Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Reply-To: Gavin Shan <gshan@redhat.com>
Subject: Re: [PATCH 5/6] KVM: selftests: memslot_perf_test: Consolidate memory
 sizes
To:     "Maciej S. Szmigiero" <mail@maciej.szmigiero.name>
Cc:     kvmarm@lists.cs.columbia.edu, kvmarm@lists.linux.dev,
        kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
        ajones@ventanamicro.com, pbonzini@redhat.com, maz@kernel.org,
        shuah@kernel.org, oliver.upton@linux.dev, seanjc@google.com,
        peterx@redhat.com, ricarkol@google.com, zhenyzha@redhat.com,
        shan.gavin@gmail.com
References: <20221014071914.227134-1-gshan@redhat.com>
 <20221014071914.227134-6-gshan@redhat.com>
 <cebafa0d-a2dc-c3f7-64c8-2637a254e3d0@maciej.szmigiero.name>
From:   Gavin Shan <gshan@redhat.com>
Message-ID: <ac09476d-da8d-db0f-1af2-93dc41251bb0@redhat.com>
Date:   Tue, 18 Oct 2022 09:13:16 +0800
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.2.0
MIME-Version: 1.0
In-Reply-To: <cebafa0d-a2dc-c3f7-64c8-2637a254e3d0@maciej.szmigiero.name>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 10/18/22 5:36 AM, Maciej S. Szmigiero wrote:
> On 14.10.2022 09:19, Gavin Shan wrote:
>> The addresses and sizes passed to madvise() and vm_userspace_mem_region_add()
>> should be aligned to host page size, which can be 64KB on aarch64. So it's
>> wrong by passing additional fixed 4KB memory area to various tests.
>>
>> Fix it by passing additional fixed 64KB memory area to various tests. After
>> it's applied, the following command works fine on 64KB-page-size-host and
>> 4KB-page-size-guest.
>>
>>    # ./memslot_perf_test -v -s 512
>>
>> Signed-off-by: Gavin Shan <gshan@redhat.com>
>> ---
>>   .../testing/selftests/kvm/memslot_perf_test.c  | 18 ++++++++++--------
>>   1 file changed, 10 insertions(+), 8 deletions(-)
>>
>> diff --git a/tools/testing/selftests/kvm/memslot_perf_test.c b/tools/testing/selftests/kvm/memslot_perf_test.c
>> index d587bd952ff9..e6d34744b45d 100644
>> --- a/tools/testing/selftests/kvm/memslot_perf_test.c
>> +++ b/tools/testing/selftests/kvm/memslot_perf_test.c
>> @@ -25,12 +25,14 @@
>>   #include <kvm_util.h>
>>   #include <processor.h>
>> -#define MEM_SIZE        ((512U << 20) + 4096)
>> -#define MEM_GPA        0x10000000UL
>> +#define MEM_EXTRA_SIZE        0x10000
> 
> So the biggest page size supported right now is 64 KiB - it would be
> good to have an assert somewhere to explicitly check for this
> (regardless of implicit checks present in other calculations).
> 
> Also, an expression like "(64 << 10)" is more readable than a "1"
> with a tail of zeroes (it's easy to add one zero too many or be one
> zero short).
> 

Yes, it makes sense to me. Lets add check in check_memory_sizes(), which
was added in the previous patch, to fail early if host/guest page size
exceeds 64KB.

    if (host_page_size > SIZE_64KiB || guest_page_size > SIZE_64KiB) {
       pr_info("Unsupported page size on host (0x%x) or guest (0x%x)\n",
               host_page_size, guest_page_size);
    }


For the macros, I think all of us agree on KiB, MiB, GiB, TiB and
their variants :)

Thanks,
Gavin