Merge branch 'for-next/net'

net: Allow hostnames for global.net.server ...

Additional to IPv4 addresses add support for global.net.server being a
hostname.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: dns: Allow to set variable with the resolved host ...

Add an additional [VARIABLE] parameter to the host command to allow
setting a variable with the resolved IP address.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: dns: leave host command with error on failure ...

When we can't resolv a host we should return an error rather than just
successfully.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: lib: Make use of ETH_ALEN ...

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: dns: return error codes ...

The resolv() function used to return the IP address. When net_udp_new()
fails we return an error code though which the callers of resolv() take
as an IP address. This is wrong of course and we could return 0 in this
case. Instead we return an error code and pass the resolved IP as a
pointer which allows us to return proper error codes.

This patch also adds error messages and error returns to the various
callers of resolv() which used to just continue with a zero IP and
let the user figure out what went wrong.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ping command: Print which host is pinged ...

When a hostname instead a IP is given to the ping command then the
IP the host is resolved to is a useful information. Add a message
printing it.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: dns: Use pr_debug ...

Use pr_debug rather than debug and add a pr_fmt string to give
the messages more context. While at it add a debug message which
prints the ip when successfully resolved.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/net-switch-mv88e6xxx'

Merge branch 'for-next/net'

Merge branch 'for-next/misc'

net: only use ethaddrs registered with device node on that device ...

If an ethaddr is registered with a device node the ethid is set
to 0 via the struct initialization. Now if a device without a node
(for eaxample a USB network adapter) with ethid 0 looks up the
ethaddr it will wrongly get that address assigned, as the ethid
matches.

Avoid this by setting the ethid to something that no device will
ever use when the addr is registered via a device node.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: dhcp: fix option 54 ...

Option 54 (Server Identifier) is something different than
SIADDR (Server IP address). Currently they are used
interchangeably. This causes DHCPNAK with DHCP servers that
have configured different values for each field.

Create new 'dhcp_serverip' field in 'struct dhcp_result', so we treat
option 54 separately. This fixes DHCP requests for servers that send
two different values for SIADDR and option 54 in DHCPOFFER.

Fixes: 528298b702a0 ("net: dhcp: rework")
Fixes: e18cc5b7e9ce ("DHCP: fix option 54 passing")
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

drivers: Introduce dev_set_name() ...

Introduce dev_set_name() in order to hide implementation details of
setting device's name so it'd be easier to change it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: make routing work after dhcp command ...

At the moment only ifup stuff turns netif up.
After the commit f0624a701513 ('net: Do not route traffic
to interfaces that are not up') the dhcp command
keeps netif->ifup == false and network subsystem
can't route packets.

How to repropduce the problem on qemu-malta_defconfig:

    qemu-system-mips -nodefaults -M malta -m 256 \
                     -nographic -serial stdio -monitor null \
                     -bios barebox-flash-image \
                     -net user -net nic,model=rtl8139

    ...

    barebox:/ dhcp
    eth0: 100Mbps full duplex link detected
    eth0: DHCP client bound to address 10.0.2.15
    barebox:/ ping 10.0.2.2
    ping failed: No route to host

However if ifup command is used for network interface
configuration then there is no network problem, e.g.

    barebox:/ ifup eth0
    eth0: 100Mbps full duplex link detected
    eth0: DHCP client bound to address 10.0.2.15
    barebox:/ ping 10.0.2.2
    host 10.0.2.2 is alive

Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
CC: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/misc'

net: dhcp: add global variable for retries ...

Signed-off-by: Oleg Karfich <oleg.karfich@wago.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: dhcp: use private extension 224 also in discover requests ...

Signed-off-by: Oleg Karfich <oleg.karfich@wago.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: Do not route traffic to interfaces that are not up ...

In the case when:

  - Board has multiple network interfaces

  - Two ore more of those interfaces are statically configured to be
    on the same network

  - Only one of those interfaces is up and it is preceeded (as far as
    for_each_netdev is concerned) by interface in the same network
    that isn't

net_route() will choose "non-up" device as a route for traffic
resulting in no network connectivity. Change the routing logic to also
consider if interface is "up", so that only such interfaces would be
considered for sending network traffic out.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: eth: Make use of ETH_ALEN ...

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: use dev_warn() when no MAC address is set ...

When a SoC such as VF610 has no addresses programmed for both its
interfaces, the "No MAC address set" warning can be confusing:

    booting 'net'
    WARNING: net: warning: No MAC address set. Using random address a2:e8:be:79:72:01
    WARNING: net: warning: No MAC address set. Using random address aa:d5:d7:10:c0:42

This patch uses dev_warn() instead and removes the redundant "warning":

    booting 'net'
    WARNING: eth0: No MAC address set. Using random address a2:e8:be:79:72:01
    WARNING: eth1: No MAC address set. Using random address 4e:be:c4:bc:ce:36

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Sam Ravnborg <sam@ravnborg.org>

net: on ifup -a only detect devices when necessary ...

For network boot we once used to hardcode eth0, but in latest changes
this was changed to work with different network devices and the 'ifup
eth0' was replaced with 'ifup -a' which lead to the result that we now
detect all devices in order to eventually also bring up USB network
adapters. In most of the cases this is not desired. When a board has
internal network support this is likely to be used. With this patch
we only detect all devices when we do not have a network device already.
For the unusual case in which a USB network adapter shall be used
even when an internal network interface is present we introduce the
global variable "global.net.ifup_force_detect" which can be used to
force detection of devices.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

DHCP: fix option 54 passing ...

Though variable net_dhcp_server_ip was declared and used as a
function parameter it was never assigned. That's why a DHCP
request was sent without option 54 set. Some DHCP server for
example dnsmasq didn't accept such a packet.

As both offered IP address and server IP are stored in the global
dhcp_result structure we don't need either net_dhcp_server_ip or
OfferedIP variables and can use the addresses from dhcp_result
directly.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: make struct bootp __packed to prevent unaligned store on MIPS ...

How to repropduce the unaligned store problem

    qemu-system-mips -nodefaults -M malta -m 256 \
                     -nographic -serial stdio -monitor null \
                     -bios barebox-flash-image \
                     -net user -net nic,model=rtl8139

    ...

    barebox:/ dhcp

    Ooops, address error on store!

    $ 0   : 00000000 00000000 01010600 697f2a2e
    $ 4   : a0850000 00000000 0000001c a040c1b8
    $ 8   : 00000000 00000002 00000002 00000000
    $12   : 00000000 00000040 00000100 00000001
    $16   : a040bba0 a0850000 a0850000 a0850000
    $20   : 00000000 00000075 00000076 a040ba20
    $24   : 00000002 a080f210
    $28   : 00000000 a03ffce0 fffffffd a0833b8c
    Hi    : 000154f8
    Lo    : 20000000
    epc   : a0833b84
    ra    : a0833b8c
    Status: 00000002
    Cause : 80000414
    Config: 80008482

    ### ERROR ### Please RESET the board ###

The unaligned store instruction is located in the bootp_request() from net/dhcp.c:

a0833b50 <bootp_request>:
..
a0833b7c:       3c020101        lui     v0,0x101
a0833b80:       24420600        addiu   v0,v0,1536 /* 0x1010600 -> v0 */
a0833b84:       0c20024a        jal     a0800928 <get_time_ns>
a0833b88:       ae02002a        sw      v0,42(s0) /* store 0x1010600 to addr s0 + 42 */

This assembler code is generated by mips-linux-gnu-gcc (Debian 7.2.0-11) 7.2.0
from this:

    bp = (struct bootp *)payload;
    bp->bp_op = OP_BOOTREQUEST; /* 0x01 */
    bp->bp_htype = HWT_ETHER; /* 0x01 */
    bp->bp_hlen = HWL_ETHER; /* 0x06 */
    bp->bp_hops = 0;

Compiler replaces four 'store byte' instruction by one 'store 32-bit word'
instruction. Alas sometimes this leads to unaligned store situation.

Making struct bootp __packed prevents this optimization and fixes the problem.

Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Tested-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: Add linuxdevname property ...

When you have a static network environment but more than one network
device on your machine it is necessary to provide the <device> parameter
to the ip parameter at kernel cmd line.

The device name assigned by Linux cannot in general be predicted as it
depends on driver bind order.

This patch introduces a new property linux.devname to eth devices.
The value is added to bootargs per interface and can be changed in
env/network/INTF

Based on patch by Gavin Schenk <g.schenk@eckelmann.de>.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: Provide new way to configure network devices ...

This provides a new way to configure network interfaces based
on nvvars. A network interface can now be configured with variables
in the nv.dev.<ethname>.* namespace. There is a new network device
parameter "mode" which specifies the mode used to obtain IP settings.
The mode can be "dhcp", "static" or "disabled":

nv.dev.eth0.mode=dhcp
(ipaddr, netmask are ignored in this setting)

nv.dev.eth0.mode=static
nv.dev.eth0.ipaddr=192.168.0.17
nv.dev.eth0.netmask=255.255.0.0

nv.dev.eth0.mode=disabled

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ifup: Use dhcp C API rather than running command ...

DHCP has a C API, so use it instead of running as command.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: ifup: Factor out a eth_discover function ...

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: remove "current" network device ...

Now that we can do routing we no longer need a "current"
network device. Remove it.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

net: Pick network device based on IP settings ...

The IP/netmask/gateway settings contain all informations
needed to pick the correct network device. This patch
adds support for that and makes specifying the "current"
network interface using the ethact command unnecessary.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>