linux-kernel - Re: [PATCH v1] mm/memory_hotplug: don't check the nid in find_(smallest|biggest)_section

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <F768F9E2-6413-437A-827F-6105D6DDCD94@lca.pw>
Date:   Thu, 28 Nov 2019 08:56:17 -0500
From:   Qian Cai <cai@....pw>
To:     David Hildenbrand <david@...hat.com>
Cc:     linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Michal Hocko <mhocko@...nel.org>,
        Oscar Salvador <osalvador@...e.de>
Subject: Re: [PATCH v1] mm/memory_hotplug: don't check the nid in find_(smallest|biggest)_section_pfn



> On Nov 28, 2019, at 3:46 AM, David Hildenbrand <david@...hat.com> wrote:
> 
> I'm sorry to say but one of the main reasons we have linux-next for is
> to find BUGs early, before they go upstream. It is a way of giving
> patches *more* testing. Yes, you are doing to dirty work (which is
> highly appreciated btw) by debugging all that crap, and I can understand
> how that can be frustrating.

It is already an expensive development practice if developers need to rely on someone else to figure out their own bugs in linux-next. linux-next is for integration testing, but majority of those regressions I had to deal with nowadays have nothing to do with integration, i.e., interaction with other subsystems.

> 
> But believe me, the world won't end if your on vacation for a couple of
> weeks, even though some BUGs could sneak in ... e.g., lately I try to
> review as much as I can on the MM list (and Michal is steadily watching
> out as well).

Sure, the world will still be running, but good luck on solely rely on reviewing with bare eyes before merging.

> 
> The solution to your problem is more review and testing, really. E.g.,
> I'd be very happy if other developers would test their patches more
> thoroughly and if there would be more review activity on the MM list in
> general (my patches barely get any review ... and I sent a lot of fixes
> lately).

Of course, that helps but it is a culture that very difficult to change now. How many times I saw even high-profile developers proudly sent out patches labeled “no testing” explicitly and implicitly ?

> 
> As soon as we stop touching our code because we are afraid of BUGs, we
> lost the battle against an unmaintainable code base.

Your generalizations of things make me sorrow.

> 
> BTW: [1] mentions "unbalanced software development culture with regard
> to quality vs quantity that supplies an endless stream of bugs". I don't
> agree to this statement. There will *always* be an endless stream of
> BUGs - and most of them come from new features and performance
> improvements IMHO. To me, cleanups and refactorings are important tools
> to improve the software quality (and reduce the code size). All we can
> do is try to minimize the number of BUGs - e.g., via more code review,
> manual testing, automatic testing, and by actually understanding the
> code. Cleanups/refactorings can even fix undiscovered BUGs (e.g., latest
> example is [2])

Surely, most of people probably don’t care about those endless bugs because Linux is a monopoly in data center and open source and it is always like this since Linux was born as a hobby project.