2005 Linux kernel Summit.

[drag]

See the summary here:
http://lwn.net/Articles/KernelSummit2005/

lots of neat stuff.

 

[bersl2]

What I think qualifies as "neat stuff":

Intel's plans for the x86 family include going to a dual-core architecture across the entire product line. The processors will also move to common internal core, based on the mobile processors. When asked, Per confirmed that this change is being driven by power management concerns. He noted that Intel is able to build processors with some fairly heavy power requirements, and that even large server systems have power management issues.

OK, further confirmation that Intel's processor division is not completely insane.

Looking further ahead, AMD plans to offer multi-core (meaning "greater than two cores") processors with an L3 cache shared between them. The processors will work with a 48-bit physical address space, and will be able to handle individual pages up to 1GB in size. There will also be a timestamp counter which continues to count at the same rate regardless of the processor's current clock speed; this announcement drew some applause from the audience.

Every other in-kernel timer is either affected by frequency change or is too slow. Applause indeed.

Arjan estimates that the ExecShield patches mitigate the effects of some 25-30% of reported vulnerabilities in Linux systems. Even so, there was no discussion of merging the remaining ExecShield patches in the near future.

Everybody loves security.

Just read the whole of the power management article.

 

[n0cmonkey]

Add PaX into the kernel proper and get the gcc guys to finally add in propolice and we're set. Those gcc guys just keep dragging their feet.

Neat stuff here though.

 

[drag]

I like the hardware manufacturers and such. Linux is starting to have a bigger impact on how people design hardware.

For instance one of the manufactures present said that "double digit' percentages of their sales are to Linux customers.

They talked about how it's very advantagous for them to get their drivers into the mainline kernel and even one guy said that they the firmware blobs are a biggest problem that customers have to deal with and that they started putting the firmware back into the hardware in order to make it easier to deal with and make it easier to make more stable drivers.

Also, I don't know if it's in there or not, (it may only be in the pay section of LWN at the moment,) but Intel gave a presentation and that they are very driven for good Linux support and that, get this, that they use Linux 100% for their proccessor design and manufacturing OS needs.

 

[n0cmonkey]

Intel doesn't care about F/OSS. They give Linux lip service because it sells. That's sad.

 

[drag]

If intel responds to their customer's demands.. that's good enough for me.

That's the ony way it's going to happen for lots of the hardware manufacturers to realise that they sell hardware rather then selling drivers.

I know intel are dicks a lot of the time and that they would be perfectly happy to support hardware-based DRM for Microsoft (and probably Apple) to allow those companies to further lock their customers down to specific software. That and refusing to release firmware for redistribution is stupid, too. And AMD proccessors are better... among other things.

But they are the only company that has modern graphics that you can get full OSS support for.. (should be included in X.org 6.8.3)

It's very difficult to get that working with AMD computers.. you'd have to pretty much use a older ATI card...

 

[bersl2]

Originally posted by: drag
But they are the only company that has modern graphics that you can get full OSS support for.. (should be included in X.org 6.8.3)

Strictly speaking, Unichrome works very well too---with little thanks to VIA for releasing code but licensing it in such a way as to be GPL-incompatable.

 

[drag]

Ya.. The Via stuff is pretty ok. Although they did that thing recently were they only released only the 2d portion of their drivers.

They have other things going on, like the mpeg2/4 accelerators and such that makes their mini-itx boards good for media playback and such even though they have very slow video chipsets and cpus.

 

[drag]

Originally posted by: n0cmonkey
Add PaX into the kernel proper and get the gcc guys to finally add in propolice and we're set. Those gcc guys just keep dragging their feet.

Neat stuff here though.

from:
http://gcc.gnu.org/

July 17, 2005
GCC 4.1 stage 2 has been closed. The following projects were contributed during stage 1 and stage 2: New C Parser, LibAda GNATTools Branch, Code Sinking, Improved phi-opt, Structure Aliasing, Autovectorization Enhancements, Hot and Cold Partitioning, SMS Improvements, Integrated Immediate Uses, Tree Optimizer Cleanups, Variable-argument Optimization, Redesigned VEC API, IPA Infrastructure, Altivec Rewrite, Warning Message Control, New SSA Operand Cache Implementation, Safe Builtins, Reimplementation of IBM Pro Police Stack Detector, New DECL hierarchy. More information about these projects can be found at GCC 4.1 projects.

Although I don't know what they mean by 'reimplementation', though.

 

[n0cmonkey]

Originally posted by: drag

Originally posted by: n0cmonkey
Add PaX into the kernel proper and get the gcc guys to finally add in propolice and we're set. Those gcc guys just keep dragging their feet.

Neat stuff here though.

from:
http://gcc.gnu.org/

July 17, 2005
GCC 4.1 stage 2 has been closed. The following projects were contributed during stage 1 and stage 2: New C Parser, LibAda GNATTools Branch, Code Sinking, Improved phi-opt, Structure Aliasing, Autovectorization Enhancements, Hot and Cold Partitioning, SMS Improvements, Integrated Immediate Uses, Tree Optimizer Cleanups, Variable-argument Optimization, Redesigned VEC API, IPA Infrastructure, Altivec Rewrite, Warning Message Control, New SSA Operand Cache Implementation, Safe Builtins, Reimplementation of IBM Pro Police Stack Detector, New DECL hierarchy. More information about these projects can be found at GCC 4.1 projects.

Although I don't know what they mean by 'reimplementation', though.

That could be a problem. Their leader said he would tell them to add propolice. Not sure if I'm going to trust something that big and new...

 

[n0cmonkey]

Originally posted by: drag
If intel responds to their customer's demands.. that's good enough for me.

But they don't. Their firmware can't be distributed without signing something, and there is little to no documentation for their hardware. The drivers suck because Intel won't open them up a bit.