Hiya,
I'm in the market for a new laptop on which to run Ubuntu, and Lenovo's Thinkpads look like a good choice.
I like the idea of running an open BIOS, but looking at past posts to the mailing list, it appears that certain chipsets (such as those for the T60) are not supported:
http://article.gmane.org/gmane.linux.bios/19651
I believe that certain (older) Thinkpad chipsets are supported, but what about the 945 or the 965 (for the T61) ?
The "Supported Chipsets" page (below) hasn't had any updates in a couple of months -- is that the best place for me to determine if Linux BIOS is compatible with a particular chipset/system?
http://linuxbios.org/index.php/Supported_Chipsets_and_Devices
Thanks, -- Robinson
On Tue, Sep 18, 2007 at 03:49:14PM -0400, Robinson Tryon wrote:
I'm in the market for a new laptop on which to run Ubuntu, and Lenovo's Thinkpads look like a good choice.
I like the idea of running an open BIOS, but looking at past posts to the mailing list, it appears that certain chipsets (such as those for the T60) are not supported:
Correct. No laptop whatsoever (except for the OLPC) is supported by LinuxBIOS at the moment. We definately like to change that, but at the moment there's no support at all.
I believe that certain (older) Thinkpad chipsets are supported, but what about the 945
Nope.
or the 965 (for the T61) ?
Nope.
The "Supported Chipsets" page (below) hasn't had any updates in a couple of months -- is that the best place for me to determine if Linux BIOS is compatible with a particular chipset/system?
Yes. That page and this mailing list (or, ultimately: the code). The last chipset which was added was the Intel 810. Supporting new chipsets is a _major_ piece of work, so don't expect this to happen every week or so...
Uwe.
On 9/18/07, Uwe Hermann uwe@hermann-uwe.de wrote:
Correct. No laptop whatsoever (except for the OLPC) is supported by LinuxBIOS at the moment. We definately like to change that, but at the moment there's no support at all. ... ... The last chipset which was added was the Intel 810. Supporting new chipsets is a _major_ piece of work, so don't expect this to happen every week or so...
So Intel doesn't want to open up the specs on their hardware, whereas AMD is willing to play ball (at least to some extent) ?
Would writing to Intel (if only to be counted as one more data point) help at all?
I thought that Intel had a pretty good track record with the FOSS community, what with their wireless cards and integrated graphics cards having nice, open drivers... but perhaps it was all about the money and they didn't actually care about maintaining a relationship with the community at all...
Robinson Tryon wrote:
On 9/18/07, Uwe Hermann uwe@hermann-uwe.de wrote:
Correct. No laptop whatsoever (except for the OLPC) is supported by LinuxBIOS at the moment. We definately like to change that, but at the moment there's no support at all. ... ... The last chipset which was added was the Intel 810. Supporting new chipsets is a _major_ piece of work, so don't expect this to happen every week or so...
So Intel doesn't want to open up the specs on their hardware, whereas AMD is willing to play ball (at least to some extent) ?
Would writing to Intel (if only to be counted as one more data point) help at all?
I thought that Intel had a pretty good track record with the FOSS community, what with their wireless cards and integrated graphics cards having nice, open drivers... but perhaps it was all about the money and they didn't actually care about maintaining a relationship with the community at all...
Intel's datasheets are available freely online. I don't know if there's more info that they send to Award, etc, but the i440bx, i810, and now i830 ports were all written from those. It's the embedded controllers, which may/may not be Intel's design, that are the biggest discouragement for most people considering laptop ports.
-Corey
On Tue, Sep 18, 2007 at 07:04:01PM -0400, Corey Osgood wrote:
Robinson Tryon wrote:
Would writing to Intel (if only to be counted as one more data point) help at all?
No.
I thought that Intel had a pretty good track record with the FOSS community, what with their wireless cards and integrated graphics cards having nice, open drivers... but perhaps it was all about the money and they didn't actually care about maintaining a relationship with the community at all...
Intel is a big company, by choice or accident the attitude towards open source software is not the same everywhere.
Intel's datasheets are available freely online. I don't know if there's more info that they send to Award, etc, but the i440bx, i810, and now i830 ports were all written from those.
These are all pretty old chipsets. Public data sheets for the newer ones have significantly less useful information.
The fact that Intel also have their own boot software EFI may also be a reason why they don't throw data sheets at us, but that is pure speculation.
It's the embedded controllers, which may/may not be Intel's design, that are the biggest discouragement for most people considering laptop ports.
Very true. Noone usually knows how the black box works and it's often involved in everything that differs between a laptop and a desktop; ie. all the reasons the user bought the laptop like power management, integrated special keys etc.
//Peter
On 9/19/07, Peter Stuge peter@stuge.se wrote:
On Tue, Sep 18, 2007 at 07:04:01PM -0400, Corey Osgood wrote:
Robinson Tryon wrote:
Would writing to Intel (if only to be counted as one more data point) help at all?
No.
I just ran across an FSF web page in support of LinuxBIOS, and on that page the FSF suggests writing directly to Intel:
http://www.fsf.org/campaigns/supportlinuxbios.html "Send e-mail to Intel, asking them to cooperate with a fully free BIOS. Also send them a snail mail letter at: Intel Corporation 2200 Mission College Blvd. Santa Clara, CA 95052 USA"
If it's the opinion of the LinuxBIOS project that writing to Intel isn't helpful, what else can I and other persons do to put pressure on Intel (and like companies) to get them to release specs on their laptop chipsets and embedded controllers?
If AMD is being much more supportive of the LinuxBIOS project than Intel, what about approaching them? If I'm not mistaken, there are a couple of AMD employees on this mailing list who are (were?) actively engaged in getting LinuxBIOS to work with various AMD chipsets...
On Thu, Sep 20, 2007 at 06:18:28PM -0400, Robinson Tryon wrote:
Would writing to Intel (if only to be counted as one more data point) help at all?
No.
I just ran across an FSF web page in support of LinuxBIOS, and on that page the FSF suggests writing directly to Intel:
http://www.fsf.org/campaigns/supportlinuxbios.html "Send e-mail to Intel, asking them to cooperate with a fully free BIOS. Also send them a snail mail letter at: Intel Corporation 2200 Mission College Blvd. Santa Clara, CA 95052 USA"
If it's the opinion of the LinuxBIOS project that writing to Intel isn't helpful, what else can I and other persons do to put pressure on Intel (and like companies) to get them to release specs on their laptop chipsets and embedded controllers?
Call me cynical but I doubt individuals can make a difference here. I suppose if there was a cut of, say, 50% in their product sales they might re-assess the situation, but again I don't think that kind of organization, marketing and education is feasible.
If AMD is being much more supportive of the LinuxBIOS project than Intel, what about approaching them? If I'm not mistaken, there are a couple of AMD employees on this mailing list who are (were?) actively engaged in getting LinuxBIOS to work with various AMD chipsets...
Yes, AMD supports LinuxBIOS as one of the boot software options on their systems. People at AMD work with LB. We think it is great of course! :)
I'm not sure I understand what we'd approach them about. Every single piece of hardware from AMD isn't supported yet, but I believe that will come with time. They do make a few chips. :)
//Peter
On Fri, Sep 21, 2007 at 01:08:40AM +0200, Peter Stuge wrote:
I just ran across an FSF web page in support of LinuxBIOS, and on that page the FSF suggests writing directly to Intel:
http://www.fsf.org/campaigns/supportlinuxbios.html "Send e-mail to Intel, asking them to cooperate with a fully free BIOS. Also send them a snail mail letter at: Intel Corporation 2200 Mission College Blvd. Santa Clara, CA 95052 USA"
Yes, please do so.
If it's the opinion of the LinuxBIOS project that writing to Intel isn't helpful, what else can I and other persons do to put pressure on Intel (and like companies) to get them to release specs on their laptop chipsets and embedded controllers?
Call me cynical but I doubt individuals can make a difference here.
Nah, don't be too pessimistic. Ongoing nagging from customers _will_ make most companies rethink their decisions and priorities sooner or later. Sure, there's no guarantee that they will react (soon, or at all), but if you don't even try you have already lost.
If AMD is being much more supportive of the LinuxBIOS project than Intel, what about approaching them? If I'm not mistaken, there are a couple of AMD employees on this mailing list who are (were?) actively engaged in getting LinuxBIOS to work with various AMD chipsets...
Yes, AMD supports LinuxBIOS as one of the boot software options on their systems. People at AMD work with LB. We think it is great of course! :)
Full ACK! The support from AMD is really great and very appreciated!
Uwe.
On 9/21/07, Uwe Hermann uwe@hermann-uwe.de wrote:
Yes, AMD supports LinuxBIOS as one of the boot software options on their systems. People at AMD work with LB. We think it is great of course! :)
Full ACK! The support from AMD is really great and very appreciated!
Marvelous! It's great to hear that AMD is so supportive of this project.
You mentioned in a previous email that it's a bit of work to get a new chipset supported:
On 9/18/07, Uwe Hermann uwe@hermann-uwe.de wrote:
...No laptop whatsoever (except for the OLPC) is supported by LinuxBIOS at the moment... Supporting new chipsets is a _major_ piece of work, so don't expect this to happen every week or so...
But what, exactly, would it take to get a chipset supported?
Let's say that our target was an AMD Turion 64 X2 processor, and we were looking at AMD's own chipsets: - AMD M690T - AMD M690G - AMD M690V - ATI Radeon Xpress 1100 - ATI Radeon Xpress 1150
What kind of time/money combination would be necessary to hack in support for one of these chipsets and get a laptop model that could run LinuxBIOS?
I'd certainly contribute money to a targeted endeavor such as that -- and I'm sure that there are several people in the FOSS community who, like RMS, ESR, and Mark Shuttleworth, would give financial support as well. I'm pretty certain about Shuttleworth, because he has an entry on his blog that is right along the lines of my goal:
http://www.markshuttleworth.com/archives/131 "To that end I'd like to build up a list of people who are interested in this idea, and would potentially buy a high-powered laptop if it were guaranteed to work completely with free software drivers and [LinuxBIOS]."
Having at least one supported, commercially-available laptop would be a HUGE win for the LinuxBIOS project.
On Mon, Sep 24, 2007 at 04:16:32AM -0400, Robinson Tryon wrote:
It's great to hear that AMD is so supportive of this project.
..
laptop
This was discussed on IRC shortly after Mark's post. For now, the graphics hardware simply isn't there to serve this market.
LB can run the binary VGA BIOS in the emulator for init but for operation going through legacy drivers/APIs is not the right way.
Until there is powerful open graphics hardware from someone the laptop wont happen. But keep pushing for it!
(Intel graphics is not available separately yet, so we would need a completely different business case in order to use Intel CPU and chipset in the laptop. Intel graphics is built-in to the Intel chipset, which only works with Intel CPUs.)
//Peter
On Tue, Sep 25, 2007 at 03:54:20AM +0200, Peter Stuge wrote:
On Mon, Sep 24, 2007 at 04:16:32AM -0400, Robinson Tryon wrote:
It's great to hear that AMD is so supportive of this project.
..
laptop
This was discussed on IRC shortly after Mark's post. For now, the
Which post? URL?
graphics hardware simply isn't there to serve this market.
LB can run the binary VGA BIOS in the emulator for init but for operation going through legacy drivers/APIs is not the right way.
Sure, it's not ideal, but as a temporary "hack" until we have something better -- why not?
Maybe a totally-free-laptop-from-scratch won't happen very soon, but an existing laptop with supported chipset, superio, EC and socketed ROM would be a (relatively) easy target to at least support _some_ laptop at all...
Personally, I would be really happy to see something like this happen.
Uwe.
-
Corey Osgood -
Peter Stuge -
Robinson Tryon -
Uwe Hermann