Any plan for armv5te support? It should compatible to Ubuntu of 30MB that seems possible [http://lists.linaro.org/pipermail/linaro-dev/2011-January/002109.html] . There are a lot of arm9 boards in the world. We have to think that arm9 is the past?
Thanks for any reply, Recalcati
Hi Rafaele,
On 01/29/2011 10:47 AM, Raffaele Recalcati wrote:
Any plan for armv5te support? It should compatible to Ubuntu of 30MB that seems possible [http://lists.linaro.org/pipermail/linaro-dev/2011-January/002109.html] . There are a lot of arm9 boards in the world. We have to think that arm9 is the past?
It's not that arm9 is the past. There are lots of products out them with such cpus.
It's just that Linaro is funded by some of the biggest chip vendors to improve Linux on the latest and future generations of ARM chips. That's an investment for them, and this will make the future products based on their chips more competitive.
Linaro has finite resources and budget. We can't work on everything, unfortunately.
Of course, we will make sure that our contributions to mainline projects don't break existing platforms. Anyway, you can still count on other community projects to continue to support all the other ARM cores. In particular, Debian is a great distro for armv5 cpus.
Cheers,
Michael.
Hi Michael,
It's just that Linaro is funded by some of the biggest chip vendors to improve Linux on the latest and future generations of ARM chips. That's an investment for them, and this will make the future products based on their chips more competitive.
Linaro has finite resources and budget. We can't work on everything, unfortunately.
I'm very curious about your job. I'm trying to understand it, but a distro compatibility for every arm cpu is a great thing, making possible, with some change in pre-bootloader, bootloader, kernel, to have the custom board derived from evm compatible to choosen distro as well.
Of course, we will make sure that our contributions to mainline projects don't break existing platforms. Anyway, you can still count on other community projects to continue to support all the other ARM cores. In particular, Debian is a great distro for armv5 cpus.
Anyway, I'll go on with Openembedded for armv5.
Thx, Raffaele
On 01/29/2011 05:40 PM, Raffaele Recalcati wrote:
I'm very curious about your job. I'm trying to understand it, but a distro compatibility for every arm cpu is a great thing, making possible, with some change in pre-bootloader, bootloader, kernel, to have the custom board derived from evm compatible to choosen distro as well.
That's right that this is very convenient for distribution developers. On the other hand, in a product with a recent ARM core, it would be a pity not to take advantage of what the latest Cortex cores can do, and miss opportunities to increase performance.
See the benchmarks on https://wiki.linaro.org/WorkingGroups/ToolChain/Benchmarks/InitialMemset, for example.
I guess this discussion has already happened countless times, and it's hard to have the last word anyway ;-)
Of course, we will make sure that our contributions to mainline projects don't break existing platforms. Anyway, you can still count on other community projects to continue to support all the other ARM cores. In particular, Debian is a great distro for armv5 cpus.
Anyway, I'll go on with Openembedded for armv5.
And if you switch to a Cortex A8 or A9 cpu one day, you should be able to rebuild your rootfs from the same sources, taking advantage of some of Linaro's contributions to mainline gcc...
Cheers,
Michael.
"Raffaele Recalcati" lamiaposta71@gmail.com napisał:
Anyway, I'll go on with Openembedded for armv5.
Good choice ;) You will get Linaro gcc 4.5 patches in OE by default.
On 01/29/11 09:47, Somebody in the thread at some point said:
Any plan for armv5te support? It should compatible to Ubuntu of 30MB that seems possible [http://lists.linaro.org/pipermail/linaro-dev/2011-January/002109.html] . There are a lot of arm9 boards in the world. We have to think that arm9 is the past?
Thanks for any reply,
Hey,
Offlist so as not to make myself unpopular ^^
You might want to look at Fedora ARM, it target armv5te.
http://fedoraproject.org/wiki/Architectures/ARM
I used it with great success.
-Andy
On Sun, Jan 30, 2011 at 09:50:10AM +0000, Andy Green wrote:
Offlist so as not to make myself unpopular ^^
You might want to look at Fedora ARM, it target armv5te.
Well, point taken though -- is anyboy talking to Fedora to see if they'd be interested in taking some of our components as their ARM upstreams?
On Wed, Feb 9, 2011 at 6:38 AM, Christian Robottom Reis kiko@linaro.org wrote:
On Sun, Jan 30, 2011 at 09:50:10AM +0000, Andy Green wrote:
Offlist so as not to make myself unpopular ^^
You might want to look at Fedora ARM, it target armv5te.
Well, point taken though -- is anyboy talking to Fedora to see if they'd be interested in taking some of our components as their ARM upstreams?
There's some interesting discussion on the list archives, like 'should we switch to ARMv7+hardfp and drop armv5te', and 'Hey, our QEMU uses versatile which is old and broken, what should we do'.
-- Michael
On 02/14/2011 03:35 AM, Somebody in the thread at some point said:
On Wed, Feb 9, 2011 at 6:38 AM, Christian Robottom Reiskiko@linaro.org wrote:
On Sun, Jan 30, 2011 at 09:50:10AM +0000, Andy Green wrote:
Offlist so as not to make myself unpopular ^^
You might want to look at Fedora ARM, it target armv5te.
Well, point taken though -- is anyboy talking to Fedora to see if they'd be interested in taking some of our components as their ARM upstreams?
There's some interesting discussion on the list archives, like 'should we switch to ARMv7+hardfp and drop armv5te', and 'Hey, our QEMU uses versatile which is old and broken, what should we do'.
I think the consensus of that was that Fedora ARM need to stick with v5 (I was certainly arguing that anyway) but are curious about if they get any general advantage with v7; in any event they see they need to support neon. So I think sooner or later they benefit from Linaro's "living in the future" work one way or another.
Hopefully a similar setup to "mainly i386 with handpicked i686 exceptions" that Fedora used on 32-bit x86 will be enough to unify the two worlds, because weak ARM9 v5 and ARM11(36) are a genuine market sector that probably won't go away for a long while.
-Andy
On Tue, Feb 8, 2011 at 6:38 PM, Christian Robottom Reis kiko@linaro.org wrote:
On Sun, Jan 30, 2011 at 09:50:10AM +0000, Andy Green wrote:
Offlist so as not to make myself unpopular ^^
You might want to look at Fedora ARM, it target armv5te.
Well, point taken though -- is anyboy talking to Fedora to see if they'd be interested in taking some of our components as their ARM upstreams?
I'm an active Fedora contributor (yes black sheep in this Ubuntu-infested Linaro world...) but haven't used the ARM port much.
I believe maybe Lennert can drop us a quick status on Fedora ARM? He was instrumental in making it happen anyways.
Yours, Linus Walleij
On 02/14/2011 10:06 AM, Somebody in the thread at some point said:
On Tue, Feb 8, 2011 at 6:38 PM, Christian Robottom Reiskiko@linaro.org wrote:
On Sun, Jan 30, 2011 at 09:50:10AM +0000, Andy Green wrote:
Offlist so as not to make myself unpopular ^^
You might want to look at Fedora ARM, it target armv5te.
Well, point taken though -- is anyboy talking to Fedora to see if they'd be interested in taking some of our components as their ARM upstreams?
I'm an active Fedora contributor (yes black sheep in this Ubuntu-infested Linaro world...) but haven't used the ARM port much.
Dude ^^ FWIW I used the ARM port for several years, Fedora since it started.
RHAT don't put much resources into it so it bumps along and a Hero like Lennert / Marvell rises now and again to keep it going. Right now they just look delivery of a dozen Pandas for their build farm so they are rising up again.
-Andy
-
Andy Green -
Christian Robottom Reis -
Linus Walleij -
Marcin Juszkiewicz -
Michael Hope -
Michael Opdenacker -
Raffaele Recalcati