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

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

Resolve "Kernel OOPS after few minutes of loading fine-delay svec module"

Closes #10

See merge request be-cem-edl/fec/hardware-modules/fmc-delay-1ns-8cha!2

In the commit:
	commit 02ec43ce74e015eed509305e4faad78c9a8f828a ("kernel:
	spec_core: Update API setup_timer() to timer_setup()")

We added the newer protoype for 'fd_update_calibration()' along-with the
older one with conditional ditrectives.

Reflect the same changes in 'fine-delay.h' to avoid compilation error.
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

This reverts commit 25f66f51.

Turns out that even though Red-Hat backported "timer_setup()" on CS7, it is
not working well and corrupting the memory somewhere. Thus, we should
stick to the API provided by vanilla kernel.

Minor cleanup

Resolve "review vaibhav changes"

Closes #8

See merge request be-cem-edl/fec/hardware-modules/fmc-delay-1ns-8cha!1

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

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

It avoids ImageMakic errors when containers have stricter policy
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: 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@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@cern.ch>

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

As per tovalds/linux, the kernel version below 4.14 support
	fd_update_calibration(unsigned long arg)
and not
	fd_update_calibration(struct timer_list *arg)
But even though CS7 is 3.10 based, patches might
have been backported, and it supports the latest one.
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

This warning is treated as error by the compiler. Mark it with
"__attribute__((unused))".
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

In struct memory_ops, the function-pointer (*read) is declared as
	u32 (*)(void *)

In its use, its object is given the address of function 'ioread32' (or
'ioread32be') which used to be of the same type.

Since linux-v5.8, their protype has been changed to
	u32 (*)(const void *)
at:
commit 8f28ca6bd8211214faf717677bbffe375c2a6072 ("iomap: constify ioreadX()
iomem argument (as in generic implementation)")

Hence, update the declaration of (*read) as well.
Signen-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

The API was modified in linux-v4.14 at

commit e99e88a9d2b067465adaa9c111ada99a041bef9a ("treewide:
setup_timer() -> timer_setup()")

In software/kernel:
- in 'fd-irq.c', the function fd_tlet() was used for both timer and
  interrupt-tasklet.
- this function  was of type: void (func)(unsigned long).
- both setup_timer() and tasklet_init() use this function-type as one
  of their arguments.
- setup_timer() was updated to timer_setup() which now accepts a
  function of type void (func)(struct timer_list*).

Thus,:
- define fd_tlet_interrupt() of type void (func)(unsigned long).
- define fd_tlet_timer() of type void (func)(struct timer_list*).
- Update fd_tlet() to type void (func)(struct fd_dev*)

After completing their specific jobs, fd_tlet_interrupt() and fd_tlet_timer()
call fd_tlet() to execute the codes common to both.
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

kernel: spec-core: definition of 'struct platform_device_id' no longer extractable from 'linux/platform_device.h'

The 'struct platform_device_id' is defined in 'linux/mod_devicetable.h';
this header was #include in 'linux/platform_device.h'. Hence, using the
latter header-file was enough. But in linux-v4.18 at

commit ac3167257b9fe16c9426c2087ead1c9f1b0992b1 ("headers: separate
linux/mod_devicetable.h from linux/platform_device.h")

the former was no more included in the latter. Thus, we need to do it
explicitly.

Since this change makes no difference if this driver is compiled against
Kernel version less than v4.18, we do not need any conditional
directives.
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

In linux-v5.6 release:

'struct timespec' got unsupported for in-kernel use at
	commit c766d1472c70d25ad475cf56042af1652e792b23 ("y2038: hide
	timeval/timespec/itimerval/itimerspec types")

And accordingly, the 32 bit interfaces were also removed at
	commit 412c53a680a97cb1ae2c0ab60230e193bee86387 ("y2038: remove
	unused time32 interfaces")

Thus, use 'struct timespec64' and related interfaces for 5.10.65 kernel.
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

Variable name 'LINUX' is very ambiguous for its purpose in the makefile.
Also, this project builds as a part of COHT project which uses another
variable name 'KERNELSRC' for the same purpose. Hence, this change makes
this project uniform with others.
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

The recipe for 'make modules' can be used for 'clean' too.
Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

Signed-off-by: Vaibhav Gupta <vaibhav.gupta@cern.ch>

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