qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Laurent Vivier <Laurent.Vivier@bull.net>
To: "Ciprian Dorin, Craciun" <ciprian.craciun@gmail.com>
Cc: nbd-general@lists.sourceforge.net, linux-newbie@vger.kernel.org,
	qemu-devel@nongnu.org
Subject: [Qemu-devel] Re: [Nbd] Transforming stdin and stdout pair into a socket
Date: Mon, 11 May 2009 10:03:43 +0200	[thread overview]
Message-ID: <1242029023.4638.3.camel@frecb07144> (raw)
In-Reply-To: <8e04b5820905101119u39ae84bfp1bae24be6e057086@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 2270 bytes --]

Le dimanche 10 mai 2009 à 21:19 +0300, Ciprian Dorin, Craciun a écrit :
> Hello all!

Hi,

perhaps the attached patch I wrote last year (november) is what you
want...

I didn't try to apply it to an up-to-date qemu-nbd.

Regards,
Laurent

>     Today I've played around with NBD (Network Block Disk), and
> qemu-nbd (a NBD client that exports QEMU disks as NBD's).
> 
>     My problem is the following: both NBD kernel module and qemu-nbd
> implementation expect to use a socket in order to communicate.
>     This means that in order to securely tunnel the connection over
> SSH (OpenSSH), I need an intermediary process that creates a socket
> and forwards all input / output between this socket and stdin / stdout
> (which are in fact pipes received from OpenSSH).
> 
>     My question is: can I somehow make the pair of stdin / stdout seem
> as a socket to the Linux syscalls (read and write)? (I would have to
> make stdin / stdout pair look like a single file descriptor.) (This
> would eliminate the intermediate process that just pipes data, and
> thus reduce the overhead.)
> 
>     Just to be clear: I know how to trick an application to have it's
> stdin and stdout be an opened socket (by using dup syscall). But in
> this case I need to trick the Linux kernel into thinking that stdin /
> stdout pair is a socket (or a single file descriptor).
> 
>     Thank you,
>     Ciprian Craciun.
> 
> ------------------------------------------------------------------------------
> The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your
> production scanning environment may not be a perfect world - but thanks to
> Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700
> Series Scanner you'll get full speed at 300 dpi even with all image 
> processing features enabled. http://p.sf.net/sfu/kodak-com
> _______________________________________________
> Nbd-general mailing list
> Nbd-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/nbd-general
> 
-- 
------------------ Laurent.Vivier@bull.net  ------------------
"Tout ce qui est impossible reste à accomplir"    Jules Verne
"Things are only impossible until they're not" Jean-Luc Picard

[-- Attachment #2: qemu-nbd-inetd.patch --]
[-- Type: text/x-vhdl, Size: 3645 bytes --]

---
 qemu-nbd.c |   41 +++++++++++++++++++++++++++++++++++------
 1 file changed, 35 insertions(+), 6 deletions(-)

Index: qemu/qemu-nbd.c
===================================================================
--- qemu.orig/qemu-nbd.c	2008-09-11 17:06:05.000000000 +0200
+++ qemu/qemu-nbd.c	2008-09-15 16:10:37.000000000 +0200
@@ -57,6 +57,7 @@ static void usage(const char *name)
 "  -d, --disconnect     disconnect the specified device\n"
 "  -e, --shared=NUM     device can be shared by NUM clients (default '1')\n"
 "  -t, --persistent     don't exit on the last connection\n"
+"  -i, --inetd          inetd interface: use stdin/stdout instead of a socke\n"
 "  -v, --verbose        display extra debugging information\n"
 "  -h, --help           display this help and exit\n"
 "  -V, --version        output version information and exit\n"
@@ -183,14 +184,14 @@ int main(int argc, char **argv)
     bool readonly = false;
     bool disconnect = false;
     const char *bindto = "0.0.0.0";
-    int port = 1024;
+    int port = 0;
     struct sockaddr_in addr;
     socklen_t addr_len = sizeof(addr);
     off_t fd_size;
     char *device = NULL;
     char *socket = NULL;
     char sockpath[128];
-    const char *sopt = "hVbo:p:rsnP:c:dvk:e:t";
+    const char *sopt = "hVbo:p:rsnP:c:dvk:e:ti";
     struct option lopt[] = {
         { "help", 0, 0, 'h' },
         { "version", 0, 0, 'V' },
@@ -207,6 +208,7 @@ int main(int argc, char **argv)
         { "shared", 1, 0, 'e' },
         { "persistent", 0, 0, 't' },
         { "verbose", 0, 0, 'v' },
+        { "inetd", 0, 0, 'i' },
         { NULL, 0, 0, 0 }
     };
     int ch;
@@ -225,6 +227,7 @@ int main(int argc, char **argv)
     int nb_fds = 0;
     int max_fd;
     int persistent = 0;
+    int inetd = 0;
 
     while ((ch = getopt_long(argc, argv, sopt, lopt, &opt_ind)) != -1) {
         switch (ch) {
@@ -289,6 +292,9 @@ int main(int argc, char **argv)
 	case 't':
 	    persistent = 1;
 	    break;
+        case 'i':
+	    inetd = 1;
+	    break;
         case 'v':
             verbose = 1;
             break;
@@ -326,6 +332,18 @@ int main(int argc, char **argv)
 	return 0;
     }
 
+    if (inetd) {
+        if (shared != 1)
+           errx(EINVAL, "You cannot use inetd and shared");
+        if (socket)
+           errx(EINVAL, "You cannot use inetd and socket");
+        if (port)
+           errx(EINVAL, "You cannot use inetd and port");
+    } else {
+        if (!socket)
+            port = 1024;
+    }
+
     bdrv_init();
 
     bs = bdrv_new("hda");
@@ -412,9 +430,24 @@ int main(int argc, char **argv)
     if (sharing_fds == NULL)
         errx(ENOMEM, "Cannot allocate sharing fds");
 
+    data = qemu_memalign(512, NBD_BUFFER_SIZE);
+    if (data == NULL)
+        errx(ENOMEM, "Cannot allocate data buffer");
+
     if (socket) {
         sharing_fds[0] = unix_socket_incoming(socket);
     } else {
+        if (inetd) {
+            /* read and write on stdin/stdout */
+            ret = nbd_negotiate(STDIN_FILENO, fd_size);
+            while (ret != -1) {
+                ret = nbd_trip(bs, STDIN_FILENO, fd_size, dev_offset,
+                               &offset, readonly, data, NBD_BUFFER_SIZE);
+            }
+            qemu_free(data);
+            bdrv_close(bs);
+            return 0;
+        }
         sharing_fds[0] = tcp_socket_incoming(bindto, port);
     }
 
@@ -423,10 +456,6 @@ int main(int argc, char **argv)
     max_fd = sharing_fds[0];
     nb_fds++;
 
-    data = qemu_memalign(512, NBD_BUFFER_SIZE);
-    if (data == NULL)
-        errx(ENOMEM, "Cannot allocate data buffer");
-
     do {
 
         FD_ZERO(&fds);

  reply	other threads:[~2009-05-11  8:00 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-05-10 18:19 [Qemu-devel] Transforming stdin and stdout pair into a socket Ciprian Dorin, Craciun
2009-05-11  8:03 ` Laurent Vivier [this message]
2009-05-11 13:28   ` [Qemu-devel] Re: [Nbd] " Ciprian Dorin, Craciun
2009-05-11 12:02 ` [Qemu-devel] " Anthony Liguori
2009-05-11 13:32   ` Ciprian Dorin, Craciun
2009-05-11 19:31     ` [Nbd] " Wouter Verhelst
2009-05-11 23:07     ` Paul Brook
2009-05-12  2:24       ` Anthony Liguori

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1242029023.4638.3.camel@frecb07144 \
    --to=laurent.vivier@bull.net \
    --cc=ciprian.craciun@gmail.com \
    --cc=linux-newbie@vger.kernel.org \
    --cc=nbd-general@lists.sourceforge.net \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).