Linus / boss speaks on buggy code in current kernels

If you want to talk about your personal paperweight, you've found the right place!

Moderator: Moderators

Linus / boss speaks on buggy code in current kernels

Postby WarraWarra » Thu Aug 16, 2007 3:02

Though a few persons might find this interesting about recent bug in the kernel's that is not Sabayon related but gives Sabayon a bad name and causes several problems.

If you read this link you will understand.
Very interesting and helpfull for Sabayon in the future.
Sagely Hen
Posts: 1989
Joined: Wed Jan 03, 2007 21:01
Location: 31.324270, -113.585511

Postby robhwill » Thu Aug 16, 2007 5:23

thanks for that, very interesting. personally i'm really frustrated by the "stepchild" treatment linux
accepts in the market from hardware vendors (my current nemesis are Broadcom & HP). as i'm
not a dev i could be way off here but i think linux aadvocacy misses great opportunity to first write, then implement, and finally enforce hardware/driver/api standards. we accept that ms$winblows can/does do it but don't expect linux to 'drive' hardware acceptance thru certification or something like it. very simply, given the preponderance of *servers* running linux and unix standards/dev apis and the like should come from the linux/unix space *first* then trickle over to ms$blows!!
just my 2cents. in any event, in the meanwhile, i'm considering an approach to consolidating and sequencing by date/version certain driver -- bug issues w/solutions. maybe something like bugtracker
but thinks it needs be a little more fluid w/very solid per-distro references. any ideas??
tia.. .. ..
Have A Healthy, Prosperous Day!!
Young Hen
Posts: 27
Joined: Mon May 07, 2007 22:54
Location: Michigan / US

Postby WarraWarra » Thu Aug 16, 2007 7:21

Same here.
Great idea not sure how-to though. If you can get into gentoo and not get nuked by them then they could likely use the driver date error list and this would help whole linux / gentoo / Sabayon community in the long run + end up back at to help them out.

I was even thinking today we should make a list of working parts in the pc's laptops and what drivers makes them tick + links to that specific wiki part that tells you how to fix it + then the working drivers / avoid this driver list.
Problem is just to keep such a list updated and filled with correct info as so many things end up with problems as a result of drivers or bugs that makes a snowball effect and all leads to the root bug in code somewhere.

This way it can end up in a stable version of SL and be build upon in future instead of reinventing the wheel with buggy driver. The effect of that link's results has been the reason of most of the bug's in SL that I encountered. No buggy drivers no buggy SL.
Sagely Hen
Posts: 1989
Joined: Wed Jan 03, 2007 21:01
Location: 31.324270, -113.585511

Postby mdmarmer » Fri Aug 17, 2007 12:43

that was about the 2.6.23 kernel and the patches have not been merged yet.

So ... Sabayon is not using these patches

Advanced Hen
Posts: 240
Joined: Thu Oct 19, 2006 1:46
Location: dayton OH usa

Postby WarraWarra » Fri Aug 17, 2007 22:29

His anger at this bugs in the kernel comes a long way already and this was just the last straw, if it happens once it is okay but it has been happening on and off over several 2.6.* kernel's and more frequently in latest kernel's. The problem is not the person doing it but what happens as a result.

Either way plenty of hours has been waisted because of this to try and fix Sabayon that was actually not broken except for bad drivers in the kernel's and it could have been spend one more important things if the bug's in the drivers was fixed as it should have been at the kernel part not at Sabayon.
Example look at the bcm43xx and knetworkmanager + several other's history over the last 6 months or more.

He could have just followed the rules and keep it clean and safe , result = sick Linux.

If the kernel has bug's / disease = distro's have the upset stomach / food poisoning and only by fixing the root / disease will the linux / distro's become healthy again + a bit of medicine for each contaminated distro.

We feel it more as we are on the cutting edge and not like the other distro now only starting to play around with the 2.6.18 kernels and their nightmares is just about to begin where we have passed that about 8 months ago.
There is good reasoning behind XEN and it's use of the older clean/cleaned out kernel's as they can not afford to have bugs and then 50% of the internet crashes as a result of big company's using XEN + bugs in the kernel's.

Code: Select all
He made it very clear that at this late stage (ready for rc3), the types of bugs being found and, more specifically the types of patches being applied, were a little too loose for comfort and wholly undesirable.

Code: Select all
"I simply want you to be more careful.  *Much* more careful. As it is, I end up always being afraid of merging your patch-series ... Bugs happen, but (a) they should happen during the merge window, not when we're in stabilization phase and (b) the percentages here were just not very good."

Code: Select all
"I end up always being afraid of merging your patch"
= Always ?? = hapenning long time already or maybe always is just the last 1 week ?? and he has been helping the out since 1994 about so why make the mistakes ?? Bad hair day ??

O well who cares as long as it gets fixed / stoped before it becomes a epidemic.
Sagely Hen
Posts: 1989
Joined: Wed Jan 03, 2007 21:01
Location: 31.324270, -113.585511

Return to Off Topic

Who is online

Users browsing this forum: No registered users and 2 guests