From: Bibo Mao <maobibo@loongson.cn>
To: Tianrui Zhao <zhaotianrui@loongson.cn>,
Huacai Chen <chenhuacai@kernel.org>
Cc: kernel@xen0n.name, kvm@vger.kernel.org,
loongarch@lists.linux.dev, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] LoongArch: KVM: Set max used FPU type with FPU exception
Date: Wed, 25 Mar 2026 09:51:57 +0800 [thread overview]
Message-ID: <d539cf62-c541-ba2e-7810-38ed85c12a62@loongson.cn> (raw)
In-Reply-To: <20260323034034.3465072-1-maobibo@loongson.cn>
another more proactive method is to enabled FPU type with max VM
supported, rather than ever used.
It is more direct and simpler, FPU none --> LASX, no middle state FPU
--> LASX or LSX --> LASX.
Regards
Bibo Mao
On 2026/3/23 上午11:40, Bibo Mao wrote:
> With FPU save and restore flow, the cost is the same with different
> FPU width 8/16/32 bytes, whatever from CPU cycle and cache line impaction.
>
> Here is to enable FPU with max used type, for example if application
> ever uses LASX instrction, enable FPU with LASX type even with FPU
> exeception. So it can avoid possible LSX/LASX exception in future.
>
> With context switch microbench which may touch FPU and LASX, there is 9%
> improvement when halt_poll_ns is disabled. The command is
> "./context --test=pipe" and source code located at:
> https://github.com/bibo-mao/context_switch/blob/main/context.c
>
> Original Wih patch improvement
> 75232 82440 9%
>
> Signed-off-by: Bibo Mao <maobibo@loongson.cn>
> ---
> arch/loongarch/include/asm/kvm_host.h | 1 +
> arch/loongarch/kvm/exit.c | 6 +++---
> arch/loongarch/kvm/vcpu.c | 25 ++++++++++---------------
> 3 files changed, 14 insertions(+), 18 deletions(-)
>
> diff --git a/arch/loongarch/include/asm/kvm_host.h b/arch/loongarch/include/asm/kvm_host.h
> index 19eb5e5c3984..7d739fd5cda6 100644
> --- a/arch/loongarch/include/asm/kvm_host.h
> +++ b/arch/loongarch/include/asm/kvm_host.h
> @@ -203,6 +203,7 @@ struct kvm_vcpu_arch {
> /* Which auxiliary state is loaded (KVM_LARCH_*) */
> unsigned int aux_inuse;
> unsigned int aux_ldtype;
> + unsigned int aux_used;
>
> /* FPU state */
> struct loongarch_fpu fpu FPU_ALIGN;
> diff --git a/arch/loongarch/kvm/exit.c b/arch/loongarch/kvm/exit.c
> index da0ad89f2eb7..da8e330c8b20 100644
> --- a/arch/loongarch/kvm/exit.c
> +++ b/arch/loongarch/kvm/exit.c
> @@ -754,7 +754,7 @@ static int kvm_handle_fpu_disabled(struct kvm_vcpu *vcpu, int ecode)
> return RESUME_HOST;
> }
>
> - vcpu->arch.aux_ldtype = KVM_LARCH_FPU;
> + vcpu->arch.aux_ldtype = vcpu->arch.aux_used | KVM_LARCH_FPU;
> kvm_make_request(KVM_REQ_AUX_LOAD, vcpu);
>
> return RESUME_GUEST;
> @@ -796,7 +796,7 @@ static int kvm_handle_lsx_disabled(struct kvm_vcpu *vcpu, int ecode)
> if (!kvm_guest_has_lsx(&vcpu->arch))
> kvm_queue_exception(vcpu, EXCCODE_INE, 0);
> else {
> - vcpu->arch.aux_ldtype = KVM_LARCH_LSX;
> + vcpu->arch.aux_ldtype = vcpu->arch.aux_used | KVM_LARCH_LSX;
> kvm_make_request(KVM_REQ_AUX_LOAD, vcpu);
> }
>
> @@ -816,7 +816,7 @@ static int kvm_handle_lasx_disabled(struct kvm_vcpu *vcpu, int ecode)
> if (!kvm_guest_has_lasx(&vcpu->arch))
> kvm_queue_exception(vcpu, EXCCODE_INE, 0);
> else {
> - vcpu->arch.aux_ldtype = KVM_LARCH_LASX;
> + vcpu->arch.aux_ldtype = vcpu->arch.aux_used | KVM_LARCH_LASX;
> kvm_make_request(KVM_REQ_AUX_LOAD, vcpu);
> }
>
> diff --git a/arch/loongarch/kvm/vcpu.c b/arch/loongarch/kvm/vcpu.c
> index 8810fcd7e26e..0a111c74a330 100644
> --- a/arch/loongarch/kvm/vcpu.c
> +++ b/arch/loongarch/kvm/vcpu.c
> @@ -239,23 +239,14 @@ static void kvm_late_check_requests(struct kvm_vcpu *vcpu)
> }
>
> if (kvm_check_request(KVM_REQ_AUX_LOAD, vcpu)) {
> - switch (vcpu->arch.aux_ldtype) {
> - case KVM_LARCH_FPU:
> - kvm_own_fpu(vcpu);
> - break;
> - case KVM_LARCH_LSX:
> - kvm_own_lsx(vcpu);
> - break;
> - case KVM_LARCH_LASX:
> + if (vcpu->arch.aux_ldtype & KVM_LARCH_LASX)
> kvm_own_lasx(vcpu);
> - break;
> - case KVM_LARCH_LBT:
> + else if (vcpu->arch.aux_ldtype & KVM_LARCH_LSX)
> + kvm_own_lsx(vcpu);
> + else if (vcpu->arch.aux_ldtype & KVM_LARCH_FPU)
> + kvm_own_fpu(vcpu);
> + else if (vcpu->arch.aux_ldtype == KVM_LARCH_LBT)
> kvm_own_lbt(vcpu);
> - break;
> - default:
> - break;
> - }
> -
> vcpu->arch.aux_ldtype = 0;
> }
> }
> @@ -956,6 +947,7 @@ static int kvm_set_one_reg(struct kvm_vcpu *vcpu,
> break;
> case KVM_REG_LOONGARCH_VCPU_RESET:
> vcpu->arch.st.guest_addr = 0;
> + vcpu->arch.aux_used = 0;
> memset(&vcpu->arch.irq_pending, 0, sizeof(vcpu->arch.irq_pending));
> memset(&vcpu->arch.irq_clear, 0, sizeof(vcpu->arch.irq_clear));
>
> @@ -1384,6 +1376,7 @@ void kvm_own_fpu(struct kvm_vcpu *vcpu)
>
> kvm_restore_fpu(&vcpu->arch.fpu);
> vcpu->arch.aux_inuse |= KVM_LARCH_FPU;
> + vcpu->arch.aux_used |= KVM_LARCH_FPU;
> trace_kvm_aux(vcpu, KVM_TRACE_AUX_RESTORE, KVM_TRACE_AUX_FPU);
> }
>
> @@ -1412,6 +1405,7 @@ int kvm_own_lsx(struct kvm_vcpu *vcpu)
>
> trace_kvm_aux(vcpu, KVM_TRACE_AUX_RESTORE, KVM_TRACE_AUX_LSX);
> vcpu->arch.aux_inuse |= KVM_LARCH_LSX | KVM_LARCH_FPU;
> + vcpu->arch.aux_used |= KVM_LARCH_LSX | KVM_LARCH_FPU;
>
> return 0;
> }
> @@ -1442,6 +1436,7 @@ int kvm_own_lasx(struct kvm_vcpu *vcpu)
>
> trace_kvm_aux(vcpu, KVM_TRACE_AUX_RESTORE, KVM_TRACE_AUX_LASX);
> vcpu->arch.aux_inuse |= KVM_LARCH_LASX | KVM_LARCH_LSX | KVM_LARCH_FPU;
> + vcpu->arch.aux_used |= KVM_LARCH_LASX | KVM_LARCH_LSX | KVM_LARCH_FPU;
>
> return 0;
> }
>
> base-commit: c369299895a591d96745d6492d4888259b004a9e
>
prev parent reply other threads:[~2026-03-25 1:54 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-23 3:40 [PATCH] LoongArch: KVM: Set max used FPU type with FPU exception Bibo Mao
2026-03-25 1:51 ` Bibo Mao [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=d539cf62-c541-ba2e-7810-38ed85c12a62@loongson.cn \
--to=maobibo@loongson.cn \
--cc=chenhuacai@kernel.org \
--cc=kernel@xen0n.name \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=loongarch@lists.linux.dev \
--cc=zhaotianrui@loongson.cn \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).