Messages in this thread |  | | | From | (bill davidsen) | | Subject | Re: Who changed /proc/<pid>/ in 2.6.0-test5-bk9? | | Date | 9 Oct 2003 18:43:00 GMT |
| |
In article <3F846183.9010205@redhat.com>, Ulrich Drepper <drepper@redhat.com> wrote:
| The kernel thread functionality is not used 1-to-1 in the userlevel | interfaces of the pthread library. One very specific combination of all | the CLONE_* flags is used in libpthread. Other users of the kernel | might use other combinations and they won't implement pthreads. That | is perfectly fine and if it fits your bill, do it. But none of this | ever has any influence on the pthread interface. The properties like | sharing of file descriptors are guaranteed.
I confess that my first reaction to this topic was that I could add an attribute when starting a thread, and then use existing code (pthreads) and just take advantage of the non-fd-sharing.
Pseudo code: (some of the if's are clearly ifdef)
if linux add_noshare_to_attributes for (;;) { fd = listen(); start_thread(fd) if linux clode fd } Clearly that isn't what's needed, I would probably define a start_service_and_close_fd_ifUcan macro/procedure. When this is stable and 2.6 is closer to production quality I will probably modify a few applications to see what benefits might be available. -- bill davidsen <davidsen@tmr.com> CTO, TMR Associates, Inc Doing interesting things with little computers since 1979. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
|  |