[lkml]   [2017]   [Jul]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 3/4] ndctl: switch to tools/include/linux/{kernel, list, bitmap}.h

* Dan Williams <> wrote:

> On Wed, Jul 26, 2017 at 4:29 AM, Ingo Molnar <> wrote:
> >
> > * Dan Williams <> wrote:
> >
> >> On Tue, Jul 25, 2017 at 4:55 PM, Arnaldo Carvalho de Melo
> >> <> wrote:
> >> > Em Tue, Jul 25, 2017 at 03:36:26PM -0700, Dan Williams escreveu:
> >> >> Replace the ccan implementation of list primitives, bitmap helpers and
> >> >> small utility macros with the common definitions available in
> >> >> tool/include/linux.
> >> >
> >> > You should first add what you need in separate patches, paving the way
> >> > to then use it, and some stuff are already there, see below:
> >>
> >> Ok, I'll break out those changes separately.
> >
> > BTW., another general observation I have is that ndctl uses autotools - while perf
> > uses its own build system, some of which is abstracted out into tools/build/ and
> > reused by other tooling projects as well.
> >
> > I despise autotools with a passion, it's slow, bloated, and encourages all sorts
> > of bad API/ABI practices that plagues many OSS projects. I know that Linus
> > explicitly did a Makefile based build system for Git for (I think) similar
> > reasons.
> >
> > It might be a good idea to not let autotools into the kernel tooling tree, not
> > because ndctl's use of autotools is bad in any fashion (it appears to be a fairly
> > straightforward use), but to generally encourage good API/ABI practices in our
> > tooling space, and to encourage enhancements to the tools/build/ infrastructure.
> That's a fair point. Regardless, autotools will be in the git history,
> but if you'd like to see the final merge product eliminate its use, I
> can't really argue otherwise. I was originally not concerned because
> tools/usb/usbip/ was an existing in tree autotools user. In any event
> if you want the autotools removal to be done out-of-tree I'll need to
> put this effort on the back burner until 4.15.

So that was another thing I wanted to suggest: why not import the current ndctl
version as a single commit?

I had a quick look, and there's quite a few of commits in the ndctl history that
don't conform to kernel standards, such as:

ce881c1e78f6: ndctl: seed tracking

which doesn't have any Signed-off-by tags.

There's also commits with ambiguous titles that would be confusing in the kernel
context - for example:

796b6f373dec: clarify copyright and license information

... which on the surface could be misunderstood as something talking about the
kernel copyright ...


e38bd36e5d0a: completion: updates for file name completion

which I could initially mistake for a commit about scheduler completions ;-)


2ad6a39c9ae9: Fix attribute sizes to match NFIT 0.8s2
cc7cb44385d3: Import initial infrastructure


I suppose all that could be corrected, SOBs added, titles clarified and prefixed
with tools/ndctl, but then it wouldn't really be unmodified history anymore,

At that point we might as well do a clean start - and not import ~500 extra
commits into the kernel tree?



 \ /
  Last update: 2017-07-26 19:20    [W:0.090 / U:3.512 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site