A Linux kernel developer found himself in a perfect storm of Linus Torvalds' sharp tongue and his intolerance for bad code. Red Hat's Mauro Carvalho Chehab was told by Linux kernel chief Torvalds to "shut the f**k up" and fix his "approach to kernel programming" after Chehab passed off a bug in the kernel as something at fault …
There are better ways of pointing out someone's mistakes, but this is much more entertaining.
Unfortunately you sometimes have to make people really angry and upset to get them to do something. They'll keep defending their position and dismissing what you say to them.
Just look at Rev Ian Paisley, for years he was defiant and dismissing everything the Irish republicans said, then shook hands with them and was laughing and joking about the whole matter.
Some people need a "parental" figure to tell them to go clean their room or tidy up their code in this instance.
Open source is unfortunately lacking a person at the top to shout at people when they produce crap. Which is probably why the Linux kernel is the best part of the whole lot since it has someone calling the shots.
"Unfortunately you sometimes have to make people really angry and upset to get them to do something."
No. This is bad management, if someone won't tow the line they can't do the job, you shouldn't have to bawl them out. This kind of management behavior just serves to alienate other employees or potential employees and is dangerously close to bullying.
Sure, if Mauro was in any way Linus' employee. Which he isn't. He can't take him aside and give him a talking to, or sack him. The only nuclear option he has is to beat on him in public, so that his employers take notice.
OTOH, there is no need for someone like Linus to take that sort of tone on-list. He could have just said something along the lines of "Mauro, please re-check this, as I am convinced you are wrong in this", which would be just as painful for a senior dev to receive.
Well you say it is bad management but I'd be fine with it.
I honestly can't stand working with delicate little flowers who need their hands held and their egos stroked the whole time.
If I balls up I don't mind a bit of a shout - get things out in the open, get them sorted and get on.
If Torvalds had done this to me I'd be embarrassed that he had to but I wouldn't resent him for having done so. (Of course if I knew I was better than him . . .)
The idea that there's one management style to rule them all is tosh.
But Mauro broke a fundamental rule of Linux kernel programming, and then tried to shift the blame elsewhere.
"But Mauro broke a fundamental rule of Linux kernel programming, and then tried to shift the blame elsewhere."
What rule is that? Backward compatibility? Oh wait...
> bad management
> dangerously close to bullying
This is the age where bad crap cannot be called out, you gotta be "careful", everyone shall have prizes, a call from the union is just a cough away and every office jerk was a cool tech genius from the instant he tried to fit a square peg into a round hole at the age of three.
As long as the criticism is not "your are $X" but "your are producing $X that needs to be fixed" you gotta say what you gotta say. Right the ship's course etc.
Which reminds me that in my tender youth, I went on a sailing expedition where you got told what's what. That's a learning experience.
It's nothing to do with "delicate little flowers", which is a wholly patronising comment, it's about basic decency and decorum in the workplace. As a manager I know that I will not get my staff on side or gain their respect by ranting and raving about what they've done or not done. Picking on someone in front of their co-workers, or worse on a public forum, is just bullying and would get me sacked were I to do it. What's more, it would have got me sacked from every job I've had, were I to have done it. I don't think that there is only one management style, but I'm do know that shouting and bullying is unacceptable and just shows weakness in the person doing it.
I really don't know what's going on with comments here these days, it's getting more and more Daily Mail by the day.
"This kind of management behavior just serves to alienate other employees or potential employees "
Sorry dude you just broke the stupidometer by driving it off the scale. Since when does Linux employ the people working on the kernel?
Welcome to the world of community based development. When you are employing someone you can fire them. When they are "volunteering" their incompetence (even as a part of a company sponsorship) in a community driven project you sometimes _HAVE_ to make them leave. Even if this involves deploying the F* word. C'est la vie. In fact, in many cases it would have been easier to employ the poor guy - in that case you can fire him.
@AC 18:48 - Re: Err...
Are you by any chance working in the public sector or in the government ?
Basic decency: praise in public, correct in private. If you cannot stick to this simple, very basic rule you have no business leading any group of people.
Sometimes it's management who isn't listening and you have to get angry for them to listen.
It's sometimes referred to as being 'passionate' but often dismissed as being 'emotional'. In some places, people getting angry is the ONLY thing management responds to....
But, hey, perhaps you've only worked in blessed places where this never happens.
>Sure, if Mauro was in any way Linus' employee. Which he isn't. He can't take him aside and give him a talking to, or sack him.
Erm... of course he could. Who do you think manages the Kernel project and thus controls the coding/contributors/contributions/maintainers?
It sounds to me like this problem has been going on for some time and Torvalds might be approaching doing exactly that!..
>How long have you been a maintainer? And you *still* haven't learnt the first rule of kernel maintenance?
>And you've shown yourself to not be competent in this issue, so I'll apply it directly and immediately myself.
>The fact that you then try to make *excuses* for breaking user space, and blaming some external program that *used* to work, is just shameful. It's not how we work.
Sounds to me like Mauro might be auditioning for a job at Microsoft. ;-)
It is bad management if they have to tow the line anywhere, after all you tow barges. What I think you meant to say was TOE the line, as in being up to the mark.
I wonder if the following equation will hold up to academic study
Torvalds = BOFH
"If a change results in user programs breaking, it's a bug in the kernel. We never EVER blame the user programs. How hard can this be to understand?"
Like when you do a 'dist-upgrade' (Debian and derivatives) and a new kernel version arrives, your programs carry on working. (Not talking about device drivers here).
You may have problems with programs depending on libraries or other (user space) code that are no longer available in your distro. That is a user space problem, and not caused by the kernel.
if you are not at work or of a nervous disposition.
I personally think a 'phone call' (remember those) might be good under these circumstances. The CentOS developer Karanbir Singh actually lists a phone number for CentOS issues now they have sponsorship.
@Ac: this is not about picking on someone this is about someone being a persistent prick in denying anything was their fault when it blatantly was. When you get someone this steadfastly in denial there comes a time when "in private" just won't do as they continually ignore you. This sound like such a case. In such situations other methods need to be used to shock them back into the real world. If you want understand this then your touchy freely only management style is of little use.
While in most situations this is bad management, the Linux Kernel project is far from typical. Whether it is good or bad depends on the outcome.
In Linus' view, Mauro's sloppy work and attitude is damaging the reputation of the team with its customers. Linus' message here is only superficially to Mauro; it is actually to reassure pulseaudio and other consumers of the kernel that this sort of thing (making sloppy changes that break stuff, then passing the buck) is not going to be tolerated. It is likewise a message to the rest of the contributors. Consequently, he does not do it privately, and cannot spare Mauro's feelings. Linus' priority is to stop the bleeding.
The customer's collaboration is what Linus feels he really needs, much more important than Mauro's contribution.
"Basic decency: praise in public, correct in private. If you cannot stick to this simple, very basic rule you have no business leading any group of people."
Bought one of those "Management for Dummies" books at the airport, have you? :-)
At least he didn't break the chairs out.
Correct in private?
You've clearly never worked on an open source project with a PUBLIC mailing list, which necessitates review feedback in a PUBLIC forum.
It is incredibly unhelpful for you to have conversations like this in PRIVATE as then anyone hoping to learn from the thread will be lost. This is why most open source projects ask for as much on list feedback as is practicable. If criticism in PUBLIC is not something you are ok with then perhaps open source is not for you.
It's nothing to do with "delicate little flowers", which is a wholly patronising comment,
It's not patronising at all. If someone has had their repeated error pointed out to them and then persists in telling the world that everyone else is wrong except them.... I'd tell them to "shut the F*CK up and grow up!".
If they can't handle it, they have no place in any kind of pressured environment, which means they're no good for my business, and will spend their lives complaining about every little thing if they get a job elsewhere.
These are the type of people who do not receive references when they leave my employment. I've had three of them try to sue because I refused to lie on recommendation letters, and just had to demonstrate their incompetence and over-sensitivity at the employment tribunals. They ended up having to pay me to leave the company...
Stupid is as stupid does.....
Re: Err...if someone won't tow the line
if someone won't toe the line
Maybe it was no security vulnerabilities in the kernel? Oh wait....
Err... like most devs you are extrapolating from your own attitude to the rest of the world. Actually, lots of people react very badly to this style of "management". They might be demotivated, or quit, or attempt sabotage, or nurse a grudge and wait for a chance of revenge...
Good for you that you're OK with this sort of treatment, but hey, lots - maybe most - are not. So unless Linus knows this person spectacularly well, it's terrible management.
Re: Correct in private?
But people NEED to have tolerance breaks, and just because someone can lick stamps, does not mean they ought to be running the post office.
There are some really amazing people doing really amazing things, and this includes really great work and also really fucking stupid shit as well.
Some people DO double check ALL of their work and then are content that it's flawless, and others are so fucking lazy, and incompetent, that they are content to leave defective and half done work as is, and expect millions of people to wear their defective bullshit.
"Oh dear, Linus called me a dumb fucking idiot"
Well chances are that there may very well be a good reason for it.
Fuck I generally make the effort to explain to many people, many things, but there are plenty who wouldn't register what a cactus was if I shoved it up their arse.
The words "can you put me through to your manager" occassionally come out, and the words like , "Go fuck yourself" and "click" do season my conversations.
People who do a shitty job of it and don't give a fuck, tend to COST immensely more than their net benefit.
It's about standards, and how hard you have to occassionally whip YOURSELF to maintain them.
If you can't dock someone's pay, withold promotions or fire someone, shouting at them is sometimes the only option. I'm not saying it should be the first response, but I've dealt with people who repeatedly do something simple wrong and only change their ways once I go through the escalation of:
polite request > polite instruction > terse instruction > impolite demand > "For the love of God, suck less!"
Seems like quite a good example of the Peter Principle.
It has been shown time and again that Torvalds is a good programmer and a very bad manager, so why has he got into a situation where he spends all his time managing (badly) and no longer does any actual coding? Complete waste of his talents, and of the managerial talents of someone else who should be doing the job instead.
After all this, what's the general standard of Linux architecture and kernel code?
Not trying to flame-bait, but I'm not really qualified to assess the Linux kernel quality, so, for those of you who are: what's your verdict?
Rev Ian Paisley
I down voted that, not because I wanted to make a comment about Irish Society, but because (hidden by the Irish politics), it didn't make any sense at all.
Let's take two possible meanings:
1) Ian Paisley made a lot of people angry and upset, and it solved the Irish problem. Sometimes you need to do that.
2) Somebody needed to make Ian Paisley angry and upset. If somebody had done that, it would have solved the Irish problem.
Was it supposed to be ironic? If anything, I think the history demonstrates that making people angry and upset doesn't get them to stop defending their position.
"If I balls up I don't mind a bit of a shout"
Ooh, aren't you manly? I bet you don't mind a bit of a shout with your balls up...
I bet you and your shouting partner are the types that have loud one-upmanship conversations while you stroke yourselves next to each other at the urinals wafting your testosterone around, just to publicly display how masculine you are.
Meanwhile the rest of us think you're wankers.
You can't assign anything to Torvalds, so for correctness, this has to be:
BOFH = Torvalds;
remember, C requires statements to be terminated by semicolons.
Sorry, but some people in life really do need a kick in the arse. If you've never met the type, you are a very lucky person.
I don't think anyone is saying decorum and civility should not be the first resort. Or even the second. But then there's the thick skulled boneheads...
I see you know the type well.
Met far too many myself.
These are usually the same people who will do those things regardless of how they are treated.
Angry outbursts as a managment tool
And yet it is used by every MILITARY in the world ...
Re: After all this, what's the general standard of Linux architecture and kernel code?
The Linux code quality is quite bad, if you listen to developers. (I talk about Linux developers complaining on the low quality of the code).
I bet you and your shouting partner are the types that have loud one-upmanship conversations while you stroke yourselves next to each other at the urinals wafting your testosterone around, just to publicly display how masculine you are
I'm manly enough to grow a beard (but then so was my grandmother) and manly enough to cope with having made someone cross without turning to jelly in the face of their ire but I'll have to disappoint you on the willy-waving contests.
I hope me saying that you have this hippyish, decidedly non-alpha male completely wrong will not prevent you from continuing to enjoy your public toilet masturbation fantasies.
"Picking on someone in front of their co-workers, or worse on a public forum, is just bullying.."
Oh really? Then I was bullied to hell and back by all of my school teachers. I agree with all the comments that sometimes you have to call people out in public to get a point across. IT IS NOT BULLYING. Attacks on personality is bullying, attacks on shoddy work is not. If someone f&*ks up, a manager should have to right to point it out, even in front of their peers.
Re: After all this, what's the general standard of Linux architecture and kernel code?
You should provide some links if you claim controversial statements. Here are links to developers complaining Linux code quality:
Linux kernel maintainer Andrew Morton says the code is bad:
"I used to think [code quality] was in decline, and I think that I might think that it still is. I see so many regressions which we never fix....it would help if people's patches were less buggy."
OpenBSD developer Theo de Radt says the code is bad
"It's terrible," De Raadt says. "Everyone is using it, and they don't realize how bad it is. And the Linux people will just stick with it and add to it rather than stepping back and saying, 'This is garbage and we should fix it.'"
Linus Torvalds, says the code is bloated
"Citing an internal INTEL corp study that tracked kernel releases, Bottomley said Linux performance had dropped about two per centage points at every release, for a cumulative drop of about 12 per cent over the last ten releases. "Is this a problem?" he asked. "We're getting bloated and huge. Yes, it's a problem," said Torvalds."
Ted Tso, creator of ext4 says the Linux developers cheat to get higher performance, on the penalty of bug free code:
"In the case of reiserfs, Chris Mason submitted a patch 4 years ago to turn on barriers by default, but Hans Reiser vetoed it. Apparently, to Hans, winning the benchmark demolition derby was more important than his user's data. (It's a sad fact that sometimes the desire to win benchmark competition will cause developers to cheat, sometimes at the expense of their users.)...We tried to get the default changed in ext3, but it was overruled by Andrew Morton, on the grounds that it would represent a big performance loss, and he didn't think the corruption happened all that often (!!!!!) --- despite the fact that Chris Mason had developed a python program that would reliably corrupt an ext3 file system if you ran it and then pulled the power plug "
"The [linux source code] tree breaks every day, and it's becomming an extremely non-fun environment to work in....We need to slow down the merging, we need to review things more, we need people to test their f--king changes!"
There are many more links if you google a bit. Unix creators have said the Linux code is bad. I can supply many more links, from Linux kernel hackers, claiming on the bad code "the kernel is going to pieces"
Also, Linux has huge scaling problems. It scales fine on a cluster, but scales very bad on a single fat server. Clusters are for HPC work (number crunching, easily parallell problems, etc), they can have 1000s of cpus and many TB of ram. For instance, the SGI Altix server has 2048 cores and 64TB RAM. But it is a cluster. As the Linux ScaleMP server with one single image Linux kernel with 8192 cores, just as the SGI Altix server:
"The vSMP hypervisor that glues systems together is not for every workload, but on workloads where there is a lot of message passing between server nodes – financial modeling, supercomputing, data analytics, and similar parallel workloads. Shai Fultheim, the company's founder and chief executive officer, says ScaleMP has over 300 customers now. "We focused on HPC as the low-hanging fruit,"
A programmer writes about this server:
"I tried running a nicely parallel shared memory workload (75% efficiency on 24 cores in a 4 socket opteron box) on a 64 core ScaleMP box with 8 2-socket boards linked by infiniband. Result: horrible. It might look like a shared memory, but access to off-board bits has huge latency."
All these cc-NUMA servers belong to the HPC server category: clusters of many comput nodes, on a fast network switch.
On the other hand, SMP servers, are a single fat server. They can have 16 cpus, 32 cpus, some has as many as 64 cpus. IBM Mainframes belong to the SMP category. IBM P795 is also a SMP server, it has as many as 32 cpus. Oracle has SMP servers with 64 cpus, the M9000. And HP has 32 cpu SMP servers. They typically can have up to 64 cpus and 2-4TB RAM. The IBM P595 used for the TPC-C record, costed 35 million USD. List price. One single SMP fat server. The HPC servers are very cheap in comparison, because it is a cluster of cheap compute nodes on a fast network.
The biggest Linux SMP server has 8 cpus (from Oracle, IBM and HP, it is just a standard 8-way x86 socket server). There are no big Linux SMP servers on the market. No one sells them. For a reason: Linux has problems handling 8 cpus in SMP fashion. There are no 16 cpu Linux servers, or 32 cpu Linux servers. Not a single one.
Thus, Linux scales good horizontally (in a HPC cluster). But scales extremely bad in a fat big server (SMP server). Linux can not handle 16 cpus, or someone would create and sell such servers for a fraction of the price of Oracle/IBM/HP - because big SMP servers costs a lot.
So, can any show me a Linux server with as many as 16 or 32 cpus? No, there are none for sale. And has never been. Only old mature Enterprise Unix like IBM AIX, Solaris, HP-UX can scale to 32 or 64 cpus.
The Irish problem is solved?
RE: Torvalds = BOFH
Needs more cattleprod...
"This is the age where bad crap cannot be called out, you gotta be "careful", everyone shall have prizes, a call from the union is just a cough away and every office jerk was a cool tech genius from the instant he tried to fit a square peg into a round hole at the age of three."
Truth be told it's one of the early warnings of you might have a psychopath on your hands, which is well worth looking out for.
Plus people should be able to expect to be not treated like shit, capable or not.
How to win friends and influence people
More mental than mentor
....would turn Ghandi into a chair-throwing Ballmerite.
(in fairness I haven't used it for a few years, perhaps it's better now)
- Crawling from the Wreckage Want a more fuel efficient car? Then redesign it – here's how
- Review Xperia Z3: Crikey, Sony – ANOTHER flagship phondleslab?
- Human spaceships dodge ALIEN BODY skimming Mars
- Downrange Are you a gun owner? Let us in OR ELSE, say Blighty's top cops
- Ex-US Navy fighter pilot MIT prof: Drones beat humans - I should know