There is a HYP enabled u-boot floating around.
http://www.virtualopensystems.com/media/chromebook/nv_u-boot-snow.kpart
So far I've been unable to get it to boot Fedora, but others are welcome to try.
Here is the git for the above: https://github.com/virtualopensystems/u-boot
They also have some documentation: https://lists.cs.columbia.edu/pipermail/kvmarm/2013-July/006481.html
Regards,
-Jon Disnard fas: parasense irc: masta
On Wed, Aug 7, 2013 at 10:51 AM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Aug 7, 2013 at 4:47 PM, Adam Goode adam@spicenitz.org wrote:
Was there ever a resolution to this? I am happy to file a bug in the chromium tracker to get this looked at, if needed.
None that I'm aware of but if you could file a bug in the chromium tracker that would be fab.
Peter
Adam
On Fri, Jul 5, 2013 at 12:16 PM, Richard W.M. Jones rjones@redhat.com wrote:
On Fri, Jul 05, 2013 at 11:30:01AM -0400, Jon Masters wrote:
On 07/05/2013 05:07 AM, Richard W.M. Jones wrote:
On Wed, Jul 03, 2013 at 11:59:28AM -0500, Jon wrote:
I'm pleased to announce the availability of Fedora 19 for the 2012 Samsung Chromebook featuring ARM Exynos dual core A15 processor.
Sorry to slightly hijack this thread. I will try your remix later.
Reading the comments on https://lwn.net/Articles/557132/#Comments it seems as if the news on KVM on the Chromebook is not good. It doesn't boot into HYP mode, and there's no way to make it boot into HYP mode, so KVM won't be supported. Is that right?
That's roughly what I'd expect to be the case. There might be a signed U-Boot someone has hacked that does enable HYP mode, but otherwise I suspect you're out of luck. I'll ask around during Linaro Connect.
I asked about this on #kvm-arm earlier today and got this long reply:
11:58 < rwmjones> I'm reading a comment here: 11:58 < rwmjones> https://lwn.net/Articles/557561/ 11:58 < rwmjones> which suggests that KVM on the Samsung Chromebook 2012 (ARM A15 version) isn't possible because the bootloader doesn't boot into HYP mode 11:58 < rwmjones> is this true? if so is there a way around it? 11:59 < pm215> IIRC the bootloader gets control in secure-SVC 11:59 < pm215> it is from there possible to get to NS-HYP 11:59 < pm215> it's just that the stock bootloader doesn't do this before booting the kernel 11:59 < rwmjones> so what's involved in making it work? 11:59 < pm215> somebody needs to write some code and get it into the bootloader 12:00 < rwmjones> ok, and the bootloader can be replaced (next comment down suggests this requires soldering)? 12:01 < pm215> I believe this to be true, though I don't have a chromebook 12:01 < pm215> I think you get the google bootloader to chain boot some other bootloader which you do have control of, and then that can actually boot your os 12:02 < suihkulokki> or maybe we could just prepend some code in front of the kernel zimage that switches to HYP mode? 12:02 < pm215> nope 12:02 < suihkulokki> damn 12:02 < pm215> we spent quite a long time being very firm that the ABI here is "bootloader's job to get this right" 12:03 < pm215> there are some u-boot patches currently going through code review to do the go-to-hyp-mode thing properly for arndale 12:03 < pm215> hopefully if they get upstream it will be more straightforward to say "ok, I have $other-board and it needs to do this too" 12:05 < apritzel> which would require that the Chromebook u-boot support is upstream as well 12:05 < apritzel> AFAIK this is not the case currently
[There's more of this, but that seems to cover the main points]
Reading around this, it does seem as if it's possible to get from secure SVC to HYP (although not easy).
Rich.
-- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://people.redhat.com/~rjones/virt-top _______________________________________________ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
--
-Jon
I'll try this out.
Enabling HYP seem conceptually easy.
From what I gather it involves initializing the cores one at a time,
then touching 16 registers. (something like that)
The nv-u-boot does not seem to support saveenv, so might be worthwhile to rebuild with better default env. (known issue)
If somebody can get this working I'll be happy to integrate with the f19 remix for chromebook. My experience so far was that the HYP enabled u-boot does not work. Meaning the bootm of the FIT image fails, and also the case of legacy uImages also fail to bootm.
regards, -Jon Disnard
On Sat, Aug 10, 2013 at 10:12 PM, Jonathan Masters jcm@redhat.com wrote:
I'll try this out.
-- Sent from my iPad
On Aug 7, 2013, at 12:02, Jon jdisnard@gmail.com wrote:
There is a HYP enabled u-boot floating around.
http://www.virtualopensystems.com/media/chromebook/nv_u-boot-snow.kpart
So far I've been unable to get it to boot Fedora, but others are welcome to try.
Here is the git for the above: https://github.com/virtualopensystems/u-boot
They also have some documentation: https://lists.cs.columbia.edu/pipermail/kvmarm/2013-July/006481.html
Regards,
-Jon Disnard fas: parasense irc: masta
On Wed, Aug 7, 2013 at 10:51 AM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Aug 7, 2013 at 4:47 PM, Adam Goode adam@spicenitz.org wrote:
Was there ever a resolution to this? I am happy to file a bug in the chromium tracker to get this looked at, if needed.
None that I'm aware of but if you could file a bug in the chromium tracker that would be fab.
Peter
Adam
On Fri, Jul 5, 2013 at 12:16 PM, Richard W.M. Jones rjones@redhat.com wrote:
On Fri, Jul 05, 2013 at 11:30:01AM -0400, Jon Masters wrote:
On 07/05/2013 05:07 AM, Richard W.M. Jones wrote: > On Wed, Jul 03, 2013 at 11:59:28AM -0500, Jon wrote: >> I'm pleased to announce the availability of Fedora 19 for the 2012 Samsung >> Chromebook featuring ARM Exynos dual core A15 processor. > > Sorry to slightly hijack this thread. I will try your remix later. > > Reading the comments on https://lwn.net/Articles/557132/#Comments > it seems as if the news on KVM on the Chromebook is not good. It > doesn't boot into HYP mode, and there's no way to make it boot into > HYP mode, so KVM won't be supported. Is that right?
That's roughly what I'd expect to be the case. There might be a signed U-Boot someone has hacked that does enable HYP mode, but otherwise I suspect you're out of luck. I'll ask around during Linaro Connect.
I asked about this on #kvm-arm earlier today and got this long reply:
11:58 < rwmjones> I'm reading a comment here: 11:58 < rwmjones> https://lwn.net/Articles/557561/ 11:58 < rwmjones> which suggests that KVM on the Samsung Chromebook 2012 (ARM A15 version) isn't possible because the bootloader doesn't boot into HYP mode 11:58 < rwmjones> is this true? if so is there a way around it? 11:59 < pm215> IIRC the bootloader gets control in secure-SVC 11:59 < pm215> it is from there possible to get to NS-HYP 11:59 < pm215> it's just that the stock bootloader doesn't do this before booting the kernel 11:59 < rwmjones> so what's involved in making it work? 11:59 < pm215> somebody needs to write some code and get it into the bootloader 12:00 < rwmjones> ok, and the bootloader can be replaced (next comment down suggests this requires soldering)? 12:01 < pm215> I believe this to be true, though I don't have a chromebook 12:01 < pm215> I think you get the google bootloader to chain boot some other bootloader which you do have control of, and then that can actually boot your os 12:02 < suihkulokki> or maybe we could just prepend some code in front of the kernel zimage that switches to HYP mode? 12:02 < pm215> nope 12:02 < suihkulokki> damn 12:02 < pm215> we spent quite a long time being very firm that the ABI here is "bootloader's job to get this right" 12:03 < pm215> there are some u-boot patches currently going through code review to do the go-to-hyp-mode thing properly for arndale 12:03 < pm215> hopefully if they get upstream it will be more straightforward to say "ok, I have $other-board and it needs to do this too" 12:05 < apritzel> which would require that the Chromebook u-boot support is upstream as well 12:05 < apritzel> AFAIK this is not the case currently
[There's more of this, but that seems to cover the main points]
Reading around this, it does seem as if it's possible to get from secure SVC to HYP (although not easy).
Rich.
-- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://people.redhat.com/~rjones/virt-top _______________________________________________ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
--
-Jon
--
-Jon _______________________________________________ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
On 08/10/2013 11:23 PM, Jon wrote:
Enabling HYP seem conceptually easy. From what I gather it involves initializing the cores one at a time, then touching 16 registers. (something like that)
The nv-u-boot does not seem to support saveenv, so might be worthwhile to rebuild with better default env. (known issue)
If somebody can get this working I'll be happy to integrate with the f19 remix for chromebook. My experience so far was that the HYP enabled u-boot does not work. Meaning the bootm of the FIT image fails, and also the case of legacy uImages also fail to bootm.
regards, -Jon Disnard
Hi Jon,
I don't know how you have things set up for uboot and the kernel. One of the problems that I found when setting up to use nv-uboot on the sd card is the uboot is looking for partition 12 (0xc) for the uboot/uboot.scr.uimg, which probably doesn't exist on the card. I found after careful reading of the following URL that I needed to adjust $script_part to point to the proper partition EFI-SYSTEM:
http://www.chromium.org/chromium-os/u-boot-porting-guide/using-nv-u-boot-on-...
That should at least allow you to set the variable more easily even if the saveenv isn't working on uboot.
There are also two different versions of nv_uboot for chrome book. From what I have read upstream kernels need to use the one that supports the simple frame buffer and that the support for the simple framebuffer needs to be enabled in the kernel.
-Will
On Sat, Aug 10, 2013 at 10:12 PM, Jonathan Masters jcm@redhat.com wrote:
I'll try this out.
-- Sent from my iPad
On Aug 7, 2013, at 12:02, Jon jdisnard@gmail.com wrote:
There is a HYP enabled u-boot floating around.
http://www.virtualopensystems.com/media/chromebook/nv_u-boot-snow.kpart
So far I've been unable to get it to boot Fedora, but others are welcome to try.
Here is the git for the above: https://github.com/virtualopensystems/u-boot
They also have some documentation: https://lists.cs.columbia.edu/pipermail/kvmarm/2013-July/006481.html
Regards,
-Jon Disnard fas: parasense irc: masta
On Wed, Aug 7, 2013 at 10:51 AM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Aug 7, 2013 at 4:47 PM, Adam Goode adam@spicenitz.org wrote:
Was there ever a resolution to this? I am happy to file a bug in the chromium tracker to get this looked at, if needed.
None that I'm aware of but if you could file a bug in the chromium tracker that would be fab.
Peter
Adam
On Fri, Jul 5, 2013 at 12:16 PM, Richard W.M. Jones rjones@redhat.com wrote:
On Fri, Jul 05, 2013 at 11:30:01AM -0400, Jon Masters wrote: > On 07/05/2013 05:07 AM, Richard W.M. Jones wrote: >> On Wed, Jul 03, 2013 at 11:59:28AM -0500, Jon wrote: >>> I'm pleased to announce the availability of Fedora 19 for the 2012 Samsung >>> Chromebook featuring ARM Exynos dual core A15 processor. >> >> Sorry to slightly hijack this thread. I will try your remix later. >> >> Reading the comments on https://lwn.net/Articles/557132/#Comments >> it seems as if the news on KVM on the Chromebook is not good. It >> doesn't boot into HYP mode, and there's no way to make it boot into >> HYP mode, so KVM won't be supported. Is that right? > > That's roughly what I'd expect to be the case. There might be a signed > U-Boot someone has hacked that does enable HYP mode, but otherwise I > suspect you're out of luck. I'll ask around during Linaro Connect.
I asked about this on #kvm-arm earlier today and got this long reply:
11:58 < rwmjones> I'm reading a comment here: 11:58 < rwmjones> https://lwn.net/Articles/557561/ 11:58 < rwmjones> which suggests that KVM on the Samsung Chromebook 2012 (ARM A15 version) isn't possible because the bootloader doesn't boot into HYP mode 11:58 < rwmjones> is this true? if so is there a way around it? 11:59 < pm215> IIRC the bootloader gets control in secure-SVC 11:59 < pm215> it is from there possible to get to NS-HYP 11:59 < pm215> it's just that the stock bootloader doesn't do this before booting the kernel 11:59 < rwmjones> so what's involved in making it work? 11:59 < pm215> somebody needs to write some code and get it into the bootloader 12:00 < rwmjones> ok, and the bootloader can be replaced (next comment down suggests this requires soldering)? 12:01 < pm215> I believe this to be true, though I don't have a chromebook 12:01 < pm215> I think you get the google bootloader to chain boot some other bootloader which you do have control of, and then that can actually boot your os 12:02 < suihkulokki> or maybe we could just prepend some code in front of the kernel zimage that switches to HYP mode? 12:02 < pm215> nope 12:02 < suihkulokki> damn 12:02 < pm215> we spent quite a long time being very firm that the ABI here is "bootloader's job to get this right" 12:03 < pm215> there are some u-boot patches currently going through code review to do the go-to-hyp-mode thing properly for arndale 12:03 < pm215> hopefully if they get upstream it will be more straightforward to say "ok, I have $other-board and it needs to do this too" 12:05 < apritzel> which would require that the Chromebook u-boot support is upstream as well 12:05 < apritzel> AFAIK this is not the case currently
[There's more of this, but that seems to cover the main points]
Reading around this, it does seem as if it's possible to get from secure SVC to HYP (although not easy).
Rich.
-- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://people.redhat.com/~rjones/virt-top _______________________________________________ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
--
-Jon
--
-Jon _______________________________________________ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Hi Will
The recently released[1] HYP enabled nv-u-boot does not allow saveenv.
We can use setenv to make changes, but they are not persistent. It might be possible to import or source the ENV from boot.scr or uEnv.txt, but haven't tried.
Regardless I was unable to boot Fedora using that firmware. I tried booting legacy uImages+uInitrd, and FIT image. Both methods did not work for me.
At some point I may try to follow the instruction to boot whatever other Linux distro is described in the link [1]. I was surprised that Fedora was not booting, and hoping I made some kind of silly mistake.
If anybody is able to get it working please let em know.
[1] https://lists.cs.columbia.edu/pipermail/kvmarm/2013-July/006481.html
On Mon, Aug 12, 2013 at 3:27 PM, William Cohen wcohen@redhat.com wrote:
On 08/10/2013 11:23 PM, Jon wrote:
Enabling HYP seem conceptually easy. From what I gather it involves initializing the cores one at a time, then touching 16 registers. (something like that)
The nv-u-boot does not seem to support saveenv, so might be worthwhile to rebuild with better default env. (known issue)
If somebody can get this working I'll be happy to integrate with the f19 remix for chromebook. My experience so far was that the HYP enabled u-boot does not work. Meaning the bootm of the FIT image fails, and also the case of legacy uImages also fail to bootm.
regards, -Jon Disnard
Hi Jon,
I don't know how you have things set up for uboot and the kernel. One of the problems that I found when setting up to use nv-uboot on the sd card is the uboot is looking for partition 12 (0xc) for the uboot/uboot.scr.uimg, which probably doesn't exist on the card. I found after careful reading of the following URL that I needed to adjust $script_part to point to the proper partition EFI-SYSTEM:
http://www.chromium.org/chromium-os/u-boot-porting-guide/using-nv-u-boot-on-...
That should at least allow you to set the variable more easily even if the saveenv isn't working on uboot.
There are also two different versions of nv_uboot for chrome book. From what I have read upstream kernels need to use the one that supports the simple frame buffer and that the support for the simple framebuffer needs to be enabled in the kernel.
-Will
On Sat, Aug 10, 2013 at 10:12 PM, Jonathan Masters jcm@redhat.com wrote:
I'll try this out.
-- Sent from my iPad
On Aug 7, 2013, at 12:02, Jon jdisnard@gmail.com wrote:
There is a HYP enabled u-boot floating around.
http://www.virtualopensystems.com/media/chromebook/nv_u-boot-snow.kpart
So far I've been unable to get it to boot Fedora, but others are welcome to try.
Here is the git for the above: https://github.com/virtualopensystems/u-boot
They also have some documentation: https://lists.cs.columbia.edu/pipermail/kvmarm/2013-July/006481.html
Regards,
-Jon Disnard fas: parasense irc: masta
On Wed, Aug 7, 2013 at 10:51 AM, Peter Robinson pbrobinson@gmail.com wrote:
On Wed, Aug 7, 2013 at 4:47 PM, Adam Goode adam@spicenitz.org wrote:
Was there ever a resolution to this? I am happy to file a bug in the chromium tracker to get this looked at, if needed.
None that I'm aware of but if you could file a bug in the chromium tracker that would be fab.
Peter
Adam
On Fri, Jul 5, 2013 at 12:16 PM, Richard W.M. Jones rjones@redhat.com wrote: > On Fri, Jul 05, 2013 at 11:30:01AM -0400, Jon Masters wrote: >> On 07/05/2013 05:07 AM, Richard W.M. Jones wrote: >>> On Wed, Jul 03, 2013 at 11:59:28AM -0500, Jon wrote: >>>> I'm pleased to announce the availability of Fedora 19 for the 2012 Samsung >>>> Chromebook featuring ARM Exynos dual core A15 processor. >>> >>> Sorry to slightly hijack this thread. I will try your remix later. >>> >>> Reading the comments on https://lwn.net/Articles/557132/#Comments >>> it seems as if the news on KVM on the Chromebook is not good. It >>> doesn't boot into HYP mode, and there's no way to make it boot into >>> HYP mode, so KVM won't be supported. Is that right? >> >> That's roughly what I'd expect to be the case. There might be a signed >> U-Boot someone has hacked that does enable HYP mode, but otherwise I >> suspect you're out of luck. I'll ask around during Linaro Connect. > > I asked about this on #kvm-arm earlier today and got this long reply: > > 11:58 < rwmjones> I'm reading a comment here: > 11:58 < rwmjones> https://lwn.net/Articles/557561/ > 11:58 < rwmjones> which suggests that KVM on the Samsung Chromebook 2012 (ARM A15 version) isn't possible because the > bootloader doesn't boot into HYP mode > 11:58 < rwmjones> is this true? if so is there a way around it? > 11:59 < pm215> IIRC the bootloader gets control in secure-SVC > 11:59 < pm215> it is from there possible to get to NS-HYP > 11:59 < pm215> it's just that the stock bootloader doesn't do this before booting the kernel > 11:59 < rwmjones> so what's involved in making it work? > 11:59 < pm215> somebody needs to write some code and get it into the bootloader > 12:00 < rwmjones> ok, and the bootloader can be replaced (next comment down suggests this requires soldering)? > 12:01 < pm215> I believe this to be true, though I don't have a chromebook > 12:01 < pm215> I think you get the google bootloader to chain boot some other bootloader which you do have control of, and then > that can actually boot your os > 12:02 < suihkulokki> or maybe we could just prepend some code in front of the kernel zimage that switches to HYP mode? > 12:02 < pm215> nope > 12:02 < suihkulokki> damn > 12:02 < pm215> we spent quite a long time being very firm that the ABI here is "bootloader's job to get this right" > 12:03 < pm215> there are some u-boot patches currently going through code review to do the go-to-hyp-mode thing properly for > arndale > 12:03 < pm215> hopefully if they get upstream it will be more straightforward to say "ok, I have $other-board and it needs to > do this too" > 12:05 < apritzel> which would require that the Chromebook u-boot support is upstream as well > 12:05 < apritzel> AFAIK this is not the case currently > > [There's more of this, but that seems to cover the main points] > > Reading around this, it does seem as if it's possible to get from > secure SVC to HYP (although not easy). > > Rich. > > -- > Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones > virt-top is 'top' for virtual machines. Tiny program with many > powerful monitoring features, net stats, disk stats, logging, etc. > http://people.redhat.com/~rjones/virt-top > _______________________________________________ > arm mailing list > arm@lists.fedoraproject.org > https://admin.fedoraproject.org/mailman/listinfo/arm _______________________________________________ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
--
-Jon
--
-Jon _______________________________________________ arm mailing list arm@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/arm
Sorry to hijack this thread a second time, but what ARM hardware is currently recommended where KVM works in a reasonably "out of the box" way?
Rich.
On 08/21/2013 12:22 PM, Richard W.M. Jones wrote:
Sorry to hijack this thread a second time, but what ARM hardware is currently recommended where KVM works in a reasonably "out of the box" way?
KVM requires hardware virtualization support which comes with Cortex-A15. This is a relatively new chip without a good full upstream implementation being generally available. The best bets right now are:
Arndale board: Hard to acquire, but once you have one there is an F19 remix for it available at:
http://fedoraproject.org/wiki/Architectures/ARM/F19/Remixes
Qemu: As of 1.6 there is Calxeda Midway (Cortex-A15) support in qemu-system-arm. Given a good kernel this will in theory allow anybody to boot a KVM-capable kernel. Not sure this has been demonstrated successfully yet though. Paul Whalen can probably comment.
Foundation model: The aarch64 foundation model, I believe, implements the kvm virt instructions, but that's aarch64.
More hardware options are just around the bend. Red Hatters should contact me off list as the arm team may be able to provide Arndales or other options. Cheers,
在 2013-8-22 AM8:43，"Brendan Conoboy" blc@redhat.com写道：
In fact arndale board is also good for Xen, which has official announced the support of this board.
On Wed, 2013-08-21 at 17:43 -0700, Brendan Conoboy wrote:
On 08/21/2013 12:22 PM, Richard W.M. Jones wrote:
Sorry to hijack this thread a second time, but what ARM hardware is currently recommended where KVM works in a reasonably "out of the box" way?
KVM requires hardware virtualization support which comes with Cortex-A15.
And here's where ARM's wonky naming schemes come in. The Cortex-A7 and Cortex-A12, which came after the Cortex-A15 but are lower-power parts, should also be able to do virtualization (though I haven't tried it). This means that devices such as the Cubieboard 2 should allow virt work, and they're easier to get than the Arndales. (I repeat, I haven't tried this yet, your mileage may vary, etc).
-Chris
-
Brendan Conoboy -
Chris Tyler -
Christopher Meng -
Jon -
Jon Masters -
Richard W.M. Jones -
William Cohen