Re: [PATCH 00/14] KVM: arm64: Parallel stage-2 fault handling

From: Oliver Upton <oliver.upton@linux.dev>
To: Marc Zyngier <maz@kernel.org>
Cc: James Morse <james.morse@arm.com>,
	Alexandru Elisei <alexandru.elisei@arm.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	Quentin Perret <qperret@google.com>,
	Ricardo Koller <ricarkol@google.com>,
	Reiji Watanabe <reijiw@google.com>,
	David Matlack <dmatlack@google.com>,
	Ben Gardon <bgardon@google.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Gavin Shan <gshan@redhat.com>, Peter Xu <peterx@redhat.com>,
	Sean Christopherson <seanjc@google.com>,
	linux-arm-kernel@lists.infradead.org,
	kvmarm@lists.cs.columbia.edu, kvm@vger.kernel.org
Subject: Re: [PATCH 00/14] KVM: arm64: Parallel stage-2 fault handling
Date: Fri, 9 Sep 2022 11:01:58 +0100	[thread overview]
Message-ID: <YxsPFltAMvls/A9n@google.com> (raw)
In-Reply-To: <87o7vsvn4m.wl-maz@kernel.org>

Hey Marc,

On Tue, Sep 06, 2022 at 11:00:09AM +0100, Marc Zyngier wrote:

[...]

> This fails to build on -rc4:
> 
>   MODPOST vmlinux.symvers
>   MODINFO modules.builtin.modinfo
>   GEN     modules.builtin
>   CC      .vmlinux.export.o
>   LD      .tmp_vmlinux.kallsyms1
> ld: Unexpected GOT/PLT entries detected!
> ld: Unexpected run-time procedure linkages detected!
> ld: ID map text too big or misaligned
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_walk':
> (.hyp.text+0xdc0c): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xdc1c): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_get_leaf':
> (.hyp.text+0xdc80): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xdc90): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_hyp_map':
> (.hyp.text+0xddb0): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xddc0): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_hyp_unmap':
> (.hyp.text+0xde44): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xde50): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_hyp_destroy':
> (.hyp.text+0xdf40): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xdf50): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_stage2_map':
> (.hyp.text+0xe16c): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xe17c): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_stage2_set_owner':
> (.hyp.text+0xe264): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xe274): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_stage2_unmap':
> (.hyp.text+0xe2d4): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xe2e4): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_stage2_flush':
> (.hyp.text+0xe5b4): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xe5c4): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> ld: arch/arm64/kvm/hyp/nvhe/kvm_nvhe.o: in function `__kvm_nvhe_kvm_pgtable_stage2_destroy':
> (.hyp.text+0xe6f0): undefined reference to `__kvm_nvhe___rcu_read_lock'
> ld: (.hyp.text+0xe700): undefined reference to `__kvm_nvhe___rcu_read_unlock'
> make[3]: *** [Makefile:1169: vmlinux] Error 1
> make[2]: *** [debian/rules:7: build-arch] Error 2
> 
> as this drags the RCU read-lock into EL2, and that's not going to
> work... The following fixes it, but I wonder how you tested it.

Ugh. I was carrying a patch on top of my series to handle compilation
issues with rseq_test, I managed to squash the equivalent of below in
that patch.

Nonetheless, I *did* actually test it to get the numbers above :)

--
Thanks,
Oliver

> diff --git a/arch/arm64/include/asm/kvm_pgtable.h b/arch/arm64/include/asm/kvm_pgtable.h
> index dc839db86a1a..adf170122daf 100644
> --- a/arch/arm64/include/asm/kvm_pgtable.h
> +++ b/arch/arm64/include/asm/kvm_pgtable.h
> @@ -580,7 +580,7 @@ enum kvm_pgtable_prot kvm_pgtable_stage2_pte_prot(kvm_pte_t pte);
>   */
>  enum kvm_pgtable_prot kvm_pgtable_hyp_pte_prot(kvm_pte_t pte);
>  
> -#if defined(__KVM_NVHE_HYPERVISOR___)
> +#if defined(__KVM_NVHE_HYPERVISOR__)
>  
>  static inline void kvm_pgtable_walk_begin(void) {}
>  static inline void kvm_pgtable_walk_end(void) {}
> 
> -- 
> Without deviation from the norm, progress is not possible.

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel