[PATCH] pci: fix compilation when CONFIG_PCI_MSI=n

[PATCH] pci: fix compilation when CONFIG_PCI_MSI=n

[Rusty Russell]

drivers/pci/pci.c:2277: error: redefinition of ‘pci_msi_off’
include/linux/pci.h:786: note: previous definition of ‘pci_msi_off’ was here

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
---
 drivers/pci/pci.c |    2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -2266,6 +2266,7 @@ pci_intx(struct pci_dev *pdev, int enabl
 	}
 }
 
+#ifdef CONFIG_PCI_MSI
 /**
  * pci_msi_off - disables any msi or msix capabilities
  * @dev: the PCI device to operate on
@@ -2293,6 +2294,7 @@ void pci_msi_off(struct pci_dev *dev)
 	}
 }
 EXPORT_SYMBOL_GPL(pci_msi_off);
+#endif
 
 #ifndef HAVE_ARCH_PCI_SET_DMA_MAX_SEGMENT_SIZE
 int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size)

Re: [PATCH] pci: fix compilation when CONFIG_PCI_MSI=n

[Hidetoshi Seto]

(2010/06/16 13:51), Rusty Russell wrote:
> drivers/pci/pci.c:2277: error: redefinition of ‘pci_msi_off’
> include/linux/pci.h:786: note: previous definition of ‘pci_msi_off’ was here
> 
> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> ---
>  drivers/pci/pci.c |    2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -2266,6 +2266,7 @@ pci_intx(struct pci_dev *pdev, int enabl
>  	}
>  }
>  
> +#ifdef CONFIG_PCI_MSI
>  /**
>   * pci_msi_off - disables any msi or msix capabilities
>   * @dev: the PCI device to operate on
> @@ -2293,6 +2294,7 @@ void pci_msi_off(struct pci_dev *dev)
>  	}
>  }
>  EXPORT_SYMBOL_GPL(pci_msi_off);
> +#endif
>  
>  #ifndef HAVE_ARCH_PCI_SET_DMA_MAX_SEGMENT_SIZE
>  int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size)

Rusty, thank you for reporting it.
But I think this fix is not correct.

It looks like a bug in Michael's patch:
  [PATCH for-2.6.35] virtio-pci: disable msi at startup

The patch includes a change as following:

> > --- a/include/linux/pci.h
> > +++ b/include/linux/pci.h
> > @@ -780,7 +780,11 @@ int __must_check pci_set_mwi(struct pci_dev *dev);
> >  int pci_try_set_mwi(struct pci_dev *dev);
> >  void pci_clear_mwi(struct pci_dev *dev);
> >  void pci_intx(struct pci_dev *dev, int enable);
> > +#ifdef CONFIG_PCI_MSI
> >  void pci_msi_off(struct pci_dev *dev);
> > +#else
> > +static inline void pci_msi_off(struct pci_dev *dev) {}
> > +#endif
> >  int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size);
> >  int pci_set_dma_seg_boundary(struct pci_dev *dev, unsigned long mask);
> >  int pcix_get_max_mmrbc(struct pci_dev *dev);

According to the old commit f5f2b13129a6541debf8851bae843cbbf48298b7,
pci_msi_off() was designed to be available even when msi support is not
compiled in.  So I think this change will make something wrong.

Michael, could you check your patch?


Thanks,
H.Seto

Re: [PATCH] pci: fix compilation when CONFIG_PCI_MSI=n

[Michael S. Tsirkin]

On Wed, Jun 16, 2010 at 03:09:11PM +0900, Hidetoshi Seto wrote:
> (2010/06/16 13:51), Rusty Russell wrote:
> > drivers/pci/pci.c:2277: error: redefinition of ‘pci_msi_off’
> > include/linux/pci.h:786: note: previous definition of ‘pci_msi_off’ was here
> > 
> > Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> > ---
> >  drivers/pci/pci.c |    2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> > --- a/drivers/pci/pci.c
> > +++ b/drivers/pci/pci.c
> > @@ -2266,6 +2266,7 @@ pci_intx(struct pci_dev *pdev, int enabl
> >  	}
> >  }
> >  
> > +#ifdef CONFIG_PCI_MSI
> >  /**
> >   * pci_msi_off - disables any msi or msix capabilities
> >   * @dev: the PCI device to operate on
> > @@ -2293,6 +2294,7 @@ void pci_msi_off(struct pci_dev *dev)
> >  	}
> >  }
> >  EXPORT_SYMBOL_GPL(pci_msi_off);
> > +#endif
> >  
> >  #ifndef HAVE_ARCH_PCI_SET_DMA_MAX_SEGMENT_SIZE
> >  int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size)
> 
> Rusty, thank you for reporting it.
> But I think this fix is not correct.
> 
> It looks like a bug in Michael's patch:
>   [PATCH for-2.6.35] virtio-pci: disable msi at startup
> 
> The patch includes a change as following:
> 
> > > --- a/include/linux/pci.h
> > > +++ b/include/linux/pci.h
> > > @@ -780,7 +780,11 @@ int __must_check pci_set_mwi(struct pci_dev *dev);
> > >  int pci_try_set_mwi(struct pci_dev *dev);
> > >  void pci_clear_mwi(struct pci_dev *dev);
> > >  void pci_intx(struct pci_dev *dev, int enable);
> > > +#ifdef CONFIG_PCI_MSI
> > >  void pci_msi_off(struct pci_dev *dev);
> > > +#else
> > > +static inline void pci_msi_off(struct pci_dev *dev) {}
> > > +#endif
> > >  int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size);
> > >  int pci_set_dma_seg_boundary(struct pci_dev *dev, unsigned long mask);
> > >  int pcix_get_max_mmrbc(struct pci_dev *dev);
> 
> According to the old commit f5f2b13129a6541debf8851bae843cbbf48298b7,
> pci_msi_off() was designed to be available even when msi support is not
> compiled in.  So I think this change will make something wrong.
> 
> Michael, could you check your patch?
> 
> 
> Thanks,
> H.Seto

You are right. I didn't think this through: we must disable
the capability at startup even if the current kernel does not
support MSI, otherwise the device won't work.

Please apply this on top.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

---

diff --git a/include/linux/pci.h b/include/linux/pci.h
index 31d8a12..7cb0084 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -780,11 +780,7 @@ int __must_check pci_set_mwi(struct pci_dev *dev);
 int pci_try_set_mwi(struct pci_dev *dev);
 void pci_clear_mwi(struct pci_dev *dev);
 void pci_intx(struct pci_dev *dev, int enable);
-#ifdef CONFIG_PCI_MSI
 void pci_msi_off(struct pci_dev *dev);
-#else
-static inline void pci_msi_off(struct pci_dev *dev) {}
-#endif
 int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size);
 int pci_set_dma_seg_boundary(struct pci_dev *dev, unsigned long mask);
 int pcix_get_max_mmrbc(struct pci_dev *dev);

Re: [PATCH] pci: fix compilation when CONFIG_PCI_MSI=n

[Jesse Barnes]

On Wed, 16 Jun 2010 11:22:40 +0300
"Michael S. Tsirkin" <mst@redhat.com> wrote:

> On Wed, Jun 16, 2010 at 03:09:11PM +0900, Hidetoshi Seto wrote:
> > (2010/06/16 13:51), Rusty Russell wrote:
> > > drivers/pci/pci.c:2277: error: redefinition of ‘pci_msi_off’
> > > include/linux/pci.h:786: note: previous definition of ‘pci_msi_off’ was here
> > > 
> > > Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
> > > ---
> > >  drivers/pci/pci.c |    2 ++
> > >  1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> > > --- a/drivers/pci/pci.c
> > > +++ b/drivers/pci/pci.c
> > > @@ -2266,6 +2266,7 @@ pci_intx(struct pci_dev *pdev, int enabl
> > >  	}
> > >  }
> > >  
> > > +#ifdef CONFIG_PCI_MSI
> > >  /**
> > >   * pci_msi_off - disables any msi or msix capabilities
> > >   * @dev: the PCI device to operate on
> > > @@ -2293,6 +2294,7 @@ void pci_msi_off(struct pci_dev *dev)
> > >  	}
> > >  }
> > >  EXPORT_SYMBOL_GPL(pci_msi_off);
> > > +#endif
> > >  
> > >  #ifndef HAVE_ARCH_PCI_SET_DMA_MAX_SEGMENT_SIZE
> > >  int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size)
> > 
> > Rusty, thank you for reporting it.
> > But I think this fix is not correct.
> > 
> > It looks like a bug in Michael's patch:
> >   [PATCH for-2.6.35] virtio-pci: disable msi at startup
> > 
> > The patch includes a change as following:
> > 
> > > > --- a/include/linux/pci.h
> > > > +++ b/include/linux/pci.h
> > > > @@ -780,7 +780,11 @@ int __must_check pci_set_mwi(struct pci_dev *dev);
> > > >  int pci_try_set_mwi(struct pci_dev *dev);
> > > >  void pci_clear_mwi(struct pci_dev *dev);
> > > >  void pci_intx(struct pci_dev *dev, int enable);
> > > > +#ifdef CONFIG_PCI_MSI
> > > >  void pci_msi_off(struct pci_dev *dev);
> > > > +#else
> > > > +static inline void pci_msi_off(struct pci_dev *dev) {}
> > > > +#endif
> > > >  int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size);
> > > >  int pci_set_dma_seg_boundary(struct pci_dev *dev, unsigned long mask);
> > > >  int pcix_get_max_mmrbc(struct pci_dev *dev);
> > 
> > According to the old commit f5f2b13129a6541debf8851bae843cbbf48298b7,
> > pci_msi_off() was designed to be available even when msi support is not
> > compiled in.  So I think this change will make something wrong.
> > 
> > Michael, could you check your patch?
> > 
> > 
> > Thanks,
> > H.Seto
> 
> You are right. I didn't think this through: we must disable
> the capability at startup even if the current kernel does not
> support MSI, otherwise the device won't work.
> 
> Please apply this on top.
> 
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> 
> ---
> 
> diff --git a/include/linux/pci.h b/include/linux/pci.h
> index 31d8a12..7cb0084 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -780,11 +780,7 @@ int __must_check pci_set_mwi(struct pci_dev *dev);
>  int pci_try_set_mwi(struct pci_dev *dev);
>  void pci_clear_mwi(struct pci_dev *dev);
>  void pci_intx(struct pci_dev *dev, int enable);
> -#ifdef CONFIG_PCI_MSI
>  void pci_msi_off(struct pci_dev *dev);
> -#else
> -static inline void pci_msi_off(struct pci_dev *dev) {}
> -#endif
>  int pci_set_dma_max_seg_size(struct pci_dev *dev, unsigned int size);
>  int pci_set_dma_seg_boundary(struct pci_dev *dev, unsigned long mask);
>  int pcix_get_max_mmrbc(struct pci_dev *dev);

I don't have the #ifdef in pci.h or the export in my tree.  Can you
guys fix this up in the virt tree and add my acked-by?

Thanks,
-- 
Jesse Barnes, Intel Open Source Technology Center