On Fri, 2 Feb 2007, Davide Libenzi wrote:
> On Fri, 2 Feb 2007, Ingo Molnar wrote:
>
> > in fact an app could also /trigger/ the execution of a syscall in a
> > different context - to create parallelism artificially - without any
> > blocking event. So we could do:
> >
> > cookie1 = sys_async(sys_read, params);
> > cookie2 = sys_async(sys_write, params);
> >
> > [ ... calculation loop ... ]
> >
> > wait_on_async_syscall(cookie1);
> > wait_on_async_syscall(cookie2);
> >
> > or something like that. Without user-space having to create threads
> > itself, etc. So basically, we'd make kernel threads more useful, and
> > we'd make threading safer - by only letting syscalls thread.
>
> Since I still think that the many-thousands potential async operations
> coming from network sockets are better handled with a classical event
> machanism [1], and since smooth integration of new async syscall into the
> standard POSIX infrastructure is IMO a huge win, I think we need to have a
> "bridge" to allow async completions being detectable through a pollable
> (by the mean of select/poll/epoll whatever) device.
> In that way you can handle async operations with the best mechanism that
> is fit for them, and gather them in a single async scheduler.
To clarify further, below are the API and the use case of my userspace
implementation. The guasi_fd() gives you back a pollable (POLLIN) fd to be
integrated in your prefered event retrieval interface. Once it fd is
signaled, you can fetch your completed requests using guasi_fetch() and
schedule work based on that.
The GUASI implementation uses pthreads, but it is clear that an in-kernel
async syscall implementation can take wiser decisions, and optimize the
heck out of it (locks, queues, ...).
- Davide
/*
* Example of async pread using GUASI
*/
static long guasi_wrap__pread(void *priv, long const *params) {
return (long) pread((int) params[0], (void *) params[1],
(size_t) params[2], (off_t) params[3]);
}
guasi_req_t guasi__pread(guasi_t hctx, void *priv, void *asid, int prio,
int fd, void *buf, size_t size, off_t off) {
return guasi_submit(hctx, priv, asid, prio, guasi_wrap__pread, 4,
(long) fd, (long) buf, (long) size, (long) off);
}
---
/*
* guasi by Davide Libenzi (generic userspace async syscall implementation)
* Copyright (C) 2003 Davide Libenzi
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program 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 General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* Davide Libenzi <[email protected]>
*
*/
#if !defined(_GUASI_H)
#define _GUASI_H
#define GUASI_MAX_PARAMS 16
#define GUASI_STATUS_PENDING 1
#define GUASI_STATUS_ACTIVE 2
#define GUASI_STATUS_COMPLETE 3
typedef long (*guasi_syscall_t)(void *, long const *);
typedef struct s_guasi { } *guasi_t;
typedef struct s_guasi_req { } *guasi_req_t;
struct guasi_reqinfo {
void *priv; /* Call private data. Passed to guasi_submit */
void *asid; /* Async request ID. Passed to guasi_submit */
long result; /* Return code of "proc" passed to guasi_submit */
long error; /* errno */
int status; /* GUASI_STATUS_* */
};
guasi_t guasi_create(int min_threads, int max_threads, int max_priority);
void guasi_free(guasi_t hctx);
int guasi_fd(guasi_t hctx);
guasi_req_t guasi_submit(guasi_t hctx, void *priv, void *asid, int prio,
guasi_syscall_t proc, int nparams, ...);
int guasi_fetch(guasi_t hctx, guasi_req_t *reqs, int nreqs);
int guasi_req_info(guasi_req_t hreq, struct guasi_reqinfo *rinf);
void guasi_req_free(guasi_req_t hreq);
#endif
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]