LOOPBACK PSEUDO INTERFACE 1 UPDATE
LOOPBACK PSEUDO INTERFACE 1 ARCHIVE
LOOPBACK PSEUDO INTERFACE 1 CODE
LOOPBACK PSEUDO INTERFACE 1 WINDOWS

11:13 ` richard -rw- weinberger 20:38 ` Martitz, Thomas Why can't you use any other pseudo interface? The only really question that matters is, why do you need a second

> * Generally, what's the proper way (if any) implement a secondary loopback > * Is there anyway around it so that I can have a custom loopback interface > particular line it begins to work, but obviously at the same time the And indeed after modifying my custom interface to include this > This suggests that each network namespace can only have one loopback > loopback.c:206 net->loopback_dev = dev

LOOPBACK PSEUDO INTERFACE 1 CODE

> After examing the code I found this line (in loopback_net_init()): > drivers/net/loopback.c the resulting device doesn't behave like the standard > However I'm facing a problem: It doesn't appear to be possible for a second > my work I want to develop a custom loopback interface (for testing > I'm Thomas Martitz, working on my Master Thesis: an Ethernet NIC. > (I'm sending this mail a second time (with a few people in CC) because I

To: Thomas Martitz +Cc: netdev, davem, edumazet, herbert, ebiederm 20:38 ` Martitz, Thomas 0 siblings, 1 reply 23+ messages in threadДЇrom: richard -rw- weinberger 11:13 UTC ( / raw) 12:05 Thomas Martitz 11:13 ` richard -rw- weinberger * Re: Trying to implement secondary loopback.* Generally, what's the proper way (if any) implement a secondary This particular line it begins to work, but obviously at the same time And indeed after modifying my custom interface to include This suggests that each network namespace can only have one loopback (I'm sending this mail a second time (with a few people in CC) because I To: netdev +Cc: davem, edumazet, herbert, ebiederm 11:13 ` richard -rw- weinberger 0 siblings, 1 reply 23+ messages in threadДЇrom: Thomas Martitz 12:05 UTC ( / raw) ^ raw reply 23+ messages in thread * Trying to implement secondary loopback 12:05 Thomas Martitz ^ raw reply 23+ messages in thread * Trying to implement secondary loopback 9:49 Thomas Martitz 0 siblings, 0 replies 23+ messages in threadДЇrom: Thomas Martitz 9:49 UTC ( / raw) OFC 2013 / March 19-21 / Anaheim Convention Center, CA, USA / booth 11807 Interface without touching lo's functionality.Д¬eBIT 2013 / March 5-9 / Hannover, Germany / hall 9, booth E08 * Is there anyway around it so that I can have a custom loopback Same time the standard lo interface stops working.

Include this particular line it begins to work, but obviously at the And indeed after modifying my custom interface to This suggests that the current network namespace can only have one Even when copying (and module'ifying)Дrivers/net/loopback.c the resulting device doesn't behave like theДЄfter examing the code I found this line (in loopback_net_init()): However I'm facing a problem: It doesn't appear to be possible for a Part of my work I want to develop a custom loopback interface (for I'm Thomas Martitz, working on my Master Thesis: an Ethernet NIC.

LOOPBACK PSEUDO INTERFACE 1 ARCHIVE

upgrade this variable to long with the next major version bump.Trying to implement secondary loopback Netdev Archive on help / color / mirror / Atom feed * Trying to implement secondary loopback 9:25 Thomas Martitz 0 siblings, 0 replies 23+ messages in threadДЇrom: Thomas Martitz 9:25 UTC ( / raw).change the toString() to use the conversion mentioned above.

LOOPBACK PSEUDO INTERFACE 1 UPDATE

update the javadoc to highlight the unsigned nature of the int.

I think the best thing to do at this point is to: However, that would be an API breaking change so will have to wait for the next major version bump. We could use OSHI to read that and overwrite the current value, but we're still stuck in the situation of having a 32-bit unsigned value that Java doesn't handle.ДЄ good long term solution would be to change OSHI's MTU to long type.

LOOPBACK PSEUDO INTERFACE 1 WINDOWS

Newer Windows versions (Vista+) use MIB_IF_ROW2 which has a 64-bit value and may correctly return the max unsigned int value inside that ULONG (or might return max unsigned long I haven't checked). So -1 does not represent an error here it faithfully reports the JDK's value, which is either correct or throws a SocketException, and reasonable to convert to a long using (long) mtu & 0xffff_ffffL (OSHI provides ParseUtil.unsignedIntToLong() that does this, and this is probably the best way to handle it at the user level right now.)ДЄlternately, you might consider that MTU is meaningless for local interfaces, and ignore the numeric value (which is essentially 'infinite' for practical purposes). This value is apparently unsigned 32-bit values at the native level (the JDK derives the value from the 32-bit dwMtu field in the MIB_IFROW structure for that interface). The MTU is a carry-through from the JDK's NetworkInterface class, in which getMTU() returns an int, which are always signed in Java.

Loopback pseudo interface 1