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

open returns -1 on error, not 0

mmap returns -1 on error, not NULL
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>

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

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

Stuff was seriously misaligned and hard to read. So Maciej agreed
about running ../../scripts/Lindent. This commit makes no modification
at all, even it changes both spaces/tabs and newlines.

Some comments still exceed 80 columns, which is bad for me, my screen and
my printer, but I'll only change if needed, at a later step.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

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

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

AT45DB642D is now obsolete and AT45DB641E will replace it
They have the same status ID (same density). Therefore we need to read
the JEDEC code instead.

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

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

Since there is no env config file, envfs_check_super() call fails to find
correct magic number and makes envfs_load function to return before setting
default MAC addresses.

  * MAXPORTS is NPORTS (so to use the usual C convention
  * parse as list of ranges like  1-5,9,12-14
  * introduce for_each_port()
  * complain of no port mask is set before config
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Conflicts:

	userspace/wrsw_rtud/rtu_fd.c
	userspace/wrsw_rtud/rtu_tru_drv.c

PSTATS_NPORTS in the kernel driver should be set by default to 18 ports, because
it's also used to create a table for cntrs values read from FPGA. If driver is
supposed to run on 8-port version of WRS, the parameter pstats_nports should be
set at load time.

This commit is cherry-picked from v4-dev-rebased only for compatibility with
v4rc FPGA firmware. TRU event has to be removed both from gateware and software
for the final release.

Conflicts:

	userspace/tools/wrsw_pstats.c

wrsw_rtud: fixed serious misconfiguration: default config was to set fixed priority (seems bad, prevented priorities from working), current default: take priority from incoming frames

initializing variables to be written to H/W - caused bug in writing frame-class to priority (switch internal)

Conflicts:

	userspace/wrsw_rtud/rtu_ep_drv.c
	userspace/wrsw_rtud/rtu_hwdu_drv.c

Conflicts:

	userspace/tools/wrsw_pstats.c

Conflicts:

	userspace/tools/Makefile

Conflicts:

	kernel/wbgen-regs/Makefile
	kernel/wbgen-regs/tru-regs.h
	kernel/wbgen-regs/tru-regs.wb

Conflicts:

	kernel/wbgen-regs/tru-regs.h
	kernel/wbgen-regs/tru-regs.wb
	userspace/wrsw_rtud/rtu_tru_drv.c
	userspace/wrsw_rtud/rtu_tru_drv.h

Again, suboptimal formatting is now an error in new makeinfo. Worse,
this looks correct, but the complain is there anyways.  Hack a fix,
killing a newline that make the parser report the error.

Hope we are done: nobody outside gnu is fool enough to use makeinfo.
Me excluded, obviously (see wr-switch-sw/doc).
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Somebody decided that suboptimal formatting is now an error. So the
new makeinfo errors out when building gcc-4.3.6 because of
documentation details.  Idiocy has no limits, and we are are bound to
suffer for that.  I remember when "gnu's not unix, we have higher standards".

This patch prevents gcc documentation to be built when buildroot prepares
the cross compiler.  Bah.

/alessandro who wasted a whole day on that