Linux KVM/arm64 development list
 help / color / mirror / Atom feed
From: Marc Zyngier <maz@kernel.org>
To: Oliver Upton <oliver.upton@linux.dev>
Cc: kvmarm@lists.linux.dev, Joey Gouly <joey.gouly@arm.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>,
	Zenghui Yu <yuzenghui@huawei.com>,
	Mingwei Zhang <mizhang@google.com>,
	Colton Lewis <coltonlewis@google.com>,
	Alexandru Elisei <alexandru.elisei@arm.com>
Subject: Re: [PATCH 08/15] KVM: arm64: Select debug state to save/restore based on debug owner
Date: Sat, 09 Nov 2024 11:57:46 +0000	[thread overview]
Message-ID: <861pzk1ux1.wl-maz@kernel.org> (raw)
In-Reply-To: <20241108222418.1677420-9-oliver.upton@linux.dev>

On Fri, 08 Nov 2024 22:24:12 +0000,
Oliver Upton <oliver.upton@linux.dev> wrote:
> 
> Select the set of debug registers to use based on the owner rather than
> relying on debug_ptr.
> 
> Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
> ---
>  arch/arm64/include/asm/kvm_host.h         | 21 +++++++++++++++++++++
>  arch/arm64/kvm/hyp/include/hyp/debug-sr.h |  8 ++++----
>  2 files changed, 25 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
> index d76e4e44b65b..2d9aee66d5d4 100644
> --- a/arch/arm64/include/asm/kvm_host.h
> +++ b/arch/arm64/include/asm/kvm_host.h
> @@ -1357,9 +1357,30 @@ void kvm_handle_debug_access(struct kvm_vcpu *vcpu);
>  #define kvm_vcpu_os_lock_enabled(vcpu)		\
>  	(!!(__vcpu_sys_reg(vcpu, OSLSR_EL1) & OSLSR_EL1_OSLK))
>  
> +#define kvm_debug_regs_in_use(vcpu)		\
> +	((vcpu)->arch.debug_owner != VCPU_DEBUG_FREE)
>  #define kvm_host_owns_debug_regs(vcpu)		\
>  	((vcpu)->arch.debug_owner == VCPU_DEBUG_HOST_OWNED)
>  
> +#define vcpu_debug_regs(vcpu)					\
> +({								\
> +	struct kvm_guest_debug_arch *__d;			\
> +								\
> +	switch ((vcpu)->arch.debug_owner) {			\
> +	case VCPU_DEBUG_FREE:					\
> +		WARN_ON_ONCE(1);				\
> +		fallthrough;					\
> +	case VCPU_DEBUG_GUEST_OWNED:				\
> +		__d = &(vcpu)->arch.vcpu_debug_state;		\
> +		break;						\
> +	case VCPU_DEBUG_HOST_OWNED:				\
> +		__d = &(vcpu)->arch.external_debug_state;	\
> +		break;						\
> +	}							\
> +								\
> +	__d;							\
> +})

Any particular reason why this can't be an actual function? Or at
least a function taking a kvm_ vcpu_arch as a parameter, and a
preprocessor wrapper for the vcpu->arch pointer chasing?

> +
>  int kvm_arm_vcpu_arch_set_attr(struct kvm_vcpu *vcpu,
>  			       struct kvm_device_attr *attr);
>  int kvm_arm_vcpu_arch_get_attr(struct kvm_vcpu *vcpu,
> diff --git a/arch/arm64/kvm/hyp/include/hyp/debug-sr.h b/arch/arm64/kvm/hyp/include/hyp/debug-sr.h
> index d00093699aaf..acc47f77b3d0 100644
> --- a/arch/arm64/kvm/hyp/include/hyp/debug-sr.h
> +++ b/arch/arm64/kvm/hyp/include/hyp/debug-sr.h
> @@ -132,13 +132,13 @@ static inline void __debug_switch_to_guest_common(struct kvm_vcpu *vcpu)
>  	struct kvm_guest_debug_arch *host_dbg;
>  	struct kvm_guest_debug_arch *guest_dbg;
>  
> -	if (!vcpu_get_flag(vcpu, DEBUG_DIRTY))
> +	if (!kvm_debug_regs_in_use(vcpu))
>  		return;
>  
>  	host_ctxt = host_data_ptr(host_ctxt);
>  	guest_ctxt = &vcpu->arch.ctxt;
>  	host_dbg = host_data_ptr(host_debug_state.regs);
> -	guest_dbg = kern_hyp_va(vcpu->arch.debug_ptr);
> +	guest_dbg = vcpu_debug_regs(vcpu);

One thing you may want to note in the commit log is that this also
allows us to drop a couple of kern_hyp_va() instances, which is always
a good thing.

>  
>  	__debug_save_state(host_dbg, host_ctxt);
>  	__debug_restore_state(guest_dbg, guest_ctxt);
> @@ -151,13 +151,13 @@ static inline void __debug_switch_to_host_common(struct kvm_vcpu *vcpu)
>  	struct kvm_guest_debug_arch *host_dbg;
>  	struct kvm_guest_debug_arch *guest_dbg;
>  
> -	if (!vcpu_get_flag(vcpu, DEBUG_DIRTY))
> +	if (!kvm_debug_regs_in_use(vcpu))
>  		return;
>  
>  	host_ctxt = host_data_ptr(host_ctxt);
>  	guest_ctxt = &vcpu->arch.ctxt;
>  	host_dbg = host_data_ptr(host_debug_state.regs);
> -	guest_dbg = kern_hyp_va(vcpu->arch.debug_ptr);
> +	guest_dbg = vcpu_debug_regs(vcpu);
>  
>  	__debug_save_state(guest_dbg, guest_ctxt);
>  	__debug_restore_state(host_dbg, host_ctxt);

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.

  reply	other threads:[~2024-11-09 11:57 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-11-08 22:24 [PATCH 00/15] KVM: arm64: Debug cleanups Oliver Upton
2024-11-08 22:24 ` [PATCH 01/15] KVM: arm64: Drop MDSCR_EL1_DEBUG_MASK Oliver Upton
2024-11-08 22:24 ` [PATCH 02/15] KVM: arm64: Get rid of __kvm_get_mdcr_el2() and related warts Oliver Upton
2024-11-11 11:00   ` Suzuki K Poulose
2024-11-12  7:22     ` Oliver Upton
2024-11-08 22:24 ` [PATCH 03/15] KVM: arm64: Track presence of SPE/TRBE in kvm_host_data instead of vCPU Oliver Upton
2024-11-11 13:47   ` Suzuki K Poulose
2024-11-11 15:58     ` Suzuki K Poulose
2024-11-11 16:09       ` James Clark
2024-11-11 18:17         ` Oliver Upton
2024-11-08 22:24 ` [PATCH 04/15] KVM: arm64: Move host SME/SVE tracking flags to host data Oliver Upton
2024-11-09 11:39   ` Marc Zyngier
2024-11-08 22:24 ` [PATCH 05/15] KVM: arm64: Evaluate debug owner at vcpu_load() Oliver Upton
2024-11-08 22:24 ` [PATCH 06/15] KVM: arm64: Advance debug_owner state machine for sysreg traps Oliver Upton
2024-11-09 11:47   ` Marc Zyngier
2024-11-08 22:24 ` [PATCH 07/15] KVM: arm64: Clean up KVM_SET_GUEST_DEBUG handler Oliver Upton
2024-11-08 22:24 ` [PATCH 08/15] KVM: arm64: Select debug state to save/restore based on debug owner Oliver Upton
2024-11-09 11:57   ` Marc Zyngier [this message]
2024-11-09 17:13     ` Oliver Upton
2024-11-08 22:24 ` [PATCH 09/15] KVM: arm64: Remove debug tracepoints Oliver Upton
2024-11-09 12:02   ` Marc Zyngier
2024-11-09 13:17     ` Marc Zyngier
2024-11-08 22:24 ` [PATCH 10/15] KVM: arm64: Remove vestiges of debug_ptr Oliver Upton
2024-11-08 22:24 ` [PATCH 11/15] KVM: arm64: Use debug_owner to track if debug regs need save/restore Oliver Upton
2024-11-09 12:11   ` Marc Zyngier
2024-11-09 17:18     ` Oliver Upton
2024-11-09 22:37       ` Marc Zyngier
2024-11-09 23:46         ` Oliver Upton
2024-11-08 22:24 ` [PATCH 12/15] KVM: arm64: Reload vCPU for accesses to OSLAR_EL1 Oliver Upton
2024-11-08 22:24 ` [PATCH 13/15] KVM: arm64: Compute MDCR_EL2 at vcpu_load() Oliver Upton
2024-11-09 12:28   ` Marc Zyngier
2024-11-08 22:24 ` [PATCH 14/15] KVM: arm64: Don't hijack guest context MDSCR_EL1 Oliver Upton
2024-11-09 12:59   ` Marc Zyngier
2024-11-08 22:24 ` [PATCH 15/15] KVM: arm64: Manage software step state at load/put Oliver Upton
2024-11-09 13:13 ` [PATCH 00/15] KVM: arm64: Debug cleanups Marc Zyngier
2024-11-09 17:08   ` Oliver Upton

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=861pzk1ux1.wl-maz@kernel.org \
    --to=maz@kernel.org \
    --cc=alexandru.elisei@arm.com \
    --cc=coltonlewis@google.com \
    --cc=joey.gouly@arm.com \
    --cc=kvmarm@lists.linux.dev \
    --cc=mizhang@google.com \
    --cc=oliver.upton@linux.dev \
    --cc=suzuki.poulose@arm.com \
    --cc=yuzenghui@huawei.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