On Sat, 2015-10-31 at 14:23 -0700, Linus Torvalds wrote:
> Mind testing something really stupid, and making the __clear_bit() in
> __clear_close_on_exec() conditiona, something like this:
>
> static inline void __clear_close_on_exec(int fd, struct fdtable *fdt)
> {
> - __clear_bit(fd, fdt->close_on_exec);
> + if (test_bit(fd, fdt->close_on_exec)
> + __clear_bit(fd, fdt->close_on_exec);
> }
>
> and see if it makes a difference.
It does ;)
About 4 % qps increase
3 runs :
lpaa24:~# taskset ff0ff ./opensock -t 16 -n 10000000 -l 10
total = 4176651
total = 4178012
total = 4105226
instead of :
total = 3910620
total = 3874567
total = 3971028
Perf profile :
69.12% opensock opensock [.] memset
|
--- memset
12.37% opensock [kernel.kallsyms] [k] queued_spin_lock_slowpath
|
--- queued_spin_lock_slowpath
|
|--99.99%-- _raw_spin_lock
| |
| |--51.99%-- __close_fd
| | sys_close
| | entry_SYSCALL_64_fastpath
| | __libc_close
| | |
| | --100.00%-- 0x0
| |
| |--47.79%-- __alloc_fd
| | get_unused_fd_flags
| | sock_map_fd
| | sys_socket
| | entry_SYSCALL_64_fastpath
| | __socket
| | |
| | --100.00%-- 0x0
| --0.21%-- [...]
--0.01%-- [...]
1.92% opensock [kernel.kallsyms] [k] _find_next_bit.part.0
|
--- _find_next_bit.part.0
|
|--66.93%-- find_next_zero_bit
| __alloc_fd
| get_unused_fd_flags
| sock_map_fd
| sys_socket
| entry_SYSCALL_64_fastpath
| __socket
|
--33.07%-- __alloc_fd
get_unused_fd_flags
sock_map_fd
sys_socket
entry_SYSCALL_64_fastpath
__socket
|
--100.00%-- 0x0
1.63% opensock [kernel.kallsyms] [k] _raw_spin_lock
|
--- _raw_spin_lock
|
|--28.66%-- get_unused_fd_flags
| sock_map_fd
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
