Date:	Thu, 7 Apr 2016 08:08:11 -0700
From:	Greg KH <gregkh@...uxfoundation.org>
To:	"Artem S. Tashkinov" <t.artem@...os.com>
Cc:	linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: The most insane proposal in regard to the Linux kernel
 development

On Thu, Apr 07, 2016 at 02:01:03PM +0500, Artem S. Tashkinov wrote:
> On 2016-04-07 01:05, Greg KH wrote:
> > On Sat, Apr 02, 2016 at 05:43:47PM +0500, Artem S. Tashkinov wrote:
> > > One very big justification of this proposal is that core Linux
> > > development
> > > (I'm talking about various subsystems like mm/ ipc/ and interfaces
> > > under
> > > block/ fs/ security/ sound/ etc. ) has slowed down significantly
> > > over the
> > > past years so radical changes which warrant new kernel API/ABI are
> > > less
> > > likely to be introduced.
> > 
> > That's not true at all, the change is constant, and increasing, just
> > look at the tree for proof of that.
> > 
> > > Please, share your opinion.
> > 
> > Please read Documentation/stable_api_nonsense.txt for my opinion, and
> > that of the current developers.
> > 
> > If you don't agree with this, that's fine, you are welcome to fork the
> > kernel at any specific point and keep that api stable, just like many
> > companies do and make money from it (SuSE, Red Hat, etc.)
> > 
> > best of luck with your kernel project,
> > 
> 
> Tell me, why no one in the Linux kernel dev team is concerned that:
> 
> 1) There is up to a hundred regressions in each kernel release where a big
> chunk of them are caused by internal API changes?

Really?  Citation please.

> 2) API changes sometimes require drastic changes in every related hardware
> driver and since there's no way you can realistically test the code or the
> hardware, people later discover that their hardware has stopped working?

Specific examples would be nice to show here, sometimes things fall
through the cracks, but usually that's just an oversight and they get
fixed once they get reported.

> 3) The core kernel developers do not have enough expertise to correctly
> update the entire kernel source tree so little things get broken?

Citation please.

> 4) Developing drivers for a moving target is a Herculean job?

Somehow more people do it than happens for any other operating system so
perhaps you just need to take the time to learn how to do it properly?
Yes, there are ways of doing this wrong, and correctly.  How have you
been doing this?

Also note, Linux drivers are, on average, 1/3 smaller than other
operating system's drivers _because_ of our development model.  You wish
to waste more driver developer time, and bloat your system, just because
you can't keep up with the rest of us?

> 5) You cannot easily bisect kernel regressions because regressions are often
> caused by things _outside_ of the problem you're experiencing.

You do know about 'git bisect skip', right?

> 6) You cannot use new drivers for you hardware on your old kernel, because
> new drivers are incompatible with an old source tree (don't remind me of
> RHEL's kernel - it's a rare exception and they usually port only the drivers
> their respective clients use).

Never use old kernels, that's your first mistake.  If you are stuck with
an old kernel, then go get support from the company that is forcing you
to do such a horrible thing, as you are paying them for this privilege.

> 7) Tech unsavvy people cannot realistically debug the kernel.

And they can debug a kernel from other operating systems?  That makes no
sense at all.

> Hey, please, do not tell me that you're doing a great job following postings
> in LKML or resolving bugs files in bugzilla. You do a very lousy job indeed
> - multiple postings in LKML get zero replies because a corresponding
> developer is either not subscribed to LKML at all, or he has missed the
> message.

You should never just blindly email lkml, that's why we have 100+ other
subsystem mailing lists, and the MAINTAINERS file, for you to figure out
who to contact.

> There are literally hundreds(!) of bugs in bugzilla which have ZERO
> replies.

Most development groups in the kernel do not use bugzilla, rather they
use the aforementioned mailing list as that scales much better.

> What's more, a great number of kernel developers do not have
> accounts in bugzilla and they don't read corresponding mailing lists.

True about bugzilla (it doesn't scale, known thing), but not true about
mailing lists.

> What the hell is wrong with you guys? You're developing the kernel like it's
> your toy project.

Really?  What other operating system is being developed better?  No one
is forcing you to use Linux, if you don't like our development model, go
use one that you like, it's that simple.

> 1) There's no accountability whatsoever.

Define "accountability" please?  Lots of companies will be glad to take
your money for that privilege.  Perhaps you should work with one of
them.

> 2) There are no unit tests. Not a single one.

You haven't looked very hard have you?  That's a totally incorrect
statement. (hint, 'make kselftest')

> 3) There's no surefire way to contact developers who have commited "bad"
> code.

Yes there is, their emails are right there in the patch, much easier to
contact than any other operating system's developers.

> 4) There's no sense of direction.

What direction do you want?

> 5) There's no easy way to debug the kernel.

Define "easy"?  Are other operating system's kernels somehow easier to
debug?  What did you try and fail doing with Linux?

> For instance, let's talk about the revoke() call. Right now, if a certain IO
> device is removed while files on it are still open (there are multiple ways
> of opening files in Linux, starting from fopen() and ending with mmap()),
> the kernel state is basically undefined(!).

If you wish to implement revoke(), please do so,  However it's
non-trivial, which is why no other operating system has it either for
all devices (BSD has it only for a limited set of devices, the trivial
ones.)

> Great! The corresponding mount
> point cannot be reused(!). Whatever program, which has its files'
> descriptors on this accidentally removed device, usually cannot gracefully
> quit or continue working. How on Earth this syscall doesn't get the utmost
> attention?

Because in real life it's not an issue.  The resources are properly
freed, userspace is notified correctly, and everyone continues happily
along with the new device descriptor.  You have written your userspace
code to properly handle this issue, right?  :)

> Then we have bug 12309(1). My last comment to this bug gives a very simple
> way of reproducing it on all Android devices.

Great, work with those developers through email please.

> Then we have bug 15875(2) which will probably take just ten man hours to be
> resolved, yet there is no interest at all, yet thousands of people have very
> real problems due to it.

If you wish to have a new feature implemented, you have to implement it
yourself, or pay someone to implement it for you.  You can't tell other
people what to do for you, which seems to be your major objection here.
That's just a matter of how the world works, no matter what group you
work with, or operating system you use.  Sorry you have to learn it this
way.

> Tell me, are you really proud of yourselves?
> Tell me, do you develop the kernel for your amusement, ego, your employee or
> for average people to use?
> Tell me, are you really interested in more people migrating from stable long
> term supported OSes to Linux?

If someone is happy with other operating systems, great, we don't care,
that works well for them, we aren't forcing anyone to use Linux.

> I want some truly honest answers. And let's not repeat this mantra "we don't
> have enough resources". You have enough resources to break API/ABIs in a
> huge way, you have enough resources to introduce regressions - you only
> don't have enough resources to have any resemblance of a responsible
> development process.

We don't have a way for people to tell other people what to do, yes, you
are correct.  Neither does any other operating system for that manner.
But if you can find such a system, great, I highly recommend you go use
that one.

Best of luck,

greg k-h

Powered by blists - more mailing lists