[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <5E55CD0A.4060209@hisilicon.com>
Date: Wed, 26 Feb 2020 09:42:34 +0800
From: Zhou Wang <wangzhou1@...ilicon.com>
To: Herbert Xu <herbert@...dor.apana.org.au>
CC: Hongbo Yao <yaohongbo@...wei.com>, <davem@...emloft.net>,
<linux-kernel@...r.kernel.org>, <linux-crypto@...r.kernel.org>,
<chenzhou10@...wei.com>, <xuzaibo@...wei.com>
Subject: Re: [PATCH -next] crypto: hisilicon - qm depends on UACCE
On 2020/2/25 18:22, Herbert Xu wrote:
> On Tue, Feb 25, 2020 at 04:44:57PM +0800, Zhou Wang wrote:
>>
>>> diff --git a/drivers/crypto/hisilicon/Kconfig b/drivers/crypto/hisilicon/Kconfig
>>> index 8851161f722f..b35c2ec15bc2 100644
>>> --- a/drivers/crypto/hisilicon/Kconfig
>>> +++ b/drivers/crypto/hisilicon/Kconfig
>>> @@ -40,6 +40,7 @@ config CRYPTO_DEV_HISI_QM
>>> tristate
>>> depends on ARM64 || COMPILE_TEST
>>> depends on PCI && PCI_MSI
>>> + depends on UACCE
>>> help
>>> HiSilicon accelerator engines use a common queue management
>>> interface. Specific engine driver may use this module.
>>>
>>
>> Indeed, this driver does not depend on uacce fully, as if there is no uacce, it still can
>> register to kernel crypto.
>>
>> Seems that changing uacce config to bool can avoid this problem.
>
> You shouldn't make it a bool. The standard way to solve this is to
> add this:
>
> depends on UACCE || UACCE=n
Thanks! Let's fix together with zip Kconfig.
Best,
Zhou
>
> Cheers,
>
Powered by blists - more mailing lists