linux-kernel - Re: [PATCH v2 17/23] KVM: arm64: Account for partitioning in PMCR

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <gsntikkm9kho.fsf@coltonlewis-kvm.c.googlers.com>
Date: Mon, 23 Jun 2025 22:11:15 +0000
From: Colton Lewis <coltonlewis@...gle.com>
To: kernel test robot <lkp@...el.com>
Cc: kvm@...r.kernel.org, oe-kbuild-all@...ts.linux.dev, pbonzini@...hat.com, 
	corbet@....net, linux@...linux.org.uk, catalin.marinas@....com, 
	will@...nel.org, maz@...nel.org, oliver.upton@...ux.dev, joey.gouly@....com, 
	suzuki.poulose@....com, yuzenghui@...wei.com, mark.rutland@....com, 
	skhan@...uxfoundation.org, linux-doc@...r.kernel.org, 
	linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org, 
	kvmarm@...ts.linux.dev, linux-perf-users@...r.kernel.org, 
	linux-kselftest@...r.kernel.org
Subject: Re: [PATCH v2 17/23] KVM: arm64: Account for partitioning in PMCR_EL0 access

kernel test robot <lkp@...el.com> writes:

>     In file included from arch/arm64/include/asm/kvm_host.h:32,
>                      from include/linux/kvm_host.h:45,
>                      from arch/arm64/kvm/sys_regs.c:16:
>     arch/arm64/include/asm/kvm_pmu.h:236:50: warning: 'struct arm_pmu'  
> declared inside parameter list will not be visible outside of this  
> definition or declaration
>      static inline bool kvm_pmu_is_partitioned(struct arm_pmu *pmu)
>                                                       ^~~~~~~
>     arch/arm64/include/asm/kvm_pmu.h:241:52: warning: 'struct arm_pmu'  
> declared inside parameter list will not be visible outside of this  
> definition or declaration
>      static inline u64 kvm_pmu_host_counter_mask(struct arm_pmu *pmu)
>                                                         ^~~~~~~
>     arch/arm64/include/asm/kvm_pmu.h:246:53: warning: 'struct arm_pmu'  
> declared inside parameter list will not be visible outside of this  
> definition or declaration
>      static inline u64 kvm_pmu_guest_counter_mask(struct arm_pmu *pmu)
>                                                          ^~~~~~~
>     arch/arm64/kvm/sys_regs.c:856:6: warning: no previous prototype  
> for 'pmu_access_el0_disabled' [-Wmissing-prototypes]
>      bool pmu_access_el0_disabled(struct kvm_vcpu *vcpu)
>           ^~~~~~~~~~~~~~~~~~~~~~~
>     arch/arm64/kvm/sys_regs.c: In function 'access_pmu_evtyper':
>     arch/arm64/kvm/sys_regs.c:1076:7: error: implicit declaration of  
> function 'kvm_vcpu_pmu_is_partitioned'; did you  
> mean 'kvm_pmu_is_partitioned'? [-Werror=implicit-function-declaration]
>        if (kvm_vcpu_pmu_is_partitioned(vcpu))
>            ^~~~~~~~~~~~~~~~~~~~~~~~~~~
>            kvm_pmu_is_partitioned
>     arch/arm64/kvm/sys_regs.c: In function 'set_pmcr':
>>> arch/arm64/kvm/sys_regs.c:1271:33: error: dereferencing pointer to  
>>> incomplete type 'struct arm_pmu'
>            new_n <= kvm->arch.arm_pmu->hpmn_max))
>                                      ^~
>     cc1: some warnings being treated as errors

Looks like the main problem here is struct arm_pmu is not defined if we
don't have CONFIG_ARM_PMU. I'll provide a dummy definition for that case
and swap out the field reference for a function call.

Secondary problem is making sure there is a prototype for
pmu_access_el0_disabled in that instance.

> vim +1271 arch/arm64/kvm/sys_regs.c

>    1253
>    1254	static int set_pmcr(struct kvm_vcpu *vcpu, const struct  
> sys_reg_desc *r,
>    1255			    u64 val)
>    1256	{
>    1257		u8 new_n = FIELD_GET(ARMV8_PMU_PMCR_N, val);
>    1258		struct kvm *kvm = vcpu->kvm;
>    1259
>    1260		mutex_lock(&kvm->arch.config_lock);
>    1261
>    1262		/*
>    1263		 * The vCPU can't have more counters than the PMU hardware
>    1264		 * implements. Ignore this error to maintain compatibility
>    1265		 * with the existing KVM behavior.
>    1266		 */
>    1267		if (!kvm_vm_has_ran_once(kvm) &&
>    1268		    !vcpu_has_nv(vcpu)	      &&
>    1269		    new_n <= kvm_arm_pmu_get_max_counters(kvm) &&
>    1270		    (!kvm_vcpu_pmu_is_partitioned(vcpu) ||
>> 1271		     new_n <= kvm->arch.arm_pmu->hpmn_max))
>    1272			kvm->arch.nr_pmu_counters = new_n;
>    1273
>    1274		mutex_unlock(&kvm->arch.config_lock);
>    1275
>    1276		/*
>    1277		 * Ignore writes to RES0 bits, read only bits that are cleared on
>    1278		 * vCPU reset, and writable bits that KVM doesn't support yet.
>    1279		 * (i.e. only PMCR.N and bits [7:0] are mutable from userspace)
>    1280		 * The LP bit is RES0 when FEAT_PMUv3p5 is not supported on the  
> vCPU.
>    1281		 * But, we leave the bit as it is here, as the vCPU's PMUver might
>    1282		 * be changed later (NOTE: the bit will be cleared on first vCPU  
> run
>    1283		 * if necessary).
>    1284		 */
>    1285		val &= ARMV8_PMU_PMCR_MASK;
>    1286
>    1287		/* The LC bit is RES1 when AArch32 is not supported */
>    1288		if (!kvm_supports_32bit_el0())
>    1289			val |= ARMV8_PMU_PMCR_LC;
>    1290
>    1291		__vcpu_assign_sys_reg(vcpu, r->reg, val);
>    1292		kvm_make_request(KVM_REQ_RELOAD_PMU, vcpu);
>    1293
>    1294		return 0;
>    1295	}
>    1296

> --
> 0-DAY CI Kernel Test Service
> https://github.com/intel/lkp-tests/wiki