[PATCH v3] sparc: Fix page alignment in dma mapping

From: stian

Date: Wed Feb 18 2026 - 07:03:12 EST


From: Stian Halseth <stian@xxxxxx>

'phys' may include an offset within the page, while 'base_paddr' is already
page-aligned. This caused incorrect DMA mapping in dma_4u_map_phys and
dma_4v_map_phys.

Fix both functions by masking phys with IO_PAGE_MASK or subtracting the
page offset, covering both generic SPARC code and sun4v.

Fixes: 38c0d0ebf520 ("sparc: Use physical address DMA mapping")
Reported-by: Stian Halseth <stian@xxxxxx>
Closes: https://github.com/sparclinux/issues/issues/75
Suggested-by: Marek Szyprowski <m.szyprowski@xxxxxxxxxxx>
Signed-off-by: Stian Halseth <stian@xxxxxx>
---
Changelog:
v3:
- Use IO_PAGE_MASK for alignment in both dma_4u_map_phys and dma_4v_map_phys
- Align phys before the loop in pci_sun4v.c
- Drop masking inside iommu_batch_add()

v2: https://lore.kernel.org/all/20260217171104.5643-3-stian@xxxxxx/
- Add Fixes tag
- Add Closes tag
- Trim commit message to 80-column lines
- Remove newlines
- Subject: Remove the "iommu/pci_sun4v:" (generic sparc patch, not specific to sun4v)

v1: https://lore.kernel.org/all/20260217150350.3541-2-stian@xxxxxx/
---
arch/sparc/kernel/iommu.c | 2 ++
arch/sparc/kernel/pci_sun4v.c | 2 ++
2 files changed, 4 insertions(+)

diff --git a/arch/sparc/kernel/iommu.c b/arch/sparc/kernel/iommu.c
index 46ef88bc9c26..7613ab0ffb89 100644
--- a/arch/sparc/kernel/iommu.c
+++ b/arch/sparc/kernel/iommu.c
@@ -312,6 +312,8 @@ static dma_addr_t dma_4u_map_phys(struct device *dev, phys_addr_t phys,
if (direction != DMA_TO_DEVICE)
iopte_protection |= IOPTE_WRITE;

+ phys &= IO_PAGE_MASK;
+
for (i = 0; i < npages; i++, base++, phys += IO_PAGE_SIZE)
iopte_val(*base) = iopte_protection | phys;

diff --git a/arch/sparc/kernel/pci_sun4v.c b/arch/sparc/kernel/pci_sun4v.c
index 791f0a76665f..58ca4148f86b 100644
--- a/arch/sparc/kernel/pci_sun4v.c
+++ b/arch/sparc/kernel/pci_sun4v.c
@@ -410,6 +410,8 @@ static dma_addr_t dma_4v_map_phys(struct device *dev, phys_addr_t phys,

iommu_batch_start(dev, prot, entry);

+ phys &= IO_PAGE_MASK;
+
for (i = 0; i < npages; i++, phys += IO_PAGE_SIZE) {
long err = iommu_batch_add(phys, mask);
if (unlikely(err < 0L))
--
2.53.0