From: Mario Smarduch <m.smarduch@samsung.com>
To: Marc Zyngier <marc.zyngier@arm.com>,
Christoffer Dall <christoffer.dall@linaro.org>
Cc: linux-arm-kernel@lists.infradead.org,
Catalin Marinas <catalin.marinas@arm.com>,
kvmarm@lists.cs.columbia.edu, kvm@vger.kernel.org,
Ard Biesheuvel <ard.biesheuvel@linaro.org>
Subject: Re: [PATCH v3 07/22] arm64: KVM: Implement system register save/restore
Date: Thu, 10 Dec 2015 19:24:59 -0800 [thread overview]
Message-ID: <566A420B.2080406@samsung.com> (raw)
In-Reply-To: <1449485618-9443-8-git-send-email-marc.zyngier@arm.com>
Hi Marc,
On 12/7/2015 2:53 AM, Marc Zyngier wrote:
> Implement the system register save/restore as a direct translation of
> the assembly code version.
>
> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
> Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
> ---
> arch/arm64/kvm/hyp/Makefile | 1 +
> arch/arm64/kvm/hyp/hyp.h | 3 ++
> arch/arm64/kvm/hyp/sysreg-sr.c | 90 ++++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 94 insertions(+)
> create mode 100644 arch/arm64/kvm/hyp/sysreg-sr.c
>
> diff --git a/arch/arm64/kvm/hyp/Makefile b/arch/arm64/kvm/hyp/Makefile
> index 455dc0a..ec94200 100644
> --- a/arch/arm64/kvm/hyp/Makefile
> +++ b/arch/arm64/kvm/hyp/Makefile
> @@ -5,3 +5,4 @@
> obj-$(CONFIG_KVM_ARM_HOST) += vgic-v2-sr.o
> obj-$(CONFIG_KVM_ARM_HOST) += vgic-v3-sr.o
> obj-$(CONFIG_KVM_ARM_HOST) += timer-sr.o
> +obj-$(CONFIG_KVM_ARM_HOST) += sysreg-sr.o
> diff --git a/arch/arm64/kvm/hyp/hyp.h b/arch/arm64/kvm/hyp/hyp.h
> index f213e46..778d56d 100644
> --- a/arch/arm64/kvm/hyp/hyp.h
> +++ b/arch/arm64/kvm/hyp/hyp.h
> @@ -38,5 +38,8 @@ void __vgic_v3_restore_state(struct kvm_vcpu *vcpu);
> void __timer_save_state(struct kvm_vcpu *vcpu);
> void __timer_restore_state(struct kvm_vcpu *vcpu);
>
> +void __sysreg_save_state(struct kvm_cpu_context *ctxt);
> +void __sysreg_restore_state(struct kvm_cpu_context *ctxt);
> +
> #endif /* __ARM64_KVM_HYP_H__ */
>
> diff --git a/arch/arm64/kvm/hyp/sysreg-sr.c b/arch/arm64/kvm/hyp/sysreg-sr.c
> new file mode 100644
> index 0000000..add8fcb
> --- /dev/null
> +++ b/arch/arm64/kvm/hyp/sysreg-sr.c
> @@ -0,0 +1,90 @@
> +/*
> + * Copyright (C) 2012-2015 - ARM Ltd
> + * Author: Marc Zyngier <marc.zyngier@arm.com>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#include <linux/compiler.h>
> +#include <linux/kvm_host.h>
> +
> +#include <asm/kvm_mmu.h>
> +
> +#include "hyp.h"
> +
I looked closer on some other ways to get better performance out of the
compiler. This code sequence performs about 35% faster for
__sysreg_save_state(..) for 5000 exits you save about 500mS or 100nS per exit.
This is on Juno.
register int volatile count asm("r2") = 0;
do {
....
} while(count);
I didn't test the restore function (ran out of time) but I suspect it should be
the same. The assembler pretty much uses all the GPRs, (a little too many, using
stp to push 4 pairs on the stack and restore) looking at the assembler it all
should execute out of order.
FWIW I gave this a try since compilers like to optimize loops. I used
'cntpct_el0' counter register to measure the intervals.
> +/* ctxt is already in the HYP VA space */
> +void __hyp_text __sysreg_save_state(struct kvm_cpu_context *ctxt)
> +{
> + ctxt->sys_regs[MPIDR_EL1] = read_sysreg(vmpidr_el2);
> + ctxt->sys_regs[CSSELR_EL1] = read_sysreg(csselr_el1);
> + ctxt->sys_regs[SCTLR_EL1] = read_sysreg(sctlr_el1);
> + ctxt->sys_regs[ACTLR_EL1] = read_sysreg(actlr_el1);
> + ctxt->sys_regs[CPACR_EL1] = read_sysreg(cpacr_el1);
> + ctxt->sys_regs[TTBR0_EL1] = read_sysreg(ttbr0_el1);
> + ctxt->sys_regs[TTBR1_EL1] = read_sysreg(ttbr1_el1);
> + ctxt->sys_regs[TCR_EL1] = read_sysreg(tcr_el1);
> + ctxt->sys_regs[ESR_EL1] = read_sysreg(esr_el1);
> + ctxt->sys_regs[AFSR0_EL1] = read_sysreg(afsr0_el1);
> + ctxt->sys_regs[AFSR1_EL1] = read_sysreg(afsr1_el1);
> + ctxt->sys_regs[FAR_EL1] = read_sysreg(far_el1);
> + ctxt->sys_regs[MAIR_EL1] = read_sysreg(mair_el1);
> + ctxt->sys_regs[VBAR_EL1] = read_sysreg(vbar_el1);
> + ctxt->sys_regs[CONTEXTIDR_EL1] = read_sysreg(contextidr_el1);
> + ctxt->sys_regs[TPIDR_EL0] = read_sysreg(tpidr_el0);
> + ctxt->sys_regs[TPIDRRO_EL0] = read_sysreg(tpidrro_el0);
> + ctxt->sys_regs[TPIDR_EL1] = read_sysreg(tpidr_el1);
> + ctxt->sys_regs[AMAIR_EL1] = read_sysreg(amair_el1);
> + ctxt->sys_regs[CNTKCTL_EL1] = read_sysreg(cntkctl_el1);
> + ctxt->sys_regs[PAR_EL1] = read_sysreg(par_el1);
> + ctxt->sys_regs[MDSCR_EL1] = read_sysreg(mdscr_el1);
> +
> + ctxt->gp_regs.regs.sp = read_sysreg(sp_el0);
> + ctxt->gp_regs.regs.pc = read_sysreg(elr_el2);
> + ctxt->gp_regs.regs.pstate = read_sysreg(spsr_el2);
> + ctxt->gp_regs.sp_el1 = read_sysreg(sp_el1);
> + ctxt->gp_regs.elr_el1 = read_sysreg(elr_el1);
> + ctxt->gp_regs.spsr[KVM_SPSR_EL1]= read_sysreg(spsr_el1);
> +}
> +
> +void __hyp_text __sysreg_restore_state(struct kvm_cpu_context *ctxt)
> +{
> + write_sysreg(ctxt->sys_regs[MPIDR_EL1], vmpidr_el2);
> + write_sysreg(ctxt->sys_regs[CSSELR_EL1], csselr_el1);
> + write_sysreg(ctxt->sys_regs[SCTLR_EL1], sctlr_el1);
> + write_sysreg(ctxt->sys_regs[ACTLR_EL1], actlr_el1);
> + write_sysreg(ctxt->sys_regs[CPACR_EL1], cpacr_el1);
> + write_sysreg(ctxt->sys_regs[TTBR0_EL1], ttbr0_el1);
> + write_sysreg(ctxt->sys_regs[TTBR1_EL1], ttbr1_el1);
> + write_sysreg(ctxt->sys_regs[TCR_EL1], tcr_el1);
> + write_sysreg(ctxt->sys_regs[ESR_EL1], esr_el1);
> + write_sysreg(ctxt->sys_regs[AFSR0_EL1], afsr0_el1);
> + write_sysreg(ctxt->sys_regs[AFSR1_EL1], afsr1_el1);
> + write_sysreg(ctxt->sys_regs[FAR_EL1], far_el1);
> + write_sysreg(ctxt->sys_regs[MAIR_EL1], mair_el1);
> + write_sysreg(ctxt->sys_regs[VBAR_EL1], vbar_el1);
> + write_sysreg(ctxt->sys_regs[CONTEXTIDR_EL1], contextidr_el1);
> + write_sysreg(ctxt->sys_regs[TPIDR_EL0], tpidr_el0);
> + write_sysreg(ctxt->sys_regs[TPIDRRO_EL0], tpidrro_el0);
> + write_sysreg(ctxt->sys_regs[TPIDR_EL1], tpidr_el1);
> + write_sysreg(ctxt->sys_regs[AMAIR_EL1], amair_el1);
> + write_sysreg(ctxt->sys_regs[CNTKCTL_EL1], cntkctl_el1);
> + write_sysreg(ctxt->sys_regs[PAR_EL1], par_el1);
> + write_sysreg(ctxt->sys_regs[MDSCR_EL1], mdscr_el1);
> +
> + write_sysreg(ctxt->gp_regs.regs.sp, sp_el0);
> + write_sysreg(ctxt->gp_regs.regs.pc, elr_el2);
> + write_sysreg(ctxt->gp_regs.regs.pstate, spsr_el2);
> + write_sysreg(ctxt->gp_regs.sp_el1, sp_el1);
> + write_sysreg(ctxt->gp_regs.elr_el1, elr_el1);
> + write_sysreg(ctxt->gp_regs.spsr[KVM_SPSR_EL1], spsr_el1);
> +}
>
next prev parent reply other threads:[~2015-12-11 3:23 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-07 10:53 [PATCH v3 00/22] arm64: KVM: Rewriting the world switch in C Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 01/22] arm64: Add macros to read/write system registers Marc Zyngier
2015-12-07 17:35 ` Catalin Marinas
2015-12-07 17:45 ` Mark Rutland
2015-12-07 17:51 ` Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 02/22] arm64: KVM: Add a HYP-specific header file Marc Zyngier
2015-12-11 21:19 ` Christoffer Dall
2015-12-07 10:53 ` [PATCH v3 03/22] arm64: KVM: Implement vgic-v2 save/restore Marc Zyngier
2015-12-11 20:55 ` Christoffer Dall
2015-12-07 10:53 ` [PATCH v3 04/22] KVM: arm/arm64: vgic-v3: Make the LR indexing macro public Marc Zyngier
2015-12-11 20:57 ` Christoffer Dall
2015-12-07 10:53 ` [PATCH v3 05/22] arm64: KVM: Implement vgic-v3 save/restore Marc Zyngier
2015-12-07 16:40 ` Mario Smarduch
2015-12-07 16:52 ` Marc Zyngier
2015-12-07 17:18 ` Mario Smarduch
2015-12-07 17:37 ` Marc Zyngier
2015-12-07 18:05 ` Mario Smarduch
2015-12-07 18:20 ` Marc Zyngier
2015-12-08 2:14 ` Mario Smarduch
2015-12-08 8:19 ` Marc Zyngier
2015-12-11 21:04 ` Christoffer Dall
2015-12-07 10:53 ` [PATCH v3 06/22] arm64: KVM: Implement timer save/restore Marc Zyngier
2015-12-08 2:18 ` Mario Smarduch
2015-12-08 10:02 ` Marc Zyngier
2015-12-11 21:20 ` Christoffer Dall
2015-12-07 10:53 ` [PATCH v3 07/22] arm64: KVM: Implement system register save/restore Marc Zyngier
2015-12-11 3:24 ` Mario Smarduch [this message]
2015-12-11 18:29 ` Marc Zyngier
2015-12-13 4:56 ` Mario Smarduch
2015-12-07 10:53 ` [PATCH v3 08/22] arm64: KVM: Implement 32bit " Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 09/22] arm64: KVM: Implement debug save/restore Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 10/22] arm64: KVM: Implement guest entry Marc Zyngier
2015-12-14 11:06 ` Christoffer Dall
2015-12-07 10:53 ` [PATCH v3 11/22] arm64: KVM: Add patchable function selector Marc Zyngier
2015-12-11 21:21 ` Christoffer Dall
2015-12-07 10:53 ` [PATCH v3 12/22] arm64: KVM: Implement the core world switch Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 13/22] arm64: KVM: Implement fpsimd save/restore Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 14/22] arm64: KVM: Implement TLB handling Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 15/22] arm64: KVM: HYP mode entry points Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 16/22] arm64: KVM: Add panic handling Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 17/22] arm64: KVM: Add compatibility aliases Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 18/22] arm64: KVM: Map the kernel RO section into HYP Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 19/22] arm64: KVM: Move away from the assembly version of the world switch Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 20/22] arm64: KVM: Turn system register numbers to an enum Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 21/22] arm64: KVM: Cleanup asm-offset.c Marc Zyngier
2015-12-07 10:53 ` [PATCH v3 22/22] arm64: KVM: Remove weak attributes Marc Zyngier
2015-12-14 11:07 ` Christoffer Dall
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=566A420B.2080406@samsung.com \
--to=m.smarduch@samsung.com \
--cc=ard.biesheuvel@linaro.org \
--cc=catalin.marinas@arm.com \
--cc=christoffer.dall@linaro.org \
--cc=kvm@vger.kernel.org \
--cc=kvmarm@lists.cs.columbia.edu \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=marc.zyngier@arm.com \
/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 NNTP newsgroup(s).