Re: [PATCH] Intel IOMMU Pass Through Support

[Fenghua Yu <fenghua.yu@intel.com>]

On Thu, Apr 30, 2009 at 05:00:34PM -0700, Andrew Morton wrote:
> On Thu, 30 Apr 2009 16:37:52 -0700
> Randy Dunlap <rdunlap@xenotime.net> wrote:
> 
> > > 
> > > arch/x86/built-in.o: In function `iommu_setup':
> > > arch/x86/kernel/pci-dma.c:215: undefined reference to `iommu_pass_through'
> > > arch/x86/built-in.o: In function `pci_swiotlb_init':
> > > arch/x86/kernel/pci-swiotlb.c:74: undefined reference to `iommu_pass_through'
> > > 
> > > Because iommu_pass_through is defined in drivers/pci/intel-iommu.c and
> > > 
> > > # CONFIG_DMAR is not set
> > > 
> > > I'll need to cook up some local hack to work around that.
> > > 
> > 
> > Patch just went to linux-next mailing list (but should have also
> > gone to lkml):
> > 
> > http://marc.info/?l=linux-next&m=124113213400748&w=2
> 
> OK, thanks.  My linux-next feed seems to have broken.
> 
> That patch modifies 
> 
> 	arch/ia64/kernel/dma-mapping.c
> 	arch/x86/kernel/pci-dma.c
> 
> but it would be more logical to modify
> 
> 	arch/ia64/kernel/pci-dma.c
> 	arch/x86/kernel/pci-dma.c
> 
> is there a reason why that cannot be done?

No particular reason. I just follow iommu_detected definition which is arch/ia64
kernel/dma-mapping.c.

Now I updated the compiling error fix patch as follows:

This updated patch should fix the compiling errors and remove the extern
iommu_pass_through from drivers/pci/intel-iommu.c file.

Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>

---

 arch/ia64/kernel/pci-dma.c |    2 ++
 arch/x86/kernel/pci-dma.c  |    4 ++--
 drivers/pci/intel-iommu.c  |    1 -
 3 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/arch/ia64/kernel/pci-dma.c b/arch/ia64/kernel/pci-dma.c
index eb98738..ecdde25 100644
--- a/arch/ia64/kernel/pci-dma.c
+++ b/arch/ia64/kernel/pci-dma.c
@@ -32,6 +32,8 @@ int force_iommu __read_mostly = 1;
 int force_iommu __read_mostly;
 #endif
 
+int iommu_pass_through;
+
 /* Dummy device used for NULL arguments (normally ISA). Better would
    be probably a smaller DMA mask, but this is bug-to-bug compatible
    to i386. */
diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c
index 8cad0d8..049005e 100644
--- a/arch/x86/kernel/pci-dma.c
+++ b/arch/x86/kernel/pci-dma.c
@@ -32,6 +32,8 @@ int no_iommu __read_mostly;
 /* Set this to 1 if there is a HW IOMMU in the system */
 int iommu_detected __read_mostly = 0;
 
+int iommu_pass_through;
+
 dma_addr_t bad_dma_address __read_mostly = 0;
 EXPORT_SYMBOL(bad_dma_address);
 
@@ -160,8 +162,6 @@ again:
 	return page_address(page);
 }
 
-extern int iommu_pass_through;
-
 /*
  * See <Documentation/x86_64/boot-options.txt> for the iommu kernel parameter
  * documentation.
diff --git a/drivers/pci/intel-iommu.c b/drivers/pci/intel-iommu.c
index d3ac5e5..66a9cba 100644
--- a/drivers/pci/intel-iommu.c
+++ b/drivers/pci/intel-iommu.c
@@ -288,7 +288,6 @@ int dmar_disabled = 1;
 static int __initdata dmar_map_gfx = 1;
 static int dmar_forcedac;
 static int intel_iommu_strict;
-int iommu_pass_through;
 
 #define DUMMY_DEVICE_DOMAIN_INFO ((struct device_domain_info *)(-1))
 static DEFINE_SPINLOCK(device_domain_lock);