mirror of
https://github.com/hanzoai/visor.git
synced 2026-09-24 17:04:14 +00:00
Tianyu Zhou
fc1afc9de5
SlimVM is a lightweight VMX-based platform for gVisor that accelerates sentry system calls by eliminating unnecessary user/kernel context switches. In a traditional gVisor + KVM setup, every guest syscall triggers a full VM exit to userspace, then re-enters the kernel via a regular syscall: gr0 -> hr0 (vm_exit) -> hr3 (ioctl return) -> hr0 (re-invoke syscall) SlimVM eliminates this overhead. When the guest executes a VMCALL, the VM exits into VMX root mode where SlimVM directly invokes the host kernel's syscall handler, then immediately resumes the guest: gr0 -> hr0 (vmcall + direct function call) SlimVM requires a companion kernel module that provides the /dev/slimvm device interface. See https://github.com/antgroup/slimvm for the kernel module and full documentation. A sentry->host getpid microbenchmark (Xeon 6982P-C, host kernel 5.10.134) shows the effect: │ ns/op │ vs base │ base 82.5 1.0x slimvm 420.9 5.1x kvm 3566.1 43.2x slimvm is ~8.5x faster than the kvm platform on this path. See the issue for details. Updates: https://github.com/google/gvisor/issues/13330 Co-developed-by: Chenggang <[email protected]> Co-developed-by: Robin Luk <[email protected]> Co-developed-by: Tiwei Bie <[email protected]> Co-developed-by: Jianfeng Tan <[email protected]> Co-developed-by: liushi.ls <[email protected]> Co-developed-by: Yong He <[email protected]> Co-developed-by: Aaron Lu <[email protected]> Co-developed-by: chris.zn <[email protected]> Co-developed-by: Lai Jiangshan <[email protected]> Co-developed-by: Lingfu <[email protected]> Co-developed-by: Min Le <[email protected]> Co-developed-by: Quan Xu <[email protected]> Co-developed-by: Zhang Haoyu <[email protected]> Co-developed-by: Dawei Shen <[email protected]> Signed-off-by: Tianyu Zhou <[email protected]>
126 lines
3.4 KiB
Go
126 lines
3.4 KiB
Go
// Copyright 2022 The gVisor Authors.
|
|
//
|
|
// Use of this source code is governed by a BSD-style
|
|
// license that can be found in the LICENSE file or at
|
|
// https://developers.google.com/open-source/licenses/bsd.
|
|
|
|
package hostarch
|
|
|
|
// Masks often used when working with alignment in constant expressions.
|
|
const (
|
|
PageMask = PageSize - 1
|
|
HugePageMask = HugePageSize - 1
|
|
CacheLineMask = CacheLineSize - 1
|
|
JumboPageMask = ^uintptr(JumboPageSize - 1)
|
|
)
|
|
|
|
type bytecount interface {
|
|
~uint | ~uint32 | ~uint64 | ~uintptr
|
|
}
|
|
|
|
type hugebytecount interface {
|
|
~uint | ~uint32 | ~uint64 | ~uintptr
|
|
}
|
|
|
|
// PageRoundDown returns x rounded down to the nearest multiple of PageSize.
|
|
func PageRoundDown[T bytecount](x T) T {
|
|
return x &^ PageMask
|
|
}
|
|
|
|
// PageRoundUp returns x rounded up to the nearest multiple of PageSize. ok is
|
|
// true iff rounding up does not overflow the range of T.
|
|
func PageRoundUp[T bytecount](x T) (val T, ok bool) {
|
|
val = PageRoundDown(x + PageMask)
|
|
ok = val >= x
|
|
return
|
|
}
|
|
|
|
// MustPageRoundUp is equivalent to PageRoundUp, but panics if rounding up
|
|
// overflows.
|
|
func MustPageRoundUp[T bytecount](x T) T {
|
|
val, ok := PageRoundUp(x)
|
|
if !ok {
|
|
panic("PageRoundUp overflows")
|
|
}
|
|
return val
|
|
}
|
|
|
|
// PageOffset returns the offset of x into its containing page.
|
|
func PageOffset[T bytecount](x T) T {
|
|
return x & PageMask
|
|
}
|
|
|
|
// IsPageAligned returns true if x is a multiple of PageSize.
|
|
func IsPageAligned[T bytecount](x T) bool {
|
|
return PageOffset(x) == 0
|
|
}
|
|
|
|
// ToPagesRoundUp returns (the number of pages equal to x bytes rounded up,
|
|
// true). If rounding x up to a multiple of PageSize overflows the range of T,
|
|
// ToPagesRoundUp returns (unspecified, false).
|
|
func ToPagesRoundUp[T bytecount](x T) (T, bool) {
|
|
y := x + PageMask
|
|
if y < x {
|
|
return x, false
|
|
}
|
|
return y / PageSize, true
|
|
}
|
|
|
|
// HugePageRoundDown returns x rounded down to the nearest multiple of
|
|
// HugePageSize.
|
|
func HugePageRoundDown[T hugebytecount](x T) T {
|
|
return x &^ HugePageMask
|
|
}
|
|
|
|
// HugePageRoundUp returns x rounded up to the nearest multiple of
|
|
// HugePageSize. ok is true iff rounding up does not overflow the range of T.
|
|
func HugePageRoundUp[T hugebytecount](x T) (val T, ok bool) {
|
|
val = HugePageRoundDown(x + HugePageMask)
|
|
ok = val >= x
|
|
return
|
|
}
|
|
|
|
// MustHugePageRoundUp is equivalent to HugePageRoundUp, but panics if rounding
|
|
// up overflows.
|
|
func MustHugePageRoundUp[T hugebytecount](x T) T {
|
|
val, ok := HugePageRoundUp(x)
|
|
if !ok {
|
|
panic("HugePageRoundUp overflows")
|
|
}
|
|
return val
|
|
}
|
|
|
|
// HugePageOffset returns the offset of x into its containing page.
|
|
func HugePageOffset[T hugebytecount](x T) T {
|
|
return x & HugePageMask
|
|
}
|
|
|
|
// IsHugePageAligned returns true if x is a multiple of HugePageSize.
|
|
func IsHugePageAligned[T hugebytecount](x T) bool {
|
|
return HugePageOffset(x) == 0
|
|
}
|
|
|
|
// CacheLineRoundDown returns the offset rounded down to the nearest multiple
|
|
// of CacheLineSize.
|
|
func CacheLineRoundDown[T bytecount](x T) T {
|
|
return x &^ CacheLineMask
|
|
}
|
|
|
|
// CacheLineRoundUp returns the offset rounded up to the nearest multiple of
|
|
// CacheLineSize. ok is true iff rounding up does not overflow the range of T.
|
|
func CacheLineRoundUp[T bytecount](x T) (val T, ok bool) {
|
|
val = CacheLineRoundDown(x + CacheLineMask)
|
|
ok = val >= x
|
|
return
|
|
}
|
|
|
|
// MustCacheLineRoundUp is equivalent to CacheLineRoundUp, but panics if
|
|
// rounding up overflows.
|
|
func MustCacheLineRoundUp[T bytecount](x T) T {
|
|
val, ok := CacheLineRoundUp(x)
|
|
if !ok {
|
|
panic("CacheLineRoundUp overflows")
|
|
}
|
|
return val
|
|
}
|