Re: [PATCH for-next 2/6] xen/common: Guard iommu symbols with CONFIG_HAS_PASSTHROUGH

From: Jan Beulich <jbeulich@suse.com>
To: Connor Davis <connojdavis@gmail.com>
Cc: Bobby Eshleman <bobbyeshleman@gmail.com>,
	Andrew Cooper <andrew.cooper3@citrix.com>,
	George Dunlap <george.dunlap@citrix.com>,
	Ian Jackson <iwj@xenproject.org>, Julien Grall <julien@xen.org>,
	Stefano Stabellini <sstabellini@kernel.org>, Wei Liu <wl@xen.org>,
	xen-devel@lists.xenproject.org
Subject: Re: [PATCH for-next 2/6] xen/common: Guard iommu symbols with CONFIG_HAS_PASSTHROUGH
Date: Fri, 26 Feb 2021 08:41:25 +0100	[thread overview]
Message-ID: <34bb915a-be60-1cc7-c802-e0491b6741df@suse.com> (raw)
In-Reply-To: <20210226025402.ryuxpicaqujmfxbu@thewall>

On 26.02.2021 03:54, Connor Davis wrote:
> On Thu, Feb 25, 2021 at 04:45:15PM +0100, Jan Beulich wrote:
>> On 25.02.2021 16:24, Connor Davis wrote:
>>> --- a/xen/common/domain.c
>>> +++ b/xen/common/domain.c
>>> @@ -501,7 +501,9 @@ static int sanitise_domain_config(struct xen_domctl_createdomain *config)
>>>              return -EINVAL;
>>>          }
>>>  
>>> +#ifdef CONFIG_HAS_PASSTHROUGH
>>>          if ( !iommu_enabled )
>>> +#endif
>>>          {
>>>              dprintk(XENLOG_INFO, "IOMMU requested but not available\n");
>>>              return -EINVAL;
>>
>> Where possible - to avoid such #ifdef-ary - the symbol instead should
>> get #define-d to a sensible value ("false" in this case) in the header.
>> The other cases here may indeed need to remain as you have them.
>>
> Do you prefer the #define in the same function near the if or
> somwhere near the top of the file?

Neither, if I understand you correctly. It should be in the same header
where the extern declaration lives, for the whole code base to consume.

Jan