[linux-6.6.y] Add support for ZXPAUSE instruction #276
Conversation
zhaoxin inclusion category: feature ------------------- ZXPAUSE instructs the processor to enter an implementation-dependent optimized state. The instruction execution wakes up when the time-stamp counter reaches or exceeds the implicit EDX:EAX 64-bit input value. The instruction execution also wakes up due to the expiration of the operating system time-limit or by an external interrupt. ZXPAUSE is available on processors with X86_FEATURE_ZXPAUSE. ZXPAUSE allows the processor to enter a light-weight power/performance optimized state (C0.1 state) for a period specified by the instruction or until the system time limit. MSR_ZX_PAUSE_CONTROL MSR register allows the OS to enable/disable C0.2 on the processor and to set the maximum time the processor can reside in C0.1 or C0.2. By default C0.2 is disabled. A sysfs interface to adjust the time and the C0.2 enablement is provided in a follow up change. Signed-off-by: leoliu-oc <leoliu-oc@zhaoxin.com>
zhaoxin inclusion category: feature -------------------- This patch introduces support for the ZXPAUSE instruction, a new addition akin to Intel's TPAUSE. Two primary distinctions set apart ZXPAUSE from TPAUSE: 1. ZXPAUSE utilizes a delta tsc, determined from the lesser value between (MSR_ZX_PAUSE_CONTROL[31:2] << 2) and the EDX:EAX input to the ZXPAUSE instruction, subtracted from the current tsc value. In contrast, TPAUSE employs a target tsc, computed from the lesser value between (MSR_IA32_UMWAIT_CONTROL[31:2] << 2) and the EDX:EAX input to the TPAUSE instruction. 2. As of now, ZXPAUSE exclusively supports the C0.1 optimization state, whereas TPAUSE potentially extends support to both C0.1 and C0.2. Successful integration of this patch hinges on QEMU's backing for ZXPAUSE, a contribution we're currently forwarding to QEMU. It also requires the preceding patch in this patchset, which offers Linux kernel support for ZXPAUSE. The choice of the name "vmx->msr_ia32_umwait_control" is deliberate. In patches for other Linux versions (e.g., 5.5), a "vmx->msr_ia32_umwait_control" already exists. By sharing this variable name with Intel, it ensures compatibility. The difference is merely software-based and poses no real-world conflicts. Currently, if the Guest writes to the ZXPAUSE/TPAUSE CONTROL MSR, we simply bypass the WRMSR instruction. If the Guest attempts to use ZXPAUSE/TPAUSE to transition the vCPU into an optimized state, it will succeed, with the duration of the optimized state being the value passed in EDX:EAX. Of course, this state can be interrupted by external interrupts and other events specified in the specification. Signed-off-by: leoliu-oc <leoliu-oc@zhaoxin.com>
|
Hi @leoliu-oc. Thanks for your PR. I'm waiting for a deepin-community member to verify that this patch is reasonable to test. If it is, they should reply with Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: opsiff The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/ok-to-test |
deepin pr auto reviewAdd Zhaoxin extended VMCS capabilities. |
Includes: