Discussion:
Drop support for x86-32
(too old to reply)
wbrana
2012-08-23 10:41:42 UTC
Permalink
Microsoft will drop support for x86-32 in Windows 9.
Linux could do same.
http://www.networkworld.com/community/blog/windows-9-details-are-already-emerging
wbrana
2012-08-23 17:22:07 UTC
Permalink
x86-32
- is deprecated since Linux supports X32.
- will slow down adoption of X32 - there won't be X32 versions of many
software - if new ABI was added, old one should be removed
- wastes time of developers who can spend their time supporting X32
instead of x86-32 or support x86-64 only as 99% of users will be able
to run x86-64 software if x86-32 will be dropped
- wouldn't be dropped this year, but there should be plan when it will
be dropped e.g. when Windows 9 will be released

On 8/23/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> Why?
Ondrej Zary
2012-08-23 17:51:27 UTC
Permalink
On Thursday 23 August 2012 19:22:07 wbrana wrote:
> x86-32
> - is deprecated since Linux supports X32.
> - will slow down adoption of X32 - there won't be X32 versions of many
> software - if new ABI was added, old one should be removed
> - wastes time of developers who can spend their time supporting X32
> instead of x86-32 or support x86-64 only as 99% of users will be able
> to run x86-64 software if x86-32 will be dropped
> - wouldn't be dropped this year, but there should be plan when it will
> be dropped e.g. when Windows 9 will be released
>
> On 8/23/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> > Why?

Please stop trolling (and top-posting). Linux is NOT Windows where people must
throw out their hardware because it stopped working in new version. There are
millions of 32-bit x86 machines all around the world. If new Windows will not
run on them, Linux will.

--
Ondrej Zary
wbrana
2012-08-23 18:07:07 UTC
Permalink
On 8/23/12, Ondrej Zary <***@rainbow-software.org> wrote:
> Please stop trolling (and top-posting). Linux is NOT Windows where people
> must
> throw out their hardware because it stopped working in new version. There
> are
> millions of 32-bit x86 machines all around the world. If new Windows will
> not
> run on them, Linux will.
Windows 9 will be probably released in 2015. Linux always has tree
with long term support, which means support for x86-32 would be
dropped in 2017. In 2017 all 32-bit machines will be trash.
Bernd Petrovitsch
2012-08-24 12:40:56 UTC
Permalink
On Don, 2012-08-23 at 20:07 +0200, wbrana wrote:
> On 8/23/12, Ondrej Zary <***@rainbow-software.org> wrote:
> > Please stop trolling (and top-posting). Linux is NOT Windows where people
> > must
> > throw out their hardware because it stopped working in new version. There
> > are
> > millions of 32-bit x86 machines all around the world. If new Windows will
> > not
> > run on them, Linux will.
> Windows 9 will be probably released in 2015. Linux always has tree

What do you mean with "Linux"? The Linux kernel as such? Some (and
which) distributions?

> with long term support, which means support for x86-32 would be
> dropped in 2017. In 2017 all 32-bit machines will be trash.

The long-term-supported *Linux kernel* won't be eternally there BTW. And
their maintainers decide when they drop the support - and others could
take over then BTW.

And you obviously never thought about embedded devices.
Servers, laptops, notebooks and desktop computers are not the whole
computing world - and from the pure numbers not even the majority BTW.

It is - obviously(?) - all the world that MSFT cares about because they
were never really present elsewhere. And we know that it is not a
technical reason that we pay the MSFT tax for each laptop ....

Please stop trolling and start sending patches. Thank you.

On Don, 2012-08-23 at 20:40 +0200, wbrana wrote:
On 8/23/12, David Daney <***@gmail.com> wrote:
> > I suspected as much. So from your point of view, this issue is of
> > exactly zero importance.
> I'm using software which is developed by others. As I already said
> many software would be developed faster if x86-32 could be dropped.

Please proof that it will be developed "faster" (whatever that means
to you). Thank you.

> Support for x86-32 can mean no support for X32.

So you want the Linux kernel people to drop support for some
architecture in the hope others follow (because you will consequently
send mails there with "the Linux kernel dropped the support, you should
too" hoping that some other software will be developed
"faster" (whatever that means to you)?

*If* you really miss something in some other parts (compilers,
virtualization, ...) or they developing to slow *for you*, help them and
send patches there but do not try to lure others into fighting your
cause.

Sorry to all others for feeding the troll,
Bernd
--
Bernd Petrovitsch Email : ***@petrovitsch.priv.at
LUGA : http://www.luga.at
wbrana
2012-08-24 12:59:29 UTC
Permalink
On 8/24/12, Bernd Petrovitsch <***@petrovitsch.priv.at> wrote:
> What do you mean with "Linux"? The Linux kernel as such? Some (and
> which) distributions?
Linux kernel first, distributions and software will follow
>
> And you obviously never thought about embedded devices.
> Servers, laptops, notebooks and desktop computers are not the whole
> computing world - and from the pure numbers not even the majority BTW.
I don't request dropping support for ARM and other platforms which are
still in production

> Please proof that it will be developed "faster" (whatever that means
> to you). Thank you.
Developers won't have to maintain e.g. x86-32 assembler code and can
work on x32 and x86-64 instead.

> So you want the Linux kernel people to drop support for some
> architecture in the hope others follow (because you will consequently
> send mails there with "the Linux kernel dropped the support, you should
> too" hoping that some other software will be developed
> "faster" (whatever that means to you)?
exactly

> *If* you really miss something in some other parts (compilers,
> virtualization, ...) or they developing to slow *for you*, help them and
> send patches there but do not try to lure others into fighting your
> cause.
I don't have knowledge about compilers and virtualization. I work on
general GUI applications (Java and C++) and web pages.
wbrana
2012-08-24 13:51:55 UTC
Permalink
On 8/24/12, wbrana <***@gmail.com> wrote:
>> *If* you really miss something in some other parts (compilers,
>> virtualization, ...) or they developing to slow *for you*, help them and
>> send patches there but do not try to lure others into fighting your
>> cause.
> I don't have knowledge about compilers and virtualization. I work on
> general GUI applications (Java and C++) and web pages.
>
Open source examples of my work
bug fixes in NetBeans
http://netbeans.org/bugzilla/show_bug.cgi?id=134281
GUI frontend to MPlayer
https://sourceforge.net/projects/qemplayer/

greetings to Phoronix
Bernd Petrovitsch
2012-08-24 15:48:00 UTC
Permalink
On Fre, 2012-08-24 at 14:59 +0200, wbrana wrote:
> On 8/24/12, Bernd Petrovitsch <***@petrovitsch.priv.at> wrote:
[...]
> > And you obviously never thought about embedded devices.
> > Servers, laptops, notebooks and desktop computers are not the whole
> > computing world - and from the pure numbers not even the majority BTW.
> I don't request dropping support for ARM and other platforms which are
> still in production

You really think that there are no 32bit x86-compatible CPUs in the
embedded world?

Bernd
--
Bernd Petrovitsch Email : ***@petrovitsch.priv.at
LUGA : http://www.luga.at
wbrana
2012-08-24 15:55:08 UTC
Permalink
On 8/24/12, Bernd Petrovitsch <***@petrovitsch.priv.at> wrote:
> You really think that there are no 32bit x86-compatible CPUs in the
> embedded world?
x86-32 would be supported by long term tree until all x86-32 CPU disappear
wbrana
2012-08-24 16:11:54 UTC
Permalink
On 8/24/12, Chen <***@gmail.com> wrote:
> Are u wasting your time on trolling?
I'm discussing my proposal.
Martin Nybo Andersen
2012-08-24 16:14:16 UTC
Permalink
On Friday 24 August 2012 17:55:08 wbrana wrote:
> On 8/24/12, Bernd Petrovitsch <***@petrovitsch.priv.at> wrote:
> > You really think that there are no 32bit x86-compatible CPUs in the
> > embedded world?
>
> x86-32 would be supported by long term tree until all x86-32 CPU disappear

Ahh..., so the development time saved by not supporting x86-32 in mainline can
now be used by backporting new features to the forementioned long term tree?

(And I'm still not sure why one would run 32-bit applications on a 64-bit
architecture...)

--
Cheers,
Martin
wbrana
2012-08-24 16:17:14 UTC
Permalink
On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> Ahh..., so the development time saved by not supporting x86-32 in mainline
> can
> now be used by backporting new features to the forementioned long term
> tree?
new features won't be backported
Bobby Powers
2012-08-24 16:20:00 UTC
Permalink
On Fri, Aug 24, 2012 at 9:17 AM, wbrana <***@gmail.com> wrote:
> On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
>> Ahh..., so the development time saved by not supporting x86-32 in mainline
>> can
>> now be used by backporting new features to the forementioned long term
>> tree?
> new features won't be backported

Did you have anyone in particular you wanted to volunteer as a
maintainer for this?

> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to ***@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
wbrana
2012-08-24 16:24:11 UTC
Permalink
On 8/24/12, Bobby Powers <***@gmail.com> wrote:
> Did you have anyone in particular you wanted to volunteer as a
> maintainer for this?
Do you mean maintainer for long term tree? There is always one.
Martin Nybo Andersen
2012-08-24 16:20:54 UTC
Permalink
On Friday 24 August 2012 18:17:14 wbrana wrote:
> On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> > Ahh..., so the development time saved by not supporting x86-32 in
> > mainline can
> > now be used by backporting new features to the forementioned long term
> > tree?
>
> new features won't be backported

Ok. Seriously.

Dropping one of the most used architectures for no apparent reason makes no
sense at all.

--
Cheers,
Martin
wbrana
2012-08-24 16:26:27 UTC
Permalink
On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> Dropping one of the most used architectures for no apparent reason makes no
> sense at all.
x86-32 won't be one of the most used architectures in 2017.
wbrana
2012-08-24 16:39:32 UTC
Permalink
On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> Then I suggest you wait another five years with this discussion.
It is important to announce this year that mainline kernel will drop
support for x86-32 in 2015 and only long term tree will support x86-32
after 2015. People shouldn't be surprised.
Chris Friesen
2012-08-24 16:24:16 UTC
Permalink
On 08/24/2012 10:14 AM, Martin Nybo Andersen wrote:
> (And I'm still not sure why one would run 32-bit applications on a 64-bit
> architecture...)

There are several architectures (powerpc comes to mind) where 32-bit
userspace on 64-bit kernel is the norm because it offers performance
advantages due to the smaller pointer size if the process doesn' t need
the larger address space. This is the rationale behind the new x32 ABI
for x86-64 kernels.

You might also want to run legacy apps (which can't be recompiled) on
new hardware.

Chris
Martin Nybo Andersen
2012-08-24 16:54:45 UTC
Permalink
On Friday 24 August 2012 18:24:16 Chris Friesen wrote:
> On 08/24/2012 10:14 AM, Martin Nybo Andersen wrote:
> > (And I'm still not sure why one would run 32-bit applications on a 64-bit
> > architecture...)
>
> There are several architectures (powerpc comes to mind) where 32-bit
> userspace on 64-bit kernel is the norm because it offers performance
> advantages due to the smaller pointer size if the process doesn' t need
> the larger address space. This is the rationale behind the new x32 ABI
> for x86-64 kernels.

For specific cases surely a specific ABI will be faster than another, and
because of that I love that we are free to choose when using Linux. Personally
I'd hate maintaining up to three versions of the same userspace code, while
other might need it (or even do it happily) -- but that is another story.

What I'd hate even more is rendering my old working hardware useless by
removing x86-32 support from the kernel. To reason the removal by saying
"Microsoft plans to do it" just makes me go bonkers...

> You might also want to run legacy apps (which can't be recompiled) on
> new hardware.

These legacy apps will most likely be compiled for x86-32 and not x32 (an
argument for not removing x86-32 support on a running x86-64 kernel).

--
Cheers,
Martin
wbrana
2012-08-24 17:05:53 UTC
Permalink
On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> What I'd hate even more is rendering my old working hardware useless by
> removing x86-32 support from the kernel. To reason the removal by saying
> "Microsoft plans to do it" just makes me go bonkers...
Your old hardware will work fine with long term kernel.

> These legacy apps will most likely be compiled for x86-32 and not x32 (an
> argument for not removing x86-32 support on a running x86-64 kernel).
Which legacy apps do you mean?
Chris Friesen
2012-08-24 17:25:17 UTC
Permalink
On 08/24/2012 11:05 AM, wbrana wrote:
> On 8/24/12, Martin Nybo Andersen<***@tweek.dk> wrote:
>> What I'd hate even more is rendering my old working hardware useless by
>> removing x86-32 support from the kernel. To reason the removal by saying
>> "Microsoft plans to do it" just makes me go bonkers...

> Your old hardware will work fine with long term kernel.

Give it up. x86-32 support isn't going away for a long time.

The kernel exists to support existing hardware, not the other way
around. As long as people want to keep using x86-32 on linux, it will
continue to be supported.

Chris
Brian Gerst
2012-08-24 17:28:41 UTC
Permalink
On Fri, Aug 24, 2012 at 1:05 PM, wbrana <***@gmail.com> wrote:
> On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
>> What I'd hate even more is rendering my old working hardware useless by
>> removing x86-32 support from the kernel. To reason the removal by saying
>> "Microsoft plans to do it" just makes me go bonkers...
> Your old hardware will work fine with long term kernel.

You still don't get it. Forget your "long term kernel" fantasy. That
is not what they are intended for. They are to provide a stable
kernel for the lifetime of a major (usually enterprise) distribution.
They provide security and bug fixes without the churn associated with
major releases.

There is absolutely no reason to remove support for hardware from
Linux unless nobody is willing to maintain it. As I said before,
there is so much more in common between x86-32 and x86-64 than there
is different, that x86-32 specific maintenance is almost nil. I
haven't seen any patches from you, so your opinions about the
maintenance burden don't carry much weight, versus those of people who
are contributors.

--
Brian Gerst
Gene Heskett
2012-08-24 18:11:42 UTC
Permalink
On Friday 24 August 2012, wbrana wrote:
>On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
>> What I'd hate even more is rendering my old working hardware useless by
>> removing x86-32 support from the kernel. To reason the removal by
>> saying "Microsoft plans to do it" just makes me go bonkers...
>
>Your old hardware will work fine with long term kernel.

Says you.

>> These legacy apps will most likely be compiled for x86-32 and not x32
>> (an argument for not removing x86-32 support on a running x86-64
>> kernel).
>
>Which legacy apps do you mean?

Any of them that have been in use, on 32 bit systems since the late 90's.
Among us 'old hands' the rate of 64 bit installs on 64 bit hardware is
minuscule at best, and here is one of them, running a 32 bit smp/pae
enabled install on amd phenom 4 core 64 bit hardware. Why? Because the
applications we run and have run for a decade or more simply have not ever
been rebuilt to run on 64 bit systems. They have no official maintainers
to do so because they were written well in the first place and continue to
do well on the 64 bit systems the makers have foisted off on us purely as a
speed requirement.

I have had 2 different 64 bit installs on this box in the last 4 years,
they were just barely noticeable as faster but they lasted less than a
month when I realized that the programs I had been using for a decade
simply were not available from the 64 bit only repo's and no one was
interested in building such old but still 100% useful stuff.

So quite trying to find a problem where there is none, and go apply
yourself to solving the lack of a full menu of the legacy apps built for 64
bit OS's and machines. Come back in a decade once you have solved that
problem and suggest this insanity again. That is approximately the right
word for this proposal today, insane.

Alan Cox probably said it better, but at my age, likely 3, maybe 4x yours,
I'll row in the same boat with Alan. Put your efforts someplace useful.
This, at this time, is not useful.

Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
My web page: <http://coyoteden.dyndns-free.com:85/gene> is up!
Results vary by individual.
wbrana
2012-08-24 18:18:49 UTC
Permalink
On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> That's right, but new hardware, that I wish to use with the old machines
> might
> not because of no backporting of new drivers. Same goes for new software
> utilising newer kernel features.
Which new hardware and which old machine do you want to use?

> Those mentioned by Chris Friesen, whose arguments you apparently ignored.
I can't find any e-mail where Chris Friesen mentioned exact legacy apps.

> You are allowed to compile most of the software running on Linux yourself.
> If
> you want a binary to use the x32 ABI, go compile.
I can't compile x32 versions of Firefox, Chromium, NVIDIA drivers, etc.
Martin Nybo Andersen
2012-08-24 18:35:38 UTC
Permalink
On Friday 24 August 2012 20:18:49 wbrana wrote:
> On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> > That's right, but new hardware, that I wish to use with the old machines
> > might
> > not because of no backporting of new drivers. Same goes for new software
> > utilising newer kernel features.
>
> Which new hardware and which old machine do you want to use?
I want to use *my* old machines (hey, I payed for them) on whatever new
hardware I can plug into them. I'm not an oracle and can't see into the
future, however USB has evolved, for instance, and will probably still do --
requiring new drivers as before.

> > Those mentioned by Chris Friesen, whose arguments you apparently ignored.
>
> I can't find any e-mail where Chris Friesen mentioned exact legacy apps.
Have a look in your inbox or one of the mailarchives. If that doesn't work,
then stop deleting half of the mails before replying.

> > You are allowed to compile most of the software running on Linux
> > yourself. If
> > you want a binary to use the x32 ABI, go compile.
>
> I can't compile x32 versions of Firefox, Chromium, NVIDIA drivers, etc.
The source code for Firefox and Chromium is available online. Go fetch and
compile. If they don't compile or execute as intended, I'm sure they would
like seeing a patch from you fixing the problem.

NVIDIA, on the other hand, is yet another reason for using nouveau.

--
Cheers,
Martin
wbrana
2012-08-24 18:59:33 UTC
Permalink
On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> I want to use *my* old machines (hey, I payed for them) on whatever new
> hardware I can plug into them. I'm not an oracle and can't see into the
> future, however USB has evolved, for instance, and will probably still do --
Your 32-bit machines will be likely dead in 2015. If it didn't happen
you will able to buy very cheap old 64-bit machine.

>> I can't find any e-mail where Chris Friesen mentioned exact legacy apps.
> Have a look in your inbox or one of the mailarchives. If that doesn't work,
>
> then stop deleting half of the mails before replying.
>
I don't delete e-mails. Can you copy-paste or give link?

> The source code for Firefox and Chromium is available online. Go fetch and
> compile. If they don't compile or execute as intended, I'm sure they would
> like seeing a patch from you fixing the problem.
Firefox and Chromium are large applications and can't be fixed by one user.

>
> NVIDIA, on the other hand, is yet another reason for using nouveau.
>
nouveau is useless garbage as most open source graphics drivers.
Martin Nybo Andersen
2012-08-24 19:53:24 UTC
Permalink
On Friday 24 August 2012 20:59:33 wbrana wrote:
> On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> > I want to use *my* old machines (hey, I payed for them) on whatever new
> > hardware I can plug into them. I'm not an oracle and can't see into the
> > future, however USB has evolved, for instance, and will probably still do
> > --
>
> Your 32-bit machines will be likely dead in 2015. If it didn't happen
> you will able to buy very cheap old 64-bit machine.
I don't think so. My trusty old Amiga 4000 from '93 still works, and I don't
think it'll die within the next three years. :-)

> >> I can't find any e-mail where Chris Friesen mentioned exact legacy apps.
> >
> > Have a look in your inbox or one of the mailarchives. If that doesn't
> > work,
> >
> > then stop deleting half of the mails before replying.
>
> I don't delete e-mails. Can you copy-paste or give link?
You delete part of them. Here's a link to Friesens mail:
https://lkml.org/lkml/2012/8/24/300

Quote:
"You might also want to run legacy apps (which can't be recompiled) on new
hardware."

I know he's not giving any names, however he's specific enough: Apps which
can't be recompiled. Ie. software for which you haven't got the source code or
a working compiler.

> > The source code for Firefox and Chromium is available online. Go fetch
> > and compile. If they don't compile or execute as intended, I'm sure they
> > would like seeing a patch from you fixing the problem.
>
> Firefox and Chromium are large applications and can't be fixed by one user.
Try telling Torvalds he couldn't create Linux, or Stallman that he couldn't
create GNU.

Seriously: Announce your work and people will help you if they feel that it is
worthwhile.

> > NVIDIA, on the other hand, is yet another reason for using nouveau.
>
> nouveau is useless garbage as most open source graphics drivers.
Quite the opposite. You should try it some day. The only thing I miss is
VDPAU. But then again, I use another computer for watching movies. It has a
Radeon card installed (using the open radeon driver -- not the flgrx). Full HD
and sound via HDMI running flawlessly. As does the integrated Intel GPU on my
eeepc 900 (the CPU seems to be the fastest chip on that netbook ... ;-) ...).

--
Cheers,
Martin
wbrana
2012-08-25 08:27:41 UTC
Permalink
On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> I know he's not giving any names, however he's specific enough: Apps which
> can't be recompiled. Ie. software for which you haven't got the source code
> or
> a working compiler.
Software used by 99% users will have alternative software which have
source code available or have 64-bit version.

> Try telling Torvalds he couldn't create Linux, or Stallman that he couldn't
> create GNU.
Chromium 22 is much more complex than Linux 1.0
linux-1.0.tar.bz2 - 20 contributors - size 1 MB
chromium-22.0.1229.14.tar.bz2 - size 200 MB

> Quite the opposite. You should try it some day. The only thing I miss is
> VDPAU. But then again, I use another computer for watching movies. It has a
> Radeon card installed (using the open radeon driver -- not the flgrx). Full
> HD
> and sound via HDMI running flawlessly. As does the integrated Intel GPU on
> my
> eeepc 900 (the CPU seems to be the fastest chip on that netbook ... ;-)
Open source drivers are black listed by Chromium because of instability
file software_rendering_list.json:
NVIDIA cards with nouveau drivers in Linux are crash-prone
The Intel Mobile 945 Express family of chipsets is not compatible with WebGL
Intel mesa drivers are crash-prone
ATI/AMD cards with third-party drivers in Linux are crash-prone
Drivers are unreliable for Optimus on Linux

http://www.phoronix.com/scan.php?page=article&item=intel_ivy_gpushow&num=10
Geforce 9800 GT with nouveau has worse performance than Ivy Bridge.
Radeon 6770 with open source drivers has far worse performance than Ivy Bridge.
There is no discrete Ivy Bridge graphics card.
Gene Heskett
2012-08-25 08:43:57 UTC
Permalink
On Saturday 25 August 2012, wbrana wrote:
>On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
>> I know he's not giving any names, however he's specific enough: Apps
>> which can't be recompiled. Ie. software for which you haven't got the
>> source code or
>> a working compiler.
>
>Software used by 99% users will have alternative software which have
>source code available or have 64-bit version.
>
>> Try telling Torvalds he couldn't create Linux, or Stallman that he
>> couldn't create GNU.
>
>Chromium 22 is much more complex than Linux 1.0
>linux-1.0.tar.bz2 - 20 contributors - size 1 MB
>chromium-22.0.1229.14.tar.bz2 - size 200 MB
>
>> Quite the opposite. You should try it some day. The only thing I miss
>> is VDPAU. But then again, I use another computer for watching movies.
>> It has a Radeon card installed (using the open radeon driver -- not
>> the flgrx). Full HD
>> and sound via HDMI running flawlessly. As does the integrated Intel GPU
>> on my
>> eeepc 900 (the CPU seems to be the fastest chip on that netbook ... ;-)
>
>Open source drivers are black listed by Chromium because of instability
>file software_rendering_list.json:
>NVIDIA cards with nouveau drivers in Linux are crash-prone

What rock did you just crawl out from under? Stuff gets fixed, I've been
using it since I installed Ubu10.04.4 LTS. Current uptime is half a day
short of 3 weeks.

>The Intel Mobile 945 Express family of chipsets is not compatible with
>WebGL Intel mesa drivers are crash-prone
>ATI/AMD cards with third-party drivers in Linux are crash-prone
>Drivers are unreliable for Optimus on Linux

Now that I can't argue with, given that what ATI/AMD has thrown over the
fence is based on their own code, and I have yet to get their own code to
run any ATI card I've ever had.

>http://www.phoronix.com/scan.php?page=article&item=intel_ivy_gpushow&num=
>10 Geforce 9800 GT with nouveau has worse performance than Ivy Bridge.
>Radeon 6770 with open source drivers has far worse performance than Ivy
>Bridge. There is no discrete Ivy Bridge graphics card.

Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
My web page: <http://coyoteden.dyndns-free.com:85/gene> is up!
... bacteriological warfare ... hard to believe we were once foolish
enough to play around with that.
-- McCoy, "The Omega Glory", stardate unknown
wbrana
2012-08-25 09:25:10 UTC
Permalink
On 8/25/12, Gene Heskett <***@wdtv.com> wrote:
> What rock did you just crawl out from under? Stuff gets fixed, I've been
> using it since I installed Ubu10.04.4 LTS. Current uptime is half a day
> short of 3 weeks.
If Chromium black listed graphics card, it means Chromium/Xserver
crashed if hardware acceleration was enabled. Chromium has to run
without hardware acceleration with nouveau and other open source
drivers.
wbrana
2012-08-25 12:20:05 UTC
Permalink
On 8/25/12, wbrana <***@gmail.com> wrote:
> Open source drivers are black listed by Chromium because of instability
> file software_rendering_list.json:
> NVIDIA cards with nouveau drivers in Linux are crash-prone
> The Intel Mobile 945 Express family of chipsets is not compatible with
> WebGL
> Intel mesa drivers are crash-prone
> ATI/AMD cards with third-party drivers in Linux are crash-prone
> Drivers are unreliable for Optimus on Linux
There is chance Ivy Bridge can accelerate Chromium, but it can't
accelerate open source game at sufficient frame rate - 60 fps
http://www.phoronix.com/scan.php?page=article&item=intel_ivy_gpushow&num=8
wbrana
2012-08-25 14:41:59 UTC
Permalink
On 8/25/12, wbrana <***@gmail.com> wrote:
> There is chance Ivy Bridge can accelerate Chromium, but it can't
> accelerate open source game at sufficient frame rate - 60 fps
> http://www.phoronix.com/scan.php?page=article&item=intel_ivy_gpushow&num=8
My 5-years-old Geforce 7300 GT provides 120 fps instead of 30 fps at 1920x1200.
Bernd Petrovitsch
2012-08-25 11:35:19 UTC
Permalink
On Fre, 2012-08-24 at 20:59 +0200, wbrana wrote:
[....]
> > The source code for Firefox and Chromium is available online. Go fetch and
> > compile. If they don't compile or execute as intended, I'm sure they would
> > like seeing a patch from you fixing the problem.
> Firefox and Chromium are large applications and can't be fixed by one user.

There are already people there. I doubt they will reject help from
others ....

> > NVIDIA, on the other hand, is yet another reason for using nouveau.
> >
> nouveau is useless garbage as most open source graphics drivers.

Then help them and send patches. Trolling does not help ....

Bernd
--
Bernd Petrovitsch Email : ***@petrovitsch.priv.at
LUGA : http://www.luga.at
wbrana
2012-08-25 11:52:42 UTC
Permalink
On 8/25/12, Bernd Petrovitsch <***@petrovitsch.priv.at> wrote:
>> Firefox and Chromium are large applications and can't be fixed by one
>> user.
>
> There are already people there. I doubt they will reject help from
> others ....
Only experienced Firefox/Chromium developer can help. Users aren't
useful as it will require major changes.

>
>> > NVIDIA, on the other hand, is yet another reason for using nouveau.
>> >
>> nouveau is useless garbage as most open source graphics drivers.
>
> Then help them and send patches. Trolling does not help ....
It is impossible that users will create patches if nouveau developers
can't create them.
Cruz Julian Bishop
2012-08-25 00:51:01 UTC
Permalink
On 25/08/12 03:05, wbrana wrote:
> On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
>> What I'd hate even more is rendering my old working hardware useless by
>> removing x86-32 support from the kernel. To reason the removal by saying
>> "Microsoft plans to do it" just makes me go bonkers...
> Your old hardware will work fine with long term kernel.
People won't want to be forced to stick with an old version of the
kernel which,
as you said, will not have any backported features.

People deserve the choice to use whatever they have, however they want.
That's the way it works. The was it has been, currently is, and always
will be.

...Unless someone at Microsoft* holds Linus hostage** in order to take over
Linux kernel development. Not that it's likely to ever happen


*Not being a troll or hurling personal insults at Microsoft - It's just
that they
currently have the majority share on the desktop (and made the original
announcement for W9)

**If this ever happens, even if it's by a terrorist group and not a company,
please don't sue me for conspiracy to kidnapping. It was just an example :)

>
>> These legacy apps will most likely be compiled for x86-32 and not x32 (an
>> argument for not removing x86-32 support on a running x86-64 kernel).
> Which legacy apps do you mean?
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to ***@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
wbrana
2012-08-25 07:46:49 UTC
Permalink
On 8/25/12, Cruz Julian Bishop <***@gmail.com> wrote:
> People won't want to be forced to stick with an old version of the
> kernel which,
> as you said, will not have any backported features.
Trash shouldn't be fully supported.
wbrana
2012-08-24 16:19:29 UTC
Permalink
On 8/24/12, Chen <***@gmail.com> wrote:
> no, u have ignore so many x86-32 boxes.I have enough reason to prove that u
> have got temperature or mental illness, or even your brain has been hit by
> somebody.LoL.
almost all x86-32 boxes will be trash in 2017, remaining boxes will
use long term tree
Alan Cox
2012-08-24 16:36:26 UTC
Permalink
> almost all x86-32 boxes will be trash in 2017, remaining boxes will
> use long term tree

People will still be manufacturing 32bit x86 processors in 2017 I'm quite
sure. You appear entirely out of touch. There are already serious
discussions going on about things like the kernel modifications needed to
make 32bit systems run past 2038.

Besides which what Linux supports is defined by what peope chose to
contribute code for. We support 32bit 680x0 machines that have been
obsolete for nigh on 20 years because someone chooses to support them.

For that matter if someone comes along with DEC-10 port and it works as
was clean without messing up the core I'm sure we'd add that too!

Alan
wbrana
2012-08-24 16:36:26 UTC
Permalink
On 8/24/12, Alan Cox <***@lxorguk.ukuu.org.uk> wrote:
>> almost all x86-32 boxes will be trash in 2017, remaining boxes will
>> use long term tree
>
> People will still be manufacturing 32bit x86 processors in 2017 I'm quite
> sure. You appear entirely out of touch. There are already serious
> discussions going on about things like the kernel modifications needed to
> make 32bit systems run past 2038.
>
You probably didn't notice: remaining boxes will use long term tree
Alan Cox
2012-08-24 16:47:12 UTC
Permalink
On Fri, 24 Aug 2012 18:36:26 +0200
wbrana <***@gmail.com> wrote:

> On 8/24/12, Alan Cox <***@lxorguk.ukuu.org.uk> wrote:
> >> almost all x86-32 boxes will be trash in 2017, remaining boxes will
> >> use long term tree
> >
> > People will still be manufacturing 32bit x86 processors in 2017 I'm quite
> > sure. You appear entirely out of touch. There are already serious
> > discussions going on about things like the kernel modifications needed to
> > make 32bit systems run past 2038.
> >
> You probably didn't notice: remaining boxes will use long term tree

That doesn't work for a variety of reasons x86 hardware is still
changing, devices are still changing. So please exit cloud cuckoo land
and go do something useful.
wbrana
2012-08-24 16:45:55 UTC
Permalink
On 8/24/12, Alan Cox <***@lxorguk.ukuu.org.uk> wrote:
> That doesn't work for a variety of reasons x86 hardware is still
> changing, devices are still changing. So please exit cloud cuckoo land
> and go do something useful.
Hardware will be discontinued if no software will support it.
Mark Lord
2012-08-29 23:03:33 UTC
Permalink
On 12-08-26 10:15 AM, wbrana wrote:
> On 8/26/12, Mark Lord <***@teksavvy.com> wrote:
>> Here are a couple of real scenarios you don't seem to have thought about.
>> A 32-bit kernel on a legacy (or even new) system in 2017 will still need
>> regular kernel updates (not "long term" un0maintained kernels)
>> in order to work with new USB devices, new 4KB+ sector hard drives,
>> newer generations of SSDs, etc..
> 12-years-old machine is trash.

There you go making assumptions again.
Who said anything about a 12-year old machine?

Much more likely is a 5-year old software installation
that gets moved to a new box.
Jeff Garzik
2012-08-29 23:42:56 UTC
Permalink
On Wed, Aug 29, 2012 at 7:03 PM, Mark Lord <***@teksavvy.com> wrote:
> On 12-08-26 10:15 AM, wbrana wrote:
>> On 8/26/12, Mark Lord <***@teksavvy.com> wrote:
>>> Here are a couple of real scenarios you don't seem to have thought about.
>>> A 32-bit kernel on a legacy (or even new) system in 2017 will still need
>>> regular kernel updates (not "long term" un0maintained kernels)
>>> in order to work with new USB devices, new 4KB+ sector hard drives,
>>> newer generations of SSDs, etc..
>> 12-years-old machine is trash.
>
> There you go making assumptions again.
> Who said anything about a 12-year old machine?
>
> Much more likely is a 5-year old software installation
> that gets moved to a new box.

Or a brand new software installation into a 32-bit virtual machine.

Jeff
Cruz Julian Bishop
2012-08-25 00:57:10 UTC
Permalink
On 25/08/12 02:36, Alan Cox wrote:
>> almost all x86-32 boxes will be trash in 2017, remaining boxes will
>> use long term tree
> People will still be manufacturing 32bit x86 processors in 2017 I'm quite
> sure. You appear entirely out of touch. There are already serious
> discussions going on about things like the kernel modifications needed to
> make 32bit systems run past 2038.
>
> Besides which what Linux supports is defined by what peope chose to
> contribute code for. We support 32bit 680x0 machines that have been
> obsolete for nigh on 20 years because someone chooses to support them.
>
> For that matter if someone comes along with DEC-10 port and it works as
> was clean without messing up the core I'm sure we'd add that too!
Is that a hint? :P
> Alan
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to ***@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
Al Viro
2012-08-23 18:08:40 UTC
Permalink
On Thu, Aug 23, 2012 at 07:22:07PM +0200, wbrana wrote:

> - wastes time of developers who can spend their time supporting X32
> instead of x86-32 or support x86-64 only as 99% of users will be able
> to run x86-64 software if x86-32 will be dropped

How much of your time is being wasted? I don't remember any patches
from you. If you mean to say that we are losing your future valuable
contributions that would happen otherwise... I think you'll find that
we will manage to muddle through, even without a visionary of such
magnitude.
wbrana
2012-08-23 18:17:27 UTC
Permalink
On 8/23/12, Al Viro <***@zeniv.linux.org.uk> wrote:
> How much of your time is being wasted? I don't remember any patches
> from you. If you mean to say that we are losing your future valuable
> contributions that would happen otherwise... I think you'll find that
> we will manage to muddle through, even without a visionary of such
> magnitude.
I'm user space developer. User space software also needs more time if
more ABIs are supported.
Pekka Enberg
2012-08-23 18:22:41 UTC
Permalink
Dear "wbrana",

On Thu, Aug 23, 2012 at 9:17 PM, wbrana <***@gmail.com> wrote:
> I'm user space developer. User space software also needs more time if
> more ABIs are supported.

I feel your pain.

As much as I appreciate your contribution here on LKML, I can't help
thinking that this discussion would be best continued on the
"linux-visionaries" mailing list.

Your pal,

Pekka
David Daney
2012-08-23 18:23:06 UTC
Permalink
On 08/23/2012 11:17 AM, wbrana wrote:
> On 8/23/12, Al Viro <***@zeniv.linux.org.uk> wrote:
>> How much of your time is being wasted? I don't remember any patches
>> from you. If you mean to say that we are losing your future valuable
>> contributions that would happen otherwise... I think you'll find that
>> we will manage to muddle through, even without a visionary of such
>> magnitude.
> I'm user space developer. User space software also needs more time if
> more ABIs are supported.

Just to keep feeding the troll:

Exactly what part of your user-space code requires special handling to
accommodate the differences between the two 32-bit x86 ABIs? Please be
specific.

David Daney
wbrana
2012-08-23 18:30:44 UTC
Permalink
On 8/23/12, David Daney <***@gmail.com> wrote:
> Exactly what part of your user-space code requires special handling to
> accommodate the differences between the two 32-bit x86 ABIs? Please be
> specific.
My code doesn't need special handling, but e.g. compilers, virtual
machines, software which use assembler need special handling.
David Daney
2012-08-23 18:33:52 UTC
Permalink
On 08/23/2012 11:30 AM, wbrana wrote:
> On 8/23/12, David Daney <***@gmail.com> wrote:
>> Exactly what part of your user-space code requires special handling to
>> accommodate the differences between the two 32-bit x86 ABIs? Please be
>> specific.
> My code doesn't need special handling,
[...]

I suspected as much. So from your point of view, this issue is of
exactly zero importance.


David Daney
wbrana
2012-08-23 18:40:19 UTC
Permalink
On 8/23/12, David Daney <***@gmail.com> wrote:
> I suspected as much. So from your point of view, this issue is of
> exactly zero importance.
I'm using software which is developed by others. As I already said
many software would be developed faster if x86-32 could be dropped.
Support for x86-32 can mean no support for X32.
Al Viro
2012-08-23 19:01:54 UTC
Permalink
On Thu, Aug 23, 2012 at 08:40:19PM +0200, wbrana wrote:
> On 8/23/12, David Daney <***@gmail.com> wrote:
> > I suspected as much. So from your point of view, this issue is of
> > exactly zero importance.

> I'm using software which is developed by others.

... which gives you no right whatsoever to demand anything. Let me
repharse what Pekka has suggested - off to the wankers' stall with you;
take it to linux-visionaries.

*plonk*
wbrana
2012-08-24 09:51:50 UTC
Permalink
On 8/23/12, Al Viro <***@zeniv.linux.org.uk> wrote:
> ... which gives you no right whatsoever to demand anything. Let me
> repharse what Pekka has suggested - off to the wankers' stall with you;
> take it to linux-visionaries.
your e-mail is off topic, try to say something relevant
Martin Nybo Andersen
2012-08-23 18:50:57 UTC
Permalink
On Thursday 23 August 2012 19:22:07 wbrana wrote:
> x86-32
> - is deprecated since Linux supports X32.
> - will slow down adoption of X32 - there won't be X32 versions of many
> software - if new ABI was added, old one should be removed
> - wastes time of developers who can spend their time supporting X32
> instead of x86-32 or support x86-64 only as 99% of users will be able
> to run x86-64 software if x86-32 will be dropped
> - wouldn't be dropped this year, but there should be plan when it will
> be dropped e.g. when Windows 9 will be released
>
> On 8/23/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> > Why?

I see that you chose to ignore the rest of my mail, which got lost on the way
to the lkml apparently because it was infected with some HTML:

----
Windows 9 will have no support for Aplha, m68k, risc, PowerPC, SPARC, ESA/390
to name a few.

They could be dropped as well, but why?
----

You could call these architectures deprecated as well. However I like the fact
that Linux doesn't try to force the users into using some specific hardware
(or software for that matter), as some companies tend to do.

BTW: This mail is written on an asus eeepc (32-bit intel thing), and my first
hands-on experience with Linux was on an amiga 1200 (mc68030). Naturally, I
love the versatility of Linux and see no reason for dropping support for old
hardware unless, it is _really_ not used anymore (token ring comes to mind).

--
Cheers,
Martin
Brian Gerst
2012-08-23 18:54:40 UTC
Permalink
On Thu, Aug 23, 2012 at 1:22 PM, wbrana <***@gmail.com> wrote:
> x86-32
> - is deprecated since Linux supports X32.
> - will slow down adoption of X32 - there won't be X32 versions of many
> software - if new ABI was added, old one should be removed

You misunderstand what the X32 ABI is. It's 64-bit code (allowing use
of the extended register set) that uses 32-bit pointers to save
memory. It has nothing to do with 32-bit kernels, and is completely
optional.

> - wastes time of developers who can spend their time supporting X32
> instead of x86-32 or support x86-64 only as 99% of users will be able
> to run x86-64 software if x86-32 will be dropped

The x86-32 arch is mature and well maintained, and shares so much in
common with x86-64, that there is little to be gained by dropping
kernel support.

> - wouldn't be dropped this year, but there should be plan when it will
> be dropped e.g. when Windows 9 will be released

Windows mostly sells with new hardware, and by the time win9 is
released all new hardware designed for it will be 64-bit capable.
Therefore it is not *profitable* for Microsoft to continue to develop
a 32-bit version. That doesn't apply to Linux. Linux is installed on
a broad range of hardware, new and old. In general, we don't drop
support for hardware unless there is nobody willing to maintain it.

--
Brian Gerst
wbrana
2012-08-23 19:03:14 UTC
Permalink
On 8/23/12, Brian Gerst <***@gmail.com> wrote:
> The x86-32 arch is mature and well maintained, and shares so much in
> common with x86-64, that there is little to be gained by dropping
> kernel support.
I would gain better chance, that NVIDIA will support X32.
Brian Gerst
2012-08-23 19:04:34 UTC
Permalink
On Thu, Aug 23, 2012 at 3:03 PM, wbrana <***@gmail.com> wrote:
> On 8/23/12, Brian Gerst <***@gmail.com> wrote:
>> The x86-32 arch is mature and well maintained, and shares so much in
>> common with x86-64, that there is little to be gained by dropping
>> kernel support.
> I would gain better chance, that NVIDIA will support X32.

Nobody here cares about closed source drivers.

--
Brian Gerst
wbrana
2012-08-23 19:08:41 UTC
Permalink
On 8/23/12, Brian Gerst <***@gmail.com> wrote:
> Nobody here cares about closed source drivers.
There are also open source software which don't support X32 like
Oracle Java, VirtualBox, Mozilla Firefox, Google Chrome.
Brian Gerst
2012-08-23 19:35:19 UTC
Permalink
On Thu, Aug 23, 2012 at 3:08 PM, wbrana <***@gmail.com> wrote:
> On 8/23/12, Brian Gerst <***@gmail.com> wrote:
>> Nobody here cares about closed source drivers.
> There are also open source software which don't support X32 like
> Oracle Java, VirtualBox, Mozilla Firefox, Google Chrome.

As I said before, X32 is an optional ABI in 64-bit kernels, for 64-bit
userspace code using 32-bit pointers to save memory. Userspace
software doesn't have to support it, it can just go full 64-bit. Any
kernel that supports X32 also supports the full 64-bit ABI.

--
Brian Gerst
wbrana
2012-08-24 09:50:16 UTC
Permalink
On 8/23/12, Brian Gerst <***@gmail.com> wrote:
> On Thu, Aug 23, 2012 at 3:08 PM, wbrana <***@gmail.com> wrote:
>> On 8/23/12, Brian Gerst <***@gmail.com> wrote:
>>> Nobody here cares about closed source drivers.
>> There are also open source software which don't support X32 like
>> Oracle Java, VirtualBox, Mozilla Firefox, Google Chrome.
>
> As I said before, X32 is an optional ABI in 64-bit kernels, for 64-bit
> userspace code using 32-bit pointers to save memory. Userspace
> software doesn't have to support it, it can just go full 64-bit. Any
> kernel that supports X32 also supports the full 64-bit ABI.
I don't understand how your e-mail is related to my e-mail.
Ronnie Collinson
2012-08-24 10:27:12 UTC
Permalink
On 8/24/12, wbrana <***@gmail.com> wrote:
> On 8/23/12, Brian Gerst <***@gmail.com> wrote:
>> On Thu, Aug 23, 2012 at 3:08 PM, wbrana <***@gmail.com> wrote:
>>> On 8/23/12, Brian Gerst <***@gmail.com> wrote:
>>>> Nobody here cares about closed source drivers.
>>> There are also open source software which don't support X32 like
>>> Oracle Java, VirtualBox, Mozilla Firefox, Google Chrome.
>>
>> As I said before, X32 is an optional ABI in 64-bit kernels, for 64-bit
>> userspace code using 32-bit pointers to save memory. Userspace
>> software doesn't have to support it, it can just go full 64-bit. Any
>> kernel that supports X32 also supports the full 64-bit ABI.
> I don't understand how your e-mail is related to my e-mail.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to ***@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>

Sort of the problem here, you dont understand what your talking about.
wbrana
2012-08-24 10:30:04 UTC
Permalink
On 8/24/12, Ronnie Collinson <***@gmail.com> wrote:
> Sort of the problem here, you dont understand what your talking about.
I understand what I'm talking about, but don't understand what Brian
Gerst is talking about.
wbrana
2012-08-24 10:36:15 UTC
Permalink
Chakra Linux will drop support for x86-32
http://chakra-linux.org/news/index.php?/archives/72-Chakra-phasing-out-i686-support.html
Ronnie Collinson
2012-08-24 10:38:40 UTC
Permalink
Hes just told you what x32 is, if you dont understand that, you cant
understand why its not a replacement for x32_64

On 8/24/12, wbrana <***@gmail.com> wrote:
> On 8/24/12, Ronnie Collinson <***@gmail.com> wrote:
>> Sort of the problem here, you dont understand what your talking about.
> I understand what I'm talking about, but don't understand what Brian
> Gerst is talking about.
>
wbrana
2012-08-24 10:42:38 UTC
Permalink
On 8/24/12, Ronnie Collinson <***@gmail.com> wrote:
> Hes just told you what x32 is, if you dont understand that, you cant
> understand why its not a replacement for x32_64
I know what is x32. x32 is replacement for x86-32, not x86-64.
Maarten Lankhorst
2012-08-24 10:49:37 UTC
Permalink
Op 24-08-12 12:42, wbrana schreef:
> On 8/24/12, Ronnie Collinson <***@gmail.com> wrote:
>> Hes just told you what x32 is, if you dont understand that, you cant
>> understand why its not a replacement for x32_64
> I know what is x32. x32 is replacement for x86-32, not x86-64.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to ***@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
Ronnie Collinson
2012-08-24 10:59:38 UTC
Permalink
No its not a replacement, its an alternative with various nuances, but
its not a direct replacement for anything
Brian Gerst
2012-08-24 11:57:31 UTC
Permalink
On Fri, Aug 24, 2012 at 6:42 AM, wbrana <***@gmail.com> wrote:
> On 8/24/12, Ronnie Collinson <***@gmail.com> wrote:
>> Hes just told you what x32 is, if you dont understand that, you cant
>> understand why its not a replacement for x32_64
> I know what is x32. x32 is replacement for x86-32, not x86-64.

Since you can't seem to do your homework:

https://sites.google.com/site/x32abi/documents/x32-abi-1.0.pdf?attredirects=0&d=1

Chapter 10
ILP32 Programming Model
"x32" is commonly used to refer to AMD64 ILP32 programming model.

10.2 Address Space
ILP32 binaries reside in the lower 32 bits of the 64-bit virtual
address space and
all addresses are 32 bits in size. They should conform to small code model or
small position independent code model (PIC) described in Section 3.5.1.

10.5 Coding Examples
Although ILP32 binaries run in the 64-bit mode...

--
Brian Gerst
wbrana
2012-08-24 09:53:37 UTC
Permalink
On 8/23/12, Brian Gerst <***@gmail.com> wrote:
> Windows mostly sells with new hardware, and by the time win9 is
> released all new hardware designed for it will be 64-bit capable.
> Therefore it is not *profitable* for Microsoft to continue to develop
> a 32-bit version. That doesn't apply to Linux. Linux is installed on
> a broad range of hardware, new and old. In general, we don't drop
> support for hardware unless there is nobody willing to maintain it.
It won't also be profitable for Linux to support trash.
Brian Gerst
2012-08-24 12:00:15 UTC
Permalink
On Fri, Aug 24, 2012 at 5:53 AM, wbrana <***@gmail.com> wrote:
> On 8/23/12, Brian Gerst <***@gmail.com> wrote:
>> Windows mostly sells with new hardware, and by the time win9 is
>> released all new hardware designed for it will be 64-bit capable.
>> Therefore it is not *profitable* for Microsoft to continue to develop
>> a 32-bit version. That doesn't apply to Linux. Linux is installed on
>> a broad range of hardware, new and old. In general, we don't drop
>> support for hardware unless there is nobody willing to maintain it.
> It won't also be profitable for Linux to support trash.

If you don't want to support it that's your opinion. Those of us who
do don't care what you think. There is no profit motive in open
source.

--
Brian Gerst
wbrana
2012-08-24 12:04:26 UTC
Permalink
On 8/24/12, Brian Gerst <***@gmail.com> wrote:
> If you don't want to support it that's your opinion. Those of us who
> do don't care what you think. There is no profit motive in open
> source.
Support needs time. Time is money.
Ondrej Zary
2012-08-24 12:14:25 UTC
Permalink
On Friday 24 August 2012, wbrana wrote:
> On 8/24/12, Brian Gerst <***@gmail.com> wrote:
> > If you don't want to support it that's your opinion. Those of us who
> > do don't care what you think. There is no profit motive in open
> > source.
>
> Support needs time. Time is money.

So stop wasting time with this trolling.

--
Ondrej Zary
wbrana
2012-08-24 12:27:39 UTC
Permalink
On 8/24/12, Ondrej Zary <***@rainbow-software.org> wrote:
> So stop wasting time with this trolling.
My and others' computers would work faster if x32 would be useful,
which isn't yet. It would save time.
I need to practise English.
H. Peter Anvin
2012-08-24 18:17:20 UTC
Permalink
On 08/23/2012 11:54 AM, Brian Gerst wrote:
>
>> - wastes time of developers who can spend their time supporting X32
>> instead of x86-32 or support x86-64 only as 99% of users will be able
>> to run x86-64 software if x86-32 will be dropped
>
> The x86-32 arch is mature and well maintained, and shares so much in
> common with x86-64, that there is little to be gained by dropping
> kernel support.
>

Speaking as one of the x86 maintainers... we are currently deciding the
cost/benefit tradeoff around removing i386 support. I don't mean
general x86-32 support, I mean i386 as opposed to i486, Pentium, and so on.

Dropping x86-32 support is decades away.

-hpa
Theodore Ts'o
2012-08-24 18:57:41 UTC
Permalink
On Fri, Aug 24, 2012 at 11:17:20AM -0700, H. Peter Anvin wrote:
>
> Speaking as one of the x86 maintainers... we are currently deciding the
> cost/benefit tradeoff around removing i386 support. I don't mean
> general x86-32 support, I mean i386 as opposed to i486, Pentium, and so on.

Random question. As I recall the Space Shuttle and the International
Space Station was only using 80386's because they have to be hardened
against radiation/cosmic rays, as well as all of the other mechnical
and thermal stresses associated with being in a spacecraft. Is there
any newer generation cpu's which are space-cerified at this point?

(Of course, I'm rather doubtful that NASA would ever be willing to use
Linux on something like the Curiosity Mars Rover, but I could imagine
Linux being used in a non-mission critcal system on the ISS....)

- Ted
Alan Cox
2012-08-24 19:13:58 UTC
Permalink
> (Of course, I'm rather doubtful that NASA would ever be willing to use
> Linux on something like the Curiosity Mars Rover, but I could imagine
> Linux being used in a non-mission critcal system on the ISS....)

GOAS, RACSI... not entirely non-mission critical stuff either.

And the ST8 project covers Linux
Theodore Ts'o
2012-08-24 19:58:56 UTC
Permalink
On Fri, Aug 24, 2012 at 08:13:58PM +0100, Alan Cox wrote:
> > (Of course, I'm rather doubtful that NASA would ever be willing to use
> > Linux on something like the Curiosity Mars Rover, but I could imagine
> > Linux being used in a non-mission critcal system on the ISS....)
>
> GOAS, RACSI... not entirely non-mission critical stuff either.
>
> And the ST8 project covers Linux

BTW, it turns out I was wrong about Linux being used on Mars.
Apparently Linux was used on the Mars Global Surveyor, as well as the
Sprit and Opportunity rovers -- and they were indeed using
space-hardened i386's from Intel.

- Ted
Dave Jones
2012-08-24 20:06:42 UTC
Permalink
On Fri, Aug 24, 2012 at 03:58:56PM -0400, Theodore Ts'o wrote:

> BTW, it turns out I was wrong about Linux being used on Mars.
> Apparently Linux was used on the Mars Global Surveyor, as well as the
> Sprit and Opportunity rovers

citation? My recollection was that they were running VxWorks.

Dave
Borislav Petkov
2012-08-24 20:25:44 UTC
Permalink
On Fri, Aug 24, 2012 at 04:06:42PM -0400, Dave Jones wrote:
> On Fri, Aug 24, 2012 at 03:58:56PM -0400, Theodore Ts'o wrote:
>
> > BTW, it turns out I was wrong about Linux being used on Mars.
> > Apparently Linux was used on the Mars Global Surveyor, as well as the
> > Sprit and Opportunity rovers
>
> citation? My recollection was that they were running VxWorks.

... on a "a radiation-hardened version of the IBM PowerPC 750." So no
x86.

http://www.pcmag.com/article2/0,2817,2408127,00.asp

--
Regards/Gruss,
Boris.
H. Peter Anvin
2012-08-24 20:47:01 UTC
Permalink
On 08/24/2012 01:25 PM, Borislav Petkov wrote:
> On Fri, Aug 24, 2012 at 04:06:42PM -0400, Dave Jones wrote:
>> On Fri, Aug 24, 2012 at 03:58:56PM -0400, Theodore Ts'o wrote:
>>
>> > BTW, it turns out I was wrong about Linux being used on Mars.
>> > Apparently Linux was used on the Mars Global Surveyor, as well as the
>> > Sprit and Opportunity rovers
>>
>> citation? My recollection was that they were running VxWorks.
>
> ... on a "a radiation-hardened version of the IBM PowerPC 750." So no
> x86.
>
> http://www.pcmag.com/article2/0,2817,2408127,00.asp
>

That is Curiosity (Mars Science Laboratory), not Spirit/Opportunity
(Mars Exploration Rover).

However, Wikipedia lists the CPU on the MERs as a 20 MHz IBM RAD6000
CPU, running VxWorks.

If you want x86, there are definitely newer things than 386:

http://www.sandia.gov/media/rhp.htm

... that was almost 14 years ago now.

-hpa
H. Peter Anvin
2012-08-24 19:38:10 UTC
Permalink
ESA has the Leon series (SPARC)...

Theodore Ts'o <***@mit.edu> wrote:

>On Fri, Aug 24, 2012 at 11:17:20AM -0700, H. Peter Anvin wrote:
>>
>> Speaking as one of the x86 maintainers... we are currently deciding
>the
>> cost/benefit tradeoff around removing i386 support. I don't mean
>> general x86-32 support, I mean i386 as opposed to i486, Pentium, and
>so on.
>
>Random question. As I recall the Space Shuttle and the International
>Space Station was only using 80386's because they have to be hardened
>against radiation/cosmic rays, as well as all of the other mechnical
>and thermal stresses associated with being in a spacecraft. Is there
>any newer generation cpu's which are space-cerified at this point?
>
>(Of course, I'm rather doubtful that NASA would ever be willing to use
>Linux on something like the Curiosity Mars Rover, but I could imagine
>Linux being used in a non-mission critcal system on the ISS....)
>
> - Ted

--
Sent from my mobile phone. Please excuse brevity and lack of formatting.
H. Peter Anvin
2012-08-24 19:40:12 UTC
Permalink
Heh... I just read about Android Nexus One phones being used as satellite controllers.

Theodore Ts'o <***@mit.edu> wrote:

>On Fri, Aug 24, 2012 at 11:17:20AM -0700, H. Peter Anvin wrote:
>>
>> Speaking as one of the x86 maintainers... we are currently deciding
>the
>> cost/benefit tradeoff around removing i386 support. I don't mean
>> general x86-32 support, I mean i386 as opposed to i486, Pentium, and
>so on.
>
>Random question. As I recall the Space Shuttle and the International
>Space Station was only using 80386's because they have to be hardened
>against radiation/cosmic rays, as well as all of the other mechnical
>and thermal stresses associated with being in a spacecraft. Is there
>any newer generation cpu's which are space-cerified at this point?
>
>(Of course, I'm rather doubtful that NASA would ever be willing to use
>Linux on something like the Curiosity Mars Rover, but I could imagine
>Linux being used in a non-mission critcal system on the ISS....)
>
> - Ted

--
Sent from my mobile phone. Please excuse brevity and lack of formatting.
Craig Milo Rogers
2012-08-24 20:18:46 UTC
Permalink
On 12.08.24, Theodore Ts'o wrote:
> Random question. As I recall the Space Shuttle and the International
> Space Station was only using 80386's because they have to be hardened
> against radiation/cosmic rays, as well as all of the other mechnical
> and thermal stresses associated with being in a spacecraft. Is there
> any newer generation cpu's which are space-cerified at this point?

The MAESTRO processor is a rad-hard-by-design variant of the Tilera
architecture, intended for space applications. Linux runs on it.

The Mongoose-V is a rad-hard MIPS R3000 processor. It can run
VxWorks, made by Wind River (a subsidiary of Intel since 2009).

http://www.synova.com/proc/mg5.html

Rad-hard Power PCs are the current space workhorse. Several variants
are available. NASA has run Linux on at least one of them in space, but I
believe that VxWorks is more the norm.

> (Of course, I'm rather doubtful that NASA would ever be willing to use
> Linux on something like the Curiosity Mars Rover, but I could imagine
> Linux being used in a non-mission critcal system on the ISS....)

Linux has been use on scientific equipment sent to the ISS.
Not, I think, on the avionics.

Curiosity runs VxWorks, as do the SpaceX Falcon 9 and Dragon.

Craig Milo Rogers
Sam Ravnborg
2012-08-24 20:49:58 UTC
Permalink
On Fri, Aug 24, 2012 at 02:57:41PM -0400, Theodore Ts'o wrote:
> On Fri, Aug 24, 2012 at 11:17:20AM -0700, H. Peter Anvin wrote:
> >
> > Speaking as one of the x86 maintainers... we are currently deciding the
> > cost/benefit tradeoff around removing i386 support. I don't mean
> > general x86-32 support, I mean i386 as opposed to i486, Pentium, and so on.
>
> Random question. As I recall the Space Shuttle and the International
> Space Station was only using 80386's because they have to be hardened
> against radiation/cosmic rays, as well as all of the other mechnical
> and thermal stresses associated with being in a spacecraft. Is there
> any newer generation cpu's which are space-cerified at this point?

I do not know if you consider the LEON variants of SPARC 32 bit as a new
generation. But they are actively developed and used in space missions.
At least to my best knowledge.

Sam
d***@lang.hm
2012-08-30 01:14:02 UTC
Permalink
On Fri, 24 Aug 2012, Theodore Ts'o wrote:

> On Fri, Aug 24, 2012 at 11:17:20AM -0700, H. Peter Anvin wrote:
>>
>> Speaking as one of the x86 maintainers... we are currently deciding the
>> cost/benefit tradeoff around removing i386 support. I don't mean
>> general x86-32 support, I mean i386 as opposed to i486, Pentium, and so on.
>
> Random question. As I recall the Space Shuttle and the International
> Space Station was only using 80386's because they have to be hardened
> against radiation/cosmic rays, as well as all of the other mechnical
> and thermal stresses associated with being in a spacecraft. Is there
> any newer generation cpu's which are space-cerified at this point?
>
> (Of course, I'm rather doubtful that NASA would ever be willing to use
> Linux on something like the Curiosity Mars Rover, but I could imagine
> Linux being used in a non-mission critcal system on the ISS....)

I've heard that there are radiation hardened versions of the 80486 (I
could be wrong, it's not something that I've ever needed to investigate)

David Lang
Raymond Jennings
2012-08-24 16:24:30 UTC
Permalink
On Thu, 2012-08-23 at 12:41 +0200, wbrana wrote:
> Microsoft will drop support for x86-32 in Windows 9.
> Linux could do same.
> http://www.networkworld.com/community/blog/windows-9-details-are-already-emerging
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to ***@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

I use an x86-32 system myself.

So do many other people.

Besides, it's not really your call to decide if x86-32 is obsolete.

If it's anyone's call, it's for companies like AMD and Intel that
actually make the chips. Microsoft doesn't make x86 chips, so their
opinion on x86-32's viability is none of our concern.

Similiarly, if I were a marketing director for pepsi, I wouldn't listen
to anything that Coca cola has to say about what flavors of soda to
make. A problem with the liquid CO2 company I buy my fizz from however
WOULD get my attention.
Martin Nybo Andersen
2012-08-24 18:09:46 UTC
Permalink
On Friday 24 August 2012 19:05:53 wbrana wrote:
> On 8/24/12, Martin Nybo Andersen <***@tweek.dk> wrote:
> > What I'd hate even more is rendering my old working hardware useless by
> > removing x86-32 support from the kernel. To reason the removal by saying
> > "Microsoft plans to do it" just makes me go bonkers...
>
> Your old hardware will work fine with long term kernel.

That's right, but new hardware, that I wish to use with the old machines might
not because of no backporting of new drivers. Same goes for new software
utilising newer kernel features.

> > These legacy apps will most likely be compiled for x86-32 and not x32 (an
> > argument for not removing x86-32 support on a running x86-64 kernel).
>
> Which legacy apps do you mean?

Those mentioned by Chris Friesen, whose arguments you apparently ignored.

Going back to your original arguments:
> x86-32
> - is deprecated since Linux supports X32.
No. X32 is merely yet another ABI supported by Linux.

> - will slow down adoption of X32
Perhaps. But that would rather be because of low benefits offered by x32 (not
being able to run on legacy hardware is not a benefit (and not its
intension)).

> - there won't be X32 versions of many software
You are allowed to compile most of the software running on Linux yourself. If
you want a binary to use the x32 ABI, go compile.

> - if new ABI was added, old one should be removed
No. If kernel.org runs out of diskspace, I'd rather sponsor some new disks.
And, x32 is an ABI for the x86-64 architecture, while x86-32 is an
architecture in itself.

> - wastes time of developers who can spend their time supporting X32
> instead of x86-32 or support x86-64 only as 99% of users will be able
> to run x86-64 software if x86-32 will be dropped
No. If, for instance, an m68k maintainer/developer stops maintaining m68k
support, nobody is telling him to continue his works on, say, sparc or
whatever hyped architecture.

In other words: It costs exactly *nothing* for us to have x86-32 support. What
is does cost, though, is the maintainers/developers spare time and goodwill.
Something we all should appreciate.

> - wouldn't be dropped this year, but there should be plan when it will
> be dropped e.g. when Windows 9 will be released
No. That plan will come automagically when x86-32 is not used anymore and when
somebody works on a patch to remove x86-32 support.

--
Cheers,
Martin
Raymond Jennings
2012-08-24 19:32:39 UTC
Permalink
Some useless troll said:
> nouveau is useless garbage as most open source graphics drivers.

Coming to an open source mailing list like LKML just to bitch about open
source being garbage? Come on...at least entertain us with better
subtlety.

I'm ready to ignore this guy, how about everyone else?

*plonk*

Ah, much better.
Boszormenyi Zoltan
2012-08-28 08:24:28 UTC
Permalink
Dear wbrana,

this would have been the perfect April 1st joke along the lines of
removing support for *all* CPU architectures and adding support
for the one true virtual CPU, the Turing machine.

Now you spoiled it, shame on you! :-D

Best regards,
Zolt=E1n B=F6sz=F6rm=E9nyi
Free Email Service
2012-08-30 08:46:30 UTC
Permalink
Or a brand new software installation into a 32-bit virtual machine.
I\'m posting replies at my blog http://kernel64only.blogspot.com/
Continue reading on narkive:
Loading...