mirror of
https://github.com/torvalds/linux.git
synced 2024-11-17 01:22:07 +00:00
fcf634098c
The basic idea behind cross memory attach is to allow MPI programs doing
intra-node communication to do a single copy of the message rather than a
double copy of the message via shared memory.
The following patch attempts to achieve this by allowing a destination
process, given an address and size from a source process, to copy memory
directly from the source process into its own address space via a system
call. There is also a symmetrical ability to copy from the current
process's address space into a destination process's address space.
- Use of /proc/pid/mem has been considered, but there are issues with
using it:
- Does not allow for specifying iovecs for both src and dest, assuming
preadv or pwritev was implemented either the area read from or
written to would need to be contiguous.
- Currently mem_read allows only processes who are currently
ptrace'ing the target and are still able to ptrace the target to read
from the target. This check could possibly be moved to the open call,
but its not clear exactly what race this restriction is stopping
(reason appears to have been lost)
- Having to send the fd of /proc/self/mem via SCM_RIGHTS on unix
domain socket is a bit ugly from a userspace point of view,
especially when you may have hundreds if not (eventually) thousands
of processes that all need to do this with each other
- Doesn't allow for some future use of the interface we would like to
consider adding in the future (see below)
- Interestingly reading from /proc/pid/mem currently actually
involves two copies! (But this could be fixed pretty easily)
As mentioned previously use of vmsplice instead was considered, but has
problems. Since you need the reader and writer working co-operatively if
the pipe is not drained then you block. Which requires some wrapping to
do non blocking on the send side or polling on the receive. In all to all
communication it requires ordering otherwise you can deadlock. And in the
example of many MPI tasks writing to one MPI task vmsplice serialises the
copying.
There are some cases of MPI collectives where even a single copy interface
does not get us the performance gain we could. For example in an
MPI_Reduce rather than copy the data from the source we would like to
instead use it directly in a mathops (say the reduce is doing a sum) as
this would save us doing a copy. We don't need to keep a copy of the data
from the source. I haven't implemented this, but I think this interface
could in the future do all this through the use of the flags - eg could
specify the math operation and type and the kernel rather than just
copying the data would apply the specified operation between the source
and destination and store it in the destination.
Although we don't have a "second user" of the interface (though I've had
some nibbles from people who may be interested in using it for intra
process messaging which is not MPI). This interface is something which
hardware vendors are already doing for their custom drivers to implement
fast local communication. And so in addition to this being useful for
OpenMPI it would mean the driver maintainers don't have to fix things up
when the mm changes.
There was some discussion about how much faster a true zero copy would
go. Here's a link back to the email with some testing I did on that:
http://marc.info/?l=linux-mm&m=130105930902915&w=2
There is a basic man page for the proposed interface here:
http://ozlabs.org/~cyeoh/cma/process_vm_readv.txt
This has been implemented for x86 and powerpc, other architecture should
mainly (I think) just need to add syscall numbers for the process_vm_readv
and process_vm_writev. There are 32 bit compatibility versions for
64-bit kernels.
For arch maintainers there are some simple tests to be able to quickly
verify that the syscalls are working correctly here:
http://ozlabs.org/~cyeoh/cma/cma-test-20110718.tgz
Signed-off-by: Chris Yeoh <yeohc@au1.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Howells <dhowells@redhat.com>
Cc: James Morris <jmorris@namei.org>
Cc: <linux-man@vger.kernel.org>
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
351 lines
8.5 KiB
ArmAsm
351 lines
8.5 KiB
ArmAsm
ENTRY(sys_call_table)
|
|
.long sys_restart_syscall /* 0 - old "setup()" system call, used for restarting */
|
|
.long sys_exit
|
|
.long ptregs_fork
|
|
.long sys_read
|
|
.long sys_write
|
|
.long sys_open /* 5 */
|
|
.long sys_close
|
|
.long sys_waitpid
|
|
.long sys_creat
|
|
.long sys_link
|
|
.long sys_unlink /* 10 */
|
|
.long ptregs_execve
|
|
.long sys_chdir
|
|
.long sys_time
|
|
.long sys_mknod
|
|
.long sys_chmod /* 15 */
|
|
.long sys_lchown16
|
|
.long sys_ni_syscall /* old break syscall holder */
|
|
.long sys_stat
|
|
.long sys_lseek
|
|
.long sys_getpid /* 20 */
|
|
.long sys_mount
|
|
.long sys_oldumount
|
|
.long sys_setuid16
|
|
.long sys_getuid16
|
|
.long sys_stime /* 25 */
|
|
.long sys_ptrace
|
|
.long sys_alarm
|
|
.long sys_fstat
|
|
.long sys_pause
|
|
.long sys_utime /* 30 */
|
|
.long sys_ni_syscall /* old stty syscall holder */
|
|
.long sys_ni_syscall /* old gtty syscall holder */
|
|
.long sys_access
|
|
.long sys_nice
|
|
.long sys_ni_syscall /* 35 - old ftime syscall holder */
|
|
.long sys_sync
|
|
.long sys_kill
|
|
.long sys_rename
|
|
.long sys_mkdir
|
|
.long sys_rmdir /* 40 */
|
|
.long sys_dup
|
|
.long sys_pipe
|
|
.long sys_times
|
|
.long sys_ni_syscall /* old prof syscall holder */
|
|
.long sys_brk /* 45 */
|
|
.long sys_setgid16
|
|
.long sys_getgid16
|
|
.long sys_signal
|
|
.long sys_geteuid16
|
|
.long sys_getegid16 /* 50 */
|
|
.long sys_acct
|
|
.long sys_umount /* recycled never used phys() */
|
|
.long sys_ni_syscall /* old lock syscall holder */
|
|
.long sys_ioctl
|
|
.long sys_fcntl /* 55 */
|
|
.long sys_ni_syscall /* old mpx syscall holder */
|
|
.long sys_setpgid
|
|
.long sys_ni_syscall /* old ulimit syscall holder */
|
|
.long sys_olduname
|
|
.long sys_umask /* 60 */
|
|
.long sys_chroot
|
|
.long sys_ustat
|
|
.long sys_dup2
|
|
.long sys_getppid
|
|
.long sys_getpgrp /* 65 */
|
|
.long sys_setsid
|
|
.long sys_sigaction
|
|
.long sys_sgetmask
|
|
.long sys_ssetmask
|
|
.long sys_setreuid16 /* 70 */
|
|
.long sys_setregid16
|
|
.long sys_sigsuspend
|
|
.long sys_sigpending
|
|
.long sys_sethostname
|
|
.long sys_setrlimit /* 75 */
|
|
.long sys_old_getrlimit
|
|
.long sys_getrusage
|
|
.long sys_gettimeofday
|
|
.long sys_settimeofday
|
|
.long sys_getgroups16 /* 80 */
|
|
.long sys_setgroups16
|
|
.long sys_old_select
|
|
.long sys_symlink
|
|
.long sys_lstat
|
|
.long sys_readlink /* 85 */
|
|
.long sys_uselib
|
|
.long sys_swapon
|
|
.long sys_reboot
|
|
.long sys_old_readdir
|
|
.long sys_old_mmap /* 90 */
|
|
.long sys_munmap
|
|
.long sys_truncate
|
|
.long sys_ftruncate
|
|
.long sys_fchmod
|
|
.long sys_fchown16 /* 95 */
|
|
.long sys_getpriority
|
|
.long sys_setpriority
|
|
.long sys_ni_syscall /* old profil syscall holder */
|
|
.long sys_statfs
|
|
.long sys_fstatfs /* 100 */
|
|
.long sys_ioperm
|
|
.long sys_socketcall
|
|
.long sys_syslog
|
|
.long sys_setitimer
|
|
.long sys_getitimer /* 105 */
|
|
.long sys_newstat
|
|
.long sys_newlstat
|
|
.long sys_newfstat
|
|
.long sys_uname
|
|
.long ptregs_iopl /* 110 */
|
|
.long sys_vhangup
|
|
.long sys_ni_syscall /* old "idle" system call */
|
|
.long ptregs_vm86old
|
|
.long sys_wait4
|
|
.long sys_swapoff /* 115 */
|
|
.long sys_sysinfo
|
|
.long sys_ipc
|
|
.long sys_fsync
|
|
.long ptregs_sigreturn
|
|
.long ptregs_clone /* 120 */
|
|
.long sys_setdomainname
|
|
.long sys_newuname
|
|
.long sys_modify_ldt
|
|
.long sys_adjtimex
|
|
.long sys_mprotect /* 125 */
|
|
.long sys_sigprocmask
|
|
.long sys_ni_syscall /* old "create_module" */
|
|
.long sys_init_module
|
|
.long sys_delete_module
|
|
.long sys_ni_syscall /* 130: old "get_kernel_syms" */
|
|
.long sys_quotactl
|
|
.long sys_getpgid
|
|
.long sys_fchdir
|
|
.long sys_bdflush
|
|
.long sys_sysfs /* 135 */
|
|
.long sys_personality
|
|
.long sys_ni_syscall /* reserved for afs_syscall */
|
|
.long sys_setfsuid16
|
|
.long sys_setfsgid16
|
|
.long sys_llseek /* 140 */
|
|
.long sys_getdents
|
|
.long sys_select
|
|
.long sys_flock
|
|
.long sys_msync
|
|
.long sys_readv /* 145 */
|
|
.long sys_writev
|
|
.long sys_getsid
|
|
.long sys_fdatasync
|
|
.long sys_sysctl
|
|
.long sys_mlock /* 150 */
|
|
.long sys_munlock
|
|
.long sys_mlockall
|
|
.long sys_munlockall
|
|
.long sys_sched_setparam
|
|
.long sys_sched_getparam /* 155 */
|
|
.long sys_sched_setscheduler
|
|
.long sys_sched_getscheduler
|
|
.long sys_sched_yield
|
|
.long sys_sched_get_priority_max
|
|
.long sys_sched_get_priority_min /* 160 */
|
|
.long sys_sched_rr_get_interval
|
|
.long sys_nanosleep
|
|
.long sys_mremap
|
|
.long sys_setresuid16
|
|
.long sys_getresuid16 /* 165 */
|
|
.long ptregs_vm86
|
|
.long sys_ni_syscall /* Old sys_query_module */
|
|
.long sys_poll
|
|
.long sys_ni_syscall /* Old nfsservctl */
|
|
.long sys_setresgid16 /* 170 */
|
|
.long sys_getresgid16
|
|
.long sys_prctl
|
|
.long ptregs_rt_sigreturn
|
|
.long sys_rt_sigaction
|
|
.long sys_rt_sigprocmask /* 175 */
|
|
.long sys_rt_sigpending
|
|
.long sys_rt_sigtimedwait
|
|
.long sys_rt_sigqueueinfo
|
|
.long sys_rt_sigsuspend
|
|
.long sys_pread64 /* 180 */
|
|
.long sys_pwrite64
|
|
.long sys_chown16
|
|
.long sys_getcwd
|
|
.long sys_capget
|
|
.long sys_capset /* 185 */
|
|
.long ptregs_sigaltstack
|
|
.long sys_sendfile
|
|
.long sys_ni_syscall /* reserved for streams1 */
|
|
.long sys_ni_syscall /* reserved for streams2 */
|
|
.long ptregs_vfork /* 190 */
|
|
.long sys_getrlimit
|
|
.long sys_mmap_pgoff
|
|
.long sys_truncate64
|
|
.long sys_ftruncate64
|
|
.long sys_stat64 /* 195 */
|
|
.long sys_lstat64
|
|
.long sys_fstat64
|
|
.long sys_lchown
|
|
.long sys_getuid
|
|
.long sys_getgid /* 200 */
|
|
.long sys_geteuid
|
|
.long sys_getegid
|
|
.long sys_setreuid
|
|
.long sys_setregid
|
|
.long sys_getgroups /* 205 */
|
|
.long sys_setgroups
|
|
.long sys_fchown
|
|
.long sys_setresuid
|
|
.long sys_getresuid
|
|
.long sys_setresgid /* 210 */
|
|
.long sys_getresgid
|
|
.long sys_chown
|
|
.long sys_setuid
|
|
.long sys_setgid
|
|
.long sys_setfsuid /* 215 */
|
|
.long sys_setfsgid
|
|
.long sys_pivot_root
|
|
.long sys_mincore
|
|
.long sys_madvise
|
|
.long sys_getdents64 /* 220 */
|
|
.long sys_fcntl64
|
|
.long sys_ni_syscall /* reserved for TUX */
|
|
.long sys_ni_syscall
|
|
.long sys_gettid
|
|
.long sys_readahead /* 225 */
|
|
.long sys_setxattr
|
|
.long sys_lsetxattr
|
|
.long sys_fsetxattr
|
|
.long sys_getxattr
|
|
.long sys_lgetxattr /* 230 */
|
|
.long sys_fgetxattr
|
|
.long sys_listxattr
|
|
.long sys_llistxattr
|
|
.long sys_flistxattr
|
|
.long sys_removexattr /* 235 */
|
|
.long sys_lremovexattr
|
|
.long sys_fremovexattr
|
|
.long sys_tkill
|
|
.long sys_sendfile64
|
|
.long sys_futex /* 240 */
|
|
.long sys_sched_setaffinity
|
|
.long sys_sched_getaffinity
|
|
.long sys_set_thread_area
|
|
.long sys_get_thread_area
|
|
.long sys_io_setup /* 245 */
|
|
.long sys_io_destroy
|
|
.long sys_io_getevents
|
|
.long sys_io_submit
|
|
.long sys_io_cancel
|
|
.long sys_fadvise64 /* 250 */
|
|
.long sys_ni_syscall
|
|
.long sys_exit_group
|
|
.long sys_lookup_dcookie
|
|
.long sys_epoll_create
|
|
.long sys_epoll_ctl /* 255 */
|
|
.long sys_epoll_wait
|
|
.long sys_remap_file_pages
|
|
.long sys_set_tid_address
|
|
.long sys_timer_create
|
|
.long sys_timer_settime /* 260 */
|
|
.long sys_timer_gettime
|
|
.long sys_timer_getoverrun
|
|
.long sys_timer_delete
|
|
.long sys_clock_settime
|
|
.long sys_clock_gettime /* 265 */
|
|
.long sys_clock_getres
|
|
.long sys_clock_nanosleep
|
|
.long sys_statfs64
|
|
.long sys_fstatfs64
|
|
.long sys_tgkill /* 270 */
|
|
.long sys_utimes
|
|
.long sys_fadvise64_64
|
|
.long sys_ni_syscall /* sys_vserver */
|
|
.long sys_mbind
|
|
.long sys_get_mempolicy
|
|
.long sys_set_mempolicy
|
|
.long sys_mq_open
|
|
.long sys_mq_unlink
|
|
.long sys_mq_timedsend
|
|
.long sys_mq_timedreceive /* 280 */
|
|
.long sys_mq_notify
|
|
.long sys_mq_getsetattr
|
|
.long sys_kexec_load
|
|
.long sys_waitid
|
|
.long sys_ni_syscall /* 285 */ /* available */
|
|
.long sys_add_key
|
|
.long sys_request_key
|
|
.long sys_keyctl
|
|
.long sys_ioprio_set
|
|
.long sys_ioprio_get /* 290 */
|
|
.long sys_inotify_init
|
|
.long sys_inotify_add_watch
|
|
.long sys_inotify_rm_watch
|
|
.long sys_migrate_pages
|
|
.long sys_openat /* 295 */
|
|
.long sys_mkdirat
|
|
.long sys_mknodat
|
|
.long sys_fchownat
|
|
.long sys_futimesat
|
|
.long sys_fstatat64 /* 300 */
|
|
.long sys_unlinkat
|
|
.long sys_renameat
|
|
.long sys_linkat
|
|
.long sys_symlinkat
|
|
.long sys_readlinkat /* 305 */
|
|
.long sys_fchmodat
|
|
.long sys_faccessat
|
|
.long sys_pselect6
|
|
.long sys_ppoll
|
|
.long sys_unshare /* 310 */
|
|
.long sys_set_robust_list
|
|
.long sys_get_robust_list
|
|
.long sys_splice
|
|
.long sys_sync_file_range
|
|
.long sys_tee /* 315 */
|
|
.long sys_vmsplice
|
|
.long sys_move_pages
|
|
.long sys_getcpu
|
|
.long sys_epoll_pwait
|
|
.long sys_utimensat /* 320 */
|
|
.long sys_signalfd
|
|
.long sys_timerfd_create
|
|
.long sys_eventfd
|
|
.long sys_fallocate
|
|
.long sys_timerfd_settime /* 325 */
|
|
.long sys_timerfd_gettime
|
|
.long sys_signalfd4
|
|
.long sys_eventfd2
|
|
.long sys_epoll_create1
|
|
.long sys_dup3 /* 330 */
|
|
.long sys_pipe2
|
|
.long sys_inotify_init1
|
|
.long sys_preadv
|
|
.long sys_pwritev
|
|
.long sys_rt_tgsigqueueinfo /* 335 */
|
|
.long sys_perf_event_open
|
|
.long sys_recvmmsg
|
|
.long sys_fanotify_init
|
|
.long sys_fanotify_mark
|
|
.long sys_prlimit64 /* 340 */
|
|
.long sys_name_to_handle_at
|
|
.long sys_open_by_handle_at
|
|
.long sys_clock_adjtime
|
|
.long sys_syncfs
|
|
.long sys_sendmmsg /* 345 */
|
|
.long sys_setns
|
|
.long sys_process_vm_readv
|
|
.long sys_process_vm_writev
|