[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <46802b9895a72c374b86399ca008b89a@kernel.org>
Date: Fri, 20 Mar 2020 09:17:30 +0000
From: Marc Zyngier <maz@...nel.org>
To: Auger Eric <eric.auger@...hat.com>
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
Jason Cooper <jason@...edaemon.net>, kvm@...r.kernel.org,
Suzuki K Poulose <suzuki.poulose@....com>,
linux-kernel@...r.kernel.org,
Robert Richter <rrichter@...vell.com>,
James Morse <james.morse@....com>,
Julien Thierry <julien.thierry.kdev@...il.com>,
Zenghui Yu <yuzenghui@...wei.com>,
Thomas Gleixner <tglx@...utronix.de>,
kvmarm@...ts.cs.columbia.edu, linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH v5 19/23] KVM: arm64: GICv4.1: Allow SGIs to switch
between HW and SW interrupts
Hi Eric,
On 2020-03-19 20:13, Auger Eric wrote:
> Hi Marc,
>
> On 3/19/20 8:52 PM, Marc Zyngier wrote:
>> The assumption here is that we're coming vgic_v4_configure_vsgis(),
>> which starts
>> by stopping the whole guest. My guess is that it should be safe
>> enough, but
>> maybe you are thinking of something else?
> I don't have a specific case in mind. Just preferred asking to make
> sure. Usually when touching those fields we take the lock (that's also
> the case in vgic_debug_show for instance).
Ah, good thing you mention the debug interface. I think it is the only
thing that can run behind our back... Fair enough, I'll add the locking.
Thanks,
M.
--
Jazz is not dead. It just smells funny...
Powered by blists - more mailing lists