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;
Message-ID: <29ed9474-748f-feb3-6cfd-15da9fca217e@linux.ibm.com>
Date:   Thu, 13 Apr 2023 08:46:16 -0500
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0)
 Gecko/20100101 Thunderbird/102.9.1
Subject: Re: [PATCH] Remove POWER10_CPU dependency and move
 PPC_MODULE_FEATURE_P10.
Content-Language: en-US
From:   Danny Tsen <dtsen@linux.ibm.com>
To:     Michael Ellerman <mpe@ellerman.id.au>, linux-crypto@vger.kernel.org
Cc:     herbert@gondor.apana.org.au, leitao@debian.org,
        nayna@linux.ibm.com, appro@cryptogams.org,
        linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
        ltcgcw@linux.vnet.ibm.com, dtsen@us.ibm.com
References: <20230412181232.2051-1-dtsen@linux.ibm.com>
 <87wn2g9b9l.fsf@mpe.ellerman.id.au>
 <71c63813-bbf1-8baa-3b94-f5184cc5e872@linux.ibm.com>
In-Reply-To: <71c63813-bbf1-8baa-3b94-f5184cc5e872@linux.ibm.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 4/13/23 8:18 AM, Danny Tsen wrote:
> Hi Michael,
>
> If I do separate patch for moving PPC_MODULE_FEATURE_P10, this will 
> break the build since it is currently defined in aes-gcm-p10-glue.c.  
> And the p10 will be detected when loading the module in 
> module_cpu_feature_match(PPC_MODULE_FEATURE_P10, p10_init); so it 
> won't load if it's not P10.
>
> Thanks.
>
> -Danny
>
> On 4/13/23 8:12 AM, Michael Ellerman wrote:
>> Danny Tsen <dtsen@linux.ibm.com> writes:
>>> Remove Power10 dependency in Kconfig and detect Power10 feature at 
>>> runtime.
>>> Move PPC_MODULE_FEATURE_P10 definition to be in
>>> arch/powerpc/include/asm/cpufeature.h.
>> This should be two patches, one for the Kconfig change and one moving
>> the feature flag.

I think I misunderstood.  I can do two patches one for Kconfig change 
and one moving the feature flag.  I'll fix it.

Thanks.

-Danny

>>
>> Also don't you need a cpu feature check in p10_init()? Otherwise the
>> driver can be loaded on non-P10 CPUs, either by being built-in, or
>> manually.
>>
>> cheers
>>
>>> Signed-off-by: Danny Tsen <dtsen@linux.ibm.com>
>>> ---
>>>   arch/powerpc/crypto/Kconfig            | 2 +-
>>>   arch/powerpc/crypto/aes-gcm-p10-glue.c | 1 -
>>>   arch/powerpc/include/asm/cpufeature.h  | 1 +
>>>   3 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/arch/powerpc/crypto/Kconfig b/arch/powerpc/crypto/Kconfig
>>> index 1f8f02b494e1..7113f9355165 100644
>>> --- a/arch/powerpc/crypto/Kconfig
>>> +++ b/arch/powerpc/crypto/Kconfig
>>> @@ -96,7 +96,7 @@ config CRYPTO_AES_PPC_SPE
>>>     config CRYPTO_AES_GCM_P10
>>>       tristate "Stitched AES/GCM acceleration support on P10 or 
>>> later CPU (PPC)"
>>> -    depends on PPC64 && POWER10_CPU && CPU_LITTLE_ENDIAN
>>> +    depends on PPC64 && CPU_LITTLE_ENDIAN
>>>       select CRYPTO_LIB_AES
>>>       select CRYPTO_ALGAPI
>>>       select CRYPTO_AEAD
>>> diff --git a/arch/powerpc/crypto/aes-gcm-p10-glue.c 
>>> b/arch/powerpc/crypto/aes-gcm-p10-glue.c
>>> index 1533c8cdd26f..bd3475f5348d 100644
>>> --- a/arch/powerpc/crypto/aes-gcm-p10-glue.c
>>> +++ b/arch/powerpc/crypto/aes-gcm-p10-glue.c
>>> @@ -22,7 +22,6 @@
>>>   #include <linux/module.h>
>>>   #include <linux/types.h>
>>>   -#define PPC_MODULE_FEATURE_P10    (32 + 
>>> ilog2(PPC_FEATURE2_ARCH_3_1))
>>>   #define    PPC_ALIGN        16
>>>   #define GCM_IV_SIZE        12
>>>   diff --git a/arch/powerpc/include/asm/cpufeature.h 
>>> b/arch/powerpc/include/asm/cpufeature.h
>>> index f6f790a90367..2dcc66225e7f 100644
>>> --- a/arch/powerpc/include/asm/cpufeature.h
>>> +++ b/arch/powerpc/include/asm/cpufeature.h
>>> @@ -22,6 +22,7 @@
>>>    */
>>>     #define PPC_MODULE_FEATURE_VEC_CRYPTO            (32 + 
>>> ilog2(PPC_FEATURE2_VEC_CRYPTO))
>>> +#define PPC_MODULE_FEATURE_P10                (32 + 
>>> ilog2(PPC_FEATURE2_ARCH_3_1))
>>>     #define cpu_feature(x)        (x)
>>>   --
>>> 2.31.1