linux-kernel - Re: [PATCH v2 3/3] KVM: s390: gaccess: Cleanup access to guest frames

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Thu, 28 Oct 2021 16:48:28 +0200
From:   Janis Schoetterl-Glausch <scgl@...ux.vnet.ibm.com>
To:     David Hildenbrand <david@...hat.com>,
        Janis Schoetterl-Glausch <scgl@...ux.ibm.com>,
        Christian Borntraeger <borntraeger@...ibm.com>,
        Janosch Frank <frankja@...ux.ibm.com>,
        Heiko Carstens <hca@...ux.ibm.com>,
        Vasily Gorbik <gor@...ux.ibm.com>
Cc:     Claudio Imbrenda <imbrenda@...ux.ibm.com>,
        Alexander Gordeev <agordeev@...ux.ibm.com>,
        kvm@...r.kernel.org, linux-s390@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 3/3] KVM: s390: gaccess: Cleanup access to guest frames

On 10/28/21 16:25, David Hildenbrand wrote:
> On 28.10.21 15:55, Janis Schoetterl-Glausch wrote:
>> Introduce a helper function for guest frame access.
> 
> "guest page access"

Ok.
> 
> But I do wonder if you actually want to call it
> 
> "access_guest_abs"
> 
> and say "guest absolute access" instead here.
> 
> Because we're dealing with absolute addresses and the fact that we are
> accessing it page-wise is just because we have to perform a page-wise
> translation in the callers (either virtual->absolute or real->absolute).
> 
> Theoretically, if you know you're across X pages but they are contiguous
> in absolute address space, nothing speaks against using that function
> directly across X pages with a single call.

There currently is no point to this, is there?
kvm_read/write_guest break the region up into pages anyway,
so no reason to try to identify larger continuous chunks.
> 
>>
>> Signed-off-by: Janis Schoetterl-Glausch <scgl@...ux.ibm.com>
>> ---
>>  arch/s390/kvm/gaccess.c | 24 ++++++++++++++++--------
>>  1 file changed, 16 insertions(+), 8 deletions(-)
>>
>> diff --git a/arch/s390/kvm/gaccess.c b/arch/s390/kvm/gaccess.c
>> index f0848c37b003..9a633310b6fe 100644
>> --- a/arch/s390/kvm/gaccess.c
>> +++ b/arch/s390/kvm/gaccess.c
>> @@ -866,6 +866,20 @@ static int guest_range_to_gpas(struct kvm_vcpu *vcpu, unsigned long ga, u8 ar,
>>  	return 0;
>>  }
>>  
>> +static int access_guest_page(struct kvm *kvm, enum gacc_mode mode, gpa_t gpa,
>> +			      void *data, unsigned int len)
>> +{
>> +	const unsigned int offset = offset_in_page(gpa);
>> +	const gfn_t gfn = gpa_to_gfn(gpa);
>> +	int rc;
>> +
>> +	if (mode == GACC_STORE)
>> +		rc = kvm_write_guest_page(kvm, gfn, data, offset, len);
>> +	else
>> +		rc = kvm_read_guest_page(kvm, gfn, data, offset, len);
>> +	return rc;
>> +}
>> +
>>  int access_guest(struct kvm_vcpu *vcpu, unsigned long ga, u8 ar, void *data,
>>  		 unsigned long len, enum gacc_mode mode)
>>  {
>> @@ -896,10 +910,7 @@ int access_guest(struct kvm_vcpu *vcpu, unsigned long ga, u8 ar, void *data,
>>  	rc = guest_range_to_gpas(vcpu, ga, ar, gpas, len, asce, mode);
>>  	for (idx = 0; idx < nr_pages && !rc; idx++) {
>>  		fragment_len = min(PAGE_SIZE - offset_in_page(gpas[idx]), len);
>> -		if (mode == GACC_STORE)
>> -			rc = kvm_write_guest(vcpu->kvm, gpas[idx], data, fragment_len);
>> -		else
>> -			rc = kvm_read_guest(vcpu->kvm, gpas[idx], data, fragment_len);
>> +		rc = access_guest_page(vcpu->kvm, mode, gpas[idx], data, fragment_len);
>>  		len -= fragment_len;
>>  		data += fragment_len;
>>  	}
>> @@ -920,10 +931,7 @@ int access_guest_real(struct kvm_vcpu *vcpu, unsigned long gra,
>>  	while (len && !rc) {
>>  		gpa = kvm_s390_real_to_abs(vcpu, gra);
>>  		fragment_len = min(PAGE_SIZE - offset_in_page(gpa), len);
>> -		if (mode)
>> -			rc = write_guest_abs(vcpu, gpa, data, fragment_len);
>> -		else
>> -			rc = read_guest_abs(vcpu, gpa, data, fragment_len);
>> +		rc = access_guest_page(vcpu->kvm, mode, gpa, data, fragment_len);
>>  		len -= fragment_len;
>>  		gra += fragment_len;
>>  		data += fragment_len;
>>
> 
>