net_interface & datalink: Fix MTU handling.

The device is what actually controls the MTU, and it has its own
field for this, so having a second one just meant the MTU never
got updated after startup.

Remove the "mtu" field from the interface, use the "device->mtu" directly,
and then actually invoke device->module->set_mtu when updating.

network stack: Fix deadlock in get_interface functions.

If these methods are called while the interface in question
is receiving data via a receive thread, we can hit a deadlock
where a receive thread is holding the receive lock and then
tries to call get_interface_address_for_link (due, e.g., to ipv4
checking is_local_link_address), which tries to acquire the interfaces
lock, while at the same time we are trying to acquire the receive lock
due to CreateDomainDatalinkIfNeeded invoking a module's datalink_init
which calls register_device_handler, so we deadlock. (Whew!)

As far as I can tell, we do not need to set Busy() here despite unlocking
the interfaces lock, as the Interface acquires its own lock in
CreateDomainDatalinkIfNeeded.

Observed in VMware when the DHCP client spins for a long time,
and the deadlock occurs upon opening Network preferences.

network stack: Add missing ASSERT_LOCKED_RECURSIVE.

More trivial syntax and logic cleanup.

Spotted by Clang. No functional change intended.

network stack: move put_device_interface() at the end of the Interface destructor.

Fixes #10919.

Removing an interface triggers this sequence:

datalink_control() calls interface->ReleaseReference() which causes the
Interface destructor to be called. The Interface destructor calls
put_device_interface(fDeviceInterface) which destroys the
net_device_interface, then calls put_domain_datalink_protocols(this, datalink->domain)
in a loop. put_domain_datalink_protocols() tries to access the deleted
net_device_interface, hence the KDL.

Change-Id: I2326bcd6d1fd80a69e5fdfa6629563b38ecdbbac
Reviewed-on: https://review.haiku-os.org/469
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>

InterfaceAddress: The real fix.

Revert "Follow-up fix for #12152 from axeld suggestions"

This reverts commit 6f5971cf52bcab0f9cfcc203bdc8b62e3c9d8fec.

* Sorry there was a misunderstanding in the Ticket and IRC.

Follow-up fix for #12152 from axeld suggestions

When realloc(address, size) returns NULL, address must be freed manually.

Signed-off-by: Dario Casalinuovo <b.vitruvio@gmail.com>

Fix naming

Improve patch in ticket #9377
Instead of locking the interface lock, set it busy and then unlock the
interface list lock.

Revert "Missing NULL check in network stack."

This reverts commit dd0ea64b5607e6310c582f62c4275dab7510c4f6.

Missing NULL check in network stack.

This would trigger a KDL, it's better to just fail the syscall call.

Add a TODO comment

Converted the "net_interfaces" mutex into a recursive locker, and
acquire it in Interface::SetDown().

Fixed up network stack and drivers for x86_64.

* Various compilation fixes.
* Fixes to the FreeBSD compatibility layer (from comparing the x86-
specific bits with the equivalent amd64 sources in FreeBSD).
* Compile all the Ethernet drivers except for sis900 and wb840, these
require a bit more work to fix (will file a ticket soon). Tested
ipro1000 and rtl81xx, no issues.

All debugger commands we've added, remove we must.
Closes #7251.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40612 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed ignoring the netmask when computing the broadcast. Renamed a few
variables to make the function a bit clearer, hopefully.
* Fixed the wrong order of releasing the reference, and removing the object
from the hash table -- thanks Rene!
* Added a bit of documentation to InterfaceAddress::Prepare().


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40455 a95241bf-73f2-0310-859d-f6bbb57e9c96

* RemoveAddresses() needs to remove the addresses from the hash table as well.
This fixes random memory access on interface removal.
* Optimized IndexOfAddress() for NULL addresses.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40447 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Allow duplicates in the sAddressTable hash table; while it makes no sense to
have duplicates there, it should certainly not panic.
* Do not add unspecified addresses to the hash table.
* The result of adding the initial address of an interface was ignored; now, the
interface is correctly destructed, if necessary.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40444 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added a bit more debug output.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@39388 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed a bug in get_interface() that would return the first interface
without acquiring a reference to it, and thus led to bug #6565.
* Added a commented out function that dumps all current reference counts.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38867 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Initialized the domain_datalink's first_{protocol|info} members. This fixes
the crashing bug reported by Atis when trying to add an interface which fails
(the bug causing this has already been fixed with the previous commit).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38180 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added ability to be able to dump a net_interface by name.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38117 a95241bf-73f2-0310-859d-f6bbb57e9c96

* B_SOCKET_SET_ALIAS handling was leaking references.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38079 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed memory leaks pointed out by Stephan, thanks!
* Automatic whitespace cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38068 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Also report the address index back to the userland.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38059 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Also create a new address for B_SOCKET_SET_ALIAS if the index is -1 - this
now has an almost equivalent logic to the old setters like SIOCSIFADDR.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38054 a95241bf-73f2-0310-859d-f6bbb57e9c96

* If the network mask, and broadcast are not specified with a B_SOCKET_SET_ALIAS
they will no longer be unset - instead they are set with defaults.
* If B_SOCKET_SET_ALIAS gets an index of -1, it will now try to find the local
address, and if that fails, will just use the first address there is.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38053 a95241bf-73f2-0310-859d-f6bbb57e9c96

* datalink_control() also need to accept structures smaller than ifreq as long
as the interface name can be specified.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38015 a95241bf-73f2-0310-859d-f6bbb57e9c96

* As Atis pointed out, the SIOCSIFNETMASK here didn't make any sense, must have
been a copy&paste error.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38001 a95241bf-73f2-0310-859d-f6bbb57e9c96

* We need to acquire references to the interface_address when copying/cloning
the buffer, and we also need to release it when freeing it.
* Adapted patch by Atis: append_cloned_data() now preserves the stored header;
this is kind of bad, because it will not notice if the original header is
removed, but this could probably just be documented this way.
* Also print the reference count of the interface addresses.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37993 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Made the work-around that Atis added to ifconfig and the net_server
in order to be able to use SIOCSIFADDR (and friends) on a net yet added
family superfluous; we now add a new address when needed on the fly.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37990 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed the _SIZEOF_ADDR_IFREQ() macro such that it can be used how we are
using it - hopefully, that is actually correct (it's adopted from FreeBSD).
* Fixed bug that cut off the sockaddr_dl reported by the system, so that DHCP
wouldn't work anymore (and ifconfig wouldn't show the actual MAC address
anymore).
* Changed the listing code to actually pad to ifreq size, and leave the length
of the sockaddr untouched.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37941 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Interface::_ChangeAddress() got the address check for equality wrong,
effectively rejecting any try to set an address.
* It now copies the new address as well in order to make sure that its
sa_len field is set correctly.
* Improved debug output.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37937 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Only actually call change_address() of the net_datalink_protocols if the
address actually changed.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37928 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Make sure the SIOCGIFCONF never returns an address with an address length
smaller than sizeof(sockaddr). This fixes a compatibility issue with other
platforms - portable software often assumes that the amount of bytes to
add to an ifreq structure is the larger amount between sizeof(ifreq), and
basically what the _SIZEOF_ADDR_IFREQ() macro returns, instead of always
relying on that macro.
* Renamed UserBuffer::Copy() to Push, ConsumedAmount() to BytesConsumed(),
added Pad() method.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37927 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Enabled removing the interface on device removal again.
* Made the return type of remove_interface() void, as it cannot fail.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37884 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed the proprietary SIOC_* ioctls to B_SOCKET_* - no reason to pollute
global name space, and have ugly identifiers for nothing :-)
* Added a flags field to struct ifaliasreq. Added flags to mark an alias that
is currently being configured, or has been automatically configured.
Those flags aren't used yet, but they will replace IFF_CONFIGURING and
friends.
* Implemented deleting addresses only from interfaces via ifconfig.
* Added more command aliases for delete to ifconfig ("del", and "delete", for
more consistency with route).
* Fixed control_routes() to only release a reference to an address if it
actually got one before.
* If an interface address is deleted, its routes are now removed as well.
* InterfaceAddress now holds a reference to its interface as planned.
* Implemented removing interfaces. Works quite nicely.
* When downing an interface, all of its routes are now removed. When upping
it again, at least the default routes are added.
* datalink.cpp's get_interface_name_or_index() leaked a reference to the
interface found.
* SIOCAIFADDR would also leak a reference when new addresses were added.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37872 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Work in progress of extending the AF_LINK protocol to be able to send and
receive raw packets (only without the ethernet framing).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37853 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added Haiku specific socket ioctls to configure the interface aliases:
SIOC_IF_ALIAS_ADD, SIOC_IF_ALIAS_REMOVE, SIOC_IF_ALIAS_GET, SIOC_ALIAS_SET,
and SIOC_IF_ALIAS_COUNT.
* Implemented all of those new ioctls, though they are yet untested.
* Added ifreq::ifr_data, and removed the hack in the FreeBSD compat if.h
header.
* Minor cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37806 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Reworked the complete stack to allow more than one address per network
interface - this caused quite a number of changes.
* Network interfaces, and its addresses are now reference counted (not yet
complete, though, InterfaceAddresses need to hold references to their
interface as well).
* There are two known regressions of this commit that I will fix later:
- you cannot remove interfaces anymore
- IPv4 multicast was broken anyway, but now it's disabled, too.
* Moved a device_interfaces.cpp|h out of interfaces.cpp.
* The datalink layer chain is now instantiated per domain per interface,
not just per interface anymore.
* When a buffer reaches the network layer, it has no known interface yet, ie.
the ipv4|6|whatever modules need to set this manually.
* Added more debug output, and some new debugger commands, the control option
is now printed in clear text.
* Added hash_address() function to the address modules. Added "const" to
set_to_defaults() where needed.
* Fixed net_buffer's restore header functions offset use as reported by Atis.
* Improved buffer dump output, use the domain module to print the address if
available.
* Moved net_buffer::type into the union, as it's not needed by the upper layers
anymore.
* Moved IPv6 specific code from {add|remove}_default_route() to where it
belongs, but disabled it for the time being.
* Completely discarded useless ipv4_datagram module.
* Added ping6 to the build.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37794 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Made some internal lists use DoublyLinkedLists instead of struct list.
* Added a few KDL commands to improve your debugging experience.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35469 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Locking order dictates that this lock is an inner lock, and shouldn't
be used everywhere.
* This also fixes a panic when deleting interfaces (an already deleted
lock was unlocked).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29902 a95241bf-73f2-0310-859d-f6bbb57e9c96

* The net_domain's lock is now a recursive lock.
* Fixed all route locking problems, of which there were numerous
({add|remove}_route(), and list_routes() did not lock at all). Added
lock assertions in functions that don't do the locking themselves.
* A route will now be removed from the list in remove_route(), not in
put_route_internal(). Before, a route could easily be removed twice, causing
remove_route() to release references it did not own. This fixes bug #2706.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29386 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed net_device_interface::rx_lock to receive_lock.
* Cleanup, improved comments, removed useless ones.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29232 a95241bf-73f2-0310-859d-f6bbb57e9c96

* More or less completed the network notification module - it does not remove
old invalid user listeners yet, though (ie. if a team dies).
* Implemented userland network monitor functions.
* Added a few notifications to the network stack, even though this part isn't
complete yet (especially notify_interface_changed()).
* Added optional debug output to the notifications module.
* Added the module to the image, it basically works now (tested).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28825 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Actually implemented the SO_BINDTODEVICE socket option I added some time ago.
* This makes it possible to select a specific device, even if no interface
has been configured for it yet. To make it work, each interface now has a
private direct route which will be returned if a socket is bound to a device.
* This will be used for example in DHCP to make it work when more than one
adapter is attached.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27983 a95241bf-73f2-0310-859d-f6bbb57e9c96

* put_device_interface() accessed "device" which was already freed to get
the module name, possibly causing a crash when you delete an interface.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27950 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Replaced all instances of benaphores in the kernel code by mutexes.
* Removed kernel benaphores.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25690 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed recursive_lock to use a mutex instead of a semaphore.
* Adjusted code using recursive locks respectively. The initialization
cannot fail anymore, and it is possible to use recursive locks in the
early boot process (even uninitialized, if in BSS), which simplifies
things a little.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25687 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed the way the socket functions work:
- The net_stack driver is no longer used. Instead we have a kernel
module which is directly used by syscall implementations in the
kernel. I.e. we no longer tunnel those functions through ioctls, but
have normal syscalls.
- Removed the superfluous net starter module.
- Implemented the FDTYPE_SOCKET type file_descriptors, that is sockets
are no longer vnode based.
- Adjusted libnetwork (the socket function implementations)
accordingly.
- Adjusted netstat accordingly.
* Socket module:
- Implemented socketpair().
- Added "kernel" parameter to the control hook. Quite a few more hooks
would actually need the parameter, but I didn't change those yet,
since that would probably also require changes to the protocol
module interface.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@24908 a95241bf-73f2-0310-859d-f6bbb57e9c96

There is absolutely no need that the interface packet consumer runs at a
priority that high.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22894 a95241bf-73f2-0310-859d-f6bbb57e9c96

net_device_interface: remove redundant fields


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21222 a95241bf-73f2-0310-859d-f6bbb57e9c96

loop: no longer requires a reader thread, it delivers directly to the device's receive queue


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21215 a95241bf-73f2-0310-859d-f6bbb57e9c96

net_interface_private: added receive queue and splited device reading from packet processing. Delivering to self no longer is executed in the sender's context, which had some problems with TCP's locking.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21214 a95241bf-73f2-0310-859d-f6bbb57e9c96

retrieve the incoming interface on the domain_receive_adapter(). This should fix potential issues with packets addressed to the IP broadcast address, such as in with some DHCP implementations.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20767 a95241bf-73f2-0310-859d-f6bbb57e9c96

packet capture sockets are now properly notified when the device is going down.

- modified LinkProtocol::ReadData() a bit so it can safely react to external changes such as the device being monitored going down.
- fixed an issue in invalidate_routes() where in some cases the default route was being kept.
- fixed another issue related with the fifo implementation, the notifier semaphore was being created with count of 1, instead of 0.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20622 a95241bf-73f2-0310-859d-f6bbb57e9c96

whenever an interface is deleted, call put_domain_datalink_protocols() so all readers are unregistered.

- Unfortunely this requires RX lock to become a recursive lock.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20621 a95241bf-73f2-0310-859d-f6bbb57e9c96

notify all monitors when a device is being removed.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20615 a95241bf-73f2-0310-859d-f6bbb57e9c96

Prepared net_device_monitor to accept device removal events.

- Introduced public net_device_monitor.
- Changed the link protocol to maintain a lock per instance instead of inside the FIFO. Now all of the link instance data is protected.
- Adapted the link protocol to use net_device_monitor.
- Introduced a private Fifo class which doesn't maintain it's own lock.
- Maybe we should add something like a public net_protocol_implementation which maintains a fifo and a benaphore? With the fifo using the structure's lock instead of maintaining it's own.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20614 a95241bf-73f2-0310-859d-f6bbb57e9c96

moved IFF_LINK handling to the device module (ethernet in this case). Now domain interfaces only keep specific flags such as IFF_UP and the configuration flags. IFF_LINK, IFF_BROADCAST etc are maintained exclusively by the device.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20612 a95241bf-73f2-0310-859d-f6bbb57e9c96

introduced net_device_interface level locking.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20611 a95241bf-73f2-0310-859d-f6bbb57e9c96

a bit more work towards proper locking including a fix to a refcount bug

- fixed a issue in add_interface_to_domain where the device interface's refcount was always incremented since that function was getting the device interface handle and not returning it unconditionlly
- if the ethernet device goes down, and the fd is close()ed, return B_FILE_ERROR instead of calling into the driver again


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20601 a95241bf-73f2-0310-859d-f6bbb57e9c96

started some work to properly handle device_removed() and setting interfaces down in general.

- remove all routes that use interfaces going down.
- when a device is going down, remove associated domain interfaces.
- interfaces weren't getting a properly referenced device interface, fixed.
- down call down_device_interface when deleting a interface, instead set it down and let the upcounts do its job.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20600 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed ETHER_GET_LINK_STATE ethernet driver interface, added ETHER_SET_LINK_STATE_SEM.
* The device interface list now uses class DoublyLinkedList instead of struct list.
* Implemented SIOC[SG]IFMEDIA for setting (not supported by any device yet), and
retrieving the device media information.
* Fixed a locking bug in list_domain_interfaces().
* Added new stack function device_link_changed() that should be called in case the
link state (media) changed.
* The ethernet device module now spawns a thread and will periodically check the media
state of all ethernet devices that support this (if any).
* Minor cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20546 a95241bf-73f2-0310-859d-f6bbb57e9c96

use UserBuffer in list_{device,domain}_interfaces


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20503 a95241bf-73f2-0310-859d-f6bbb57e9c96

SIOCGIFCONF will now also report the size of the written buffer in ifconf.ifc_len
to cover the case the list of interfaces changed since SIOCGIFCOUNT was called.
Patch by Hugo Santos.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20491 a95241bf-73f2-0310-859d-f6bbb57e9c96

No longer crashes when deleting "certain" interfaces (couldn't reproduce the
crash in Qemu for some reason).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20397 a95241bf-73f2-0310-859d-f6bbb57e9c96

When removing an interface, we must make sure it's not still up (ie. we need to
bring it down in that case). This fixes the problem that you couldn't bring an
ethernet interface up again, after you deleted it once.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19524 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Completed the previous commit and merger of the team/network/new_stack branch.
* Removed ppp_up and pppcontrol from the image for now.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18457 a95241bf-73f2-0310-859d-f6bbb57e9c96

Fix naming

Improve patch in ticket #9377
Instead of locking the interface lock, set it busy and then unlock the
interface list lock.

Revert "Missing NULL check in network stack."

This reverts commit dd0ea64b5607e6310c582f62c4275dab7510c4f6.

Missing NULL check in network stack.

This would trigger a KDL, it's better to just fail the syscall call.

Add a TODO comment

Converted the "net_interfaces" mutex into a recursive locker, and
acquire it in Interface::SetDown().

Fixed up network stack and drivers for x86_64.

* Various compilation fixes.
* Fixes to the FreeBSD compatibility layer (from comparing the x86-
specific bits with the equivalent amd64 sources in FreeBSD).
* Compile all the Ethernet drivers except for sis900 and wb840, these
require a bit more work to fix (will file a ticket soon). Tested
ipro1000 and rtl81xx, no issues.

All debugger commands we've added, remove we must.
Closes #7251.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40612 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed ignoring the netmask when computing the broadcast. Renamed a few
variables to make the function a bit clearer, hopefully.
* Fixed the wrong order of releasing the reference, and removing the object
from the hash table -- thanks Rene!
* Added a bit of documentation to InterfaceAddress::Prepare().


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40455 a95241bf-73f2-0310-859d-f6bbb57e9c96

* RemoveAddresses() needs to remove the addresses from the hash table as well.
This fixes random memory access on interface removal.
* Optimized IndexOfAddress() for NULL addresses.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40447 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Allow duplicates in the sAddressTable hash table; while it makes no sense to
have duplicates there, it should certainly not panic.
* Do not add unspecified addresses to the hash table.
* The result of adding the initial address of an interface was ignored; now, the
interface is correctly destructed, if necessary.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@40444 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added a bit more debug output.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@39388 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed a bug in get_interface() that would return the first interface
without acquiring a reference to it, and thus led to bug #6565.
* Added a commented out function that dumps all current reference counts.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38867 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Initialized the domain_datalink's first_{protocol|info} members. This fixes
the crashing bug reported by Atis when trying to add an interface which fails
(the bug causing this has already been fixed with the previous commit).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38180 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added ability to be able to dump a net_interface by name.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38117 a95241bf-73f2-0310-859d-f6bbb57e9c96

* B_SOCKET_SET_ALIAS handling was leaking references.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38079 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Fixed memory leaks pointed out by Stephan, thanks!
* Automatic whitespace cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38068 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Also report the address index back to the userland.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38059 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Also create a new address for B_SOCKET_SET_ALIAS if the index is -1 - this
now has an almost equivalent logic to the old setters like SIOCSIFADDR.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38054 a95241bf-73f2-0310-859d-f6bbb57e9c96

* If the network mask, and broadcast are not specified with a B_SOCKET_SET_ALIAS
they will no longer be unset - instead they are set with defaults.
* If B_SOCKET_SET_ALIAS gets an index of -1, it will now try to find the local
address, and if that fails, will just use the first address there is.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38053 a95241bf-73f2-0310-859d-f6bbb57e9c96

* datalink_control() also need to accept structures smaller than ifreq as long
as the interface name can be specified.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38015 a95241bf-73f2-0310-859d-f6bbb57e9c96

* As Atis pointed out, the SIOCSIFNETMASK here didn't make any sense, must have
been a copy&paste error.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@38001 a95241bf-73f2-0310-859d-f6bbb57e9c96

* We need to acquire references to the interface_address when copying/cloning
the buffer, and we also need to release it when freeing it.
* Adapted patch by Atis: append_cloned_data() now preserves the stored header;
this is kind of bad, because it will not notice if the original header is
removed, but this could probably just be documented this way.
* Also print the reference count of the interface addresses.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37993 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Made the work-around that Atis added to ifconfig and the net_server
in order to be able to use SIOCSIFADDR (and friends) on a net yet added
family superfluous; we now add a new address when needed on the fly.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37990 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed the _SIZEOF_ADDR_IFREQ() macro such that it can be used how we are
using it - hopefully, that is actually correct (it's adopted from FreeBSD).
* Fixed bug that cut off the sockaddr_dl reported by the system, so that DHCP
wouldn't work anymore (and ifconfig wouldn't show the actual MAC address
anymore).
* Changed the listing code to actually pad to ifreq size, and leave the length
of the sockaddr untouched.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37941 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Interface::_ChangeAddress() got the address check for equality wrong,
effectively rejecting any try to set an address.
* It now copies the new address as well in order to make sure that its
sa_len field is set correctly.
* Improved debug output.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37937 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Only actually call change_address() of the net_datalink_protocols if the
address actually changed.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37928 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Make sure the SIOCGIFCONF never returns an address with an address length
smaller than sizeof(sockaddr). This fixes a compatibility issue with other
platforms - portable software often assumes that the amount of bytes to
add to an ifreq structure is the larger amount between sizeof(ifreq), and
basically what the _SIZEOF_ADDR_IFREQ() macro returns, instead of always
relying on that macro.
* Renamed UserBuffer::Copy() to Push, ConsumedAmount() to BytesConsumed(),
added Pad() method.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37927 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Enabled removing the interface on device removal again.
* Made the return type of remove_interface() void, as it cannot fail.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37884 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed the proprietary SIOC_* ioctls to B_SOCKET_* - no reason to pollute
global name space, and have ugly identifiers for nothing :-)
* Added a flags field to struct ifaliasreq. Added flags to mark an alias that
is currently being configured, or has been automatically configured.
Those flags aren't used yet, but they will replace IFF_CONFIGURING and
friends.
* Implemented deleting addresses only from interfaces via ifconfig.
* Added more command aliases for delete to ifconfig ("del", and "delete", for
more consistency with route).
* Fixed control_routes() to only release a reference to an address if it
actually got one before.
* If an interface address is deleted, its routes are now removed as well.
* InterfaceAddress now holds a reference to its interface as planned.
* Implemented removing interfaces. Works quite nicely.
* When downing an interface, all of its routes are now removed. When upping
it again, at least the default routes are added.
* datalink.cpp's get_interface_name_or_index() leaked a reference to the
interface found.
* SIOCAIFADDR would also leak a reference when new addresses were added.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37872 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Work in progress of extending the AF_LINK protocol to be able to send and
receive raw packets (only without the ethernet framing).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37853 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Added Haiku specific socket ioctls to configure the interface aliases:
SIOC_IF_ALIAS_ADD, SIOC_IF_ALIAS_REMOVE, SIOC_IF_ALIAS_GET, SIOC_ALIAS_SET,
and SIOC_IF_ALIAS_COUNT.
* Implemented all of those new ioctls, though they are yet untested.
* Added ifreq::ifr_data, and removed the hack in the FreeBSD compat if.h
header.
* Minor cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37806 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Reworked the complete stack to allow more than one address per network
interface - this caused quite a number of changes.
* Network interfaces, and its addresses are now reference counted (not yet
complete, though, InterfaceAddresses need to hold references to their
interface as well).
* There are two known regressions of this commit that I will fix later:
- you cannot remove interfaces anymore
- IPv4 multicast was broken anyway, but now it's disabled, too.
* Moved a device_interfaces.cpp|h out of interfaces.cpp.
* The datalink layer chain is now instantiated per domain per interface,
not just per interface anymore.
* When a buffer reaches the network layer, it has no known interface yet, ie.
the ipv4|6|whatever modules need to set this manually.
* Added more debug output, and some new debugger commands, the control option
is now printed in clear text.
* Added hash_address() function to the address modules. Added "const" to
set_to_defaults() where needed.
* Fixed net_buffer's restore header functions offset use as reported by Atis.
* Improved buffer dump output, use the domain module to print the address if
available.
* Moved net_buffer::type into the union, as it's not needed by the upper layers
anymore.
* Moved IPv6 specific code from {add|remove}_default_route() to where it
belongs, but disabled it for the time being.
* Completely discarded useless ipv4_datagram module.
* Added ping6 to the build.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@37794 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Made some internal lists use DoublyLinkedLists instead of struct list.
* Added a few KDL commands to improve your debugging experience.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@35469 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Locking order dictates that this lock is an inner lock, and shouldn't
be used everywhere.
* This also fixes a panic when deleting interfaces (an already deleted
lock was unlocked).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29902 a95241bf-73f2-0310-859d-f6bbb57e9c96

* The net_domain's lock is now a recursive lock.
* Fixed all route locking problems, of which there were numerous
({add|remove}_route(), and list_routes() did not lock at all). Added
lock assertions in functions that don't do the locking themselves.
* A route will now be removed from the list in remove_route(), not in
put_route_internal(). Before, a route could easily be removed twice, causing
remove_route() to release references it did not own. This fixes bug #2706.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29386 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Renamed net_device_interface::rx_lock to receive_lock.
* Cleanup, improved comments, removed useless ones.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29232 a95241bf-73f2-0310-859d-f6bbb57e9c96

* More or less completed the network notification module - it does not remove
old invalid user listeners yet, though (ie. if a team dies).
* Implemented userland network monitor functions.
* Added a few notifications to the network stack, even though this part isn't
complete yet (especially notify_interface_changed()).
* Added optional debug output to the notifications module.
* Added the module to the image, it basically works now (tested).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28825 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Actually implemented the SO_BINDTODEVICE socket option I added some time ago.
* This makes it possible to select a specific device, even if no interface
has been configured for it yet. To make it work, each interface now has a
private direct route which will be returned if a socket is bound to a device.
* This will be used for example in DHCP to make it work when more than one
adapter is attached.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27983 a95241bf-73f2-0310-859d-f6bbb57e9c96

* put_device_interface() accessed "device" which was already freed to get
the module name, possibly causing a crash when you delete an interface.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27950 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Replaced all instances of benaphores in the kernel code by mutexes.
* Removed kernel benaphores.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25690 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed recursive_lock to use a mutex instead of a semaphore.
* Adjusted code using recursive locks respectively. The initialization
cannot fail anymore, and it is possible to use recursive locks in the
early boot process (even uninitialized, if in BSS), which simplifies
things a little.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25687 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed the way the socket functions work:
- The net_stack driver is no longer used. Instead we have a kernel
module which is directly used by syscall implementations in the
kernel. I.e. we no longer tunnel those functions through ioctls, but
have normal syscalls.
- Removed the superfluous net starter module.
- Implemented the FDTYPE_SOCKET type file_descriptors, that is sockets
are no longer vnode based.
- Adjusted libnetwork (the socket function implementations)
accordingly.
- Adjusted netstat accordingly.
* Socket module:
- Implemented socketpair().
- Added "kernel" parameter to the control hook. Quite a few more hooks
would actually need the parameter, but I didn't change those yet,
since that would probably also require changes to the protocol
module interface.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@24908 a95241bf-73f2-0310-859d-f6bbb57e9c96

There is absolutely no need that the interface packet consumer runs at a
priority that high.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22894 a95241bf-73f2-0310-859d-f6bbb57e9c96

net_device_interface: remove redundant fields


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21222 a95241bf-73f2-0310-859d-f6bbb57e9c96

loop: no longer requires a reader thread, it delivers directly to the device's receive queue


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21215 a95241bf-73f2-0310-859d-f6bbb57e9c96

net_interface_private: added receive queue and splited device reading from packet processing. Delivering to self no longer is executed in the sender's context, which had some problems with TCP's locking.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21214 a95241bf-73f2-0310-859d-f6bbb57e9c96

retrieve the incoming interface on the domain_receive_adapter(). This should fix potential issues with packets addressed to the IP broadcast address, such as in with some DHCP implementations.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20767 a95241bf-73f2-0310-859d-f6bbb57e9c96

packet capture sockets are now properly notified when the device is going down.

- modified LinkProtocol::ReadData() a bit so it can safely react to external changes such as the device being monitored going down.
- fixed an issue in invalidate_routes() where in some cases the default route was being kept.
- fixed another issue related with the fifo implementation, the notifier semaphore was being created with count of 1, instead of 0.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20622 a95241bf-73f2-0310-859d-f6bbb57e9c96

whenever an interface is deleted, call put_domain_datalink_protocols() so all readers are unregistered.

- Unfortunely this requires RX lock to become a recursive lock.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20621 a95241bf-73f2-0310-859d-f6bbb57e9c96

notify all monitors when a device is being removed.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20615 a95241bf-73f2-0310-859d-f6bbb57e9c96

Prepared net_device_monitor to accept device removal events.

- Introduced public net_device_monitor.
- Changed the link protocol to maintain a lock per instance instead of inside the FIFO. Now all of the link instance data is protected.
- Adapted the link protocol to use net_device_monitor.
- Introduced a private Fifo class which doesn't maintain it's own lock.
- Maybe we should add something like a public net_protocol_implementation which maintains a fifo and a benaphore? With the fifo using the structure's lock instead of maintaining it's own.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20614 a95241bf-73f2-0310-859d-f6bbb57e9c96

moved IFF_LINK handling to the device module (ethernet in this case). Now domain interfaces only keep specific flags such as IFF_UP and the configuration flags. IFF_LINK, IFF_BROADCAST etc are maintained exclusively by the device.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20612 a95241bf-73f2-0310-859d-f6bbb57e9c96

introduced net_device_interface level locking.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20611 a95241bf-73f2-0310-859d-f6bbb57e9c96

a bit more work towards proper locking including a fix to a refcount bug

- fixed a issue in add_interface_to_domain where the device interface's refcount was always incremented since that function was getting the device interface handle and not returning it unconditionlly
- if the ethernet device goes down, and the fd is close()ed, return B_FILE_ERROR instead of calling into the driver again


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20601 a95241bf-73f2-0310-859d-f6bbb57e9c96

started some work to properly handle device_removed() and setting interfaces down in general.

- remove all routes that use interfaces going down.
- when a device is going down, remove associated domain interfaces.
- interfaces weren't getting a properly referenced device interface, fixed.
- down call down_device_interface when deleting a interface, instead set it down and let the upcounts do its job.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20600 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Changed ETHER_GET_LINK_STATE ethernet driver interface, added ETHER_SET_LINK_STATE_SEM.
* The device interface list now uses class DoublyLinkedList instead of struct list.
* Implemented SIOC[SG]IFMEDIA for setting (not supported by any device yet), and
retrieving the device media information.
* Fixed a locking bug in list_domain_interfaces().
* Added new stack function device_link_changed() that should be called in case the
link state (media) changed.
* The ethernet device module now spawns a thread and will periodically check the media
state of all ethernet devices that support this (if any).
* Minor cleanup.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20546 a95241bf-73f2-0310-859d-f6bbb57e9c96

use UserBuffer in list_{device,domain}_interfaces


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20503 a95241bf-73f2-0310-859d-f6bbb57e9c96

SIOCGIFCONF will now also report the size of the written buffer in ifconf.ifc_len
to cover the case the list of interfaces changed since SIOCGIFCOUNT was called.
Patch by Hugo Santos.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20491 a95241bf-73f2-0310-859d-f6bbb57e9c96

No longer crashes when deleting "certain" interfaces (couldn't reproduce the
crash in Qemu for some reason).


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20397 a95241bf-73f2-0310-859d-f6bbb57e9c96

When removing an interface, we must make sure it's not still up (ie. we need to
bring it down in that case). This fixes the problem that you couldn't bring an
ethernet interface up again, after you deleted it once.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19524 a95241bf-73f2-0310-859d-f6bbb57e9c96

* Completed the previous commit and merger of the team/network/new_stack branch.
* Removed ppp_up and pppcontrol from the image for now.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18457 a95241bf-73f2-0310-859d-f6bbb57e9c96