[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Zwj4AllH_JjH5xEb@linux.ibm.com>
Date: Fri, 11 Oct 2024 15:33:46 +0530
From: Vishal Chourasia <vishalc@...ux.ibm.com>
To: Paolo Bonzini <pbonzini@...hat.com>
Cc: linuxppc-dev@...ts.ozlabs.org, Michael Ellerman <mpe@...erman.id.au>,
Nicholas Piggin <npiggin@...il.com>,
Christophe Leroy <christophe.leroy@...roup.eu>,
Naveen N Rao <naveen@...nel.org>,
Madhavan Srinivasan <maddy@...ux.ibm.com>, kvm@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [RFC] powerpc/kvm: Fix spinlock member access for PREEMPT_RT
On Thu, Oct 10, 2024 at 11:23:55PM +0200, Paolo Bonzini wrote:
> On 10/10/24 20:09, Vishal Chourasia wrote:
> > Hi,
> >
> > While building the kernel with CONFIG_PREEMPT_RT, I encountered several
> > compilation errors in the PowerPC KVM code. The issues appear in
> > book3s_hv_rm_mmu.c where it tries to access the 'rlock' member of struct
> > spinlock, which doesn't exist in the RT configuration.
>
> How was this tested? I suspect that putting to sleep a task that is running
> in real mode is a huge no-no. The actual solution would have to be to split
> mmu_lock into a spin_lock and a raw_spin_lock, but that's a huge amount of
> work probably. I'd just add a "depends on !PPC || !KVM_BOOK3S_64_HV" or
> something like that, to prevent enabling KVM-HV on PREEMPT_RT kernels.
Hi Paolo,
This is a build time error, I didn't boot the kernel.
I used pseries_le_defconfig with some other configs enabled. I was trying
to see if the kernel would compile with ARCH_SUPPORTS_RT and
CONFIG_PREEMPT_RT enabled.
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 8094a01974cca..568dc856f0dfa 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -168,6 +168,7 @@ config PPC
select ARCH_STACKWALK
select ARCH_SUPPORTS_ATOMIC_RMW
select ARCH_SUPPORTS_DEBUG_PAGEALLOC if PPC_BOOK3S || PPC_8xx
+ select ARCH_SUPPORTS_RT if !PPC || !KVM_BOOK3S_64_HV
select ARCH_USE_BUILTIN_BSWAP
select ARCH_USE_CMPXCHG_LOCKREF if PPC64
select ARCH_USE_MEMTEST
I tried rebuilding with the above diff as per your suggestion
though it works when KVM_BOOK3S_64_HV is set to N, but for
pseries_le_defconfig, it's set to M, by default, which then requires setting it
to N explicitly.
Will something like below be a better solution? This will set
KVM_BOOK3S_64_HV to N if ARCH_SUPPORTS_RT is set.
diff --git a/arch/powerpc/kvm/Kconfig b/arch/powerpc/kvm/Kconfig
index dbfdc126bf144..33e0d50b08b14 100644
--- a/arch/powerpc/kvm/Kconfig
+++ b/arch/powerpc/kvm/Kconfig
@@ -80,7 +80,7 @@ config KVM_BOOK3S_64
config KVM_BOOK3S_64_HV
tristate "KVM for POWER7 and later using hypervisor mode in host"
- depends on KVM_BOOK3S_64 && PPC_POWERNV
+ depends on KVM_BOOK3S_64 && PPC_POWERNV && !ARCH_SUPPORTS_RT
select KVM_BOOK3S_HV_POSSIBLE
select KVM_GENERIC_MMU_NOTIFIER
select CMA
Thanks
Powered by blists - more mailing lists