| Message ID | 20240405103027.2704728-1-Sergiy_Kibrik@epam.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [XEN,V1] x86/ucode: optional amd/intel ucode build & load | expand |
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
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
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.
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.
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 )
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(-)