lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <aLlGeLOE_PVsYDAU@linux.dev>
Date: Thu, 4 Sep 2025 00:57:44 -0700
From: Oliver Upton <oliver.upton@...ux.dev>
To: Sascha Bischoff <Sascha.Bischoff@....com>
Cc: "linux-arm-kernel@...ts.infradead.org" <linux-arm-kernel@...ts.infradead.org>,
	"kvmarm@...ts.linux.dev" <kvmarm@...ts.linux.dev>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"kvm@...r.kernel.org" <kvm@...r.kernel.org>, nd <nd@....com>,
	"maz@...nel.org" <maz@...nel.org>, Joey Gouly <Joey.Gouly@....com>,
	Suzuki Poulose <Suzuki.Poulose@....com>,
	"yuzenghui@...wei.com" <yuzenghui@...wei.com>,
	"will@...nel.org" <will@...nel.org>,
	"tglx@...utronix.de" <tglx@...utronix.de>,
	"lpieralisi@...nel.org" <lpieralisi@...nel.org>,
	Timothy Hayes <Timothy.Hayes@....com>
Subject: Re: [PATCH 0/5] KVM: arm64: GICv5 legacy (GCIE_LEGACY) NV enablement
 and cleanup

On Thu, Aug 28, 2025 at 10:59:41AM +0000, Sascha Bischoff wrote:
> Hi all,
> 
> This series enables nested virtualization for GICv3-based VMs on GICv5
> hosts (w/ FEAT_GCIE_LEGACY) in KVM/arm64. In addition, it adds a CPU
> capability to track support for FEAT_GCIE_LEGACY across all CPUs.
> 
> The series fixes ICC_SRE_EL2 access handling for GICv5 hosts (to match
> the updated bet1+ specification [1]), and extends nested
> virtualization support to vGICv3 guests running on compatible GICv5
> systems. With these changes, it becomes possible to run with
> kvm-arm.mode=nested, and these changes have been tested with three
> levels of nesting on simulated hardware (Arm FVP).
> 
> Previously, the presence of FEAT_GCIE_LEGACY was tracked in the GICv5
> driver via gic_kvm_info, and the probing logic could incorrectly
> enable legacy support if the boot CPU exposed the feature while others
> did not. This created the risk of mismatched configurations,
> particularly when late-onlining CPUs without FEAT_GCIE_LEGACY.
> 
> To address this, the series introduces a proper ARM64_HAS_GICV5_LEGACY
> CPU capability, and moves KVM to use cpus_have_final_cap() to ensure
> consistent system-wide enablement. With this, late-onlined but
> mismatched CPUs are cleanly rejected at bring-up.
> 
> Patch summary
> 
> KVM: arm64: allow ICC_SRE_EL2 accesses on a GICv5 host
>     Update handling to reflect the corrected GICv5 specification.
> 
> KVM: arm64: Enable nested for GICv5 host with FEAT_GCIE_LEGACY
>     Allow nested virtualization for vGICv3 guests on GICv5 hosts with
>     legacy support.
> 
> arm64: cpucaps: Add GICv5 Legacy vCPU interface (GCIE_LEGACY) capability
>     Introduce a new CPU capability that prevents mismatched
>     configurations.
> 
> KVM: arm64: Use ARM64_HAS_GICV5_LEGACY for GICv5 probing
>     Ensure probing is consistent across all CPUs by using cpucaps.
> 
> irqchip/gic-v5: Drop has_gcie_v3_compat from gic_kvm_info
>     Remove obsolete compatibility flag, as FEAT_GCIE_LEGACY is now a
>     CPU feature.
> 
> Comments and reviews are very welcome.
> 
> Thanks,
> Sascha
> 
> [1] https://developer.arm.com/documentation/aes0070/latest/
> 
> Sascha Bischoff (5):
>   KVM: arm64: Allow ICC_SRE_EL2 accesses on a GICv5 host
>   KVM: arm64: Enable nested for GICv5 host with FEAT_GCIE_LEGACY
>   arm64: cpucaps: Add GICv5 Legacy vCPU interface (GCIE_LEGACY)
>     capability
>   KVM: arm64: Use ARM64_HAS_GICV5_LEGACY for GICv5 probing
>   irqchip/gic-v5: Drop has_gcie_v3_compat from gic_kvm_info

For the series:

Reviewed-by: Oliver Upton <oliver.upton@...ux.dev>

Thanks,
Oliver

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ