[dpdk-dev] [PATCH] bus/pci: fix vfio mode

[Gaëtan Rivet]

On Mon, Oct 30, 2017 at 09:57:08AM -0700, Ferruh Yigit wrote:
> On 10/30/2017 4:24 AM, Thomas Monjalon wrote:
> > 30/10/2017 10:27, Thomas Monjalon:
> >> 30/10/2017 10:17, Gaëtan Rivet:
> >>> Hi Ferruh,
> >>>
> >>> On Mon, Oct 30, 2017 at 02:00:43AM -0700, Ferruh Yigit wrote:
> >>>> On 10/30/2017 1:06 AM, Gaëtan Rivet wrote:
> >>>>> Hi Jerin,
> >>>>>
> >>>>> On Sat, Oct 28, 2017 at 11:50:52AM +0530, Jerin Jacob wrote:
> >>>>>> The definition of VFIO_PRESENT is "eal_vfio.h", Fail to
> >>>>>> include eal_vfio.h will result in disabling vfio.
> >>>>>>
> >>>>>> Fixes: 279b581c897d ("vfio: expose functions")
> >>>>>>
> >>>>>
> >>>>> Thanks for the fix, sorry for VFIO.
> >>>>> I tried to let go of VFIO_PRESENT in the PCI patchset, unfortunately I did
> >>>>> not do a good-enough job.
> >>>>>
> >>>>> Instead of reinstating the dependency on the private eal_vfio.h header,
> >>>>> I'd suggest replacing all VFIO_PRESENT references within the PCI bus by
> >>>>> RTE_EAL_VFIO, and make the pci_vfio.c compilation depend on it within
> >>>>> the linux Makefile. Something like:
> >>>>>
> >>>>> ---8<---
> >>>>>
> >>>>> grep -rl VFIO_PRESENT drivers/bus/pci/linux/ |while read -r file
> >>>>> do sed -i 's;VFIO_PRESENT;RTE_EAL_VFIO;' $file
> >>>>> done
> >>>>
> >>>> VFIO_PRESENT is the combination of the if user enabled VFIO and if Linux kernel
> >>>> supports it.
> >>>>
> >>>> Why not add same check and VFIO_PRESENT definition to rte_vfio.h:
> >>>>
> >>>>   --- a/lib/librte_eal/common/include/rte_vfio.h
> >>>>   +++ b/lib/librte_eal/common/include/rte_vfio.h
> >>>>   @@ -34,7 +34,13 @@
> >>>>    #ifndef _RTE_VFIO_H_
> >>>>    #define _RTE_VFIO_H_
> >>>>
> >>>>   +#if !defined(VFIO_PRESENT) && defined(RTE_EAL_VFIO)
> >>>>   +#include <linux/version.h>
> >>>>   +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 6, 0)
> >>>>    #include <linux/vfio.h>
> >>>>   +#define VFIO_PRESENT
> >>>>   +#endif /* >= 3.6.0 */
> >>>>   +#endif
> >>>>   +
> >>>>   +#ifdef VFIO_PRESENT
> >>>>
> >>>>   ... Need to wrap here in case VFIO disabled ...
> >>>>
> >>>>   #endif
> >>>>
> >>>
> >>> It would work indeed.
> >>>
> >>> I mostly dislike having whole compilation units disabled silently on a
> >>> linux version check. I think that if someone wanted to support kernels <
> >>> 3.6, they ought to do the work of disabling RTE_EAL_VFIO.
> >>>
> >>> A build error could be thrown to help those users toward the right
> >>> solution, but I think that the meaning of having this option enabled
> >>> should be enforced: if it is enabled, it is compiled. If dependencies
> >>> are not met, then the option should be disabled.
> >>
> >> +1 to avoid implicit disabling.
> > 
> > To make it clear, we can disable VFIO automatically if not supported by
> > the kernel at compilation time, but there should be a warning.
> 
> Gaetan suggest letting build error and push user to disable the feature. If we
> let the build error, I think feature should be disabled by default.
> 
> And for VFIO, instead of disabling it by default I am for auto disable it if
> kernel requirement not met, and +1 for adding a warning.
> 
> btw, that kernel version check is already in eal_vfio.h, we are already
> disabling some vfio code based on kernel version check, weather add here or not.
> 
> Thanks,
> ferruh

The new version of the patches should behave like you described :) .

I still think VFIO_PRESENT should be changed, but not now.

-- 
Gaëtan Rivet
6WIND