[PATCH] eal: fix compilation without VFIO

[PATCH] eal: fix compilation without VFIO

[Shahaf Shuler]

a compilation error occurred when compiling with CONFIG_RTE_EAL_VFIO=n

== Build lib/librte_eal/linuxapp/eal
  CC eal_vfio.o
/download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1535:1: error: no
previous prototype for 'rte_vfio_dma_map' [-Werror=missing-prototypes]
 rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t
iova,
 ^
/download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1542:1: error: no
previous prototype for 'rte_vfio_dma_unmap' [-Werror=missing-prototypes]
 rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused
iova,
 ^

As there is no use for those dummy functions without VFIO removing them
completely.

Fixes: 73a639085938 ("vfio: allow to map other memory regions")
Cc: anatoly.burakov@intel.com

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
---
 lib/librte_eal/linuxapp/eal/eal_vfio.c | 16 ----------------
 1 file changed, 16 deletions(-)

diff --git a/lib/librte_eal/linuxapp/eal/eal_vfio.c b/lib/librte_eal/linuxapp/eal/eal_vfio.c
index 589d7d4787..4163bd4e08 100644
--- a/lib/librte_eal/linuxapp/eal/eal_vfio.c
+++ b/lib/librte_eal/linuxapp/eal/eal_vfio.c
@@ -1529,20 +1529,4 @@ rte_vfio_noiommu_is_enabled(void)
 	return c == 'Y';
 }
 
-#else
-
-int __rte_experimental
-rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t iova,
-		  __rte_unused uint64_t len)
-{
-	return -1;
-}
-
-int __rte_experimental
-rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused iova,
-		    __rte_unused uint64_t len)
-{
-	return -1;
-}
-
 #endif
-- 
2.12.0

Re: [PATCH] eal: fix compilation without VFIO

[Burakov, Anatoly]

On 12-Apr-18 2:34 PM, Shahaf Shuler wrote:
> a compilation error occurred when compiling with CONFIG_RTE_EAL_VFIO=n
> 
> == Build lib/librte_eal/linuxapp/eal
>    CC eal_vfio.o
> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1535:1: error: no
> previous prototype for 'rte_vfio_dma_map' [-Werror=missing-prototypes]
>   rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t
> iova,
>   ^
> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1542:1: error: no
> previous prototype for 'rte_vfio_dma_unmap' [-Werror=missing-prototypes]
>   rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused
> iova,
>   ^
> 
> As there is no use for those dummy functions without VFIO removing them
> completely.
> 
> Fixes: 73a639085938 ("vfio: allow to map other memory regions")
> Cc: anatoly.burakov@intel.com
> 
> Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
> ---
>   lib/librte_eal/linuxapp/eal/eal_vfio.c | 16 ----------------
>   1 file changed, 16 deletions(-)
> 
> diff --git a/lib/librte_eal/linuxapp/eal/eal_vfio.c b/lib/librte_eal/linuxapp/eal/eal_vfio.c
> index 589d7d4787..4163bd4e08 100644
> --- a/lib/librte_eal/linuxapp/eal/eal_vfio.c
> +++ b/lib/librte_eal/linuxapp/eal/eal_vfio.c
> @@ -1529,20 +1529,4 @@ rte_vfio_noiommu_is_enabled(void)
>   	return c == 'Y';
>   }
>   
> -#else
> -
> -int __rte_experimental
> -rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t iova,
> -		  __rte_unused uint64_t len)
> -{
> -	return -1;
> -}
> -
> -int __rte_experimental
> -rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused iova,
> -		    __rte_unused uint64_t len)
> -{
> -	return -1;
> -}
> -
>   #endif
> 

These functions are part of public API, like rest of functions in this 
header. They're in the map file. Should we perhaps go the BSD way and 
provide EAL with dummy prototypes as well? See bsdapp/eal/eal.c:763 onwards.

-- 
Thanks,
Anatoly

Re: [PATCH] eal: fix compilation without VFIO

[Thomas Monjalon]

12/04/2018 16:13, Burakov, Anatoly:
> On 12-Apr-18 2:34 PM, Shahaf Shuler wrote:
> > a compilation error occurred when compiling with CONFIG_RTE_EAL_VFIO=n
> > 
> > == Build lib/librte_eal/linuxapp/eal
> >    CC eal_vfio.o
> > /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1535:1: error: no
> > previous prototype for 'rte_vfio_dma_map' [-Werror=missing-prototypes]
> >   rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t
> > iova,
> >   ^
> > /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1542:1: error: no
> > previous prototype for 'rte_vfio_dma_unmap' [-Werror=missing-prototypes]
> >   rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused
> > iova,
> >   ^
> > 
> > As there is no use for those dummy functions without VFIO removing them
> > completely.
> 
> These functions are part of public API, like rest of functions in this 
> header. They're in the map file. Should we perhaps go the BSD way and 
> provide EAL with dummy prototypes as well? See bsdapp/eal/eal.c:763 onwards.

Why using dummy prototypes?
Because the prototypes in rte_vfio.h are under #ifdef VFIO_PRESENT ?
Is it possible to always define the prototypes in rte_vfio.h ?

Re: [PATCH] eal: fix compilation without VFIO

[Burakov, Anatoly]

On 13-Apr-18 12:39 AM, Thomas Monjalon wrote:
> 12/04/2018 16:13, Burakov, Anatoly:
>> On 12-Apr-18 2:34 PM, Shahaf Shuler wrote:
>>> a compilation error occurred when compiling with CONFIG_RTE_EAL_VFIO=n
>>>
>>> == Build lib/librte_eal/linuxapp/eal
>>>     CC eal_vfio.o
>>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1535:1: error: no
>>> previous prototype for 'rte_vfio_dma_map' [-Werror=missing-prototypes]
>>>    rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t
>>> iova,
>>>    ^
>>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1542:1: error: no
>>> previous prototype for 'rte_vfio_dma_unmap' [-Werror=missing-prototypes]
>>>    rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused
>>> iova,
>>>    ^
>>>
>>> As there is no use for those dummy functions without VFIO removing them
>>> completely.
>>
>> These functions are part of public API, like rest of functions in this
>> header. They're in the map file. Should we perhaps go the BSD way and
>> provide EAL with dummy prototypes as well? See bsdapp/eal/eal.c:763 onwards.
> 
> Why using dummy prototypes?
> Because the prototypes in rte_vfio.h are under #ifdef VFIO_PRESENT ?
> Is it possible to always define the prototypes in rte_vfio.h ?
> 

Well, technically, yes, we could. There is one function that uses a 
VFIO-specific struct definition:

int rte_vfio_setup_device(const char *sysfs_base, const char *dev_addr,
		int *vfio_dev_fd, struct vfio_device_info *device_info);

I'm sure we can work around that.

-- 
Thanks,
Anatoly

Re: [PATCH] eal: fix compilation without VFIO

[Thomas Monjalon]

13/04/2018 11:11, Burakov, Anatoly:
> On 13-Apr-18 12:39 AM, Thomas Monjalon wrote:
> > 12/04/2018 16:13, Burakov, Anatoly:
> >> On 12-Apr-18 2:34 PM, Shahaf Shuler wrote:
> >>> a compilation error occurred when compiling with CONFIG_RTE_EAL_VFIO=n
> >>>
> >>> == Build lib/librte_eal/linuxapp/eal
> >>>     CC eal_vfio.o
> >>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1535:1: error: no
> >>> previous prototype for 'rte_vfio_dma_map' [-Werror=missing-prototypes]
> >>>    rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t
> >>> iova,
> >>>    ^
> >>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1542:1: error: no
> >>> previous prototype for 'rte_vfio_dma_unmap' [-Werror=missing-prototypes]
> >>>    rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused
> >>> iova,
> >>>    ^
> >>>
> >>> As there is no use for those dummy functions without VFIO removing them
> >>> completely.
> >>
> >> These functions are part of public API, like rest of functions in this
> >> header. They're in the map file. Should we perhaps go the BSD way and
> >> provide EAL with dummy prototypes as well? See bsdapp/eal/eal.c:763 onwards.
> > 
> > Why using dummy prototypes?
> > Because the prototypes in rte_vfio.h are under #ifdef VFIO_PRESENT ?
> > Is it possible to always define the prototypes in rte_vfio.h ?
> > 
> 
> Well, technically, yes, we could. There is one function that uses a 
> VFIO-specific struct definition:
> 
> int rte_vfio_setup_device(const char *sysfs_base, const char *dev_addr,
> 		int *vfio_dev_fd, struct vfio_device_info *device_info);
> 
> I'm sure we can work around that.


Removing dummy prototypes need, would be a nicer fix.

Re: [PATCH] eal: fix compilation without VFIO

[Burakov, Anatoly]

On 13-Apr-18 2:08 PM, Thomas Monjalon wrote:
> 13/04/2018 11:11, Burakov, Anatoly:
>> On 13-Apr-18 12:39 AM, Thomas Monjalon wrote:
>>> 12/04/2018 16:13, Burakov, Anatoly:
>>>> On 12-Apr-18 2:34 PM, Shahaf Shuler wrote:
>>>>> a compilation error occurred when compiling with CONFIG_RTE_EAL_VFIO=n
>>>>>
>>>>> == Build lib/librte_eal/linuxapp/eal
>>>>>      CC eal_vfio.o
>>>>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1535:1: error: no
>>>>> previous prototype for 'rte_vfio_dma_map' [-Werror=missing-prototypes]
>>>>>     rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t
>>>>> iova,
>>>>>     ^
>>>>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1542:1: error: no
>>>>> previous prototype for 'rte_vfio_dma_unmap' [-Werror=missing-prototypes]
>>>>>     rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused
>>>>> iova,
>>>>>     ^
>>>>>
>>>>> As there is no use for those dummy functions without VFIO removing them
>>>>> completely.
>>>>
>>>> These functions are part of public API, like rest of functions in this
>>>> header. They're in the map file. Should we perhaps go the BSD way and
>>>> provide EAL with dummy prototypes as well? See bsdapp/eal/eal.c:763 onwards.
>>>
>>> Why using dummy prototypes?
>>> Because the prototypes in rte_vfio.h are under #ifdef VFIO_PRESENT ?
>>> Is it possible to always define the prototypes in rte_vfio.h ?
>>>
>>
>> Well, technically, yes, we could. There is one function that uses a
>> VFIO-specific struct definition:
>>
>> int rte_vfio_setup_device(const char *sysfs_base, const char *dev_addr,
>> 		int *vfio_dev_fd, struct vfio_device_info *device_info);
>>
>> I'm sure we can work around that.
> 
> 
> Removing dummy prototypes need, would be a nicer fix.
> 

OK. Shahaf, will you submit a v2 with this, or should i do it? I think 
it should be just a matter of #ifndef VFIO_PRESENT //define 
vfio_device_info struct #endif - this should take care of the problem of 
hiding the function definitions.

FreeBSD will also need to be adjusted to remove dummy prototypes.

-- 
Thanks,
Anatoly

Re: [PATCH] eal: fix compilation without VFIO

[Thomas Monjalon]

13/04/2018 15:37, Burakov, Anatoly:
> On 13-Apr-18 2:08 PM, Thomas Monjalon wrote:
> > 13/04/2018 11:11, Burakov, Anatoly:
> >> On 13-Apr-18 12:39 AM, Thomas Monjalon wrote:
> >>> 12/04/2018 16:13, Burakov, Anatoly:
> >>>> On 12-Apr-18 2:34 PM, Shahaf Shuler wrote:
> >>>>> a compilation error occurred when compiling with CONFIG_RTE_EAL_VFIO=n
> >>>>>
> >>>>> == Build lib/librte_eal/linuxapp/eal
> >>>>>      CC eal_vfio.o
> >>>>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1535:1: error: no
> >>>>> previous prototype for 'rte_vfio_dma_map' [-Werror=missing-prototypes]
> >>>>>     rte_vfio_dma_map(uint64_t __rte_unused vaddr, __rte_unused uint64_t
> >>>>> iova,
> >>>>>     ^
> >>>>> /download/dpdk/lib/librte_eal/linuxapp/eal/eal_vfio.c:1542:1: error: no
> >>>>> previous prototype for 'rte_vfio_dma_unmap' [-Werror=missing-prototypes]
> >>>>>     rte_vfio_dma_unmap(uint64_t __rte_unused vaddr, uint64_t __rte_unused
> >>>>> iova,
> >>>>>     ^
> >>>>>
> >>>>> As there is no use for those dummy functions without VFIO removing them
> >>>>> completely.
> >>>>
> >>>> These functions are part of public API, like rest of functions in this
> >>>> header. They're in the map file. Should we perhaps go the BSD way and
> >>>> provide EAL with dummy prototypes as well? See bsdapp/eal/eal.c:763 onwards.
> >>>
> >>> Why using dummy prototypes?
> >>> Because the prototypes in rte_vfio.h are under #ifdef VFIO_PRESENT ?
> >>> Is it possible to always define the prototypes in rte_vfio.h ?
> >>>
> >>
> >> Well, technically, yes, we could. There is one function that uses a
> >> VFIO-specific struct definition:
> >>
> >> int rte_vfio_setup_device(const char *sysfs_base, const char *dev_addr,
> >> 		int *vfio_dev_fd, struct vfio_device_info *device_info);
> >>
> >> I'm sure we can work around that.
> > 
> > 
> > Removing dummy prototypes need, would be a nicer fix.
> > 
> 
> OK. Shahaf, will you submit a v2 with this, or should i do it? I think 
> it should be just a matter of #ifndef VFIO_PRESENT //define 
> vfio_device_info struct #endif - this should take care of the problem of 
> hiding the function definitions.
> 
> FreeBSD will also need to be adjusted to remove dummy prototypes.

I think you are more familiar with VFIO than any of us.
It is better to let you do, think about the implications and do the tests.
Thanks :)

Re: [PATCH] eal: fix compilation without VFIO

[Shahaf Shuler]

Friday, April 13, 2018 4:59 PM, Thomas Monjalon:
> >
> > OK. Shahaf, will you submit a v2 with this, or should i do it? I think
> > it should be just a matter of #ifndef VFIO_PRESENT //define
> > vfio_device_info struct #endif - this should take care of the problem
> > of hiding the function definitions.
> >
> > FreeBSD will also need to be adjusted to remove dummy prototypes.
> 
> I think you are more familiar with VFIO than any of us.
> It is better to let you do, think about the implications and do the tests.
> Thanks :)
> 

I don't mind whom of us will do it, as long as it will be done quickly. Currently there are some tests which cannot run on our regression due to it. 

Re: [PATCH] eal: fix compilation without VFIO

[Burakov, Anatoly]

On 16-Apr-18 6:50 AM, Shahaf Shuler wrote:
> Friday, April 13, 2018 4:59 PM, Thomas Monjalon:
>>>
>>> OK. Shahaf, will you submit a v2 with this, or should i do it? I think
>>> it should be just a matter of #ifndef VFIO_PRESENT //define
>>> vfio_device_info struct #endif - this should take care of the problem
>>> of hiding the function definitions.
>>>
>>> FreeBSD will also need to be adjusted to remove dummy prototypes.
>>
>> I think you are more familiar with VFIO than any of us.
>> It is better to let you do, think about the implications and do the tests.
>> Thanks :)
>>
> 
> I don't mind whom of us will do it, as long as it will be done quickly. Currently there are some tests which cannot run on our regression due to it.
> 
> 

Yes, i'll be submitting the patch shortly.

-- 
Thanks,
Anatoly