struct vm_fault changed its definition before kernel 4.10, at commit
1a29d85eb0 to be more precise. This makes code compatible with pre- and
post- 4.10 kernels.

Signed-off by: Juan David Gonzalez Cobas <dcobas@cern.ch>

The reason to extended it is that it makes easier the identification
of a single zio object from user-space.
It is, in a way, similar to the zio_addr, but this is ascii and with
less information (indeed it can be used only locally on the same machine)

Signed-off-by: federico.vaga@cern.ch <federico.vaga@cern.ch>

In output configuration, 0 will mean "do not change seq_num". Any other
value will be assigned to current_ctrl->seq_num

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
Acked-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: federico.vaga@cern.ch <federico.vaga@gmail.com>

Signed-off-by: federico.vaga@cern.ch <federico.vaga@gmail.com>

The introduction on symbolic link was not 100% correct on patch

af1c35a5

  Enumerate channels and channel-sets in sysfs

If a custom name is not assigned, DO NOT force one (like we did). Leave it
empty so that we can easily recognize this condition.

Change the sysfs 'name' attribute in order to use the proper variable

Signed-off-by: federico.vaga@cern.ch <federico.vaga@gmail.com>

with -V switch

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Ignore temporary files created during build.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Changes in the buildsystem.

Include parent_common.mk from up directory. Which may contain various
environment definitions such compiler, linker, etc. If such file is not
present it will not affect build.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

remove all *.ko.cmd and *.o.cmd files

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Remove defining variables LINUX and GIT_VERSION from Makefiles in
subdirectories.

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Signed-off-by: Adam Wujek <adam.wujek@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

ZIO_OBJ_NAME_LEN is *just* the limit for names, but sometimes we want to append
thing to a given name. Typically, index to identify the component in the
hierarchy. So this patch introduce ZIO_OBJ_NAME_FULL_LEN which should be used as
size for any string that modify the original name

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Create a sysfs symlink for backward compatibility using cset/chan name

Signed-off-by: Paolo Savini <il.paolo.savini@gmail.com>
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

The check is already done by the change functions (and the buffer one
was wrong but I didn't noticed it because we had a double check)

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

In a big industrial environment we may have several different ZIO devices
each of them compiled with a different version of ZIO. With this versioning
field we can check che compatibility among devices and ZIO framework.

This avoid to recompile everytime all device drivers for each change.

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Miguel Jimenez Lopez <klyone@ugr.es>
Acked-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Alessandro Rubini <rubini@gnudd.com>

Add to zio_device, zio_cset and zio_channel a new operation: change_flags.
The core uses this function to signal to the driver that some flags are changed.

This patch also improve the comment on the __zobj_enable function

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Alessandro Rubini <rubini@gnudd.com>

Currently the trigger abort function returns only the previous value
of the status flag, but actually we had always used it as if it was
the entire flag, and this can break "self timed" applications on
trigger configuration.

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Use a generic implementation of ti->push_block() in order to avoid
code duplication and different behaviors on output configuration.
(it will be more clear with multishot implementation)

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

The new option is "prefer-new" and it is boolean.

0 means that you prefer old blocks. So, when buffer is full all new
incoming blocks will be discarder.

1 means that you prefer new blocks. So, when buffer is full the oldest
block is removed in order to store the new one

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
Acked-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

If the driver call data_done when the trigger was un-armed, then print a
debug message. This is useful on driver development.

Theoretically, this code should not be compiled if we are not setting -DDEBUG

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>