Skip to contentSkip to conversation

i386: KVM: expand Hyper-V features early

RFC PATCH Superseded Vitaly Kuznetsov 2 participants 37 messages Sep 4, 2020 – Sep 22, 2020

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

cover i386: KVM: expand Hyper-V features early

This pairs with https://lore.kernel.org/kvm/[email protected]/
and as KVM part is not yet accepted, sending this as RFC.

Upper layer tools like libvirt want to figure out which Hyper-V features are supported by the underlying stack (QEMU/KVM) but currently they are unable to do so. We have a nice 'hv_passthrough' CPU flag supported by QEMU but it has no effect on e.g. QMP's

query-cpu-model-expansion type=full model={"name":"host","props":{"hv-passthrough":true}}

command as we parse Hyper-V features after creating KVM vCPUs and not at feature expansion time. To support the use-case we first need to make KVM_GET_SUPPORTED_HV_CPUID ioctl a system-wide ioctl as the existing vCPU version can't be used that early. This is what KVM part does. With that done, we can make early Hyper-V feature expansion (this series).

Vitaly Kuznetsov (22):
WIP: update linux/headers
i386: drop x86_cpu_get_supported_feature_word() forward declaration
i386: move hyperv_vendor_id initialization to x86_cpu_realizefn()
i386: move hyperv_interface_id initialization to x86_cpu_realizefn()
i386: move hyperv_version_id initialization to x86_cpu_realizefn()
i386: move hyperv_limits initialization to x86_cpu_realizefn()
i386: fill in FEAT_HYPERV_EDX from edx instead of eax
i386: invert hyperv_spinlock_attempts setting logic with
hv_passthrough
i386: add reserved FEAT_HYPERV_ECX CPUID leaf
i386: add reserved FEAT_HV_RECOMM_ECX/FEAT_HV_RECOMM_EDX CPUID leaves
i386: add reserved FEAT_HV_NESTED_EBX/ECX/EDX CPUID leaves
i386: always fill Hyper-V CPUID feature leaves from X86CPU data
i386: split hyperv_handle_properties() into
hyperv_expand_features()/hyperv_fill_cpuids()
i386: move eVMCS enablement to hyperv_init_vcpu()
i386: switch hyperv_expand_features() to using error_setg()
i386: make hyperv_expand_features() return void
i386: adjust the expected KVM_GET_SUPPORTED_HV_CPUID array size
i386: prefer system KVM_GET_SUPPORTED_HV_CPUID ioctl over vCPU's one
i386: prepare hyperv_expand_features() to be called at CPU feature
expansion time
i386: use global kvm_state in hyperv_enabled() check
i386: record if Hyper-V features were already expanded
i386: expand Hyper-V features early

 include/standard-headers/drm/drm_fourcc.h     | 140 +++++++
 include/standard-headers/linux/ethtool.h      |  87 +++++
 .../linux/input-event-codes.h                 |   3 +-
 linux-headers/asm-generic/unistd.h            |   6 +-
 linux-headers/asm-mips/unistd_n32.h           |   1 +
 linux-headers/asm-mips/unistd_n64.h           |   1 +
 linux-headers/asm-mips/unistd_o32.h           |   1 +
 linux-headers/asm-powerpc/kvm.h               |   5 +
 linux-headers/asm-powerpc/mman.h              |   2 +-
 linux-headers/asm-powerpc/unistd_32.h         |   1 +
 linux-headers/asm-powerpc/unistd_64.h         |   1 +
 linux-headers/asm-s390/kvm.h                  |   7 +-
 linux-headers/asm-s390/unistd_32.h            |   1 +
 linux-headers/asm-s390/unistd_64.h            |   1 +
 linux-headers/asm-x86/unistd_32.h             |   1 +
 linux-headers/asm-x86/unistd_64.h             |   1 +
 linux-headers/asm-x86/unistd_x32.h            |   1 +
 linux-headers/linux/kvm.h                     |   7 +-
 linux-headers/linux/vfio.h                    |   2 +-
 target/i386/cpu.c                             |  78 +++-
 target/i386/cpu.h                             |  13 +-
 target/i386/kvm.c                             | 347 ++++++++++++------
 target/i386/kvm_i386.h                        |   7 +
 23 files changed, 579 insertions(+), 135 deletions(-)
Signature · 1 line
2.25.4
#1 lore raw

RFC PATCH 1/22 WIP: update linux/headers

KVM_CAP_SYS_HYPERV_CPUID definition is needed for this series.

---
 include/standard-headers/drm/drm_fourcc.h     | 140 ++++++++++++++++++
 include/standard-headers/linux/ethtool.h      |  87 +++++++++++
 .../linux/input-event-codes.h                 |   3 +-
 linux-headers/asm-generic/unistd.h            |   6 +-
 linux-headers/asm-mips/unistd_n32.h           |   1 +
 linux-headers/asm-mips/unistd_n64.h           |   1 +
 linux-headers/asm-mips/unistd_o32.h           |   1 +
 linux-headers/asm-powerpc/kvm.h               |   5 +
 linux-headers/asm-powerpc/mman.h              |   2 +-
 linux-headers/asm-powerpc/unistd_32.h         |   1 +
 linux-headers/asm-powerpc/unistd_64.h         |   1 +
 linux-headers/asm-s390/kvm.h                  |   7 +-
 linux-headers/asm-s390/unistd_32.h            |   1 +
 linux-headers/asm-s390/unistd_64.h            |   1 +
 linux-headers/asm-x86/unistd_32.h             |   1 +
 linux-headers/asm-x86/unistd_64.h             |   1 +
 linux-headers/asm-x86/unistd_x32.h            |   1 +
 linux-headers/linux/kvm.h                     |   7 +-
 linux-headers/linux/vfio.h                    |   2 +-
 19 files changed, 261 insertions(+), 8 deletions(-)
include/standard-headers/drm/drm_fourcc.h +140 -0
index 909a66753c03..2757bdf6fa70 100644
@@ -235,6 +235,12 @@ extern "C" {
 #define DRM_FORMAT_NV61		fourcc_code('N', 'V', '6', '1') /* 2x1 subsampled Cb:Cr plane */
 #define DRM_FORMAT_NV24		fourcc_code('N', 'V', '2', '4') /* non-subsampled Cr:Cb plane */
 #define DRM_FORMAT_NV42		fourcc_code('N', 'V', '4', '2') /* non-subsampled Cb:Cr plane */
+/*
+ * 2 plane YCbCr
+ * index 0 = Y plane, [39:0] Y3:Y2:Y1:Y0 little endian
+ * index 1 = Cr:Cb plane, [39:0] Cr1:Cb1:Cr0:Cb0 little endian
+ */
+#define DRM_FORMAT_NV15		fourcc_code('N', 'V', '1', '5') /* 2x2 subsampled Cr:Cb plane */
 
 /*
  * 2 plane YCbCr MSB aligned
@@ -264,6 +270,22 @@ extern "C" {
  */
 #define DRM_FORMAT_P016		fourcc_code('P', '0', '1', '6') /* 2x2 subsampled Cr:Cb plane 16 bits per channel */
 
+/* 3 plane non-subsampled (444) YCbCr
+ * 16 bits per component, but only 10 bits are used and 6 bits are padded
+ * index 0: Y plane, [15:0] Y:x [10:6] little endian
+ * index 1: Cb plane, [15:0] Cb:x [10:6] little endian
+ * index 2: Cr plane, [15:0] Cr:x [10:6] little endian
+ */
+#define DRM_FORMAT_Q410		fourcc_code('Q', '4', '1', '0')
+
+/* 3 plane non-subsampled (444) YCrCb
+ * 16 bits per component, but only 10 bits are used and 6 bits are padded
+ * index 0: Y plane, [15:0] Y:x [10:6] little endian
+ * index 1: Cr plane, [15:0] Cr:x [10:6] little endian
+ * index 2: Cb plane, [15:0] Cb:x [10:6] little endian
+ */
+#define DRM_FORMAT_Q401		fourcc_code('Q', '4', '0', '1')
+
 /*
  * 3 plane YCbCr
  * index 0: Y plane, [7:0] Y
@@ -308,6 +330,7 @@ extern "C" {
 #define DRM_FORMAT_MOD_VENDOR_BROADCOM 0x07
 #define DRM_FORMAT_MOD_VENDOR_ARM     0x08
 #define DRM_FORMAT_MOD_VENDOR_ALLWINNER 0x09
+#define DRM_FORMAT_MOD_VENDOR_AMLOGIC 0x0a
 
 /* add more to the end as needed */
 
@@ -322,8 +345,33 @@ extern "C" {
  * When adding a new token please document the layout with a code comment,
  * similar to the fourcc codes above. drm_fourcc.h is considered the
  * authoritative source for all of these.
+ *
+ * Generic modifier names:
+ *
+ * DRM_FORMAT_MOD_GENERIC_* definitions are used to provide vendor-neutral names
+ * for layouts which are common across multiple vendors. To preserve
+ * compatibility, in cases where a vendor-specific definition already exists and
+ * a generic name for it is desired, the common name is a purely symbolic alias
+ * and must use the same numerical value as the original definition.
+ *
+ * Note that generic names should only be used for modifiers which describe
+ * generic layouts (such as pixel re-ordering), which may have
+ * independently-developed support across multiple vendors.
+ *
+ * In future cases where a generic layout is identified before merging with a
+ * vendor-specific modifier, a new 'GENERIC' vendor or modifier using vendor
+ * 'NONE' could be considered. This should only be for obvious, exceptional
+ * cases to avoid polluting the 'GENERIC' namespace with modifiers which only
+ * apply to a single vendor.
+ *
+ * Generic names should not be used for cases where multiple hardware vendors
+ * have implementations of the same standardised compression scheme (such as
+ * AFBC). In those cases, all implementations should use the same format
+ * modifier(s), reflecting the vendor of the standard.
  */
 
+#define DRM_FORMAT_MOD_GENERIC_16_16_TILE DRM_FORMAT_MOD_SAMSUNG_16_16_TILE
+
 /*
  * Invalid Modifier
  *
@@ -891,6 +939,18 @@ drm_fourcc_canonicalize_nvidia_format_mod(uint64_t modifier)
  */
 #define AFBC_FORMAT_MOD_BCH     (1ULL << 11)
 
+/* AFBC uncompressed storage mode
+ *
+ * Indicates that the buffer is using AFBC uncompressed storage mode.
+ * In this mode all superblock payloads in the buffer use the uncompressed
+ * storage mode, which is usually only used for data which cannot be compressed.
+ * The buffer layout is the same as for AFBC buffers without USM set, this only
+ * affects the storage mode of the individual superblocks. Note that even a
+ * buffer without USM set may use uncompressed storage mode for some or all
+ * superblocks, USM just guarantees it for all.
+ */
+#define AFBC_FORMAT_MOD_USM	(1ULL << 12)
+
 /*
  * Arm 16x16 Block U-Interleaved modifier
  *
@@ -915,6 +975,86 @@ drm_fourcc_canonicalize_nvidia_format_mod(uint64_t modifier)
  */
 #define DRM_FORMAT_MOD_ALLWINNER_TILED fourcc_mod_code(ALLWINNER, 1)
 
+/*
+ * Amlogic Video Framebuffer Compression modifiers
+ *
+ * Amlogic uses a proprietary lossless image compression protocol and format
+ * for their hardware video codec accelerators, either video decoders or
+ * video input encoders.
+ *
+ * It considerably reduces memory bandwidth while writing and reading
+ * frames in memory.
+ *
+ * The underlying storage is considered to be 3 components, 8bit or 10-bit
+ * per component YCbCr 420, single plane :
+ * - DRM_FORMAT_YUV420_8BIT
+ * - DRM_FORMAT_YUV420_10BIT
+ *
+ * The first 8 bits of the mode defines the layout, then the following 8 bits
+ * defines the options changing the layout.
+ *
+ * Not all combinations are valid, and different SoCs may support different
+ * combinations of layout and options.
+ */
+#define __fourcc_mod_amlogic_layout_mask 0xf
+#define __fourcc_mod_amlogic_options_shift 8
+#define __fourcc_mod_amlogic_options_mask 0xf
+
+#define DRM_FORMAT_MOD_AMLOGIC_FBC(__layout, __options) \
+	fourcc_mod_code(AMLOGIC, \
+			((__layout) & __fourcc_mod_amlogic_layout_mask) | \
+			((__options) & __fourcc_mod_amlogic_options_mask \
+			 << __fourcc_mod_amlogic_options_shift))
+
+/* Amlogic FBC Layouts */
+
+/*
+ * Amlogic FBC Basic Layout
+ *
+ * The basic layout is composed of:
+ * - a body content organized in 64x32 superblocks with 4096 bytes per
+ *   superblock in default mode.
+ * - a 32 bytes per 128x64 header block
+ *
+ * This layout is transferrable between Amlogic SoCs supporting this modifier.
+ */
+#define AMLOGIC_FBC_LAYOUT_BASIC		(1ULL)
+
+/*
+ * Amlogic FBC Scatter Memory layout
+ *
+ * Indicates the header contains IOMMU references to the compressed
+ * frames content to optimize memory access and layout.
+ *
+ * In this mode, only the header memory address is needed, thus the
+ * content memory organization is tied to the current producer
+ * execution and cannot be saved/dumped neither transferrable between
+ * Amlogic SoCs supporting this modifier.
+ *
+ * Due to the nature of the layout, these buffers are not expected to
+ * be accessible by the user-space clients, but only accessible by the
+ * hardware producers and consumers.
+ *
+ * The user-space clients should expect a failure while trying to mmap
+ * the DMA-BUF handle returned by the producer.
+ */
+#define AMLOGIC_FBC_LAYOUT_SCATTER		(2ULL)
+
+/* Amlogic FBC Layout Options Bit Mask */
+
+/*
+ * Amlogic FBC Memory Saving mode
+ *
+ * Indicates the storage is packed when pixel size is multiple of word
+ * boudaries, i.e. 8bit should be stored in this mode to save allocation
+ * memory.
+ *
+ * This mode reduces body layout to 3072 bytes per 64x32 superblock with
+ * the basic layout and 3200 bytes per 64x32 superblock combined with
+ * the scatter layout.
+ */
+#define AMLOGIC_FBC_OPTION_MEM_SAVING		(1ULL << 0)
+
 #if defined(__cplusplus)
 }
 #endif
include/standard-headers/linux/ethtool.h +87 -0

111 lines not on this page · raw message

include/standard-headers/linux/input-event-codes.h +2 -1
index ebf72c10317b..e740ad9f2e01 100644
@@ -888,7 +888,8 @@
 #define SW_LINEIN_INSERT	0x0d  /* set = inserted */
 #define SW_MUTE_DEVICE		0x0e  /* set = device disabled */
 #define SW_PEN_INSERTED		0x0f  /* set = pen inserted */
-#define SW_MAX_			0x0f
+#define SW_MACHINE_COVER	0x10  /* set = cover closed */
+#define SW_MAX_			0x10
 #define SW_CNT			(SW_MAX_+1)
 
 /*
linux-headers/asm-generic/unistd.h +4 -2
index f4a01305d9a6..995b36c2ea7d 100644
@@ -606,9 +606,9 @@ __SYSCALL(__NR_sendto, sys_sendto)
 #define __NR_recvfrom 207
 __SC_COMP(__NR_recvfrom, sys_recvfrom, compat_sys_recvfrom)
 #define __NR_setsockopt 208
-__SC_COMP(__NR_setsockopt, sys_setsockopt, compat_sys_setsockopt)
+__SC_COMP(__NR_setsockopt, sys_setsockopt, sys_setsockopt)
 #define __NR_getsockopt 209
-__SC_COMP(__NR_getsockopt, sys_getsockopt, compat_sys_getsockopt)
+__SC_COMP(__NR_getsockopt, sys_getsockopt, sys_getsockopt)
 #define __NR_shutdown 210
 __SYSCALL(__NR_shutdown, sys_shutdown)
 #define __NR_sendmsg 211
@@ -850,6 +850,8 @@ __SYSCALL(__NR_pidfd_open, sys_pidfd_open)
 #define __NR_clone3 435
 __SYSCALL(__NR_clone3, sys_clone3)
 #endif
+#define __NR_close_range 436
+__SYSCALL(__NR_close_range, sys_close_range)
 
 #define __NR_openat2 437
 __SYSCALL(__NR_openat2, sys_openat2)
linux-headers/asm-mips/unistd_n32.h +1 -0
index 3b9eda7e7d8f..246fbb6a7885 100644
@@ -365,6 +365,7 @@
 #define __NR_fspick	(__NR_Linux + 433)
 #define __NR_pidfd_open	(__NR_Linux + 434)
 #define __NR_clone3	(__NR_Linux + 435)
+#define __NR_close_range	(__NR_Linux + 436)
 #define __NR_openat2	(__NR_Linux + 437)
 #define __NR_pidfd_getfd	(__NR_Linux + 438)
 #define __NR_faccessat2	(__NR_Linux + 439)
linux-headers/asm-mips/unistd_n64.h +1 -0
index 9cdf9b6c60df..194d777dfd42 100644
@@ -341,6 +341,7 @@
 #define __NR_fspick	(__NR_Linux + 433)
 #define __NR_pidfd_open	(__NR_Linux + 434)
 #define __NR_clone3	(__NR_Linux + 435)
+#define __NR_close_range	(__NR_Linux + 436)
 #define __NR_openat2	(__NR_Linux + 437)
 #define __NR_pidfd_getfd	(__NR_Linux + 438)
 #define __NR_faccessat2	(__NR_Linux + 439)
linux-headers/asm-mips/unistd_o32.h +1 -0
index e3e5e238f026..3e093dd9134d 100644
@@ -411,6 +411,7 @@
 #define __NR_fspick	(__NR_Linux + 433)
 #define __NR_pidfd_open	(__NR_Linux + 434)
 #define __NR_clone3	(__NR_Linux + 435)
+#define __NR_close_range	(__NR_Linux + 436)
 #define __NR_openat2	(__NR_Linux + 437)
 #define __NR_pidfd_getfd	(__NR_Linux + 438)
 #define __NR_faccessat2	(__NR_Linux + 439)
linux-headers/asm-powerpc/kvm.h +5 -0
index 264e266a85bf..c3af3f324c5a 100644
@@ -640,6 +640,11 @@ struct kvm_ppc_cpu_char {
 #define KVM_REG_PPC_ONLINE	(KVM_REG_PPC | KVM_REG_SIZE_U32 | 0xbf)
 #define KVM_REG_PPC_PTCR	(KVM_REG_PPC | KVM_REG_SIZE_U64 | 0xc0)
 
+/* POWER10 registers */
+#define KVM_REG_PPC_MMCR3	(KVM_REG_PPC | KVM_REG_SIZE_U64 | 0xc1)
+#define KVM_REG_PPC_SIER2	(KVM_REG_PPC | KVM_REG_SIZE_U64 | 0xc2)
+#define KVM_REG_PPC_SIER3	(KVM_REG_PPC | KVM_REG_SIZE_U64 | 0xc3)
+
 /* Transactional Memory checkpointed state:
  * This is all GPRs, all VSX regs and a subset of SPRs
  */
linux-headers/asm-powerpc/mman.h +1 -1
index 8db7c2a3be30..c16840d33bfd 100644
@@ -11,7 +11,7 @@
 #include <asm-generic/mman-common.h>
 
 
-#define PROT_SAO	0x10		/* Strong Access Ordering */
+#define PROT_SAO	0x10		/* Unsupported since v5.9 */
 
 #define MAP_RENAME      MAP_ANONYMOUS   /* In SunOS terminology */
 #define MAP_NORESERVE   0x40            /* don't reserve swap pages */
linux-headers/asm-powerpc/unistd_32.h +1 -0
index 862edb7448c5..0db9481d4962 100644
@@ -418,6 +418,7 @@
 #define __NR_fspick	433
 #define __NR_pidfd_open	434
 #define __NR_clone3	435
+#define __NR_close_range	436
 #define __NR_openat2	437
 #define __NR_pidfd_getfd	438
 #define __NR_faccessat2	439
linux-headers/asm-powerpc/unistd_64.h +1 -0
index f553224ce408..9f74310988e1 100644
@@ -390,6 +390,7 @@
 #define __NR_fspick	433
 #define __NR_pidfd_open	434
 #define __NR_clone3	435
+#define __NR_close_range	436
 #define __NR_openat2	437
 #define __NR_pidfd_getfd	438
 #define __NR_faccessat2	439
linux-headers/asm-s390/kvm.h +5 -2
index 0138ccb0d892..f053b8304a85 100644
@@ -231,11 +231,13 @@ struct kvm_guest_debug_arch {
 #define KVM_SYNC_GSCB   (1UL << 9)
 #define KVM_SYNC_BPBC   (1UL << 10)
 #define KVM_SYNC_ETOKEN (1UL << 11)
+#define KVM_SYNC_DIAG318 (1UL << 12)
 
 #define KVM_SYNC_S390_VALID_FIELDS \
 	(KVM_SYNC_PREFIX | KVM_SYNC_GPRS | KVM_SYNC_ACRS | KVM_SYNC_CRS | \
 	 KVM_SYNC_ARCH0 | KVM_SYNC_PFAULT | KVM_SYNC_VRS | KVM_SYNC_RICCB | \
-	 KVM_SYNC_FPRS | KVM_SYNC_GSCB | KVM_SYNC_BPBC | KVM_SYNC_ETOKEN)
+	 KVM_SYNC_FPRS | KVM_SYNC_GSCB | KVM_SYNC_BPBC | KVM_SYNC_ETOKEN | \
+	 KVM_SYNC_DIAG318)
 
 /* length and alignment of the sdnx as a power of two */
 #define SDNXC 8
@@ -264,7 +266,8 @@ struct kvm_sync_regs {
 	__u8 reserved2 : 7;
 	__u8 padding1[51];	/* riccb needs to be 64byte aligned */
 	__u8 riccb[64];		/* runtime instrumentation controls block */
-	__u8 padding2[192];	/* sdnx needs to be 256byte aligned */
+	__u64 diag318;		/* diagnose 0x318 info */
+	__u8 padding2[184];	/* sdnx needs to be 256byte aligned */
 	union {
 		__u8 sdnx[SDNXL];  /* state description annex */
 		struct {
linux-headers/asm-s390/unistd_32.h +1 -0
index e08233c0c377..1803cd0c3ba6 100644
@@ -408,6 +408,7 @@
 #define __NR_fspick 433
 #define __NR_pidfd_open 434
 #define __NR_clone3 435
+#define __NR_close_range 436
 #define __NR_openat2 437
 #define __NR_pidfd_getfd 438
 #define __NR_faccessat2 439
linux-headers/asm-s390/unistd_64.h +1 -0
index 560e19ae2bb4..228d5004e5a8 100644
@@ -356,6 +356,7 @@
 #define __NR_fspick 433
 #define __NR_pidfd_open 434
 #define __NR_clone3 435
+#define __NR_close_range 436
 #define __NR_openat2 437
 #define __NR_pidfd_getfd 438
 #define __NR_faccessat2 439
linux-headers/asm-x86/unistd_32.h +1 -0
index c727981d4a3a..356c12c2dbce 100644
@@ -426,6 +426,7 @@
 #define __NR_fspick 433
 #define __NR_pidfd_open 434
 #define __NR_clone3 435
+#define __NR_close_range 436
 #define __NR_openat2 437
 #define __NR_pidfd_getfd 438
 #define __NR_faccessat2 439
linux-headers/asm-x86/unistd_64.h +1 -0
index 843fa6274584..ef70e1c7c93f 100644
@@ -348,6 +348,7 @@
 #define __NR_fspick 433
 #define __NR_pidfd_open 434
 #define __NR_clone3 435
+#define __NR_close_range 436
 #define __NR_openat2 437
 #define __NR_pidfd_getfd 438
 #define __NR_faccessat2 439
linux-headers/asm-x86/unistd_x32.h +1 -0
index 7d63d703cab4..84ae8e9f5fca 100644
@@ -301,6 +301,7 @@
 #define __NR_fspick (__X32_SYSCALL_BIT + 433)
 #define __NR_pidfd_open (__X32_SYSCALL_BIT + 434)
 #define __NR_clone3 (__X32_SYSCALL_BIT + 435)
+#define __NR_close_range (__X32_SYSCALL_BIT + 436)
 #define __NR_openat2 (__X32_SYSCALL_BIT + 437)
 #define __NR_pidfd_getfd (__X32_SYSCALL_BIT + 438)
 #define __NR_faccessat2 (__X32_SYSCALL_BIT + 439)
linux-headers/linux/kvm.h +6 -1
index a28c3667370b..b4bd05d09135 100644
@@ -289,6 +289,7 @@ struct kvm_run {
 		/* KVM_EXIT_FAIL_ENTRY */
 		struct {
 			__u64 hardware_entry_failure_reason;
+			__u32 cpu;
 		} fail_entry;
 		/* KVM_EXIT_EXCEPTION */
 		struct {
@@ -1031,6 +1032,10 @@ struct kvm_ppc_resize_hpt {
 #define KVM_CAP_PPC_SECURE_GUEST 181
 #define KVM_CAP_HALT_POLL 182
 #define KVM_CAP_ASYNC_PF_INT 183
+#define KVM_CAP_LAST_CPU 184
+#define KVM_CAP_SMALLER_MAXPHYADDR 185
+#define KVM_CAP_S390_DIAG318 186
+#define KVM_CAP_SYS_HYPERV_CPUID 187
 
 #ifdef KVM_CAP_IRQ_ROUTING
 
@@ -1489,7 +1494,7 @@ struct kvm_enc_region {
 /* Available with KVM_CAP_MANUAL_DIRTY_LOG_PROTECT_2 */
 #define KVM_CLEAR_DIRTY_LOG          _IOWR(KVMIO, 0xc0, struct kvm_clear_dirty_log)
 
-/* Available with KVM_CAP_HYPERV_CPUID */
+/* Available with KVM_CAP_HYPERV_CPUID (vcpu) / KVM_CAP_SYS_HYPERV_CPUID (system) */
 #define KVM_GET_SUPPORTED_HV_CPUID _IOWR(KVMIO, 0xc1, struct kvm_cpuid2)
 
 /* Available with KVM_CAP_ARM_SVE */
linux-headers/linux/vfio.h +1 -1
index f09df262c4b5..a90672494dc5 100644
@@ -1030,7 +1030,7 @@ struct vfio_iommu_type1_info_cap_iova_range {
  * size in bytes that can be used by user applications when getting the dirty
  * bitmap.
  */
-#define VFIO_IOMMU_TYPE1_INFO_CAP_MIGRATION  1
+#define VFIO_IOMMU_TYPE1_INFO_CAP_MIGRATION  2
 
 struct vfio_iommu_type1_info_cap_migration {
 	struct	vfio_info_cap_header header;
Signature · 1 line
2.25.4
#2 lore raw

RFC PATCH 2/22 i386: drop x86_cpu_get_supported_feature_word() forward declaration

In reply to Vitaly Kuznetsov

We only use x86_cpu_get_supported_feature_word() after its implementation, no forward declaration needed.

---
 target/i386/cpu.c | 3 ---
 1 file changed, 3 deletions(-)
target/i386/cpu.c +0 -3
index 49d89585288d..14489def2177 100644
@@ -4283,9 +4283,6 @@ void x86_cpu_change_kvm_default(const char *prop, const char *value)
     assert(pv->prop);
 }
 
-static uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
-                                                   bool migratable_only);
-
 static bool lmce_supported(void)
 {
     uint64_t mce_cap = 0;
Signature · 1 line
2.25.4
#3 lore raw

RFC PATCH 3/22 i386: move hyperv_vendor_id initialization to x86_cpu_realizefn()

In reply to Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_vendor_id initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c | 15 ++++++++++++++-
 target/i386/cpu.h |  3 ++-
 target/i386/kvm.c | 25 ++++++++++---------------
 3 files changed, 26 insertions(+), 17 deletions(-)
target/i386/cpu.c +14 -1
index 14489def2177..07e9da9e567e 100644
@@ -6625,6 +6625,19 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
         }
     }
 
+    if (!cpu->hyperv_vendor) {
+        memcpy(cpu->hyperv_vendor_id, "Microsoft Hv", 12);
+    } else {
+        size_t len = strlen(cpu->hyperv_vendor);
+
+        if (len > 12) {
+            warn_report("hv-vendor-id truncated to 12 characters");
+            len = 12;
+        }
+        memset(cpu->hyperv_vendor_id, 0, 12);
+        memcpy(cpu->hyperv_vendor_id, cpu->hyperv_vendor, len);
+    }
+
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
@@ -7313,7 +7326,7 @@ static Property x86_cpu_properties[] = {
     DEFINE_PROP_UINT32("min-xlevel2", X86CPU, env.cpuid_min_xlevel2, 0),
     DEFINE_PROP_UINT64("ucode-rev", X86CPU, ucode_rev, 0),
     DEFINE_PROP_BOOL("full-cpuid-auto-level", X86CPU, full_cpuid_auto_level, true),
-    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor_id),
+    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor),
     DEFINE_PROP_BOOL("cpuid-0xb", X86CPU, enable_cpuid_0xb, true),
     DEFINE_PROP_BOOL("lmce", X86CPU, enable_lmce, false),
     DEFINE_PROP_BOOL("l3-cache", X86CPU, enable_l3_cache, true),
target/i386/cpu.h +2 -1
index d3097be6a50a..903994818093 100644
@@ -1654,11 +1654,12 @@ struct X86CPU {
     uint64_t ucode_rev;
 
     uint32_t hyperv_spinlock_attempts;
-    char *hyperv_vendor_id;
+    char *hyperv_vendor;
     bool hyperv_synic_kvm_only;
     uint64_t hyperv_features;
     bool hyperv_passthrough;
     OnOffAuto hyperv_no_nonarch_cs;
+    uint32_t hyperv_vendor_id[3];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +10 -15
index 205b68bc0ce8..47779c5e1efd 100644
@@ -1225,6 +1225,13 @@ static int hyperv_handle_properties(CPUState *cs,
         memcpy(cpuid_ent, &cpuid->entries[0],
                cpuid->nent * sizeof(cpuid->entries[0]));
 
+        c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
+        if (c) {
+            cpu->hyperv_vendor_id[0] = c->ebx;
+            cpu->hyperv_vendor_id[1] = c->ecx;
+            cpu->hyperv_vendor_id[2] = c->edx;
+        }
+
         c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
@@ -1299,23 +1306,11 @@ static int hyperv_handle_properties(CPUState *cs,
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
-    if (!cpu->hyperv_vendor_id) {
-        memcpy(signature, "Microsoft Hv", 12);
-    } else {
-        size_t len = strlen(cpu->hyperv_vendor_id);
-
-        if (len > 12) {
-            error_report("hv-vendor-id truncated to 12 characters");
-            len = 12;
-        }
-        memset(signature, 0, 12);
-        memcpy(signature, cpu->hyperv_vendor_id, len);
-    }
     c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
         HV_CPUID_NESTED_FEATURES : HV_CPUID_IMPLEMENT_LIMITS;
-    c->ebx = signature[0];
-    c->ecx = signature[1];
-    c->edx = signature[2];
+    c->ebx = cpu->hyperv_vendor_id[0];
+    c->ecx = cpu->hyperv_vendor_id[1];
+    c->edx = cpu->hyperv_vendor_id[2];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_INTERFACE;
Signature · 1 line
2.25.4
#4 lore raw

Re: [PATCH RFC 03/22] i386: move hyperv_vendor_id initialization to x86_cpu_realizefn()

Quoted 21 lines · Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_vendor_id initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c | 15 ++++++++++++++-
 target/i386/cpu.h |  3 ++-
 target/i386/kvm.c | 25 ++++++++++---------------
 3 files changed, 26 insertions(+), 17 deletions(-)
target/i386/cpu.c +3 -0
index 14489def2177..07e9da9e567e 100644
@@ -6625,6 +6625,19 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
         }
     }
 
+    if (!cpu->hyperv_vendor) {
+        memcpy(cpu->hyperv_vendor_id, "Microsoft Hv", 12);
+    } else {

target/i386/cpu.c static void x86_cpu_realizefn(DeviceState *dev, Error **errp) · line 6640

Quoted 10 lines
target/i386/cpu.c +10 -0
+        size_t len = strlen(cpu->hyperv_vendor);
+
+        if (len > 12) {
+            warn_report("hv-vendor-id truncated to 12 characters");
+            len = 12;
+        }
+        memset(cpu->hyperv_vendor_id, 0, 12);
+        memcpy(cpu->hyperv_vendor_id, cpu->hyperv_vendor, len);
+    }
+

The change makes sense, but considering that we'll have a lot of new code added to x86_cpu_realizefn(), I would prefer to create a separate x86_cpu_hyperv_realize() function to make x86_cpu_realizefn() a bit more readable.

Quoted 35 lines
target/i386/cpu.c +1 -1
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
@@ -7313,7 +7326,7 @@ static Property x86_cpu_properties[] = {
     DEFINE_PROP_UINT32("min-xlevel2", X86CPU, env.cpuid_min_xlevel2, 0),
     DEFINE_PROP_UINT64("ucode-rev", X86CPU, ucode_rev, 0),
     DEFINE_PROP_BOOL("full-cpuid-auto-level", X86CPU, full_cpuid_auto_level, true),
-    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor_id),
+    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor),
     DEFINE_PROP_BOOL("cpuid-0xb", X86CPU, enable_cpuid_0xb, true),
     DEFINE_PROP_BOOL("lmce", X86CPU, enable_lmce, false),
     DEFINE_PROP_BOOL("l3-cache", X86CPU, enable_l3_cache, true),
target/i386/cpu.h +2 -1
index d3097be6a50a..903994818093 100644
@@ -1654,11 +1654,12 @@ struct X86CPU {
     uint64_t ucode_rev;
 
     uint32_t hyperv_spinlock_attempts;
-    char *hyperv_vendor_id;
+    char *hyperv_vendor;
     bool hyperv_synic_kvm_only;
     uint64_t hyperv_features;
     bool hyperv_passthrough;
     OnOffAuto hyperv_no_nonarch_cs;
+    uint32_t hyperv_vendor_id[3];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +0 -0
index 205b68bc0ce8..47779c5e1efd 100644

target/i386/kvm.c static int hyperv_handle_properties(CPUState *cs, · line 1234

Quoted 10 lines
target/i386/kvm.c +7 -0
@@ -1225,6 +1225,13 @@ static int hyperv_handle_properties(CPUState *cs,
         memcpy(cpuid_ent, &cpuid->entries[0],
                cpuid->nent * sizeof(cpuid->entries[0]));
 
+        c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
+        if (c) {
+            cpu->hyperv_vendor_id[0] = c->ebx;
+            cpu->hyperv_vendor_id[1] = c->ecx;
+            cpu->hyperv_vendor_id[2] = c->edx;
+        }
+

I can't find the equivalent of this code in the current tree? Is hyperv vendor ID broken when using hv-passthrough today?

Maybe this could be done as a separate patch, as it changes behavior of hv-passthrough?

Quoted 33 lines
target/i386/kvm.c +3 -15
         c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
@@ -1299,23 +1306,11 @@ static int hyperv_handle_properties(CPUState *cs,
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
-    if (!cpu->hyperv_vendor_id) {
-        memcpy(signature, "Microsoft Hv", 12);
-    } else {
-        size_t len = strlen(cpu->hyperv_vendor_id);
-
-        if (len > 12) {
-            error_report("hv-vendor-id truncated to 12 characters");
-            len = 12;
-        }
-        memset(signature, 0, 12);
-        memcpy(signature, cpu->hyperv_vendor_id, len);
-    }
     c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
         HV_CPUID_NESTED_FEATURES : HV_CPUID_IMPLEMENT_LIMITS;
-    c->ebx = signature[0];
-    c->ecx = signature[1];
-    c->edx = signature[2];
+    c->ebx = cpu->hyperv_vendor_id[0];
+    c->ecx = cpu->hyperv_vendor_id[1];
+    c->edx = cpu->hyperv_vendor_id[2];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_INTERFACE;
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#5 lore raw

Re: [PATCH RFC 03/22] i386: move hyperv_vendor_id initialization to x86_cpu_realizefn()

Quoted 29 lines · Eduardo Habkost
Quoted 28 lines · Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_vendor_id initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c | 15 ++++++++++++++-
 target/i386/cpu.h |  3 ++-
 target/i386/kvm.c | 25 ++++++++++---------------
 3 files changed, 26 insertions(+), 17 deletions(-)
target/i386/cpu.c +10 -0
index 14489def2177..07e9da9e567e 100644
@@ -6625,6 +6625,19 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
         }
     }
 
+    if (!cpu->hyperv_vendor) {
+        memcpy(cpu->hyperv_vendor_id, "Microsoft Hv", 12);
+    } else {
+        size_t len = strlen(cpu->hyperv_vendor);
+
+        if (len > 12) {
+            warn_report("hv-vendor-id truncated to 12 characters");
+            len = 12;
+        }
+        memset(cpu->hyperv_vendor_id, 0, 12);
Quoted 9 lines

target/i386/cpu.c static void x86_cpu_realizefn(DeviceState *dev, Error **errp) · line 6640

Quoted 3 lines
target/i386/cpu.c +3 -0
+        memcpy(cpu->hyperv_vendor_id, cpu->hyperv_vendor, len);
+    }
+

The change makes sense, but considering that we'll have a lot of new code added to x86_cpu_realizefn(), I would prefer to create a separate x86_cpu_hyperv_realize() function to make x86_cpu_realizefn() a bit more readable.

Agreed.

Quoted 50 lines
Quoted 36 lines
target/i386/cpu.c +1 -1
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
@@ -7313,7 +7326,7 @@ static Property x86_cpu_properties[] = {
     DEFINE_PROP_UINT32("min-xlevel2", X86CPU, env.cpuid_min_xlevel2, 0),
     DEFINE_PROP_UINT64("ucode-rev", X86CPU, ucode_rev, 0),
     DEFINE_PROP_BOOL("full-cpuid-auto-level", X86CPU, full_cpuid_auto_level, true),
-    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor_id),
+    DEFINE_PROP_STRING("hv-vendor-id", X86CPU, hyperv_vendor),
     DEFINE_PROP_BOOL("cpuid-0xb", X86CPU, enable_cpuid_0xb, true),
     DEFINE_PROP_BOOL("lmce", X86CPU, enable_lmce, false),
     DEFINE_PROP_BOOL("l3-cache", X86CPU, enable_l3_cache, true),
target/i386/cpu.h +2 -1
index d3097be6a50a..903994818093 100644
@@ -1654,11 +1654,12 @@ struct X86CPU {
     uint64_t ucode_rev;
 
     uint32_t hyperv_spinlock_attempts;
-    char *hyperv_vendor_id;
+    char *hyperv_vendor;
     bool hyperv_synic_kvm_only;
     uint64_t hyperv_features;
     bool hyperv_passthrough;
     OnOffAuto hyperv_no_nonarch_cs;
+    uint32_t hyperv_vendor_id[3];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +0 -0
index 205b68bc0ce8..47779c5e1efd 100644

target/i386/kvm.c static int hyperv_handle_properties(CPUState *cs, · line 1234

Quoted 10 lines
target/i386/kvm.c +7 -0
@@ -1225,6 +1225,13 @@ static int hyperv_handle_properties(CPUState *cs,
         memcpy(cpuid_ent, &cpuid->entries[0],
                cpuid->nent * sizeof(cpuid->entries[0]));
 
+        c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
+        if (c) {
+            cpu->hyperv_vendor_id[0] = c->ebx;
+            cpu->hyperv_vendor_id[1] = c->ecx;
+            cpu->hyperv_vendor_id[2] = c->edx;
+        }
+

I can't find the equivalent of this code in the current tree? Is hyperv vendor ID broken when using hv-passthrough today?

Quoted 2 lines

Maybe this could be done as a separate patch, as it changes behavior of hv-passthrough?

(this and similar changes in other patches) Actually we don't change anything. Before this series and with hv-passthrough we just don't reflect host's CPUIDs in our internal QEMU structures so e.g. X86CPU->hyperv_vendor remains 'Microsoft Hv' while in reality guest sees what kernel told us ("Linux KVM Hv" BTW). We just copy everything we get from KVM_GET_SUPPORTED_HV_CPUID into guest's CPUIDs. This is fine as we didn't actually need the information in QEMU but to achieve the goal of the series we need to keep proper in-QEMU representation.

The real change is that post-series QEMU is not enabling any Hyper-V features which it doesn't know about while pre-series it was actually doing this. This is arguably a good change: enabling new features may require some additional work (e.g. enabling capabilities in KVM) and without it just passing CPUID feature bits the guest may get confused.

Quoted 34 lines
Quoted 34 lines
target/i386/kvm.c +3 -15
         c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
@@ -1299,23 +1306,11 @@ static int hyperv_handle_properties(CPUState *cs,
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
-    if (!cpu->hyperv_vendor_id) {
-        memcpy(signature, "Microsoft Hv", 12);
-    } else {
-        size_t len = strlen(cpu->hyperv_vendor_id);
-
-        if (len > 12) {
-            error_report("hv-vendor-id truncated to 12 characters");
-            len = 12;
-        }
-        memset(signature, 0, 12);
-        memcpy(signature, cpu->hyperv_vendor_id, len);
-    }
     c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
         HV_CPUID_NESTED_FEATURES : HV_CPUID_IMPLEMENT_LIMITS;
-    c->ebx = signature[0];
-    c->ecx = signature[1];
-    c->edx = signature[2];
+    c->ebx = cpu->hyperv_vendor_id[0];
+    c->ecx = cpu->hyperv_vendor_id[1];
+    c->edx = cpu->hyperv_vendor_id[2];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_INTERFACE;
Signature · 1 line
2.25.4
Signature · 1 line
Vitaly
#6 lore raw

RFC PATCH 4/22 i386: move hyperv_interface_id initialization to x86_cpu_realizefn()

In reply to Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_interface_id initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c |  6 ++++++
 target/i386/cpu.h |  1 +
 target/i386/kvm.c | 18 ++++++++++++------
 3 files changed, 19 insertions(+), 6 deletions(-)
target/i386/cpu.c +6 -0
index 07e9da9e567e..16888125a30a 100644
@@ -6638,6 +6638,12 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
         memcpy(cpu->hyperv_vendor_id, cpu->hyperv_vendor, len);
     }
 
+    /* 'Hv#1' interface identification*/
+    cpu->hyperv_interface_id[0] = 0x31237648;
+    cpu->hyperv_interface_id[1] = 0;
+    cpu->hyperv_interface_id[2] = 0;
+    cpu->hyperv_interface_id[3] = 0;
+
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
target/i386/cpu.h +1 -0
index 903994818093..91edc54a268c 100644
@@ -1660,6 +1660,7 @@ struct X86CPU {
     bool hyperv_passthrough;
     OnOffAuto hyperv_no_nonarch_cs;
     uint32_t hyperv_vendor_id[3];
+    uint32_t hyperv_interface_id[4];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +12 -6
index 47779c5e1efd..a36c65100cd0 100644
@@ -1188,7 +1188,6 @@ static int hyperv_handle_properties(CPUState *cs,
     CPUX86State *env = &cpu->env;
     struct kvm_cpuid2 *cpuid;
     struct kvm_cpuid_entry2 *c;
-    uint32_t signature[3];
     uint32_t cpuid_i = 0;
     int r;
 
@@ -1232,6 +1231,14 @@ static int hyperv_handle_properties(CPUState *cs,
             cpu->hyperv_vendor_id[2] = c->edx;
         }
 
+        c = cpuid_find_entry(cpuid, HV_CPUID_INTERFACE, 0);
+        if (c) {
+            cpu->hyperv_interface_id[0] = c->eax;
+            cpu->hyperv_interface_id[1] = c->ebx;
+            cpu->hyperv_interface_id[2] = c->ecx;
+            cpu->hyperv_interface_id[3] = c->edx;
+        }
+
         c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
@@ -1314,11 +1321,10 @@ static int hyperv_handle_properties(CPUState *cs,
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_INTERFACE;
-    memcpy(signature, "Hv#1\0\0\0\0\0\0\0\0", 12);
-    c->eax = signature[0];
-    c->ebx = 0;
-    c->ecx = 0;
-    c->edx = 0;
+    c->eax = cpu->hyperv_interface_id[0];
+    c->ebx = cpu->hyperv_interface_id[1];
+    c->ecx = cpu->hyperv_interface_id[2];
+    c->edx = cpu->hyperv_interface_id[3];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VERSION;
Signature · 1 line
2.25.4
#7 lore raw

Re: [PATCH RFC 04/22] i386: move hyperv_interface_id initialization to x86_cpu_realizefn()

Quoted 53 lines · Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_interface_id initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c |  6 ++++++
 target/i386/cpu.h |  1 +
 target/i386/kvm.c | 18 ++++++++++++------
 3 files changed, 19 insertions(+), 6 deletions(-)
target/i386/cpu.c +6 -0
index 07e9da9e567e..16888125a30a 100644
@@ -6638,6 +6638,12 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
         memcpy(cpu->hyperv_vendor_id, cpu->hyperv_vendor, len);
     }
 
+    /* 'Hv#1' interface identification*/
+    cpu->hyperv_interface_id[0] = 0x31237648;
+    cpu->hyperv_interface_id[1] = 0;
+    cpu->hyperv_interface_id[2] = 0;
+    cpu->hyperv_interface_id[3] = 0;
+
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
target/i386/cpu.h +1 -0
index 903994818093..91edc54a268c 100644
@@ -1660,6 +1660,7 @@ struct X86CPU {
     bool hyperv_passthrough;
     OnOffAuto hyperv_no_nonarch_cs;
     uint32_t hyperv_vendor_id[3];
+    uint32_t hyperv_interface_id[4];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +0 -1
index 47779c5e1efd..a36c65100cd0 100644
@@ -1188,7 +1188,6 @@ static int hyperv_handle_properties(CPUState *cs,
     CPUX86State *env = &cpu->env;
     struct kvm_cpuid2 *cpuid;
     struct kvm_cpuid_entry2 *c;
-    uint32_t signature[3];
     uint32_t cpuid_i = 0;
     int r;
 
@@ -1232,6 +1231,14 @@ static int hyperv_handle_properties(CPUState *cs,
             cpu->hyperv_vendor_id[2] = c->edx;

target/i386/kvm.c static int hyperv_handle_properties(CPUState *cs, · line 1241

Quoted 10 lines
target/i386/kvm.c +8 -0
         }
 
+        c = cpuid_find_entry(cpuid, HV_CPUID_INTERFACE, 0);
+        if (c) {
+            cpu->hyperv_interface_id[0] = c->eax;
+            cpu->hyperv_interface_id[1] = c->ebx;
+            cpu->hyperv_interface_id[2] = c->ecx;
+            cpu->hyperv_interface_id[3] = c->edx;
+        }
+

Same question I sent for vendor id, version id, hyperv_limits:
should this go to a separate patch, as it changes behavior of hv-passthrough?

Quoted 22 lines
target/i386/kvm.c +4 -5
         c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
@@ -1314,11 +1321,10 @@ static int hyperv_handle_properties(CPUState *cs,
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_INTERFACE;
-    memcpy(signature, "Hv#1\0\0\0\0\0\0\0\0", 12);
-    c->eax = signature[0];
-    c->ebx = 0;
-    c->ecx = 0;
-    c->edx = 0;
+    c->eax = cpu->hyperv_interface_id[0];
+    c->ebx = cpu->hyperv_interface_id[1];
+    c->ecx = cpu->hyperv_interface_id[2];
+    c->edx = cpu->hyperv_interface_id[3];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VERSION;
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#8 lore raw

RFC PATCH 5/22 i386: move hyperv_version_id initialization to x86_cpu_realizefn()

In reply to Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_version_id initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c |  4 ++++
 target/i386/cpu.h |  1 +
 target/i386/kvm.c | 14 ++++++++++++--
 3 files changed, 17 insertions(+), 2 deletions(-)
target/i386/cpu.c +4 -0
index 16888125a30a..e605399eb8c0 100644
@@ -6644,6 +6644,10 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
     cpu->hyperv_interface_id[2] = 0;
     cpu->hyperv_interface_id[3] = 0;
 
+    /* Hypervisor system identity */
+    cpu->hyperv_version_id[0] = 0x00001bbc;
+    cpu->hyperv_version_id[1] = 0x00060001;
+
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
target/i386/cpu.h +1 -0
index 91edc54a268c..2630ffd2d4b2 100644
@@ -1661,6 +1661,7 @@ struct X86CPU {
     OnOffAuto hyperv_no_nonarch_cs;
     uint32_t hyperv_vendor_id[3];
     uint32_t hyperv_interface_id[4];
+    uint32_t hyperv_version_id[4];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +12 -2
index a36c65100cd0..169bae2779a4 100644
@@ -1239,6 +1239,14 @@ static int hyperv_handle_properties(CPUState *cs,
             cpu->hyperv_interface_id[3] = c->edx;
         }
 
+        c = cpuid_find_entry(cpuid, HV_CPUID_VERSION, 0);
+        if (c) {
+            cpu->hyperv_version_id[0] = c->eax;
+            cpu->hyperv_version_id[1] = c->ebx;
+            cpu->hyperv_version_id[2] = c->ecx;
+            cpu->hyperv_version_id[3] = c->edx;
+        }
+
         c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
@@ -1328,8 +1336,10 @@ static int hyperv_handle_properties(CPUState *cs,
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VERSION;
-    c->eax = 0x00001bbc;
-    c->ebx = 0x00060001;
+    c->eax = cpu->hyperv_version_id[0];
+    c->ebx = cpu->hyperv_version_id[1];
+    c->ecx = cpu->hyperv_version_id[2];
+    c->edx = cpu->hyperv_version_id[3];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_FEATURES;
Signature · 1 line
2.25.4
#9 lore raw

Re: [PATCH RFC 05/22] i386: move hyperv_version_id initialization to x86_cpu_realizefn()

Quoted 43 lines · Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_version_id initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c |  4 ++++
 target/i386/cpu.h |  1 +
 target/i386/kvm.c | 14 ++++++++++++--
 3 files changed, 17 insertions(+), 2 deletions(-)
target/i386/cpu.c +4 -0
index 16888125a30a..e605399eb8c0 100644
@@ -6644,6 +6644,10 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
     cpu->hyperv_interface_id[2] = 0;
     cpu->hyperv_interface_id[3] = 0;
 
+    /* Hypervisor system identity */
+    cpu->hyperv_version_id[0] = 0x00001bbc;
+    cpu->hyperv_version_id[1] = 0x00060001;
+
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
target/i386/cpu.h +1 -0
index 91edc54a268c..2630ffd2d4b2 100644
@@ -1661,6 +1661,7 @@ struct X86CPU {
     OnOffAuto hyperv_no_nonarch_cs;
     uint32_t hyperv_vendor_id[3];
     uint32_t hyperv_interface_id[4];
+    uint32_t hyperv_version_id[4];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +0 -0
index a36c65100cd0..169bae2779a4 100644
@@ -1239,6 +1239,14 @@ static int hyperv_handle_properties(CPUState *cs,
             cpu->hyperv_interface_id[3] = c->edx;

target/i386/kvm.c static int hyperv_handle_properties(CPUState *cs, · line 1249

Quoted 10 lines
target/i386/kvm.c +8 -0
         }
 
+        c = cpuid_find_entry(cpuid, HV_CPUID_VERSION, 0);
+        if (c) {
+            cpu->hyperv_version_id[0] = c->eax;
+            cpu->hyperv_version_id[1] = c->ebx;
+            cpu->hyperv_version_id[2] = c->ecx;
+            cpu->hyperv_version_id[3] = c->edx;
+        }
+

Same questions I had for vendor id: is hyperv_version_id broken when using hv-passthrough today? Should this go to a separate patch?

Quoted 19 lines
target/i386/kvm.c +4 -2
         c = cpuid_find_entry(cpuid, HV_CPUID_FEATURES, 0);
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
@@ -1328,8 +1336,10 @@ static int hyperv_handle_properties(CPUState *cs,
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VERSION;
-    c->eax = 0x00001bbc;
-    c->ebx = 0x00060001;
+    c->eax = cpu->hyperv_version_id[0];
+    c->ebx = cpu->hyperv_version_id[1];
+    c->ecx = cpu->hyperv_version_id[2];
+    c->edx = cpu->hyperv_version_id[3];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_FEATURES;
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#10 lore raw

RFC PATCH 6/22 i386: move hyperv_limits initialization to x86_cpu_realizefn()

In reply to Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_limits initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c |  5 +++++
 target/i386/cpu.h |  1 +
 target/i386/kvm.c | 13 ++++++++++++-
 3 files changed, 18 insertions(+), 1 deletion(-)
target/i386/cpu.c +5 -0
index e605399eb8c0..ef3c672cf415 100644
@@ -6648,6 +6648,11 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
     cpu->hyperv_version_id[0] = 0x00001bbc;
     cpu->hyperv_version_id[1] = 0x00060001;
 
+    /* Hypervisor implementation limits */
+    cpu->hyperv_limits[0] = 64;
+    cpu->hyperv_limits[1] = 0;
+    cpu->hyperv_limits[2] = 0;
+
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
target/i386/cpu.h +1 -0
index 2630ffd2d4b2..095d0bf75493 100644
@@ -1662,6 +1662,7 @@ struct X86CPU {
     uint32_t hyperv_vendor_id[3];
     uint32_t hyperv_interface_id[4];
     uint32_t hyperv_version_id[4];
+    uint32_t hyperv_limits[3];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +12 -1
index 169bae2779a4..720c30e9df17 100644
@@ -1253,6 +1253,15 @@ static int hyperv_handle_properties(CPUState *cs,
             env->features[FEAT_HYPERV_EBX] = c->ebx;
             env->features[FEAT_HYPERV_EDX] = c->eax;
         }
+
+        c = cpuid_find_entry(cpuid, HV_CPUID_IMPLEMENT_LIMITS, 0);
+        if (c) {
+            cpu->hv_max_vps = c->eax;
+            cpu->hyperv_limits[0] = c->ebx;
+            cpu->hyperv_limits[1] = c->ecx;
+            cpu->hyperv_limits[2] = c->edx;
+        }
+
         c = cpuid_find_entry(cpuid, HV_CPUID_ENLIGHTMENT_INFO, 0);
         if (c) {
             env->features[FEAT_HV_RECOMM_EAX] = c->eax;
@@ -1355,7 +1364,9 @@ static int hyperv_handle_properties(CPUState *cs,
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_IMPLEMENT_LIMITS;
     c->eax = cpu->hv_max_vps;
-    c->ebx = 0x40;
+    c->ebx = cpu->hyperv_limits[0];
+    c->ecx = cpu->hyperv_limits[1];
+    c->edx = cpu->hyperv_limits[2];
 
     if (hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS)) {
         __u32 function;
Signature · 1 line
2.25.4
#11 lore raw

Re: [PATCH RFC 06/22] i386: move hyperv_limits initialization to x86_cpu_realizefn()

Quoted 45 lines · Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, move hyperv_limits initialization to x86_cpu_realizefn().

---
 target/i386/cpu.c |  5 +++++
 target/i386/cpu.h |  1 +
 target/i386/kvm.c | 13 ++++++++++++-
 3 files changed, 18 insertions(+), 1 deletion(-)
target/i386/cpu.c +5 -0
index e605399eb8c0..ef3c672cf415 100644
@@ -6648,6 +6648,11 @@ static void x86_cpu_realizefn(DeviceState *dev, Error **errp)
     cpu->hyperv_version_id[0] = 0x00001bbc;
     cpu->hyperv_version_id[1] = 0x00060001;
 
+    /* Hypervisor implementation limits */
+    cpu->hyperv_limits[0] = 64;
+    cpu->hyperv_limits[1] = 0;
+    cpu->hyperv_limits[2] = 0;
+
     if (cpu->ucode_rev == 0) {
         /* The default is the same as KVM's.  */
         if (IS_AMD_CPU(env)) {
target/i386/cpu.h +1 -0
index 2630ffd2d4b2..095d0bf75493 100644
@@ -1662,6 +1662,7 @@ struct X86CPU {
     uint32_t hyperv_vendor_id[3];
     uint32_t hyperv_interface_id[4];
     uint32_t hyperv_version_id[4];
+    uint32_t hyperv_limits[3];
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +0 -0
index 169bae2779a4..720c30e9df17 100644
@@ -1253,6 +1253,15 @@ static int hyperv_handle_properties(CPUState *cs,
             env->features[FEAT_HYPERV_EBX] = c->ebx;
             env->features[FEAT_HYPERV_EDX] = c->eax;

target/i386/kvm.c static int hyperv_handle_properties(CPUState *cs, · line 1264

Quoted 10 lines
target/i386/kvm.c +9 -0
         }
+
+        c = cpuid_find_entry(cpuid, HV_CPUID_IMPLEMENT_LIMITS, 0);
+        if (c) {
+            cpu->hv_max_vps = c->eax;
+            cpu->hyperv_limits[0] = c->ebx;
+            cpu->hyperv_limits[1] = c->ecx;
+            cpu->hyperv_limits[2] = c->edx;
+        }
+

Sam question I had for vendor id and version id: should it go to a separate patch as it changes behavior of hv-passthrough?

Quoted 17 lines
target/i386/kvm.c +3 -1
         c = cpuid_find_entry(cpuid, HV_CPUID_ENLIGHTMENT_INFO, 0);
         if (c) {
             env->features[FEAT_HV_RECOMM_EAX] = c->eax;
@@ -1355,7 +1364,9 @@ static int hyperv_handle_properties(CPUState *cs,
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_IMPLEMENT_LIMITS;
     c->eax = cpu->hv_max_vps;
-    c->ebx = 0x40;
+    c->ebx = cpu->hyperv_limits[0];
+    c->ecx = cpu->hyperv_limits[1];
+    c->edx = cpu->hyperv_limits[2];
 
     if (hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS)) {
         __u32 function;
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#12 lore raw

RFC PATCH 7/22 i386: fill in FEAT_HYPERV_EDX from edx instead of eax

In reply to Vitaly Kuznetsov

There was a typo which went unnoticed.

Fixes e48ddcc6ce13 ("i386/kvm: implement 'hv-passthrough' mode")
---
 target/i386/kvm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
target/i386/kvm.c +1 -1
index 720c30e9df17..1cb5592d4f7d 100644
@@ -1251,7 +1251,7 @@ static int hyperv_handle_properties(CPUState *cs,
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
             env->features[FEAT_HYPERV_EBX] = c->ebx;
-            env->features[FEAT_HYPERV_EDX] = c->eax;
+            env->features[FEAT_HYPERV_EDX] = c->edx;
         }
 
         c = cpuid_find_entry(cpuid, HV_CPUID_IMPLEMENT_LIMITS, 0);
Signature · 1 line
2.25.4
#13 lore raw

Re: [PATCH RFC 07/22] i386: fill in FEAT_HYPERV_EDX from edx instead of eax

Quoted 4 lines · Vitaly Kuznetsov

There was a typo which went unnoticed.

Fixes e48ddcc6ce13 ("i386/kvm: implement 'hv-passthrough' mode")

Similar fix had been submitted at:
https://lore.kernel.org/qemu-devel/[email protected]/

Quoted 20 lines
---
 target/i386/kvm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
target/i386/kvm.c +1 -1
index 720c30e9df17..1cb5592d4f7d 100644
@@ -1251,7 +1251,7 @@ static int hyperv_handle_properties(CPUState *cs,
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
             env->features[FEAT_HYPERV_EBX] = c->ebx;
-            env->features[FEAT_HYPERV_EDX] = c->eax;
+            env->features[FEAT_HYPERV_EDX] = c->edx;
         }
 
         c = cpuid_find_entry(cpuid, HV_CPUID_IMPLEMENT_LIMITS, 0);
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#14 lore raw

RFC PATCH 8/22 i386: invert hyperv_spinlock_attempts setting logic with hv_passthrough

In reply to Vitaly Kuznetsov

There is no need to have this special case: like all other Hyper-V enlightenments we can just use kernel's supplied value in hv_passthrough mode.

---
 target/i386/kvm.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
target/i386/kvm.c +1 -5
index 1cb5592d4f7d..96ac719adca1 100644
@@ -1265,11 +1265,7 @@ static int hyperv_handle_properties(CPUState *cs,
         c = cpuid_find_entry(cpuid, HV_CPUID_ENLIGHTMENT_INFO, 0);
         if (c) {
             env->features[FEAT_HV_RECOMM_EAX] = c->eax;
-
-            /* hv-spinlocks may have been overriden */
-            if (cpu->hyperv_spinlock_attempts != HYPERV_SPINLOCK_NEVER_RETRY) {
-                c->ebx = cpu->hyperv_spinlock_attempts;
-            }
+            cpu->hyperv_spinlock_attempts = c->ebx;
         }
         c = cpuid_find_entry(cpuid, HV_CPUID_NESTED_FEATURES, 0);
         if (c) {
Signature · 1 line
2.25.4
#15 lore raw

RFC PATCH 9/22 i386: add reserved FEAT_HYPERV_ECX CPUID leaf

In reply to Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, add reserved FEAT_HYPERV_ECX CPUID leaf.

---
 target/i386/cpu.c | 5 +++++
 target/i386/cpu.h | 1 +
 target/i386/kvm.c | 2 ++
 3 files changed, 8 insertions(+)
target/i386/cpu.c +5 -0
index ef3c672cf415..70588571ccb1 100644
@@ -974,6 +974,11 @@ static FeatureWordInfo feature_word_info[FEATURE_WORDS] = {
         },
         .cpuid = { .eax = 0x40000003, .reg = R_EBX, },
     },
+    [FEAT_HYPERV_ECX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */
+        .cpuid = { .eax = 0x40000003, .reg = R_ECX, },
+    },
     [FEAT_HYPERV_EDX] = {
         .type = CPUID_FEATURE_WORD,
         .feat_names = {
target/i386/cpu.h +1 -0
index 095d0bf75493..39e0e89aa41f 100644
@@ -521,6 +521,7 @@ typedef enum FeatureWord {
     FEAT_KVM_HINTS,     /* CPUID[4000_0001].EDX */
     FEAT_HYPERV_EAX,    /* CPUID[4000_0003].EAX */
     FEAT_HYPERV_EBX,    /* CPUID[4000_0003].EBX */
+    FEAT_HYPERV_ECX,    /* CPUID[4000_0003].ECX */
     FEAT_HYPERV_EDX,    /* CPUID[4000_0003].EDX */
     FEAT_HV_RECOMM_EAX, /* CPUID[4000_0004].EAX */
     FEAT_HV_NESTED_EAX, /* CPUID[4000_000A].EAX */
target/i386/kvm.c +2 -0
index 96ac719adca1..259be2d514dd 100644
@@ -1251,6 +1251,7 @@ static int hyperv_handle_properties(CPUState *cs,
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
             env->features[FEAT_HYPERV_EBX] = c->ebx;
+            env->features[FEAT_HYPERV_ECX] = c->ecx;
             env->features[FEAT_HYPERV_EDX] = c->edx;
         }
 
@@ -1350,6 +1351,7 @@ static int hyperv_handle_properties(CPUState *cs,
     c->function = HV_CPUID_FEATURES;
     c->eax = env->features[FEAT_HYPERV_EAX];
     c->ebx = env->features[FEAT_HYPERV_EBX];
+    c->ecx = env->features[FEAT_HYPERV_ECX];
     c->edx = env->features[FEAT_HYPERV_EDX];
 
     c = &cpuid_ent[cpuid_i++];
Signature · 1 line
2.25.4
#16 lore raw

Re: [PATCH RFC 09/22] i386: add reserved FEAT_HYPERV_ECX CPUID leaf

Quoted 15 lines · Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, add reserved FEAT_HYPERV_ECX CPUID leaf.

---
 target/i386/cpu.c | 5 +++++
 target/i386/cpu.h | 1 +
 target/i386/kvm.c | 2 ++
 3 files changed, 8 insertions(+)
target/i386/cpu.c +0 -0
index ef3c672cf415..70588571ccb1 100644

target/i386/cpu.c static FeatureWordInfo feature_word_info[FEATURE_WORDS] = { · line 979

Quoted 6 lines
target/i386/cpu.c +3 -0
@@ -974,6 +974,11 @@ static FeatureWordInfo feature_word_info[FEATURE_WORDS] = {
         },
         .cpuid = { .eax = 0x40000003, .reg = R_EBX, },
     },
+    [FEAT_HYPERV_ECX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */

What does "reserved" mean here?

Quoted 40 lines
target/i386/cpu.c +2 -0
+        .cpuid = { .eax = 0x40000003, .reg = R_ECX, },
+    },
     [FEAT_HYPERV_EDX] = {
         .type = CPUID_FEATURE_WORD,
         .feat_names = {
target/i386/cpu.h +1 -0
index 095d0bf75493..39e0e89aa41f 100644
@@ -521,6 +521,7 @@ typedef enum FeatureWord {
     FEAT_KVM_HINTS,     /* CPUID[4000_0001].EDX */
     FEAT_HYPERV_EAX,    /* CPUID[4000_0003].EAX */
     FEAT_HYPERV_EBX,    /* CPUID[4000_0003].EBX */
+    FEAT_HYPERV_ECX,    /* CPUID[4000_0003].ECX */
     FEAT_HYPERV_EDX,    /* CPUID[4000_0003].EDX */
     FEAT_HV_RECOMM_EAX, /* CPUID[4000_0004].EAX */
     FEAT_HV_NESTED_EAX, /* CPUID[4000_000A].EAX */
target/i386/kvm.c +2 -0
index 96ac719adca1..259be2d514dd 100644
@@ -1251,6 +1251,7 @@ static int hyperv_handle_properties(CPUState *cs,
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
             env->features[FEAT_HYPERV_EBX] = c->ebx;
+            env->features[FEAT_HYPERV_ECX] = c->ecx;
             env->features[FEAT_HYPERV_EDX] = c->edx;
         }
 
@@ -1350,6 +1351,7 @@ static int hyperv_handle_properties(CPUState *cs,
     c->function = HV_CPUID_FEATURES;
     c->eax = env->features[FEAT_HYPERV_EAX];
     c->ebx = env->features[FEAT_HYPERV_EBX];
+    c->ecx = env->features[FEAT_HYPERV_ECX];
     c->edx = env->features[FEAT_HYPERV_EDX];
 
     c = &cpuid_ent[cpuid_i++];
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#17 lore raw

Re: [PATCH RFC 09/22] i386: add reserved FEAT_HYPERV_ECX CPUID leaf

Quoted 11 lines · Eduardo Habkost
Quoted 10 lines · Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, add reserved FEAT_HYPERV_ECX CPUID leaf.

---
 target/i386/cpu.c | 5 +++++
 target/i386/cpu.h | 1 +
 target/i386/kvm.c | 2 ++
 3 files changed, 8 insertions(+)
Quoted 13 lines

target/i386/cpu.c static FeatureWordInfo feature_word_info[FEATURE_WORDS] = { · line 979

Quoted 11 lines
target/i386/cpu.c +3 -0
index ef3c672cf415..70588571ccb1 100644
@@ -974,6 +974,11 @@ static FeatureWordInfo feature_word_info[FEATURE_WORDS] = {
         },
         .cpuid = { .eax = 0x40000003, .reg = R_EBX, },
     },
+    [FEAT_HYPERV_ECX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */

What does "reserved" mean here?

This is actually from TLFS, these bits are currently reserved and there are no features represented by them. We can, of course, short-circuit this to '0' but I'd prefer to keep it for consistency. I can change to something like 'Reserved as of TLFS v6.0b' to make it clear.

Quoted 41 lines
Quoted 41 lines
target/i386/cpu.c +2 -0
+        .cpuid = { .eax = 0x40000003, .reg = R_ECX, },
+    },
     [FEAT_HYPERV_EDX] = {
         .type = CPUID_FEATURE_WORD,
         .feat_names = {
target/i386/cpu.h +1 -0
index 095d0bf75493..39e0e89aa41f 100644
@@ -521,6 +521,7 @@ typedef enum FeatureWord {
     FEAT_KVM_HINTS,     /* CPUID[4000_0001].EDX */
     FEAT_HYPERV_EAX,    /* CPUID[4000_0003].EAX */
     FEAT_HYPERV_EBX,    /* CPUID[4000_0003].EBX */
+    FEAT_HYPERV_ECX,    /* CPUID[4000_0003].ECX */
     FEAT_HYPERV_EDX,    /* CPUID[4000_0003].EDX */
     FEAT_HV_RECOMM_EAX, /* CPUID[4000_0004].EAX */
     FEAT_HV_NESTED_EAX, /* CPUID[4000_000A].EAX */
target/i386/kvm.c +2 -0
index 96ac719adca1..259be2d514dd 100644
@@ -1251,6 +1251,7 @@ static int hyperv_handle_properties(CPUState *cs,
         if (c) {
             env->features[FEAT_HYPERV_EAX] = c->eax;
             env->features[FEAT_HYPERV_EBX] = c->ebx;
+            env->features[FEAT_HYPERV_ECX] = c->ecx;
             env->features[FEAT_HYPERV_EDX] = c->edx;
         }
 
@@ -1350,6 +1351,7 @@ static int hyperv_handle_properties(CPUState *cs,
     c->function = HV_CPUID_FEATURES;
     c->eax = env->features[FEAT_HYPERV_EAX];
     c->ebx = env->features[FEAT_HYPERV_EBX];
+    c->ecx = env->features[FEAT_HYPERV_ECX];
     c->edx = env->features[FEAT_HYPERV_EDX];
 
     c = &cpuid_ent[cpuid_i++];
Signature · 1 line
2.25.4
Signature · 1 line
Vitaly
#18 lore raw

RFC PATCH 10/22 i386: add reserved FEAT_HV_RECOMM_ECX/FEAT_HV_RECOMM_EDX CPUID leaves

In reply to Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, add reserved FEAT_HV_RECOMM_ECX/FEAT_HV_RECOMM_EDX CPUID leaves.

---
 target/i386/cpu.c | 10 ++++++++++
 target/i386/cpu.h |  2 ++
 target/i386/kvm.c |  5 +++++
 3 files changed, 17 insertions(+)
target/i386/cpu.c +10 -0
index 70588571ccb1..3665f22093e9 100644
@@ -1021,6 +1021,16 @@ static FeatureWordInfo feature_word_info[FEATURE_WORDS] = {
         },
         .cpuid = { .eax = 0x40000004, .reg = R_EAX, },
     },
+    [FEAT_HV_RECOMM_ECX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */
+        .cpuid = { .eax = 0x40000004, .reg = R_ECX, },
+    },
+    [FEAT_HV_RECOMM_EDX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */
+        .cpuid = { .eax = 0x40000004, .reg = R_EDX, },
+    },
     [FEAT_HV_NESTED_EAX] = {
         .type = CPUID_FEATURE_WORD,
         .cpuid = { .eax = 0x4000000A, .reg = R_EAX, },
target/i386/cpu.h +2 -0
index 39e0e89aa41f..0668fe74df4f 100644
@@ -524,6 +524,8 @@ typedef enum FeatureWord {
     FEAT_HYPERV_ECX,    /* CPUID[4000_0003].ECX */
     FEAT_HYPERV_EDX,    /* CPUID[4000_0003].EDX */
     FEAT_HV_RECOMM_EAX, /* CPUID[4000_0004].EAX */
+    FEAT_HV_RECOMM_ECX, /* CPUID[4000_0004].ECX */
+    FEAT_HV_RECOMM_EDX, /* CPUID[4000_0004].EDX */
     FEAT_HV_NESTED_EAX, /* CPUID[4000_000A].EAX */
     FEAT_SVM,           /* CPUID[8000_000A].EDX */
     FEAT_XSAVE,         /* CPUID[EAX=0xd,ECX=1].EAX */
target/i386/kvm.c +5 -0
index 259be2d514dd..c18f221cd301 100644
@@ -1267,7 +1267,10 @@ static int hyperv_handle_properties(CPUState *cs,
         if (c) {
             env->features[FEAT_HV_RECOMM_EAX] = c->eax;
             cpu->hyperv_spinlock_attempts = c->ebx;
+            env->features[FEAT_HV_RECOMM_ECX] = c->ecx;
+            env->features[FEAT_HV_RECOMM_EDX] = c->edx;
         }
+
         c = cpuid_find_entry(cpuid, HV_CPUID_NESTED_FEATURES, 0);
         if (c) {
             env->features[FEAT_HV_NESTED_EAX] = c->eax;
@@ -1358,6 +1361,8 @@ static int hyperv_handle_properties(CPUState *cs,
     c->function = HV_CPUID_ENLIGHTMENT_INFO;
     c->eax = env->features[FEAT_HV_RECOMM_EAX];
     c->ebx = cpu->hyperv_spinlock_attempts;
+    c->ecx = env->features[FEAT_HV_RECOMM_ECX];
+    c->edx = env->features[FEAT_HV_RECOMM_EDX];
 
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_IMPLEMENT_LIMITS;
Signature · 1 line
2.25.4
#19 lore raw

RFC PATCH 11/22 i386: add reserved FEAT_HV_NESTED_EBX/ECX/EDX CPUID leaves

In reply to Vitaly Kuznetsov

As a preparation to expanding Hyper-V CPU features early, add reserved FEAT_HV_NESTED_EBX/ECX/EDX CPUID leaves.

---
 target/i386/cpu.c | 15 +++++++++++++++
 target/i386/cpu.h |  3 +++
 target/i386/kvm.c |  6 ++++++
 3 files changed, 24 insertions(+)
target/i386/cpu.c +15 -0
index 3665f22093e9..479c4bbbf459 100644
@@ -1035,6 +1035,21 @@ static FeatureWordInfo feature_word_info[FEATURE_WORDS] = {
         .type = CPUID_FEATURE_WORD,
         .cpuid = { .eax = 0x4000000A, .reg = R_EAX, },
     },
+    [FEAT_HV_NESTED_EBX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */
+        .cpuid = { .eax = 0x4000000A, .reg = R_EBX, },
+    },
+    [FEAT_HV_NESTED_ECX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */
+        .cpuid = { .eax = 0x4000000A, .reg = R_ECX, },
+    },
+    [FEAT_HV_NESTED_EDX] = {
+        .type = CPUID_FEATURE_WORD,
+        /* reserved */
+        .cpuid = { .eax = 0x4000000A, .reg = R_EDX, },
+    },
     [FEAT_SVM] = {
         .type = CPUID_FEATURE_WORD,
         .feat_names = {
target/i386/cpu.h +3 -0
index 0668fe74df4f..0aad60e0c707 100644
@@ -527,6 +527,9 @@ typedef enum FeatureWord {
     FEAT_HV_RECOMM_ECX, /* CPUID[4000_0004].ECX */
     FEAT_HV_RECOMM_EDX, /* CPUID[4000_0004].EDX */
     FEAT_HV_NESTED_EAX, /* CPUID[4000_000A].EAX */
+    FEAT_HV_NESTED_EBX, /* CPUID[4000_000A].EBX */
+    FEAT_HV_NESTED_ECX, /* CPUID[4000_000A].ECX */
+    FEAT_HV_NESTED_EDX, /* CPUID[4000_000A].EDX */
     FEAT_SVM,           /* CPUID[8000_000A].EDX */
     FEAT_XSAVE,         /* CPUID[EAX=0xd,ECX=1].EAX */
     FEAT_6_EAX,         /* CPUID[6].EAX */
target/i386/kvm.c +6 -0
index c18f221cd301..87b83a2aa2cb 100644
@@ -1274,6 +1274,9 @@ static int hyperv_handle_properties(CPUState *cs,
         c = cpuid_find_entry(cpuid, HV_CPUID_NESTED_FEATURES, 0);
         if (c) {
             env->features[FEAT_HV_NESTED_EAX] = c->eax;
+            env->features[FEAT_HV_NESTED_EBX] = c->ebx;
+            env->features[FEAT_HV_NESTED_ECX] = c->ecx;
+            env->features[FEAT_HV_NESTED_EDX] = c->edx;
         }
     }
 
@@ -1384,6 +1387,9 @@ static int hyperv_handle_properties(CPUState *cs,
         c = &cpuid_ent[cpuid_i++];
         c->function = HV_CPUID_NESTED_FEATURES;
         c->eax = env->features[FEAT_HV_NESTED_EAX];
+        c->ebx = env->features[FEAT_HV_NESTED_EBX];
+        c->ecx = env->features[FEAT_HV_NESTED_ECX];
+        c->edx = env->features[FEAT_HV_NESTED_EDX];
     }
     r = cpuid_i;
 
Signature · 1 line
2.25.4
#20 lore raw

RFC PATCH 12/22 i386: always fill Hyper-V CPUID feature leaves from X86CPU data

In reply to Vitaly Kuznetsov

We have all the required data in X86CPU already and as we are about to split hyperv_handle_properties() into hyperv_expand_features()/ hyperv_fill_cpuids() we can remove the blind copy. The functional change is that QEMU won't pass CPUID leaves it doesn't currently know about to the guest but arguably this is a good change.

---
 target/i386/kvm.c | 9 ---------
 1 file changed, 9 deletions(-)
target/i386/kvm.c +0 -9
index 87b83a2aa2cb..b7f0e200a75f 100644
@@ -1221,9 +1221,6 @@ static int hyperv_handle_properties(CPUState *cs,
     }
 
     if (cpu->hyperv_passthrough) {
-        memcpy(cpuid_ent, &cpuid->entries[0],
-               cpuid->nent * sizeof(cpuid->entries[0]));
-
         c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
         if (c) {
             cpu->hyperv_vendor_id[0] = c->ebx;
@@ -1325,12 +1322,6 @@ static int hyperv_handle_properties(CPUState *cs,
         goto free;
     }
 
-    if (cpu->hyperv_passthrough) {
-        /* We already copied all feature words from KVM as is */
-        r = cpuid->nent;
-        goto free;
-    }
-
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
     c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
Signature · 1 line
2.25.4
#21 lore raw

Re: [PATCH RFC 12/22] i386: always fill Hyper-V CPUID feature leaves from X86CPU data

Quoted 5 lines · Vitaly Kuznetsov

We have all the required data in X86CPU already and as we are about to split hyperv_handle_properties() into hyperv_expand_features()/ hyperv_fill_cpuids() we can remove the blind copy. The functional change is that QEMU won't pass CPUID leaves it doesn't currently know about to the guest but arguably this is a good change.

Is it, though? Maybe we don't want it to be the default, but a blind passthrough mode that copies all CPUID leaves would be useful for testing new features.

Quoted 36 lines
---
 target/i386/kvm.c | 9 ---------
 1 file changed, 9 deletions(-)
target/i386/kvm.c +0 -9
index 87b83a2aa2cb..b7f0e200a75f 100644
@@ -1221,9 +1221,6 @@ static int hyperv_handle_properties(CPUState *cs,
     }
 
     if (cpu->hyperv_passthrough) {
-        memcpy(cpuid_ent, &cpuid->entries[0],
-               cpuid->nent * sizeof(cpuid->entries[0]));
-
         c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
         if (c) {
             cpu->hyperv_vendor_id[0] = c->ebx;
@@ -1325,12 +1322,6 @@ static int hyperv_handle_properties(CPUState *cs,
         goto free;
     }
 
-    if (cpu->hyperv_passthrough) {
-        /* We already copied all feature words from KVM as is */
-        r = cpuid->nent;
-        goto free;
-    }
-
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
     c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#22 lore raw

Re: [PATCH RFC 12/22] i386: always fill Hyper-V CPUID feature leaves from X86CPU data

Quoted 11 lines · Eduardo Habkost
Quoted 5 lines · Vitaly Kuznetsov

We have all the required data in X86CPU already and as we are about to split hyperv_handle_properties() into hyperv_expand_features()/ hyperv_fill_cpuids() we can remove the blind copy. The functional change is that QEMU won't pass CPUID leaves it doesn't currently know about to the guest but arguably this is a good change.

Is it, though? Maybe we don't want it to be the default, but a blind passthrough mode that copies all CPUID leaves would be useful for testing new features.

We already have Hyper-V features which require extra work to get enabled (SynIC, EVMCS). Passing corresponding feature bits to guests without actually enabling these features in KVM will likely break them. So a pure 'blind passthrough' mode is not possible.

Quoted 36 lines
Quoted 36 lines
---
 target/i386/kvm.c | 9 ---------
 1 file changed, 9 deletions(-)
target/i386/kvm.c +0 -9
index 87b83a2aa2cb..b7f0e200a75f 100644
@@ -1221,9 +1221,6 @@ static int hyperv_handle_properties(CPUState *cs,
     }
 
     if (cpu->hyperv_passthrough) {
-        memcpy(cpuid_ent, &cpuid->entries[0],
-               cpuid->nent * sizeof(cpuid->entries[0]));
-
         c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
         if (c) {
             cpu->hyperv_vendor_id[0] = c->ebx;
@@ -1325,12 +1322,6 @@ static int hyperv_handle_properties(CPUState *cs,
         goto free;
     }
 
-    if (cpu->hyperv_passthrough) {
-        /* We already copied all feature words from KVM as is */
-        r = cpuid->nent;
-        goto free;
-    }
-
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
     c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
Signature · 1 line
2.25.4
Signature · 1 line
Vitaly
#23 lore raw

RFC PATCH 13/22 i386: split hyperv_handle_properties() into hyperv_expand_features()/hyperv_fill_cpuids()

In reply to Vitaly Kuznetsov

The intention is to call hyperv_expand_features() early, before vCPUs are created and use the acquired data later when we set guest visible CPUID data.

---
 target/i386/kvm.c | 45 +++++++++++++++++++++++++++++----------------
 1 file changed, 29 insertions(+), 16 deletions(-)
target/i386/kvm.c +29 -16
index b7f0e200a75f..c4bb147090a9 100644
@@ -1177,18 +1177,18 @@ static int hv_cpuid_check_and_set(CPUState *cs, struct kvm_cpuid2 *cpuid,
 }
 
 /*
- * Fill in Hyper-V CPUIDs. Returns the number of entries filled in cpuid_ent in
- * case of success, errno < 0 in case of failure and 0 when no Hyper-V
- * extentions are enabled.
+ * Expand Hyper-V CPU features. In partucular, check that all the requested
+ * features are supported by the host and the sanity of the configuration
+ * (that all the required dependencies are included). Also, this takes care
+ * of 'hv_passthrough' mode and fills the environment with all supported
+ * Hyper-V features.
  */
-static int hyperv_handle_properties(CPUState *cs,
-                                    struct kvm_cpuid_entry2 *cpuid_ent)
+static int hyperv_expand_features(CPUState *cs)
 {
     X86CPU *cpu = X86_CPU(cs);
     CPUX86State *env = &cpu->env;
     struct kvm_cpuid2 *cpuid;
     struct kvm_cpuid_entry2 *c;
-    uint32_t cpuid_i = 0;
     int r;
 
     if (!hyperv_enabled(cpu))
@@ -1317,11 +1317,26 @@ static int hyperv_handle_properties(CPUState *cs,
     /* Not exposed by KVM but needed to make CPU hotplug in Windows work */
     env->features[FEAT_HYPERV_EDX] |= HV_CPU_DYNAMIC_PARTITIONING_AVAILABLE;
 
+    g_free(cpuid);
+
     if (r) {
-        r = -ENOSYS;
-        goto free;
+        return -ENOSYS;
     }
 
+    return 0;
+}
+
+/*
+ * Fill in Hyper-V CPUIDs. Returns the number of entries filled in cpuid_ent.
+ */
+static int hyperv_fill_cpuids(CPUState *cs,
+                              struct kvm_cpuid_entry2 *cpuid_ent)
+{
+    X86CPU *cpu = X86_CPU(cs);
+    CPUX86State *env = &cpu->env;
+    struct kvm_cpuid_entry2 *c;
+    uint32_t cpuid_i = 0;
+
     c = &cpuid_ent[cpuid_i++];
     c->function = HV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
     c->eax = hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ?
@@ -1382,12 +1397,8 @@ static int hyperv_handle_properties(CPUState *cs,
         c->ecx = env->features[FEAT_HV_NESTED_ECX];
         c->edx = env->features[FEAT_HV_NESTED_EDX];
     }
-    r = cpuid_i;
 
-free:
-    g_free(cpuid);
-
-    return r;
+    return cpuid_i;
 }
 
 static Error *hv_passthrough_mig_blocker;
@@ -1529,11 +1540,13 @@ int kvm_arch_init_vcpu(CPUState *cs)
     env->apic_bus_freq = KVM_APIC_BUS_FREQUENCY;
 
     /* Paravirtualization CPUIDs */
-    r = hyperv_handle_properties(cs, cpuid_data.entries);
+    r = hyperv_expand_features(cs);
     if (r < 0) {
         return r;
-    } else if (r > 0) {
-        cpuid_i = r;
+    }
+
+    if (hyperv_enabled(cpu)) {
+        cpuid_i = hyperv_fill_cpuids(cs, cpuid_data.entries);
         kvm_base = KVM_CPUID_SIGNATURE_NEXT;
         has_msr_hv_hypercall = true;
     }
Signature · 1 line
2.25.4
#24 lore raw

RFC PATCH 14/22 i386: move eVMCS enablement to hyperv_init_vcpu()

In reply to Vitaly Kuznetsov

hyperv_expand_features() will be called before we create vCPU so evmcs enablement should go away. hyperv_init_vcpu() looks like the right place.

---
 target/i386/kvm.c | 63 ++++++++++++++++++++++++++++-------------------
 1 file changed, 38 insertions(+), 25 deletions(-)
target/i386/kvm.c +38 -25
index c4bb147090a9..2e9ceafa2421 100644
@@ -974,6 +974,7 @@ static struct kvm_cpuid2 *get_supported_hv_cpuid(CPUState *cs)
 {
     struct kvm_cpuid2 *cpuid;
     int max = 7; /* 0x40000000..0x40000005, 0x4000000A */
+    int i;
 
     /*
      * When the buffer is too small, KVM_GET_SUPPORTED_HV_CPUID fails with
@@ -983,6 +984,22 @@ static struct kvm_cpuid2 *get_supported_hv_cpuid(CPUState *cs)
     while ((cpuid = try_get_hv_cpuid(cs, max)) == NULL) {
         max++;
     }
+
+    /*
+     * KVM_GET_SUPPORTED_HV_CPUID does not set EVMCS CPUID bit before
+     * KVM_CAP_HYPERV_ENLIGHTENED_VMCS is enabled but we want to get the
+     * information early, just check for the capability and set the bit
+     * manually.
+     */
+    if (kvm_check_extension(cs->kvm_state,
+                            KVM_CAP_HYPERV_ENLIGHTENED_VMCS) > 0) {
+        for (i = 0; i < cpuid->nent; i++) {
+            if (cpuid->entries[i].function == HV_CPUID_ENLIGHTMENT_INFO) {
+                cpuid->entries[i].eax |= HV_ENLIGHTENED_VMCS_RECOMMENDED;
+            }
+        }
+    }
+
     return cpuid;
 }
 
@@ -1194,26 +1211,6 @@ static int hyperv_expand_features(CPUState *cs)
     if (!hyperv_enabled(cpu))
         return 0;
 
-    if (hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) ||
-        cpu->hyperv_passthrough) {
-        uint16_t evmcs_version;
-
-        r = kvm_vcpu_enable_cap(cs, KVM_CAP_HYPERV_ENLIGHTENED_VMCS, 0,
-                                (uintptr_t)&evmcs_version);
-
-        if (hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS) && r) {
-            fprintf(stderr, "Hyper-V %s is not supported by kernel\n",
-                    kvm_hyperv_properties[HYPERV_FEAT_EVMCS].desc);
-            return -ENOSYS;
-        }
-
-        if (!r) {
-            env->features[FEAT_HV_RECOMM_EAX] |=
-                HV_ENLIGHTENED_VMCS_RECOMMENDED;
-            env->features[FEAT_HV_NESTED_EAX] = evmcs_version;
-        }
-    }
-
     if (kvm_check_extension(cs->kvm_state, KVM_CAP_HYPERV_CPUID) > 0) {
         cpuid = get_supported_hv_cpuid(cs);
     } else {
@@ -1407,6 +1404,7 @@ static Error *hv_no_nonarch_cs_mig_blocker;
 static int hyperv_init_vcpu(X86CPU *cpu)
 {
     CPUState *cs = CPU(cpu);
+    CPUX86State *env = &cpu->env;
     Error *local_err = NULL;
     int ret;
 
@@ -1482,6 +1480,21 @@ static int hyperv_init_vcpu(X86CPU *cpu)
         }
     }
 
+    if (hyperv_feat_enabled(cpu, HYPERV_FEAT_EVMCS)) {
+        uint16_t evmcs_version;
+
+        ret = kvm_vcpu_enable_cap(cs, KVM_CAP_HYPERV_ENLIGHTENED_VMCS, 0,
+                                  (uintptr_t)&evmcs_version);
+
+        if (ret < 0) {
+            fprintf(stderr, "Hyper-V %s is not supported by kernel\n",
+                    kvm_hyperv_properties[HYPERV_FEAT_EVMCS].desc);
+            return ret;
+        }
+
+        env->features[FEAT_HV_NESTED_EAX] = evmcs_version;
+    }
+
     return 0;
 }
 
@@ -1546,6 +1559,11 @@ int kvm_arch_init_vcpu(CPUState *cs)
     }
 
     if (hyperv_enabled(cpu)) {
+        r = hyperv_init_vcpu(cpu);
+        if (r) {
+            return r;
+        }
+
         cpuid_i = hyperv_fill_cpuids(cs, cpuid_data.entries);
         kvm_base = KVM_CPUID_SIGNATURE_NEXT;
         has_msr_hv_hypercall = true;
@@ -1892,11 +1910,6 @@ int kvm_arch_init_vcpu(CPUState *cs)
 
     kvm_init_msrs(cpu);
 
-    r = hyperv_init_vcpu(cpu);
-    if (r) {
-        goto fail;
-    }
-
     return 0;
 
  fail:
Signature · 1 line
2.25.4
#25 lore raw

RFC PATCH 15/22 i386: switch hyperv_expand_features() to using error_setg()

In reply to Vitaly Kuznetsov

Use standard error_setg() mechanism in hyperv_expand_features().

---
 target/i386/kvm.c | 96 +++++++++++++++++++++++++++++++----------------
 1 file changed, 64 insertions(+), 32 deletions(-)
target/i386/kvm.c +64 -32
index 2e9ceafa2421..760b93091fea 100644
@@ -1139,7 +1139,7 @@ static int hv_cpuid_get_fw(struct kvm_cpuid2 *cpuid, int fw, uint32_t *r)
 }
 
 static int hv_cpuid_check_and_set(CPUState *cs, struct kvm_cpuid2 *cpuid,
-                                  int feature)
+                                  int feature, Error **errp)
 {
     X86CPU *cpu = X86_CPU(cs);
     CPUX86State *env = &cpu->env;
@@ -1155,11 +1155,10 @@ static int hv_cpuid_check_and_set(CPUState *cs, struct kvm_cpuid2 *cpuid,
     while (deps) {
         dep_feat = ctz64(deps);
         if (!(hyperv_feat_enabled(cpu, dep_feat))) {
-                fprintf(stderr,
-                        "Hyper-V %s requires Hyper-V %s\n",
-                        kvm_hyperv_properties[feature].desc,
-                        kvm_hyperv_properties[dep_feat].desc);
-                return 1;
+            error_setg(errp, "Hyper-V %s requires Hyper-V %s",
+                       kvm_hyperv_properties[feature].desc,
+                       kvm_hyperv_properties[dep_feat].desc);
+            return 1;
         }
         deps &= ~(1ull << dep_feat);
     }
@@ -1174,9 +1173,8 @@ static int hv_cpuid_check_and_set(CPUState *cs, struct kvm_cpuid2 *cpuid,
 
         if (hv_cpuid_get_fw(cpuid, fw, &r) || (r & bits) != bits) {
             if (hyperv_feat_enabled(cpu, feature)) {
-                fprintf(stderr,
-                        "Hyper-V %s is not supported by kernel\n",
-                        kvm_hyperv_properties[feature].desc);
+                error_setg(errp, "Hyper-V %s is not supported by kernel",
+                           kvm_hyperv_properties[feature].desc);
                 return 1;
             } else {
                 return 0;
@@ -1200,13 +1198,13 @@ static int hv_cpuid_check_and_set(CPUState *cs, struct kvm_cpuid2 *cpuid,
  * of 'hv_passthrough' mode and fills the environment with all supported
  * Hyper-V features.
  */
-static int hyperv_expand_features(CPUState *cs)
+static int hyperv_expand_features(CPUState *cs, Error **errp)
 {
     X86CPU *cpu = X86_CPU(cs);
     CPUX86State *env = &cpu->env;
     struct kvm_cpuid2 *cpuid;
     struct kvm_cpuid_entry2 *c;
-    int r;
+    int r = 1;
 
     if (!hyperv_enabled(cpu))
         return 0;
@@ -1285,34 +1283,67 @@ static int hyperv_expand_features(CPUState *cs)
     }
 
     /* Features */
-    r = hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_RELAXED);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_VAPIC);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_TIME);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_CRASH);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_RESET);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_VPINDEX);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_RUNTIME);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_SYNIC);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_STIMER);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_FREQUENCIES);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_REENLIGHTENMENT);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_TLBFLUSH);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_EVMCS);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_IPI);
-    r |= hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_STIMER_DIRECT);
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_RELAXED, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_VAPIC, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_TIME, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_CRASH, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_RESET, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_VPINDEX, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_RUNTIME, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_SYNIC, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_STIMER, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_FREQUENCIES, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_REENLIGHTENMENT, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_TLBFLUSH, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_EVMCS, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_IPI, errp)) {
+        goto out;
+    }
+    if (hv_cpuid_check_and_set(cs, cpuid, HYPERV_FEAT_STIMER_DIRECT, errp)) {
+        goto out;
+    }
 
     /* Additional dependencies not covered by kvm_hyperv_properties[] */
     if (hyperv_feat_enabled(cpu, HYPERV_FEAT_SYNIC) &&
         !cpu->hyperv_synic_kvm_only &&
         !hyperv_feat_enabled(cpu, HYPERV_FEAT_VPINDEX)) {
-        fprintf(stderr, "Hyper-V %s requires Hyper-V %s\n",
-                kvm_hyperv_properties[HYPERV_FEAT_SYNIC].desc,
-                kvm_hyperv_properties[HYPERV_FEAT_VPINDEX].desc);
-        r |= 1;
+        error_setg(errp, "Hyper-V %s requires Hyper-V %s",
+                   kvm_hyperv_properties[HYPERV_FEAT_SYNIC].desc,
+                   kvm_hyperv_properties[HYPERV_FEAT_VPINDEX].desc);
+        goto out;
     }
 
     /* Not exposed by KVM but needed to make CPU hotplug in Windows work */
     env->features[FEAT_HYPERV_EDX] |= HV_CPU_DYNAMIC_PARTITIONING_AVAILABLE;
+    r = 0;
+
+out:
 
     g_free(cpuid);
 
@@ -1553,8 +1584,9 @@ int kvm_arch_init_vcpu(CPUState *cs)
     env->apic_bus_freq = KVM_APIC_BUS_FREQUENCY;
 
     /* Paravirtualization CPUIDs */
-    r = hyperv_expand_features(cs);
-    if (r < 0) {
+    r = hyperv_expand_features(cs, &local_err);
+    if (local_err) {
+        error_report_err(local_err);
         return r;
     }
 
Signature · 1 line
2.25.4
#26 lore raw

RFC PATCH 16/22 i386: make hyperv_expand_features() return void

In reply to Vitaly Kuznetsov

As hyperv_expand_features() now uses standard error_setg() mechanism we can make it return void.

---
 target/i386/kvm.c | 16 +++++-----------
 1 file changed, 5 insertions(+), 11 deletions(-)
target/i386/kvm.c +5 -11
index 760b93091fea..cb35f759acaa 100644
@@ -1198,16 +1198,15 @@ static int hv_cpuid_check_and_set(CPUState *cs, struct kvm_cpuid2 *cpuid,
  * of 'hv_passthrough' mode and fills the environment with all supported
  * Hyper-V features.
  */
-static int hyperv_expand_features(CPUState *cs, Error **errp)
+static void hyperv_expand_features(CPUState *cs, Error **errp)
 {
     X86CPU *cpu = X86_CPU(cs);
     CPUX86State *env = &cpu->env;
     struct kvm_cpuid2 *cpuid;
     struct kvm_cpuid_entry2 *c;
-    int r = 1;
 
     if (!hyperv_enabled(cpu))
-        return 0;
+        return;
 
     if (kvm_check_extension(cs->kvm_state, KVM_CAP_HYPERV_CPUID) > 0) {
         cpuid = get_supported_hv_cpuid(cs);
@@ -1341,17 +1340,12 @@ static int hyperv_expand_features(CPUState *cs, Error **errp)
 
     /* Not exposed by KVM but needed to make CPU hotplug in Windows work */
     env->features[FEAT_HYPERV_EDX] |= HV_CPU_DYNAMIC_PARTITIONING_AVAILABLE;
-    r = 0;
 
 out:
 
     g_free(cpuid);
 
-    if (r) {
-        return -ENOSYS;
-    }
-
-    return 0;
+    return;
 }
 
 /*
@@ -1584,10 +1578,10 @@ int kvm_arch_init_vcpu(CPUState *cs)
     env->apic_bus_freq = KVM_APIC_BUS_FREQUENCY;
 
     /* Paravirtualization CPUIDs */
-    r = hyperv_expand_features(cs, &local_err);
+    hyperv_expand_features(cs, &local_err);
     if (local_err) {
         error_report_err(local_err);
-        return r;
+        return -ENOSYS;
     }
 
     if (hyperv_enabled(cpu)) {
Signature · 1 line
2.25.4
#27 lore raw

RFC PATCH 17/22 i386: adjust the expected KVM_GET_SUPPORTED_HV_CPUID array size

In reply to Vitaly Kuznetsov

SYNDBG leaves were recently (Linux-5.8) added to KVM but we haven't updated the expected size of KVM_GET_SUPPORTED_HV_CPUID output in KVM so we now make serveral tries before succeeding. Update the default.

---
 target/i386/kvm.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
target/i386/kvm.c +2 -1
index cb35f759acaa..70db2ca26313 100644
@@ -973,7 +973,8 @@ static struct kvm_cpuid2 *try_get_hv_cpuid(CPUState *cs, int max)
 static struct kvm_cpuid2 *get_supported_hv_cpuid(CPUState *cs)
 {
     struct kvm_cpuid2 *cpuid;
-    int max = 7; /* 0x40000000..0x40000005, 0x4000000A */
+    /* 0x40000000..0x40000005, 0x4000000A, 0x40000080..0x40000080 leaves */
+    int max = 10;
     int i;
 
     /*
Signature · 1 line
2.25.4
#28 lore raw

RFC PATCH 18/22 i386: prefer system KVM_GET_SUPPORTED_HV_CPUID ioctl over vCPU's one

In reply to Vitaly Kuznetsov

KVM_GET_SUPPORTED_HV_CPUID was made a system wide ioctl which can be called prior to creating vCPUs and we are going to use that to expand Hyper-V cpu features early. Use it when it is supported by KVM.

---
 target/i386/kvm.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)
target/i386/kvm.c +13 -4
index 70db2ca26313..b76d4a5a147b 100644
@@ -940,7 +940,8 @@ static struct {
     },
 };
 
-static struct kvm_cpuid2 *try_get_hv_cpuid(CPUState *cs, int max)
+static struct kvm_cpuid2 *try_get_hv_cpuid(CPUState *cs, int max,
+                                           bool do_sys_ioctl)
 {
     struct kvm_cpuid2 *cpuid;
     int r, size;
@@ -949,7 +950,11 @@ static struct kvm_cpuid2 *try_get_hv_cpuid(CPUState *cs, int max)
     cpuid = g_malloc0(size);
     cpuid->nent = max;
 
-    r = kvm_vcpu_ioctl(cs, KVM_GET_SUPPORTED_HV_CPUID, cpuid);
+    if (do_sys_ioctl) {
+        r = kvm_ioctl(kvm_state, KVM_GET_SUPPORTED_HV_CPUID, cpuid);
+    } else {
+        r = kvm_vcpu_ioctl(cs, KVM_GET_SUPPORTED_HV_CPUID, cpuid);
+    }
     if (r == 0 && cpuid->nent >= max) {
         r = -E2BIG;
     }
@@ -976,13 +981,17 @@ static struct kvm_cpuid2 *get_supported_hv_cpuid(CPUState *cs)
     /* 0x40000000..0x40000005, 0x4000000A, 0x40000080..0x40000080 leaves */
     int max = 10;
     int i;
+    bool do_sys_ioctl;
+
+    do_sys_ioctl =
+        kvm_check_extension(kvm_state, KVM_CAP_SYS_HYPERV_CPUID) > 0;
 
     /*
      * When the buffer is too small, KVM_GET_SUPPORTED_HV_CPUID fails with
      * -E2BIG, however, it doesn't report back the right size. Keep increasing
      * it and re-trying until we succeed.
      */
-    while ((cpuid = try_get_hv_cpuid(cs, max)) == NULL) {
+    while ((cpuid = try_get_hv_cpuid(cs, max, do_sys_ioctl)) == NULL) {
         max++;
     }
 
@@ -992,7 +1001,7 @@ static struct kvm_cpuid2 *get_supported_hv_cpuid(CPUState *cs)
      * information early, just check for the capability and set the bit
      * manually.
      */
-    if (kvm_check_extension(cs->kvm_state,
+    if (!do_sys_ioctl && kvm_check_extension(cs->kvm_state,
                             KVM_CAP_HYPERV_ENLIGHTENED_VMCS) > 0) {
         for (i = 0; i < cpuid->nent; i++) {
             if (cpuid->entries[i].function == HV_CPUID_ENLIGHTMENT_INFO) {
Signature · 1 line
2.25.4
#29 lore raw

RFC PATCH 19/22 i386: prepare hyperv_expand_features() to be called at CPU feature expansion time

In reply to Vitaly Kuznetsov

If hyperv_expand_features() is called before vCPU is created, cs->kvm_state is NULL. We can only do the expantion if system wide KVM_GET_SUPPORTED_HV_CPUID is supported.

---
 target/i386/kvm.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)
target/i386/kvm.c +12 -1
index b76d4a5a147b..c17e7db5e627 100644
@@ -986,6 +986,10 @@ static struct kvm_cpuid2 *get_supported_hv_cpuid(CPUState *cs)
     do_sys_ioctl =
         kvm_check_extension(kvm_state, KVM_CAP_SYS_HYPERV_CPUID) > 0;
 
+    if (!do_sys_ioctl && !cs->kvm_state) {
+        return NULL;
+    }
+
     /*
      * When the buffer is too small, KVM_GET_SUPPORTED_HV_CPUID fails with
      * -E2BIG, however, it doesn't report back the right size. Keep increasing
@@ -1023,6 +1027,10 @@ static struct kvm_cpuid2 *get_supported_hv_cpuid_legacy(CPUState *cs)
     struct kvm_cpuid2 *cpuid;
     struct kvm_cpuid_entry2 *entry_feat, *entry_recomm;
 
+    if (!cs->kvm_state) {
+        return NULL;
+    }
+
     /* HV_CPUID_FEATURES, HV_CPUID_ENLIGHTMENT_INFO */
     cpuid = g_malloc0(sizeof(*cpuid) + 2 * sizeof(*cpuid->entries));
     cpuid->nent = 2;
@@ -1218,12 +1226,15 @@ static void hyperv_expand_features(CPUState *cs, Error **errp)
     if (!hyperv_enabled(cpu))
         return;
 
-    if (kvm_check_extension(cs->kvm_state, KVM_CAP_HYPERV_CPUID) > 0) {
+    if (kvm_check_extension(kvm_state, KVM_CAP_HYPERV_CPUID) > 0) {
         cpuid = get_supported_hv_cpuid(cs);
     } else {
         cpuid = get_supported_hv_cpuid_legacy(cs);
     }
 
+    if (!cpuid)
+        return;
+
     if (cpu->hyperv_passthrough) {
         c = cpuid_find_entry(cpuid, HV_CPUID_VENDOR_AND_MAX_FUNCTIONS, 0);
         if (c) {
Signature · 1 line
2.25.4
#30 lore raw

RFC PATCH 20/22 i386: use global kvm_state in hyperv_enabled() check

In reply to Vitaly Kuznetsov

There is no need to use vCPU-specific kvm state in hyperv_enabled() check and we need to do that when feature expansion happens early, before vCPU specific KVM state is created.

---
 target/i386/kvm.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
target/i386/kvm.c +1 -2
index c17e7db5e627..0945983498b2 100644
@@ -728,8 +728,7 @@ unsigned long kvm_arch_vcpu_id(CPUState *cs)
 
 static bool hyperv_enabled(X86CPU *cpu)
 {
-    CPUState *cs = CPU(cpu);
-    return kvm_check_extension(cs->kvm_state, KVM_CAP_HYPERV) > 0 &&
+    return kvm_check_extension(kvm_state, KVM_CAP_HYPERV) > 0 &&
         ((cpu->hyperv_spinlock_attempts != HYPERV_SPINLOCK_NEVER_RETRY) ||
          cpu->hyperv_features || cpu->hyperv_passthrough);
 }
Signature · 1 line
2.25.4
#31 lore raw

RFC PATCH 21/22 i386: record if Hyper-V features were already expanded

In reply to Vitaly Kuznetsov

Introduce a per cpu hyperv_features_expaned flag indicating that Hyper-V related CPUIDs were already acquired from KVM. We are going to start doing the expansion prior to creating KVM vCPU but in case KVM doesn't support sustem wide KVM_GET_SUPPORTED_HV_CPUID ioctl this can't happen.

---
 target/i386/cpu.h | 1 +
 target/i386/kvm.c | 6 ++++++
 2 files changed, 7 insertions(+)
target/i386/cpu.h +1 -0
index 0aad60e0c707..c72755139047 100644
@@ -1669,6 +1669,7 @@ struct X86CPU {
     uint32_t hyperv_interface_id[4];
     uint32_t hyperv_version_id[4];
     uint32_t hyperv_limits[3];
+    bool hyperv_features_expaned;
 
     bool check_cpuid;
     bool enforce_cpuid;
target/i386/kvm.c +6 -0
index 0945983498b2..961241528a5c 100644
@@ -1225,6 +1225,10 @@ static void hyperv_expand_features(CPUState *cs, Error **errp)
     if (!hyperv_enabled(cpu))
         return;
 
+    if (cpu->hyperv_features_expaned) {
+        return;
+    }
+
     if (kvm_check_extension(kvm_state, KVM_CAP_HYPERV_CPUID) > 0) {
         cpuid = get_supported_hv_cpuid(cs);
     } else {
@@ -1361,6 +1365,8 @@ static void hyperv_expand_features(CPUState *cs, Error **errp)
     /* Not exposed by KVM but needed to make CPU hotplug in Windows work */
     env->features[FEAT_HYPERV_EDX] |= HV_CPU_DYNAMIC_PARTITIONING_AVAILABLE;
 
+    cpu->hyperv_features_expaned = true;
+
 out:
 
     g_free(cpuid);
Signature · 1 line
2.25.4
#32 lore raw

RFC PATCH 22/22 i386: expand Hyper-V features early

In reply to Vitaly Kuznetsov

To make Hyper-V features appear in e.g. QMP query-cpu-model-expansion we need to expand and set the corresponding CPUID leaves early. Modify x86_cpu_get_supported_feature_word() to call newly intoduced Hyper-V specific kvm_hv_get_supported_cpuid() instead of kvm_arch_get_supported_cpuid(). We can't use kvm_arch_get_supported_cpuid() as Hyper-V specific CPUID leaves intersect with KVM's.

Note, early expansion will only happen when KVM supports system wide KVM_GET_SUPPORTED_HV_CPUID ioctl (KVM_CAP_SYS_HYPERV_CPUID).

---
 target/i386/cpu.c      | 15 +++++++++------
 target/i386/kvm.c      | 15 +++++++++++++++
 target/i386/kvm_i386.h |  7 +++++++
 3 files changed, 31 insertions(+), 6 deletions(-)
target/i386/cpu.c +9 -6
index 479c4bbbf459..d3c4ecb3535c 100644
@@ -5147,7 +5147,7 @@ CpuDefinitionInfoList *qmp_query_cpu_definitions(Error **errp)
     return cpu_list;
 }
 
-static uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
+static uint64_t x86_cpu_get_supported_feature_word(X86CPU *cpu, FeatureWord w,
                                                    bool migratable_only)
 {
     FeatureWordInfo *wi = &feature_word_info[w];
@@ -5156,9 +5156,12 @@ static uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
     if (kvm_enabled()) {
         switch (wi->type) {
         case CPUID_FEATURE_WORD:
-            r = kvm_arch_get_supported_cpuid(kvm_state, wi->cpuid.eax,
-                                                        wi->cpuid.ecx,
-                                                        wi->cpuid.reg);
+            if (hyperv_feature_word(w))
+                r = kvm_hv_get_supported_cpuid(cpu, w);
+            else
+                r = kvm_arch_get_supported_cpuid(kvm_state, wi->cpuid.eax,
+                                                 wi->cpuid.ecx,
+                                                 wi->cpuid.reg);
             break;
         case MSR_FEATURE_WORD:
             r = kvm_arch_get_supported_msr_feature(kvm_state,
@@ -6485,7 +6488,7 @@ static void x86_cpu_expand_features(X86CPU *cpu, Error **errp)
              * by the user.
              */
             env->features[w] |=
-                x86_cpu_get_supported_feature_word(w, cpu->migratable) &
+                x86_cpu_get_supported_feature_word(cpu, w, cpu->migratable) &
                 ~env->user_features[w] &
                 ~feature_word_info[w].no_autoenable_flags;
         }
@@ -6589,7 +6592,7 @@ static void x86_cpu_filter_features(X86CPU *cpu, bool verbose)
 
     for (w = 0; w < FEATURE_WORDS; w++) {
         uint64_t host_feat =
-            x86_cpu_get_supported_feature_word(w, false);
+            x86_cpu_get_supported_feature_word(cpu, w, false);
         uint64_t requested_features = env->features[w];
         uint64_t unavailable_features = requested_features & ~host_feat;
         mark_unavailable_features(cpu, w, unavailable_features, prefix);
target/i386/kvm.c +15 -0
index 961241528a5c..764b96fbbb7d 100644
@@ -1449,6 +1449,21 @@ static int hyperv_fill_cpuids(CPUState *cs,
     return cpuid_i;
 }
 
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w)
+{
+    CPUState *cs = CPU(cpu);
+    CPUX86State *env = &cpu->env;
+    Error *local_err = NULL;
+
+    hyperv_expand_features(cs, &local_err);
+
+    if (local_err) {
+            error_report_err(local_err);
+    }
+
+    return env->features[w];
+}
+
 static Error *hv_passthrough_mig_blocker;
 static Error *hv_no_nonarch_cs_mig_blocker;
 
target/i386/kvm_i386.h +7 -0
index 064b8798a26c..2e7da4f39668 100644
@@ -48,4 +48,11 @@ bool kvm_has_waitpkg(void);
 
 bool kvm_hv_vpindex_settable(void);
 
+static inline bool hyperv_feature_word(enum FeatureWord w)
+{
+    return w >= FEAT_HYPERV_EAX && w <= FEAT_HV_NESTED_EDX;
+}
+
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w);
+
 #endif
Signature · 1 line
2.25.4
#33 lore raw

Re: [PATCH RFC 22/22] i386: expand Hyper-V features early

Quoted 9 lines · Vitaly Kuznetsov

To make Hyper-V features appear in e.g. QMP query-cpu-model-expansion we need to expand and set the corresponding CPUID leaves early. Modify x86_cpu_get_supported_feature_word() to call newly intoduced Hyper-V specific kvm_hv_get_supported_cpuid() instead of kvm_arch_get_supported_cpuid(). We can't use kvm_arch_get_supported_cpuid() as Hyper-V specific CPUID leaves intersect with KVM's.

Note, early expansion will only happen when KVM supports system wide KVM_GET_SUPPORTED_HV_CPUID ioctl (KVM_CAP_SYS_HYPERV_CPUID).

Will this implicitly enable hyperv CPUID passthrough when using "-cpu host"? Do we want it to?

Quoted 100 lines
---
 target/i386/cpu.c      | 15 +++++++++------
 target/i386/kvm.c      | 15 +++++++++++++++
 target/i386/kvm_i386.h |  7 +++++++
 3 files changed, 31 insertions(+), 6 deletions(-)
target/i386/cpu.c +9 -6
index 479c4bbbf459..d3c4ecb3535c 100644
@@ -5147,7 +5147,7 @@ CpuDefinitionInfoList *qmp_query_cpu_definitions(Error **errp)
     return cpu_list;
 }
 
-static uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
+static uint64_t x86_cpu_get_supported_feature_word(X86CPU *cpu, FeatureWord w,
                                                    bool migratable_only)
 {
     FeatureWordInfo *wi = &feature_word_info[w];
@@ -5156,9 +5156,12 @@ static uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
     if (kvm_enabled()) {
         switch (wi->type) {
         case CPUID_FEATURE_WORD:
-            r = kvm_arch_get_supported_cpuid(kvm_state, wi->cpuid.eax,
-                                                        wi->cpuid.ecx,
-                                                        wi->cpuid.reg);
+            if (hyperv_feature_word(w))
+                r = kvm_hv_get_supported_cpuid(cpu, w);
+            else
+                r = kvm_arch_get_supported_cpuid(kvm_state, wi->cpuid.eax,
+                                                 wi->cpuid.ecx,
+                                                 wi->cpuid.reg);
             break;
         case MSR_FEATURE_WORD:
             r = kvm_arch_get_supported_msr_feature(kvm_state,
@@ -6485,7 +6488,7 @@ static void x86_cpu_expand_features(X86CPU *cpu, Error **errp)
              * by the user.
              */
             env->features[w] |=
-                x86_cpu_get_supported_feature_word(w, cpu->migratable) &
+                x86_cpu_get_supported_feature_word(cpu, w, cpu->migratable) &
                 ~env->user_features[w] &
                 ~feature_word_info[w].no_autoenable_flags;
         }
@@ -6589,7 +6592,7 @@ static void x86_cpu_filter_features(X86CPU *cpu, bool verbose)
 
     for (w = 0; w < FEATURE_WORDS; w++) {
         uint64_t host_feat =
-            x86_cpu_get_supported_feature_word(w, false);
+            x86_cpu_get_supported_feature_word(cpu, w, false);
         uint64_t requested_features = env->features[w];
         uint64_t unavailable_features = requested_features & ~host_feat;
         mark_unavailable_features(cpu, w, unavailable_features, prefix);
target/i386/kvm.c +15 -0
index 961241528a5c..764b96fbbb7d 100644
@@ -1449,6 +1449,21 @@ static int hyperv_fill_cpuids(CPUState *cs,
     return cpuid_i;
 }
 
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w)
+{
+    CPUState *cs = CPU(cpu);
+    CPUX86State *env = &cpu->env;
+    Error *local_err = NULL;
+
+    hyperv_expand_features(cs, &local_err);
+
+    if (local_err) {
+            error_report_err(local_err);
+    }
+
+    return env->features[w];
+}
+
 static Error *hv_passthrough_mig_blocker;
 static Error *hv_no_nonarch_cs_mig_blocker;
 
target/i386/kvm_i386.h +7 -0
index 064b8798a26c..2e7da4f39668 100644
@@ -48,4 +48,11 @@ bool kvm_has_waitpkg(void);
 
 bool kvm_hv_vpindex_settable(void);
 
+static inline bool hyperv_feature_word(enum FeatureWord w)
+{
+    return w >= FEAT_HYPERV_EAX && w <= FEAT_HV_NESTED_EDX;
+}
+
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w);
+
 #endif
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#34 lore raw

Re: [PATCH RFC 22/22] i386: expand Hyper-V features early

Quoted 8 lines · Eduardo Habkost
Quoted 7 lines · Vitaly Kuznetsov

To make Hyper-V features appear in e.g. QMP query-cpu-model-expansion we need to expand and set the corresponding CPUID leaves early. Modify x86_cpu_get_supported_feature_word() to call newly intoduced Hyper-V specific kvm_hv_get_supported_cpuid() instead of kvm_arch_get_supported_cpuid(). We can't use kvm_arch_get_supported_cpuid() as Hyper-V specific CPUID leaves intersect with KVM's.

Quoted 6 lines
Quoted 2 lines

Note, early expansion will only happen when KVM supports system wide KVM_GET_SUPPORTED_HV_CPUID ioctl (KVM_CAP_SYS_HYPERV_CPUID).

Will this implicitly enable hyperv CPUID passthrough when using "-cpu host"? Do we want it to?

I hope it won't. With 'hv_*' features set explicitly we will get just them (if the host supports), with 'hv_passthrough' we will get everything supported by the host. With no 'hv_*' parameters on the command line we should get no Hyper-V features. I'll check this.

Personally, I don't think '-host' mode should enable any Hyper-V features as these features have side-effects. E.g. if you enable SynIC you won't be able to use vAPIC. Probably not a reasonable default for the majority of the guests which (hopefully) are Linux.

Quoted 100 lines
Quoted 100 lines
---
 target/i386/cpu.c      | 15 +++++++++------
 target/i386/kvm.c      | 15 +++++++++++++++
 target/i386/kvm_i386.h |  7 +++++++
 3 files changed, 31 insertions(+), 6 deletions(-)
target/i386/cpu.c +9 -6
index 479c4bbbf459..d3c4ecb3535c 100644
@@ -5147,7 +5147,7 @@ CpuDefinitionInfoList *qmp_query_cpu_definitions(Error **errp)
     return cpu_list;
 }
 
-static uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
+static uint64_t x86_cpu_get_supported_feature_word(X86CPU *cpu, FeatureWord w,
                                                    bool migratable_only)
 {
     FeatureWordInfo *wi = &feature_word_info[w];
@@ -5156,9 +5156,12 @@ static uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
     if (kvm_enabled()) {
         switch (wi->type) {
         case CPUID_FEATURE_WORD:
-            r = kvm_arch_get_supported_cpuid(kvm_state, wi->cpuid.eax,
-                                                        wi->cpuid.ecx,
-                                                        wi->cpuid.reg);
+            if (hyperv_feature_word(w))
+                r = kvm_hv_get_supported_cpuid(cpu, w);
+            else
+                r = kvm_arch_get_supported_cpuid(kvm_state, wi->cpuid.eax,
+                                                 wi->cpuid.ecx,
+                                                 wi->cpuid.reg);
             break;
         case MSR_FEATURE_WORD:
             r = kvm_arch_get_supported_msr_feature(kvm_state,
@@ -6485,7 +6488,7 @@ static void x86_cpu_expand_features(X86CPU *cpu, Error **errp)
              * by the user.
              */
             env->features[w] |=
-                x86_cpu_get_supported_feature_word(w, cpu->migratable) &
+                x86_cpu_get_supported_feature_word(cpu, w, cpu->migratable) &
                 ~env->user_features[w] &
                 ~feature_word_info[w].no_autoenable_flags;
         }
@@ -6589,7 +6592,7 @@ static void x86_cpu_filter_features(X86CPU *cpu, bool verbose)
 
     for (w = 0; w < FEATURE_WORDS; w++) {
         uint64_t host_feat =
-            x86_cpu_get_supported_feature_word(w, false);
+            x86_cpu_get_supported_feature_word(cpu, w, false);
         uint64_t requested_features = env->features[w];
         uint64_t unavailable_features = requested_features & ~host_feat;
         mark_unavailable_features(cpu, w, unavailable_features, prefix);
target/i386/kvm.c +15 -0
index 961241528a5c..764b96fbbb7d 100644
@@ -1449,6 +1449,21 @@ static int hyperv_fill_cpuids(CPUState *cs,
     return cpuid_i;
 }
 
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w)
+{
+    CPUState *cs = CPU(cpu);
+    CPUX86State *env = &cpu->env;
+    Error *local_err = NULL;
+
+    hyperv_expand_features(cs, &local_err);
+
+    if (local_err) {
+            error_report_err(local_err);
+    }
+
+    return env->features[w];
+}
+
 static Error *hv_passthrough_mig_blocker;
 static Error *hv_no_nonarch_cs_mig_blocker;
 
target/i386/kvm_i386.h +7 -0
index 064b8798a26c..2e7da4f39668 100644
@@ -48,4 +48,11 @@ bool kvm_has_waitpkg(void);
 
 bool kvm_hv_vpindex_settable(void);
 
+static inline bool hyperv_feature_word(enum FeatureWord w)
+{
+    return w >= FEAT_HYPERV_EAX && w <= FEAT_HV_NESTED_EDX;
+}
+
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w);
+
 #endif
Signature · 1 line
2.25.4
Signature · 1 line
Vitaly
#35 lore raw

Re: [PATCH RFC 22/22] i386: expand Hyper-V features early

In reply to Vitaly Kuznetsov
Quoted 12 lines · Vitaly Kuznetsov

To make Hyper-V features appear in e.g. QMP query-cpu-model-expansion we need to expand and set the corresponding CPUID leaves early. Modify x86_cpu_get_supported_feature_word() to call newly intoduced Hyper-V specific kvm_hv_get_supported_cpuid() instead of kvm_arch_get_supported_cpuid(). We can't use kvm_arch_get_supported_cpuid() as Hyper-V specific CPUID leaves intersect with KVM's.

Note, early expansion will only happen when KVM supports system wide KVM_GET_SUPPORTED_HV_CPUID ioctl (KVM_CAP_SYS_HYPERV_CPUID).

---

[...]

target/i386/kvm.c static int hyperv_fill_cpuids(CPUState *cs, · line 1458

Quoted 7 lines
target/i386/kvm.c +7 -0
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w)
+{
+    CPUState *cs = CPU(cpu);
+    CPUX86State *env = &cpu->env;
+    Error *local_err = NULL;
+
+    hyperv_expand_features(cs, &local_err);

This makes a function that sounds like it doesn't have any side effect ("get supported cpuid") have an unintended side effect (hyperv_expand_features() will change all CPUID data inside the cpu object).

What about making it more similar to kvm_arch_get_supported_cpuid(), and be just a wrapper to get_supported_hv_cpuid()?

I would also make sure get_supported_hv_cpuid() doesn't get CPUState as argument, just to be sure it will never touch the CPU object state.

Quoted 30 lines
target/i386/kvm.c +8 -0
+
+    if (local_err) {
+            error_report_err(local_err);
+    }
+
+    return env->features[w];
+}
+
 static Error *hv_passthrough_mig_blocker;
 static Error *hv_no_nonarch_cs_mig_blocker;
target/i386/kvm_i386.h +7 -0
index 064b8798a26c..2e7da4f39668 100644
@@ -48,4 +48,11 @@ bool kvm_has_waitpkg(void);
 
 bool kvm_hv_vpindex_settable(void);
 
+static inline bool hyperv_feature_word(enum FeatureWord w)
+{
+    return w >= FEAT_HYPERV_EAX && w <= FEAT_HV_NESTED_EDX;
+}
+
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w);
+
 #endif
Signature · 1 line
2.25.4
Signature · 1 line
Eduardo
#36 lore raw

Re: [PATCH RFC 22/22] i386: expand Hyper-V features early

Quoted 27 lines · Eduardo Habkost
Quoted 12 lines · Vitaly Kuznetsov

To make Hyper-V features appear in e.g. QMP query-cpu-model-expansion we need to expand and set the corresponding CPUID leaves early. Modify x86_cpu_get_supported_feature_word() to call newly intoduced Hyper-V specific kvm_hv_get_supported_cpuid() instead of kvm_arch_get_supported_cpuid(). We can't use kvm_arch_get_supported_cpuid() as Hyper-V specific CPUID leaves intersect with KVM's.

Note, early expansion will only happen when KVM supports system wide KVM_GET_SUPPORTED_HV_CPUID ioctl (KVM_CAP_SYS_HYPERV_CPUID).

---

[...]

target/i386/kvm.c static int hyperv_fill_cpuids(CPUState *cs, · line 1458

Quoted 7 lines
target/i386/kvm.c +7 -0
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w)
+{
+    CPUState *cs = CPU(cpu);
+    CPUX86State *env = &cpu->env;
+    Error *local_err = NULL;
+
+    hyperv_expand_features(cs, &local_err);

This makes a function that sounds like it doesn't have any side effect ("get supported cpuid") have an unintended side effect (hyperv_expand_features() will change all CPUID data inside the cpu object).

Quoted 7 lines

What about making it more similar to kvm_arch_get_supported_cpuid(), and be just a wrapper to get_supported_hv_cpuid()?

I would also make sure get_supported_hv_cpuid() doesn't get CPUState as argument, just to be sure it will never touch the CPU object state.

Sure, I can try. The difference with get_supported_cpuid() is that get_supported_hv_cpuid() currently doesn't have cache and we don't probably want to call KVM_GET_SUPPORTED_HV_CPUID ioctl for every CPUID leaf. I don't see why it can't be introduced though.

Quoted 31 lines
Quoted 31 lines
target/i386/kvm.c +8 -0
+
+    if (local_err) {
+            error_report_err(local_err);
+    }
+
+    return env->features[w];
+}
+
 static Error *hv_passthrough_mig_blocker;
 static Error *hv_no_nonarch_cs_mig_blocker;
target/i386/kvm_i386.h +7 -0
index 064b8798a26c..2e7da4f39668 100644
@@ -48,4 +48,11 @@ bool kvm_has_waitpkg(void);
 
 bool kvm_hv_vpindex_settable(void);
 
+static inline bool hyperv_feature_word(enum FeatureWord w)
+{
+    return w >= FEAT_HYPERV_EAX && w <= FEAT_HV_NESTED_EDX;
+}
+
+uint32_t kvm_hv_get_supported_cpuid(X86CPU *cpu, enum FeatureWord w);
+
 #endif
Signature · 1 line
2.25.4
Signature · 1 line
Vitaly