra
>> diff --git a/src/thread/loongarch64/__unmapself.s 
>> b/src/thread/loongarch64/__unmapself.s
>> new file mode 100644
>> index 00000000..3a44b850
>> --- /dev/null
>> +++ b/src/thread/loongarch64/__unmapself.s
>> @@ -0,0 +1,7 @@
>> +.global    __unmapself
>> +.type    __unmapself, @function
>> +__unmapself:
>> +    li.d    $a7, 215                        # call munmap
>> +    syscall    0
>> +    li.d    $a7, 93                         # call exit
>> +    syscall    0
>> diff --git a/src/thread/loongarch64/clone.s 
>> b/src/thread/loongarch64/clone.s
>> new file mode 100644
>> index 00000000..86e69cfa
>> --- /dev/null
>> +++ b/src/thread/loongarch64/clone.s
>> @@ -0,0 +1,47 @@
>> +#__clone(func, stack, flags, arg, ptid, tls, ctid)
>> +#         a0,    a1,   a2,    a3,  a4,  a5,   a6
>> +# sys_clone3(struct clone_args *cl_args, size_t size)
>> +#                                 a0             a1
>> +
>> +.global    __clone
>> +.hidden __clone
>> +.type    __clone,@function
>> +__clone:
>> +    # Save function pointer and argument pointer on new thread stack
>> +    addi.d    $a1, $a1, -16
>> +    st.d    $a0, $a1, 0    # save function pointer
>> +    st.d    $a3, $a1, 8    # save argument pointer
>> +
>> +    li.d    $t0, ~0x004000ff  # mask CSIGNAL and CLONE_DETACHED
>> +    and    $t1, $a2, $t0     # cl_args.flags
>> +    li.d    $t0, 0x000000ff   # CSIGNAL
>> +    and    $t2, $a2, $t0     # cl_args.exit_signal
>> +
>> +    bstrins.d $sp, $zero, 3, 0  # align stack to 16 bytes
>> +    addi.d    $sp, $sp, -88   # struct clone_args
>> +    st.d    $t1, $sp, 0     # flags
>> +    st.d    $a4, $sp, 8     # pidfd
>> +    st.d    $a6, $sp, 16    # child_tid
>> +    st.d    $a4, $sp, 24    # parent_tid
>> +    st.d    $t2, $sp, 32    # exit_signal
>> +    st.d    $a1, $sp, 40    # stack
>> +    st.d    $zero, $sp, 48  # stack_size
>> +    st.d    $a5, $sp, 56    # tls
>> +    st.d    $zero, $sp, 64  # set_tid
>> +    st.d    $zero, $sp, 72  # set_tid_size
>> +    st.d    $zero, $sp, 80  # cgroup
>> +
>> +    move    $a0, $sp
>> +    li.d    $a1, 88
>> +    li.d    $a7, 435    # __NR_clone3
>> +    syscall 0        # call clone3
>> +
>> +    beqz    $a0, 1f        # whether child process
>> +    addi.d    $sp, $sp, 88
>> +    jr    $ra            # parent process return
>> +1:
>> +    ld.d    $t8, $sp, 0     # function pointer
>> +    ld.d    $a0, $sp, 8     # argument pointer
>> +    jirl    $ra, $t8, 0     # call the user's function
>> +    li.d    $a7, 93
>> +    syscall    0        # child process exit
>
> Do we still want to implement clone functionality entirely with 
> clone3, given the
> previous conclusion of keeping clone(2) for existing sandboxes to 
> continue to work?
>
>> diff --git a/src/thread/loongarch64/syscall_cp.s 
>> b/src/thread/loongarch64/syscall_cp.s
>> new file mode 100644
>> index 00000000..9f57d254
>> --- /dev/null
>> +++ b/src/thread/loongarch64/syscall_cp.s
>> @@ -0,0 +1,29 @@
>> +.global    __cp_begin
>> +.hidden    __cp_begin
>> +.global    __cp_end
>> +.hidden    __cp_end
>> +.global    __cp_cancel
>> +.hidden    __cp_cancel
>> +.hidden    __cancel
>> +.global    __syscall_cp_asm
>> +.hidden    __syscall_cp_asm
>> +.type    __syscall_cp_asm,@function
>> +
>> +__syscall_cp_asm:
>> +__cp_begin:
>> +    ld.w    $a0, $a0, 0
>> +    bnez    $a0, __cp_cancel
>> +    move    $t8, $a1     # reserve system call number
>> +    move    $a0, $a2
>> +    move    $a1, $a3
>> +    move    $a2, $a4
>> +    move    $a3, $a5
>> +    move    $a4, $a6
>> +    move    $a5, $a7
>> +    move    $a7, $t8
>> +    syscall    0
>> +__cp_end:
>> +    jr    $ra
>> +__cp_cancel:
>> +    la.local $t8, __cancel
>> +    jr    $t8
>> -- 
>> 2.36.0
>>
>>

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.