mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Suzuki K Poulose <suzuki.poulose@arm.com>
To: Steven Price <steven.price@arm.com>,
	Gavin Shan <gshan@redhat.com>,
	kvm@vger.kernel.org, kvmarm@lists.linux.dev
Cc: Catalin Marinas <catalin.marinas@arm.com>,
	Marc Zyngier <maz@kernel.org>, Will Deacon <will@kernel.org>,
	James Morse <james.morse@arm.com>,
	Oliver Upton <oliver.upton@linux.dev>,
	Zenghui Yu <yuzenghui@huawei.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, Joey Gouly <joey.gouly@arm.com>,
	Alexandru Elisei <alexandru.elisei@arm.com>,
	Christoffer Dall <christoffer.dall@arm.com>,
	Fuad Tabba <tabba@google.com>,
	linux-coco@lists.linux.dev,
	Ganapatrao Kulkarni <gankulkarni@os.amperecomputing.com>,
	Shanker Donthineni <sdonthineni@nvidia.com>,
	Alper Gun <alpergun@google.com>,
	"Aneesh Kumar K . V" <aneesh.kumar@kernel.org>
Subject: Re: [PATCH v6 12/43] arm64: RME: Allocate/free RECs to match vCPUs
Date: Mon, 3 Feb 2025 11:18:06 +0000	[thread overview]
Message-ID: <d499c3e7-cc0f-467a-8401-bc53b70259f7@arm.com> (raw)
In-Reply-To: <9174bc24-6217-4663-a370-291d4790a212@arm.com>

On 30/01/2025 16:40, Steven Price wrote:
> On 29/01/2025 04:50, Gavin Shan wrote:
>> On 12/13/24 1:55 AM, Steven Price wrote:
>>> The RMM maintains a data structure known as the Realm Execution Context
>>> (or REC). It is similar to struct kvm_vcpu and tracks the state of the
>>> virtual CPUs. KVM must delegate memory and request the structures are
>>> created when vCPUs are created, and suitably tear down on destruction.
>>>
>>> RECs must also be supplied with addition pages - auxiliary (or AUX)
>>> granules - for storing the larger registers state (e.g. for SVE). The
>>> number of AUX granules for a REC depends on the parameters with which
>>> the Realm was created - the RMM makes this information available via the
>>> RMI_REC_AUX_COUNT call performed after creating the Realm Descriptor
>>> (RD).
>>>
>>> Note that only some of register state for the REC can be set by KVM, the
>>> rest is defined by the RMM (zeroed). The register state then cannot be
>>> changed by KVM after the REC is created (except when the guest
>>> explicitly requests this e.g. by performing a PSCI call). The RMM also
>>> requires that the VMM creates RECs in ascending order of the MPIDR.
>>>
>>> See Realm Management Monitor specification (DEN0137) for more
>>> information:
>>> https://developer.arm.com/documentation/den0137/
>>>
>>> Signed-off-by: Steven Price <steven.price@arm.com>
>>> ---
>>> Changes since v5:
>>>    * Separate the concept of vcpu_is_rec() and
>>>      kvm_arm_vcpu_rec_finalized() by using the KVM_ARM_VCPU_REC feature as
>>>      the indication that the VCPU is a REC.
>>> Changes since v2:
>>>    * Free rec->run earlier in kvm_destroy_realm() and adapt to previous
>>> patches.
>>> ---
>>>    arch/arm64/include/asm/kvm_emulate.h |   7 ++
>>>    arch/arm64/include/asm/kvm_host.h    |   3 +
>>>    arch/arm64/include/asm/kvm_rme.h     |  18 ++++
>>>    arch/arm64/kvm/arm.c                 |   9 ++
>>>    arch/arm64/kvm/reset.c               |  11 ++
>>>    arch/arm64/kvm/rme.c                 | 144 +++++++++++++++++++++++++++
>>>    6 files changed, 192 insertions(+)
>>>
>>> diff --git a/arch/arm64/include/asm/kvm_emulate.h b/arch/arm64/
>>> include/asm/kvm_emulate.h
>>> index 27f54a7778aa..ec2b6d9c9c07 100644
>>> --- a/arch/arm64/include/asm/kvm_emulate.h
>>> +++ b/arch/arm64/include/asm/kvm_emulate.h
>>> @@ -722,7 +722,14 @@ static inline bool kvm_realm_is_created(struct
>>> kvm *kvm)
>>>      static inline bool vcpu_is_rec(struct kvm_vcpu *vcpu)
>>>    {
>>> +    if (static_branch_unlikely(&kvm_rme_is_available))
>>> +        return vcpu_has_feature(vcpu, KVM_ARM_VCPU_REC);
>>>        return false;
>>>    }
>>>    
>>
>> It seems the check on kvm_rme_is_available is unnecessary because
>> KVM_ARM_VCPU_REC
>> is possible to be true only when kvm_rme_is_available is true.
> 
> Similar to a previous patch - the check of the static key is to avoid
> overhead on systems without RME.
> 
>>> +static inline bool kvm_arm_vcpu_rec_finalized(struct kvm_vcpu *vcpu)
>>> +{
>>> +    return vcpu->arch.rec.mpidr != INVALID_HWID;
>>> +}
>>> +
>>>    #endif /* __ARM64_KVM_EMULATE_H__ */
>>
>> I would suggest to rename to kvm_arm_rec_finalized() since vCPU and REC are
>> similar objects at the same level. It'd better to avoid duplicate object
>> name reference in the function name.
> 
> Ack
> 
>>> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/
>>> asm/kvm_host.h
>>> index 8482638dce3b..220195c727ef 100644
>>> --- a/arch/arm64/include/asm/kvm_host.h
>>> +++ b/arch/arm64/include/asm/kvm_host.h
>>> @@ -789,6 +789,9 @@ struct kvm_vcpu_arch {
>>>          /* Per-vcpu CCSIDR override or NULL */
>>>        u32 *ccsidr;
>>> +
>>> +    /* Realm meta data */
>>> +    struct realm_rec rec;
>>>    };
>>>      /*
>>> diff --git a/arch/arm64/include/asm/kvm_rme.h b/arch/arm64/include/
>>> asm/kvm_rme.h
>>> index 32bdedf1d866..62d4a63d3035 100644
>>> --- a/arch/arm64/include/asm/kvm_rme.h
>>> +++ b/arch/arm64/include/asm/kvm_rme.h
>>> @@ -6,6 +6,7 @@
>>>    #ifndef __ASM_KVM_RME_H
>>>    #define __ASM_KVM_RME_H
>>>    +#include <asm/rmi_smc.h>
>>>    #include <uapi/linux/kvm.h>
>>>      /**
>>> @@ -65,6 +66,21 @@ struct realm {
>>>        unsigned int ia_bits;
>>>    };
>>>    +/**
>>> + * struct realm_rec - Additional per VCPU data for a Realm
>>> + *
>>> + * @mpidr: MPIDR (Multiprocessor Affinity Register) value to identify
>>> this VCPU
>>> + * @rec_page: Kernel VA of the RMM's private page for this REC
>>> + * @aux_pages: Additional pages private to the RMM for this REC
>>> + * @run: Kernel VA of the RmiRecRun structure shared with the RMM
>>> + */
>>> +struct realm_rec {
>>> +    unsigned long mpidr;
>>> +    void *rec_page;
>>> +    struct page *aux_pages[REC_PARAMS_AUX_GRANULES];
>>> +    struct rec_run *run;
>>> +};
>>> +
>>>    void kvm_init_rme(void);
>>>    u32 kvm_realm_ipa_limit(void);
>>>    @@ -72,6 +88,8 @@ int kvm_realm_enable_cap(struct kvm *kvm, struct
>>> kvm_enable_cap *cap);
>>>    int kvm_init_realm_vm(struct kvm *kvm);
>>>    void kvm_destroy_realm(struct kvm *kvm);
>>>    void kvm_realm_destroy_rtts(struct kvm *kvm, u32 ia_bits);
>>> +int kvm_create_rec(struct kvm_vcpu *vcpu);
>>> +void kvm_destroy_rec(struct kvm_vcpu *vcpu);
>>>      #define RMM_RTT_BLOCK_LEVEL    2
>>>    #define RMM_RTT_MAX_LEVEL    3
>>> diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
>>> index 73016e1e0067..2d97147651be 100644
>>> --- a/arch/arm64/kvm/arm.c
>>> +++ b/arch/arm64/kvm/arm.c
>>> @@ -525,6 +525,8 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu)
>>>        /* Force users to call KVM_ARM_VCPU_INIT */
>>>        vcpu_clear_flag(vcpu, VCPU_INITIALIZED);
>>>    +    vcpu->arch.rec.mpidr = INVALID_HWID;
>>> +
>>>        vcpu->arch.mmu_page_cache.gfp_zero = __GFP_ZERO;
>>>          /* Set up the timer */
>>> @@ -1467,6 +1469,9 @@ static unsigned long
>>> system_supported_vcpu_features(void)
>>>        if (!cpus_have_final_cap(ARM64_HAS_NESTED_VIRT))
>>>            clear_bit(KVM_ARM_VCPU_HAS_EL2, &features);
>>>    +    if (!static_branch_unlikely(&kvm_rme_is_available))
>>> +        clear_bit(KVM_ARM_VCPU_REC, &features);
>>> +
>>>        return features;
>>>    }
>>>    @@ -1506,6 +1511,10 @@ static int
>>> kvm_vcpu_init_check_features(struct kvm_vcpu *vcpu,
>>>        if (test_bit(KVM_ARM_VCPU_HAS_EL2, &features))
>>>            return -EINVAL;
>>>    +    /* RME is incompatible with AArch32 */
>>> +    if (test_bit(KVM_ARM_VCPU_REC, &features))
>>> +        return -EINVAL;
>>> +
>>>        return 0;
>>>    }
>>>    
>>
>> One corner case seems missed in kvm_vcpu_init_check_features(). It's
>> allowed to
>> initialize a vCPU with REC feature even kvm_is_realm(kvm) is false.
>> Hopefully,
>> I didn't miss something.
> 
> Ah, yes good point. I'll pass a kvm pointer to
> kvm_vcpu_init_check_features() and use kvm_is_realm() in there.

nit: kvm is available from the VCPU via vcpu->kvm

Cheers
Suzuki



  reply	other threads:[~2025-02-03 11:18 UTC|newest]

Thread overview: 90+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-12 15:55 [PATCH v6 00/43] arm64: Support for Arm CCA in KVM Steven Price
2024-12-12 15:55 ` [PATCH v6 01/43] KVM: Prepare for handling only shared mappings in mmu_notifier events Steven Price
2024-12-12 15:55 ` [PATCH v6 02/43] kvm: arm64: Include kvm_emulate.h in kvm/arm_psci.h Steven Price
2024-12-12 15:55 ` [PATCH v6 03/43] arm64: RME: Handle Granule Protection Faults (GPFs) Steven Price
2024-12-16 16:45   ` Suzuki K Poulose
2024-12-12 15:55 ` [PATCH v6 04/43] arm64: RME: Add SMC definitions for calling the RMM Steven Price
2024-12-16 17:51   ` Suzuki K Poulose
2024-12-12 15:55 ` [PATCH v6 05/43] arm64: RME: Add wrappers for RMI calls Steven Price
2024-12-12 15:55 ` [PATCH v6 06/43] arm64: RME: Check for RME support at KVM init Steven Price
2024-12-19  5:44   ` Aneesh Kumar K.V
2025-01-28  5:47     ` Gavin Shan
2025-01-29  3:57   ` Gavin Shan
2025-01-29 15:24     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 07/43] arm64: RME: Define the user ABI Steven Price
2025-01-28 23:51   ` Gavin Shan
2025-01-29 16:31     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 08/43] arm64: RME: ioctls to create and configure realms Steven Price
2025-01-29  0:43   ` Gavin Shan
2025-01-30 11:57     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 09/43] kvm: arm64: Expose debug HW register numbers for Realm Steven Price
2024-12-12 15:55 ` [PATCH v6 10/43] arm64: kvm: Allow passing machine type in KVM creation Steven Price
2025-01-29  4:07   ` Gavin Shan
2025-01-30 14:14     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 11/43] arm64: RME: RTT tear down Steven Price
2025-01-29 23:06   ` Gavin Shan
2024-12-12 15:55 ` [PATCH v6 12/43] arm64: RME: Allocate/free RECs to match vCPUs Steven Price
2024-12-19  5:37   ` Aneesh Kumar K.V
2025-01-29  4:50   ` Gavin Shan
2025-01-30 16:40     ` Steven Price
2025-02-03 11:18       ` Suzuki K Poulose [this message]
2025-02-03 16:34         ` Steven Price
2024-12-12 15:55 ` [PATCH v6 13/43] KVM: arm64: vgic: Provide helper for number of list registers Steven Price
2024-12-12 15:55 ` [PATCH v6 14/43] arm64: RME: Support for the VGIC in realms Steven Price
2024-12-12 15:55 ` [PATCH v6 15/43] KVM: arm64: Support timers in realm RECs Steven Price
2024-12-12 15:55 ` [PATCH v6 16/43] arm64: RME: Allow VMM to set RIPAS Steven Price
2025-01-29 23:25   ` Gavin Shan
2025-01-30 16:56     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 17/43] arm64: RME: Handle realm enter/exit Steven Price
2025-01-29 23:41   ` Gavin Shan
2025-02-03 16:34     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 18/43] KVM: arm64: Handle realm MMIO emulation Steven Price
2024-12-12 15:55 ` [PATCH v6 19/43] arm64: RME: Allow populating initial contents Steven Price
2025-01-30  4:38   ` Gavin Shan
2025-02-03 16:52     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 20/43] arm64: RME: Runtime faulting of memory Steven Price
2025-01-30  5:22   ` Gavin Shan
2025-02-06  4:33     ` Aneesh Kumar K.V
2025-02-07 17:04     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 21/43] KVM: arm64: Handle realm VCPU load Steven Price
2024-12-12 15:55 ` [PATCH v6 22/43] KVM: arm64: Validate register access for a Realm VM Steven Price
2025-02-02  1:22   ` Gavin Shan
2025-02-07 17:04     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 23/43] KVM: arm64: Handle Realm PSCI requests Steven Price
2025-02-02  2:06   ` Gavin Shan
2024-12-12 15:55 ` [PATCH v6 24/43] KVM: arm64: WARN on injected undef exceptions Steven Price
2025-02-02  2:11   ` Gavin Shan
2024-12-12 15:55 ` [PATCH v6 25/43] arm64: Don't expose stolen time for realm guests Steven Price
2025-02-02  2:15   ` Gavin Shan
2025-02-07 17:05     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 26/43] arm64: rme: allow userspace to inject aborts Steven Price
2024-12-12 15:55 ` [PATCH v6 27/43] arm64: rme: support RSI_HOST_CALL Steven Price
2025-02-02  6:41   ` Gavin Shan
2025-02-07 17:05     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 28/43] arm64: rme: Allow checking SVE on VM instance Steven Price
2025-02-02  6:00   ` Gavin Shan
2025-02-07 17:05     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 29/43] arm64: RME: Always use 4k pages for realms Steven Price
2025-02-02  6:52   ` Gavin Shan
2025-02-07 17:05     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 30/43] arm64: rme: Prevent Device mappings for Realms Steven Price
2025-02-02  7:12   ` Gavin Shan
2025-02-07 17:08     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 31/43] arm_pmu: Provide a mechanism for disabling the physical IRQ Steven Price
2024-12-12 15:55 ` [PATCH v6 32/43] arm64: rme: Enable PMU support with a realm guest Steven Price
2024-12-12 16:54   ` Joey Gouly
2025-02-10 11:58     ` Steven Price
2024-12-12 15:55 ` [PATCH v6 33/43] kvm: rme: Hide KVM_CAP_READONLY_MEM for realm guests Steven Price
2024-12-12 15:55 ` [PATCH v6 34/43] arm64: RME: Propagate number of breakpoints and watchpoints to userspace Steven Price
2025-02-02 23:15   ` Gavin Shan
2025-02-02 23:17     ` Gavin Shan
2024-12-12 15:56 ` [PATCH v6 35/43] arm64: RME: Set breakpoint parameters through SET_ONE_REG Steven Price
2024-12-12 15:56 ` [PATCH v6 36/43] arm64: RME: Initialize PMCR.N with number counter supported by RMM Steven Price
2024-12-12 15:56 ` [PATCH v6 37/43] arm64: RME: Propagate max SVE vector length from RMM Steven Price
2024-12-12 15:56 ` [PATCH v6 38/43] arm64: RME: Configure max SVE vector length for a Realm Steven Price
2025-02-02 23:26   ` Gavin Shan
2024-12-12 15:56 ` [PATCH v6 39/43] arm64: RME: Provide register list for unfinalized RME RECs Steven Price
2024-12-12 15:56 ` [PATCH v6 40/43] arm64: RME: Provide accurate register list Steven Price
2024-12-12 15:56 ` [PATCH v6 41/43] arm64: kvm: Expose support for private memory Steven Price
2024-12-12 15:56 ` [PATCH v6 42/43] KVM: arm64: Expose KVM_ARM_VCPU_REC to user space Steven Price
2024-12-12 15:56 ` [PATCH v6 43/43] KVM: arm64: Allow activating realms Steven Price

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=d499c3e7-cc0f-467a-8401-bc53b70259f7@arm.com \
    --to=suzuki.poulose@arm.com \
    --cc=alexandru.elisei@arm.com \
    --cc=alpergun@google.com \
    --cc=aneesh.kumar@kernel.org \
    --cc=catalin.marinas@arm.com \
    --cc=christoffer.dall@arm.com \
    --cc=gankulkarni@os.amperecomputing.com \
    --cc=gshan@redhat.com \
    --cc=james.morse@arm.com \
    --cc=joey.gouly@arm.com \
    --cc=kvm@vger.kernel.org \
    --cc=kvmarm@lists.linux.dev \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-coco@lists.linux.dev \
    --cc=linux-kernel@vger.kernel.org \
    --cc=maz@kernel.org \
    --cc=oliver.upton@linux.dev \
    --cc=sdonthineni@nvidia.com \
    --cc=steven.price@arm.com \
    --cc=tabba@google.com \
    --cc=will@kernel.org \
    --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

all inboxes | Powered by JetHome®