Needed to ensure that writer is available before reader.
Note that data might not be available, when lock is released.
Read data with wrs_shm_seqbegin and wrs_shm_seqend to ensure proper data!
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Even function halexp_query_ports was called results were not used.
Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

It is not used anymore.
Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

wr_phytool needs to read number of ports from HAL, now it uses shm for this read.
Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

rtu_stat needs to read number of ports from HAL, now it uses shm for this read.
Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

wr-mon didn't update ports' information read from hal while
running in interactive mode.

Now also explicitly call read_ports before show_all

Bug was introduced by commit:
"userspace/tools: change wr-mon to use shm for HAL"
Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

run hal_ports_update_all() only 5 times per second (with some
jitter).  However the overall load is not lowered.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Every two months I've been wondering what this 100%-cpu main loop
was, to later acknowledge it was not really 100%.

This commit reorders items and makes it clear where the delay is.
I tried raising it to 100ms from 25ms, but the cpu load of wrsw_hal
didn't change, because most of the time we return to the main
loop before the timeout, due to an RPC call being served.

In my opinion we should add a check in the main loop, to call
port_update_all only 10 times per second, to actually
lower the CPU load of this process.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

bug introduced in c45f9161.

The new check for dotconfig params was added after setting
port->in_use to 1.  Thus, at each loop, the hal was calling

   ioctl(6, SIOCGIFFLAGS, {ifr_name="", ???}) = -1 ENODEV (No such device)

but that ENODEV was after a call to modprobe for the "" interface.
Actually this command:

    /sbin/modprobe -q -- netdev-

By means of /sbin/modprobe being the string in /proc/sys/kernel/modprobe .
But modprobe is busybox, so over NFS we had a zillion stat() accesses
for ld.so, uclibc.so and busybox.

Thanks Adam for noting the NFS traffic increase at that commit.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Since commit "userspace/snmpd: make snmpd use shm for HAL requests"
snmpd uses shm to get information from HAL instead of wr_mon.
Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

snmpd still uses minipc to communicate with ppsi
Symlink to shmem.c is created since I couldn't find a good way to link libwr
into shared library wrsSnmp.so. I managed only to link entire library not
only used functions.
Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

endpointcalibration.php: migrated to dotconfig and data structures.
endpointmode.php: migrated to dotconfig and data structures.
(modifymode.php is part of this change)
sfpcalibration.php: migrated to dotconfig and data structures.

wrs-data.php: removed NETWORK_SETUP duplicated info

dns.php: configures DNS server and domain using the web interface and
dotconfig
logs.php: configures HAL, PTP and RTU logs.
network.php: tables and forms are now generated automatically and
processed (it required a modification in style.css. Same for DHCP
configuration. (static conf. still undone)
management.php: UTC removed. Old config files form removed. Now we only
backup and restore dot-config file.
functions.php:
- print_info() enhanced.
- process_form() created. It process automatically the data from forms.
- apply_kconfig() created. Calls apply_dot-config after saving.
- management process: removed previous actions (files, halt, etc)

wrs-info-generator.php: command for kernel compile dateh was wrong
command, fixed.
wrs-info.php: adding the generated file by default. It should be better
to have when flashing.

- functions.php: Calling wrs-info-generator.php to generate wrs-info.php
at the beginning of each session.
- functions.php: print_form() and print_info() functions with data from
wrs-data.php
- functions.php: added KCONFIG load and save functions.
- functions.php:  creation of $_SESSION data structres.
- ptp.php: added PPSi at boot (Enable/Disable) and form for PPSi
Parameter.
- contact.php: info is now displayed using a table from wrs-data.
- network.php: print_info(data) for eth0 configuration.
- index.php: minor format fix.

Because of using data structures to load and display configuration, at
the end of each session all web interface user data must be destroyed.

wrs-data.php declares a different data structures for the different
sections of the web interface.
There are tables by default, but user will be allowed to customize
DASHBOARD display, Contact Us, Network Setup... etc

- data/ folder has been created. It will contain all data structures
declaration for the web interface.
- wrs-info-generator.php => generates wrs-info.php automatically. Users
can select the fields.
- wrs-info-generator.php is run once per user session. It declares
$WRS_INFO, which will contain wrs-info (uname info + wrs_version info)
during the session.

This adds the possibility to use PERL parsing from PHP extension PCRE.
It is usefull for future file parsing changes from PHP.

Our web pages use the "jquery" javascript library, from
code.jquery.com.  With this commit we refer to a local copy of the
file, instead of always referring to the upstream instance -- which
failed when the WRS was not connected to the outside world, and
generated extra traffic anyways.

Now head.php refers to the local copy; the WRS filesystem includes both
the source form (http://code.jquery.com/jquery-1.10.1.js)
and the minimized form, unreadable but shorter, which is actually
served to the browser (http://code.jquery.com/jquery-1.10.1.min.js).

Note: commit author is benoit but then alessandro added the unminimized
javascript and wrote the commit message: any error or imprecision is a's,
not b's.

- white rabbit figure added
- style.css: adding style for the image and the text
- reboot.php: adding new <div> for image and text

Removing images not used:
footer.jpg
footnote.gif
header.jpg

footer.php minor fix (footnote.gif link removed)

New reboot system. It shows now a loading gif image without crashing
because of not accessing the WRS anymore.

- Reboot javascript function: displays image, message & reboots without
web crash
- Redirect javascript function: When rebooting, it redirects to index.hp
after 50s.
- All shell_exec("reboot") calls have been replaced by wrs_reboot() in
all files.
- loader.gif added
- new entry to style.css, one for the gif animation and other for the
text.
- reboot.php & rebooter.php carry out the new reboot.

Javascript folder created (/js) with scripts.js file.
This file contains all javascript functions.

head.php now includes this file instead of include the JS functions
directly.

File newlines were a mixture of Windows/Unix.
All php files are now in Unix-like format
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

If no specific alpha is passed in configuration, we can't even know if
it is positive or negative. 0 is a safe default, and it is the exact
correct value for dual-fiber situations.

The next commit, which adds documentation, explains the thing better.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This commit adds the timing mode to Kconfig, but it is only
used in later commits.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This replicates the bevaviour of previous code, unchanged. Only,
Kconfig is now used to specify the alpha for the pair of wavelengths
being used.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This commit adds fiber parameters to Kconfig. To be used to specify
fiber-specific alpha values.  Use in the next commits.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>