linux-kernel - Re: [PATCH v6 26/29] KVM: Optimize gfn lookup in kvm_zap_gfn

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 for Android: free password hash cracker in your pocket

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

Message-ID: <ea99b9ba-ea30-f0fc-8c08-34d0aabf57a1@maciej.szmigiero.name>
Date:   Thu, 2 Dec 2021 00:08:31 +0100
From:   "Maciej S. Szmigiero" <mail@...iej.szmigiero.name>
To:     Sean Christopherson <seanjc@...gle.com>,
        Paolo Bonzini <pbonzini@...hat.com>
Cc:     Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>,
        Igor Mammedov <imammedo@...hat.com>,
        Marc Zyngier <maz@...nel.org>,
        James Morse <james.morse@....com>,
        Julien Thierry <julien.thierry.kdev@...il.com>,
        Suzuki K Poulose <suzuki.poulose@....com>,
        Huacai Chen <chenhuacai@...nel.org>,
        Aleksandar Markovic <aleksandar.qemu.devel@...il.com>,
        Paul Mackerras <paulus@...abs.org>,
        Christian Borntraeger <borntraeger@...ibm.com>,
        Janosch Frank <frankja@...ux.ibm.com>,
        David Hildenbrand <david@...hat.com>,
        Cornelia Huck <cohuck@...hat.com>,
        Claudio Imbrenda <imbrenda@...ux.ibm.com>,
        Anup Patel <anup.patel@....com>,
        Paul Walmsley <paul.walmsley@...ive.com>,
        Palmer Dabbelt <palmer@...belt.com>,
        Albert Ou <aou@...s.berkeley.edu>,
        Alexandru Elisei <alexandru.elisei@....com>,
        Ben Gardon <bgardon@...gle.com>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v6 26/29] KVM: Optimize gfn lookup in kvm_zap_gfn_range()

On 01.12.2021 17:36, Sean Christopherson wrote:
> On Wed, Dec 01, 2021, Maciej S. Szmigiero wrote:
>> On 01.12.2021 04:41, Sean Christopherson wrote:
>>> On Tue, Nov 30, 2021, Maciej S. Szmigiero wrote:
>>>> diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
>>>> index 41efe53cf150..6fce6eb797a7 100644
>>>> --- a/include/linux/kvm_host.h
>>>> +++ b/include/linux/kvm_host.h
>>>> @@ -848,6 +848,105 @@ struct kvm_memory_slot *id_to_memslot(struct kvm_memslots *slots, int id)
>>>>    	return NULL;
>>>>    }
>>>> +/* Iterator used for walking memslots that overlap a gfn range. */
>>>> +struct kvm_memslot_iter {
>>>> +	struct kvm_memslots *slots;
>>>> +	gfn_t end;
>>>> +	struct rb_node *node;
>>>> +};
>>>
>>> ...
>>>
>>>> +static inline struct kvm_memory_slot *kvm_memslot_iter_slot(struct kvm_memslot_iter *iter)
>>>> +{
>>>> +	return container_of(iter->node, struct kvm_memory_slot, gfn_node[iter->slots->node_idx]);
>>>
>>> Having to use a helper in callers of kvm_for_each_memslot_in_gfn_range() is a bit
>>> ugly, any reason not to grab @slot as well?  Then the callers just do iter.slot,
>>> which IMO is much more readable.
>>
>> "slot" can be easily calculated from "node" together with either "slots" or
>> "node_idx" (the code above just adjusts a pointer) so storing it in the
>> iterator makes little sense if the later are already stored there.
> 
> I don't want the callers to have to calculate the slot.  It's mostly syntatic
> sugar, but I really do think it improves readability.  And since the first thing
> every caller will do is retrieve the slot, I see no benefit in forcing the caller
> to do the work.
> 
> E.g. in the simple kvm_check_memslot_overlap() usage, iter.slot->id is intuitive
> and easy to parse, whereas kvm_memslot_iter_slot(&iter)->id is slightly more
> difficult to parse and raises questions about why a function call is necessary
> and what the function might be doing.

Personally, I don't think it's that much less readable, but I will change
the code to store "slots" instead (as you wish) since it's the last remaining
change - other than Paolo's call whether we should keep or drop the
kvm_arch_flush_shadow_memslot()-related patch 25.

Thanks,
Maciej