Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Message-ID: <0527baa0-c136-662d-a493-aa5ba423c22e@linux.ibm.com>
Date:   Wed, 9 Feb 2022 14:08:25 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.5.0
Subject: Re: [PATCH v2 05/11] KVM: s390: Add optional storage key checking to
 MEMOP IOCTL
Content-Language: en-US
To:     Christian Borntraeger <borntraeger@linux.ibm.com>,
        Heiko Carstens <hca@linux.ibm.com>,
        Janosch Frank <frankja@linux.ibm.com>
Cc:     Alexander Gordeev <agordeev@linux.ibm.com>,
        Claudio Imbrenda <imbrenda@linux.ibm.com>,
        David Hildenbrand <david@redhat.com>,
        Jonathan Corbet <corbet@lwn.net>, kvm@vger.kernel.org,
        linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-s390@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>,
        Sven Schnelle <svens@linux.ibm.com>,
        Vasily Gorbik <gor@linux.ibm.com>
References: <20220207165930.1608621-1-scgl@linux.ibm.com>
 <20220207165930.1608621-6-scgl@linux.ibm.com>
 <48d1678f-746c-dab6-5ec3-56397277f752@linux.ibm.com>
 <71f07914-d0b2-e98b-22b2-bc05f04df2da@linux.ibm.com>
 <6ea27647-fbbe-3962-03a0-8ca5340fc7fd@linux.ibm.com>
 <8d502356c3a624847c0dd2fe5d5f60e72923a141.camel@linux.ibm.com>
 <3ec91f7a-10ca-b984-d852-1327f965b1e8@linux.ibm.com>
 <83408abf-86fe-20b0-564c-8cf840757e76@linux.ibm.com>
 <ef35d70a-c0dc-e5ae-8182-79847085d593@linux.ibm.com>
 <1eb6ae828fd02340ff30bfab6a949fff90e85d3b.camel@linux.ibm.com>
 <61d9aa7b-4474-fce9-4884-275d1f6dee99@linux.ibm.com>
From:   Janis Schoetterl-Glausch <scgl@linux.ibm.com>
In-Reply-To: <61d9aa7b-4474-fce9-4884-275d1f6dee99@linux.ibm.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 2/9/22 13:11, Christian Borntraeger wrote:
> 
> 
> Am 09.02.22 um 12:04 schrieb Janis Schoetterl-Glausch:
>> On Wed, 2022-02-09 at 11:48 +0100, Christian Borntraeger wrote:
>>>
>>> Am 09.02.22 um 11:39 schrieb Janis Schoetterl-Glausch:
>>>> On 2/9/22 11:08, Christian Borntraeger wrote:
>>>>>
>>>>> Am 09.02.22 um 11:01 schrieb Janis Schoetterl-Glausch:
>>>>>> On Wed, 2022-02-09 at 10:08 +0100, Christian Borntraeger wrote:
>>>>>>> Am 09.02.22 um 09:49 schrieb Janis Schoetterl-Glausch:
>>>>>>>> On 2/9/22 08:34, Christian Borntraeger wrote:
>>>>>>>>> Am 07.02.22 um 17:59 schrieb Janis Schoetterl-Glausch:
>>>>>>>>>> User space needs a mechanism to perform key checked accesses when
>>>>>>>>>> emulating instructions.
>>>>>>>>>>
>>>>>>>>>> The key can be passed as an additional argument.
>>>>>>>>>> Having an additional argument is flexible, as user space can
>>>>>>>>>> pass the guest PSW's key, in order to make an access the same way the
>>>>>>>>>> CPU would, or pass another key if necessary.
>>>>>>>>>>
>>>>>>>>>> Signed-off-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com>
>>>>>>>>>> Acked-by: Janosch Frank <frankja@linux.ibm.com>
>>>>>>>>>> Reviewed-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
>>>>>>>>>> ---
>>>>>>>>>>       arch/s390/kvm/kvm-s390.c | 49 +++++++++++++++++++++++++++++++---------
>>>>>>>>>>       include/uapi/linux/kvm.h |  8 +++++--
>>>>>>>>>>       2 files changed, 44 insertions(+), 13 deletions(-)
>>>>>>>>>>
>>>>>>>>>> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
>>>>>>>>>> index cf347e1a4f17..71e61fb3f0d9 100644
>>>>>>>>>> --- a/arch/s390/kvm/kvm-s390.c
>>>>>>>>>> +++ b/arch/s390/kvm/kvm-s390.c
>>>>>>>>>> @@ -32,6 +32,7 @@
>>>>>>>>>>       #include <linux/sched/signal.h>
>>>>>>>>>>       #include <linux/string.h>
>>>>>>>>>>       #include <linux/pgtable.h>
>>>>>>>>>> +#include <linux/bitfield.h>
>>>>>>>>>>         #include <asm/asm-offsets.h>
>>>>>>>>>>       #include <asm/lowcore.h>
>>>>>>>>>> @@ -2359,6 +2360,11 @@ static int kvm_s390_handle_pv(struct kvm *kvm, struct kvm_pv_cmd *cmd)
>>>>>>>>>>           return r;
>>>>>>>>>>       }
>>>>>>>>>>       +static bool access_key_invalid(u8 access_key)
>>>>>>>>>> +{
>>>>>>>>>> +    return access_key > 0xf;
>>>>>>>>>> +}
>>>>>>>>>> +
>>>>>>>>>>       long kvm_arch_vm_ioctl(struct file *filp,
>>>>>>>>>>                      unsigned int ioctl, unsigned long arg)
>>>>>>>>>>       {
>>>>>>>>>> @@ -4687,34 +4693,54 @@ static long kvm_s390_guest_mem_op(struct kvm_vcpu *vcpu,
>>>>>>>>>>                         struct kvm_s390_mem_op *mop)
>>>>>>>>>>       {
>>>>>>>>>>           void __user *uaddr = (void __user *)mop->buf;
>>>>>>>>>> +    u8 access_key = 0, ar = 0;
>>>>>>>>>>           void *tmpbuf = NULL;
>>>>>>>>>> +    bool check_reserved;
>>>>>>>>>>           int r = 0;
>>>>>>>>>>           const u64 supported_flags = KVM_S390_MEMOP_F_INJECT_EXCEPTION
>>>>>>>>>> -                    | KVM_S390_MEMOP_F_CHECK_ONLY;
>>>>>>>>>> +                    | KVM_S390_MEMOP_F_CHECK_ONLY
>>>>>>>>>> +                    | KVM_S390_MEMOP_F_SKEY_PROTECTION;
>>>>>>>>>>       -    if (mop->flags & ~supported_flags || mop->ar >= NUM_ACRS || !mop->size)
>>>>>>>>>> +    if (mop->flags & ~supported_flags || !mop->size)
>>>>>>>>>>               return -EINVAL;
>>>>>>>>>> -
>>>>>>>>>>           if (mop->size > MEM_OP_MAX_SIZE)
>>>>>>>>>>               return -E2BIG;
>>>>>>>>>> -
>>>>>>>>>>           if (kvm_s390_pv_cpu_is_protected(vcpu))
>>>>>>>>>>               return -EINVAL;
>>>>>>>>>> -
>>>>>>>>>>           if (!(mop->flags & KVM_S390_MEMOP_F_CHECK_ONLY)) {
>>>>>>>>>>               tmpbuf = vmalloc(mop->size);
>>>>>>>>>>               if (!tmpbuf)
>>>>>>>>>>                   return -ENOMEM;
>>>>>>>>>>           }
>>>>>>>>>> +    ar = mop->ar;
>>>>>>>>>> +    mop->ar = 0;
>>>>>>>>>
>>>>>>>>> Why this assignment to 0?
>>>>>>>>
>>>>>>>> It's so the check of reserved below works like that, they're all part of the anonymous union.
>>>>>>>
>>>>>>> Ah, I see. This is ugly :-)
>>>>>>
>>>>>> Yes :)
>>>>>>>>>> +    if (ar >= NUM_ACRS)
>>>>>>>>>> +        return -EINVAL;
>>>>>>>>>> +    if (mop->flags & KVM_S390_MEMOP_F_SKEY_PROTECTION) {
>>>>>>>>>> +        access_key = mop->key;
>>>>>>>>>> +        mop->key = 0;
>>>>>>>>>
>>>>>>>>> and this? I think we can leave mop unchanged.
>>>>>>>>>
>>>>>>>>> In fact, why do we add the ar and access_key variable?
>>>>>>>>> This breaks the check from above (if (mop->flags & ~supported_flags || mop->ar >= NUM_ACRS || !mop->size))  into two checks
>>>>>>>>> and it will create a memleak for tmpbuf.
>>>>>>>>
>>>>>>>> I can move the allocation down, goto out or get rid of the reserved check and keep everything as before.
>>>>>>>> First is simpler, but second makes handling that case more explicit and might help in the future.
>>>>>>>
>>>>>>> Maybe add a reserved_02 field in the anon struct and check this for being zero and get rid of the local variables?
>>>>>>
>>>>>> I think that would require us adding new fields in the struct by putting them in a union with reserved_02 and so on,
>>>>>> which could get rather messy.
>>>>>
>>>>> I think it is fine to rename reserved_02. Maybe rename that to dont_use_02 ?
>>>>
>>>> I don't know what kind of stability guarantees we give here, since it can only happen when recompiling with
>>>> a new header. dont_use is a lot better than reserved here, after all we tell user space to set
>>>> reserved bytes to 0, using reserved_02 to do that would be quite handy and therefore likely.
>>>>
>>>> The question is also what semantic we want for the check.
>>>> The way it works right now, user space also needs to set unused fields to 0, e.g. key if the flag is not set.
>>>> At least this is the case for the vm memop, the vcpu memop cannot do that because of backward compatibility.
>>>
>>> As an alternative just remove the check for reserved == 0 and do that later on as an add-on patch?
>>
>> That would kinda defeat the purpose of the check, since misbehaving user space programs would
>> get an error then but not now.
> 
> 
> As a matter of fact, we do not check today. What about the following.

We don't do it for the vcpu memop, but since we're newly introducing the vm memop we are free to decide what we want.
It's purely about future proofing, e.g. we would have had the possibility to add the key checking feature without a flag,
if the existing memop did the check. Committing ourselves to always adding a flag is fine by me, but I don't like the
previous state of affairs, where user space should set reserved bytes to 0 but it's not enforced.

> 1. remove the checkreserved logic. its too complicated
> 2. do not check for reserved to be zero
> 4. state that the reserved fields are ignored without the appropriate flag
> 5. add the necessary flag as comment to the fields
> 6. check for unkmown flags and bail out

I'll implement this, except maybe 5, since the documentation covers that and the availability of the flags themselves
is conditional on other factors.