* [XEN PATCH V1] x86/ucode: optional amd/intel ucode build & load
@ 2024-04-05 10:30 Sergiy Kibrik
2024-04-05 10:57 ` Andrew Cooper
0 siblings, 1 reply; 5+ messages in thread
From: Sergiy Kibrik @ 2024-04-05 10:30 UTC (permalink / raw
To: xen-devel; +Cc: Jan Beulich, Andrew Cooper, Sergiy Kibrik
Introduce configuration variables to make it possible to selectively turn
on/off CPU microcode management code in the build for AMD and Intel CPUs.
This is to allow build configuration for a specific CPU, not compile and
load what will not be used anyway.
Signed-off-by: Sergiy Kibrik <Sergiy_Kibrik@epam.com>
---
xen/arch/x86/Kconfig | 12 ++++++++++++
xen/arch/x86/cpu/microcode/Makefile | 4 ++--
xen/arch/x86/cpu/microcode/core.c | 5 ++++-
3 files changed, 18 insertions(+), 3 deletions(-)
diff --git a/xen/arch/x86/Kconfig b/xen/arch/x86/Kconfig
index d6f3128588..1ee5ae793d 100644
--- a/xen/arch/x86/Kconfig
+++ b/xen/arch/x86/Kconfig
@@ -350,6 +350,18 @@ config REQUIRE_NX
was unavailable. However, if enabled, Xen will no longer boot on
any CPU which is lacking NX support.
+config MICROCODE_INTEL
+ bool "Build with Intel CPU ucode support" if EXPERT
+ default y
+ help
+ Support microcode management for Intel processors. If unsure, say Y.
+
+config MICROCODE_AMD
+ bool "Build with AMD CPU ucode support" if EXPERT
+ default y
+ help
+ Support microcode management for AMD K10 family of processors
+ and later. If unsure, say Y.
endmenu
source "common/Kconfig"
diff --git a/xen/arch/x86/cpu/microcode/Makefile b/xen/arch/x86/cpu/microcode/Makefile
index aae235245b..abd0afe8c5 100644
--- a/xen/arch/x86/cpu/microcode/Makefile
+++ b/xen/arch/x86/cpu/microcode/Makefile
@@ -1,3 +1,3 @@
-obj-y += amd.o
obj-y += core.o
-obj-y += intel.o
+obj-$(CONFIG_MICROCODE_AMD) += amd.o
+obj-$(CONFIG_MICROCODE_INTEL) += intel.o
diff --git a/xen/arch/x86/cpu/microcode/core.c b/xen/arch/x86/cpu/microcode/core.c
index 1c9f66ea8a..b7c108f68b 100644
--- a/xen/arch/x86/cpu/microcode/core.c
+++ b/xen/arch/x86/cpu/microcode/core.c
@@ -865,6 +865,7 @@ int __init early_microcode_init(unsigned long *module_map,
switch ( c->x86_vendor )
{
+#ifdef CONFIG_MICROCODE_AMD
case X86_VENDOR_AMD:
if ( c->x86 >= 0x10 )
{
@@ -872,11 +873,13 @@ int __init early_microcode_init(unsigned long *module_map,
can_load = true;
}
break;
-
+#endif
+#ifdef CONFIG_MICROCODE_INTEL
case X86_VENDOR_INTEL:
ucode_ops = intel_ucode_ops;
can_load = intel_can_load_microcode();
break;
+#endif
}
if ( !ucode_ops.apply_microcode )
--
2.25.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [XEN PATCH V1] x86/ucode: optional amd/intel ucode build & load
2024-04-05 10:30 [XEN PATCH V1] x86/ucode: optional amd/intel ucode build & load Sergiy Kibrik
@ 2024-04-05 10:57 ` Andrew Cooper
2024-04-09 10:34 ` Sergiy Kibrik
0 siblings, 1 reply; 5+ messages in thread
From: Andrew Cooper @ 2024-04-05 10:57 UTC (permalink / raw
To: Sergiy Kibrik, xen-devel; +Cc: Jan Beulich
On 05/04/2024 11:30 am, Sergiy Kibrik wrote:
> Introduce configuration variables to make it possible to selectively turn
> on/off CPU microcode management code in the build for AMD and Intel CPUs.
>
> This is to allow build configuration for a specific CPU, not compile and
> load what will not be used anyway.
>
> Signed-off-by: Sergiy Kibrik <Sergiy_Kibrik@epam.com>
Hmm... Stefano didn't check up with me first.
https://lore.kernel.org/xen-devel/20231027191926.3283871-1-andrew.cooper3@citrix.com/
I've already got a series out for this, although its blocked on naming
and IMO, particularly unhelpful replies. I've not had time to apply the
community-resolution vote on naming issues. (Also, you should be CC-ing
Roger on x86 patches).
For microcode loading specifically, there's no situation even for safety
certification where it's reasonable to remove this functionality.
Therefore it wants tying to generic Intel / AMD only, and not a
ucode-loading specific option.
~Andrew
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [XEN PATCH V1] x86/ucode: optional amd/intel ucode build & load
2024-04-05 10:57 ` Andrew Cooper
@ 2024-04-09 10:34 ` Sergiy Kibrik
2024-04-09 21:49 ` Stefano Stabellini
0 siblings, 1 reply; 5+ messages in thread
From: Sergiy Kibrik @ 2024-04-09 10:34 UTC (permalink / raw
To: Andrew Cooper, Stefano Stabellini, Roger Pau Monné
Cc: Jan Beulich, xen-devel
05.04.24 13:57, Andrew Cooper:
> On 05/04/2024 11:30 am, Sergiy Kibrik wrote:
>> Introduce configuration variables to make it possible to selectively turn
>> on/off CPU microcode management code in the build for AMD and Intel CPUs.
>>
>> This is to allow build configuration for a specific CPU, not compile and
>> load what will not be used anyway.
>>
>> Signed-off-by: Sergiy Kibrik <Sergiy_Kibrik@epam.com>
>
> Hmm... Stefano didn't check up with me first.
>
> _https://lore.kernel.org/xen-devel/20231027191926.3283871-1-andrew.cooper3@citrix.com/
>
> I've already got a series out for this, although its blocked on naming
> and IMO, particularly unhelpful replies. I've not had time to apply the
> community-resolution vote on naming issues. (Also, you should be CC-ing
> Roger on x86 patches).
+ Stefano & Roger
should we revisit this series then?
-Sergiy
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [XEN PATCH V1] x86/ucode: optional amd/intel ucode build & load
2024-04-09 10:34 ` Sergiy Kibrik
@ 2024-04-09 21:49 ` Stefano Stabellini
2024-04-17 15:01 ` Jan Beulich
0 siblings, 1 reply; 5+ messages in thread
From: Stefano Stabellini @ 2024-04-09 21:49 UTC (permalink / raw
To: Sergiy Kibrik
Cc: Andrew Cooper, Stefano Stabellini, Roger Pau Monné,
Jan Beulich, xen-devel
[-- Attachment #1: Type: text/plain, Size: 1377 bytes --]
On Tue, 9 Apr 2024, Sergiy Kibrik wrote:
> 05.04.24 13:57, Andrew Cooper:
> > On 05/04/2024 11:30 am, Sergiy Kibrik wrote:
> > > Introduce configuration variables to make it possible to selectively turn
> > > on/off CPU microcode management code in the build for AMD and Intel CPUs.
> > >
> > > This is to allow build configuration for a specific CPU, not compile and
> > > load what will not be used anyway.
> > >
> > > Signed-off-by: Sergiy Kibrik <Sergiy_Kibrik@epam.com>
> >
> > Hmm... Stefano didn't check up with me first.
> >
> > _https://lore.kernel.org/xen-devel/20231027191926.3283871-1-andrew.cooper3@citrix.com/
> >
> > I've already got a series out for this, although its blocked on naming
> > and IMO, particularly unhelpful replies. I've not had time to apply the
> > community-resolution vote on naming issues. (Also, you should be CC-ing
> > Roger on x86 patches).
>
> + Stefano & Roger
>
> should we revisit this series then?
Yes, I replied. I think there are three naming options:
Option-1)
CONFIG_{AMD,INTEL}
nothing else
Option-2)
CONFIG_{AMD,INTEL}_IOMMU
CONFIG_{AMD,INTEL} selects CONFIG_{AMD,INTEL}_IOMMU
Option-3)
CONFIG_{AMD,INTEL}_IOMMU
CONFIG_{AMD,INTEL}_CPU
CONFIG_{AMD,INTEL} selects both CPU and IOMMU options
My preference is Option-1 (best), Option-3, Option-2 (worse)
I am fine with anything but please let move this forward.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [XEN PATCH V1] x86/ucode: optional amd/intel ucode build & load
2024-04-09 21:49 ` Stefano Stabellini
@ 2024-04-17 15:01 ` Jan Beulich
0 siblings, 0 replies; 5+ messages in thread
From: Jan Beulich @ 2024-04-17 15:01 UTC (permalink / raw
To: Andrew Cooper
Cc: xen-devel, Sergiy Kibrik, Stefano Stabellini,
Roger Pau Monné
On 09.04.2024 23:49, Stefano Stabellini wrote:
> On Tue, 9 Apr 2024, Sergiy Kibrik wrote:
>> 05.04.24 13:57, Andrew Cooper:
>>> On 05/04/2024 11:30 am, Sergiy Kibrik wrote:
>>>> Introduce configuration variables to make it possible to selectively turn
>>>> on/off CPU microcode management code in the build for AMD and Intel CPUs.
>>>>
>>>> This is to allow build configuration for a specific CPU, not compile and
>>>> load what will not be used anyway.
>>>>
>>>> Signed-off-by: Sergiy Kibrik <Sergiy_Kibrik@epam.com>
>>>
>>> Hmm... Stefano didn't check up with me first.
>>>
>>> _https://lore.kernel.org/xen-devel/20231027191926.3283871-1-andrew.cooper3@citrix.com/
>>>
>>> I've already got a series out for this, although its blocked on naming
>>> and IMO, particularly unhelpful replies. I've not had time to apply the
>>> community-resolution vote on naming issues. (Also, you should be CC-ing
>>> Roger on x86 patches).
I can't remain silent here. Roger has now changed his mind, so formally
things are correct. Yet judging from the earlier example where Roger
was agreeing with you, shouldn't it have been this time the other way
around: A majority was of different opinion than you, and you should
have accepted that? Instead you've waited for a time when I was away,
got Stefano to agree and Roger to change his mind, and once again you
got what you want. It feels increasingly like not everyone among the
REST maintainers and not everyone among the x86 ones are equal.
Jan
>> + Stefano & Roger
>>
>> should we revisit this series then?
>
> Yes, I replied. I think there are three naming options:
>
> Option-1)
> CONFIG_{AMD,INTEL}
> nothing else
>
> Option-2)
> CONFIG_{AMD,INTEL}_IOMMU
> CONFIG_{AMD,INTEL} selects CONFIG_{AMD,INTEL}_IOMMU
>
> Option-3)
> CONFIG_{AMD,INTEL}_IOMMU
> CONFIG_{AMD,INTEL}_CPU
> CONFIG_{AMD,INTEL} selects both CPU and IOMMU options
>
>
> My preference is Option-1 (best), Option-3, Option-2 (worse)
>
> I am fine with anything but please let move this forward.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2024-04-17 15:01 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-04-05 10:30 [XEN PATCH V1] x86/ucode: optional amd/intel ucode build & load Sergiy Kibrik
2024-04-05 10:57 ` Andrew Cooper
2024-04-09 10:34 ` Sergiy Kibrik
2024-04-09 21:49 ` Stefano Stabellini
2024-04-17 15:01 ` Jan Beulich
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.