[PATCH] DMA 4GB boundary protection

[Jake Moilanen]

There are many adapters which can not handle DMAing acrosss any 4 GB
boundary.  For instance the latest Emulex adapters.  

This normally is not an issue as firmware gives us dma-windows under
4gigs.  However, some of the new System-P boxes have dma-windows above
4gigs, and this present a problem.

I propose fixing it in the IOMMU allocation instead of making each
driver protect against it as it is more efficient, and won't require
changing every driver which has not considered this issue.

This patch checks to see if the mapping spans a 4 gig boundary, and if
it does, retries the allocation.  It tries the next allocation at the
start of the crossed 4 gig boundary.

Signed-off-by: Jake Moilanen <moilanen at austin.ibm.com>
---

 arch/powerpc/kernel/iommu.c |   10 ++++++++++
 1 files changed, 10 insertions(+)

Index: powerpc/arch/powerpc/kernel/iommu.c
===================================================================
--- powerpc.orig/arch/powerpc/kernel/iommu.c
+++ powerpc/arch/powerpc/kernel/iommu.c
@@ -76,6 +76,7 @@ static unsigned long iommu_range_alloc(s
                                        unsigned int align_order)
 { 
 	unsigned long n, end, i, start;
+	unsigned long start_addr, end_addr;
 	unsigned long limit;
 	int largealloc = npages > 15;
 	int pass = 0;
@@ -146,6 +147,15 @@ static unsigned long iommu_range_alloc(s
 		}
 	}
 
+	/* DMA cannot cross 4 GB boundary */
+	start_addr = (n + tbl->it_offset) << PAGE_SHIFT;
+	end_addr = (end + tbl->it_offset) << PAGE_SHIFT;
+	if ((start_addr >> 32) != (end_addr >> 32)) {
+		end_addr &= 0xffffffff00000000l;
+		start = (end_addr >> PAGE_SHIFT) - tbl->it_offset;
+		goto again;
+	}
+
 	for (i = n; i < end; i++)
 		if (test_bit(i, tbl->it_map)) {
 			start = i+1;