Hi Robin,

On 11/29/2017 6:29 AM, Robin Murphy wrote:
Hi Nate,

On 29/11/17 07:07, Nate Watterson wrote:
Hi Robin,

On 11/28/2017 12:27 PM, Robin Murphy wrote:
Bring io-pgtable-arm in line with the ARMv8.2-LPA feature allowing
52-bit physical addresses when using the 64KB translation granule.
This will be supported by SMMUv3.1.

Signed-off-by: Robin Murphy <robin.mur...@arm.com>
---
  drivers/iommu/io-pgtable-arm.c | 65 ++++++++++++++++++++++++++++++------------
  1 file changed, 47 insertions(+), 18 deletions(-)

diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
index 51e5c43caed1..4d46017b3262 100644
--- a/drivers/iommu/io-pgtable-arm.c
+++ b/drivers/iommu/io-pgtable-arm.c
@@ -21,6 +21,7 @@
  #define pr_fmt(fmt)    "arm-lpae io-pgtable: " fmt
  #include <linux/atomic.h>
+#include <linux/bitops.h>
  #include <linux/iommu.h>
  #include <linux/kernel.h>
  #include <linux/sizes.h>
@@ -32,7 +33,7 @@
  #include "io-pgtable.h"
-#define ARM_LPAE_MAX_ADDR_BITS        48
+#define ARM_LPAE_MAX_ADDR_BITS        52
  #define ARM_LPAE_S2_MAX_CONCAT_PAGES    16
  #define ARM_LPAE_MAX_LEVELS        4
@@ -86,6 +87,8 @@
  #define ARM_LPAE_PTE_TYPE_TABLE        3
  #define ARM_LPAE_PTE_TYPE_PAGE        3
+#define ARM_LPAE_PTE_ADDR_MASK        GENMASK_ULL(47,12)
+
  #define ARM_LPAE_PTE_NSTABLE        (((arm_lpae_iopte)1) << 63)
  #define ARM_LPAE_PTE_XN            (((arm_lpae_iopte)3) << 53)
  #define ARM_LPAE_PTE_AF            (((arm_lpae_iopte)1) << 10)
@@ -159,6 +162,7 @@
  #define ARM_LPAE_TCR_PS_42_BIT        0x3ULL
  #define ARM_LPAE_TCR_PS_44_BIT        0x4ULL
  #define ARM_LPAE_TCR_PS_48_BIT        0x5ULL
+#define ARM_LPAE_TCR_PS_52_BIT        0x6ULL
  #define ARM_LPAE_MAIR_ATTR_SHIFT(n)    ((n) << 3)
  #define ARM_LPAE_MAIR_ATTR_MASK        0xff
@@ -170,9 +174,7 @@
  #define ARM_LPAE_MAIR_ATTR_IDX_DEV    2
  /* IOPTE accessors */
-#define iopte_deref(pte,d)                    \
-    (__va((pte) & ((1ULL << ARM_LPAE_MAX_ADDR_BITS) - 1)    \
-    & ~(ARM_LPAE_GRANULE(d) - 1ULL)))
+#define iopte_deref(pte,d) __va(iopte_to_paddr(pte, d))
  #define iopte_type(pte,l)                    \
      (((pte) >> ARM_LPAE_PTE_TYPE_SHIFT) & ARM_LPAE_PTE_TYPE_MASK)
@@ -184,12 +186,6 @@
          (iopte_type(pte,l) == ARM_LPAE_PTE_TYPE_PAGE) :    \
          (iopte_type(pte,l) == ARM_LPAE_PTE_TYPE_BLOCK))
-#define iopte_to_pfn(pte,d)                    \
-    (((pte) & ((1ULL << ARM_LPAE_MAX_ADDR_BITS) - 1)) >> (d)->pg_shift)
-
-#define pfn_to_iopte(pfn,d)                    \
-    (((pfn) << (d)->pg_shift) & ((1ULL << ARM_LPAE_MAX_ADDR_BITS) - 1))
-
  struct arm_lpae_io_pgtable {
      struct io_pgtable    iop;
@@ -203,6 +199,25 @@ struct arm_lpae_io_pgtable {
  typedef u64 arm_lpae_iopte;
+static arm_lpae_iopte paddr_to_iopte(phys_addr_t paddr,
+                     struct arm_lpae_io_pgtable *data)
+{
+    arm_lpae_iopte pte = paddr;
+
+    /* Of the bits which overlap, either 51:48 or 15:12 are always RES0 */
+    return (pte | (pte >> 36)) & ARM_LPAE_PTE_ADDR_MASK;
+}
+
+static phys_addr_t iopte_to_paddr(arm_lpae_iopte pte,
+                  struct arm_lpae_io_pgtable *data)
+{
+    phys_addr_t paddr = pte & ARM_LPAE_PTE_ADDR_MASK;
+    phys_addr_t paddr_hi = paddr & (ARM_LPAE_GRANULE(data) - 1);
+
+    /* paddr_hi spans nothing for 4K granule, and only RES0 bits for 16K */
+    return (paddr ^ paddr_hi) | (paddr_hi << 36);
+}
+
  static bool selftest_running = false;
  static dma_addr_t __arm_lpae_dma_addr(void *pages)
@@ -287,7 +302,7 @@ static void __arm_lpae_init_pte(struct arm_lpae_io_pgtable 
*data,
          pte |= ARM_LPAE_PTE_TYPE_BLOCK;
      pte |= ARM_LPAE_PTE_AF | ARM_LPAE_PTE_SH_IS;
-    pte |= pfn_to_iopte(paddr >> data->pg_shift, data);
+    pte |= paddr_to_iopte(paddr, data);
      __arm_lpae_set_pte(ptep, pte, &data->iop.cfg);
  }
@@ -528,7 +543,7 @@ static int arm_lpae_split_blk_unmap(struct 
arm_lpae_io_pgtable *data,
      if (size == split_sz)
          unmap_idx = ARM_LPAE_LVL_IDX(iova, lvl, data);
-    blk_paddr = iopte_to_pfn(blk_pte, data) << data->pg_shift;
+    blk_paddr = iopte_to_paddr(blk_pte, data);
      pte = iopte_prot(blk_pte);
      for (i = 0; i < tablesz / sizeof(pte); i++, blk_paddr += split_sz) {
@@ -652,12 +667,13 @@ static phys_addr_t arm_lpae_iova_to_phys(struct 
io_pgtable_ops *ops,
  found_translation:
      iova &= (ARM_LPAE_BLOCK_SIZE(lvl, data) - 1);
-    return ((phys_addr_t)iopte_to_pfn(pte,data) << data->pg_shift) | iova;
+    return iopte_to_paddr(pte, data) | iova;
  }
  static void arm_lpae_restrict_pgsizes(struct io_pgtable_cfg *cfg)
  {
-    unsigned long granule;
+    unsigned long granule, page_sizes;
+    unsigned int max_addr_bits = 48;
      /*
       * We need to restrict the supported page sizes to match the
@@ -677,17 +693,24 @@ static void arm_lpae_restrict_pgsizes(struct 
io_pgtable_cfg *cfg)
      switch (granule) {
      case SZ_4K:
-        cfg->pgsize_bitmap &= (SZ_4K | SZ_2M | SZ_1G);
+        page_sizes = (SZ_4K | SZ_2M | SZ_1G);
          break;
      case SZ_16K:
-        cfg->pgsize_bitmap &= (SZ_16K | SZ_32M);
+        page_sizes = (SZ_16K | SZ_32M);
          break;
      case SZ_64K:
-        cfg->pgsize_bitmap &= (SZ_64K | SZ_512M);
+        max_addr_bits = 52;
+        page_sizes = (SZ_64K | SZ_512M);
+        if (cfg->oas > 48)
+            page_sizes |= 1ULL << 42; /* 4TB */
          break;
      default:
-        cfg->pgsize_bitmap = 0;
+        page_sizes = 0;
      }
+
+    cfg->pgsize_bitmap &= page_sizes;
+    cfg->ias = min(cfg->ias, max_addr_bits);
+    cfg->oas = min(cfg->oas, max_addr_bits);
  }
  static struct arm_lpae_io_pgtable *
@@ -784,6 +807,9 @@ arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, 
void *cookie)
      case 48:
          reg |= (ARM_LPAE_TCR_PS_48_BIT << ARM_LPAE_TCR_IPS_SHIFT);
          break;
+    case 52:
+        reg |= (ARM_LPAE_TCR_PS_52_BIT << ARM_LPAE_TCR_PS_SHIFT);
I think you probably intended to use ARM_LPAE_TCR_[I]PS_SHIFT here.

As originally written, I see F_ADDR_SIZE faults for transactions whose
output address exceeds 32-bits when testing on a 52-bit capable system.
Because of the apparent shift typo, TCR[IPS] was being left unconfigured
and this setting was in turn carried into CD[IPS] effectively limiting
OAS to 32-bits.

Oops, you're quite right - this was indeed a silly copy-paste error. Perhaps I 
should have mentioned that I don't have a model nor any hardware which does 
52-bit, so this is written to spec and only regression-tested for <=48-bit.

After fixing the shift typo, things seem to be working as expected.

Thanks! May I take that as a Tested-by for the series when I post a corrected 
v2?
Sure, I tested SMMU attached platform and PCIe devices in 48-bit and
52-bit IAS configurations. For the SMMU attached platform devices I
dumped the TLB and verified that 52-bit IOVAs were being used. I didn't
lean on the PCI devices to try to induce allocations beyond 32-bits,
however I expect that would work just fine.

Nate

Robin.


-Nate
+        break;
      default:
          goto out_free_data;
      }
@@ -891,6 +917,9 @@ arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, 
void *cookie)
      case 48:
          reg |= (ARM_LPAE_TCR_PS_48_BIT << ARM_LPAE_TCR_PS_SHIFT);
          break;
+    case 52:
+        reg |= (ARM_LPAE_TCR_PS_52_BIT << ARM_LPAE_TCR_PS_SHIFT);
+        break;
      default:
          goto out_free_data;
      }



--
Qualcomm Datacenter Technologies as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux 
Foundation Collaborative Project.
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Reply via email to