Install nvmem(9) manpage

[ceri]

Summary

There was a manpage and device nvmem(9) created in July but the manpage was never added to the Makefile.
I have guessed at the appropriate architectures, but also adding fernape as a reviewer in respect of the ongoing work to install all manpages regardless.

[manu]

share/man/man9/Makefile

2414

This should be included for anything fdt based.
So riscv and powerpc too.

[ceri]

share/man/man9/Makefile

2414

Thanks; updated.

[manu]

Just realized that we install most OF_* manpages on all arches.
And always installing nvmem.9 would help people developing on amd64 for embedded (most people do that) so I think that this should always be installed.
Sorry for not noticing before.

[ceri]

In D30998#697421, @manu wrote:

Just realized that we install most OF_* manpages on all arches.
And always installing nvmem.9 would help people developing on amd64 for embedded (most people do that) so I think that this should always be installed.
Sorry for not noticing before.

That's ok; the work that @fernape is doing on installing all manpages anyway should take care of that and will be the default, so the logic here would then just allow people who didn't want the page to leave it out.

[fernape]

In D30998#697422, @ceri wrote:

In D30998#697421, @manu wrote:

Just realized that we install most OF_* manpages on all arches.
And always installing nvmem.9 would help people developing on amd64 for embedded (most people do that) so I think that this should always be installed.
Sorry for not noticing before.

That's ok; the work that @fernape is doing on installing all manpages anyway should take care of that and will be the default, so the logic here would then just allow people who didn't want the page to leave it out.

Since MAN_ARCH=all by default starting at 0a0f7486413c147d56808b38055c40c64cff61f5 maybe we could add || empty(${MAN_ARCH}) || ${MAN_ARCH} == "all" to that condition. That is what I plan to do with the remaining pages not covered by the initial commit.

[ceri]

In D30998#697446, @fernape wrote:

In D30998#697422, @ceri wrote:

In D30998#697421, @manu wrote:

Just realized that we install most OF_* manpages on all arches.
And always installing nvmem.9 would help people developing on amd64 for embedded (most people do that) so I think that this should always be installed.
Sorry for not noticing before.

That's ok; the work that @fernape is doing on installing all manpages anyway should take care of that and will be the default, so the logic here would then just allow people who didn't want the page to leave it out.

Since MAN_ARCH=all by default starting at 0a0f7486413c147d56808b38055c40c64cff61f5 maybe we could add || empty(${MAN_ARCH}) || ${MAN_ARCH} == "all" to that condition. That is what I plan to do with the remaining pages not covered by the initial commit.

On reflection, would it be more correct to just install this all the time given manu's comment?
I'm happy either way.

[fernape]

In D30998#697546, @ceri wrote:

In D30998#697446, @fernape wrote:

In D30998#697422, @ceri wrote:

In D30998#697421, @manu wrote:

Just realized that we install most OF_* manpages on all arches.
And always installing nvmem.9 would help people developing on amd64 for embedded (most people do that) so I think that this should always be installed.
Sorry for not noticing before.

That's ok; the work that @fernape is doing on installing all manpages anyway should take care of that and will be the default, so the logic here would then just allow people who didn't want the page to leave it out.

Since MAN_ARCH=all by default starting at 0a0f7486413c147d56808b38055c40c64cff61f5 maybe we could add || empty(${MAN_ARCH}) || ${MAN_ARCH} == "all" to that condition. That is what I plan to do with the remaining pages not covered by the initial commit.

On reflection, would it be more correct to just install this all the time given manu's comment?
I'm happy either way.

I don't have a strong opinion about this.
For instance, in sbin/Makefile, sconfig.8 is installed unconditionally. AFAIK, the only place where MAN_ARCH is truly respected is in share/man/man4 in which Makefile there is a small calculation to decide which arch-specific manpages install.
Probably removing the ${MACHINE_CPUARCH} conditions is the best option: it does what we want and it simplifies things.

[ceri]

OK; I'm going to abandon this review then and just hook up the page as is.
Thanks both.