Linus chews out a kernel maintainer

[mcveigh]

http://news.softpedia.com/news/Linu...er-on-the-Official-Mainling-List-317880.shtml

thought this was funny.

 

[AnonymouseUser]

You can't argue with someone who bitches you out and then fixes your mistake himself.

 

[Nothinman]

It's also not like the sound system in Linux is awesome and that this is the first issue. We've been dealing with huge changes every few years from OSS to ALSA in the kernel and to userland crap like ESD and Pulse to work around issues that that arguably shouldn't even exist because the kernel should just deal with them, like soft mixing for concurrent access.

 

[lxskllr]

Linus Torvalds is not usually this angry, especially not in mailing list. Most likely, he wanted to turn the kernel maintainer into an example and we can safely say that his mission was accomplished.

I've gotta disagree with this. He has a quick temper, and I've seen it turned on kernel hackers more than once. Probably all deserved, but I wouldn't want to be the recipient of a Torvalds teaching moment :^D

 

[Red Squirrel]

Haha yeah Linus tends to get pissed off, but he usually has a good point. The best is still when he trashed Unity lol.

 

[Kiska]

HAHAHAHAHAHAHAHAHAHAHA. Just what I needed to make my day. Can't stop laughing.

 

[Praetor]

Mauro did a remarkable job of staying on target and did not stoop to replying with expletives to Linus on his replies. I'm impressed.

 

[VinDSL]

Bwahahaha! That jackball needed to have his ass handed to him! 3.8-rc1 sucked!

3.8-rc1 worked for a couple of days, then I stated getting Perl5 errors, followed by total libpam breakage. Most of the AMD guys couldn't even get it to boot.

I was scratching my head for a week, trying to figure out what happened -- until I read Linus' tirade.

3.8-rc2 ironed out these problems. 3.8-rc3 is working a treat!

I'm glad Linus dressed-down this rogue "maintainer" in public...

 

[Elixer]

That was really uncalled for.
Most of the people who work on these projects do so without getting paid.

Now, he could have had a normal, level headed comment about how this was wrong and how this fix was better, but noooo... let's just burn this guy on a public ML.

Then people wonder why there are so many forks out there of duplicated effort just because idiots can't behave in public.

 

[Kiska]

If it was uncalled for then I would have started my own triad with him. Since I couldn't even get into BIOS with 3.8-rc1. Good thing that Linus done it before me and also many other people from what I heard.

 

[LumbergTech]

Lost a bit of respect for Linus there. I don't always do the right thing and lose my temper sometimes as well, but that doesn't mean that its OK. That was pretty much abusive. Sure, you can take the "thick skinned" approach where anyone can say any rude thing to each other, but to me it just seemed like a little much. I don't know if there is any back story where that maintainer did other things wrong, but just based on that it was out of line even if he was technically correct.

 

[AnonymouseUser]

That was really uncalled for.
Most of the people who work on these projects do so without getting paid.

Well this guy does get paid to do this, by Red Hat. It's not unusual for big companies to pay their employees to maintain free software, especially the Linux kernel. When you consider that this paid maintainer was breaking software that is mostly maintained by volunteers, then yes, he deserved it.

 

[_Rick_]

Since I couldn't even get into BIOS with 3.8-rc1.

Wat?

You do know, that this was a bug a in an A/V subsystem, that exclusively lead to issues, when a specific API was called?
And that the kernel isn't even invoked, until after the BIOS hands control of the system over to the boot loader?

If you don't, then maybe you should refrain from commenting, until you do.

On topic: Everybody knows, that Torvalds is a bit of a nut job.
Yes, this was an...unfortunate bug, and a not so well worded response to the bug report.
But the involvement of Mr. Kernel was stupidly verbose. A simple, short mail, that denying that a bug is present is not the way to deal with reports should have sufficed. The insistence on the kernel breaking user land always being a bug is idiotic. If user land depends on bugs or undocumented behavior, then the stability of these interfaces can not be guaranteed.

Absolute power will always be abused. I think the problem is that Linux has not transcended into a true community project, because in the end Torvalds or one of the appointed maintainers has to sign it off.

Open source is nice, an open source project is hell.

 

[Nothinman]

That was really uncalled for.
Most of the people who work on these projects do so without getting paid.

Not any longer, the large majority of contributions to the kernel have been made by corporate employees for a while now. Mauro works for RedHat, for example.

 

[Crusty]

If it was uncalled for then I would have started my own triad with him. Since I couldn't even get into BIOS with 3.8-rc1. Good thing that Linus done it before me and also many other people from what I heard.

You've got something else wrong, and it's not your Linux kernel.

 

[Elixer]

Not any longer, the large majority of contributions to the kernel have been made by corporate employees for a while now. Mauro works for RedHat, for example.

There is no peer review being done then ?

 

[Nothinman]

There is no peer review being done then ?

How would you get that from what I said? Everything in that respect is just as it was because the git source tree is still publicly available, patches are still posted to lkml for review and Linus still gets his say as we see from this thread. Most of the primary contributors are paid by companies but the companies are unrelated like RedHat, IBM, Oracle, MS, Google, etc so it's not like there's anything untoward happening behind the scenes.

http://www.linuxfoundation.org/news...tion-releases-annual-linux-development-report

 

[Kiska]

Wat?
{SNIP}
Open source is nice, an open source project is hell.

Its a laptop and no its took me 2 hours to reset the BIOS of my Lenovo SL500c laptop by pulling it out and erasing it then downloaded the standard BIOS firmware from Lenovo inserting it into another computer where I can do a reset.
Yes I must agree with you.

 

[_Rick_]

Why would the kernel flash your BIOS EEPROM?
If it did, that's a about as serious a bug as it gets.

While there is a driver, to access nvram, it's rarely used, and shouldn't touch the actual BIOS code.

 

[Kiska]

Why would the kernel flash your BIOS EEPROM?
If it did, that's a about as serious a bug as it gets.

While there is a driver, to access nvram, it's rarely used, and shouldn't touch the actual BIOS code.

I have no idea. As you can see I was pretty angry about the whole kernel and I hate Windows but until I got the BIOS back I had to use Windows to flash the BIOS back to firmware version 1.37 for Lenovo that firmware is outdated but still works. BTW that firmware is from 2006. Well good night
James aka Kiska

 

[Elixer]

How would you get that from what I said? Everything in that respect is just as it was because the git source tree is still publicly available, patches are still posted to lkml for review and Linus still gets his say as we see from this thread. Most of the primary contributors are paid by companies but the companies are unrelated like RedHat, IBM, Oracle, MS, Google, etc so it's not like there's anything untoward happening behind the scenes.

http://www.linuxfoundation.org/news...tion-releases-annual-linux-development-report

No, just trying to understand the chain of events.
I am assuming that since this person works at RedHat, that they have a checklist / peer review / tests being done before the commit.