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

On higher storm of traffic, in the time between scheduling the tasklet and
executing it, I was loosing cntrs overflow information already (was overwritten
by all-the-time-incrementing counters in HDL).

File provides information about pstats version, number of counters per word
stored in memory and number of counters per each port.

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

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

We don't need a ppsi_defconfig any more, as it is now the default.
Also, use a full defconfig, so it's explicit what the options are.
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 fills the information by parsing the output of "wr_mon -g"
and "wr_mon -p".

Parsing is data-struct driven, and the strings must match the one
printed by wr_mon. If they don't a log message is generated:

   18:02:54 192 snmpd[1468]: wrs_ppsi_parse_line: can't find key "seource:"

To test, you can export WRS_SNMP_MON_FNAME and WRS_SNMP_MON_FNAME_P to
name two text files that can be opened instead of a pipe from wr_mon.

A problem with COUNTER64 remains, but it is a bug in netsnmp. See next
commit.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This is needed if for example it is called in a pipe that is then closed.

meanwhile, remove trailing white space.
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>

Version items are now fake strings, as I still miss the code
that reads if from the device.  But they work, and the mib is
updated:

   tornado% snmpwalk -c public -v 2c -m +`pwd`/WR-SWITCH-MIB.txt wrs \
            .1.3.6.1.4.1.96.100.4
   WR-SWITCH-MIB::wrsVersionSw.0 = STRING:  fake-v4.0-rc1
   WR-SWITCH-MIB::wrsVersionGw1.0 = STRING: fake-7cce708
   WR-SWITCH-MIB::wrsVersionGw2.0 = STRING: fake-5118070
   WR-SWITCH-MIB::wrsVersionGw3.0 = STRING: fake-7efeb16
   WR-SWITCH-MIB::wrsVersionHw1.0 = STRING: fake-3.30
   WR-SWITCH-MIB::wrsVersionHw2.0 = STRING: fake-LX240T
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

thanks to greg for a 1-byte fix I squashed here
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

This is a squash-up of the whole netsnmp-pain branch, that I kept
on the repository, so to document the steps.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

I want to keep this scalar thing in place as a minimal working thing.
However, to rename an object I need to redo everything from scratch.
No comment on the quality of this all.
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 series of commits changes completely the installation and runtime,
moving to UBI.  I used it several times, and there are still minor problems
related to flash access, but it is generally stable.  I plan to find
the flash bug and fix it in a separate commit. This is ready for master,
in my opinion.

Note: a number of intermediate commits won't work, but I chose to split
the changes according to the package (barebox, kernel, userspace, ...)

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>

The information is not up to date with the new installation.  We are
not using JTAG ourselves these times, so we'd better remove this
chapter, stating it is still in the history of the package, if it is
needed by someone.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

The release tar file (wrs-firmware.tar) is not compressed any more, so
fix the name all around, and the extracting command too.

Similarly, the official release is a .tar file, and it is generated by
the "wraprootfs" step, so this removes "--pack" as a build option (it
was a duplicate of wrapping it up anyways).

Release files are already the ones used for installation, so flash-wrs
doesn't need to build the cpio file any more.

The "macigstring" for installation is concatenated to the cpio file by
tellling the usb-loader to copy it to RAM, so no temporary cpio archive
is created any more: release files are always used unchanged.

Wrapping rootfs creates the release tar dire
"--pack" i
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

This brings everything into /etc/init.d/wrs-boot-procedure that is
executed as a "sysinit" task, before anything else.  The procedure
runs from the initrd/initramfs and is concerned with the complete
update of the device, if an update is pending.

See documentation (next commit) for details about partitions and boot
procedure.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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