Pad has no len byte, so the normal parsing code fails.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
RFC1542, 3.2:
"The 'secs' field of a BOOTREQUEST message SHOULD represent the
elapsed time, in seconds, since the client sent its first BOOTREQUEST
message. Note that this implies that the 'secs' field of the first
BOOTREQUEST message SHOULD be set to zero."
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Rename check_packet to check_reply_packet to make its function more
obvious.
The check for DHCP_* values is completely off, as it should
compare against DHCP option 53 (Message Type). Only valid value for
any Bootp/DHCP reply is BOOTREPLY.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
This fixes the error when STATUS_LED_BOOT is not defined.
Signed-off-by: Thomas Chou <thomas@wytron.com.tw>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
If CONFIG_CMD_DHCP is enabled, the vci (vendor-class-identifier) string
isn't inserted into the bootp-packet during SPL stage because the
CONFIG_BOOTP_VCI_STRING
instead
CONFIG_SPL_NET_VCI_STRING
We fix this with testing for CONFIG_SPL_BUILD and testing for existing
CONFIG_SPL_NET_VCI_STRING.
Signed-off-by: Hannes Schmelzer <oe5hpm@oevsv.at>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
TFTP source and destination port variable names are
'tftpsrcp' and 'tftpdstp' in the code, but 'tftpsrcport'
and 'tftpdstport' in the README file. Fix the README.
Add environment variable 'tftptimeoutcountmax'. As per the
comments about the global variable tftp_timeout_count_max,
make sure tftptimeoutcountmax is nonnegative.
Introduce configuration option CONFIG_NET_TFTP_VARS,
which controls whether environment variables tftpblocksize,
tftptimeout, and tftptimoueoutcountmax are read by the TFTP
client code. CONFIG_NET_TFTP_VARS defaults to y but can be
set to n by targets with to tight size contraints.
Make bf527-ezkit set CONFIG_NET_TFTP_VARS to n to keep the
target size below limit.
This patch introduces CONFIG_NETCONSOLE as an option to the
Kconfig system.
Joe Hershberger pointed out that it may not be entirely free of
problems, as many boards predating the driver model define this
symbol directly via include files. In case they're not properly
migrated, their NetConsole might 'vanish' if they start to use
CONFIG_NET or CONFIG_NETDEVICES.
Signed-off-by: Bernhard Nortmann <bernhard.nortmann@web.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
The previous eth_device struct returned by eth_get_dev() allowed
code to directly query the state member field. However, with
CONFIG_DM_ETH this data gets encapsulated (i.e. private), and
eth_get_dev() returns a udevice struct 'abstraction' instead.
This breaks legacy code relying on the former behaviour - e.g.
netconsole.
(see http://lists.denx.de/pipermail/u-boot/2015-June/216528.html)
The patch introduces a method to retrieve the ethernet device
state in a 'clean' and uniform way, supporting both legacy code
and driver model. The new function eth_is_active() accepts a
device struct pointer and tests it for ETH_STATE_ACTIVE.
Signed-off-by: Bernhard Nortmann <bernhard.nortmann@web.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
When given a device name string, we should test to see if it is
really an alias like "eth#".
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
With driver model, board_eth_init() or cpu_eth_init() is not a must.
Thus we don't need print a misleading "Net Initialization Skipped".
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Commit 620776d "tftp: adjust settings to be suitable for 100Mbit ethernet"
causes the following error message when trying to load a file using 'tftp'
command via a tftp server.
TFTP error: 'Unsupported option(s) requested' (8)
This is due to with commit 620776d changes, the tftp option 'timeout'
value is now set to zero which is an invalid value as per RFC2349 [1].
Valid values range between "1" and "255" seconds, inclusive. With some
tftp servers that strictly implement the RFC requirement, it reports
such an error message.
Revert commit 620776d for RFC compliance.
[1] https://www.ietf.org/rfc/rfc2349.txt
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
This change gives the ability to reuse the <tftp.h> header file by other
subsystems (like e.g. dfu).
Without this change compilation error emerges for the legacy update.c file.
Signed-off-by: Lukasz Majewski <l.majewski@majess.pl>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Current behavior is that if CTRL+C is pressed command returns 0 that was
successful which is not correct behavior.
The easiest test case is "tftpboot 80000 uImage && echo yes"
and press CTRL+C. Then the second command is called which is incorrect.
Error log:
zynq-uboot> tftpb 80000 uImage && echo yes
Gem.e000b000:7 is connected to Gem.e000b000. Reconnecting to
Gem.e000b000
Gem.e000b000 Waiting for PHY auto negotiation to complete....... done
Using Gem.e000b000 device
TFTP from server 192.168.0.102; our IP address is 192.168.0.101
Filename 'uImage'.
Load address: 0x80000
Loading: ################
Abort
yes
zynq-uboot>
This patch adds -EINTR return value when CTRL+C is pressed.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Adjust timouts and retry counts to be suitable for loaded ethernet
network. With 5 seconds timeout, 10 retries maximum, tftp is
impossible even on local network with single full-speed TCP
connection.
100msec timeout should be suitable for most networks tftp is used on,
that is local ethernets. Timeout count really needs to be way higher,
as lost packets are normal when TCP is running over the same network.
Enforce 10msec minimum.
Signed-off-by: Pavel Machek <pavel@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
CONFIG_TFTP_TSIZE should limit a tftp downloads progress to 50 '#'
chars. Make this work also for small files.
If the file size is small, i.e. smaller than 2 tftp block sizes the
number of '#' can get much larger. i.e. with a 1 byte file 65000
characters are printed, with a 512 byte file around 500.
When using CONFIG TFTP BLOCKSIZE together with CONFIG_IP_DEFRAG the
issue is more notable.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Some drivers may want to implement this method for some of their devices but
not for others. So it is not possible to just leave the operation out of
the table. Drivers could get around this by masquerading as two separate
drivers but that seems unpleasant.
Allow the driver to return an error when it does not want to process the
write_hwaddr() method.
Signed-off-by: Simon Glass <sjg@chromium.org>
Some devices can take a long time to work out whether they have a new packet
or now. For example the ASIX USB Ethernet dongle can take 5 seconds to do
this, since it waits until it gets a new packet on the wire before allowing
the USB bulk read packet to be submitted.
At present with driver mode the Ethernet receive code reads 32 packets. This
can take a very long time if we must wait for all 32 packets. The old code
(before driver model) worked by reading a single set of packets from the USB
device, then processing all the packets with in. It would be nice to use
the same behaviour with driver model.
Add a flag to the receive method which indicates that the driver should try
to find a packet if available, by consulting the hardware. When the flag is
not set, it should just return any packet data it has already received. If
there is none, it should return -EAGAIN so that the loop will terminate.
Signed-off-by: Simon Glass <sjg@chromium.org>
Instead of selecting REGEX when NET is enabled, make it the default, but
allow boards that are tiny to disable it and lose functionality on all
but the first Ethernet adapter.
cm-bf548, bf538f-ezkit, and bf533-stamp need this. None appear to have
more than one Ethernet interface.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
NET_RANDOM_ETHADDR depends on lib/rand.c. This patch adds dependency to
Kconfig to ensure that library is also compiled.
Remove the definitions from Blackfin boards' include/configs.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
When the ethaddr is changed in the env, update the device pdata at the
same time (only if it is probed for the DM case; only if registered for
the non-DM case). Again this gets us closer to completely non-polled
env needed to simplify the net_loop.
This requires that the NET feature select the REGEX feature.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
The bootfile env var is already kept up to date by the callback in net.c
so there is no need to poll it too.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Instead of checking for changes to the env each time we enter the
net_loop, use the env callbacks to update the values of the variables.
Don't update the variables when the source was programmatic, since the
variables were the source of the new value.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Implement the random ethaddr fallback in eth.c so it is in a common
place and not reimplemented in each board or driver that wants this
behavior.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
When the ethaddr changes in the env, the hardware should also be updated
so that MAC filtering will work properly without resetting U-Boot.
Also remove the manual calls to set the hwaddr that was included in a
few drivers as a result of the framework not doing it.
Reported-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Tested-by: Michal Simek <michal.simek@xilinx.com>
Rename this function so that it is clear that it is provided by the RTC.
Also return an error when it cannot function as expected. This is unlikely
to occur since it works for dates since 1752 and many RTCs do not support
such old dates. Still it is better to be accurate.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Heiko Schocher <hs@denx.de>
At present even with driver model is used there is still much manual init
of related devices: PHY, environment and board init. Until these requirements
are dealt with in another way we need to keep them around.
Break out the init portion of the legacy eth_initialize() into a separate
function and call it from both the legacy and driver model eth_initialize()
functions.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
When SNTP is enabled and DATE is not, to_tm() is not built in. It could
be defined when TIMESTAMP is defined, so check for that.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
This commit fixes incorrect DHCP/BOOTP packet layout caused by
'ulong' type size difference on 64 and 32-bit architectures.
It also renames NetReadLong()/NetCopyLong() to
net_read_u32/net_copy_u32() accordingly.
Signed-off-by: Radha Mohan Chintakuntla <rchintakuntla@cavium.com>
Signed-off-by: Sergey Temerkhanov <s.temerkhanov@gmail.com>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Finish eliminating CamelCase from net.c and other failures
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
There were still a few failures in net/eth.c, especially in the legacy
part of the code.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
A few new rules in checkpatch.pl since linklocal.c was added.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within common/cmd_net.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within netconsole.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within dns.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within cdp.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within ping.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within sntp.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within rarp.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within nfs.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within bootp.c and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within arp and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Make a thorough pass through all variables and function names contained
within tftp and remove CamelCase and improve naming.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Update the naming convention used in the network stack functions and
variables that Ethernet drivers use to interact with it.
This cleans up the temporary hacks that were added to this interface
along with the DM support.
This patch has a few remaining checkpatch.pl failures that would be out
of the scope of this patch to fix (drivers that are in gross violation
of checkpatch.pl).
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
This patch cleans up the names of internal packet buffer names that are
used within the network stack and the functions that use them.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Use "_ethaddr" at the end of variables and drop CamelCase.
Make constant values actually 'const'.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Remove CamelCase variable naming.
Move the definition to the same compilation unit as the primary use.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
The variables around the bootfile were inconsistent and used CamelCase.
Update them to make the code more readable.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
This patch is simply clean-up to make the IPv4 type that is used match
what Linux uses. It also attempts to move all variables that are IP
addresses use good naming instead of CamelCase. No functional change.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Some drivers need a chance to manage their receive buffers after the
packet has been handled by the network stack. Add an operation that
will allow the driver to be called in that case.
Reported-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Tested-on: pcduino3
Take a pass at plumbing errors through to the users of the network stack
Currently only the start() function errors will be returned from
NetLoop(). recv() tends not to have errors, so that is likely not worth
adding. send() certainly can return errors, but this patch does not
attempt to plumb them yet. halt() is not expected to error.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The ethprime env var is used to indicate the starting device if none is
specified in ethact. Also support aliases specified in the ethprime var.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Allow network devices to be referred to as "eth0" instead of
"eth@12345678" when specified in ethact.
Add tests to verify this behavior.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Stop forcing drivers to call net_process_received_packet() - formerly
called NetReceive(). Now the uclass will handle calling the driver for
each packet until the driver errors or has nothing to return. The uclass
will then pass the good packets off to the network stack by calling
net_process_received_packet().
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Take the opportunity to enforce better names on newly written or
retrofitted Ethernet drivers.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Previously the net functions would access memory assuming physmem did
not need to be mapped. In sandbox, that's not the case.
Now we map the physmem specified by the user in loadaddr to the buffer
that represents that space.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
netretry previously would only retry in one specific case (your MAC
address is not set) and no other. This is basically useless. In the DM
implementation for eth it turns this into a completely useless case
since an un-configured MAC address results in not even entering the
NetLoop. The behavior is now changed to retry any failed command
(rotating through the eth adapters if ethrotate != no).
It also defaulted to retry forever. It is now changed to default to not
retry
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This value is not used by the network stack and is available in the
global data, so stop passing it around. For the one legacy function
that still expects it (init op on old Ethernet drivers) pass in the
global pointer version directly to avoid changing that interface.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reported-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Simon Glass <sjg@chromium.org>
(Trival fix to remove an unneeded variable declaration in 4xx_enet.c)
On some archs masking the parameter is inefficient, so don't use u8.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reported-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Many functions returned -1 previously. Change them to return appropriate error
codes.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reported-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Move some things around and organize things so that the driver model
implementation will fit in more easily.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Make it clear that the helper is checking the addr, not setting it.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The current implementation exposes the eth_device struct to code that
needs to access the MAC address. Add a wrapper function for this to
abstract away the pointer for this operation.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Drop the old checksum functions in favour of the new ones.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Move the checksum code out into its own file so it can be used elsewhere.
Also use a new version which supports a length which is not a multiple of
2 and add a new function to add two checksums.
Signed-off-by: Simon Glass <sjg@chromium.org>
Currenly when CONFIG_BOOTP_SERVERIP is defined, the SERVERIP is not changed
when receive the BOOTP packet. But BOOTFILE is changed via BOOTP packet.
As we will load the BOOTFILE from SERVERIP, if the BOOTFILE is modified
by bootp packet but SERVERIP is not, that is not make sense.
This patch make SERVERIP and BOOTFILE be consistent. If we define the
CONFIG_BOOTP_SERVERIP, then SERVERIP and BOOTFILE will not changed by
BOOTP packet. Only IP address is changed.
Signed-off-by: Josh Wu <josh.wu@atmel.com>
A number of network related files were imported from the LiMon
project; these contain a somewhat unclear license statement:
Copyright 1994 - 2000 Neil Russell.
(See License)
I analyzed the source code of LiMon v1.4.2 which was used for this
import. It does not contain any "License" file, but the top level
directory contains a file "COPYING", which turns out to be GPL v2
of June 1991. So it is legitimate to conclude that the LiMon derived
files are also to be released under GPLv2. Mark them as such.
Signed-off-by: Wolfgang Denk <wd@denx.de>
This would be useful to start moving various config options.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
When a DNS query is sent out, the ethernet packet can get directed to
the MAC address of a server that was communicated to before. This is
wrong when the previously stored MAC address corresponds to a different
server's IP address, i.e. when the IP address of the previous and the
current communication are different.
The error can get reproduced by running a sequence of e.g. a TFTP
download and a DNS query, where the TFTP and DNS servers reside on
individual machines.
The fix is to clear the server's MAC address that might be left from a
previous operation, and to fetch the peer's MAC address in a new ARP
lookup, before the DNS query is sent. This is the approach taken in
other network services, like 8e52533d10 ("net: tftpsrv: Get correct
client MAC address").
Reported-by: Dirk Zimoch <dirk.zimoch@psi.ch>
Signed-off-by: Gerhard Sittig <gsi@denx.de>
It's not unusual for DHCP servers to take a couple hundred milliseconds
to respond to DHCP discover messages. One possible reason for the delay
can be that the server checks (typically using an ARP request) that the
IP it's about to hand out isn't in use yet. To make matters worse, some
servers may also queue up requests and process them sequentially, which
can cause excessively long delays if clients retry too fast.
Commit f59be6e850 ("net: BOOTP retry timeout improvements") shortened
the retry timeouts significantly, but the BOOTP/DHCP implementation in
U-Boot doesn't handle that well because it will ignore incoming replies
to earlier requests. In one particular setup this increases the time it
takes to obtain a DHCP lease from 630 ms to 8313 ms.
This commit attempts to fix this in two ways. First it increases the
initial retry timeout from 10 ms to 250 ms to give DHCP servers some
more time to respond. At the same time a cache of outstanding DHCP
request IDs is kept so that the implementation will know to continue
transactions even after a retransmission of the DISCOVER message. The
maximum retry timeout is also increased from 1 second to 2 seconds. An
ID cache of size 4 will keep DHCP requests around for 8 seconds (once
the maximum retry timeout has been reached) before dropping them. This
should give servers plenty of time to respond. If it ever turns out
that this isn't enough, the size of the cache can easily be increased.
With this commit the DHCP lease on the above-mentioned setup still takes
longer (1230 ms) than originally, but that's an acceptable compromise to
improve DHCP lease acquisition time for a broader range of setups.
To make it easier to benchmark DHCP in the future, this commit also adds
the time it took to obtain a lease to the final "DHCP client bound to
address x.x.x.x" message.
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Currently, the BOOTP code sends out its initial request as soon as the
Ethernet driver indicates "link up". If this packet is lost or not
replied to for some reason, the code waits for a 1s timeout before
retrying. For some reason, such early packets are often lost on my
system, so this causes an annoying delay.
To optimize this, modify the BOOTP code to have very short timeouts for
the first packet transmitted, but gradually increase the timeout each
time a timeout occurs. This way, if the first packet is lost, the second
packet is transmitted quite quickly and hence the overall delay is low.
However, if there's still no response, we don't keep spewing out packets
at an insane speed.
It's arguably more correct to try and find out why the first packet is
lost. However, it seems to disappear inside my Ethenet chip; the TX chip
indicates no error during TX (not that it has much in the way of
reporting...), yet wireshark on the RX side doesn't see any packet.
FWIW, I'm using an ASIX USB Ethernet adapter. Perhaps "link up" is
reported too early or based on the wrong condition in HW, and we should
add some fixed extra delay into the driver. However, this would slow down
every link up event even if it ends up not being needed in some cases.
Having BOOTP retry quickly applies the fix/WAR to every possible
Ethernet device, and is quite simple to implement, so seems a better
solution.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
SPL stage does not support various networking things, and therefore
CONFIG_NETCONSOLE cannot be built within SPL.
Signed-off-by: Hannes Petermaier <oe5hpm@oevsv.at>
Some functions in include/net.h are ported from
include/linux/etherdevice.h of Linux Kernel.
For ex.
is_zero_ether_addr()
is_multicast_ether_addr()
is_broadcast_ether_addr()
is_valid_ether_addr();
So, we should use the same function name as that of Linux Kernel,
eth_rand_addr(), for consistency.
Besides, eth_rand_addr() has been implemented as an inline function.
So it should not be surrounded by #ifdef CONFIG_RANDOM_MACADDR.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Changes in lib/uuid.c to:
- uuid_str_to_bin()
- uuid_bin_to_str()
New parameter is added to specify input/output string format in listed functions
This change allows easy recognize which UUID type is or should be stored in given
string array. Binary data of UUID and GUID is always stored in big endian, only
string representations are different as follows.
String byte: 0 36
String char: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
string UUID: be be be be be
string GUID: le le le be be
This patch also updates functions calls and declarations in a whole code.
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: trini@ti.com
Now we are ready to switch over to real Kbuild.
This commit disables temporary scripts:
scripts/{Makefile.build.tmp, Makefile.host.tmp}
and enables real Kbuild scripts:
scripts/{Makefile.build,Makefile.host,Makefile.lib}.
This switch is triggered by the line in scripts/Kbuild.include
-build := -f $(if $(KBUILD_SRC),$(srctree)/)scripts/Makefile.build.tmp obj
+build := -f $(if $(KBUILD_SRC),$(srctree)/)scripts/Makefile.build obj
We need to adjust some build scripts for U-Boot.
But smaller amount of modification is preferable.
Additionally, we need to fix compiler flags which are
locally added or removed.
In Kbuild, it is not allowed to change CFLAGS locally.
Instead, ccflags-y, asflags-y, cppflags-y,
CFLAGS_$(basetarget).o, CFLAGS_REMOVE_$(basetarget).o
are prepared for that purpose.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Tested-by: Gerhard Sittig <gsi@denx.de>
In "common/Makefile" "miiphyutil.o" gets built if any of the following
items enabled:
* CONFIG_PHYLIB
* CONFIG_MII
* CONFIG_CMD_MII
So it's possible to not define CONFIG_MII or CONFIG_CMD_MII and still
use functions like "miiphy_get_dev_by_name".
In its turn "miiphy_get_dev_by_name" traverses "mii_devs" list which is
not initialized because "miiphy_init" never got called.
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Wolfgang Denk <wd@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
NetServerEther was not being cleared in the tftp server code, so the
destination MAC address would be whatever the last destination MAC
address was.
Scenario:
U-Boot:
dhcp
tftpsrv
Host:
Send device WRQ
Device:
Responds with ACK to dhcp server mac address with
host ip address
By clearing NetServerEther, we force a lookup of the host MAC address
to go with the associated host IP.
Signed-off-by: Andrew Ruder <andrew.ruder@elecsyscorp.com>
When the block 0 store to the memory of client and timeout at this
moment. Because of no ACK packet, the server will send block 0 again,
if this client reconnect to the server at this time,
TftpBlockWrapOffset will become larger than it should be.
Signed-off-by: Rockly <rocklygnome@gmail.com>
Patch: 264417
Make the link-local code conform more completely with the RFC.
This will prevent ARP queries for the target (such as while it is
rebooting) from causing the device to choose a different link-local
address, thinking that its address is in use by another machine.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
This patch tackles the time out problem which leads to break the
boot process, when loading file over nfs. The patch does two things.
First of all, we just ignore messages that arrive with a rpc_id smaller
then the client id. We just interpret this messages as answers to
formaly timed out messages.
Second, when a time out occurs we double the time to wait, so that we
do not stress the server resending the last message.
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
Tested-by: Enric Balletbo i Serra <eballetbo@gmail.com>
If we try to boot from NET device, NetInitLoop in net.c will be invoked.
If NET device is not installed, eth_get_dev() function will return
eth_current value, which is NULL.
When NetInitLoop is called, "eth_get_dev->enetaddr" will access
restricted memory area and therefore cause hanging.
This issue is found on Tegra30 Cardhu platform after adding
CONFIG_CMD_NET and CONFIG_CMD_DHCP in config header file.
Signed-off-by: Jim Lin <jilin@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
The ip is stored in network order, so we can't test it in host order.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Kim Phillips <kim.phillips@freescale.com>
This prints a tftp speed indication after the download completes. This
is the 3.6 MiB/s indicator below.
Tegra2 (SeaBoard) # tftp ...
Using asx0 device
TFTP from server 172.22.72.144; our IP address is 172.22.73.81
Filename '/tftpboot/uImage-user-seaboard-1'.
Load address: 0x408000
Loading: #################################################
3.6 MiB/s
done
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Igor Grinberg <grinberg@compulab.co.il>
Move the getenv_yesno() to env_common.c and change most checks for
'y' or 'n' to use this helper.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
bootp.c:44:14: warning: symbol 'dhcp_state' was not declared. Should it be static?
bootp.c:45:15: warning: symbol 'dhcp_leasetime' was not declared. Should it be static?
bootp.c:46:10: warning: symbol 'NetDHCPServerIP' was not declared. Should it be static?
arp.c:30:17: warning: symbol 'NetArpWaitReplyIP' was not declared. Should it be static?
arp.c:37:16: warning: symbol 'NetArpTxPacket' was not declared. Should it be static?
arp.c:38:17: warning: symbol 'NetArpPacketBuf' was not declared. Should it be static?
atheros.c:33:19: warning: symbol 'AR8021_driver' was not declared. Should it be static?
net.c:183:7: warning: symbol 'PktBuf' was not declared. Should it be static?
net.c:159:21: warning: symbol 'net_state' was not declared. Should it be static?
ping.c:73:6: warning: symbol 'ping_start' was not declared. Should it be static?
ping.c:82:13: warning: symbol 'ping_receive' was not declared. Should it be static?
tftp.c:53:7: warning: symbol 'TftpRRQTimeoutMSecs' was not declared. Should it be static?
tftp.c:54:5: warning: symbol 'TftpRRQTimeoutCountMax' was not declared. Should it be static?
eth.c:125:19: warning: symbol 'eth_current' was not declared. Should it be static?
Note: in the ping.c fix, commit a36b12f95a
"net: Move PING out of net.c" mistakenly carried the ifdef CMD_PING
clause from when it was necessary to avoid warnings when it was embedded
in net.c.
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Adjustment of Michael Walle's fix patch
Commit 8a0eccb105 breaks netconsole. src_ip
must not be converted to host byte order, because nc_ip is already stored
in network byte order (see string_to_ip(), called by getenv_IPaddr()).
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
This patch adds support for networking in SPL. Some devices are
capable of loading SPL via network so it makes sense to load the
main U-Boot binary via network too. This patch tries to use
existing network code as much as possible. Unfortunately, it depends
on environment which in turn depends on other code so SPL size
is increased significantly. No effort was done to decouple network
code and environment so far.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Tom Rini <trini@ti.com>
Vendor Class Identifier option is common to BOOTP and DHCP and
can be useful without PXE. So send VCI in both BOOTP and DHCP
requests if CONFIG_BOOTP_VCI_STRING is defined.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
Signed-off-by: Tom Rini <trini@ti.com>
Some DHCP servers (notably dnsmasq) always transmit DHCP Option 28,
Broadcast Address as specified in RFC 2132. Without this patch u-boot
displays the warning:
*** Unhandled DHCP Option in OFFER/ACK: 28
The patch suppresses the warning and ignores DHCP Option 28. There is
no environment variable to set the broadcast address into and if for
some reason u-boot needs the broadcast it can be calculated from
ipaddr and netmask.
Signed-off-by: Brian Rzycki <bmr@freescale.com>
If dev->enetaddr was supposed to be set with dev->write_hwaddr() but the MAC
address was not valid, return an error.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Check the incoming packets' source IP address... if ncip isn't set to a
broadcast address, only listen to the client at ncip.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Previously u-boot would initialize the network interface for every
network operation and then shut it down again. This makes sense for
most operations where the network in not known to be needed soon after
the operation is complete. In the case of netconsole, it will use the
network for every interaction with the shell or every printf. This
means that the network is being reinitialized very often. On many
devices, this intialization is very slow.
This patch checks for consecutive netconsole actions and leaves the
ethernet hardware initialized between them. It will still behave the
same old way for all other network operations and any time another
network operation happens between netconsole operations.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Cc: Stefano Babic <sbabic@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
If the requested length is too small to hold the received packet,
eth_receive() will return -1 and will leave the packet in the receive
buffers. Instead of returning an error in this case, we return the first
portion of the received packet and remove it from the receive buffers.
This fixes FreeBSD's ubldr. Without this patch it will just stop receiving
packets if the NIC receives more than PKTBUFSRX too large packets.
Signed-off-by: Michael Walle <michael@walle.cc>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Cc: Rafal Jaworowski <raj@semihalf.com>
Cc: Piotr Kruszynski <ppk@semihalf.com>
* 'next' of git://git.denx.de/u-boot:
MPC8xx: Fixup warning in arch/powerpc/cpu/mpc8xx/cpu.c
doc: cleanup - move board READMEs into respective board directories
net: sh_eth: add support for SH7757's GETHER
net: sh_eth: modify the definitions of regsiter
net: sh_eth: add SH_ETH_TYPE_ condition
net: sh_eth: clean up for the SH7757's code
net: fec_mxc: Fix MDC for xMII
net: fec_mxc: Fix setting of RCR for xMII
net: nfs: make NFS_TIMEOUT configurable
net: Inline the new eth_setenv_enetaddr_by_index function
net: allow setting env enetaddr from net device setting
net/designware: Consecutive writes to the same register to be avoided
CACHE: net: asix: Fix asix driver to work with data cache on
net: phy: micrel: make ksz9021 phy accessible
net: abort network initialization if the PHY driver fails
phylib: phy_startup() should return an error code on failure
net: tftp: fix type of block arg to store_block
Signed-off-by: Wolfgang Denk <wd@denx.de>
NetConsole may call NetSendUDPPacket before NetLoop is called. This
will cause the source MAC address (NetOurEther) to be wrong. Instead
of only changing it in NetLoop, move it to NetLoopInit so that it is
also updated when net_init() is called (especially by nc_start()).
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reported-by: Michael Walle <michael@walle.cc>
Acked-by: Michael Walle <michael@walle.cc>
Fix NetSetState function name used with CONFIG_BOOTP_MAY_FAIL.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
NFS_TIMEOUT is constant value defined in net/nfs.c. But sometimes it needs to adjust.
This patch enables to override NFS_TIMEOUT by defining CONFIG_NFS_TIMEOUT in a board specific config file.
Signed-off-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
* 'next' of git://git.denx.de/u-boot-net:
net: Inline the new eth_setenv_enetaddr_by_index function
net: allow setting env enetaddr from net device setting
net/designware: Consecutive writes to the same register to be avoided
CACHE: net: asix: Fix asix driver to work with data cache on
net: phy: micrel: make ksz9021 phy accessible
net: abort network initialization if the PHY driver fails
phylib: phy_startup() should return an error code on failure
net: tftp: fix type of block arg to store_block
Signed-off-by: Wolfgang Denk <wd@denx.de>
If the net driver has setup a valid ethernet address and an ethernet
address is not set in the environment already, then set the environment
variables from the net driver setting.
This enables pxe booting on boards which don't set ethaddr env variable.
Signed-off-by: Rob Herring <rob.herring@calxeda.com>
The block argument for store_block can be -1 when the tftp sequence
number rolls over (i.e TftpBlock == 0), so the first argument to
store_block has to be of type 'int' instead of 'unsigned'.
In our environment (gcc 4.4.5 mips toolchain), this causes incorrect
'offset' to be generated for storing the block, and the tftp block
with number 0 will be written elsewhere, resulting in a bad block in
the downloaded file and a memory corruption.
Signed-off-by: Jayachandran Chandrasekharan Nair <jayachandranc@netlogicmicro.com>
The clean up patch missed an &, so we end up passing an int rather than
a pointer to the sprintf function.
arp.c: In function 'ArpReceive':
arp.c:197: warning: format '%p' expects type 'void *', but argument 3 has type 'int'
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
commit "net: use common rand()/srand() functions" introduced the following
build warning on the current u-boot-arm tree:
$ ./MAKEALL MPC8313ERDB_66
Configuring for MPC8313ERDB_66 - Board: MPC8313ERDB, Options: SYS_66MHZ
text data bss dec hex filename
271988 13976 41768 327732 50034 ./u-boot
In file included from bootp.c:15:0:
net_rand.h: In function 'srand_mac':
net_rand.h:40:2: warning: implicit declaration of function 'srand' [-Wimplicit-function-declaration]
adding this dependency fixes it.
Cc: Michael Walle <michael@walle.cc>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Acked-by: Michael Walle <michael@walle.cc>
Add new function eth_random_enetaddr() to generate a locally administered
ethernet address.
Signed-off-by: Michael Walle <michael@walle.cc>
Acked-by: Joe Hershberger <joe.hershberger@gmail.com>
Replace rand() with the functions from lib/. The link-local network code
stores its own seed, derived from the MAC address. Thus making it
independent from calls to srand() in other modules.
Signed-off-by: Michael Walle <michael@walle.cc>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Add several levels of DEBUG prints so that you can limit the noise to
the severety of your problem.
DEBUG_LL_STATE = Link local state machine changes
DEBUG_DEV_PKT = Packets or info directed to the device
DEBUG_NET_PKT = Packets on info on the network at large
DEBUG_INT_STATE = Internal network state changes
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Cisco's arp-proxy feature fails to ignore the link-local address range
This means that a link-local device on a network with this Cisco
equipment will reply to ARP requests for our device (in addition to
our reply).
If we happen to reply first, the requester's ARP table will be
populated with our MAC address, and one packet will be sent to us...
shortly following this, the requester will get an ARP reply from the
Cisco equipment telling the requester to send packets their way
instead of to our device from now on.
This work-around detects this link-local condition and will delay
replying to the ARP request for 5ms so that the first packet is sent
to the Cisco equipment and all following packets are sent to our
device.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Link-local support will need to send ARP packets, but needs more
fine-grained control over the contents. Split the implementation
into 2 parts so link-local can share the code.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Use the NetArpTxPacket for the ARP packet, not to hold what used to
be in NetTxPacket.
This saves a copy and makes the code easier to understand.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
A new non-static function net_init() will initialize buffers and
read from the environment. Only update from the env on each entry
to NetLoop().
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Don't force ARP clients to return the MAC address if they don't care
(such as ping)
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
This is useful if you want to look for a DHCP server, but try some
other settings if not available.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
If CONFIG_BOOTP_SERVERIP is not defined, unused variable warning is
reported. This was fixed upstream using a compiler feature instead
of a simple reorder of the statements.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
When the network is VLAN or SNAP, net_update_ether() will preserve
the original Ethernet packet header and simply replace the src and
dest MACs and the protocol
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Call a built-in dummy if none is registered... don't require
protocols to register a handler (eliminating dummies)
NetConsole now uses the ARP handler when waiting on arp
(instead of needing a #define hack in arp.c)
Clear handlers at the end of net loop
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Use this entry-point consistently across the net/ code
Use a static inline function to preserve code size
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
There is no need to call through the handler... inline it
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Save the length when it is computed instead of forgetting it and
subtracting pointers to figure it out again.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Share more of the code that is common between ARP vs not.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
ICMP (ping) was reimplementing IP header code... it now shares code.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Clearer constant name.
Also remove related BOOTP_SIZE which was unused and doesn't take
into account VLAN packets.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Make the comment more accurate about the header including SNAP
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
This name more explicitly claims that it does not include the
header size
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Acked-by: Simon Glass <sjg@chromium.org>
