[PATCH] arch-powerpc64.inc: disable the use of qemu usermode on ppc64

[PATCH] arch-powerpc64.inc: disable the use of qemu usermode on ppc64

[Alexander Kanavin]

It simply does not work at all:
https://lists.yoctoproject.org/pipermail/yocto/2016-April/029698.html

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
---
 meta/conf/machine/include/powerpc/arch-powerpc64.inc | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/meta/conf/machine/include/powerpc/arch-powerpc64.inc b/meta/conf/machine/include/powerpc/arch-powerpc64.inc
index 49419b6..f751c6b 100644
--- a/meta/conf/machine/include/powerpc/arch-powerpc64.inc
+++ b/meta/conf/machine/include/powerpc/arch-powerpc64.inc
@@ -7,6 +7,9 @@ TUNECONFLICTS[m64] = "m32 nf"
 TUNE_CCARGS .= "${@bb.utils.contains('TUNE_FEATURES', 'm64', ' -m64', '', d)}"
 TUNE_ARCH .= "${@bb.utils.contains('TUNE_FEATURES', [ 'm64' ], 'powerpc64', '', d)}"
 
+# user mode qemu doesn't support ppc64
+MACHINE_FEATURES_BACKFILL_CONSIDERED_append = " ${@bb.utils.contains('TUNE_FEATURES', 'm64', 'qemu-usermode', '', d)}"
+
 AVAILTUNES += "powerpc64"
 TUNE_FEATURES_tune-powerpc64 = "m64 fpu-hard"
 BASE_LIB_tune-powerpc64 = "lib64"
-- 
2.8.0.rc3

Re: [PATCH] arch-powerpc64.inc: disable the use of qemu usermode on ppc64

[Khem Raj]

On Wed, Apr 27, 2016 at 5:44 AM, Alexander Kanavin
<alexander.kanavin@linux.intel.com> wrote:
> It simply does not work at all:
> https://lists.yoctoproject.org/pipermail/yocto/2016-April/029698.html
>
> Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
> ---
>  meta/conf/machine/include/powerpc/arch-powerpc64.inc | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/meta/conf/machine/include/powerpc/arch-powerpc64.inc b/meta/conf/machine/include/powerpc/arch-powerpc64.inc
> index 49419b6..f751c6b 100644
> --- a/meta/conf/machine/include/powerpc/arch-powerpc64.inc
> +++ b/meta/conf/machine/include/powerpc/arch-powerpc64.inc
> @@ -7,6 +7,9 @@ TUNECONFLICTS[m64] = "m32 nf"
>  TUNE_CCARGS .= "${@bb.utils.contains('TUNE_FEATURES', 'm64', ' -m64', '', d)}"
>  TUNE_ARCH .= "${@bb.utils.contains('TUNE_FEATURES', [ 'm64' ], 'powerpc64', '', d)}"
>
> +# user mode qemu doesn't support ppc64
> +MACHINE_FEATURES_BACKFILL_CONSIDERED_append = " ${@bb.utils.contains('TUNE_FEATURES', 'm64', 'qemu-usermode', '', d)}"
> +

should we replace qemu-ppc with qemu-ppc64 instead of disabling it

>  AVAILTUNES += "powerpc64"
>  TUNE_FEATURES_tune-powerpc64 = "m64 fpu-hard"
>  BASE_LIB_tune-powerpc64 = "lib64"
> --
> 2.8.0.rc3
>
> --
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [PATCH] arch-powerpc64.inc: disable the use of qemu usermode on ppc64

[Alexander Kanavin]

On 04/29/2016 11:34 PM, Khem Raj wrote:
> should we replace qemu-ppc with qemu-ppc64 instead of disabling it

qemu-ppc64 is already used on 64 bit powerpc, the problem is that it 
always crashes.

Alex

Re: [PATCH] arch-powerpc64.inc: disable the use of qemu usermode on ppc64

[Khem Raj]

[-- Attachment #1: Type: text/plain, Size: 361 bytes --]


> On May 2, 2016, at 3:06 AM, Alexander Kanavin <alexander.kanavin@linux.intel.com> wrote:
> 
> On 04/29/2016 11:34 PM, Khem Raj wrote:
>> should we replace qemu-ppc with qemu-ppc64 instead of disabling it
> 
> qemu-ppc64 is already used on 64 bit powerpc, the problem is that it always crashes.

Do we need to adjust the emulated machine parameter ?

[-- Attachment #2: Message signed with OpenPGP using GPGMail --]
[-- Type: application/pgp-signature, Size: 211 bytes --]

Re: [PATCH] arch-powerpc64.inc: disable the use of qemu usermode on ppc64

[Alexander Kanavin]

On 05/02/2016 06:46 PM, Khem Raj wrote:

>>> should we replace qemu-ppc with qemu-ppc64 instead of disabling it
>>
>> qemu-ppc64 is already used on 64 bit powerpc, the problem is that it always crashes.
>
> Do we need to adjust the emulated machine parameter ?

I don't see how. The CPU list produced by 'qemu-ppc64 -cpu help' does 
not include anything that resembles e5500/6500 or Power5/6/7 (the 64 bit 
ppc chips that oe-core has definitions for), and I'm not a powerpc 
expert by any measure, so I can't investigate the problem any further. 
It needs to be looked at by someone who is actually working with that 
hardware.

The best we can do now is disable use of qemu, so that builds that 
target ppc64 don't fail.

Alex