On Fri, 2020-10-23 at 12:21 +0100, Will Deacon wrote:

On Wed, Sep 30, 2020 at 03:06:34PM +0800, Yong Wu wrote:

quoted

The standard input iova bits is 32. MediaTek quad the lvl1 pagetable
(4 * lvl1). No change for lvl2 pagetable. Then the iova bits can reach
34bit.

Signed-off-by: Yong Wu <yong.wu@mediatek.com>
---
 drivers/iommu/io-pgtable-arm-v7s.c | 13 ++++++++++---
 drivers/iommu/mtk_iommu.c          |  2 +-
 2 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
index 8362fdf76657..306bae2755ed 100644
--- a/drivers/iommu/io-pgtable-arm-v7s.c
+++ b/drivers/iommu/io-pgtable-arm-v7s.c

@@ -50,10 +50,17 @@
  */
 #define ARM_V7S_ADDR_BITS		32

If we rename this to _ARM_V7S_ADDR_BITS can we then have ARM_V7S_ADDR_BITS
take a cfg parameter and move the arm_v7s_is_mtk_enabled() check in there?
Same for _ARM_V7S_LVL_BITS.

That would avoid scattering arm_v7s_is_mtk_enabled() into all the users.

I added "cfg" for _ARM_V7S_LVL_BITS in Robin's mail. is that ok?

Regarding ARM_V7S_ADDR_BITS, I'd like to keep it as is(Don't add cfg),
this macro only is used in ARM_V7S_LVL_SHIFT and checking the value of
ias/oas.

a) ARM_V7S_LVL_SHIFT always expect ARM_V7S_ADDR_BITS is 32.

b) our ias/oas is different(ias is 34 while oas is 35). If we define a
new macro, we need two about this, like:
#define ARM_V7S_IADDR_BITS(cfg)	(!arm_v7s_is_mtk_enabled(cfg) ? 32 : 34)
#define ARM_V7S_OADDR_BITS(cfg)	(!arm_v7s_is_mtk_enabled(cfg) ? 32 : 35)
and the two will only are used in the checking of ias/oas.

thus it looks unnecessary?

Will

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel