| Message ID | 20201117202308.7568-1-info@metux.net |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [1/2] x86: make vmware support optional | expand |
On Tue, Nov 17, 2020 at 09:23:07PM +0100, Enrico Weigelt, metux IT consult wrote: > Make it possible to opt-out from vmware support Why? I can think of a couple of reasons but maybe yours might not be the one I'm thinking of. > Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net> > --- > arch/x86/Kconfig | 7 +++++++ > arch/x86/kernel/cpu/Makefile | 4 +++- > arch/x86/kernel/cpu/hypervisor.c | 2 ++ > drivers/input/mouse/Kconfig | 2 +- > drivers/misc/Kconfig | 2 +- > drivers/ptp/Kconfig | 2 +- > 6 files changed, 15 insertions(+), 4 deletions(-) > > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig > index f6946b81f74a..c227c1fa0091 100644 > --- a/arch/x86/Kconfig > +++ b/arch/x86/Kconfig > @@ -801,6 +801,13 @@ config X86_HV_CALLBACK_VECTOR > > source "arch/x86/xen/Kconfig" > > +config VMWARE_GUEST > + bool "Vmware Guest support" > + default y depends on HYPERVISOR_GUEST. The hyperv one too.
On 17.11.20 21:31, Borislav Petkov wrote: > On Tue, Nov 17, 2020 at 09:23:07PM +0100, Enrico Weigelt, metux IT consult wrote: >> Make it possible to opt-out from vmware support > > Why? Reducing the kernel size. Think of very high density virtualization (w/ specially stripped-down workloads) or embedded systems. For example, I'm running bare minimum kernels w/ only kvm and virtio (not even pci, etc) in such scenarios. Of course, that's nothing for an average distro, therefore leaving default y. --mtx > > I can think of a couple of reasons but maybe yours might not be the one > I'm thinking of. > >> Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net> >> --- >> arch/x86/Kconfig | 7 +++++++ >> arch/x86/kernel/cpu/Makefile | 4 +++- >> arch/x86/kernel/cpu/hypervisor.c | 2 ++ >> drivers/input/mouse/Kconfig | 2 +- >> drivers/misc/Kconfig | 2 +- >> drivers/ptp/Kconfig | 2 +- >> 6 files changed, 15 insertions(+), 4 deletions(-) >> >> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig >> index f6946b81f74a..c227c1fa0091 100644 >> --- a/arch/x86/Kconfig >> +++ b/arch/x86/Kconfig >> @@ -801,6 +801,13 @@ config X86_HV_CALLBACK_VECTOR >> >> source "arch/x86/xen/Kconfig" >> >> +config VMWARE_GUEST >> + bool "Vmware Guest support" >> + default y > > depends on HYPERVISOR_GUEST. The hyperv one too. >
On Wed, Dec 02, 2020 at 08:17:23PM +0100, Enrico Weigelt, metux IT consult wrote: > Reducing the kernel size. Think of very high density virtualization > (w/ specially stripped-down workloads) or embedded systems. > > For example, I'm running bare minimum kernels w/ only kvm and virtio > (not even pci, etc) in such scenarios. > > Of course, that's nothing for an average distro, therefore leaving > default y. Ok, pls put the reasoning for the change in the next revision's commit message along with how much KB savings we're talking about. Thx.
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index f6946b81f74a..c227c1fa0091 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -801,6 +801,13 @@ config X86_HV_CALLBACK_VECTOR source "arch/x86/xen/Kconfig" +config VMWARE_GUEST + bool "Vmware Guest support" + default y + help + This option enables several optimizations for running under the + VMware hypervisor. + config KVM_GUEST bool "KVM Guest support (including kvmclock)" depends on PARAVIRT diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index 93792b457b81..a615b0152bf0 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -51,7 +51,9 @@ obj-$(CONFIG_X86_CPU_RESCTRL) += resctrl/ obj-$(CONFIG_X86_LOCAL_APIC) += perfctr-watchdog.o -obj-$(CONFIG_HYPERVISOR_GUEST) += vmware.o hypervisor.o mshyperv.o +obj-$(CONFIG_HYPERVISOR_GUEST) += hypervisor.o mshyperv.o +obj-$(CONFIG_VMWARE_GUEST) += vmware.o + obj-$(CONFIG_ACRN_GUEST) += acrn.o ifdef CONFIG_X86_FEATURE_NAMES diff --git a/arch/x86/kernel/cpu/hypervisor.c b/arch/x86/kernel/cpu/hypervisor.c index 553bfbfc3a1b..c0e770a224aa 100644 --- a/arch/x86/kernel/cpu/hypervisor.c +++ b/arch/x86/kernel/cpu/hypervisor.c @@ -34,7 +34,9 @@ static const __initconst struct hypervisor_x86 * const hypervisors[] = #ifdef CONFIG_XEN_PVHVM &x86_hyper_xen_hvm, #endif +#ifdef CONFIG_VMWARE_GUEST &x86_hyper_vmware, +#endif &x86_hyper_ms_hyperv, #ifdef CONFIG_KVM_GUEST &x86_hyper_kvm, diff --git a/drivers/input/mouse/Kconfig b/drivers/input/mouse/Kconfig index d8b6a5dab190..29ced69d5c85 100644 --- a/drivers/input/mouse/Kconfig +++ b/drivers/input/mouse/Kconfig @@ -186,7 +186,7 @@ config MOUSE_PS2_FOCALTECH config MOUSE_PS2_VMMOUSE bool "Virtual mouse (vmmouse)" - depends on MOUSE_PS2 && X86 && HYPERVISOR_GUEST + depends on MOUSE_PS2 && X86 && VMWARE_GUEST help Say Y here if you are running under control of VMware hypervisor (ESXi, Workstation or Fusion). Also make sure that when you enable diff --git a/drivers/misc/Kconfig b/drivers/misc/Kconfig index fafa8b0d8099..d2bd8eff6eb6 100644 --- a/drivers/misc/Kconfig +++ b/drivers/misc/Kconfig @@ -363,7 +363,7 @@ config DS1682 config VMWARE_BALLOON tristate "VMware Balloon Driver" - depends on VMWARE_VMCI && X86 && HYPERVISOR_GUEST + depends on VMWARE_VMCI && X86 && VMWARE_GUEST select MEMORY_BALLOON help This is VMware physical memory management driver which acts diff --git a/drivers/ptp/Kconfig b/drivers/ptp/Kconfig index 942f72d8151d..6bf30153270e 100644 --- a/drivers/ptp/Kconfig +++ b/drivers/ptp/Kconfig @@ -141,7 +141,7 @@ config PTP_1588_CLOCK_IDTCM config PTP_1588_CLOCK_VMW tristate "VMware virtual PTP clock" - depends on ACPI && HYPERVISOR_GUEST && X86 + depends on ACPI && VMWARE_GUEST && X86 depends on PTP_1588_CLOCK help This driver adds support for using VMware virtual precision
Make it possible to opt-out from vmware support Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net> --- arch/x86/Kconfig | 7 +++++++ arch/x86/kernel/cpu/Makefile | 4 +++- arch/x86/kernel/cpu/hypervisor.c | 2 ++ drivers/input/mouse/Kconfig | 2 +- drivers/misc/Kconfig | 2 +- drivers/ptp/Kconfig | 2 +- 6 files changed, 15 insertions(+), 4 deletions(-)