[Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Kazu]

Hi,

I updated a patch to support tap for win32.

Here is patches and a binary.
These patches are applied by patch -p1 option. Tap patch can be applied on
top of VLAN patch. VLAN patch is updated.
http://www.h7.dion.ne.jp/~qemu-win/download/qemu-20060111-vlan.patch
http://www.h7.dion.ne.jp/~qemu-win/download/qemu-20060111-tap.patch

http://www.h6.dion.ne.jp/~kazuw/qemu-win/qemu-20060111-vlan-tap.zip

Options are:
-net nic -net tap,ifname=my-tap
Use ifname to set a name of Tap.
my-tap is the name of  my TAP-Win32 Adapter.

Don't use the same network address as a physical network.
For example, if the physical network is 192.168.0.x, use different network 
192.168.10.y for TAP-Win32 Adapter.

For more information,
http://www.h7.dion.ne.jp/~qemu-win/TapWin32-en.html#vlan

Regards,
Kazu

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[André Braga]

Wow, thanks! I almost gave up on seeing someone port over this patch :)


--
"I decry the current tendency to seek patents on algorithms. There are
better ways to earn a living than to prevent other people from making
use of one's contributions to computer science."
Donald Knuth



On 1/31/06, Kazu <kazoo@r3.dion.ne.jp> wrote:
> Hi,
>
> I updated a patch to support tap for win32.
>
> Here is patches and a binary.
> These patches are applied by patch -p1 option. Tap patch can be applied on
> top of VLAN patch. VLAN patch is updated.
> http://www.h7.dion.ne.jp/~qemu-win/download/qemu-20060111-vlan.patch
> http://www.h7.dion.ne.jp/~qemu-win/download/qemu-20060111-tap.patch
>
> http://www.h6.dion.ne.jp/~kazuw/qemu-win/qemu-20060111-vlan-tap.zip
>
> Options are:
> -net nic -net tap,ifname=my-tap
> Use ifname to set a name of Tap.
> my-tap is the name of  my TAP-Win32 Adapter.
>
> Don't use the same network address as a physical network.
> For example, if the physical network is 192.168.0.x, use different network
> 192.168.10.y for TAP-Win32 Adapter.
>
> For more information,
> http://www.h7.dion.ne.jp/~qemu-win/TapWin32-en.html#vlan
>
> Regards,
> Kazu

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Fabrice Bellard]

Hi,

I merged your patches and I made important changes to simplify them. I
did not do any tests so tell me if you see problems.

Regards,

Fabrice.

Kazu wrote:
> Hi,
> 
> I updated a patch to support tap for win32.
> 
> Here is patches and a binary.
> These patches are applied by patch -p1 option. Tap patch can be applied on
> top of VLAN patch. VLAN patch is updated.
> http://www.h7.dion.ne.jp/~qemu-win/download/qemu-20060111-vlan.patch
> http://www.h7.dion.ne.jp/~qemu-win/download/qemu-20060111-tap.patch
> 
> http://www.h6.dion.ne.jp/~kazuw/qemu-win/qemu-20060111-vlan-tap.zip
> 
> Options are:
> -net nic -net tap,ifname=my-tap
> Use ifname to set a name of Tap.
> my-tap is the name of  my TAP-Win32 Adapter.
> 
> Don't use the same network address as a physical network.
> For example, if the physical network is 192.168.0.x, use different 
> network 192.168.10.y for TAP-Win32 Adapter.
> 
> For more information,
> http://www.h7.dion.ne.jp/~qemu-win/TapWin32-en.html#vlan
> 
> Regards,
> Kazu
> 
> 
> 
> 
> 
> _______________________________________________
> Qemu-devel mailing list
> Qemu-devel@nongnu.org
> http://lists.nongnu.org/mailman/listinfo/qemu-devel
> 
> 

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Kazu]

[-- Attachment #1: Type: text/plain, Size: 1706 bytes --]

Thursday, February 02, 2006 8:10 AM Fabrice Bellard wrote:

> Hi,
>
> I merged your patches and I made important changes to simplify them. I
> did not do any tests so tell me if you see problems.
>

-net socket,connect doesn't work. On Windows host, connect returns with err
= WSAEWOULDBLOCK and second time err = WSAEINVAL. I think changing the place
of EWOULDBLOCK would be good. On Linux host, EWOULDBLOCK is the same as
EAGAIN but a patch works on both Linux and Windows.

For -net socket,mcast, bind have to be done by sin_addr.s_addr = INADDR_ANY.
It seems that it works on Linux host.

Regards,
Kazu

Index: vl.c
===================================================================
RCS file: /sources/qemu/qemu/vl.c,v
retrieving revision 1.160
diff -u -r1.160 vl.c
--- vl.c 1 Feb 2006 23:06:55 -0000 1.160
+++ vl.c 2 Feb 2006 09:28:19 -0000
@@ -2335,7 +2335,13 @@
  goto fail;
     }

-    ret = bind(fd, (struct sockaddr *)mcastaddr, sizeof(*mcastaddr));
+    struct sockaddr_in addr;
+    memset(&addr, 0, sizeof(addr));
+    addr.sin_family = AF_INET;
+    addr.sin_port = mcastaddr->sin_port;
+    addr.sin_addr.s_addr = INADDR_ANY;
+
+    ret = bind(fd, (struct sockaddr *)&addr, sizeof(addr));
     if (ret < 0) {
         perror("bind");
         goto fail;
@@ -2561,8 +2567,8 @@
         ret = connect(fd, (struct sockaddr *)&saddr, sizeof(saddr));
         if (ret < 0) {
             err = socket_error();
-            if (err == EINTR || err == EWOULDBLOCK) {
-            } else if (err == EINPROGRESS) {
+            if (err == EINTR) {
+            } else if (err == EINPROGRESS || err == EWOULDBLOCK) {
                 break;
             } else {
                 perror("connect");

[-- Attachment #2: qemu-20060202-vlan.patch --]
[-- Type: application/octet-stream, Size: 1132 bytes --]

Index: vl.c
===================================================================
RCS file: /sources/qemu/qemu/vl.c,v
retrieving revision 1.160
diff -u -r1.160 vl.c
--- vl.c	1 Feb 2006 23:06:55 -0000	1.160
+++ vl.c	2 Feb 2006 09:28:19 -0000
@@ -2335,7 +2335,13 @@
 	goto fail;
     }
 
-    ret = bind(fd, (struct sockaddr *)mcastaddr, sizeof(*mcastaddr));
+    struct sockaddr_in addr;
+    memset(&addr, 0, sizeof(addr));
+    addr.sin_family = AF_INET;
+    addr.sin_port = mcastaddr->sin_port;
+    addr.sin_addr.s_addr = INADDR_ANY;
+
+    ret = bind(fd, (struct sockaddr *)&addr, sizeof(addr));
     if (ret < 0) {
         perror("bind");
         goto fail;
@@ -2561,8 +2567,8 @@
         ret = connect(fd, (struct sockaddr *)&saddr, sizeof(saddr));
         if (ret < 0) {
             err = socket_error();
-            if (err == EINTR || err == EWOULDBLOCK) {
-            } else if (err == EINPROGRESS) {
+            if (err == EINTR) {
+            } else if (err == EINPROGRESS || err == EWOULDBLOCK) {
                 break;
             } else {
                 perror("connect");

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Fabrice Bellard]

Kazu wrote:
> Thursday, February 02, 2006 8:10 AM Fabrice Bellard wrote:
> 
>> Hi,
>>
>> I merged your patches and I made important changes to simplify them. I
>> did not do any tests so tell me if you see problems.
>>
> 
> -net socket,connect doesn't work. On Windows host, connect returns with err
> = WSAEWOULDBLOCK and second time err = WSAEINVAL. I think changing the 
> place
> of EWOULDBLOCK would be good. On Linux host, EWOULDBLOCK is the same as
> EAGAIN but a patch works on both Linux and Windows.

Unfortunately on Linux the correct return value we are expecting is 
EINPROGRESS. EAGAIN means that the 'connect' was not initiated so it is 
necessary to redo it.

> 
> For -net socket,mcast, bind have to be done by sin_addr.s_addr = 
> INADDR_ANY.
> It seems that it works on Linux host.

It works on Linux but it is not correct because it prevents from 
listening to several multicast addresses at the same time. If doing the 
same on Windows is not possible I agree to make a special case.

Fabrice.

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Jim C. Brown]

On Thu, Feb 02, 2006 at 12:10:36AM +0100, Fabrice Bellard wrote:
> Hi,
> 
> I merged your patches and I made important changes to simplify them. I
> did not do any tests so tell me if you see problems.
> 
> Regards,
> 
> Fabrice.
> 

Have you decided to accept the GPL license on it then?

http://lists.gnu.org/archive/html/qemu-devel/2004-10/msg00217.html

-- 
Infinite complexity begets infinite beauty.
Infinite precision begets infinite perfection.

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Jim C. Brown]

On Thu, Feb 02, 2006 at 12:10:36AM +0100, Fabrice Bellard wrote:
> Hi,
> 
> I merged your patches and I made important changes to simplify them. I
> did not do any tests so tell me if you see problems.
> 
> Regards,
> 
> Fabrice.
> 

Have you decided to accept the GPL license on it then?

http://lists.gnu.org/archive/html/qemu-devel/2004-10/msg00217.html

-- 
Infinite complexity begets infinite beauty.
Infinite precision begets infinite perfection.

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Fabrice Bellard]

Jim C. Brown wrote:
> On Thu, Feb 02, 2006 at 12:10:36AM +0100, Fabrice Bellard wrote:
> 
>>Hi,
>>
>>I merged your patches and I made important changes to simplify them. I
>>did not do any tests so tell me if you see problems.
>>
>>Regards,
>>
>>Fabrice.
>>
> 
> 
> Have you decided to accept the GPL license on it then?
> 
> http://lists.gnu.org/archive/html/qemu-devel/2004-10/msg00217.html

Yes because the features are packed in a single file which can be 
removed if needed.

Fabrice.

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[David Fraser]

Fabrice Bellard wrote:
> Jim C. Brown wrote:
>> On Thu, Feb 02, 2006 at 12:10:36AM +0100, Fabrice Bellard wrote:
>>
>>> Hi,
>>>
>>> I merged your patches and I made important changes to simplify them. I
>>> did not do any tests so tell me if you see problems.
>>>
>>> Regards,
>>>
>>> Fabrice.
>>>
>>
>>
>> Have you decided to accept the GPL license on it then?
>>
>> http://lists.gnu.org/archive/html/qemu-devel/2004-10/msg00217.html
>
> Yes because the features are packed in a single file which can be 
> removed if needed.
Just checking, would that make it illegal to use this feature together 
with kqemu?


David

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[M. Warner Losh]

In message: <43E99C81.6040302@sjsoft.com>
            David Fraser <davidf@sjsoft.com> writes:
: Fabrice Bellard wrote:
: > Jim C. Brown wrote:
: >> On Thu, Feb 02, 2006 at 12:10:36AM +0100, Fabrice Bellard wrote:
: >>
: >>> Hi,
: >>>
: >>> I merged your patches and I made important changes to simplify them. I
: >>> did not do any tests so tell me if you see problems.
: >>>
: >>> Regards,
: >>>
: >>> Fabrice.
: >>>
: >>
: >>
: >> Have you decided to accept the GPL license on it then?
: >>
: >> http://lists.gnu.org/archive/html/qemu-devel/2004-10/msg00217.html
: >
: > Yes because the features are packed in a single file which can be 
: > removed if needed.
: Just checking, would that make it illegal to use this feature together 
: with kqemu?

No.  The GPL is about distribution, not use.

It might make certain binary packages contrary to rights granted by
the GPL.

Warner

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Kazu]

[-- Attachment #1: Type: text/plain, Size: 5582 bytes --]

Sunday, February 05, 2006 3:28 AM Fabrice  Bellard wrote:
> Kazu wrote:
>> Thursday, February 02, 2006 8:10 AM Fabrice Bellard wrote:
>>
>>> Hi,
>>>
>>> I merged your patches and I made important changes to simplify them. I
>>> did not do any tests so tell me if you see problems.
>>>
>>
>> -net socket,connect doesn't work. On Windows host, connect returns with
>> err
>> = WSAEWOULDBLOCK and second time err = WSAEINVAL. I think changing the
>> place
>> of EWOULDBLOCK would be good. On Linux host, EWOULDBLOCK is the same as
>> EAGAIN but a patch works on both Linux and Windows.
>
> Unfortunately on Linux the correct return value we are expecting is
> EINPROGRESS. EAGAIN means that the 'connect' was not initiated so it is
> necessary to redo it.
>

There is not good way to detect a completion of asynchronous connect. So I
used an event object to detect it. A patch is attached.

>>
>> For -net socket,mcast, bind have to be done by sin_addr.s_addr =
>> INADDR_ANY.
>> It seems that it works on Linux host.
>
> It works on Linux but it is not correct because it prevents from listening
> to several multicast addresses at the same time. If doing the same on
> Windows is not possible I agree to make a special case.
>

I couldn't find a way to set a multicast address. I made it a special case
in the patch

Regard,
Kazu

Index: vl.c
===================================================================
RCS file: /sources/qemu/qemu/vl.c,v
retrieving revision 1.162
diff -u -r1.162 vl.c
--- vl.c 5 Feb 2006 04:14:41 -0000 1.162
+++ vl.c 8 Feb 2006 05:50:28 -0000
@@ -1083,9 +1083,10 @@

 #define socket_error() WSAGetLastError()
 #undef EINTR
-#define EWOULDBLOCK WSAEWOULDBLOCK
-#define EINTR       WSAEINTR
-#define EINPROGRESS WSAEINPROGRESS
+#define EWOULDBLOCK   WSAEWOULDBLOCK
+#define EINTR         WSAEINTR
+#define EINPROGRESS   WSAEINPROGRESS
+#define EADDRNOTAVAIL WSAEADDRNOTAVAIL

 static void socket_cleanup(void)
 {
@@ -1136,6 +1137,59 @@
     ioctlsocket(fd, FIONBIO, &opt);
 }

+static int socket_set_event(SOCKET fd, WSAEVENT *phEvent)
+{
+    int ret;
+
+    *phEvent = WSACreateEvent();
+    if (*phEvent == WSA_INVALID_EVENT) {
+        perror("connect: CreateEvent");
+        return -1;
+    }
+
+    ret = WSAEventSelect(fd, *phEvent, FD_CONNECT);
+    if (ret == SOCKET_ERROR) {
+        perror("connect: EventSelect");
+        return -1;
+    }
+    return 0;
+}
+
+static int socket_wait_event(SOCKET fd, WSAEVENT *phEvent)
+{
+    WSANETWORKEVENTS events;
+    int ret;
+
+    ret = WSAWaitForMultipleEvents(1, phEvent, FALSE, WSA_INFINITE, FALSE);
+    if (ret == WSA_WAIT_FAILED) {
+        perror("connect: Wait");
+        goto fail;
+    }
+
+    ret = WSAEnumNetworkEvents(fd, *phEvent, &events);
+    if (ret == SOCKET_ERROR) {
+        perror("connect: EnumEvent");
+        goto fail;
+    } else {
+        if (events.lNetworkEvents & FD_CONNECT) {
+            if (events.iErrorCode[FD_CONNECT_BIT] == 0) {
+                if (*phEvent != WSA_INVALID_EVENT)
+                    WSACloseEvent(*phEvent);
+                return 1;
+            } else {
+                perror("connect: refused");
+                goto fail;
+            }
+        } else {
+            perror("connect: fd_connect");
+            goto fail;
+        }
+    }
+ fail:
+    if (*phEvent != WSA_INVALID_EVENT)
+        WSACloseEvent(*phEvent);
+    return 0;
+}
 #else

 #define socket_error() errno
@@ -2330,7 +2384,9 @@
 {
     struct ip_mreq imr;
     int fd;
-    int val, ret;
+    int val, ret, err;
+    struct sockaddr_in addr;
+
     if (!IN_MULTICAST(ntohl(mcastaddr->sin_addr.s_addr))) {
  fprintf(stderr, "qemu: error: specified mcastaddr \"%s\" (0x%08x) does not
contain a multicast address\n",
   inet_ntoa(mcastaddr->sin_addr),
@@ -2354,8 +2410,22 @@

     ret = bind(fd, (struct sockaddr *)mcastaddr, sizeof(*mcastaddr));
     if (ret < 0) {
-        perror("bind");
-        goto fail;
+        err = socket_error();
+        if (err == EADDRNOTAVAIL) {
+            memset(&addr, 0, sizeof(addr));
+            addr.sin_family = AF_INET;
+            addr.sin_port = mcastaddr->sin_port;
+            addr.sin_addr.s_addr = INADDR_ANY;
+
+            ret = bind(fd, (struct sockaddr *)&addr, sizeof(addr));
+            if (ret < 0) {
+                perror("bind");
+                goto fail;
+            }
+        } else {
+            perror("bind");
+            goto fail;
+        }
     }

     /* Add host to multicast group */
@@ -2557,11 +2627,15 @@
     return 0;
 }

+
 static int net_socket_connect_init(VLANState *vlan, const char *host_str)
 {
     NetSocketState *s;
     int fd, connected, ret, err;
     struct sockaddr_in saddr;
+#ifdef _WIN32
+    WSAEVENT hEvent;
+#endif

     if (parse_host_port(&saddr, host_str) < 0)
         return -1;
@@ -2573,6 +2647,26 @@
     }
     socket_set_nonblock(fd);

+#ifdef _WIN32
+    ret = socket_set_event(fd, &hEvent);
+    if (ret < 0) {
+        perror("connect: set_event");
+        closesocket(fd);
+        return -1;
+    }
+
+    connected = 0;
+    ret = connect(fd, (struct sockaddr *)&saddr, sizeof(saddr));
+
+    err = socket_wait_event(fd, &hEvent);
+    if (err > 0) {
+        connected = 1;
+    } else {
+        perror("connect");
+        closesocket(fd);
+        return -1;
+    }
+#else
     connected = 0;
     for(;;) {
         ret = connect(fd, (struct sockaddr *)&saddr, sizeof(saddr));
@@ -2591,6 +2685,8 @@
             break;
         }
     }
+#endif
+
     s = net_socket_fd_init(vlan, fd, connected);
     if (!s)
         return -1;


[-- Attachment #2: qemu-20060208-vlan-2.patch.gz --]
[-- Type: application/x-gzip, Size: 1475 bytes --]

Re: [Qemu-devel][PATCH] Tap and VLAN socket support for win32

[Kazu]

[-- Attachment #1: Type: text/plain, Size: 1431 bytes --]

Hi,

I send this mail again because it doesn't reach qemu-devel list by unknown
security reason.

Sent: Sunday, February 05, 2006 3:28 AM Fabrice Bellard wrote:

> Kazu wrote:
>> Thursday, February 02, 2006 8:10 AM Fabrice Bellard wrote:
>>
>>> Hi,
>>>
>>> I merged your patches and I made important changes to simplify them. I
>>> did not do any tests so tell me if you see problems.
>>>
>>
>> -net socket,connect doesn't work. On Windows host, connect returns with
err
>> = WSAEWOULDBLOCK and second time err = WSAEINVAL. I think changing the
>> place
>> of EWOULDBLOCK would be good. On Linux host, EWOULDBLOCK is the same as
>> EAGAIN but a patch works on both Linux and Windows.
>
> Unfortunately on Linux the correct return value we are expecting is
> EINPROGRESS. EAGAIN means that the 'connect' was not initiated so it is
> necessary to redo it.
>

There is not good way to detect a completion of asynchronous connect. So I
used an event object to detect it. A patch is attached.

>>
>> For -net socket,mcast, bind have to be done by sin_addr.s_addr =
>> INADDR_ANY.
>> It seems that it works on Linux host.
>
> It works on Linux but it is not correct because it prevents from
> listening to several multicast addresses at the same time. If doing the
> same on Windows is not possible I agree to make a special case.
>

I couldn't find a way to set a multicast address. I made it a special case
in the patch

Regards,
Kazu

[-- Attachment #2: qemu-20060208-vlan-2.patch.gz --]
[-- Type: application/x-gzip, Size: 1475 bytes --]