Message ID | 20090122194000.GA14767@austin.ibm.com (mailing list archive) |
---|---|
State | Accepted, archived |
Delegated to: | Benjamin Herrenschmidt |
Headers | show |
Acked by: Brian King <brking@linux.vnet.ibm.com> Robert Jennings wrote: > In the VIO bus code the wrappers for dma alloc_coherent and free_coherent > calls are rounding to IOMMU_PAGE_SIZE. Taking a look at the underlying > calls, the actual mapping is promoted to PAGE_SIZE. Changing the > rounding in these two functions fixes under-reporting the entitlement > used by the system. Without this change, the system could run out of > entitlement before it believes it has and incur mapping failures at the > firmware level. > > Also in the VIO bus code, the wrapper for dma map_sg is not exiting in > an error path where it should. Rather than fall through to code for the > success case, this patch adds the return that is needed in the error path. > > Signed-off-by: Robert Jennings <rcj@linux.vnet.ibm.com> > > --- > arch/powerpc/kernel/vio.c | 7 ++++--- > 1 file changed, 4 insertions(+), 3 deletions(-) > > Index: b/arch/powerpc/kernel/vio.c > =================================================================== > --- a/arch/powerpc/kernel/vio.c > +++ b/arch/powerpc/kernel/vio.c > @@ -492,14 +492,14 @@ static void *vio_dma_iommu_alloc_coheren > struct vio_dev *viodev = to_vio_dev(dev); > void *ret; > > - if (vio_cmo_alloc(viodev, roundup(size, IOMMU_PAGE_SIZE))) { > + if (vio_cmo_alloc(viodev, roundup(size, PAGE_SIZE))) { > atomic_inc(&viodev->cmo.allocs_failed); > return NULL; > } > > ret = dma_iommu_ops.alloc_coherent(dev, size, dma_handle, flag); > if (unlikely(ret == NULL)) { > - vio_cmo_dealloc(viodev, roundup(size, IOMMU_PAGE_SIZE)); > + vio_cmo_dealloc(viodev, roundup(size, PAGE_SIZE)); > atomic_inc(&viodev->cmo.allocs_failed); > } > > @@ -513,7 +513,7 @@ static void vio_dma_iommu_free_coherent( > > dma_iommu_ops.free_coherent(dev, size, vaddr, dma_handle); > > - vio_cmo_dealloc(viodev, roundup(size, IOMMU_PAGE_SIZE)); > + vio_cmo_dealloc(viodev, roundup(size, PAGE_SIZE)); > } > > static dma_addr_t vio_dma_iommu_map_page(struct device *dev, struct page *page, > @@ -572,6 +572,7 @@ static int vio_dma_iommu_map_sg(struct d > if (unlikely(!ret)) { > vio_cmo_dealloc(viodev, alloc_size); > atomic_inc(&viodev->cmo.allocs_failed); > + return ret; > } > > for (sgl = sglist, count = 0; count < ret; count++, sgl++) > _______________________________________________ > Linuxppc-dev mailing list > Linuxppc-dev@ozlabs.org > https://ozlabs.org/mailman/listinfo/linuxppc-dev
Index: b/arch/powerpc/kernel/vio.c =================================================================== --- a/arch/powerpc/kernel/vio.c +++ b/arch/powerpc/kernel/vio.c @@ -492,14 +492,14 @@ static void *vio_dma_iommu_alloc_coheren struct vio_dev *viodev = to_vio_dev(dev); void *ret; - if (vio_cmo_alloc(viodev, roundup(size, IOMMU_PAGE_SIZE))) { + if (vio_cmo_alloc(viodev, roundup(size, PAGE_SIZE))) { atomic_inc(&viodev->cmo.allocs_failed); return NULL; } ret = dma_iommu_ops.alloc_coherent(dev, size, dma_handle, flag); if (unlikely(ret == NULL)) { - vio_cmo_dealloc(viodev, roundup(size, IOMMU_PAGE_SIZE)); + vio_cmo_dealloc(viodev, roundup(size, PAGE_SIZE)); atomic_inc(&viodev->cmo.allocs_failed); } @@ -513,7 +513,7 @@ static void vio_dma_iommu_free_coherent( dma_iommu_ops.free_coherent(dev, size, vaddr, dma_handle); - vio_cmo_dealloc(viodev, roundup(size, IOMMU_PAGE_SIZE)); + vio_cmo_dealloc(viodev, roundup(size, PAGE_SIZE)); } static dma_addr_t vio_dma_iommu_map_page(struct device *dev, struct page *page, @@ -572,6 +572,7 @@ static int vio_dma_iommu_map_sg(struct d if (unlikely(!ret)) { vio_cmo_dealloc(viodev, alloc_size); atomic_inc(&viodev->cmo.allocs_failed); + return ret; } for (sgl = sglist, count = 0; count < ret; count++, sgl++)
In the VIO bus code the wrappers for dma alloc_coherent and free_coherent calls are rounding to IOMMU_PAGE_SIZE. Taking a look at the underlying calls, the actual mapping is promoted to PAGE_SIZE. Changing the rounding in these two functions fixes under-reporting the entitlement used by the system. Without this change, the system could run out of entitlement before it believes it has and incur mapping failures at the firmware level. Also in the VIO bus code, the wrapper for dma map_sg is not exiting in an error path where it should. Rather than fall through to code for the success case, this patch adds the return that is needed in the error path. Signed-off-by: Robert Jennings <rcj@linux.vnet.ibm.com> --- arch/powerpc/kernel/vio.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)