* [PATCH v3 0/3] io-pgtable ARM short descriptor format
@ 2016-01-26 17:13 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: will.deacon-5wv7dgnIgG8, yong.wu-NuS5LvNUpcJWk0Htik3J/w,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Cc: laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
Hi all,
As promised, here's v3 based on 4.5-rc1, with all the previous comments
addressed other than the logic of split_blk_unmap() - having looked into
it, trying to simplify that essentially just pushes the complexity out
into map() and complicates the general case. Ultimately it's an awkward
edge case to handle however it's implemented, and keeping it as close as
possible to the LPAE version makes it somewhat easier to reason about.
This is currently just a dependency for the Mediatek IOMMU series[1] -
the ARM SMMU driver support I've been testing with is now much closer to
"proper patch" than "horrible development hack", but still depends on a
bunch of other work in that driver I need to post first.
Thanks,
Robin.
Changes since v2[2]:
- #include dma-mapping.h before I break ARC again...
- Fix PTE type bit swizzling to not collide with XN bit in cont_to_pte().
- Correct init_pte to increment the IOVA, rather than the table pointer,
when freeing L2 tables being replaced by a supersection entry.
- Rename sync_flag to the much better tlb_sync_pending.
- Acks on patches 2 and 3.
[1]:http://thread.gmane.org/gmane.linux.kernel.iommu/11974
[2]:http://thread.gmane.org/gmane.linux.kernel.iommu/11752
Robin Murphy (3):
iommu/io-pgtable: Add ARMv7 short descriptor support
iommu/io-pgtable: Add helper functions for TLB ops
iommu/io-pgtable: Avoid redundant TLB syncs
drivers/iommu/Kconfig | 19 +
drivers/iommu/Makefile | 1 +
drivers/iommu/io-pgtable-arm-v7s.c | 841 +++++++++++++++++++++++++++++++++++++
drivers/iommu/io-pgtable-arm.c | 21 +-
drivers/iommu/io-pgtable.c | 5 +-
drivers/iommu/io-pgtable.h | 37 +-
6 files changed, 910 insertions(+), 14 deletions(-)
create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
--
2.7.0.25.gfc10eb5
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 0/3] io-pgtable ARM short descriptor format
@ 2016-01-26 17:13 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: linux-arm-kernel
Hi all,
As promised, here's v3 based on 4.5-rc1, with all the previous comments
addressed other than the logic of split_blk_unmap() - having looked into
it, trying to simplify that essentially just pushes the complexity out
into map() and complicates the general case. Ultimately it's an awkward
edge case to handle however it's implemented, and keeping it as close as
possible to the LPAE version makes it somewhat easier to reason about.
This is currently just a dependency for the Mediatek IOMMU series[1] -
the ARM SMMU driver support I've been testing with is now much closer to
"proper patch" than "horrible development hack", but still depends on a
bunch of other work in that driver I need to post first.
Thanks,
Robin.
Changes since v2[2]:
- #include dma-mapping.h before I break ARC again...
- Fix PTE type bit swizzling to not collide with XN bit in cont_to_pte().
- Correct init_pte to increment the IOVA, rather than the table pointer,
when freeing L2 tables being replaced by a supersection entry.
- Rename sync_flag to the much better tlb_sync_pending.
- Acks on patches 2 and 3.
[1]:http://thread.gmane.org/gmane.linux.kernel.iommu/11974
[2]:http://thread.gmane.org/gmane.linux.kernel.iommu/11752
Robin Murphy (3):
iommu/io-pgtable: Add ARMv7 short descriptor support
iommu/io-pgtable: Add helper functions for TLB ops
iommu/io-pgtable: Avoid redundant TLB syncs
drivers/iommu/Kconfig | 19 +
drivers/iommu/Makefile | 1 +
drivers/iommu/io-pgtable-arm-v7s.c | 841 +++++++++++++++++++++++++++++++++++++
drivers/iommu/io-pgtable-arm.c | 21 +-
drivers/iommu/io-pgtable.c | 5 +-
drivers/iommu/io-pgtable.h | 37 +-
6 files changed, 910 insertions(+), 14 deletions(-)
create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
--
2.7.0.25.gfc10eb5
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-01-26 17:13 ` Robin Murphy
@ 2016-01-26 17:13 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: will.deacon-5wv7dgnIgG8, yong.wu-NuS5LvNUpcJWk0Htik3J/w,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Cc: laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
Add a nearly-complete ARMv7 short descriptor implementation, omitting
only a few legacy and CPU-centric aspects which shouldn't be necessary
for IOMMU API use anyway.
Signed-off-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
---
drivers/iommu/Kconfig | 19 +
drivers/iommu/Makefile | 1 +
drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
drivers/iommu/io-pgtable.c | 3 +
drivers/iommu/io-pgtable.h | 14 +-
5 files changed, 885 insertions(+), 1 deletion(-)
create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index a1e75cb..dc1aaa5 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
If unsure, say N here.
+config IOMMU_IO_PGTABLE_ARMV7S
+ bool "ARMv7/v8 Short Descriptor Format"
+ select IOMMU_IO_PGTABLE
+ depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
+ help
+ Enable support for the ARM Short-descriptor pagetable format.
+ This supports 32-bit virtual and physical addresses mapped using
+ 2-level tables with 4KB pages/1MB sections, and contiguous entries
+ for 64KB pages/16MB supersections if indicated by the IOMMU driver.
+
+config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
+ bool "ARMv7s selftests"
+ depends on IOMMU_IO_PGTABLE_ARMV7S
+ help
+ Enable self-tests for ARMv7s page table allocator. This performs
+ a series of page-table consistency checks during boot.
+
+ If unsure, say N here.
+
endmenu
config IOMMU_IOVA
diff --git a/drivers/iommu/Makefile b/drivers/iommu/Makefile
index 42fc0c2..2f9bfbc 100644
--- a/drivers/iommu/Makefile
+++ b/drivers/iommu/Makefile
@@ -3,6 +3,7 @@ obj-$(CONFIG_IOMMU_API) += iommu-traces.o
obj-$(CONFIG_IOMMU_API) += iommu-sysfs.o
obj-$(CONFIG_IOMMU_DMA) += dma-iommu.o
obj-$(CONFIG_IOMMU_IO_PGTABLE) += io-pgtable.o
+obj-$(CONFIG_IOMMU_IO_PGTABLE_ARMV7S) += io-pgtable-arm-v7s.o
obj-$(CONFIG_IOMMU_IO_PGTABLE_LPAE) += io-pgtable-arm.o
obj-$(CONFIG_IOMMU_IOVA) += iova.o
obj-$(CONFIG_OF_IOMMU) += of_iommu.o
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
new file mode 100644
index 0000000..efc7d1e
--- /dev/null
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -0,0 +1,849 @@
+/*
+ * CPU-agnostic ARM page table allocator.
+ *
+ * ARMv7 Short-descriptor format, supporting
+ * - Basic memory attributes
+ * - Simplified access permissions (AP[2:1] model)
+ * - Backwards-compatible TEX remap
+ * - Large pages/supersections (if indicated by the caller)
+ *
+ * Not supporting:
+ * - Legacy access permissions (AP[2:0] model)
+ *
+ * Almost certainly never supporting:
+ * - PXN
+ * - Domains
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (C) 2014-2015 ARM Limited
+ * Copyright (c) 2014-2015 MediaTek Inc.
+ */
+
+#define pr_fmt(fmt) "arm-v7s io-pgtable: " fmt
+
+#include <linux/dma-mapping.h>
+#include <linux/gfp.h>
+#include <linux/iommu.h>
+#include <linux/kernel.h>
+#include <linux/kmemleak.h>
+#include <linux/sizes.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include <asm/barrier.h>
+
+#include "io-pgtable.h"
+
+/* Struct accessors */
+#define io_pgtable_to_data(x) \
+ container_of((x), struct arm_v7s_io_pgtable, iop)
+
+#define io_pgtable_ops_to_data(x) \
+ io_pgtable_to_data(io_pgtable_ops_to_pgtable(x))
+
+/*
+ * We have 32 bits total; 12 bits resolved at level 1, 8 bits at level 2,
+ * and 12 bits in a page. With some carefully-chosen coefficients we can
+ * hide the ugly inconsistencies behind these macros and at least let the
+ * rest of the code pretend to be somewhat sane.
+ */
+#define ARM_V7S_ADDR_BITS 32
+#define _ARM_V7S_LVL_BITS(lvl) (16 - (lvl) * 4)
+#define ARM_V7S_LVL_SHIFT(lvl) (ARM_V7S_ADDR_BITS - (4 + 8 * (lvl)))
+#define ARM_V7S_TABLE_SHIFT 10
+
+#define ARM_V7S_PTES_PER_LVL(lvl) (1 << _ARM_V7S_LVL_BITS(lvl))
+#define ARM_V7S_TABLE_SIZE(lvl) \
+ (ARM_V7S_PTES_PER_LVL(lvl) * sizeof(arm_v7s_iopte))
+
+#define ARM_V7S_BLOCK_SIZE(lvl) (1UL << ARM_V7S_LVL_SHIFT(lvl))
+#define ARM_V7S_LVL_MASK(lvl) ((u32)(~0U << ARM_V7S_LVL_SHIFT(lvl)))
+#define ARM_V7S_TABLE_MASK ((u32)(~0U << ARM_V7S_TABLE_SHIFT))
+#define _ARM_V7S_IDX_MASK(lvl) (ARM_V7S_PTES_PER_LVL(lvl) - 1)
+#define ARM_V7S_LVL_IDX(addr, lvl) ({ \
+ int _l = lvl; \
+ ((u32)(addr) >> ARM_V7S_LVL_SHIFT(_l)) & _ARM_V7S_IDX_MASK(_l); \
+})
+
+/*
+ * Large page/supersection entries are effectively a block of 16 page/section
+ * entries, along the lines of the LPAE contiguous hint, but all with the
+ * same output address. For want of a better common name we'll call them
+ * "contiguous" versions of their respective page/section entries here, but
+ * noting the distinction (WRT to TLB maintenance) that they represent *one*
+ * entry repeated 16 times, not 16 separate entries (as in the LPAE case).
+ */
+#define ARM_V7S_CONT_PAGES 16
+
+/* PTE type bits: these are all mixed up with XN/PXN bits in most cases */
+#define ARM_V7S_PTE_TYPE_TABLE 0x1
+#define ARM_V7S_PTE_TYPE_PAGE 0x2
+#define ARM_V7S_PTE_TYPE_CONT_PAGE 0x1
+
+#define ARM_V7S_PTE_IS_VALID(pte) (((pte) & 0x3) != 0)
+#define ARM_V7S_PTE_IS_TABLE(pte, lvl) (lvl == 1 && ((pte) & ARM_V7S_PTE_TYPE_TABLE))
+
+/* Page table bits */
+#define ARM_V7S_ATTR_XN(lvl) BIT(4 * (2 - (lvl)))
+#define ARM_V7S_ATTR_B BIT(2)
+#define ARM_V7S_ATTR_C BIT(3)
+#define ARM_V7S_ATTR_NS_TABLE BIT(3)
+#define ARM_V7S_ATTR_NS_SECTION BIT(19)
+
+#define ARM_V7S_CONT_SECTION BIT(18)
+#define ARM_V7S_CONT_PAGE_XN_SHIFT 15
+
+/*
+ * The attribute bits are consistently ordered*, but occupy bits [17:10] of
+ * a level 1 PTE vs. bits [11:4] at level 2. Thus we define the individual
+ * fields relative to that 8-bit block, plus a total shift relative to the PTE.
+ */
+#define ARM_V7S_ATTR_SHIFT(lvl) (16 - (lvl) * 6)
+
+#define ARM_V7S_ATTR_MASK 0xff
+#define ARM_V7S_ATTR_AP0 BIT(0)
+#define ARM_V7S_ATTR_AP1 BIT(1)
+#define ARM_V7S_ATTR_AP2 BIT(5)
+#define ARM_V7S_ATTR_S BIT(6)
+#define ARM_V7S_ATTR_NG BIT(7)
+#define ARM_V7S_TEX_SHIFT 2
+#define ARM_V7S_TEX_MASK 0x7
+#define ARM_V7S_ATTR_TEX(val) (((val) & ARM_V7S_TEX_MASK) << ARM_V7S_TEX_SHIFT)
+
+/* *well, except for TEX on level 2 large pages, of course :( */
+#define ARM_V7S_CONT_PAGE_TEX_SHIFT 6
+#define ARM_V7S_CONT_PAGE_TEX_MASK (ARM_V7S_TEX_MASK << ARM_V7S_CONT_PAGE_TEX_SHIFT)
+
+/* Simplified access permissions */
+#define ARM_V7S_PTE_AF ARM_V7S_ATTR_AP0
+#define ARM_V7S_PTE_AP_UNPRIV ARM_V7S_ATTR_AP1
+#define ARM_V7S_PTE_AP_RDONLY ARM_V7S_ATTR_AP2
+
+/* Register bits */
+#define ARM_V7S_RGN_NC 0
+#define ARM_V7S_RGN_WBWA 1
+#define ARM_V7S_RGN_WT 2
+#define ARM_V7S_RGN_WB 3
+
+#define ARM_V7S_PRRR_TYPE_DEVICE 1
+#define ARM_V7S_PRRR_TYPE_NORMAL 2
+#define ARM_V7S_PRRR_TR(n, type) (((type) & 0x3) << ((n) * 2))
+#define ARM_V7S_PRRR_DS0 BIT(16)
+#define ARM_V7S_PRRR_DS1 BIT(17)
+#define ARM_V7S_PRRR_NS0 BIT(18)
+#define ARM_V7S_PRRR_NS1 BIT(19)
+#define ARM_V7S_PRRR_NOS(n) BIT((n) + 24)
+
+#define ARM_V7S_NMRR_IR(n, attr) (((attr) & 0x3) << ((n) * 2))
+#define ARM_V7S_NMRR_OR(n, attr) (((attr) & 0x3) << ((n) * 2 + 16))
+
+#define ARM_V7S_TTBR_S BIT(1)
+#define ARM_V7S_TTBR_NOS BIT(5)
+#define ARM_V7S_TTBR_ORGN_ATTR(attr) (((attr) & 0x3) << 3)
+#define ARM_V7S_TTBR_IRGN_ATTR(attr) \
+ ((((attr) & 0x1) << 6) | (((attr) & 0x2) >> 1))
+
+#define ARM_V7S_TCR_PD1 BIT(5)
+
+typedef u32 arm_v7s_iopte;
+
+static bool selftest_running;
+
+struct arm_v7s_io_pgtable {
+ struct io_pgtable iop;
+
+ arm_v7s_iopte *pgd;
+ struct kmem_cache *l2_tables;
+};
+
+static dma_addr_t __arm_v7s_dma_addr(void *pages)
+{
+ return (dma_addr_t)virt_to_phys(pages);
+}
+
+static arm_v7s_iopte *iopte_deref(arm_v7s_iopte pte, int lvl)
+{
+ if (ARM_V7S_PTE_IS_TABLE(pte, lvl))
+ pte &= ARM_V7S_TABLE_MASK;
+ else
+ pte &= ARM_V7S_LVL_MASK(lvl);
+ return phys_to_virt(pte);
+}
+
+static void *__arm_v7s_alloc_table(int lvl, gfp_t gfp,
+ struct arm_v7s_io_pgtable *data)
+{
+ struct device *dev = data->iop.cfg.iommu_dev;
+ dma_addr_t dma;
+ size_t size = ARM_V7S_TABLE_SIZE(lvl);
+ void *table = NULL;
+
+ if (lvl == 1)
+ table = (void *)__get_dma_pages(__GFP_ZERO, get_order(size));
+ else if (lvl == 2)
+ table = kmem_cache_zalloc(data->l2_tables, gfp);
+ if (table && !selftest_running) {
+ dma = dma_map_single(dev, table, size, DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, dma))
+ goto out_free;
+ /*
+ * We depend on the IOMMU being able to work with any physical
+ * address directly, so if the DMA layer suggests otherwise by
+ * translating or truncating them, that bodes very badly...
+ */
+ if (dma != virt_to_phys(table))
+ goto out_unmap;
+ }
+ kmemleak_ignore(table);
+ return table;
+
+out_unmap:
+ dev_err(dev, "Cannot accommodate DMA translation for IOMMU page tables\n");
+ dma_unmap_single(dev, dma, size, DMA_TO_DEVICE);
+out_free:
+ if (lvl == 1)
+ free_pages((unsigned long)table, get_order(size));
+ else
+ kmem_cache_free(data->l2_tables, table);
+ return NULL;
+}
+
+static void __arm_v7s_free_table(void *table, int lvl,
+ struct arm_v7s_io_pgtable *data)
+{
+ struct device *dev = data->iop.cfg.iommu_dev;
+ size_t size = ARM_V7S_TABLE_SIZE(lvl);
+
+ if (!selftest_running)
+ dma_unmap_single(dev, __arm_v7s_dma_addr(table), size,
+ DMA_TO_DEVICE);
+ if (lvl == 1)
+ free_pages((unsigned long)table, get_order(size));
+ else
+ kmem_cache_free(data->l2_tables, table);
+}
+
+static void __arm_v7s_pte_sync(arm_v7s_iopte *ptep, int num_entries,
+ struct io_pgtable_cfg *cfg)
+{
+ if (selftest_running)
+ return;
+
+ dma_sync_single_for_device(cfg->iommu_dev, __arm_v7s_dma_addr(ptep),
+ num_entries * sizeof(*ptep), DMA_TO_DEVICE);
+}
+static void __arm_v7s_set_pte(arm_v7s_iopte *ptep, arm_v7s_iopte pte,
+ int num_entries, struct io_pgtable_cfg *cfg)
+{
+ int i;
+
+ for (i = 0; i < num_entries; i++)
+ ptep[i] = pte;
+
+ __arm_v7s_pte_sync(ptep, num_entries, cfg);
+}
+
+static arm_v7s_iopte arm_v7s_prot_to_pte(int prot, int lvl,
+ struct io_pgtable_cfg *cfg)
+{
+ bool ap = !(cfg->quirks & IO_PGTABLE_QUIRK_NO_PERMS);
+ arm_v7s_iopte pte = ARM_V7S_ATTR_NG | ARM_V7S_ATTR_S |
+ ARM_V7S_ATTR_TEX(1);
+
+ if (ap) {
+ pte |= ARM_V7S_PTE_AF | ARM_V7S_PTE_AP_UNPRIV;
+ if (!(prot & IOMMU_WRITE))
+ pte |= ARM_V7S_PTE_AP_RDONLY;
+ }
+ pte <<= ARM_V7S_ATTR_SHIFT(lvl);
+
+ if ((prot & IOMMU_NOEXEC) && ap)
+ pte |= ARM_V7S_ATTR_XN(lvl);
+ if (prot & IOMMU_CACHE)
+ pte |= ARM_V7S_ATTR_B | ARM_V7S_ATTR_C;
+
+ return pte;
+}
+
+static int arm_v7s_pte_to_prot(arm_v7s_iopte pte, int lvl)
+{
+ int prot = IOMMU_READ;
+
+ if (pte & (ARM_V7S_PTE_AP_RDONLY << ARM_V7S_ATTR_SHIFT(lvl)))
+ prot |= IOMMU_WRITE;
+ if (pte & ARM_V7S_ATTR_C)
+ prot |= IOMMU_CACHE;
+
+ return prot;
+}
+
+static arm_v7s_iopte arm_v7s_pte_to_cont(arm_v7s_iopte pte, int lvl)
+{
+ if (lvl == 1) {
+ pte |= ARM_V7S_CONT_SECTION;
+ } else if (lvl == 2) {
+ arm_v7s_iopte xn = pte & ARM_V7S_ATTR_XN(lvl);
+ arm_v7s_iopte tex = pte & ARM_V7S_CONT_PAGE_TEX_MASK;
+
+ pte ^= xn | tex | ARM_V7S_PTE_TYPE_PAGE;
+ pte |= (xn << ARM_V7S_CONT_PAGE_XN_SHIFT) |
+ (tex << ARM_V7S_CONT_PAGE_TEX_SHIFT) |
+ ARM_V7S_PTE_TYPE_CONT_PAGE;
+ }
+ return pte;
+}
+
+static arm_v7s_iopte arm_v7s_cont_to_pte(arm_v7s_iopte pte, int lvl)
+{
+ if (lvl == 1) {
+ pte &= ~ARM_V7S_CONT_SECTION;
+ } else if (lvl == 2) {
+ arm_v7s_iopte xn = pte & BIT(ARM_V7S_CONT_PAGE_XN_SHIFT);
+ arm_v7s_iopte tex = pte & (ARM_V7S_CONT_PAGE_TEX_MASK <<
+ ARM_V7S_CONT_PAGE_TEX_SHIFT);
+
+ pte ^= xn | tex | ARM_V7S_PTE_TYPE_CONT_PAGE;
+ pte |= (xn >> ARM_V7S_CONT_PAGE_XN_SHIFT) |
+ (tex >> ARM_V7S_CONT_PAGE_TEX_SHIFT) |
+ ARM_V7S_PTE_TYPE_PAGE;
+ }
+ return pte;
+}
+
+static bool arm_v7s_pte_is_cont(arm_v7s_iopte pte, int lvl)
+{
+ if (lvl == 1 && !ARM_V7S_PTE_IS_TABLE(pte, lvl))
+ return pte & ARM_V7S_CONT_SECTION;
+ else if (lvl == 2)
+ return !(pte & ARM_V7S_PTE_TYPE_PAGE);
+ return false;
+}
+
+static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *, unsigned long,
+ size_t, int, arm_v7s_iopte *);
+
+static int arm_v7s_init_pte(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, phys_addr_t paddr, int prot,
+ int lvl, int num_entries, arm_v7s_iopte *ptep)
+{
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ arm_v7s_iopte pte = arm_v7s_prot_to_pte(prot, lvl, cfg);
+ int i;
+
+ for (i = 0; i < num_entries; i++)
+ if (ARM_V7S_PTE_IS_TABLE(ptep[i], lvl)) {
+ /*
+ * We need to unmap and free the old table before
+ * overwriting it with a block entry.
+ */
+ arm_v7s_iopte *tblp;
+ size_t sz = ARM_V7S_BLOCK_SIZE(lvl);
+
+ tblp = ptep - ARM_V7S_LVL_IDX(iova, lvl);
+ if (WARN_ON(__arm_v7s_unmap(data, iova + i * sz,
+ sz, lvl, tblp) != sz))
+ return -EINVAL;
+ } else if (ptep[i]) {
+ /* We require an unmap first */
+ WARN_ON(!selftest_running);
+ return -EEXIST;
+ }
+
+ pte |= ARM_V7S_PTE_TYPE_PAGE;
+ if (lvl == 1 && (cfg->quirks & IO_PGTABLE_QUIRK_ARM_NS))
+ pte |= ARM_V7S_ATTR_NS_SECTION;
+
+ if (num_entries > 1)
+ pte = arm_v7s_pte_to_cont(pte, lvl);
+
+ pte |= paddr & ARM_V7S_LVL_MASK(lvl);
+
+ __arm_v7s_set_pte(ptep, pte, num_entries, cfg);
+ return 0;
+}
+
+static int __arm_v7s_map(struct arm_v7s_io_pgtable *data, unsigned long iova,
+ phys_addr_t paddr, size_t size, int prot,
+ int lvl, arm_v7s_iopte *ptep)
+{
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ arm_v7s_iopte pte, *cptep;
+ int num_entries = size >> ARM_V7S_LVL_SHIFT(lvl);
+
+ /* Find our entry at the current level */
+ ptep += ARM_V7S_LVL_IDX(iova, lvl);
+
+ /* If we can install a leaf entry at this level, then do so */
+ if (num_entries)
+ return arm_v7s_init_pte(data, iova, paddr, prot,
+ lvl, num_entries, ptep);
+
+ /* We can't allocate tables at the final level */
+ if (WARN_ON(lvl == 2))
+ return -EINVAL;
+
+ /* Grab a pointer to the next level */
+ pte = *ptep;
+ if (!pte) {
+ cptep = __arm_v7s_alloc_table(lvl + 1, GFP_ATOMIC, data);
+ if (!cptep)
+ return -ENOMEM;
+
+ pte = virt_to_phys(cptep) | ARM_V7S_PTE_TYPE_TABLE;
+ if (cfg->quirks & IO_PGTABLE_QUIRK_ARM_NS)
+ pte |= ARM_V7S_ATTR_NS_TABLE;
+
+ __arm_v7s_set_pte(ptep, pte, 1, cfg);
+ } else {
+ cptep = iopte_deref(pte, lvl);
+ }
+
+ /* Rinse, repeat */
+ return __arm_v7s_map(data, iova, paddr, size, prot, lvl + 1, cptep);
+}
+
+static int arm_v7s_map(struct io_pgtable_ops *ops, unsigned long iova,
+ phys_addr_t paddr, size_t size, int prot)
+{
+ struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ const struct iommu_gather_ops *tlb = cfg->tlb;
+ void *cookie = data->iop.cookie;
+ int ret;
+
+ /* If no access, then nothing to do */
+ if (!(prot & (IOMMU_READ | IOMMU_WRITE)))
+ return 0;
+
+ ret = __arm_v7s_map(data, iova, paddr, size, prot, 1, data->pgd);
+ /*
+ * Synchronise all PTE updates for the new mapping before there's
+ * a chance for anything to kick off a table walk for the new iova.
+ */
+ if (cfg->quirks & IO_PGTABLE_QUIRK_TLBI_ON_MAP) {
+ tlb->tlb_add_flush(iova, size, ARM_V7S_BLOCK_SIZE(2), false,
+ cookie);
+ tlb->tlb_sync(cookie);
+ } else {
+ wmb();
+ }
+
+ return ret;
+}
+
+static void arm_v7s_free_pgtable(struct io_pgtable *iop)
+{
+ struct arm_v7s_io_pgtable *data = io_pgtable_to_data(iop);
+ int i;
+
+ for (i = 0; i < ARM_V7S_PTES_PER_LVL(1); i++) {
+ arm_v7s_iopte pte = data->pgd[i];
+
+ if (ARM_V7S_PTE_IS_TABLE(pte, 1))
+ __arm_v7s_free_table(iopte_deref(pte, 1), 2, data);
+ }
+ __arm_v7s_free_table(data->pgd, 1, data);
+ kmem_cache_destroy(data->l2_tables);
+ kfree(data);
+}
+
+static void arm_v7s_split_cont(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, int idx, int lvl,
+ arm_v7s_iopte *ptep)
+{
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ void *cookie = data->iop.cookie;
+ arm_v7s_iopte pte;
+ size_t size = ARM_V7S_BLOCK_SIZE(lvl);
+ int i;
+
+ ptep -= idx & (ARM_V7S_CONT_PAGES - 1);
+ pte = arm_v7s_cont_to_pte(*ptep, lvl);
+ for (i = 0; i < ARM_V7S_CONT_PAGES; i++) {
+ ptep[i] = pte;
+ pte += size;
+ }
+
+ __arm_v7s_pte_sync(ptep, ARM_V7S_CONT_PAGES, cfg);
+
+ size *= ARM_V7S_CONT_PAGES;
+ cfg->tlb->tlb_add_flush(iova, size, size, true, cookie);
+ cfg->tlb->tlb_sync(cookie);
+}
+
+static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, size_t size,
+ arm_v7s_iopte *ptep)
+{
+ unsigned long blk_start, blk_end, blk_size;
+ phys_addr_t blk_paddr;
+ arm_v7s_iopte table = 0;
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ int prot = arm_v7s_pte_to_prot(*ptep, 1);
+
+ blk_size = ARM_V7S_BLOCK_SIZE(1);
+ blk_start = iova & ARM_V7S_LVL_MASK(1);
+ blk_end = blk_start + ARM_V7S_BLOCK_SIZE(1);
+ blk_paddr = *ptep & ARM_V7S_LVL_MASK(1);
+
+ for (; blk_start < blk_end; blk_start += size, blk_paddr += size) {
+ arm_v7s_iopte *tablep;
+
+ /* Unmap! */
+ if (blk_start == iova)
+ continue;
+
+ /* __arm_v7s_map expects a pointer to the start of the table */
+ tablep = &table - ARM_V7S_LVL_IDX(blk_start, 1);
+ if (__arm_v7s_map(data, blk_start, blk_paddr, size, prot, 1,
+ tablep) < 0) {
+ if (table) {
+ /* Free the table we allocated */
+ tablep = iopte_deref(table, 1);
+ __arm_v7s_free_table(tablep, 2, data);
+ }
+ return 0; /* Bytes unmapped */
+ }
+ }
+
+ __arm_v7s_set_pte(ptep, table, 1, cfg);
+ iova &= ~(blk_size - 1);
+ cfg->tlb->tlb_add_flush(iova, blk_size, blk_size, true, data->iop.cookie);
+ return size;
+}
+
+static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, size_t size, int lvl,
+ arm_v7s_iopte *ptep)
+{
+ arm_v7s_iopte pte[ARM_V7S_CONT_PAGES];
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ const struct iommu_gather_ops *tlb = cfg->tlb;
+ void *cookie = data->iop.cookie;
+ int idx, i = 0, num_entries = size >> ARM_V7S_LVL_SHIFT(lvl);
+
+ /* Something went horribly wrong and we ran out of page table */
+ if (WARN_ON(lvl > 2))
+ return 0;
+
+ idx = ARM_V7S_LVL_IDX(iova, lvl);
+ ptep += idx;
+ do {
+ if (WARN_ON(!ARM_V7S_PTE_IS_VALID(ptep[i])))
+ return 0;
+ pte[i] = ptep[i];
+ } while (++i < num_entries);
+
+ /*
+ * If we've hit a contiguous 'large page' entry at this level, it
+ * needs splitting first, unless we're unmapping the whole lot.
+ */
+ if (num_entries <= 1 && arm_v7s_pte_is_cont(pte[0], lvl))
+ arm_v7s_split_cont(data, iova, idx, lvl, ptep);
+
+ /* If the size matches this level, we're in the right place */
+ if (num_entries) {
+ size_t blk_size = ARM_V7S_BLOCK_SIZE(lvl);
+
+ __arm_v7s_set_pte(ptep, 0, num_entries, cfg);
+
+ for (i = 0; i < num_entries; i++) {
+ if (ARM_V7S_PTE_IS_TABLE(pte[i], lvl)) {
+ /* Also flush any partial walks */
+ tlb->tlb_add_flush(iova, blk_size,
+ ARM_V7S_BLOCK_SIZE(lvl + 1),
+ false, cookie);
+ tlb->tlb_sync(cookie);
+ ptep = iopte_deref(pte[i], lvl);
+ __arm_v7s_free_table(ptep, lvl + 1, data);
+ } else {
+ tlb->tlb_add_flush(iova, blk_size, blk_size,
+ true, cookie);
+ }
+ iova += blk_size;
+ }
+ return size;
+ } else if (lvl == 1 && !ARM_V7S_PTE_IS_TABLE(pte[0], lvl)) {
+ /*
+ * Insert a table at the next level to map the old region,
+ * minus the part we want to unmap
+ */
+ return arm_v7s_split_blk_unmap(data, iova, size, ptep);
+ }
+
+ /* Keep on walkin' */
+ ptep = iopte_deref(pte[0], lvl);
+ return __arm_v7s_unmap(data, iova, size, lvl + 1, ptep);
+}
+
+static int arm_v7s_unmap(struct io_pgtable_ops *ops, unsigned long iova,
+ size_t size)
+{
+ size_t unmapped;
+ struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
+ struct io_pgtable *iop = &data->iop;
+
+ unmapped = __arm_v7s_unmap(data, iova, size, 1, data->pgd);
+ if (unmapped)
+ iop->cfg.tlb->tlb_sync(iop->cookie);
+
+ return unmapped;
+}
+
+static phys_addr_t arm_v7s_iova_to_phys(struct io_pgtable_ops *ops,
+ unsigned long iova)
+{
+ struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
+ arm_v7s_iopte *ptep = data->pgd, pte;
+ int lvl = 0;
+ u32 mask;
+
+ do {
+ pte = ptep[ARM_V7S_LVL_IDX(iova, ++lvl)];
+ ptep = iopte_deref(pte, lvl);
+ } while (ARM_V7S_PTE_IS_TABLE(pte, lvl));
+
+ if (!ARM_V7S_PTE_IS_VALID(pte))
+ return 0;
+
+ mask = ARM_V7S_LVL_MASK(lvl);
+ if (arm_v7s_pte_is_cont(pte, lvl))
+ mask *= ARM_V7S_CONT_PAGES;
+ return (pte & mask) | (iova & ~mask);
+}
+
+static struct io_pgtable *arm_v7s_alloc_pgtable(struct io_pgtable_cfg *cfg,
+ void *cookie)
+{
+ struct arm_v7s_io_pgtable *data;
+
+ if (cfg->ias > ARM_V7S_ADDR_BITS || cfg->oas > ARM_V7S_ADDR_BITS)
+ return NULL;
+
+ data = kmalloc(sizeof(*data), GFP_KERNEL);
+ if (!data)
+ return NULL;
+
+ data->l2_tables = kmem_cache_create("io-pgtable_armv7s_l2",
+ ARM_V7S_TABLE_SIZE(2),
+ ARM_V7S_TABLE_SIZE(2),
+ SLAB_CACHE_DMA, NULL);
+ if (!data->l2_tables)
+ goto out_free_data;
+
+ data->iop.ops = (struct io_pgtable_ops) {
+ .map = arm_v7s_map,
+ .unmap = arm_v7s_unmap,
+ .iova_to_phys = arm_v7s_iova_to_phys,
+ };
+
+ /* We have to do this early for __arm_v7s_alloc_table to work... */
+ data->iop.cfg = *cfg;
+
+ /*
+ * Unless the IOMMU driver indicates supersection support by
+ * having SZ_16M set in the initial bitmap, they won't be used.
+ */
+ cfg->pgsize_bitmap &= SZ_4K | SZ_64K | SZ_1M | SZ_16M;
+
+ /* TCR: T0SZ=0, disable TTBR1 */
+ cfg->arm_v7s_cfg.tcr = ARM_V7S_TCR_PD1;
+
+ /*
+ * TEX remap: the indices used map to the closest equivalent types
+ * under the non-TEX-remap interpretation of those attribute bits,
+ * excepting various implementation-defined aspects of shareability.
+ */
+ cfg->arm_v7s_cfg.prrr = ARM_V7S_PRRR_TR(1, ARM_V7S_PRRR_TYPE_DEVICE) |
+ ARM_V7S_PRRR_TR(4, ARM_V7S_PRRR_TYPE_NORMAL) |
+ ARM_V7S_PRRR_TR(7, ARM_V7S_PRRR_TYPE_NORMAL) |
+ ARM_V7S_PRRR_DS0 | ARM_V7S_PRRR_DS1 |
+ ARM_V7S_PRRR_NS1 | ARM_V7S_PRRR_NOS(7);
+ cfg->arm_v7s_cfg.nmrr = ARM_V7S_NMRR_IR(7, ARM_V7S_RGN_WBWA) |
+ ARM_V7S_NMRR_OR(7, ARM_V7S_RGN_WBWA);
+
+ /* Looking good; allocate a pgd */
+ data->pgd = __arm_v7s_alloc_table(1, GFP_KERNEL, data);
+ if (!data->pgd)
+ goto out_free_data;
+
+ /* Ensure the empty pgd is visible before any actual TTBR write */
+ wmb();
+
+ /* TTBRs */
+ cfg->arm_v7s_cfg.ttbr[0] = virt_to_phys(data->pgd) |
+ ARM_V7S_TTBR_S | ARM_V7S_TTBR_NOS |
+ ARM_V7S_TTBR_IRGN_ATTR(ARM_V7S_RGN_WBWA) |
+ ARM_V7S_TTBR_ORGN_ATTR(ARM_V7S_RGN_WBWA);
+ cfg->arm_v7s_cfg.ttbr[1] = 0;
+ return &data->iop;
+
+out_free_data:
+ kmem_cache_destroy(data->l2_tables);
+ kfree(data);
+ return NULL;
+}
+
+struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
+ .alloc = arm_v7s_alloc_pgtable,
+ .free = arm_v7s_free_pgtable,
+};
+
+#ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
+
+static struct io_pgtable_cfg *cfg_cookie;
+
+static void dummy_tlb_flush_all(void *cookie)
+{
+ WARN_ON(cookie != cfg_cookie);
+}
+
+static void dummy_tlb_add_flush(unsigned long iova, size_t size,
+ size_t granule, bool leaf, void *cookie)
+{
+ WARN_ON(cookie != cfg_cookie);
+ WARN_ON(!(size & cfg_cookie->pgsize_bitmap));
+}
+
+static void dummy_tlb_sync(void *cookie)
+{
+ WARN_ON(cookie != cfg_cookie);
+}
+
+static struct iommu_gather_ops dummy_tlb_ops = {
+ .tlb_flush_all = dummy_tlb_flush_all,
+ .tlb_add_flush = dummy_tlb_add_flush,
+ .tlb_sync = dummy_tlb_sync,
+};
+
+#define __FAIL(ops) ({ \
+ WARN(1, "selftest: test failed\n"); \
+ selftest_running = false; \
+ -EFAULT; \
+})
+
+static int __init arm_v7s_do_selftests(void)
+{
+ struct io_pgtable_ops *ops;
+ struct io_pgtable_cfg cfg = {
+ .tlb = &dummy_tlb_ops,
+ .oas = 32,
+ .ias = 32,
+ .quirks = IO_PGTABLE_QUIRK_ARM_NS,
+ .pgsize_bitmap = SZ_4K | SZ_64K | SZ_1M | SZ_16M,
+ };
+ unsigned int iova, size, iova_start;
+ unsigned int i, loopnr = 0;
+
+ selftest_running = true;
+
+ cfg_cookie = &cfg;
+
+ ops = alloc_io_pgtable_ops(ARM_V7S, &cfg, &cfg);
+ if (!ops) {
+ pr_err("selftest: failed to allocate io pgtable ops\n");
+ return -EINVAL;
+ }
+
+ /*
+ * Initial sanity checks.
+ * Empty page tables shouldn't provide any translations.
+ */
+ if (ops->iova_to_phys(ops, 42))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, SZ_1G + 42))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, SZ_2G + 42))
+ return __FAIL(ops);
+
+ /*
+ * Distinct mappings of different granule sizes.
+ */
+ iova = 0;
+ i = find_first_bit(&cfg.pgsize_bitmap, BITS_PER_LONG);
+ while (i != BITS_PER_LONG) {
+ size = 1UL << i;
+ if (ops->map(ops, iova, iova, size, IOMMU_READ |
+ IOMMU_WRITE |
+ IOMMU_NOEXEC |
+ IOMMU_CACHE))
+ return __FAIL(ops);
+
+ /* Overlapping mappings */
+ if (!ops->map(ops, iova, iova + size, size,
+ IOMMU_READ | IOMMU_NOEXEC))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova + 42) != (iova + 42))
+ return __FAIL(ops);
+
+ iova += SZ_16M;
+ i++;
+ i = find_next_bit(&cfg.pgsize_bitmap, BITS_PER_LONG, i);
+ loopnr++;
+ }
+
+ /* Partial unmap */
+ i = 1;
+ size = 1UL << __ffs(cfg.pgsize_bitmap);
+ while (i < loopnr) {
+ iova_start = i * SZ_16M;
+ if (ops->unmap(ops, iova_start + size, size) != size)
+ return __FAIL(ops);
+
+ /* Remap of partial unmap */
+ if (ops->map(ops, iova_start + size, size, size, IOMMU_READ))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova_start + size + 42)
+ != (size + 42))
+ return __FAIL(ops);
+ i++;
+ }
+
+ /* Full unmap */
+ iova = 0;
+ i = find_first_bit(&cfg.pgsize_bitmap, BITS_PER_LONG);
+ while (i != BITS_PER_LONG) {
+ size = 1UL << i;
+
+ if (ops->unmap(ops, iova, size) != size)
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova + 42))
+ return __FAIL(ops);
+
+ /* Remap full block */
+ if (ops->map(ops, iova, iova, size, IOMMU_WRITE))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova + 42) != (iova + 42))
+ return __FAIL(ops);
+
+ iova += SZ_16M;
+ i++;
+ i = find_next_bit(&cfg.pgsize_bitmap, BITS_PER_LONG, i);
+ }
+
+ free_io_pgtable_ops(ops);
+
+ selftest_running = false;
+
+ pr_info("self test ok\n");
+ return 0;
+}
+subsys_initcall(arm_v7s_do_selftests);
+#endif
diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
index 6f2e319..8c615b7 100644
--- a/drivers/iommu/io-pgtable.c
+++ b/drivers/iommu/io-pgtable.c
@@ -33,6 +33,9 @@ io_pgtable_init_table[IO_PGTABLE_NUM_FMTS] =
[ARM_64_LPAE_S1] = &io_pgtable_arm_64_lpae_s1_init_fns,
[ARM_64_LPAE_S2] = &io_pgtable_arm_64_lpae_s2_init_fns,
#endif
+#ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S
+ [ARM_V7S] = &io_pgtable_arm_v7s_init_fns,
+#endif
};
struct io_pgtable_ops *alloc_io_pgtable_ops(enum io_pgtable_fmt fmt,
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 36673c8..aa57073 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -1,5 +1,6 @@
#ifndef __IO_PGTABLE_H
#define __IO_PGTABLE_H
+#include <linux/bitops.h>
/*
* Public API for use by IOMMU drivers
@@ -9,6 +10,7 @@ enum io_pgtable_fmt {
ARM_32_LPAE_S2,
ARM_64_LPAE_S1,
ARM_64_LPAE_S2,
+ ARM_V7S,
IO_PGTABLE_NUM_FMTS,
};
@@ -45,7 +47,9 @@ struct iommu_gather_ops {
* page table walker.
*/
struct io_pgtable_cfg {
- #define IO_PGTABLE_QUIRK_ARM_NS (1 << 0) /* Set NS bit in PTEs */
+ #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
+ #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
+ #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
int quirks;
unsigned long pgsize_bitmap;
unsigned int ias;
@@ -65,6 +69,13 @@ struct io_pgtable_cfg {
u64 vttbr;
u64 vtcr;
} arm_lpae_s2_cfg;
+
+ struct {
+ u32 ttbr[2];
+ u32 tcr;
+ u32 nmrr;
+ u32 prrr;
+ } arm_v7s_cfg;
};
};
@@ -149,5 +160,6 @@ extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns;
extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns;
extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns;
+extern struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns;
#endif /* __IO_PGTABLE_H */
--
2.7.0.25.gfc10eb5
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-01-26 17:13 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: linux-arm-kernel
Add a nearly-complete ARMv7 short descriptor implementation, omitting
only a few legacy and CPU-centric aspects which shouldn't be necessary
for IOMMU API use anyway.
Signed-off-by: Yong Wu <yong.wu@mediatek.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
drivers/iommu/Kconfig | 19 +
drivers/iommu/Makefile | 1 +
drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
drivers/iommu/io-pgtable.c | 3 +
drivers/iommu/io-pgtable.h | 14 +-
5 files changed, 885 insertions(+), 1 deletion(-)
create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index a1e75cb..dc1aaa5 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
If unsure, say N here.
+config IOMMU_IO_PGTABLE_ARMV7S
+ bool "ARMv7/v8 Short Descriptor Format"
+ select IOMMU_IO_PGTABLE
+ depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
+ help
+ Enable support for the ARM Short-descriptor pagetable format.
+ This supports 32-bit virtual and physical addresses mapped using
+ 2-level tables with 4KB pages/1MB sections, and contiguous entries
+ for 64KB pages/16MB supersections if indicated by the IOMMU driver.
+
+config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
+ bool "ARMv7s selftests"
+ depends on IOMMU_IO_PGTABLE_ARMV7S
+ help
+ Enable self-tests for ARMv7s page table allocator. This performs
+ a series of page-table consistency checks during boot.
+
+ If unsure, say N here.
+
endmenu
config IOMMU_IOVA
diff --git a/drivers/iommu/Makefile b/drivers/iommu/Makefile
index 42fc0c2..2f9bfbc 100644
--- a/drivers/iommu/Makefile
+++ b/drivers/iommu/Makefile
@@ -3,6 +3,7 @@ obj-$(CONFIG_IOMMU_API) += iommu-traces.o
obj-$(CONFIG_IOMMU_API) += iommu-sysfs.o
obj-$(CONFIG_IOMMU_DMA) += dma-iommu.o
obj-$(CONFIG_IOMMU_IO_PGTABLE) += io-pgtable.o
+obj-$(CONFIG_IOMMU_IO_PGTABLE_ARMV7S) += io-pgtable-arm-v7s.o
obj-$(CONFIG_IOMMU_IO_PGTABLE_LPAE) += io-pgtable-arm.o
obj-$(CONFIG_IOMMU_IOVA) += iova.o
obj-$(CONFIG_OF_IOMMU) += of_iommu.o
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
new file mode 100644
index 0000000..efc7d1e
--- /dev/null
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -0,0 +1,849 @@
+/*
+ * CPU-agnostic ARM page table allocator.
+ *
+ * ARMv7 Short-descriptor format, supporting
+ * - Basic memory attributes
+ * - Simplified access permissions (AP[2:1] model)
+ * - Backwards-compatible TEX remap
+ * - Large pages/supersections (if indicated by the caller)
+ *
+ * Not supporting:
+ * - Legacy access permissions (AP[2:0] model)
+ *
+ * Almost certainly never supporting:
+ * - PXN
+ * - Domains
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (C) 2014-2015 ARM Limited
+ * Copyright (c) 2014-2015 MediaTek Inc.
+ */
+
+#define pr_fmt(fmt) "arm-v7s io-pgtable: " fmt
+
+#include <linux/dma-mapping.h>
+#include <linux/gfp.h>
+#include <linux/iommu.h>
+#include <linux/kernel.h>
+#include <linux/kmemleak.h>
+#include <linux/sizes.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include <asm/barrier.h>
+
+#include "io-pgtable.h"
+
+/* Struct accessors */
+#define io_pgtable_to_data(x) \
+ container_of((x), struct arm_v7s_io_pgtable, iop)
+
+#define io_pgtable_ops_to_data(x) \
+ io_pgtable_to_data(io_pgtable_ops_to_pgtable(x))
+
+/*
+ * We have 32 bits total; 12 bits resolved at level 1, 8 bits at level 2,
+ * and 12 bits in a page. With some carefully-chosen coefficients we can
+ * hide the ugly inconsistencies behind these macros and at least let the
+ * rest of the code pretend to be somewhat sane.
+ */
+#define ARM_V7S_ADDR_BITS 32
+#define _ARM_V7S_LVL_BITS(lvl) (16 - (lvl) * 4)
+#define ARM_V7S_LVL_SHIFT(lvl) (ARM_V7S_ADDR_BITS - (4 + 8 * (lvl)))
+#define ARM_V7S_TABLE_SHIFT 10
+
+#define ARM_V7S_PTES_PER_LVL(lvl) (1 << _ARM_V7S_LVL_BITS(lvl))
+#define ARM_V7S_TABLE_SIZE(lvl) \
+ (ARM_V7S_PTES_PER_LVL(lvl) * sizeof(arm_v7s_iopte))
+
+#define ARM_V7S_BLOCK_SIZE(lvl) (1UL << ARM_V7S_LVL_SHIFT(lvl))
+#define ARM_V7S_LVL_MASK(lvl) ((u32)(~0U << ARM_V7S_LVL_SHIFT(lvl)))
+#define ARM_V7S_TABLE_MASK ((u32)(~0U << ARM_V7S_TABLE_SHIFT))
+#define _ARM_V7S_IDX_MASK(lvl) (ARM_V7S_PTES_PER_LVL(lvl) - 1)
+#define ARM_V7S_LVL_IDX(addr, lvl) ({ \
+ int _l = lvl; \
+ ((u32)(addr) >> ARM_V7S_LVL_SHIFT(_l)) & _ARM_V7S_IDX_MASK(_l); \
+})
+
+/*
+ * Large page/supersection entries are effectively a block of 16 page/section
+ * entries, along the lines of the LPAE contiguous hint, but all with the
+ * same output address. For want of a better common name we'll call them
+ * "contiguous" versions of their respective page/section entries here, but
+ * noting the distinction (WRT to TLB maintenance) that they represent *one*
+ * entry repeated 16 times, not 16 separate entries (as in the LPAE case).
+ */
+#define ARM_V7S_CONT_PAGES 16
+
+/* PTE type bits: these are all mixed up with XN/PXN bits in most cases */
+#define ARM_V7S_PTE_TYPE_TABLE 0x1
+#define ARM_V7S_PTE_TYPE_PAGE 0x2
+#define ARM_V7S_PTE_TYPE_CONT_PAGE 0x1
+
+#define ARM_V7S_PTE_IS_VALID(pte) (((pte) & 0x3) != 0)
+#define ARM_V7S_PTE_IS_TABLE(pte, lvl) (lvl == 1 && ((pte) & ARM_V7S_PTE_TYPE_TABLE))
+
+/* Page table bits */
+#define ARM_V7S_ATTR_XN(lvl) BIT(4 * (2 - (lvl)))
+#define ARM_V7S_ATTR_B BIT(2)
+#define ARM_V7S_ATTR_C BIT(3)
+#define ARM_V7S_ATTR_NS_TABLE BIT(3)
+#define ARM_V7S_ATTR_NS_SECTION BIT(19)
+
+#define ARM_V7S_CONT_SECTION BIT(18)
+#define ARM_V7S_CONT_PAGE_XN_SHIFT 15
+
+/*
+ * The attribute bits are consistently ordered*, but occupy bits [17:10] of
+ * a level 1 PTE vs. bits [11:4] at level 2. Thus we define the individual
+ * fields relative to that 8-bit block, plus a total shift relative to the PTE.
+ */
+#define ARM_V7S_ATTR_SHIFT(lvl) (16 - (lvl) * 6)
+
+#define ARM_V7S_ATTR_MASK 0xff
+#define ARM_V7S_ATTR_AP0 BIT(0)
+#define ARM_V7S_ATTR_AP1 BIT(1)
+#define ARM_V7S_ATTR_AP2 BIT(5)
+#define ARM_V7S_ATTR_S BIT(6)
+#define ARM_V7S_ATTR_NG BIT(7)
+#define ARM_V7S_TEX_SHIFT 2
+#define ARM_V7S_TEX_MASK 0x7
+#define ARM_V7S_ATTR_TEX(val) (((val) & ARM_V7S_TEX_MASK) << ARM_V7S_TEX_SHIFT)
+
+/* *well, except for TEX on level 2 large pages, of course :( */
+#define ARM_V7S_CONT_PAGE_TEX_SHIFT 6
+#define ARM_V7S_CONT_PAGE_TEX_MASK (ARM_V7S_TEX_MASK << ARM_V7S_CONT_PAGE_TEX_SHIFT)
+
+/* Simplified access permissions */
+#define ARM_V7S_PTE_AF ARM_V7S_ATTR_AP0
+#define ARM_V7S_PTE_AP_UNPRIV ARM_V7S_ATTR_AP1
+#define ARM_V7S_PTE_AP_RDONLY ARM_V7S_ATTR_AP2
+
+/* Register bits */
+#define ARM_V7S_RGN_NC 0
+#define ARM_V7S_RGN_WBWA 1
+#define ARM_V7S_RGN_WT 2
+#define ARM_V7S_RGN_WB 3
+
+#define ARM_V7S_PRRR_TYPE_DEVICE 1
+#define ARM_V7S_PRRR_TYPE_NORMAL 2
+#define ARM_V7S_PRRR_TR(n, type) (((type) & 0x3) << ((n) * 2))
+#define ARM_V7S_PRRR_DS0 BIT(16)
+#define ARM_V7S_PRRR_DS1 BIT(17)
+#define ARM_V7S_PRRR_NS0 BIT(18)
+#define ARM_V7S_PRRR_NS1 BIT(19)
+#define ARM_V7S_PRRR_NOS(n) BIT((n) + 24)
+
+#define ARM_V7S_NMRR_IR(n, attr) (((attr) & 0x3) << ((n) * 2))
+#define ARM_V7S_NMRR_OR(n, attr) (((attr) & 0x3) << ((n) * 2 + 16))
+
+#define ARM_V7S_TTBR_S BIT(1)
+#define ARM_V7S_TTBR_NOS BIT(5)
+#define ARM_V7S_TTBR_ORGN_ATTR(attr) (((attr) & 0x3) << 3)
+#define ARM_V7S_TTBR_IRGN_ATTR(attr) \
+ ((((attr) & 0x1) << 6) | (((attr) & 0x2) >> 1))
+
+#define ARM_V7S_TCR_PD1 BIT(5)
+
+typedef u32 arm_v7s_iopte;
+
+static bool selftest_running;
+
+struct arm_v7s_io_pgtable {
+ struct io_pgtable iop;
+
+ arm_v7s_iopte *pgd;
+ struct kmem_cache *l2_tables;
+};
+
+static dma_addr_t __arm_v7s_dma_addr(void *pages)
+{
+ return (dma_addr_t)virt_to_phys(pages);
+}
+
+static arm_v7s_iopte *iopte_deref(arm_v7s_iopte pte, int lvl)
+{
+ if (ARM_V7S_PTE_IS_TABLE(pte, lvl))
+ pte &= ARM_V7S_TABLE_MASK;
+ else
+ pte &= ARM_V7S_LVL_MASK(lvl);
+ return phys_to_virt(pte);
+}
+
+static void *__arm_v7s_alloc_table(int lvl, gfp_t gfp,
+ struct arm_v7s_io_pgtable *data)
+{
+ struct device *dev = data->iop.cfg.iommu_dev;
+ dma_addr_t dma;
+ size_t size = ARM_V7S_TABLE_SIZE(lvl);
+ void *table = NULL;
+
+ if (lvl == 1)
+ table = (void *)__get_dma_pages(__GFP_ZERO, get_order(size));
+ else if (lvl == 2)
+ table = kmem_cache_zalloc(data->l2_tables, gfp);
+ if (table && !selftest_running) {
+ dma = dma_map_single(dev, table, size, DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, dma))
+ goto out_free;
+ /*
+ * We depend on the IOMMU being able to work with any physical
+ * address directly, so if the DMA layer suggests otherwise by
+ * translating or truncating them, that bodes very badly...
+ */
+ if (dma != virt_to_phys(table))
+ goto out_unmap;
+ }
+ kmemleak_ignore(table);
+ return table;
+
+out_unmap:
+ dev_err(dev, "Cannot accommodate DMA translation for IOMMU page tables\n");
+ dma_unmap_single(dev, dma, size, DMA_TO_DEVICE);
+out_free:
+ if (lvl == 1)
+ free_pages((unsigned long)table, get_order(size));
+ else
+ kmem_cache_free(data->l2_tables, table);
+ return NULL;
+}
+
+static void __arm_v7s_free_table(void *table, int lvl,
+ struct arm_v7s_io_pgtable *data)
+{
+ struct device *dev = data->iop.cfg.iommu_dev;
+ size_t size = ARM_V7S_TABLE_SIZE(lvl);
+
+ if (!selftest_running)
+ dma_unmap_single(dev, __arm_v7s_dma_addr(table), size,
+ DMA_TO_DEVICE);
+ if (lvl == 1)
+ free_pages((unsigned long)table, get_order(size));
+ else
+ kmem_cache_free(data->l2_tables, table);
+}
+
+static void __arm_v7s_pte_sync(arm_v7s_iopte *ptep, int num_entries,
+ struct io_pgtable_cfg *cfg)
+{
+ if (selftest_running)
+ return;
+
+ dma_sync_single_for_device(cfg->iommu_dev, __arm_v7s_dma_addr(ptep),
+ num_entries * sizeof(*ptep), DMA_TO_DEVICE);
+}
+static void __arm_v7s_set_pte(arm_v7s_iopte *ptep, arm_v7s_iopte pte,
+ int num_entries, struct io_pgtable_cfg *cfg)
+{
+ int i;
+
+ for (i = 0; i < num_entries; i++)
+ ptep[i] = pte;
+
+ __arm_v7s_pte_sync(ptep, num_entries, cfg);
+}
+
+static arm_v7s_iopte arm_v7s_prot_to_pte(int prot, int lvl,
+ struct io_pgtable_cfg *cfg)
+{
+ bool ap = !(cfg->quirks & IO_PGTABLE_QUIRK_NO_PERMS);
+ arm_v7s_iopte pte = ARM_V7S_ATTR_NG | ARM_V7S_ATTR_S |
+ ARM_V7S_ATTR_TEX(1);
+
+ if (ap) {
+ pte |= ARM_V7S_PTE_AF | ARM_V7S_PTE_AP_UNPRIV;
+ if (!(prot & IOMMU_WRITE))
+ pte |= ARM_V7S_PTE_AP_RDONLY;
+ }
+ pte <<= ARM_V7S_ATTR_SHIFT(lvl);
+
+ if ((prot & IOMMU_NOEXEC) && ap)
+ pte |= ARM_V7S_ATTR_XN(lvl);
+ if (prot & IOMMU_CACHE)
+ pte |= ARM_V7S_ATTR_B | ARM_V7S_ATTR_C;
+
+ return pte;
+}
+
+static int arm_v7s_pte_to_prot(arm_v7s_iopte pte, int lvl)
+{
+ int prot = IOMMU_READ;
+
+ if (pte & (ARM_V7S_PTE_AP_RDONLY << ARM_V7S_ATTR_SHIFT(lvl)))
+ prot |= IOMMU_WRITE;
+ if (pte & ARM_V7S_ATTR_C)
+ prot |= IOMMU_CACHE;
+
+ return prot;
+}
+
+static arm_v7s_iopte arm_v7s_pte_to_cont(arm_v7s_iopte pte, int lvl)
+{
+ if (lvl == 1) {
+ pte |= ARM_V7S_CONT_SECTION;
+ } else if (lvl == 2) {
+ arm_v7s_iopte xn = pte & ARM_V7S_ATTR_XN(lvl);
+ arm_v7s_iopte tex = pte & ARM_V7S_CONT_PAGE_TEX_MASK;
+
+ pte ^= xn | tex | ARM_V7S_PTE_TYPE_PAGE;
+ pte |= (xn << ARM_V7S_CONT_PAGE_XN_SHIFT) |
+ (tex << ARM_V7S_CONT_PAGE_TEX_SHIFT) |
+ ARM_V7S_PTE_TYPE_CONT_PAGE;
+ }
+ return pte;
+}
+
+static arm_v7s_iopte arm_v7s_cont_to_pte(arm_v7s_iopte pte, int lvl)
+{
+ if (lvl == 1) {
+ pte &= ~ARM_V7S_CONT_SECTION;
+ } else if (lvl == 2) {
+ arm_v7s_iopte xn = pte & BIT(ARM_V7S_CONT_PAGE_XN_SHIFT);
+ arm_v7s_iopte tex = pte & (ARM_V7S_CONT_PAGE_TEX_MASK <<
+ ARM_V7S_CONT_PAGE_TEX_SHIFT);
+
+ pte ^= xn | tex | ARM_V7S_PTE_TYPE_CONT_PAGE;
+ pte |= (xn >> ARM_V7S_CONT_PAGE_XN_SHIFT) |
+ (tex >> ARM_V7S_CONT_PAGE_TEX_SHIFT) |
+ ARM_V7S_PTE_TYPE_PAGE;
+ }
+ return pte;
+}
+
+static bool arm_v7s_pte_is_cont(arm_v7s_iopte pte, int lvl)
+{
+ if (lvl == 1 && !ARM_V7S_PTE_IS_TABLE(pte, lvl))
+ return pte & ARM_V7S_CONT_SECTION;
+ else if (lvl == 2)
+ return !(pte & ARM_V7S_PTE_TYPE_PAGE);
+ return false;
+}
+
+static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *, unsigned long,
+ size_t, int, arm_v7s_iopte *);
+
+static int arm_v7s_init_pte(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, phys_addr_t paddr, int prot,
+ int lvl, int num_entries, arm_v7s_iopte *ptep)
+{
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ arm_v7s_iopte pte = arm_v7s_prot_to_pte(prot, lvl, cfg);
+ int i;
+
+ for (i = 0; i < num_entries; i++)
+ if (ARM_V7S_PTE_IS_TABLE(ptep[i], lvl)) {
+ /*
+ * We need to unmap and free the old table before
+ * overwriting it with a block entry.
+ */
+ arm_v7s_iopte *tblp;
+ size_t sz = ARM_V7S_BLOCK_SIZE(lvl);
+
+ tblp = ptep - ARM_V7S_LVL_IDX(iova, lvl);
+ if (WARN_ON(__arm_v7s_unmap(data, iova + i * sz,
+ sz, lvl, tblp) != sz))
+ return -EINVAL;
+ } else if (ptep[i]) {
+ /* We require an unmap first */
+ WARN_ON(!selftest_running);
+ return -EEXIST;
+ }
+
+ pte |= ARM_V7S_PTE_TYPE_PAGE;
+ if (lvl == 1 && (cfg->quirks & IO_PGTABLE_QUIRK_ARM_NS))
+ pte |= ARM_V7S_ATTR_NS_SECTION;
+
+ if (num_entries > 1)
+ pte = arm_v7s_pte_to_cont(pte, lvl);
+
+ pte |= paddr & ARM_V7S_LVL_MASK(lvl);
+
+ __arm_v7s_set_pte(ptep, pte, num_entries, cfg);
+ return 0;
+}
+
+static int __arm_v7s_map(struct arm_v7s_io_pgtable *data, unsigned long iova,
+ phys_addr_t paddr, size_t size, int prot,
+ int lvl, arm_v7s_iopte *ptep)
+{
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ arm_v7s_iopte pte, *cptep;
+ int num_entries = size >> ARM_V7S_LVL_SHIFT(lvl);
+
+ /* Find our entry at the current level */
+ ptep += ARM_V7S_LVL_IDX(iova, lvl);
+
+ /* If we can install a leaf entry at this level, then do so */
+ if (num_entries)
+ return arm_v7s_init_pte(data, iova, paddr, prot,
+ lvl, num_entries, ptep);
+
+ /* We can't allocate tables at the final level */
+ if (WARN_ON(lvl == 2))
+ return -EINVAL;
+
+ /* Grab a pointer to the next level */
+ pte = *ptep;
+ if (!pte) {
+ cptep = __arm_v7s_alloc_table(lvl + 1, GFP_ATOMIC, data);
+ if (!cptep)
+ return -ENOMEM;
+
+ pte = virt_to_phys(cptep) | ARM_V7S_PTE_TYPE_TABLE;
+ if (cfg->quirks & IO_PGTABLE_QUIRK_ARM_NS)
+ pte |= ARM_V7S_ATTR_NS_TABLE;
+
+ __arm_v7s_set_pte(ptep, pte, 1, cfg);
+ } else {
+ cptep = iopte_deref(pte, lvl);
+ }
+
+ /* Rinse, repeat */
+ return __arm_v7s_map(data, iova, paddr, size, prot, lvl + 1, cptep);
+}
+
+static int arm_v7s_map(struct io_pgtable_ops *ops, unsigned long iova,
+ phys_addr_t paddr, size_t size, int prot)
+{
+ struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ const struct iommu_gather_ops *tlb = cfg->tlb;
+ void *cookie = data->iop.cookie;
+ int ret;
+
+ /* If no access, then nothing to do */
+ if (!(prot & (IOMMU_READ | IOMMU_WRITE)))
+ return 0;
+
+ ret = __arm_v7s_map(data, iova, paddr, size, prot, 1, data->pgd);
+ /*
+ * Synchronise all PTE updates for the new mapping before there's
+ * a chance for anything to kick off a table walk for the new iova.
+ */
+ if (cfg->quirks & IO_PGTABLE_QUIRK_TLBI_ON_MAP) {
+ tlb->tlb_add_flush(iova, size, ARM_V7S_BLOCK_SIZE(2), false,
+ cookie);
+ tlb->tlb_sync(cookie);
+ } else {
+ wmb();
+ }
+
+ return ret;
+}
+
+static void arm_v7s_free_pgtable(struct io_pgtable *iop)
+{
+ struct arm_v7s_io_pgtable *data = io_pgtable_to_data(iop);
+ int i;
+
+ for (i = 0; i < ARM_V7S_PTES_PER_LVL(1); i++) {
+ arm_v7s_iopte pte = data->pgd[i];
+
+ if (ARM_V7S_PTE_IS_TABLE(pte, 1))
+ __arm_v7s_free_table(iopte_deref(pte, 1), 2, data);
+ }
+ __arm_v7s_free_table(data->pgd, 1, data);
+ kmem_cache_destroy(data->l2_tables);
+ kfree(data);
+}
+
+static void arm_v7s_split_cont(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, int idx, int lvl,
+ arm_v7s_iopte *ptep)
+{
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ void *cookie = data->iop.cookie;
+ arm_v7s_iopte pte;
+ size_t size = ARM_V7S_BLOCK_SIZE(lvl);
+ int i;
+
+ ptep -= idx & (ARM_V7S_CONT_PAGES - 1);
+ pte = arm_v7s_cont_to_pte(*ptep, lvl);
+ for (i = 0; i < ARM_V7S_CONT_PAGES; i++) {
+ ptep[i] = pte;
+ pte += size;
+ }
+
+ __arm_v7s_pte_sync(ptep, ARM_V7S_CONT_PAGES, cfg);
+
+ size *= ARM_V7S_CONT_PAGES;
+ cfg->tlb->tlb_add_flush(iova, size, size, true, cookie);
+ cfg->tlb->tlb_sync(cookie);
+}
+
+static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, size_t size,
+ arm_v7s_iopte *ptep)
+{
+ unsigned long blk_start, blk_end, blk_size;
+ phys_addr_t blk_paddr;
+ arm_v7s_iopte table = 0;
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ int prot = arm_v7s_pte_to_prot(*ptep, 1);
+
+ blk_size = ARM_V7S_BLOCK_SIZE(1);
+ blk_start = iova & ARM_V7S_LVL_MASK(1);
+ blk_end = blk_start + ARM_V7S_BLOCK_SIZE(1);
+ blk_paddr = *ptep & ARM_V7S_LVL_MASK(1);
+
+ for (; blk_start < blk_end; blk_start += size, blk_paddr += size) {
+ arm_v7s_iopte *tablep;
+
+ /* Unmap! */
+ if (blk_start == iova)
+ continue;
+
+ /* __arm_v7s_map expects a pointer to the start of the table */
+ tablep = &table - ARM_V7S_LVL_IDX(blk_start, 1);
+ if (__arm_v7s_map(data, blk_start, blk_paddr, size, prot, 1,
+ tablep) < 0) {
+ if (table) {
+ /* Free the table we allocated */
+ tablep = iopte_deref(table, 1);
+ __arm_v7s_free_table(tablep, 2, data);
+ }
+ return 0; /* Bytes unmapped */
+ }
+ }
+
+ __arm_v7s_set_pte(ptep, table, 1, cfg);
+ iova &= ~(blk_size - 1);
+ cfg->tlb->tlb_add_flush(iova, blk_size, blk_size, true, data->iop.cookie);
+ return size;
+}
+
+static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
+ unsigned long iova, size_t size, int lvl,
+ arm_v7s_iopte *ptep)
+{
+ arm_v7s_iopte pte[ARM_V7S_CONT_PAGES];
+ struct io_pgtable_cfg *cfg = &data->iop.cfg;
+ const struct iommu_gather_ops *tlb = cfg->tlb;
+ void *cookie = data->iop.cookie;
+ int idx, i = 0, num_entries = size >> ARM_V7S_LVL_SHIFT(lvl);
+
+ /* Something went horribly wrong and we ran out of page table */
+ if (WARN_ON(lvl > 2))
+ return 0;
+
+ idx = ARM_V7S_LVL_IDX(iova, lvl);
+ ptep += idx;
+ do {
+ if (WARN_ON(!ARM_V7S_PTE_IS_VALID(ptep[i])))
+ return 0;
+ pte[i] = ptep[i];
+ } while (++i < num_entries);
+
+ /*
+ * If we've hit a contiguous 'large page' entry@this level, it
+ * needs splitting first, unless we're unmapping the whole lot.
+ */
+ if (num_entries <= 1 && arm_v7s_pte_is_cont(pte[0], lvl))
+ arm_v7s_split_cont(data, iova, idx, lvl, ptep);
+
+ /* If the size matches this level, we're in the right place */
+ if (num_entries) {
+ size_t blk_size = ARM_V7S_BLOCK_SIZE(lvl);
+
+ __arm_v7s_set_pte(ptep, 0, num_entries, cfg);
+
+ for (i = 0; i < num_entries; i++) {
+ if (ARM_V7S_PTE_IS_TABLE(pte[i], lvl)) {
+ /* Also flush any partial walks */
+ tlb->tlb_add_flush(iova, blk_size,
+ ARM_V7S_BLOCK_SIZE(lvl + 1),
+ false, cookie);
+ tlb->tlb_sync(cookie);
+ ptep = iopte_deref(pte[i], lvl);
+ __arm_v7s_free_table(ptep, lvl + 1, data);
+ } else {
+ tlb->tlb_add_flush(iova, blk_size, blk_size,
+ true, cookie);
+ }
+ iova += blk_size;
+ }
+ return size;
+ } else if (lvl == 1 && !ARM_V7S_PTE_IS_TABLE(pte[0], lvl)) {
+ /*
+ * Insert a table at the next level to map the old region,
+ * minus the part we want to unmap
+ */
+ return arm_v7s_split_blk_unmap(data, iova, size, ptep);
+ }
+
+ /* Keep on walkin' */
+ ptep = iopte_deref(pte[0], lvl);
+ return __arm_v7s_unmap(data, iova, size, lvl + 1, ptep);
+}
+
+static int arm_v7s_unmap(struct io_pgtable_ops *ops, unsigned long iova,
+ size_t size)
+{
+ size_t unmapped;
+ struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
+ struct io_pgtable *iop = &data->iop;
+
+ unmapped = __arm_v7s_unmap(data, iova, size, 1, data->pgd);
+ if (unmapped)
+ iop->cfg.tlb->tlb_sync(iop->cookie);
+
+ return unmapped;
+}
+
+static phys_addr_t arm_v7s_iova_to_phys(struct io_pgtable_ops *ops,
+ unsigned long iova)
+{
+ struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
+ arm_v7s_iopte *ptep = data->pgd, pte;
+ int lvl = 0;
+ u32 mask;
+
+ do {
+ pte = ptep[ARM_V7S_LVL_IDX(iova, ++lvl)];
+ ptep = iopte_deref(pte, lvl);
+ } while (ARM_V7S_PTE_IS_TABLE(pte, lvl));
+
+ if (!ARM_V7S_PTE_IS_VALID(pte))
+ return 0;
+
+ mask = ARM_V7S_LVL_MASK(lvl);
+ if (arm_v7s_pte_is_cont(pte, lvl))
+ mask *= ARM_V7S_CONT_PAGES;
+ return (pte & mask) | (iova & ~mask);
+}
+
+static struct io_pgtable *arm_v7s_alloc_pgtable(struct io_pgtable_cfg *cfg,
+ void *cookie)
+{
+ struct arm_v7s_io_pgtable *data;
+
+ if (cfg->ias > ARM_V7S_ADDR_BITS || cfg->oas > ARM_V7S_ADDR_BITS)
+ return NULL;
+
+ data = kmalloc(sizeof(*data), GFP_KERNEL);
+ if (!data)
+ return NULL;
+
+ data->l2_tables = kmem_cache_create("io-pgtable_armv7s_l2",
+ ARM_V7S_TABLE_SIZE(2),
+ ARM_V7S_TABLE_SIZE(2),
+ SLAB_CACHE_DMA, NULL);
+ if (!data->l2_tables)
+ goto out_free_data;
+
+ data->iop.ops = (struct io_pgtable_ops) {
+ .map = arm_v7s_map,
+ .unmap = arm_v7s_unmap,
+ .iova_to_phys = arm_v7s_iova_to_phys,
+ };
+
+ /* We have to do this early for __arm_v7s_alloc_table to work... */
+ data->iop.cfg = *cfg;
+
+ /*
+ * Unless the IOMMU driver indicates supersection support by
+ * having SZ_16M set in the initial bitmap, they won't be used.
+ */
+ cfg->pgsize_bitmap &= SZ_4K | SZ_64K | SZ_1M | SZ_16M;
+
+ /* TCR: T0SZ=0, disable TTBR1 */
+ cfg->arm_v7s_cfg.tcr = ARM_V7S_TCR_PD1;
+
+ /*
+ * TEX remap: the indices used map to the closest equivalent types
+ * under the non-TEX-remap interpretation of those attribute bits,
+ * excepting various implementation-defined aspects of shareability.
+ */
+ cfg->arm_v7s_cfg.prrr = ARM_V7S_PRRR_TR(1, ARM_V7S_PRRR_TYPE_DEVICE) |
+ ARM_V7S_PRRR_TR(4, ARM_V7S_PRRR_TYPE_NORMAL) |
+ ARM_V7S_PRRR_TR(7, ARM_V7S_PRRR_TYPE_NORMAL) |
+ ARM_V7S_PRRR_DS0 | ARM_V7S_PRRR_DS1 |
+ ARM_V7S_PRRR_NS1 | ARM_V7S_PRRR_NOS(7);
+ cfg->arm_v7s_cfg.nmrr = ARM_V7S_NMRR_IR(7, ARM_V7S_RGN_WBWA) |
+ ARM_V7S_NMRR_OR(7, ARM_V7S_RGN_WBWA);
+
+ /* Looking good; allocate a pgd */
+ data->pgd = __arm_v7s_alloc_table(1, GFP_KERNEL, data);
+ if (!data->pgd)
+ goto out_free_data;
+
+ /* Ensure the empty pgd is visible before any actual TTBR write */
+ wmb();
+
+ /* TTBRs */
+ cfg->arm_v7s_cfg.ttbr[0] = virt_to_phys(data->pgd) |
+ ARM_V7S_TTBR_S | ARM_V7S_TTBR_NOS |
+ ARM_V7S_TTBR_IRGN_ATTR(ARM_V7S_RGN_WBWA) |
+ ARM_V7S_TTBR_ORGN_ATTR(ARM_V7S_RGN_WBWA);
+ cfg->arm_v7s_cfg.ttbr[1] = 0;
+ return &data->iop;
+
+out_free_data:
+ kmem_cache_destroy(data->l2_tables);
+ kfree(data);
+ return NULL;
+}
+
+struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
+ .alloc = arm_v7s_alloc_pgtable,
+ .free = arm_v7s_free_pgtable,
+};
+
+#ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
+
+static struct io_pgtable_cfg *cfg_cookie;
+
+static void dummy_tlb_flush_all(void *cookie)
+{
+ WARN_ON(cookie != cfg_cookie);
+}
+
+static void dummy_tlb_add_flush(unsigned long iova, size_t size,
+ size_t granule, bool leaf, void *cookie)
+{
+ WARN_ON(cookie != cfg_cookie);
+ WARN_ON(!(size & cfg_cookie->pgsize_bitmap));
+}
+
+static void dummy_tlb_sync(void *cookie)
+{
+ WARN_ON(cookie != cfg_cookie);
+}
+
+static struct iommu_gather_ops dummy_tlb_ops = {
+ .tlb_flush_all = dummy_tlb_flush_all,
+ .tlb_add_flush = dummy_tlb_add_flush,
+ .tlb_sync = dummy_tlb_sync,
+};
+
+#define __FAIL(ops) ({ \
+ WARN(1, "selftest: test failed\n"); \
+ selftest_running = false; \
+ -EFAULT; \
+})
+
+static int __init arm_v7s_do_selftests(void)
+{
+ struct io_pgtable_ops *ops;
+ struct io_pgtable_cfg cfg = {
+ .tlb = &dummy_tlb_ops,
+ .oas = 32,
+ .ias = 32,
+ .quirks = IO_PGTABLE_QUIRK_ARM_NS,
+ .pgsize_bitmap = SZ_4K | SZ_64K | SZ_1M | SZ_16M,
+ };
+ unsigned int iova, size, iova_start;
+ unsigned int i, loopnr = 0;
+
+ selftest_running = true;
+
+ cfg_cookie = &cfg;
+
+ ops = alloc_io_pgtable_ops(ARM_V7S, &cfg, &cfg);
+ if (!ops) {
+ pr_err("selftest: failed to allocate io pgtable ops\n");
+ return -EINVAL;
+ }
+
+ /*
+ * Initial sanity checks.
+ * Empty page tables shouldn't provide any translations.
+ */
+ if (ops->iova_to_phys(ops, 42))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, SZ_1G + 42))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, SZ_2G + 42))
+ return __FAIL(ops);
+
+ /*
+ * Distinct mappings of different granule sizes.
+ */
+ iova = 0;
+ i = find_first_bit(&cfg.pgsize_bitmap, BITS_PER_LONG);
+ while (i != BITS_PER_LONG) {
+ size = 1UL << i;
+ if (ops->map(ops, iova, iova, size, IOMMU_READ |
+ IOMMU_WRITE |
+ IOMMU_NOEXEC |
+ IOMMU_CACHE))
+ return __FAIL(ops);
+
+ /* Overlapping mappings */
+ if (!ops->map(ops, iova, iova + size, size,
+ IOMMU_READ | IOMMU_NOEXEC))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova + 42) != (iova + 42))
+ return __FAIL(ops);
+
+ iova += SZ_16M;
+ i++;
+ i = find_next_bit(&cfg.pgsize_bitmap, BITS_PER_LONG, i);
+ loopnr++;
+ }
+
+ /* Partial unmap */
+ i = 1;
+ size = 1UL << __ffs(cfg.pgsize_bitmap);
+ while (i < loopnr) {
+ iova_start = i * SZ_16M;
+ if (ops->unmap(ops, iova_start + size, size) != size)
+ return __FAIL(ops);
+
+ /* Remap of partial unmap */
+ if (ops->map(ops, iova_start + size, size, size, IOMMU_READ))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova_start + size + 42)
+ != (size + 42))
+ return __FAIL(ops);
+ i++;
+ }
+
+ /* Full unmap */
+ iova = 0;
+ i = find_first_bit(&cfg.pgsize_bitmap, BITS_PER_LONG);
+ while (i != BITS_PER_LONG) {
+ size = 1UL << i;
+
+ if (ops->unmap(ops, iova, size) != size)
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova + 42))
+ return __FAIL(ops);
+
+ /* Remap full block */
+ if (ops->map(ops, iova, iova, size, IOMMU_WRITE))
+ return __FAIL(ops);
+
+ if (ops->iova_to_phys(ops, iova + 42) != (iova + 42))
+ return __FAIL(ops);
+
+ iova += SZ_16M;
+ i++;
+ i = find_next_bit(&cfg.pgsize_bitmap, BITS_PER_LONG, i);
+ }
+
+ free_io_pgtable_ops(ops);
+
+ selftest_running = false;
+
+ pr_info("self test ok\n");
+ return 0;
+}
+subsys_initcall(arm_v7s_do_selftests);
+#endif
diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
index 6f2e319..8c615b7 100644
--- a/drivers/iommu/io-pgtable.c
+++ b/drivers/iommu/io-pgtable.c
@@ -33,6 +33,9 @@ io_pgtable_init_table[IO_PGTABLE_NUM_FMTS] =
[ARM_64_LPAE_S1] = &io_pgtable_arm_64_lpae_s1_init_fns,
[ARM_64_LPAE_S2] = &io_pgtable_arm_64_lpae_s2_init_fns,
#endif
+#ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S
+ [ARM_V7S] = &io_pgtable_arm_v7s_init_fns,
+#endif
};
struct io_pgtable_ops *alloc_io_pgtable_ops(enum io_pgtable_fmt fmt,
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 36673c8..aa57073 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -1,5 +1,6 @@
#ifndef __IO_PGTABLE_H
#define __IO_PGTABLE_H
+#include <linux/bitops.h>
/*
* Public API for use by IOMMU drivers
@@ -9,6 +10,7 @@ enum io_pgtable_fmt {
ARM_32_LPAE_S2,
ARM_64_LPAE_S1,
ARM_64_LPAE_S2,
+ ARM_V7S,
IO_PGTABLE_NUM_FMTS,
};
@@ -45,7 +47,9 @@ struct iommu_gather_ops {
* page table walker.
*/
struct io_pgtable_cfg {
- #define IO_PGTABLE_QUIRK_ARM_NS (1 << 0) /* Set NS bit in PTEs */
+ #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
+ #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
+ #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
int quirks;
unsigned long pgsize_bitmap;
unsigned int ias;
@@ -65,6 +69,13 @@ struct io_pgtable_cfg {
u64 vttbr;
u64 vtcr;
} arm_lpae_s2_cfg;
+
+ struct {
+ u32 ttbr[2];
+ u32 tcr;
+ u32 nmrr;
+ u32 prrr;
+ } arm_v7s_cfg;
};
};
@@ -149,5 +160,6 @@ extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns;
extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns;
extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns;
+extern struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns;
#endif /* __IO_PGTABLE_H */
--
2.7.0.25.gfc10eb5
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH v3 2/3] iommu/io-pgtable: Add helper functions for TLB ops
2016-01-26 17:13 ` Robin Murphy
@ 2016-01-26 17:13 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: will.deacon-5wv7dgnIgG8, yong.wu-NuS5LvNUpcJWk0Htik3J/w,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Cc: laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
Add some simple wrappers to avoid having the guts of the TLB operations
spilled all over the page table implementations, and to provide a point
to implement extra common functionality.
Acked-by: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
Acked-by: Laurent Pinchart <laurent.pinchart-ryLnwIuWjnjg/C1BVhZhaw@public.gmane.org>
Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
---
drivers/iommu/io-pgtable-arm-v7s.c | 48 ++++++++++++++++----------------------
drivers/iommu/io-pgtable-arm.c | 21 +++++++----------
drivers/iommu/io-pgtable.c | 2 +-
drivers/iommu/io-pgtable.h | 16 +++++++++++++
4 files changed, 46 insertions(+), 41 deletions(-)
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
index efc7d1e..d39a021 100644
--- a/drivers/iommu/io-pgtable-arm-v7s.c
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -417,9 +417,7 @@ static int arm_v7s_map(struct io_pgtable_ops *ops, unsigned long iova,
phys_addr_t paddr, size_t size, int prot)
{
struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
- const struct iommu_gather_ops *tlb = cfg->tlb;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
int ret;
/* If no access, then nothing to do */
@@ -431,10 +429,10 @@ static int arm_v7s_map(struct io_pgtable_ops *ops, unsigned long iova,
* Synchronise all PTE updates for the new mapping before there's
* a chance for anything to kick off a table walk for the new iova.
*/
- if (cfg->quirks & IO_PGTABLE_QUIRK_TLBI_ON_MAP) {
- tlb->tlb_add_flush(iova, size, ARM_V7S_BLOCK_SIZE(2), false,
- cookie);
- tlb->tlb_sync(cookie);
+ if (iop->cfg.quirks & IO_PGTABLE_QUIRK_TLBI_ON_MAP) {
+ io_pgtable_tlb_add_flush(iop, iova, size,
+ ARM_V7S_BLOCK_SIZE(2), false);
+ io_pgtable_tlb_sync(iop);
} else {
wmb();
}
@@ -462,8 +460,7 @@ static void arm_v7s_split_cont(struct arm_v7s_io_pgtable *data,
unsigned long iova, int idx, int lvl,
arm_v7s_iopte *ptep)
{
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
arm_v7s_iopte pte;
size_t size = ARM_V7S_BLOCK_SIZE(lvl);
int i;
@@ -475,11 +472,11 @@ static void arm_v7s_split_cont(struct arm_v7s_io_pgtable *data,
pte += size;
}
- __arm_v7s_pte_sync(ptep, ARM_V7S_CONT_PAGES, cfg);
+ __arm_v7s_pte_sync(ptep, ARM_V7S_CONT_PAGES, &iop->cfg);
size *= ARM_V7S_CONT_PAGES;
- cfg->tlb->tlb_add_flush(iova, size, size, true, cookie);
- cfg->tlb->tlb_sync(cookie);
+ io_pgtable_tlb_add_flush(iop, iova, size, size, true);
+ io_pgtable_tlb_sync(iop);
}
static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
@@ -489,7 +486,6 @@ static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
unsigned long blk_start, blk_end, blk_size;
phys_addr_t blk_paddr;
arm_v7s_iopte table = 0;
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
int prot = arm_v7s_pte_to_prot(*ptep, 1);
blk_size = ARM_V7S_BLOCK_SIZE(1);
@@ -517,9 +513,9 @@ static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
}
}
- __arm_v7s_set_pte(ptep, table, 1, cfg);
+ __arm_v7s_set_pte(ptep, table, 1, &data->iop.cfg);
iova &= ~(blk_size - 1);
- cfg->tlb->tlb_add_flush(iova, blk_size, blk_size, true, data->iop.cookie);
+ io_pgtable_tlb_add_flush(&data->iop, iova, blk_size, blk_size, true);
return size;
}
@@ -528,9 +524,7 @@ static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
arm_v7s_iopte *ptep)
{
arm_v7s_iopte pte[ARM_V7S_CONT_PAGES];
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
- const struct iommu_gather_ops *tlb = cfg->tlb;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
int idx, i = 0, num_entries = size >> ARM_V7S_LVL_SHIFT(lvl);
/* Something went horribly wrong and we ran out of page table */
@@ -556,20 +550,19 @@ static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
if (num_entries) {
size_t blk_size = ARM_V7S_BLOCK_SIZE(lvl);
- __arm_v7s_set_pte(ptep, 0, num_entries, cfg);
+ __arm_v7s_set_pte(ptep, 0, num_entries, &iop->cfg);
for (i = 0; i < num_entries; i++) {
if (ARM_V7S_PTE_IS_TABLE(pte[i], lvl)) {
/* Also flush any partial walks */
- tlb->tlb_add_flush(iova, blk_size,
- ARM_V7S_BLOCK_SIZE(lvl + 1),
- false, cookie);
- tlb->tlb_sync(cookie);
+ io_pgtable_tlb_add_flush(iop, iova, blk_size,
+ ARM_V7S_BLOCK_SIZE(lvl + 1), false);
+ io_pgtable_tlb_sync(iop);
ptep = iopte_deref(pte[i], lvl);
__arm_v7s_free_table(ptep, lvl + 1, data);
} else {
- tlb->tlb_add_flush(iova, blk_size, blk_size,
- true, cookie);
+ io_pgtable_tlb_add_flush(iop, iova, blk_size,
+ blk_size, true);
}
iova += blk_size;
}
@@ -590,13 +583,12 @@ static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
static int arm_v7s_unmap(struct io_pgtable_ops *ops, unsigned long iova,
size_t size)
{
- size_t unmapped;
struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
- struct io_pgtable *iop = &data->iop;
+ size_t unmapped;
unmapped = __arm_v7s_unmap(data, iova, size, 1, data->pgd);
if (unmapped)
- iop->cfg.tlb->tlb_sync(iop->cookie);
+ io_pgtable_tlb_sync(&data->iop);
return unmapped;
}
diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
index 8bbcbfe..4095af2 100644
--- a/drivers/iommu/io-pgtable-arm.c
+++ b/drivers/iommu/io-pgtable-arm.c
@@ -445,7 +445,6 @@ static int arm_lpae_split_blk_unmap(struct arm_lpae_io_pgtable *data,
unsigned long blk_start, blk_end;
phys_addr_t blk_paddr;
arm_lpae_iopte table = 0;
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
blk_start = iova & ~(blk_size - 1);
blk_end = blk_start + blk_size;
@@ -471,9 +470,9 @@ static int arm_lpae_split_blk_unmap(struct arm_lpae_io_pgtable *data,
}
}
- __arm_lpae_set_pte(ptep, table, cfg);
+ __arm_lpae_set_pte(ptep, table, &data->iop.cfg);
iova &= ~(blk_size - 1);
- cfg->tlb->tlb_add_flush(iova, blk_size, blk_size, true, data->iop.cookie);
+ io_pgtable_tlb_add_flush(&data->iop, iova, blk_size, blk_size, true);
return size;
}
@@ -482,8 +481,7 @@ static int __arm_lpae_unmap(struct arm_lpae_io_pgtable *data,
arm_lpae_iopte *ptep)
{
arm_lpae_iopte pte;
- const struct iommu_gather_ops *tlb = data->iop.cfg.tlb;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
size_t blk_size = ARM_LPAE_BLOCK_SIZE(lvl, data);
/* Something went horribly wrong and we ran out of page table */
@@ -497,17 +495,17 @@ static int __arm_lpae_unmap(struct arm_lpae_io_pgtable *data,
/* If the size matches this level, we're in the right place */
if (size == blk_size) {
- __arm_lpae_set_pte(ptep, 0, &data->iop.cfg);
+ __arm_lpae_set_pte(ptep, 0, &iop->cfg);
if (!iopte_leaf(pte, lvl)) {
/* Also flush any partial walks */
- tlb->tlb_add_flush(iova, size, ARM_LPAE_GRANULE(data),
- false, cookie);
- tlb->tlb_sync(cookie);
+ io_pgtable_tlb_add_flush(iop, iova, size,
+ ARM_LPAE_GRANULE(data), false);
+ io_pgtable_tlb_sync(iop);
ptep = iopte_deref(pte, data);
__arm_lpae_free_pgtable(data, lvl + 1, ptep);
} else {
- tlb->tlb_add_flush(iova, size, size, true, cookie);
+ io_pgtable_tlb_add_flush(iop, iova, size, size, true);
}
return size;
@@ -531,13 +529,12 @@ static int arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
{
size_t unmapped;
struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
- struct io_pgtable *iop = &data->iop;
arm_lpae_iopte *ptep = data->pgd;
int lvl = ARM_LPAE_START_LVL(data);
unmapped = __arm_lpae_unmap(data, iova, size, lvl, ptep);
if (unmapped)
- iop->cfg.tlb->tlb_sync(iop->cookie);
+ io_pgtable_tlb_sync(&data->iop);
return unmapped;
}
diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
index 8c615b7..876f6a7 100644
--- a/drivers/iommu/io-pgtable.c
+++ b/drivers/iommu/io-pgtable.c
@@ -75,6 +75,6 @@ void free_io_pgtable_ops(struct io_pgtable_ops *ops)
return;
iop = container_of(ops, struct io_pgtable, ops);
- iop->cfg.tlb->tlb_flush_all(iop->cookie);
+ io_pgtable_tlb_flush_all(iop);
io_pgtable_init_table[iop->fmt]->free(iop);
}
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index aa57073..95c5565 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -144,6 +144,22 @@ struct io_pgtable {
#define io_pgtable_ops_to_pgtable(x) container_of((x), struct io_pgtable, ops)
+static inline void io_pgtable_tlb_flush_all(struct io_pgtable *iop)
+{
+ iop->cfg.tlb->tlb_flush_all(iop->cookie);
+}
+
+static inline void io_pgtable_tlb_add_flush(struct io_pgtable *iop,
+ unsigned long iova, size_t size, size_t granule, bool leaf)
+{
+ iop->cfg.tlb->tlb_add_flush(iova, size, granule, leaf, iop->cookie);
+}
+
+static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
+{
+ iop->cfg.tlb->tlb_sync(iop->cookie);
+}
+
/**
* struct io_pgtable_init_fns - Alloc/free a set of page tables for a
* particular format.
--
2.7.0.25.gfc10eb5
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH v3 2/3] iommu/io-pgtable: Add helper functions for TLB ops
@ 2016-01-26 17:13 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: linux-arm-kernel
Add some simple wrappers to avoid having the guts of the TLB operations
spilled all over the page table implementations, and to provide a point
to implement extra common functionality.
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
drivers/iommu/io-pgtable-arm-v7s.c | 48 ++++++++++++++++----------------------
drivers/iommu/io-pgtable-arm.c | 21 +++++++----------
drivers/iommu/io-pgtable.c | 2 +-
drivers/iommu/io-pgtable.h | 16 +++++++++++++
4 files changed, 46 insertions(+), 41 deletions(-)
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
index efc7d1e..d39a021 100644
--- a/drivers/iommu/io-pgtable-arm-v7s.c
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -417,9 +417,7 @@ static int arm_v7s_map(struct io_pgtable_ops *ops, unsigned long iova,
phys_addr_t paddr, size_t size, int prot)
{
struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
- const struct iommu_gather_ops *tlb = cfg->tlb;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
int ret;
/* If no access, then nothing to do */
@@ -431,10 +429,10 @@ static int arm_v7s_map(struct io_pgtable_ops *ops, unsigned long iova,
* Synchronise all PTE updates for the new mapping before there's
* a chance for anything to kick off a table walk for the new iova.
*/
- if (cfg->quirks & IO_PGTABLE_QUIRK_TLBI_ON_MAP) {
- tlb->tlb_add_flush(iova, size, ARM_V7S_BLOCK_SIZE(2), false,
- cookie);
- tlb->tlb_sync(cookie);
+ if (iop->cfg.quirks & IO_PGTABLE_QUIRK_TLBI_ON_MAP) {
+ io_pgtable_tlb_add_flush(iop, iova, size,
+ ARM_V7S_BLOCK_SIZE(2), false);
+ io_pgtable_tlb_sync(iop);
} else {
wmb();
}
@@ -462,8 +460,7 @@ static void arm_v7s_split_cont(struct arm_v7s_io_pgtable *data,
unsigned long iova, int idx, int lvl,
arm_v7s_iopte *ptep)
{
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
arm_v7s_iopte pte;
size_t size = ARM_V7S_BLOCK_SIZE(lvl);
int i;
@@ -475,11 +472,11 @@ static void arm_v7s_split_cont(struct arm_v7s_io_pgtable *data,
pte += size;
}
- __arm_v7s_pte_sync(ptep, ARM_V7S_CONT_PAGES, cfg);
+ __arm_v7s_pte_sync(ptep, ARM_V7S_CONT_PAGES, &iop->cfg);
size *= ARM_V7S_CONT_PAGES;
- cfg->tlb->tlb_add_flush(iova, size, size, true, cookie);
- cfg->tlb->tlb_sync(cookie);
+ io_pgtable_tlb_add_flush(iop, iova, size, size, true);
+ io_pgtable_tlb_sync(iop);
}
static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
@@ -489,7 +486,6 @@ static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
unsigned long blk_start, blk_end, blk_size;
phys_addr_t blk_paddr;
arm_v7s_iopte table = 0;
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
int prot = arm_v7s_pte_to_prot(*ptep, 1);
blk_size = ARM_V7S_BLOCK_SIZE(1);
@@ -517,9 +513,9 @@ static int arm_v7s_split_blk_unmap(struct arm_v7s_io_pgtable *data,
}
}
- __arm_v7s_set_pte(ptep, table, 1, cfg);
+ __arm_v7s_set_pte(ptep, table, 1, &data->iop.cfg);
iova &= ~(blk_size - 1);
- cfg->tlb->tlb_add_flush(iova, blk_size, blk_size, true, data->iop.cookie);
+ io_pgtable_tlb_add_flush(&data->iop, iova, blk_size, blk_size, true);
return size;
}
@@ -528,9 +524,7 @@ static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
arm_v7s_iopte *ptep)
{
arm_v7s_iopte pte[ARM_V7S_CONT_PAGES];
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
- const struct iommu_gather_ops *tlb = cfg->tlb;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
int idx, i = 0, num_entries = size >> ARM_V7S_LVL_SHIFT(lvl);
/* Something went horribly wrong and we ran out of page table */
@@ -556,20 +550,19 @@ static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
if (num_entries) {
size_t blk_size = ARM_V7S_BLOCK_SIZE(lvl);
- __arm_v7s_set_pte(ptep, 0, num_entries, cfg);
+ __arm_v7s_set_pte(ptep, 0, num_entries, &iop->cfg);
for (i = 0; i < num_entries; i++) {
if (ARM_V7S_PTE_IS_TABLE(pte[i], lvl)) {
/* Also flush any partial walks */
- tlb->tlb_add_flush(iova, blk_size,
- ARM_V7S_BLOCK_SIZE(lvl + 1),
- false, cookie);
- tlb->tlb_sync(cookie);
+ io_pgtable_tlb_add_flush(iop, iova, blk_size,
+ ARM_V7S_BLOCK_SIZE(lvl + 1), false);
+ io_pgtable_tlb_sync(iop);
ptep = iopte_deref(pte[i], lvl);
__arm_v7s_free_table(ptep, lvl + 1, data);
} else {
- tlb->tlb_add_flush(iova, blk_size, blk_size,
- true, cookie);
+ io_pgtable_tlb_add_flush(iop, iova, blk_size,
+ blk_size, true);
}
iova += blk_size;
}
@@ -590,13 +583,12 @@ static int __arm_v7s_unmap(struct arm_v7s_io_pgtable *data,
static int arm_v7s_unmap(struct io_pgtable_ops *ops, unsigned long iova,
size_t size)
{
- size_t unmapped;
struct arm_v7s_io_pgtable *data = io_pgtable_ops_to_data(ops);
- struct io_pgtable *iop = &data->iop;
+ size_t unmapped;
unmapped = __arm_v7s_unmap(data, iova, size, 1, data->pgd);
if (unmapped)
- iop->cfg.tlb->tlb_sync(iop->cookie);
+ io_pgtable_tlb_sync(&data->iop);
return unmapped;
}
diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
index 8bbcbfe..4095af2 100644
--- a/drivers/iommu/io-pgtable-arm.c
+++ b/drivers/iommu/io-pgtable-arm.c
@@ -445,7 +445,6 @@ static int arm_lpae_split_blk_unmap(struct arm_lpae_io_pgtable *data,
unsigned long blk_start, blk_end;
phys_addr_t blk_paddr;
arm_lpae_iopte table = 0;
- struct io_pgtable_cfg *cfg = &data->iop.cfg;
blk_start = iova & ~(blk_size - 1);
blk_end = blk_start + blk_size;
@@ -471,9 +470,9 @@ static int arm_lpae_split_blk_unmap(struct arm_lpae_io_pgtable *data,
}
}
- __arm_lpae_set_pte(ptep, table, cfg);
+ __arm_lpae_set_pte(ptep, table, &data->iop.cfg);
iova &= ~(blk_size - 1);
- cfg->tlb->tlb_add_flush(iova, blk_size, blk_size, true, data->iop.cookie);
+ io_pgtable_tlb_add_flush(&data->iop, iova, blk_size, blk_size, true);
return size;
}
@@ -482,8 +481,7 @@ static int __arm_lpae_unmap(struct arm_lpae_io_pgtable *data,
arm_lpae_iopte *ptep)
{
arm_lpae_iopte pte;
- const struct iommu_gather_ops *tlb = data->iop.cfg.tlb;
- void *cookie = data->iop.cookie;
+ struct io_pgtable *iop = &data->iop;
size_t blk_size = ARM_LPAE_BLOCK_SIZE(lvl, data);
/* Something went horribly wrong and we ran out of page table */
@@ -497,17 +495,17 @@ static int __arm_lpae_unmap(struct arm_lpae_io_pgtable *data,
/* If the size matches this level, we're in the right place */
if (size == blk_size) {
- __arm_lpae_set_pte(ptep, 0, &data->iop.cfg);
+ __arm_lpae_set_pte(ptep, 0, &iop->cfg);
if (!iopte_leaf(pte, lvl)) {
/* Also flush any partial walks */
- tlb->tlb_add_flush(iova, size, ARM_LPAE_GRANULE(data),
- false, cookie);
- tlb->tlb_sync(cookie);
+ io_pgtable_tlb_add_flush(iop, iova, size,
+ ARM_LPAE_GRANULE(data), false);
+ io_pgtable_tlb_sync(iop);
ptep = iopte_deref(pte, data);
__arm_lpae_free_pgtable(data, lvl + 1, ptep);
} else {
- tlb->tlb_add_flush(iova, size, size, true, cookie);
+ io_pgtable_tlb_add_flush(iop, iova, size, size, true);
}
return size;
@@ -531,13 +529,12 @@ static int arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
{
size_t unmapped;
struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
- struct io_pgtable *iop = &data->iop;
arm_lpae_iopte *ptep = data->pgd;
int lvl = ARM_LPAE_START_LVL(data);
unmapped = __arm_lpae_unmap(data, iova, size, lvl, ptep);
if (unmapped)
- iop->cfg.tlb->tlb_sync(iop->cookie);
+ io_pgtable_tlb_sync(&data->iop);
return unmapped;
}
diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
index 8c615b7..876f6a7 100644
--- a/drivers/iommu/io-pgtable.c
+++ b/drivers/iommu/io-pgtable.c
@@ -75,6 +75,6 @@ void free_io_pgtable_ops(struct io_pgtable_ops *ops)
return;
iop = container_of(ops, struct io_pgtable, ops);
- iop->cfg.tlb->tlb_flush_all(iop->cookie);
+ io_pgtable_tlb_flush_all(iop);
io_pgtable_init_table[iop->fmt]->free(iop);
}
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index aa57073..95c5565 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -144,6 +144,22 @@ struct io_pgtable {
#define io_pgtable_ops_to_pgtable(x) container_of((x), struct io_pgtable, ops)
+static inline void io_pgtable_tlb_flush_all(struct io_pgtable *iop)
+{
+ iop->cfg.tlb->tlb_flush_all(iop->cookie);
+}
+
+static inline void io_pgtable_tlb_add_flush(struct io_pgtable *iop,
+ unsigned long iova, size_t size, size_t granule, bool leaf)
+{
+ iop->cfg.tlb->tlb_add_flush(iova, size, granule, leaf, iop->cookie);
+}
+
+static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
+{
+ iop->cfg.tlb->tlb_sync(iop->cookie);
+}
+
/**
* struct io_pgtable_init_fns - Alloc/free a set of page tables for a
* particular format.
--
2.7.0.25.gfc10eb5
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH v3 3/3] iommu/io-pgtable: Avoid redundant TLB syncs
2016-01-26 17:13 ` Robin Murphy
@ 2016-01-26 17:13 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: will.deacon-5wv7dgnIgG8, yong.wu-NuS5LvNUpcJWk0Htik3J/w,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Cc: laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
In certain unmapping situations it is quite possible to end up issuing
back-to-back TLB synchronisations, which at best is a waste of time and
effort, and at worst causes some hardware to get rather confused. Whilst
the pagetable implementations, or the IOMMU drivers, or both, could keep
track of things to avoid this happening, it seems to make the most sense
to prevent code duplication and add some simple state tracking in the
common interface between the two.
Reviewed-by: Laurent Pinchart <laurent.pinchart-ryLnwIuWjnjg/C1BVhZhaw@public.gmane.org>
Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
---
drivers/iommu/io-pgtable.h | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 95c5565..4faee7d 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -132,12 +132,14 @@ void free_io_pgtable_ops(struct io_pgtable_ops *ops);
* @fmt: The page table format.
* @cookie: An opaque token provided by the IOMMU driver and passed back to
* any callback routines.
+ * @tlb_sync_pending: Private flag for optimising out redundant syncs.
* @cfg: A copy of the page table configuration.
* @ops: The page table operations in use for this set of page tables.
*/
struct io_pgtable {
enum io_pgtable_fmt fmt;
void *cookie;
+ bool tlb_sync_pending;
struct io_pgtable_cfg cfg;
struct io_pgtable_ops ops;
};
@@ -147,17 +149,22 @@ struct io_pgtable {
static inline void io_pgtable_tlb_flush_all(struct io_pgtable *iop)
{
iop->cfg.tlb->tlb_flush_all(iop->cookie);
+ iop->tlb_sync_pending = true;
}
static inline void io_pgtable_tlb_add_flush(struct io_pgtable *iop,
unsigned long iova, size_t size, size_t granule, bool leaf)
{
iop->cfg.tlb->tlb_add_flush(iova, size, granule, leaf, iop->cookie);
+ iop->tlb_sync_pending = true;
}
static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
{
- iop->cfg.tlb->tlb_sync(iop->cookie);
+ if (iop->tlb_sync_pending) {
+ iop->cfg.tlb->tlb_sync(iop->cookie);
+ iop->tlb_sync_pending = false;
+ }
}
/**
--
2.7.0.25.gfc10eb5
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH v3 3/3] iommu/io-pgtable: Avoid redundant TLB syncs
@ 2016-01-26 17:13 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-26 17:13 UTC (permalink / raw)
To: linux-arm-kernel
In certain unmapping situations it is quite possible to end up issuing
back-to-back TLB synchronisations, which at best is a waste of time and
effort, and at worst causes some hardware to get rather confused. Whilst
the pagetable implementations, or the IOMMU drivers, or both, could keep
track of things to avoid this happening, it seems to make the most sense
to prevent code duplication and add some simple state tracking in the
common interface between the two.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
drivers/iommu/io-pgtable.h | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 95c5565..4faee7d 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -132,12 +132,14 @@ void free_io_pgtable_ops(struct io_pgtable_ops *ops);
* @fmt: The page table format.
* @cookie: An opaque token provided by the IOMMU driver and passed back to
* any callback routines.
+ * @tlb_sync_pending: Private flag for optimising out redundant syncs.
* @cfg: A copy of the page table configuration.
* @ops: The page table operations in use for this set of page tables.
*/
struct io_pgtable {
enum io_pgtable_fmt fmt;
void *cookie;
+ bool tlb_sync_pending;
struct io_pgtable_cfg cfg;
struct io_pgtable_ops ops;
};
@@ -147,17 +149,22 @@ struct io_pgtable {
static inline void io_pgtable_tlb_flush_all(struct io_pgtable *iop)
{
iop->cfg.tlb->tlb_flush_all(iop->cookie);
+ iop->tlb_sync_pending = true;
}
static inline void io_pgtable_tlb_add_flush(struct io_pgtable *iop,
unsigned long iova, size_t size, size_t granule, bool leaf)
{
iop->cfg.tlb->tlb_add_flush(iova, size, granule, leaf, iop->cookie);
+ iop->tlb_sync_pending = true;
}
static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
{
- iop->cfg.tlb->tlb_sync(iop->cookie);
+ if (iop->tlb_sync_pending) {
+ iop->cfg.tlb->tlb_sync(iop->cookie);
+ iop->tlb_sync_pending = false;
+ }
}
/**
--
2.7.0.25.gfc10eb5
^ permalink raw reply related [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-01-26 17:13 ` Robin Murphy
@ 2016-01-27 1:16 ` Yong Wu
-1 siblings, 0 replies; 46+ messages in thread
From: Yong Wu @ 2016-01-27 1:16 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
will.deacon-5wv7dgnIgG8, srv_heupstream-NuS5LvNUpcJWk0Htik3J/w,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
> Add a nearly-complete ARMv7 short descriptor implementation, omitting
> only a few legacy and CPU-centric aspects which shouldn't be necessary
> for IOMMU API use anyway.
>
> Signed-off-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
This looks good to me.
And actually our disp/vdec/venc works well on your v2.
You could choose the two below,
Reviewed-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
Tested-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
Thanks.
> ---
> drivers/iommu/Kconfig | 19 +
> drivers/iommu/Makefile | 1 +
> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
> drivers/iommu/io-pgtable.c | 3 +
> drivers/iommu/io-pgtable.h | 14 +-
> 5 files changed, 885 insertions(+), 1 deletion(-)
> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-01-27 1:16 ` Yong Wu
0 siblings, 0 replies; 46+ messages in thread
From: Yong Wu @ 2016-01-27 1:16 UTC (permalink / raw)
To: linux-arm-kernel
On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
> Add a nearly-complete ARMv7 short descriptor implementation, omitting
> only a few legacy and CPU-centric aspects which shouldn't be necessary
> for IOMMU API use anyway.
>
> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
This looks good to me.
And actually our disp/vdec/venc works well on your v2.
You could choose the two below,
Reviewed-by: Yong Wu <yong.wu@mediatek.com>
Tested-by: Yong Wu <yong.wu@mediatek.com>
Thanks.
> ---
> drivers/iommu/Kconfig | 19 +
> drivers/iommu/Makefile | 1 +
> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
> drivers/iommu/io-pgtable.c | 3 +
> drivers/iommu/io-pgtable.h | 14 +-
> 5 files changed, 885 insertions(+), 1 deletion(-)
> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-01-27 1:16 ` Yong Wu
@ 2016-01-28 13:41 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-28 13:41 UTC (permalink / raw)
To: Yong Wu, will.deacon-5wv7dgnIgG8
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
srv_heupstream-NuS5LvNUpcJWk0Htik3J/w,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On 27/01/16 01:16, Yong Wu wrote:
> On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
>> Add a nearly-complete ARMv7 short descriptor implementation, omitting
>> only a few legacy and CPU-centric aspects which shouldn't be necessary
>> for IOMMU API use anyway.
>>
>> Signed-off-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
>> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
>
> This looks good to me.
> And actually our disp/vdec/venc works well on your v2.
>
> You could choose the two below,
> Reviewed-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
> Tested-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
Great, thanks!
Will, it seems to make the most sense for Joerg to pick these patches up
directly with the MTK series that depends on them - any objections?
Robin.
>
> Thanks.
>
>> ---
>> drivers/iommu/Kconfig | 19 +
>> drivers/iommu/Makefile | 1 +
>> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
>> drivers/iommu/io-pgtable.c | 3 +
>> drivers/iommu/io-pgtable.h | 14 +-
>> 5 files changed, 885 insertions(+), 1 deletion(-)
>> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
>
>
>
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-01-28 13:41 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-01-28 13:41 UTC (permalink / raw)
To: linux-arm-kernel
On 27/01/16 01:16, Yong Wu wrote:
> On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
>> Add a nearly-complete ARMv7 short descriptor implementation, omitting
>> only a few legacy and CPU-centric aspects which shouldn't be necessary
>> for IOMMU API use anyway.
>>
>> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>
> This looks good to me.
> And actually our disp/vdec/venc works well on your v2.
>
> You could choose the two below,
> Reviewed-by: Yong Wu <yong.wu@mediatek.com>
> Tested-by: Yong Wu <yong.wu@mediatek.com>
Great, thanks!
Will, it seems to make the most sense for Joerg to pick these patches up
directly with the MTK series that depends on them - any objections?
Robin.
>
> Thanks.
>
>> ---
>> drivers/iommu/Kconfig | 19 +
>> drivers/iommu/Makefile | 1 +
>> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
>> drivers/iommu/io-pgtable.c | 3 +
>> drivers/iommu/io-pgtable.h | 14 +-
>> 5 files changed, 885 insertions(+), 1 deletion(-)
>> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
>
>
>
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v3 3/3] iommu/io-pgtable: Avoid redundant TLB syncs
2016-01-26 17:13 ` Robin Murphy
@ 2016-02-10 14:21 ` Will Deacon
-1 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-10 14:21 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On Tue, Jan 26, 2016 at 05:13:15PM +0000, Robin Murphy wrote:
> In certain unmapping situations it is quite possible to end up issuing
> back-to-back TLB synchronisations, which at best is a waste of time and
> effort, and at worst causes some hardware to get rather confused. Whilst
> the pagetable implementations, or the IOMMU drivers, or both, could keep
> track of things to avoid this happening, it seems to make the most sense
> to prevent code duplication and add some simple state tracking in the
> common interface between the two.
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart-ryLnwIuWjnjg/C1BVhZhaw@public.gmane.org>
> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> ---
> drivers/iommu/io-pgtable.h | 9 ++++++++-
> 1 file changed, 8 insertions(+), 1 deletion(-)
Acked-by: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 3/3] iommu/io-pgtable: Avoid redundant TLB syncs
@ 2016-02-10 14:21 ` Will Deacon
0 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-10 14:21 UTC (permalink / raw)
To: linux-arm-kernel
On Tue, Jan 26, 2016 at 05:13:15PM +0000, Robin Murphy wrote:
> In certain unmapping situations it is quite possible to end up issuing
> back-to-back TLB synchronisations, which at best is a waste of time and
> effort, and at worst causes some hardware to get rather confused. Whilst
> the pagetable implementations, or the IOMMU drivers, or both, could keep
> track of things to avoid this happening, it seems to make the most sense
> to prevent code duplication and add some simple state tracking in the
> common interface between the two.
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/iommu/io-pgtable.h | 9 ++++++++-
> 1 file changed, 8 insertions(+), 1 deletion(-)
Acked-by: Will Deacon <will.deacon@arm.com>
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-01-26 17:13 ` Robin Murphy
@ 2016-02-10 15:48 ` Will Deacon
-1 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-10 15:48 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
Hi Robin,
On Tue, Jan 26, 2016 at 05:13:13PM +0000, Robin Murphy wrote:
> Add a nearly-complete ARMv7 short descriptor implementation, omitting
> only a few legacy and CPU-centric aspects which shouldn't be necessary
> for IOMMU API use anyway.
>
> Signed-off-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> ---
> drivers/iommu/Kconfig | 19 +
> drivers/iommu/Makefile | 1 +
> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
> drivers/iommu/io-pgtable.c | 3 +
> drivers/iommu/io-pgtable.h | 14 +-
> 5 files changed, 885 insertions(+), 1 deletion(-)
> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
This looks good to me, with one minor comment below.
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 36673c8..aa57073 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -1,5 +1,6 @@
> #ifndef __IO_PGTABLE_H
> #define __IO_PGTABLE_H
> +#include <linux/bitops.h>
>
> /*
> * Public API for use by IOMMU drivers
> @@ -9,6 +10,7 @@ enum io_pgtable_fmt {
> ARM_32_LPAE_S2,
> ARM_64_LPAE_S1,
> ARM_64_LPAE_S2,
> + ARM_V7S,
> IO_PGTABLE_NUM_FMTS,
> };
>
> @@ -45,7 +47,9 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS (1 << 0) /* Set NS bit in PTEs */
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
We should either try to keep these generic or add _ARM_ in the names. I
don't see why we couldn't have them as generic quicks, but that means
you should fix the AP/XN comment to be architecture agnostic and also
hack the lpae iopgtable alloc code to reject a cfg with unsupported
quirks.
With that change, I'm happy to queue this via my tree or whatever is
easiest for Yong/Joerg.
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-02-10 15:48 ` Will Deacon
0 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-10 15:48 UTC (permalink / raw)
To: linux-arm-kernel
Hi Robin,
On Tue, Jan 26, 2016 at 05:13:13PM +0000, Robin Murphy wrote:
> Add a nearly-complete ARMv7 short descriptor implementation, omitting
> only a few legacy and CPU-centric aspects which shouldn't be necessary
> for IOMMU API use anyway.
>
> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/iommu/Kconfig | 19 +
> drivers/iommu/Makefile | 1 +
> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
> drivers/iommu/io-pgtable.c | 3 +
> drivers/iommu/io-pgtable.h | 14 +-
> 5 files changed, 885 insertions(+), 1 deletion(-)
> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
This looks good to me, with one minor comment below.
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 36673c8..aa57073 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -1,5 +1,6 @@
> #ifndef __IO_PGTABLE_H
> #define __IO_PGTABLE_H
> +#include <linux/bitops.h>
>
> /*
> * Public API for use by IOMMU drivers
> @@ -9,6 +10,7 @@ enum io_pgtable_fmt {
> ARM_32_LPAE_S2,
> ARM_64_LPAE_S1,
> ARM_64_LPAE_S2,
> + ARM_V7S,
> IO_PGTABLE_NUM_FMTS,
> };
>
> @@ -45,7 +47,9 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS (1 << 0) /* Set NS bit in PTEs */
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
We should either try to keep these generic or add _ARM_ in the names. I
don't see why we couldn't have them as generic quicks, but that means
you should fix the AP/XN comment to be architecture agnostic and also
hack the lpae iopgtable alloc code to reject a cfg with unsupported
quirks.
With that change, I'm happy to queue this via my tree or whatever is
easiest for Yong/Joerg.
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
2016-02-10 15:48 ` Will Deacon
@ 2016-02-11 16:13 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-11 16:13 UTC (permalink / raw)
To: will.deacon-5wv7dgnIgG8
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
As the number of io-pgtable implementations grows beyond 1, it's time
to rationalise the quirks mechanism before things have a chance to
start getting really ugly and out-of-hand.
To that end:
- Indicate exactly which quirks each format can/does support.
- Fail creating a table if a caller wants unsupported quirks.
- Properly document where each quirk applies and why.
Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
---
drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
drivers/iommu/io-pgtable-arm.c | 2 ++
drivers/iommu/io-pgtable.c | 3 +++
drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
4 files changed, 28 insertions(+), 4 deletions(-)
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
index d39a021..9bc607a 100644
--- a/drivers/iommu/io-pgtable-arm-v7s.c
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -690,6 +690,9 @@ out_free_data:
struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
.alloc = arm_v7s_alloc_pgtable,
.free = arm_v7s_free_pgtable,
+ .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
+ IO_PGTABLE_QUIRK_NO_PERMS |
+ IO_PGTABLE_QUIRK_TLBI_ON_MAP,
};
#ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
index 4095af2..32f94f1 100644
--- a/drivers/iommu/io-pgtable-arm.c
+++ b/drivers/iommu/io-pgtable-arm.c
@@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
.alloc = arm_64_lpae_alloc_pgtable_s1,
.free = arm_lpae_free_pgtable,
+ .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
};
struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
@@ -873,6 +874,7 @@ struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
.alloc = arm_32_lpae_alloc_pgtable_s1,
.free = arm_lpae_free_pgtable,
+ .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
};
struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
index 876f6a7..0f57a45 100644
--- a/drivers/iommu/io-pgtable.c
+++ b/drivers/iommu/io-pgtable.c
@@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum io_pgtable_fmt fmt,
if (!fns)
return NULL;
+ if (cfg->quirks & ~fns->supported_quirks)
+ return NULL;
+
iop = fns->alloc(cfg, cookie);
if (!iop)
return NULL;
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 4faee7d..6e7f11e 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -47,10 +47,24 @@ struct iommu_gather_ops {
* page table walker.
*/
struct io_pgtable_cfg {
- #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
- #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
- #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
- int quirks;
+ /*
+ * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
+ * stage 1 PTEs, for hardware which insists on validating them
+ * even in non-secure state where they should normally be ignored.
+ *
+ * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
+ * IOMMU_NOEXEC flags and map everything with full access, for
+ * hardware which does not implement the permissions of a given
+ * format, and/or requires some format-specific default value.
+ *
+ * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
+ * (unmapped) entries but the hardware might do so anyway, perform
+ * TLB maintenance when mapping as well as when unmapping.
+ */
+ #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
+ #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
+ #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
+ unsigned int quirks;
unsigned long pgsize_bitmap;
unsigned int ias;
unsigned int oas;
@@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
*
* @alloc: Allocate a set of page tables described by cfg.
* @free: Free the page tables associated with iop.
+ * @supported_quirks: Bitmap of quirks supported by the format.
*/
struct io_pgtable_init_fns {
struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void *cookie);
void (*free)(struct io_pgtable *iop);
+ unsigned int supported_quirks;
};
extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
--
2.7.0.25.gfc10eb5.dirty
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-11 16:13 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-11 16:13 UTC (permalink / raw)
To: linux-arm-kernel
As the number of io-pgtable implementations grows beyond 1, it's time
to rationalise the quirks mechanism before things have a chance to
start getting really ugly and out-of-hand.
To that end:
- Indicate exactly which quirks each format can/does support.
- Fail creating a table if a caller wants unsupported quirks.
- Properly document where each quirk applies and why.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
drivers/iommu/io-pgtable-arm.c | 2 ++
drivers/iommu/io-pgtable.c | 3 +++
drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
4 files changed, 28 insertions(+), 4 deletions(-)
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
index d39a021..9bc607a 100644
--- a/drivers/iommu/io-pgtable-arm-v7s.c
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -690,6 +690,9 @@ out_free_data:
struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
.alloc = arm_v7s_alloc_pgtable,
.free = arm_v7s_free_pgtable,
+ .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
+ IO_PGTABLE_QUIRK_NO_PERMS |
+ IO_PGTABLE_QUIRK_TLBI_ON_MAP,
};
#ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
index 4095af2..32f94f1 100644
--- a/drivers/iommu/io-pgtable-arm.c
+++ b/drivers/iommu/io-pgtable-arm.c
@@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
.alloc = arm_64_lpae_alloc_pgtable_s1,
.free = arm_lpae_free_pgtable,
+ .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
};
struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
@@ -873,6 +874,7 @@ struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
.alloc = arm_32_lpae_alloc_pgtable_s1,
.free = arm_lpae_free_pgtable,
+ .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
};
struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
index 876f6a7..0f57a45 100644
--- a/drivers/iommu/io-pgtable.c
+++ b/drivers/iommu/io-pgtable.c
@@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum io_pgtable_fmt fmt,
if (!fns)
return NULL;
+ if (cfg->quirks & ~fns->supported_quirks)
+ return NULL;
+
iop = fns->alloc(cfg, cookie);
if (!iop)
return NULL;
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 4faee7d..6e7f11e 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -47,10 +47,24 @@ struct iommu_gather_ops {
* page table walker.
*/
struct io_pgtable_cfg {
- #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
- #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
- #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
- int quirks;
+ /*
+ * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
+ * stage 1 PTEs, for hardware which insists on validating them
+ * even in non-secure state where they should normally be ignored.
+ *
+ * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
+ * IOMMU_NOEXEC flags and map everything with full access, for
+ * hardware which does not implement the permissions of a given
+ * format, and/or requires some format-specific default value.
+ *
+ * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
+ * (unmapped) entries but the hardware might do so anyway, perform
+ * TLB maintenance when mapping as well as when unmapping.
+ */
+ #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
+ #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
+ #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
+ unsigned int quirks;
unsigned long pgsize_bitmap;
unsigned int ias;
unsigned int oas;
@@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
*
* @alloc: Allocate a set of page tables described by cfg.
* @free: Free the page tables associated with iop.
+ * @supported_quirks: Bitmap of quirks supported by the format.
*/
struct io_pgtable_init_fns {
struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void *cookie);
void (*free)(struct io_pgtable *iop);
+ unsigned int supported_quirks;
};
extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
--
2.7.0.25.gfc10eb5.dirty
^ permalink raw reply related [flat|nested] 46+ messages in thread
* Re: [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
2016-02-11 16:13 ` Robin Murphy
@ 2016-02-11 20:39 ` Laurent Pinchart
-1 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-02-11 20:39 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
will.deacon-5wv7dgnIgG8,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
Hi Robin,
Thank you for the patch.
On Thursday 11 February 2016 16:13:45 Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart-ryLnwIuWjnjg/C1BVhZhaw@public.gmane.org>
> ---
> drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
> drivers/iommu/io-pgtable-arm.c | 2 ++
> drivers/iommu/io-pgtable.c | 3 +++
> drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
> 4 files changed, 28 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
> b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9bc607a 100644
> --- a/drivers/iommu/io-pgtable-arm-v7s.c
> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> @@ -690,6 +690,9 @@ out_free_data:
> struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
> .alloc = arm_v7s_alloc_pgtable,
> .free = arm_v7s_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
> + IO_PGTABLE_QUIRK_NO_PERMS |
> + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
> };
>
> #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
> index 4095af2..32f94f1 100644
> --- a/drivers/iommu/io-pgtable-arm.c
> +++ b/drivers/iommu/io-pgtable-arm.c
> @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg,
> void *cookie) struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns
> = { .alloc = arm_64_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> @@ -873,6 +874,7 @@ struct io_pgtable_init_fns
> io_pgtable_arm_64_lpae_s2_init_fns = { struct io_pgtable_init_fns
> io_pgtable_arm_32_lpae_s1_init_fns = { .alloc =
> arm_32_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
> diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
> index 876f6a7..0f57a45 100644
> --- a/drivers/iommu/io-pgtable.c
> +++ b/drivers/iommu/io-pgtable.c
> @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum
> io_pgtable_fmt fmt, if (!fns)
> return NULL;
>
> + if (cfg->quirks & ~fns->supported_quirks)
> + return NULL;
> +
> iop = fns->alloc(cfg, cookie);
> if (!iop)
> return NULL;
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 4faee7d..6e7f11e 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> - int quirks;
> + /*
> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> + * stage 1 PTEs, for hardware which insists on validating them
> + * even in non-secure state where they should normally be ignored.
> + *
> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> + * IOMMU_NOEXEC flags and map everything with full access, for
> + * hardware which does not implement the permissions of a given
> + * format, and/or requires some format-specific default value.
> + *
> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
> + * (unmapped) entries but the hardware might do so anyway, perform
> + * TLB maintenance when mapping as well as when unmapping.
> + */
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> + unsigned int quirks;
> unsigned long pgsize_bitmap;
> unsigned int ias;
> unsigned int oas;
> @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
> io_pgtable *iop) *
> * @alloc: Allocate a set of page tables described by cfg.
> * @free: Free the page tables associated with iop.
> + * @supported_quirks: Bitmap of quirks supported by the format.
> */
> struct io_pgtable_init_fns {
> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void *cookie);
> void (*free)(struct io_pgtable *iop);
> + unsigned int supported_quirks;
> };
>
> extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-11 20:39 ` Laurent Pinchart
0 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-02-11 20:39 UTC (permalink / raw)
To: linux-arm-kernel
Hi Robin,
Thank you for the patch.
On Thursday 11 February 2016 16:13:45 Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> ---
> drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
> drivers/iommu/io-pgtable-arm.c | 2 ++
> drivers/iommu/io-pgtable.c | 3 +++
> drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
> 4 files changed, 28 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
> b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9bc607a 100644
> --- a/drivers/iommu/io-pgtable-arm-v7s.c
> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> @@ -690,6 +690,9 @@ out_free_data:
> struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
> .alloc = arm_v7s_alloc_pgtable,
> .free = arm_v7s_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
> + IO_PGTABLE_QUIRK_NO_PERMS |
> + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
> };
>
> #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
> index 4095af2..32f94f1 100644
> --- a/drivers/iommu/io-pgtable-arm.c
> +++ b/drivers/iommu/io-pgtable-arm.c
> @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg,
> void *cookie) struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns
> = { .alloc = arm_64_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> @@ -873,6 +874,7 @@ struct io_pgtable_init_fns
> io_pgtable_arm_64_lpae_s2_init_fns = { struct io_pgtable_init_fns
> io_pgtable_arm_32_lpae_s1_init_fns = { .alloc =
> arm_32_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
> diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
> index 876f6a7..0f57a45 100644
> --- a/drivers/iommu/io-pgtable.c
> +++ b/drivers/iommu/io-pgtable.c
> @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum
> io_pgtable_fmt fmt, if (!fns)
> return NULL;
>
> + if (cfg->quirks & ~fns->supported_quirks)
> + return NULL;
> +
> iop = fns->alloc(cfg, cookie);
> if (!iop)
> return NULL;
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 4faee7d..6e7f11e 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> - int quirks;
> + /*
> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> + * stage 1 PTEs, for hardware which insists on validating them
> + * even in non-secure state where they should normally be ignored.
> + *
> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> + * IOMMU_NOEXEC flags and map everything with full access, for
> + * hardware which does not implement the permissions of a given
> + * format, and/or requires some format-specific default value.
> + *
> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
> + * (unmapped) entries but the hardware might do so anyway, perform
> + * TLB maintenance when mapping as well as when unmapping.
> + */
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> + unsigned int quirks;
> unsigned long pgsize_bitmap;
> unsigned int ias;
> unsigned int oas;
> @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
> io_pgtable *iop) *
> * @alloc: Allocate a set of page tables described by cfg.
> * @free: Free the page tables associated with iop.
> + * @supported_quirks: Bitmap of quirks supported by the format.
> */
> struct io_pgtable_init_fns {
> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void *cookie);
> void (*free)(struct io_pgtable *iop);
> + unsigned int supported_quirks;
> };
>
> extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
2016-02-11 16:13 ` Robin Murphy
@ 2016-02-12 12:08 ` Will Deacon
-1 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-12 12:08 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> ---
> drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
> drivers/iommu/io-pgtable-arm.c | 2 ++
> drivers/iommu/io-pgtable.c | 3 +++
> drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
> 4 files changed, 28 insertions(+), 4 deletions(-)
Note that I can't merge this until we've figured out the plan for Yong's
driver.
> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
> index d39a021..9bc607a 100644
> --- a/drivers/iommu/io-pgtable-arm-v7s.c
> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> @@ -690,6 +690,9 @@ out_free_data:
> struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
> .alloc = arm_v7s_alloc_pgtable,
> .free = arm_v7s_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
> + IO_PGTABLE_QUIRK_NO_PERMS |
> + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
> };
>
> #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
> index 4095af2..32f94f1 100644
> --- a/drivers/iommu/io-pgtable-arm.c
> +++ b/drivers/iommu/io-pgtable-arm.c
> @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
> .alloc = arm_64_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> @@ -873,6 +874,7 @@ struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
> .alloc = arm_32_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
> diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
> index 876f6a7..0f57a45 100644
> --- a/drivers/iommu/io-pgtable.c
> +++ b/drivers/iommu/io-pgtable.c
> @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum io_pgtable_fmt fmt,
> if (!fns)
> return NULL;
>
> + if (cfg->quirks & ~fns->supported_quirks)
> + return NULL;
> +
> iop = fns->alloc(cfg, cookie);
> if (!iop)
> return NULL;
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 4faee7d..6e7f11e 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> - int quirks;
> + /*
> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> + * stage 1 PTEs, for hardware which insists on validating them
> + * even in non-secure state where they should normally be ignored.
> + *
> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> + * IOMMU_NOEXEC flags and map everything with full access, for
> + * hardware which does not implement the permissions of a given
> + * format, and/or requires some format-specific default value.
> + *
> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
> + * (unmapped) entries but the hardware might do so anyway, perform
> + * TLB maintenance when mapping as well as when unmapping.
> + */
Much better, cheers.
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> + unsigned int quirks;
Let's just make this unsigned long.
> unsigned long pgsize_bitmap;
> unsigned int ias;
> unsigned int oas;
> @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
> *
> * @alloc: Allocate a set of page tables described by cfg.
> * @free: Free the page tables associated with iop.
> + * @supported_quirks: Bitmap of quirks supported by the format.
> */
> struct io_pgtable_init_fns {
> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void *cookie);
> void (*free)(struct io_pgtable *iop);
> + unsigned int supported_quirks;
> };
My only gripe here is that this is supposed to be a struct of function
pointers... get_supported_quirks() ?
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 12:08 ` Will Deacon
0 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-12 12:08 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
> drivers/iommu/io-pgtable-arm.c | 2 ++
> drivers/iommu/io-pgtable.c | 3 +++
> drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
> 4 files changed, 28 insertions(+), 4 deletions(-)
Note that I can't merge this until we've figured out the plan for Yong's
driver.
> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
> index d39a021..9bc607a 100644
> --- a/drivers/iommu/io-pgtable-arm-v7s.c
> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> @@ -690,6 +690,9 @@ out_free_data:
> struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
> .alloc = arm_v7s_alloc_pgtable,
> .free = arm_v7s_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
> + IO_PGTABLE_QUIRK_NO_PERMS |
> + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
> };
>
> #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
> index 4095af2..32f94f1 100644
> --- a/drivers/iommu/io-pgtable-arm.c
> +++ b/drivers/iommu/io-pgtable-arm.c
> @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
> .alloc = arm_64_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> @@ -873,6 +874,7 @@ struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
> .alloc = arm_32_lpae_alloc_pgtable_s1,
> .free = arm_lpae_free_pgtable,
> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> };
>
> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
> diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
> index 876f6a7..0f57a45 100644
> --- a/drivers/iommu/io-pgtable.c
> +++ b/drivers/iommu/io-pgtable.c
> @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum io_pgtable_fmt fmt,
> if (!fns)
> return NULL;
>
> + if (cfg->quirks & ~fns->supported_quirks)
> + return NULL;
> +
> iop = fns->alloc(cfg, cookie);
> if (!iop)
> return NULL;
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 4faee7d..6e7f11e 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> - int quirks;
> + /*
> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> + * stage 1 PTEs, for hardware which insists on validating them
> + * even in non-secure state where they should normally be ignored.
> + *
> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> + * IOMMU_NOEXEC flags and map everything with full access, for
> + * hardware which does not implement the permissions of a given
> + * format, and/or requires some format-specific default value.
> + *
> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
> + * (unmapped) entries but the hardware might do so anyway, perform
> + * TLB maintenance when mapping as well as when unmapping.
> + */
Much better, cheers.
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> + unsigned int quirks;
Let's just make this unsigned long.
> unsigned long pgsize_bitmap;
> unsigned int ias;
> unsigned int oas;
> @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct io_pgtable *iop)
> *
> * @alloc: Allocate a set of page tables described by cfg.
> * @free: Free the page tables associated with iop.
> + * @supported_quirks: Bitmap of quirks supported by the format.
> */
> struct io_pgtable_init_fns {
> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void *cookie);
> void (*free)(struct io_pgtable *iop);
> + unsigned int supported_quirks;
> };
My only gripe here is that this is supposed to be a struct of function
pointers... get_supported_quirks() ?
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
2016-02-12 12:08 ` Will Deacon
@ 2016-02-12 12:32 ` Laurent Pinchart
-1 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-02-12 12:32 UTC (permalink / raw)
To: Will Deacon
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On Friday 12 February 2016 12:08:58 Will Deacon wrote:
> On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
> > As the number of io-pgtable implementations grows beyond 1, it's time
> > to rationalise the quirks mechanism before things have a chance to
> > start getting really ugly and out-of-hand.
> >
> > To that end:
> > - Indicate exactly which quirks each format can/does support.
> > - Fail creating a table if a caller wants unsupported quirks.
> > - Properly document where each quirk applies and why.
> >
> > Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> > ---
> >
> > drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
> > drivers/iommu/io-pgtable-arm.c | 2 ++
> > drivers/iommu/io-pgtable.c | 3 +++
> > drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
> > 4 files changed, 28 insertions(+), 4 deletions(-)
>
> Note that I can't merge this until we've figured out the plan for Yong's
> driver.
>
> > diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
> > b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9bc607a 100644
> > --- a/drivers/iommu/io-pgtable-arm-v7s.c
> > +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> >
> > @@ -690,6 +690,9 @@ out_free_data:
> > struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
> >
> > .alloc = arm_v7s_alloc_pgtable,
> > .free = arm_v7s_free_pgtable,
> >
> > + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
> > + IO_PGTABLE_QUIRK_NO_PERMS |
> > + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
> >
> > };
> >
> > #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> >
> > diff --git a/drivers/iommu/io-pgtable-arm.c
> > b/drivers/iommu/io-pgtable-arm.c index 4095af2..32f94f1 100644
> > --- a/drivers/iommu/io-pgtable-arm.c
> > +++ b/drivers/iommu/io-pgtable-arm.c
> > @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg
> > *cfg, void *cookie)>
> > struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
> >
> > .alloc = arm_64_lpae_alloc_pgtable_s1,
> > .free = arm_lpae_free_pgtable,
> >
> > + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> >
> > };
> >
> > struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> >
> > @@ -873,6 +874,7 @@ struct io_pgtable_init_fns
> > io_pgtable_arm_64_lpae_s2_init_fns = {>
> > struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
> >
> > .alloc = arm_32_lpae_alloc_pgtable_s1,
> > .free = arm_lpae_free_pgtable,
> >
> > + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> >
> > };
> >
> > struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
> >
> > diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
> > index 876f6a7..0f57a45 100644
> > --- a/drivers/iommu/io-pgtable.c
> > +++ b/drivers/iommu/io-pgtable.c
> > @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum
> > io_pgtable_fmt fmt,>
> > if (!fns)
> >
> > return NULL;
> >
> > + if (cfg->quirks & ~fns->supported_quirks)
> > + return NULL;
> > +
> >
> > iop = fns->alloc(cfg, cookie);
> > if (!iop)
> >
> > return NULL;
> >
> > diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> > index 4faee7d..6e7f11e 100644
> > --- a/drivers/iommu/io-pgtable.h
> > +++ b/drivers/iommu/io-pgtable.h
> > @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> >
> > * page table walker.
> > */
> >
> > struct io_pgtable_cfg {
> >
> > - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> > - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> > - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> > - int quirks;
> > + /*
> > + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> > + * stage 1 PTEs, for hardware which insists on validating them
> > + * even in non-secure state where they should normally be ignored.
> > + *
> > + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> > + * IOMMU_NOEXEC flags and map everything with full access, for
> > + * hardware which does not implement the permissions of a given
> > + * format, and/or requires some format-specific default value.
> > + *
> > + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching
invalid
> > + * (unmapped) entries but the hardware might do so anyway, perform
> > + * TLB maintenance when mapping as well as when unmapping.
> > + */
>
> Much better, cheers.
>
> > + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> > + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> > + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> > + unsigned int quirks;
>
> Let's just make this unsigned long.
>
> > unsigned long pgsize_bitmap;
> > unsigned int ias;
> > unsigned int oas;
> >
> > @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
> > io_pgtable *iop)
> > *
> > * @alloc: Allocate a set of page tables described by cfg.
> > * @free: Free the page tables associated with iop.
> > + * @supported_quirks: Bitmap of quirks supported by the format.
> > */
> > struct io_pgtable_init_fns {
> > struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void
*cookie);
> > void (*free)(struct io_pgtable *iop);
> > + unsigned int supported_quirks;
> > };
>
> My only gripe here is that this is supposed to be a struct of function
> pointers... get_supported_quirks() ?
Can't we instead say it's a struct of function pointers and static data ? :-)
A function to retrieve the supported quirks will just contribute to kernel
bloat without adding any extra benefit.
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 12:32 ` Laurent Pinchart
0 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-02-12 12:32 UTC (permalink / raw)
To: linux-arm-kernel
On Friday 12 February 2016 12:08:58 Will Deacon wrote:
> On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
> > As the number of io-pgtable implementations grows beyond 1, it's time
> > to rationalise the quirks mechanism before things have a chance to
> > start getting really ugly and out-of-hand.
> >
> > To that end:
> > - Indicate exactly which quirks each format can/does support.
> > - Fail creating a table if a caller wants unsupported quirks.
> > - Properly document where each quirk applies and why.
> >
> > Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> > ---
> >
> > drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
> > drivers/iommu/io-pgtable-arm.c | 2 ++
> > drivers/iommu/io-pgtable.c | 3 +++
> > drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
> > 4 files changed, 28 insertions(+), 4 deletions(-)
>
> Note that I can't merge this until we've figured out the plan for Yong's
> driver.
>
> > diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
> > b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9bc607a 100644
> > --- a/drivers/iommu/io-pgtable-arm-v7s.c
> > +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> >
> > @@ -690,6 +690,9 @@ out_free_data:
> > struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
> >
> > .alloc = arm_v7s_alloc_pgtable,
> > .free = arm_v7s_free_pgtable,
> >
> > + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
> > + IO_PGTABLE_QUIRK_NO_PERMS |
> > + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
> >
> > };
> >
> > #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> >
> > diff --git a/drivers/iommu/io-pgtable-arm.c
> > b/drivers/iommu/io-pgtable-arm.c index 4095af2..32f94f1 100644
> > --- a/drivers/iommu/io-pgtable-arm.c
> > +++ b/drivers/iommu/io-pgtable-arm.c
> > @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg
> > *cfg, void *cookie)>
> > struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
> >
> > .alloc = arm_64_lpae_alloc_pgtable_s1,
> > .free = arm_lpae_free_pgtable,
> >
> > + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> >
> > };
> >
> > struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
> >
> > @@ -873,6 +874,7 @@ struct io_pgtable_init_fns
> > io_pgtable_arm_64_lpae_s2_init_fns = {>
> > struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
> >
> > .alloc = arm_32_lpae_alloc_pgtable_s1,
> > .free = arm_lpae_free_pgtable,
> >
> > + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
> >
> > };
> >
> > struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
> >
> > diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
> > index 876f6a7..0f57a45 100644
> > --- a/drivers/iommu/io-pgtable.c
> > +++ b/drivers/iommu/io-pgtable.c
> > @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum
> > io_pgtable_fmt fmt,>
> > if (!fns)
> >
> > return NULL;
> >
> > + if (cfg->quirks & ~fns->supported_quirks)
> > + return NULL;
> > +
> >
> > iop = fns->alloc(cfg, cookie);
> > if (!iop)
> >
> > return NULL;
> >
> > diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> > index 4faee7d..6e7f11e 100644
> > --- a/drivers/iommu/io-pgtable.h
> > +++ b/drivers/iommu/io-pgtable.h
> > @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> >
> > * page table walker.
> > */
> >
> > struct io_pgtable_cfg {
> >
> > - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> > - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> > - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> > - int quirks;
> > + /*
> > + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> > + * stage 1 PTEs, for hardware which insists on validating them
> > + * even in non-secure state where they should normally be ignored.
> > + *
> > + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> > + * IOMMU_NOEXEC flags and map everything with full access, for
> > + * hardware which does not implement the permissions of a given
> > + * format, and/or requires some format-specific default value.
> > + *
> > + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching
invalid
> > + * (unmapped) entries but the hardware might do so anyway, perform
> > + * TLB maintenance when mapping as well as when unmapping.
> > + */
>
> Much better, cheers.
>
> > + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> > + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> > + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> > + unsigned int quirks;
>
> Let's just make this unsigned long.
>
> > unsigned long pgsize_bitmap;
> > unsigned int ias;
> > unsigned int oas;
> >
> > @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
> > io_pgtable *iop)
> > *
> > * @alloc: Allocate a set of page tables described by cfg.
> > * @free: Free the page tables associated with iop.
> > + * @supported_quirks: Bitmap of quirks supported by the format.
> > */
> > struct io_pgtable_init_fns {
> > struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void
*cookie);
> > void (*free)(struct io_pgtable *iop);
> > + unsigned int supported_quirks;
> > };
>
> My only gripe here is that this is supposed to be a struct of function
> pointers... get_supported_quirks() ?
Can't we instead say it's a struct of function pointers and static data ? :-)
A function to retrieve the supported quirks will just contribute to kernel
bloat without adding any extra benefit.
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
2016-02-12 12:32 ` Laurent Pinchart
@ 2016-02-12 15:23 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-12 15:23 UTC (permalink / raw)
To: Laurent Pinchart, Will Deacon
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On 12/02/16 12:32, Laurent Pinchart wrote:
> On Friday 12 February 2016 12:08:58 Will Deacon wrote:
>> On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
>>> As the number of io-pgtable implementations grows beyond 1, it's time
>>> to rationalise the quirks mechanism before things have a chance to
>>> start getting really ugly and out-of-hand.
>>>
>>> To that end:
>>> - Indicate exactly which quirks each format can/does support.
>>> - Fail creating a table if a caller wants unsupported quirks.
>>> - Properly document where each quirk applies and why.
>>>
>>> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
>>> ---
>>>
>>> drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
>>> drivers/iommu/io-pgtable-arm.c | 2 ++
>>> drivers/iommu/io-pgtable.c | 3 +++
>>> drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
>>> 4 files changed, 28 insertions(+), 4 deletions(-)
>>
>> Note that I can't merge this until we've figured out the plan for Yong's
>> driver.
Indeed (hence 4/3) - it just ended up seeming like a significant enough
change to warrant its own patch, but not critical enough to warrant
splitting it for the sake of applying the parts touching the existing
code right now.
>>> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
>>> b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9bc607a 100644
>>> --- a/drivers/iommu/io-pgtable-arm-v7s.c
>>> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
>>>
>>> @@ -690,6 +690,9 @@ out_free_data:
>>> struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
>>>
>>> .alloc = arm_v7s_alloc_pgtable,
>>> .free = arm_v7s_free_pgtable,
>>>
>>> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
>>> + IO_PGTABLE_QUIRK_NO_PERMS |
>>> + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
>>>
>>> };
>>>
>>> #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
>>>
>>> diff --git a/drivers/iommu/io-pgtable-arm.c
>>> b/drivers/iommu/io-pgtable-arm.c index 4095af2..32f94f1 100644
>>> --- a/drivers/iommu/io-pgtable-arm.c
>>> +++ b/drivers/iommu/io-pgtable-arm.c
>>> @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg
>>> *cfg, void *cookie)>
>>> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
>>>
>>> .alloc = arm_64_lpae_alloc_pgtable_s1,
>>> .free = arm_lpae_free_pgtable,
>>>
>>> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
>>>
>>> };
>>>
>>> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
>>>
>>> @@ -873,6 +874,7 @@ struct io_pgtable_init_fns
>>> io_pgtable_arm_64_lpae_s2_init_fns = {>
>>> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
>>>
>>> .alloc = arm_32_lpae_alloc_pgtable_s1,
>>> .free = arm_lpae_free_pgtable,
>>>
>>> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
>>>
>>> };
>>>
>>> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
>>>
>>> diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
>>> index 876f6a7..0f57a45 100644
>>> --- a/drivers/iommu/io-pgtable.c
>>> +++ b/drivers/iommu/io-pgtable.c
>>> @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum
>>> io_pgtable_fmt fmt,>
>>> if (!fns)
>>>
>>> return NULL;
>>>
>>> + if (cfg->quirks & ~fns->supported_quirks)
>>> + return NULL;
>>> +
>>>
>>> iop = fns->alloc(cfg, cookie);
>>> if (!iop)
>>>
>>> return NULL;
>>>
>>> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
>>> index 4faee7d..6e7f11e 100644
>>> --- a/drivers/iommu/io-pgtable.h
>>> +++ b/drivers/iommu/io-pgtable.h
>>> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
>>>
>>> * page table walker.
>>> */
>>>
>>> struct io_pgtable_cfg {
>>>
>>> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
>>> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
>>> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
>>> - int quirks;
>>> + /*
>>> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
>>> + * stage 1 PTEs, for hardware which insists on validating them
>>> + * even in non-secure state where they should normally be ignored.
>>> + *
>>> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
>>> + * IOMMU_NOEXEC flags and map everything with full access, for
>>> + * hardware which does not implement the permissions of a given
>>> + * format, and/or requires some format-specific default value.
>>> + *
>>> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching
> invalid
>>> + * (unmapped) entries but the hardware might do so anyway, perform
>>> + * TLB maintenance when mapping as well as when unmapping.
>>> + */
>>
>> Much better, cheers.
>>
>>> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
>>> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
>>> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
>>> + unsigned int quirks;
>>
>> Let's just make this unsigned long.
>>
>>> unsigned long pgsize_bitmap;
>>> unsigned int ias;
>>> unsigned int oas;
>>>
>>> @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
>>> io_pgtable *iop)
>>> *
>>> * @alloc: Allocate a set of page tables described by cfg.
>>> * @free: Free the page tables associated with iop.
>>> + * @supported_quirks: Bitmap of quirks supported by the format.
>>> */
>>> struct io_pgtable_init_fns {
>>> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void
> *cookie);
>>> void (*free)(struct io_pgtable *iop);
>>> + unsigned int supported_quirks;
>>> };
>>
>> My only gripe here is that this is supposed to be a struct of function
>> pointers... get_supported_quirks() ?
>
> Can't we instead say it's a struct of function pointers and static data ? :-)
> A function to retrieve the supported quirks will just contribute to kernel
> bloat without adding any extra benefit.
Bleh - personally I think it's marginally more obvious and readable to
have the information looking like data, but I'd much sooner just have a
hard-coded "if (cfq->quirks & ..." check in every format's alloc
function than introduce a whole new callback. Does that sound like a
reasonable compromise?
Robin.
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 15:23 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-12 15:23 UTC (permalink / raw)
To: linux-arm-kernel
On 12/02/16 12:32, Laurent Pinchart wrote:
> On Friday 12 February 2016 12:08:58 Will Deacon wrote:
>> On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
>>> As the number of io-pgtable implementations grows beyond 1, it's time
>>> to rationalise the quirks mechanism before things have a chance to
>>> start getting really ugly and out-of-hand.
>>>
>>> To that end:
>>> - Indicate exactly which quirks each format can/does support.
>>> - Fail creating a table if a caller wants unsupported quirks.
>>> - Properly document where each quirk applies and why.
>>>
>>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>>> ---
>>>
>>> drivers/iommu/io-pgtable-arm-v7s.c | 3 +++
>>> drivers/iommu/io-pgtable-arm.c | 2 ++
>>> drivers/iommu/io-pgtable.c | 3 +++
>>> drivers/iommu/io-pgtable.h | 24 ++++++++++++++++++++----
>>> 4 files changed, 28 insertions(+), 4 deletions(-)
>>
>> Note that I can't merge this until we've figured out the plan for Yong's
>> driver.
Indeed (hence 4/3) - it just ended up seeming like a significant enough
change to warrant its own patch, but not critical enough to warrant
splitting it for the sake of applying the parts touching the existing
code right now.
>>> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
>>> b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9bc607a 100644
>>> --- a/drivers/iommu/io-pgtable-arm-v7s.c
>>> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
>>>
>>> @@ -690,6 +690,9 @@ out_free_data:
>>> struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns = {
>>>
>>> .alloc = arm_v7s_alloc_pgtable,
>>> .free = arm_v7s_free_pgtable,
>>>
>>> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS |
>>> + IO_PGTABLE_QUIRK_NO_PERMS |
>>> + IO_PGTABLE_QUIRK_TLBI_ON_MAP,
>>>
>>> };
>>>
>>> #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
>>>
>>> diff --git a/drivers/iommu/io-pgtable-arm.c
>>> b/drivers/iommu/io-pgtable-arm.c index 4095af2..32f94f1 100644
>>> --- a/drivers/iommu/io-pgtable-arm.c
>>> +++ b/drivers/iommu/io-pgtable-arm.c
>>> @@ -863,6 +863,7 @@ arm_32_lpae_alloc_pgtable_s2(struct io_pgtable_cfg
>>> *cfg, void *cookie)>
>>> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
>>>
>>> .alloc = arm_64_lpae_alloc_pgtable_s1,
>>> .free = arm_lpae_free_pgtable,
>>>
>>> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
>>>
>>> };
>>>
>>> struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
>>>
>>> @@ -873,6 +874,7 @@ struct io_pgtable_init_fns
>>> io_pgtable_arm_64_lpae_s2_init_fns = {>
>>> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns = {
>>>
>>> .alloc = arm_32_lpae_alloc_pgtable_s1,
>>> .free = arm_lpae_free_pgtable,
>>>
>>> + .supported_quirks = IO_PGTABLE_QUIRK_ARM_NS,
>>>
>>> };
>>>
>>> struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns = {
>>>
>>> diff --git a/drivers/iommu/io-pgtable.c b/drivers/iommu/io-pgtable.c
>>> index 876f6a7..0f57a45 100644
>>> --- a/drivers/iommu/io-pgtable.c
>>> +++ b/drivers/iommu/io-pgtable.c
>>> @@ -52,6 +52,9 @@ struct io_pgtable_ops *alloc_io_pgtable_ops(enum
>>> io_pgtable_fmt fmt,>
>>> if (!fns)
>>>
>>> return NULL;
>>>
>>> + if (cfg->quirks & ~fns->supported_quirks)
>>> + return NULL;
>>> +
>>>
>>> iop = fns->alloc(cfg, cookie);
>>> if (!iop)
>>>
>>> return NULL;
>>>
>>> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
>>> index 4faee7d..6e7f11e 100644
>>> --- a/drivers/iommu/io-pgtable.h
>>> +++ b/drivers/iommu/io-pgtable.h
>>> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
>>>
>>> * page table walker.
>>> */
>>>
>>> struct io_pgtable_cfg {
>>>
>>> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
>>> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
>>> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
>>> - int quirks;
>>> + /*
>>> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
>>> + * stage 1 PTEs, for hardware which insists on validating them
>>> + * even in non-secure state where they should normally be ignored.
>>> + *
>>> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
>>> + * IOMMU_NOEXEC flags and map everything with full access, for
>>> + * hardware which does not implement the permissions of a given
>>> + * format, and/or requires some format-specific default value.
>>> + *
>>> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching
> invalid
>>> + * (unmapped) entries but the hardware might do so anyway, perform
>>> + * TLB maintenance when mapping as well as when unmapping.
>>> + */
>>
>> Much better, cheers.
>>
>>> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
>>> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
>>> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
>>> + unsigned int quirks;
>>
>> Let's just make this unsigned long.
>>
>>> unsigned long pgsize_bitmap;
>>> unsigned int ias;
>>> unsigned int oas;
>>>
>>> @@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
>>> io_pgtable *iop)
>>> *
>>> * @alloc: Allocate a set of page tables described by cfg.
>>> * @free: Free the page tables associated with iop.
>>> + * @supported_quirks: Bitmap of quirks supported by the format.
>>> */
>>> struct io_pgtable_init_fns {
>>> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void
> *cookie);
>>> void (*free)(struct io_pgtable *iop);
>>> + unsigned int supported_quirks;
>>> };
>>
>> My only gripe here is that this is supposed to be a struct of function
>> pointers... get_supported_quirks() ?
>
> Can't we instead say it's a struct of function pointers and static data ? :-)
> A function to retrieve the supported quirks will just contribute to kernel
> bloat without adding any extra benefit.
Bleh - personally I think it's marginally more obvious and readable to
have the information looking like data, but I'd much sooner just have a
hard-coded "if (cfq->quirks & ..." check in every format's alloc
function than introduce a whole new callback. Does that sound like a
reasonable compromise?
Robin.
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
2016-02-12 15:23 ` Robin Murphy
@ 2016-02-12 15:55 ` Will Deacon
-1 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-12 15:55 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
Laurent Pinchart,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On Fri, Feb 12, 2016 at 03:23:04PM +0000, Robin Murphy wrote:
> On 12/02/16 12:32, Laurent Pinchart wrote:
> >On Friday 12 February 2016 12:08:58 Will Deacon wrote:
> >>On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
> >>>@@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
> >>>io_pgtable *iop)
> >>> *
> >>> * @alloc: Allocate a set of page tables described by cfg.
> >>> * @free: Free the page tables associated with iop.
> >>>+ * @supported_quirks: Bitmap of quirks supported by the format.
> >>> */
> >>> struct io_pgtable_init_fns {
> >>> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void
> >*cookie);
> >>> void (*free)(struct io_pgtable *iop);
> >>>+ unsigned int supported_quirks;
> >>> };
> >>
> >>My only gripe here is that this is supposed to be a struct of function
> >>pointers... get_supported_quirks() ?
> >
> >Can't we instead say it's a struct of function pointers and static data ? :-)
> >A function to retrieve the supported quirks will just contribute to kernel
> >bloat without adding any extra benefit.
>
> Bleh - personally I think it's marginally more obvious and readable to have
> the information looking like data, but I'd much sooner just have a
> hard-coded "if (cfq->quirks & ..." check in every format's alloc function
> than introduce a whole new callback. Does that sound like a reasonable
> compromise?
Yeah, do that. You never know, there could be quirks that are only supported
based on other things in the cfg, so it's more extensible that way too.
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 15:55 ` Will Deacon
0 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-12 15:55 UTC (permalink / raw)
To: linux-arm-kernel
On Fri, Feb 12, 2016 at 03:23:04PM +0000, Robin Murphy wrote:
> On 12/02/16 12:32, Laurent Pinchart wrote:
> >On Friday 12 February 2016 12:08:58 Will Deacon wrote:
> >>On Thu, Feb 11, 2016 at 04:13:45PM +0000, Robin Murphy wrote:
> >>>@@ -173,10 +187,12 @@ static inline void io_pgtable_tlb_sync(struct
> >>>io_pgtable *iop)
> >>> *
> >>> * @alloc: Allocate a set of page tables described by cfg.
> >>> * @free: Free the page tables associated with iop.
> >>>+ * @supported_quirks: Bitmap of quirks supported by the format.
> >>> */
> >>> struct io_pgtable_init_fns {
> >>> struct io_pgtable *(*alloc)(struct io_pgtable_cfg *cfg, void
> >*cookie);
> >>> void (*free)(struct io_pgtable *iop);
> >>>+ unsigned int supported_quirks;
> >>> };
> >>
> >>My only gripe here is that this is supposed to be a struct of function
> >>pointers... get_supported_quirks() ?
> >
> >Can't we instead say it's a struct of function pointers and static data ? :-)
> >A function to retrieve the supported quirks will just contribute to kernel
> >bloat without adding any extra benefit.
>
> Bleh - personally I think it's marginally more obvious and readable to have
> the information looking like data, but I'd much sooner just have a
> hard-coded "if (cfq->quirks & ..." check in every format's alloc function
> than introduce a whole new callback. Does that sound like a reasonable
> compromise?
Yeah, do that. You never know, there could be quirks that are only supported
based on other things in the cfg, so it's more extensible that way too.
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
2016-01-26 17:13 ` Robin Murphy
@ 2016-02-12 17:09 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-12 17:09 UTC (permalink / raw)
To: will.deacon-5wv7dgnIgG8
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
As the number of io-pgtable implementations grows beyond 1, it's time
to rationalise the quirks mechanism before things have a chance to
start getting really ugly and out-of-hand.
To that end:
- Indicate exactly which quirks each format can/does support.
- Fail creating a table if a caller wants unsupported quirks.
- Properly document where each quirk applies and why.
Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
---
Just to allay any confusion, I intend this as a followup to the
short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
crucial addition, so it can either get picked up along with those
or wait 'til later.
v2:
- Use unsigned long for the bitmap.
- Do format-specific checks inline.
- Enough of a difference that I don't feel comfortable putting
Laurent's reviewed-by from v1 on it...
Robin.
[1]:http://thread.gmane.org/gmane.linux.kernel.iommu/12007
[2]:http://thread.gmane.org/gmane.linux.kernel.iommu/11974
drivers/iommu/io-pgtable-arm-v7s.c | 5 +++++
drivers/iommu/io-pgtable-arm.c | 13 +++++++++++--
drivers/iommu/io-pgtable.h | 22 ++++++++++++++++++----
3 files changed, 34 insertions(+), 6 deletions(-)
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
index d39a021..9fcceb1 100644
--- a/drivers/iommu/io-pgtable-arm-v7s.c
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -623,6 +623,11 @@ static struct io_pgtable *arm_v7s_alloc_pgtable(struct io_pgtable_cfg *cfg,
if (cfg->ias > ARM_V7S_ADDR_BITS || cfg->oas > ARM_V7S_ADDR_BITS)
return NULL;
+ if (cfg->quirks & ~(IO_PGTABLE_QUIRK_ARM_NS |
+ IO_PGTABLE_QUIRK_NO_PERMS |
+ IO_PGTABLE_QUIRK_TLBI_ON_MAP))
+ return NULL;
+
data = kmalloc(sizeof(*data), GFP_KERNEL);
if (!data)
return NULL;
diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
index 4095af2..debdf04 100644
--- a/drivers/iommu/io-pgtable-arm.c
+++ b/drivers/iommu/io-pgtable-arm.c
@@ -658,8 +658,12 @@ static struct io_pgtable *
arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
{
u64 reg;
- struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
+ struct arm_lpae_io_pgtable *data;
+ if (cfg->quirks & ~IO_PGTABLE_QUIRK_ARM_NS)
+ return NULL;
+
+ data = arm_lpae_alloc_pgtable(cfg);
if (!data)
return NULL;
@@ -742,8 +746,13 @@ static struct io_pgtable *
arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
{
u64 reg, sl;
- struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
+ struct arm_lpae_io_pgtable *data;
+ /* The NS quirk doesn't apply at stage 2 */
+ if (cfg->quirks)
+ return NULL;
+
+ data = arm_lpae_alloc_pgtable(cfg);
if (!data)
return NULL;
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 4faee7d..d4f5027 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -47,10 +47,24 @@ struct iommu_gather_ops {
* page table walker.
*/
struct io_pgtable_cfg {
- #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
- #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
- #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
- int quirks;
+ /*
+ * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
+ * stage 1 PTEs, for hardware which insists on validating them
+ * even in non-secure state where they should normally be ignored.
+ *
+ * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
+ * IOMMU_NOEXEC flags and map everything with full access, for
+ * hardware which does not implement the permissions of a given
+ * format, and/or requires some format-specific default value.
+ *
+ * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
+ * (unmapped) entries but the hardware might do so anyway, perform
+ * TLB maintenance when mapping as well as when unmapping.
+ */
+ #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
+ #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
+ #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
+ unsigned long quirks;
unsigned long pgsize_bitmap;
unsigned int ias;
unsigned int oas;
--
2.7.0.25.gfc10eb5.dirty
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 17:09 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-12 17:09 UTC (permalink / raw)
To: linux-arm-kernel
As the number of io-pgtable implementations grows beyond 1, it's time
to rationalise the quirks mechanism before things have a chance to
start getting really ugly and out-of-hand.
To that end:
- Indicate exactly which quirks each format can/does support.
- Fail creating a table if a caller wants unsupported quirks.
- Properly document where each quirk applies and why.
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
Just to allay any confusion, I intend this as a followup to the
short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
crucial addition, so it can either get picked up along with those
or wait 'til later.
v2:
- Use unsigned long for the bitmap.
- Do format-specific checks inline.
- Enough of a difference that I don't feel comfortable putting
Laurent's reviewed-by from v1 on it...
Robin.
[1]:http://thread.gmane.org/gmane.linux.kernel.iommu/12007
[2]:http://thread.gmane.org/gmane.linux.kernel.iommu/11974
drivers/iommu/io-pgtable-arm-v7s.c | 5 +++++
drivers/iommu/io-pgtable-arm.c | 13 +++++++++++--
drivers/iommu/io-pgtable.h | 22 ++++++++++++++++++----
3 files changed, 34 insertions(+), 6 deletions(-)
diff --git a/drivers/iommu/io-pgtable-arm-v7s.c b/drivers/iommu/io-pgtable-arm-v7s.c
index d39a021..9fcceb1 100644
--- a/drivers/iommu/io-pgtable-arm-v7s.c
+++ b/drivers/iommu/io-pgtable-arm-v7s.c
@@ -623,6 +623,11 @@ static struct io_pgtable *arm_v7s_alloc_pgtable(struct io_pgtable_cfg *cfg,
if (cfg->ias > ARM_V7S_ADDR_BITS || cfg->oas > ARM_V7S_ADDR_BITS)
return NULL;
+ if (cfg->quirks & ~(IO_PGTABLE_QUIRK_ARM_NS |
+ IO_PGTABLE_QUIRK_NO_PERMS |
+ IO_PGTABLE_QUIRK_TLBI_ON_MAP))
+ return NULL;
+
data = kmalloc(sizeof(*data), GFP_KERNEL);
if (!data)
return NULL;
diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
index 4095af2..debdf04 100644
--- a/drivers/iommu/io-pgtable-arm.c
+++ b/drivers/iommu/io-pgtable-arm.c
@@ -658,8 +658,12 @@ static struct io_pgtable *
arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
{
u64 reg;
- struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
+ struct arm_lpae_io_pgtable *data;
+ if (cfg->quirks & ~IO_PGTABLE_QUIRK_ARM_NS)
+ return NULL;
+
+ data = arm_lpae_alloc_pgtable(cfg);
if (!data)
return NULL;
@@ -742,8 +746,13 @@ static struct io_pgtable *
arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
{
u64 reg, sl;
- struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
+ struct arm_lpae_io_pgtable *data;
+ /* The NS quirk doesn't apply at stage 2 */
+ if (cfg->quirks)
+ return NULL;
+
+ data = arm_lpae_alloc_pgtable(cfg);
if (!data)
return NULL;
diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
index 4faee7d..d4f5027 100644
--- a/drivers/iommu/io-pgtable.h
+++ b/drivers/iommu/io-pgtable.h
@@ -47,10 +47,24 @@ struct iommu_gather_ops {
* page table walker.
*/
struct io_pgtable_cfg {
- #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
- #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
- #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
- int quirks;
+ /*
+ * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
+ * stage 1 PTEs, for hardware which insists on validating them
+ * even in non-secure state where they should normally be ignored.
+ *
+ * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
+ * IOMMU_NOEXEC flags and map everything with full access, for
+ * hardware which does not implement the permissions of a given
+ * format, and/or requires some format-specific default value.
+ *
+ * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
+ * (unmapped) entries but the hardware might do so anyway, perform
+ * TLB maintenance when mapping as well as when unmapping.
+ */
+ #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
+ #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
+ #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
+ unsigned long quirks;
unsigned long pgsize_bitmap;
unsigned int ias;
unsigned int oas;
--
2.7.0.25.gfc10eb5.dirty
^ permalink raw reply related [flat|nested] 46+ messages in thread
* Re: [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
2016-02-12 17:09 ` Robin Murphy
@ 2016-02-12 17:24 ` Laurent Pinchart
-1 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-02-12 17:24 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
will.deacon-5wv7dgnIgG8,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
Hi Robin,
Thank you for the patch.
On Friday 12 February 2016 17:09:46 Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> ---
>
> Just to allay any confusion, I intend this as a followup to the
> short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
> crucial addition, so it can either get picked up along with those
> or wait 'til later.
>
> v2:
> - Use unsigned long for the bitmap.
> - Do format-specific checks inline.
> - Enough of a difference that I don't feel comfortable putting
> Laurent's reviewed-by from v1 on it...
Here it is again.
Reviewed-by: Laurent Pinchart <laurent.pinchart-ryLnwIuWjnjg/C1BVhZhaw@public.gmane.org>
> [1]:http://thread.gmane.org/gmane.linux.kernel.iommu/12007
> [2]:http://thread.gmane.org/gmane.linux.kernel.iommu/11974
>
> drivers/iommu/io-pgtable-arm-v7s.c | 5 +++++
> drivers/iommu/io-pgtable-arm.c | 13 +++++++++++--
> drivers/iommu/io-pgtable.h | 22 ++++++++++++++++++----
> 3 files changed, 34 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
> b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9fcceb1 100644
> --- a/drivers/iommu/io-pgtable-arm-v7s.c
> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> @@ -623,6 +623,11 @@ static struct io_pgtable *arm_v7s_alloc_pgtable(struct
> io_pgtable_cfg *cfg, if (cfg->ias > ARM_V7S_ADDR_BITS || cfg->oas >
> ARM_V7S_ADDR_BITS) return NULL;
>
> + if (cfg->quirks & ~(IO_PGTABLE_QUIRK_ARM_NS |
> + IO_PGTABLE_QUIRK_NO_PERMS |
> + IO_PGTABLE_QUIRK_TLBI_ON_MAP))
> + return NULL;
> +
> data = kmalloc(sizeof(*data), GFP_KERNEL);
> if (!data)
> return NULL;
> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
> index 4095af2..debdf04 100644
> --- a/drivers/iommu/io-pgtable-arm.c
> +++ b/drivers/iommu/io-pgtable-arm.c
> @@ -658,8 +658,12 @@ static struct io_pgtable *
> arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
> {
> u64 reg;
> - struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
> + struct arm_lpae_io_pgtable *data;
>
> + if (cfg->quirks & ~IO_PGTABLE_QUIRK_ARM_NS)
> + return NULL;
> +
> + data = arm_lpae_alloc_pgtable(cfg);
> if (!data)
> return NULL;
>
> @@ -742,8 +746,13 @@ static struct io_pgtable *
> arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
> {
> u64 reg, sl;
> - struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
> + struct arm_lpae_io_pgtable *data;
>
> + /* The NS quirk doesn't apply at stage 2 */
> + if (cfg->quirks)
> + return NULL;
> +
> + data = arm_lpae_alloc_pgtable(cfg);
> if (!data)
> return NULL;
>
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 4faee7d..d4f5027 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> - int quirks;
> + /*
> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> + * stage 1 PTEs, for hardware which insists on validating them
> + * even in non-secure state where they should normally be ignored.
> + *
> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> + * IOMMU_NOEXEC flags and map everything with full access, for
> + * hardware which does not implement the permissions of a given
> + * format, and/or requires some format-specific default value.
> + *
> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
> + * (unmapped) entries but the hardware might do so anyway, perform
> + * TLB maintenance when mapping as well as when unmapping.
> + */
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> + unsigned long quirks;
> unsigned long pgsize_bitmap;
> unsigned int ias;
> unsigned int oas;
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 17:24 ` Laurent Pinchart
0 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-02-12 17:24 UTC (permalink / raw)
To: linux-arm-kernel
Hi Robin,
Thank you for the patch.
On Friday 12 February 2016 17:09:46 Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
>
> Just to allay any confusion, I intend this as a followup to the
> short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
> crucial addition, so it can either get picked up along with those
> or wait 'til later.
>
> v2:
> - Use unsigned long for the bitmap.
> - Do format-specific checks inline.
> - Enough of a difference that I don't feel comfortable putting
> Laurent's reviewed-by from v1 on it...
Here it is again.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> [1]:http://thread.gmane.org/gmane.linux.kernel.iommu/12007
> [2]:http://thread.gmane.org/gmane.linux.kernel.iommu/11974
>
> drivers/iommu/io-pgtable-arm-v7s.c | 5 +++++
> drivers/iommu/io-pgtable-arm.c | 13 +++++++++++--
> drivers/iommu/io-pgtable.h | 22 ++++++++++++++++++----
> 3 files changed, 34 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/iommu/io-pgtable-arm-v7s.c
> b/drivers/iommu/io-pgtable-arm-v7s.c index d39a021..9fcceb1 100644
> --- a/drivers/iommu/io-pgtable-arm-v7s.c
> +++ b/drivers/iommu/io-pgtable-arm-v7s.c
> @@ -623,6 +623,11 @@ static struct io_pgtable *arm_v7s_alloc_pgtable(struct
> io_pgtable_cfg *cfg, if (cfg->ias > ARM_V7S_ADDR_BITS || cfg->oas >
> ARM_V7S_ADDR_BITS) return NULL;
>
> + if (cfg->quirks & ~(IO_PGTABLE_QUIRK_ARM_NS |
> + IO_PGTABLE_QUIRK_NO_PERMS |
> + IO_PGTABLE_QUIRK_TLBI_ON_MAP))
> + return NULL;
> +
> data = kmalloc(sizeof(*data), GFP_KERNEL);
> if (!data)
> return NULL;
> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c
> index 4095af2..debdf04 100644
> --- a/drivers/iommu/io-pgtable-arm.c
> +++ b/drivers/iommu/io-pgtable-arm.c
> @@ -658,8 +658,12 @@ static struct io_pgtable *
> arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
> {
> u64 reg;
> - struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
> + struct arm_lpae_io_pgtable *data;
>
> + if (cfg->quirks & ~IO_PGTABLE_QUIRK_ARM_NS)
> + return NULL;
> +
> + data = arm_lpae_alloc_pgtable(cfg);
> if (!data)
> return NULL;
>
> @@ -742,8 +746,13 @@ static struct io_pgtable *
> arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
> {
> u64 reg, sl;
> - struct arm_lpae_io_pgtable *data = arm_lpae_alloc_pgtable(cfg);
> + struct arm_lpae_io_pgtable *data;
>
> + /* The NS quirk doesn't apply at stage 2 */
> + if (cfg->quirks)
> + return NULL;
> +
> + data = arm_lpae_alloc_pgtable(cfg);
> if (!data)
> return NULL;
>
> diff --git a/drivers/iommu/io-pgtable.h b/drivers/iommu/io-pgtable.h
> index 4faee7d..d4f5027 100644
> --- a/drivers/iommu/io-pgtable.h
> +++ b/drivers/iommu/io-pgtable.h
> @@ -47,10 +47,24 @@ struct iommu_gather_ops {
> * page table walker.
> */
> struct io_pgtable_cfg {
> - #define IO_PGTABLE_QUIRK_ARM_NS BIT(0) /* Set NS bit in PTEs */
> - #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1) /* No AP/XN bits */
> - #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2) /* TLB Inv. on map */
> - int quirks;
> + /*
> + * IO_PGTABLE_QUIRK_ARM_NS: (ARM formats) Set NS and NSTABLE bits in
> + * stage 1 PTEs, for hardware which insists on validating them
> + * even in non-secure state where they should normally be ignored.
> + *
> + * IO_PGTABLE_QUIRK_NO_PERMS: Ignore the IOMMU_READ, IOMMU_WRITE and
> + * IOMMU_NOEXEC flags and map everything with full access, for
> + * hardware which does not implement the permissions of a given
> + * format, and/or requires some format-specific default value.
> + *
> + * IO_PGTABLE_QUIRK_TLBI_ON_MAP: If the format forbids caching invalid
> + * (unmapped) entries but the hardware might do so anyway, perform
> + * TLB maintenance when mapping as well as when unmapping.
> + */
> + #define IO_PGTABLE_QUIRK_ARM_NS BIT(0)
> + #define IO_PGTABLE_QUIRK_NO_PERMS BIT(1)
> + #define IO_PGTABLE_QUIRK_TLBI_ON_MAP BIT(2)
> + unsigned long quirks;
> unsigned long pgsize_bitmap;
> unsigned int ias;
> unsigned int oas;
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
2016-02-12 17:24 ` Laurent Pinchart
@ 2016-02-12 17:25 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-12 17:25 UTC (permalink / raw)
To: Laurent Pinchart
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
will.deacon-5wv7dgnIgG8,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On 12/02/16 17:24, Laurent Pinchart wrote:
> Hi Robin,
>
> Thank you for the patch.
>
> On Friday 12 February 2016 17:09:46 Robin Murphy wrote:
>> As the number of io-pgtable implementations grows beyond 1, it's time
>> to rationalise the quirks mechanism before things have a chance to
>> start getting really ugly and out-of-hand.
>>
>> To that end:
>> - Indicate exactly which quirks each format can/does support.
>> - Fail creating a table if a caller wants unsupported quirks.
>> - Properly document where each quirk applies and why.
>>
>> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
>> ---
>>
>> Just to allay any confusion, I intend this as a followup to the
>> short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
>> crucial addition, so it can either get picked up along with those
>> or wait 'til later.
>>
>> v2:
>> - Use unsigned long for the bitmap.
>> - Do format-specific checks inline.
>> - Enough of a difference that I don't feel comfortable putting
>> Laurent's reviewed-by from v1 on it...
>
> Here it is again.
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart-ryLnwIuWjnjg/C1BVhZhaw@public.gmane.org>
Thanks!
Robin.
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 17:25 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-12 17:25 UTC (permalink / raw)
To: linux-arm-kernel
On 12/02/16 17:24, Laurent Pinchart wrote:
> Hi Robin,
>
> Thank you for the patch.
>
> On Friday 12 February 2016 17:09:46 Robin Murphy wrote:
>> As the number of io-pgtable implementations grows beyond 1, it's time
>> to rationalise the quirks mechanism before things have a chance to
>> start getting really ugly and out-of-hand.
>>
>> To that end:
>> - Indicate exactly which quirks each format can/does support.
>> - Fail creating a table if a caller wants unsupported quirks.
>> - Properly document where each quirk applies and why.
>>
>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>> ---
>>
>> Just to allay any confusion, I intend this as a followup to the
>> short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
>> crucial addition, so it can either get picked up along with those
>> or wait 'til later.
>>
>> v2:
>> - Use unsigned long for the bitmap.
>> - Do format-specific checks inline.
>> - Enough of a difference that I don't feel comfortable putting
>> Laurent's reviewed-by from v1 on it...
>
> Here it is again.
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Thanks!
Robin.
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
2016-02-12 17:09 ` Robin Murphy
@ 2016-02-12 17:27 ` Will Deacon
-1 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-12 17:27 UTC (permalink / raw)
To: Robin Murphy
Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On Fri, Feb 12, 2016 at 05:09:46PM +0000, Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> ---
>
> Just to allay any confusion, I intend this as a followup to the
> short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
> crucial addition, so it can either get picked up along with those
> or wait 'til later.
I'd like to see it included in whichever tree takes the short-descriptor
code so, in case that's not my tree:
Acked-by: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v2] iommu/io-pgtable: Rationalise quirk handling
@ 2016-02-12 17:27 ` Will Deacon
0 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-12 17:27 UTC (permalink / raw)
To: linux-arm-kernel
On Fri, Feb 12, 2016 at 05:09:46PM +0000, Robin Murphy wrote:
> As the number of io-pgtable implementations grows beyond 1, it's time
> to rationalise the quirks mechanism before things have a chance to
> start getting really ugly and out-of-hand.
>
> To that end:
> - Indicate exactly which quirks each format can/does support.
> - Fail creating a table if a caller wants unsupported quirks.
> - Properly document where each quirk applies and why.
>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
>
> Just to allay any confusion, I intend this as a followup to the
> short-descriptor[1] and Mediatek IOMMU[2] patches, rather than a
> crucial addition, so it can either get picked up along with those
> or wait 'til later.
I'd like to see it included in whichever tree takes the short-descriptor
code so, in case that's not my tree:
Acked-by: Will Deacon <will.deacon@arm.com>
Will
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-01-28 13:41 ` Robin Murphy
@ 2016-02-17 14:31 ` Will Deacon
-1 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-17 14:31 UTC (permalink / raw)
To: Robin Murphy
Cc: srv_heupstream-NuS5LvNUpcJWk0Htik3J/w,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
Hi Robin,
On Thu, Jan 28, 2016 at 01:41:29PM +0000, Robin Murphy wrote:
> On 27/01/16 01:16, Yong Wu wrote:
> >On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
> >>Add a nearly-complete ARMv7 short descriptor implementation, omitting
> >>only a few legacy and CPU-centric aspects which shouldn't be necessary
> >>for IOMMU API use anyway.
> >>
> >>Signed-off-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
> >>Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> >
> >This looks good to me.
> >And actually our disp/vdec/venc works well on your v2.
> >
> >You could choose the two below,
> >Reviewed-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
> >Tested-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
>
> Great, thanks!
>
> Will, it seems to make the most sense for Joerg to pick these patches up
> directly with the MTK series that depends on them - any objections?
As discussed with Joerg, I'll pick these up and send him a pull request
containing all the pgtable bits. I've taken the liberty of updating
MAINTAINERS too, assuming you don't mind being a reviewer for this?
Will
--->8
>From bccb08d096703bd790fcf2f934553dd9c8503fc9 Mon Sep 17 00:00:00 2001
From: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
Date: Wed, 17 Feb 2016 14:18:00 +0000
Subject: [PATCH] MAINTAINERS: update ARM SMMU entry
Ensure that the short-descriptor page table code is included under the
SMMU entry, and add Robin as a designated reviewer.
Cc: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
Signed-off-by: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
---
MAINTAINERS | 2 ++
1 file changed, 2 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index cc2f753cb357..a01bae148b89 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1793,11 +1793,13 @@ F: drivers/edac/synopsys_edac.c
ARM SMMU DRIVERS
M: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
+R: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
L: linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org (moderated for non-subscribers)
S: Maintained
F: drivers/iommu/arm-smmu.c
F: drivers/iommu/arm-smmu-v3.c
F: drivers/iommu/io-pgtable-arm.c
+F: drivers/iommu/io-pgtable-arm-v7s.c
ARM64 PORT (AARCH64 ARCHITECTURE)
M: Catalin Marinas <catalin.marinas-5wv7dgnIgG8@public.gmane.org>
--
2.1.4
^ permalink raw reply related [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-02-17 14:31 ` Will Deacon
0 siblings, 0 replies; 46+ messages in thread
From: Will Deacon @ 2016-02-17 14:31 UTC (permalink / raw)
To: linux-arm-kernel
Hi Robin,
On Thu, Jan 28, 2016 at 01:41:29PM +0000, Robin Murphy wrote:
> On 27/01/16 01:16, Yong Wu wrote:
> >On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
> >>Add a nearly-complete ARMv7 short descriptor implementation, omitting
> >>only a few legacy and CPU-centric aspects which shouldn't be necessary
> >>for IOMMU API use anyway.
> >>
> >>Signed-off-by: Yong Wu <yong.wu@mediatek.com>
> >>Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> >
> >This looks good to me.
> >And actually our disp/vdec/venc works well on your v2.
> >
> >You could choose the two below,
> >Reviewed-by: Yong Wu <yong.wu@mediatek.com>
> >Tested-by: Yong Wu <yong.wu@mediatek.com>
>
> Great, thanks!
>
> Will, it seems to make the most sense for Joerg to pick these patches up
> directly with the MTK series that depends on them - any objections?
As discussed with Joerg, I'll pick these up and send him a pull request
containing all the pgtable bits. I've taken the liberty of updating
MAINTAINERS too, assuming you don't mind being a reviewer for this?
Will
--->8
>From bccb08d096703bd790fcf2f934553dd9c8503fc9 Mon Sep 17 00:00:00 2001
From: Will Deacon <will.deacon@arm.com>
Date: Wed, 17 Feb 2016 14:18:00 +0000
Subject: [PATCH] MAINTAINERS: update ARM SMMU entry
Ensure that the short-descriptor page table code is included under the
SMMU entry, and add Robin as a designated reviewer.
Cc: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
MAINTAINERS | 2 ++
1 file changed, 2 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index cc2f753cb357..a01bae148b89 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1793,11 +1793,13 @@ F: drivers/edac/synopsys_edac.c
ARM SMMU DRIVERS
M: Will Deacon <will.deacon@arm.com>
+R: Robin Murphy <robin.murphy@arm.com>
L: linux-arm-kernel at lists.infradead.org (moderated for non-subscribers)
S: Maintained
F: drivers/iommu/arm-smmu.c
F: drivers/iommu/arm-smmu-v3.c
F: drivers/iommu/io-pgtable-arm.c
+F: drivers/iommu/io-pgtable-arm-v7s.c
ARM64 PORT (AARCH64 ARCHITECTURE)
M: Catalin Marinas <catalin.marinas@arm.com>
--
2.1.4
^ permalink raw reply related [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-02-17 14:31 ` Will Deacon
@ 2016-02-17 14:34 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-17 14:34 UTC (permalink / raw)
To: Will Deacon
Cc: srv_heupstream-NuS5LvNUpcJWk0Htik3J/w,
iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
laurent.pinchart+renesas-ryLnwIuWjnjg/C1BVhZhaw,
linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r
On 17/02/16 14:31, Will Deacon wrote:
> Hi Robin,
>
> On Thu, Jan 28, 2016 at 01:41:29PM +0000, Robin Murphy wrote:
>> On 27/01/16 01:16, Yong Wu wrote:
>>> On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
>>>> Add a nearly-complete ARMv7 short descriptor implementation, omitting
>>>> only a few legacy and CPU-centric aspects which shouldn't be necessary
>>>> for IOMMU API use anyway.
>>>>
>>>> Signed-off-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
>>>> Signed-off-by: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
>>>
>>> This looks good to me.
>>> And actually our disp/vdec/venc works well on your v2.
>>>
>>> You could choose the two below,
>>> Reviewed-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
>>> Tested-by: Yong Wu <yong.wu-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
>>
>> Great, thanks!
>>
>> Will, it seems to make the most sense for Joerg to pick these patches up
>> directly with the MTK series that depends on them - any objections?
>
> As discussed with Joerg, I'll pick these up and send him a pull request
> containing all the pgtable bits. I've taken the liberty of updating
> MAINTAINERS too, assuming you don't mind being a reviewer for this?
Sure, that's fine by me.
> Will
>
> --->8
>
> From bccb08d096703bd790fcf2f934553dd9c8503fc9 Mon Sep 17 00:00:00 2001
> From: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
> Date: Wed, 17 Feb 2016 14:18:00 +0000
> Subject: [PATCH] MAINTAINERS: update ARM SMMU entry
>
> Ensure that the short-descriptor page table code is included under the
> SMMU entry, and add Robin as a designated reviewer.
>
> Cc: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
s/Cc/Acked-by/
Robin.
> Signed-off-by: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
> ---
> MAINTAINERS | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index cc2f753cb357..a01bae148b89 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -1793,11 +1793,13 @@ F: drivers/edac/synopsys_edac.c
>
> ARM SMMU DRIVERS
> M: Will Deacon <will.deacon-5wv7dgnIgG8@public.gmane.org>
> +R: Robin Murphy <robin.murphy-5wv7dgnIgG8@public.gmane.org>
> L: linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org (moderated for non-subscribers)
> S: Maintained
> F: drivers/iommu/arm-smmu.c
> F: drivers/iommu/arm-smmu-v3.c
> F: drivers/iommu/io-pgtable-arm.c
> +F: drivers/iommu/io-pgtable-arm-v7s.c
>
> ARM64 PORT (AARCH64 ARCHITECTURE)
> M: Catalin Marinas <catalin.marinas-5wv7dgnIgG8@public.gmane.org>
>
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-02-17 14:34 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-02-17 14:34 UTC (permalink / raw)
To: linux-arm-kernel
On 17/02/16 14:31, Will Deacon wrote:
> Hi Robin,
>
> On Thu, Jan 28, 2016 at 01:41:29PM +0000, Robin Murphy wrote:
>> On 27/01/16 01:16, Yong Wu wrote:
>>> On Tue, 2016-01-26 at 17:13 +0000, Robin Murphy wrote:
>>>> Add a nearly-complete ARMv7 short descriptor implementation, omitting
>>>> only a few legacy and CPU-centric aspects which shouldn't be necessary
>>>> for IOMMU API use anyway.
>>>>
>>>> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
>>>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>>>
>>> This looks good to me.
>>> And actually our disp/vdec/venc works well on your v2.
>>>
>>> You could choose the two below,
>>> Reviewed-by: Yong Wu <yong.wu@mediatek.com>
>>> Tested-by: Yong Wu <yong.wu@mediatek.com>
>>
>> Great, thanks!
>>
>> Will, it seems to make the most sense for Joerg to pick these patches up
>> directly with the MTK series that depends on them - any objections?
>
> As discussed with Joerg, I'll pick these up and send him a pull request
> containing all the pgtable bits. I've taken the liberty of updating
> MAINTAINERS too, assuming you don't mind being a reviewer for this?
Sure, that's fine by me.
> Will
>
> --->8
>
> From bccb08d096703bd790fcf2f934553dd9c8503fc9 Mon Sep 17 00:00:00 2001
> From: Will Deacon <will.deacon@arm.com>
> Date: Wed, 17 Feb 2016 14:18:00 +0000
> Subject: [PATCH] MAINTAINERS: update ARM SMMU entry
>
> Ensure that the short-descriptor page table code is included under the
> SMMU entry, and add Robin as a designated reviewer.
>
> Cc: Robin Murphy <robin.murphy@arm.com>
s/Cc/Acked-by/
Robin.
> Signed-off-by: Will Deacon <will.deacon@arm.com>
> ---
> MAINTAINERS | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index cc2f753cb357..a01bae148b89 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -1793,11 +1793,13 @@ F: drivers/edac/synopsys_edac.c
>
> ARM SMMU DRIVERS
> M: Will Deacon <will.deacon@arm.com>
> +R: Robin Murphy <robin.murphy@arm.com>
> L: linux-arm-kernel at lists.infradead.org (moderated for non-subscribers)
> S: Maintained
> F: drivers/iommu/arm-smmu.c
> F: drivers/iommu/arm-smmu-v3.c
> F: drivers/iommu/io-pgtable-arm.c
> +F: drivers/iommu/io-pgtable-arm-v7s.c
>
> ARM64 PORT (AARCH64 ARCHITECTURE)
> M: Catalin Marinas <catalin.marinas@arm.com>
>
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-01-26 17:13 ` Robin Murphy
@ 2016-03-01 12:01 ` Geert Uytterhoeven
-1 siblings, 0 replies; 46+ messages in thread
From: Geert Uytterhoeven @ 2016-03-01 12:01 UTC (permalink / raw)
To: Robin Murphy
Cc: Will Deacon, yong.wu, iommu, Laurent Pinchart, linux-arm-kernel,
linux-renesas-soc
Hi Robin,
On Tue, Jan 26, 2016 at 6:13 PM, Robin Murphy <robin.murphy@arm.com> wrote:
> Add a nearly-complete ARMv7 short descriptor implementation, omitting
> only a few legacy and CPU-centric aspects which shouldn't be necessary
> for IOMMU API use anyway.
>
> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/iommu/Kconfig | 19 +
> drivers/iommu/Makefile | 1 +
> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
> drivers/iommu/io-pgtable.c | 3 +
> drivers/iommu/io-pgtable.h | 14 +-
> 5 files changed, 885 insertions(+), 1 deletion(-)
> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
>
> diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
> index a1e75cb..dc1aaa5 100644
> --- a/drivers/iommu/Kconfig
> +++ b/drivers/iommu/Kconfig
> @@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
>
> If unsure, say N here.
>
> +config IOMMU_IO_PGTABLE_ARMV7S
> + bool "ARMv7/v8 Short Descriptor Format"
> + select IOMMU_IO_PGTABLE
> + depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
> + help
> + Enable support for the ARM Short-descriptor pagetable format.
> + This supports 32-bit virtual and physical addresses mapped using
> + 2-level tables with 4KB pages/1MB sections, and contiguous entries
> + for 64KB pages/16MB supersections if indicated by the IOMMU driver.
> +
> +config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> + bool "ARMv7s selftests"
> + depends on IOMMU_IO_PGTABLE_ARMV7S
> + help
> + Enable self-tests for ARMv7s page table allocator. This performs
> + a series of page-table consistency checks during boot.
> +
> + If unsure, say N here.
> +
When enabling IOMMU_IO_PGTABLE_ARMV7S_SELFTEST on r8a7791 (dual Cortex A15),
the kernel crashes with:
kernel BUG at mm/slab.c:2536!
Internal error: Oops - BUG: 0 [#1] SMP ARM
Modules linked in:[ 1.299311] Modules linked in:
CPU: 1 PID: 1 Comm: swapper/0 Not tainted
4.5.0-rc6-koelsch-05892-ge7e45ad53ab6795e #2270
Hardware name: Generic R8A7791 (Flattened Device Tree)
task: ef422040 ti: ef442000 task.ti: ef442000
PC is at cache_alloc_refill+0x2a0/0x530
LR is at _raw_spin_unlock+0x8/0xc
pc : [<c02c6928>] lr : [<c067f97c>] psr: 00000093
sp : ef443d88 ip : d0a7b1f0 fp : 02000020
r10: 00000010 r9 : 00000000 r8 : 02088020
r7 : 00000000 r6 : ef5172c0 r5 : 00000000 r4 : ef5d7f40
r3 : 00004001 r2 : 00000001 r1 : 00020001 r0 : ef5172c0
Flags: nzcv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
Control: 30c5387d Table: 40003000 DAC: 55555555
Process swapper/0 (pid: 1, stack limit = 0xef442210)
Stack: (0xef443d88 to 0xef444000)
3d80: ef48b4c0 00000000 ef5172d0 c0e38280 00000000 00001000
3da0: 02088020 60000013 ef5d7f40 c0445010 00000000 ef517340 00001000 c02c6630
3dc0: 00000008 00001000 00000000 00000400 00000002 c04444bc 00000000 00000000
3de0: 024000c0 ef517340 00000000 00004000 00000001 00000001 c0e3b000 ef517340
3e00: 00000000 00001000 00000000 ef5f8000 ef517340 c0445010 00000000 0000000f
3e20: 00001000 c0444e1c 00000000 00000000 ef517350 00000000 ef520f15 ef517340
3e40: ef443ea0 ef517390 00000000 00000000 00001000 c0445010 00000000 ef517390
3e60: 00001000 c0445044 00001000 0000000f ef5f8000 ef443ea0 c0e92a00 c0e92a00
3e80: 00000000 00000000 0000000c c0c18ee4 00001000 0000000f 00000001 00000014
3ea0: 00000001 01111000 00000020 00000020 c0e25bc8 00000000 6f5f806a 00000000
3ec0: 00000020 40004000 800b8204 00000000 00000000 00000000 00000000 00000000
3ee0: c0c3ce94 ef520f00 c0c18da4 00000000 c0e05550 c0e05550 c0e3b000 c0c2d848
3f00: 00000000 c0201760 ef5036c0 c0804b44 c0e5fdd0 00000000 c0e0be00 c0e0be7c
3f20: 00000000 60000013 00000000 efffec51 efffec59 c023893c c09dc204 00000004
3f40: 00000004 c09dce98 000000a0 c09dce98 c0e09e78 00000004 c0c2d834 00000004
3f60: c0c2d838 000000a1 c0c3ce94 c0e3b000 c0e3b000 c0c00d4c 00000004 00000004
3f80: 00000000 c0c0058c 00000000 c067a35c 00000000 00000000 00000000 00000000
3fa0: 00000000 c067a364 00000000 c0206b68 00000000 00000000 00000000 00000000
3fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
3fe0: 00000000 00000000 00000000 00000000 00000013 00000000 ef443ff4 00000000
[<c02c6928>] (cache_alloc_refill) from [<c02c6630>] (kmem_cache_alloc+0x7c/0xd4)
[<c02c6630>] (kmem_cache_alloc) from [<c04444bc>]
(__arm_v7s_alloc_table+0x5c/0x278)
[<c04444bc>] (__arm_v7s_alloc_table) from [<c0444e1c>]
(__arm_v7s_map.constprop.6+0x68/0x25c)
[<c0444e1c>] (__arm_v7s_map.constprop.6) from [<c0445044>]
(arm_v7s_map+0x34/0xa4)
[<c0445044>] (arm_v7s_map) from [<c0c18ee4>] (arm_v7s_do_selftests+0x140/0x418)
[<c0c18ee4>] (arm_v7s_do_selftests) from [<c0201760>]
(do_one_initcall+0x100/0x1b4)
[<c0201760>] (do_one_initcall) from [<c0c00d4c>]
(kernel_init_freeable+0x120/0x1e8)
[<c0c00d4c>] (kernel_init_freeable) from [<c067a364>] (kernel_init+0x8/0xec)
[<c067a364>] (kernel_init) from [<c0206b68>] (ret_from_fork+0x14/0x2c)
Code: 1a000003 e7f001f2 e3130001 0a000000 (e7f001f2)
---[ end trace 190f6f6b84352efd ]---
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
I don't know if r8a7791 supports v7s. Still, I don't expect a crash.
The LPAE selftest runs fine.
On r8a7795 (arm64), both the LPAE and V7S selftests run fine.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-03-01 12:01 ` Geert Uytterhoeven
0 siblings, 0 replies; 46+ messages in thread
From: Geert Uytterhoeven @ 2016-03-01 12:01 UTC (permalink / raw)
To: linux-arm-kernel
Hi Robin,
On Tue, Jan 26, 2016 at 6:13 PM, Robin Murphy <robin.murphy@arm.com> wrote:
> Add a nearly-complete ARMv7 short descriptor implementation, omitting
> only a few legacy and CPU-centric aspects which shouldn't be necessary
> for IOMMU API use anyway.
>
> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> ---
> drivers/iommu/Kconfig | 19 +
> drivers/iommu/Makefile | 1 +
> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
> drivers/iommu/io-pgtable.c | 3 +
> drivers/iommu/io-pgtable.h | 14 +-
> 5 files changed, 885 insertions(+), 1 deletion(-)
> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
>
> diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
> index a1e75cb..dc1aaa5 100644
> --- a/drivers/iommu/Kconfig
> +++ b/drivers/iommu/Kconfig
> @@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
>
> If unsure, say N here.
>
> +config IOMMU_IO_PGTABLE_ARMV7S
> + bool "ARMv7/v8 Short Descriptor Format"
> + select IOMMU_IO_PGTABLE
> + depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
> + help
> + Enable support for the ARM Short-descriptor pagetable format.
> + This supports 32-bit virtual and physical addresses mapped using
> + 2-level tables with 4KB pages/1MB sections, and contiguous entries
> + for 64KB pages/16MB supersections if indicated by the IOMMU driver.
> +
> +config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> + bool "ARMv7s selftests"
> + depends on IOMMU_IO_PGTABLE_ARMV7S
> + help
> + Enable self-tests for ARMv7s page table allocator. This performs
> + a series of page-table consistency checks during boot.
> +
> + If unsure, say N here.
> +
When enabling IOMMU_IO_PGTABLE_ARMV7S_SELFTEST on r8a7791 (dual Cortex A15),
the kernel crashes with:
kernel BUG at mm/slab.c:2536!
Internal error: Oops - BUG: 0 [#1] SMP ARM
Modules linked in:[ 1.299311] Modules linked in:
CPU: 1 PID: 1 Comm: swapper/0 Not tainted
4.5.0-rc6-koelsch-05892-ge7e45ad53ab6795e #2270
Hardware name: Generic R8A7791 (Flattened Device Tree)
task: ef422040 ti: ef442000 task.ti: ef442000
PC is at cache_alloc_refill+0x2a0/0x530
LR is at _raw_spin_unlock+0x8/0xc
pc : [<c02c6928>] lr : [<c067f97c>] psr: 00000093
sp : ef443d88 ip : d0a7b1f0 fp : 02000020
r10: 00000010 r9 : 00000000 r8 : 02088020
r7 : 00000000 r6 : ef5172c0 r5 : 00000000 r4 : ef5d7f40
r3 : 00004001 r2 : 00000001 r1 : 00020001 r0 : ef5172c0
Flags: nzcv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
Control: 30c5387d Table: 40003000 DAC: 55555555
Process swapper/0 (pid: 1, stack limit = 0xef442210)
Stack: (0xef443d88 to 0xef444000)
3d80: ef48b4c0 00000000 ef5172d0 c0e38280 00000000 00001000
3da0: 02088020 60000013 ef5d7f40 c0445010 00000000 ef517340 00001000 c02c6630
3dc0: 00000008 00001000 00000000 00000400 00000002 c04444bc 00000000 00000000
3de0: 024000c0 ef517340 00000000 00004000 00000001 00000001 c0e3b000 ef517340
3e00: 00000000 00001000 00000000 ef5f8000 ef517340 c0445010 00000000 0000000f
3e20: 00001000 c0444e1c 00000000 00000000 ef517350 00000000 ef520f15 ef517340
3e40: ef443ea0 ef517390 00000000 00000000 00001000 c0445010 00000000 ef517390
3e60: 00001000 c0445044 00001000 0000000f ef5f8000 ef443ea0 c0e92a00 c0e92a00
3e80: 00000000 00000000 0000000c c0c18ee4 00001000 0000000f 00000001 00000014
3ea0: 00000001 01111000 00000020 00000020 c0e25bc8 00000000 6f5f806a 00000000
3ec0: 00000020 40004000 800b8204 00000000 00000000 00000000 00000000 00000000
3ee0: c0c3ce94 ef520f00 c0c18da4 00000000 c0e05550 c0e05550 c0e3b000 c0c2d848
3f00: 00000000 c0201760 ef5036c0 c0804b44 c0e5fdd0 00000000 c0e0be00 c0e0be7c
3f20: 00000000 60000013 00000000 efffec51 efffec59 c023893c c09dc204 00000004
3f40: 00000004 c09dce98 000000a0 c09dce98 c0e09e78 00000004 c0c2d834 00000004
3f60: c0c2d838 000000a1 c0c3ce94 c0e3b000 c0e3b000 c0c00d4c 00000004 00000004
3f80: 00000000 c0c0058c 00000000 c067a35c 00000000 00000000 00000000 00000000
3fa0: 00000000 c067a364 00000000 c0206b68 00000000 00000000 00000000 00000000
3fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
3fe0: 00000000 00000000 00000000 00000000 00000013 00000000 ef443ff4 00000000
[<c02c6928>] (cache_alloc_refill) from [<c02c6630>] (kmem_cache_alloc+0x7c/0xd4)
[<c02c6630>] (kmem_cache_alloc) from [<c04444bc>]
(__arm_v7s_alloc_table+0x5c/0x278)
[<c04444bc>] (__arm_v7s_alloc_table) from [<c0444e1c>]
(__arm_v7s_map.constprop.6+0x68/0x25c)
[<c0444e1c>] (__arm_v7s_map.constprop.6) from [<c0445044>]
(arm_v7s_map+0x34/0xa4)
[<c0445044>] (arm_v7s_map) from [<c0c18ee4>] (arm_v7s_do_selftests+0x140/0x418)
[<c0c18ee4>] (arm_v7s_do_selftests) from [<c0201760>]
(do_one_initcall+0x100/0x1b4)
[<c0201760>] (do_one_initcall) from [<c0c00d4c>]
(kernel_init_freeable+0x120/0x1e8)
[<c0c00d4c>] (kernel_init_freeable) from [<c067a364>] (kernel_init+0x8/0xec)
[<c067a364>] (kernel_init) from [<c0206b68>] (ret_from_fork+0x14/0x2c)
Code: 1a000003 e7f001f2 e3130001 0a000000 (e7f001f2)
---[ end trace 190f6f6b84352efd ]---
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
I don't know if r8a7791 supports v7s. Still, I don't expect a crash.
The LPAE selftest runs fine.
On r8a7795 (arm64), both the LPAE and V7S selftests run fine.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-03-01 12:01 ` Geert Uytterhoeven
@ 2016-03-01 13:13 ` Laurent Pinchart
-1 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-03-01 13:13 UTC (permalink / raw)
To: Geert Uytterhoeven
Cc: Robin Murphy, Will Deacon, yong.wu, iommu, Laurent Pinchart,
linux-arm-kernel, linux-renesas-soc
Hi Geert,
On Tuesday 01 March 2016 13:01:39 Geert Uytterhoeven wrote:
> On Tue, Jan 26, 2016 at 6:13 PM, Robin Murphy <robin.murphy@arm.com> wrote:
> > Add a nearly-complete ARMv7 short descriptor implementation, omitting
> > only a few legacy and CPU-centric aspects which shouldn't be necessary
> > for IOMMU API use anyway.
> >
> > Signed-off-by: Yong Wu <yong.wu@mediatek.com>
> > Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> > ---
> >
> > drivers/iommu/Kconfig | 19 +
> > drivers/iommu/Makefile | 1 +
> > drivers/iommu/io-pgtable-arm-v7s.c | 849 ++++++++++++++++++++++++++++++++
> > drivers/iommu/io-pgtable.c | 3 +
> > drivers/iommu/io-pgtable.h | 14 +-
> > 5 files changed, 885 insertions(+), 1 deletion(-)
> > create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
> >
> > diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
> > index a1e75cb..dc1aaa5 100644
> > --- a/drivers/iommu/Kconfig
> > +++ b/drivers/iommu/Kconfig
> > @@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
> >
> > If unsure, say N here.
> >
> > +config IOMMU_IO_PGTABLE_ARMV7S
> > + bool "ARMv7/v8 Short Descriptor Format"
> > + select IOMMU_IO_PGTABLE
> > + depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
> > + help
> > + Enable support for the ARM Short-descriptor pagetable format.
> > + This supports 32-bit virtual and physical addresses mapped using
> > + 2-level tables with 4KB pages/1MB sections, and contiguous
> > entries + for 64KB pages/16MB supersections if indicated by the
> > IOMMU driver. +
> > +config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> > + bool "ARMv7s selftests"
> > + depends on IOMMU_IO_PGTABLE_ARMV7S
> > + help
> > + Enable self-tests for ARMv7s page table allocator. This performs
> > + a series of page-table consistency checks during boot.
> > +
> > + If unsure, say N here.
> > +
>
> When enabling IOMMU_IO_PGTABLE_ARMV7S_SELFTEST on r8a7791 (dual Cortex A15),
> the kernel crashes with:
>
> kernel BUG at mm/slab.c:2536!
> Internal error: Oops - BUG: 0 [#1] SMP ARM
> Modules linked in:[ 1.299311] Modules linked in:
>
> CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> 4.5.0-rc6-koelsch-05892-ge7e45ad53ab6795e #2270
> Hardware name: Generic R8A7791 (Flattened Device Tree)
> task: ef422040 ti: ef442000 task.ti: ef442000
> PC is at cache_alloc_refill+0x2a0/0x530
> LR is at _raw_spin_unlock+0x8/0xc
> pc : [<c02c6928>] lr : [<c067f97c>] psr: 00000093
> sp : ef443d88 ip : d0a7b1f0 fp : 02000020
> r10: 00000010 r9 : 00000000 r8 : 02088020
> r7 : 00000000 r6 : ef5172c0 r5 : 00000000 r4 : ef5d7f40
> r3 : 00004001 r2 : 00000001 r1 : 00020001 r0 : ef5172c0
> Flags: nzcv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
> Control: 30c5387d Table: 40003000 DAC: 55555555
> Process swapper/0 (pid: 1, stack limit = 0xef442210)
> Stack: (0xef443d88 to 0xef444000)
> 3d80: ef48b4c0 00000000 ef5172d0 c0e38280 00000000
> 00001000 3da0: 02088020 60000013 ef5d7f40 c0445010 00000000 ef517340
> 00001000 c02c6630 3dc0: 00000008 00001000 00000000 00000400 00000002
> c04444bc 00000000 00000000 3de0: 024000c0 ef517340 00000000 00004000
> 00000001 00000001 c0e3b000 ef517340 3e00: 00000000 00001000 00000000
> ef5f8000 ef517340 c0445010 00000000 0000000f 3e20: 00001000 c0444e1c
> 00000000 00000000 ef517350 00000000 ef520f15 ef517340 3e40: ef443ea0
> ef517390 00000000 00000000 00001000 c0445010 00000000 ef517390 3e60:
> 00001000 c0445044 00001000 0000000f ef5f8000 ef443ea0 c0e92a00 c0e92a00
> 3e80: 00000000 00000000 0000000c c0c18ee4 00001000 0000000f 00000001
> 00000014 3ea0: 00000001 01111000 00000020 00000020 c0e25bc8 00000000
> 6f5f806a 00000000 3ec0: 00000020 40004000 800b8204 00000000 00000000
> 00000000 00000000 00000000 3ee0: c0c3ce94 ef520f00 c0c18da4 00000000
> c0e05550 c0e05550 c0e3b000 c0c2d848 3f00: 00000000 c0201760 ef5036c0
> c0804b44 c0e5fdd0 00000000 c0e0be00 c0e0be7c 3f20: 00000000 60000013
> 00000000 efffec51 efffec59 c023893c c09dc204 00000004 3f40: 00000004
> c09dce98 000000a0 c09dce98 c0e09e78 00000004 c0c2d834 00000004 3f60:
> c0c2d838 000000a1 c0c3ce94 c0e3b000 c0e3b000 c0c00d4c 00000004 00000004
> 3f80: 00000000 c0c0058c 00000000 c067a35c 00000000 00000000 00000000
> 00000000 3fa0: 00000000 c067a364 00000000 c0206b68 00000000 00000000
> 00000000 00000000 3fc0: 00000000 00000000 00000000 00000000 00000000
> 00000000 00000000 00000000 3fe0: 00000000 00000000 00000000 00000000
> 00000013 00000000 ef443ff4 00000000 [<c02c6928>] (cache_alloc_refill) from
> [<c02c6630>] (kmem_cache_alloc+0x7c/0xd4) [<c02c6630>] (kmem_cache_alloc)
> from [<c04444bc>]
> (__arm_v7s_alloc_table+0x5c/0x278)
> [<c04444bc>] (__arm_v7s_alloc_table) from [<c0444e1c>]
> (__arm_v7s_map.constprop.6+0x68/0x25c)
> [<c0444e1c>] (__arm_v7s_map.constprop.6) from [<c0445044>]
> (arm_v7s_map+0x34/0xa4)
> [<c0445044>] (arm_v7s_map) from [<c0c18ee4>]
> (arm_v7s_do_selftests+0x140/0x418) [<c0c18ee4>] (arm_v7s_do_selftests) from
> [<c0201760>]
> (do_one_initcall+0x100/0x1b4)
> [<c0201760>] (do_one_initcall) from [<c0c00d4c>]
> (kernel_init_freeable+0x120/0x1e8)
> [<c0c00d4c>] (kernel_init_freeable) from [<c067a364>] (kernel_init+0x8/0xec)
> [<c067a364>] (kernel_init) from [<c0206b68>] (ret_from_fork+0x14/0x2c)
> Code: 1a000003 e7f001f2 e3130001 0a000000 (e7f001f2)
> ---[ end trace 190f6f6b84352efd ]---
> Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
>
> I don't know if r8a7791 supports v7s.
The IPMMU does, but that's not implemented in the driver at the moment.
> Still, I don't expect a crash.
I agree.
> The LPAE selftest runs fine.
>
> On r8a7795 (arm64), both the LPAE and V7S selftests run fine.
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-03-01 13:13 ` Laurent Pinchart
0 siblings, 0 replies; 46+ messages in thread
From: Laurent Pinchart @ 2016-03-01 13:13 UTC (permalink / raw)
To: linux-arm-kernel
Hi Geert,
On Tuesday 01 March 2016 13:01:39 Geert Uytterhoeven wrote:
> On Tue, Jan 26, 2016 at 6:13 PM, Robin Murphy <robin.murphy@arm.com> wrote:
> > Add a nearly-complete ARMv7 short descriptor implementation, omitting
> > only a few legacy and CPU-centric aspects which shouldn't be necessary
> > for IOMMU API use anyway.
> >
> > Signed-off-by: Yong Wu <yong.wu@mediatek.com>
> > Signed-off-by: Robin Murphy <robin.murphy@arm.com>
> > ---
> >
> > drivers/iommu/Kconfig | 19 +
> > drivers/iommu/Makefile | 1 +
> > drivers/iommu/io-pgtable-arm-v7s.c | 849 ++++++++++++++++++++++++++++++++
> > drivers/iommu/io-pgtable.c | 3 +
> > drivers/iommu/io-pgtable.h | 14 +-
> > 5 files changed, 885 insertions(+), 1 deletion(-)
> > create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
> >
> > diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
> > index a1e75cb..dc1aaa5 100644
> > --- a/drivers/iommu/Kconfig
> > +++ b/drivers/iommu/Kconfig
> > @@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
> >
> > If unsure, say N here.
> >
> > +config IOMMU_IO_PGTABLE_ARMV7S
> > + bool "ARMv7/v8 Short Descriptor Format"
> > + select IOMMU_IO_PGTABLE
> > + depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
> > + help
> > + Enable support for the ARM Short-descriptor pagetable format.
> > + This supports 32-bit virtual and physical addresses mapped using
> > + 2-level tables with 4KB pages/1MB sections, and contiguous
> > entries + for 64KB pages/16MB supersections if indicated by the
> > IOMMU driver. +
> > +config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
> > + bool "ARMv7s selftests"
> > + depends on IOMMU_IO_PGTABLE_ARMV7S
> > + help
> > + Enable self-tests for ARMv7s page table allocator. This performs
> > + a series of page-table consistency checks during boot.
> > +
> > + If unsure, say N here.
> > +
>
> When enabling IOMMU_IO_PGTABLE_ARMV7S_SELFTEST on r8a7791 (dual Cortex A15),
> the kernel crashes with:
>
> kernel BUG at mm/slab.c:2536!
> Internal error: Oops - BUG: 0 [#1] SMP ARM
> Modules linked in:[ 1.299311] Modules linked in:
>
> CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> 4.5.0-rc6-koelsch-05892-ge7e45ad53ab6795e #2270
> Hardware name: Generic R8A7791 (Flattened Device Tree)
> task: ef422040 ti: ef442000 task.ti: ef442000
> PC is at cache_alloc_refill+0x2a0/0x530
> LR is at _raw_spin_unlock+0x8/0xc
> pc : [<c02c6928>] lr : [<c067f97c>] psr: 00000093
> sp : ef443d88 ip : d0a7b1f0 fp : 02000020
> r10: 00000010 r9 : 00000000 r8 : 02088020
> r7 : 00000000 r6 : ef5172c0 r5 : 00000000 r4 : ef5d7f40
> r3 : 00004001 r2 : 00000001 r1 : 00020001 r0 : ef5172c0
> Flags: nzcv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
> Control: 30c5387d Table: 40003000 DAC: 55555555
> Process swapper/0 (pid: 1, stack limit = 0xef442210)
> Stack: (0xef443d88 to 0xef444000)
> 3d80: ef48b4c0 00000000 ef5172d0 c0e38280 00000000
> 00001000 3da0: 02088020 60000013 ef5d7f40 c0445010 00000000 ef517340
> 00001000 c02c6630 3dc0: 00000008 00001000 00000000 00000400 00000002
> c04444bc 00000000 00000000 3de0: 024000c0 ef517340 00000000 00004000
> 00000001 00000001 c0e3b000 ef517340 3e00: 00000000 00001000 00000000
> ef5f8000 ef517340 c0445010 00000000 0000000f 3e20: 00001000 c0444e1c
> 00000000 00000000 ef517350 00000000 ef520f15 ef517340 3e40: ef443ea0
> ef517390 00000000 00000000 00001000 c0445010 00000000 ef517390 3e60:
> 00001000 c0445044 00001000 0000000f ef5f8000 ef443ea0 c0e92a00 c0e92a00
> 3e80: 00000000 00000000 0000000c c0c18ee4 00001000 0000000f 00000001
> 00000014 3ea0: 00000001 01111000 00000020 00000020 c0e25bc8 00000000
> 6f5f806a 00000000 3ec0: 00000020 40004000 800b8204 00000000 00000000
> 00000000 00000000 00000000 3ee0: c0c3ce94 ef520f00 c0c18da4 00000000
> c0e05550 c0e05550 c0e3b000 c0c2d848 3f00: 00000000 c0201760 ef5036c0
> c0804b44 c0e5fdd0 00000000 c0e0be00 c0e0be7c 3f20: 00000000 60000013
> 00000000 efffec51 efffec59 c023893c c09dc204 00000004 3f40: 00000004
> c09dce98 000000a0 c09dce98 c0e09e78 00000004 c0c2d834 00000004 3f60:
> c0c2d838 000000a1 c0c3ce94 c0e3b000 c0e3b000 c0c00d4c 00000004 00000004
> 3f80: 00000000 c0c0058c 00000000 c067a35c 00000000 00000000 00000000
> 00000000 3fa0: 00000000 c067a364 00000000 c0206b68 00000000 00000000
> 00000000 00000000 3fc0: 00000000 00000000 00000000 00000000 00000000
> 00000000 00000000 00000000 3fe0: 00000000 00000000 00000000 00000000
> 00000013 00000000 ef443ff4 00000000 [<c02c6928>] (cache_alloc_refill) from
> [<c02c6630>] (kmem_cache_alloc+0x7c/0xd4) [<c02c6630>] (kmem_cache_alloc)
> from [<c04444bc>]
> (__arm_v7s_alloc_table+0x5c/0x278)
> [<c04444bc>] (__arm_v7s_alloc_table) from [<c0444e1c>]
> (__arm_v7s_map.constprop.6+0x68/0x25c)
> [<c0444e1c>] (__arm_v7s_map.constprop.6) from [<c0445044>]
> (arm_v7s_map+0x34/0xa4)
> [<c0445044>] (arm_v7s_map) from [<c0c18ee4>]
> (arm_v7s_do_selftests+0x140/0x418) [<c0c18ee4>] (arm_v7s_do_selftests) from
> [<c0201760>]
> (do_one_initcall+0x100/0x1b4)
> [<c0201760>] (do_one_initcall) from [<c0c00d4c>]
> (kernel_init_freeable+0x120/0x1e8)
> [<c0c00d4c>] (kernel_init_freeable) from [<c067a364>] (kernel_init+0x8/0xec)
> [<c067a364>] (kernel_init) from [<c0206b68>] (ret_from_fork+0x14/0x2c)
> Code: 1a000003 e7f001f2 e3130001 0a000000 (e7f001f2)
> ---[ end trace 190f6f6b84352efd ]---
> Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
>
> I don't know if r8a7791 supports v7s.
The IPMMU does, but that's not implemented in the driver at the moment.
> Still, I don't expect a crash.
I agree.
> The LPAE selftest runs fine.
>
> On r8a7795 (arm64), both the LPAE and V7S selftests run fine.
--
Regards,
Laurent Pinchart
^ permalink raw reply [flat|nested] 46+ messages in thread
* Re: [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
2016-03-01 12:01 ` Geert Uytterhoeven
@ 2016-03-01 13:51 ` Robin Murphy
-1 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-03-01 13:51 UTC (permalink / raw)
To: Geert Uytterhoeven
Cc: Will Deacon, yong.wu, iommu, Laurent Pinchart, linux-arm-kernel,
linux-renesas-soc
Hi Geert,
On 01/03/16 12:01, Geert Uytterhoeven wrote:
> Hi Robin,
>
> On Tue, Jan 26, 2016 at 6:13 PM, Robin Murphy <robin.murphy@arm.com> wrote:
>> Add a nearly-complete ARMv7 short descriptor implementation, omitting
>> only a few legacy and CPU-centric aspects which shouldn't be necessary
>> for IOMMU API use anyway.
>>
>> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>> ---
>> drivers/iommu/Kconfig | 19 +
>> drivers/iommu/Makefile | 1 +
>> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
>> drivers/iommu/io-pgtable.c | 3 +
>> drivers/iommu/io-pgtable.h | 14 +-
>> 5 files changed, 885 insertions(+), 1 deletion(-)
>> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
>>
>> diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
>> index a1e75cb..dc1aaa5 100644
>> --- a/drivers/iommu/Kconfig
>> +++ b/drivers/iommu/Kconfig
>> @@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
>>
>> If unsure, say N here.
>>
>> +config IOMMU_IO_PGTABLE_ARMV7S
>> + bool "ARMv7/v8 Short Descriptor Format"
>> + select IOMMU_IO_PGTABLE
>> + depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
>> + help
>> + Enable support for the ARM Short-descriptor pagetable format.
>> + This supports 32-bit virtual and physical addresses mapped using
>> + 2-level tables with 4KB pages/1MB sections, and contiguous entries
>> + for 64KB pages/16MB supersections if indicated by the IOMMU driver.
>> +
>> +config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
>> + bool "ARMv7s selftests"
>> + depends on IOMMU_IO_PGTABLE_ARMV7S
>> + help
>> + Enable self-tests for ARMv7s page table allocator. This performs
>> + a series of page-table consistency checks during boot.
>> +
>> + If unsure, say N here.
>> +
>
> When enabling IOMMU_IO_PGTABLE_ARMV7S_SELFTEST on r8a7791 (dual Cortex A15),
> the kernel crashes with:
>
> kernel BUG at mm/slab.c:2536!
> Internal error: Oops - BUG: 0 [#1] SMP ARM
> Modules linked in:[ 1.299311] Modules linked in:
>
> CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> 4.5.0-rc6-koelsch-05892-ge7e45ad53ab6795e #2270
> Hardware name: Generic R8A7791 (Flattened Device Tree)
> task: ef422040 ti: ef442000 task.ti: ef442000
> PC is at cache_alloc_refill+0x2a0/0x530
> LR is at _raw_spin_unlock+0x8/0xc
> pc : [<c02c6928>] lr : [<c067f97c>] psr: 00000093
> sp : ef443d88 ip : d0a7b1f0 fp : 02000020
> r10: 00000010 r9 : 00000000 r8 : 02088020
> r7 : 00000000 r6 : ef5172c0 r5 : 00000000 r4 : ef5d7f40
> r3 : 00004001 r2 : 00000001 r1 : 00020001 r0 : ef5172c0
> Flags: nzcv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
> Control: 30c5387d Table: 40003000 DAC: 55555555
> Process swapper/0 (pid: 1, stack limit = 0xef442210)
> Stack: (0xef443d88 to 0xef444000)
> 3d80: ef48b4c0 00000000 ef5172d0 c0e38280 00000000 00001000
> 3da0: 02088020 60000013 ef5d7f40 c0445010 00000000 ef517340 00001000 c02c6630
> 3dc0: 00000008 00001000 00000000 00000400 00000002 c04444bc 00000000 00000000
> 3de0: 024000c0 ef517340 00000000 00004000 00000001 00000001 c0e3b000 ef517340
> 3e00: 00000000 00001000 00000000 ef5f8000 ef517340 c0445010 00000000 0000000f
> 3e20: 00001000 c0444e1c 00000000 00000000 ef517350 00000000 ef520f15 ef517340
> 3e40: ef443ea0 ef517390 00000000 00000000 00001000 c0445010 00000000 ef517390
> 3e60: 00001000 c0445044 00001000 0000000f ef5f8000 ef443ea0 c0e92a00 c0e92a00
> 3e80: 00000000 00000000 0000000c c0c18ee4 00001000 0000000f 00000001 00000014
> 3ea0: 00000001 01111000 00000020 00000020 c0e25bc8 00000000 6f5f806a 00000000
> 3ec0: 00000020 40004000 800b8204 00000000 00000000 00000000 00000000 00000000
> 3ee0: c0c3ce94 ef520f00 c0c18da4 00000000 c0e05550 c0e05550 c0e3b000 c0c2d848
> 3f00: 00000000 c0201760 ef5036c0 c0804b44 c0e5fdd0 00000000 c0e0be00 c0e0be7c
> 3f20: 00000000 60000013 00000000 efffec51 efffec59 c023893c c09dc204 00000004
> 3f40: 00000004 c09dce98 000000a0 c09dce98 c0e09e78 00000004 c0c2d834 00000004
> 3f60: c0c2d838 000000a1 c0c3ce94 c0e3b000 c0e3b000 c0c00d4c 00000004 00000004
> 3f80: 00000000 c0c0058c 00000000 c067a35c 00000000 00000000 00000000 00000000
> 3fa0: 00000000 c067a364 00000000 c0206b68 00000000 00000000 00000000 00000000
> 3fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> 3fe0: 00000000 00000000 00000000 00000000 00000013 00000000 ef443ff4 00000000
> [<c02c6928>] (cache_alloc_refill) from [<c02c6630>] (kmem_cache_alloc+0x7c/0xd4)
> [<c02c6630>] (kmem_cache_alloc) from [<c04444bc>]
> (__arm_v7s_alloc_table+0x5c/0x278)
> [<c04444bc>] (__arm_v7s_alloc_table) from [<c0444e1c>]
> (__arm_v7s_map.constprop.6+0x68/0x25c)
> [<c0444e1c>] (__arm_v7s_map.constprop.6) from [<c0445044>]
> (arm_v7s_map+0x34/0xa4)
> [<c0445044>] (arm_v7s_map) from [<c0c18ee4>] (arm_v7s_do_selftests+0x140/0x418)
> [<c0c18ee4>] (arm_v7s_do_selftests) from [<c0201760>]
> (do_one_initcall+0x100/0x1b4)
> [<c0201760>] (do_one_initcall) from [<c0c00d4c>]
> (kernel_init_freeable+0x120/0x1e8)
> [<c0c00d4c>] (kernel_init_freeable) from [<c067a364>] (kernel_init+0x8/0xec)
> [<c067a364>] (kernel_init) from [<c0206b68>] (ret_from_fork+0x14/0x2c)
> Code: 1a000003 e7f001f2 e3130001 0a000000 (e7f001f2)
> ---[ end trace 190f6f6b84352efd ]---
> Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
>
> I don't know if r8a7791 supports v7s. Still, I don't expect a crash.
> The LPAE selftest runs fine.
>
> On r8a7795 (arm64), both the LPAE and V7S selftests run fine.
The hardware itself shouldn't matter, as the selftests run without
reference to any device at all. At a glance this looks like a difference
in behaviour between SLAB and SLUB - I see shmobile_defconfig sets
CONFIG_SLAB=y. I'll reproduce it locally to make sure.
Thanks for the report,
Robin.
> Gr{oetje,eeting}s,
>
> Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
>
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
> -- Linus Torvalds
>
^ permalink raw reply [flat|nested] 46+ messages in thread
* [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support
@ 2016-03-01 13:51 ` Robin Murphy
0 siblings, 0 replies; 46+ messages in thread
From: Robin Murphy @ 2016-03-01 13:51 UTC (permalink / raw)
To: linux-arm-kernel
Hi Geert,
On 01/03/16 12:01, Geert Uytterhoeven wrote:
> Hi Robin,
>
> On Tue, Jan 26, 2016 at 6:13 PM, Robin Murphy <robin.murphy@arm.com> wrote:
>> Add a nearly-complete ARMv7 short descriptor implementation, omitting
>> only a few legacy and CPU-centric aspects which shouldn't be necessary
>> for IOMMU API use anyway.
>>
>> Signed-off-by: Yong Wu <yong.wu@mediatek.com>
>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>> ---
>> drivers/iommu/Kconfig | 19 +
>> drivers/iommu/Makefile | 1 +
>> drivers/iommu/io-pgtable-arm-v7s.c | 849 +++++++++++++++++++++++++++++++++++++
>> drivers/iommu/io-pgtable.c | 3 +
>> drivers/iommu/io-pgtable.h | 14 +-
>> 5 files changed, 885 insertions(+), 1 deletion(-)
>> create mode 100644 drivers/iommu/io-pgtable-arm-v7s.c
>>
>> diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
>> index a1e75cb..dc1aaa5 100644
>> --- a/drivers/iommu/Kconfig
>> +++ b/drivers/iommu/Kconfig
>> @@ -39,6 +39,25 @@ config IOMMU_IO_PGTABLE_LPAE_SELFTEST
>>
>> If unsure, say N here.
>>
>> +config IOMMU_IO_PGTABLE_ARMV7S
>> + bool "ARMv7/v8 Short Descriptor Format"
>> + select IOMMU_IO_PGTABLE
>> + depends on HAS_DMA && (ARM || ARM64 || COMPILE_TEST)
>> + help
>> + Enable support for the ARM Short-descriptor pagetable format.
>> + This supports 32-bit virtual and physical addresses mapped using
>> + 2-level tables with 4KB pages/1MB sections, and contiguous entries
>> + for 64KB pages/16MB supersections if indicated by the IOMMU driver.
>> +
>> +config IOMMU_IO_PGTABLE_ARMV7S_SELFTEST
>> + bool "ARMv7s selftests"
>> + depends on IOMMU_IO_PGTABLE_ARMV7S
>> + help
>> + Enable self-tests for ARMv7s page table allocator. This performs
>> + a series of page-table consistency checks during boot.
>> +
>> + If unsure, say N here.
>> +
>
> When enabling IOMMU_IO_PGTABLE_ARMV7S_SELFTEST on r8a7791 (dual Cortex A15),
> the kernel crashes with:
>
> kernel BUG at mm/slab.c:2536!
> Internal error: Oops - BUG: 0 [#1] SMP ARM
> Modules linked in:[ 1.299311] Modules linked in:
>
> CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> 4.5.0-rc6-koelsch-05892-ge7e45ad53ab6795e #2270
> Hardware name: Generic R8A7791 (Flattened Device Tree)
> task: ef422040 ti: ef442000 task.ti: ef442000
> PC is at cache_alloc_refill+0x2a0/0x530
> LR is at _raw_spin_unlock+0x8/0xc
> pc : [<c02c6928>] lr : [<c067f97c>] psr: 00000093
> sp : ef443d88 ip : d0a7b1f0 fp : 02000020
> r10: 00000010 r9 : 00000000 r8 : 02088020
> r7 : 00000000 r6 : ef5172c0 r5 : 00000000 r4 : ef5d7f40
> r3 : 00004001 r2 : 00000001 r1 : 00020001 r0 : ef5172c0
> Flags: nzcv IRQs off FIQs on Mode SVC_32 ISA ARM Segment kernel
> Control: 30c5387d Table: 40003000 DAC: 55555555
> Process swapper/0 (pid: 1, stack limit = 0xef442210)
> Stack: (0xef443d88 to 0xef444000)
> 3d80: ef48b4c0 00000000 ef5172d0 c0e38280 00000000 00001000
> 3da0: 02088020 60000013 ef5d7f40 c0445010 00000000 ef517340 00001000 c02c6630
> 3dc0: 00000008 00001000 00000000 00000400 00000002 c04444bc 00000000 00000000
> 3de0: 024000c0 ef517340 00000000 00004000 00000001 00000001 c0e3b000 ef517340
> 3e00: 00000000 00001000 00000000 ef5f8000 ef517340 c0445010 00000000 0000000f
> 3e20: 00001000 c0444e1c 00000000 00000000 ef517350 00000000 ef520f15 ef517340
> 3e40: ef443ea0 ef517390 00000000 00000000 00001000 c0445010 00000000 ef517390
> 3e60: 00001000 c0445044 00001000 0000000f ef5f8000 ef443ea0 c0e92a00 c0e92a00
> 3e80: 00000000 00000000 0000000c c0c18ee4 00001000 0000000f 00000001 00000014
> 3ea0: 00000001 01111000 00000020 00000020 c0e25bc8 00000000 6f5f806a 00000000
> 3ec0: 00000020 40004000 800b8204 00000000 00000000 00000000 00000000 00000000
> 3ee0: c0c3ce94 ef520f00 c0c18da4 00000000 c0e05550 c0e05550 c0e3b000 c0c2d848
> 3f00: 00000000 c0201760 ef5036c0 c0804b44 c0e5fdd0 00000000 c0e0be00 c0e0be7c
> 3f20: 00000000 60000013 00000000 efffec51 efffec59 c023893c c09dc204 00000004
> 3f40: 00000004 c09dce98 000000a0 c09dce98 c0e09e78 00000004 c0c2d834 00000004
> 3f60: c0c2d838 000000a1 c0c3ce94 c0e3b000 c0e3b000 c0c00d4c 00000004 00000004
> 3f80: 00000000 c0c0058c 00000000 c067a35c 00000000 00000000 00000000 00000000
> 3fa0: 00000000 c067a364 00000000 c0206b68 00000000 00000000 00000000 00000000
> 3fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> 3fe0: 00000000 00000000 00000000 00000000 00000013 00000000 ef443ff4 00000000
> [<c02c6928>] (cache_alloc_refill) from [<c02c6630>] (kmem_cache_alloc+0x7c/0xd4)
> [<c02c6630>] (kmem_cache_alloc) from [<c04444bc>]
> (__arm_v7s_alloc_table+0x5c/0x278)
> [<c04444bc>] (__arm_v7s_alloc_table) from [<c0444e1c>]
> (__arm_v7s_map.constprop.6+0x68/0x25c)
> [<c0444e1c>] (__arm_v7s_map.constprop.6) from [<c0445044>]
> (arm_v7s_map+0x34/0xa4)
> [<c0445044>] (arm_v7s_map) from [<c0c18ee4>] (arm_v7s_do_selftests+0x140/0x418)
> [<c0c18ee4>] (arm_v7s_do_selftests) from [<c0201760>]
> (do_one_initcall+0x100/0x1b4)
> [<c0201760>] (do_one_initcall) from [<c0c00d4c>]
> (kernel_init_freeable+0x120/0x1e8)
> [<c0c00d4c>] (kernel_init_freeable) from [<c067a364>] (kernel_init+0x8/0xec)
> [<c067a364>] (kernel_init) from [<c0206b68>] (ret_from_fork+0x14/0x2c)
> Code: 1a000003 e7f001f2 e3130001 0a000000 (e7f001f2)
> ---[ end trace 190f6f6b84352efd ]---
> Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
>
> I don't know if r8a7791 supports v7s. Still, I don't expect a crash.
> The LPAE selftest runs fine.
>
> On r8a7795 (arm64), both the LPAE and V7S selftests run fine.
The hardware itself shouldn't matter, as the selftests run without
reference to any device at all. At a glance this looks like a difference
in behaviour between SLAB and SLUB - I see shmobile_defconfig sets
CONFIG_SLAB=y. I'll reproduce it locally to make sure.
Thanks for the report,
Robin.
> Gr{oetje,eeting}s,
>
> Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org
>
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
> -- Linus Torvalds
>
^ permalink raw reply [flat|nested] 46+ messages in thread
end of thread, other threads:[~2016-03-01 13:51 UTC | newest]
Thread overview: 46+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-01-26 17:13 [PATCH v3 0/3] io-pgtable ARM short descriptor format Robin Murphy
2016-01-26 17:13 ` Robin Murphy
[not found] ` <cover.1453723096.git.robin.murphy-5wv7dgnIgG8@public.gmane.org>
2016-01-26 17:13 ` [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support Robin Murphy
2016-01-26 17:13 ` Robin Murphy
[not found] ` <066c4af0e29c9cc658bd5e48edc5e2d7ba7140a1.1453723096.git.robin.murphy-5wv7dgnIgG8@public.gmane.org>
2016-01-27 1:16 ` Yong Wu
2016-01-27 1:16 ` Yong Wu
2016-01-28 13:41 ` Robin Murphy
2016-01-28 13:41 ` Robin Murphy
[not found] ` <56AA1A89.6080706-5wv7dgnIgG8@public.gmane.org>
2016-02-17 14:31 ` Will Deacon
2016-02-17 14:31 ` Will Deacon
[not found] ` <20160217143106.GC1459-5wv7dgnIgG8@public.gmane.org>
2016-02-17 14:34 ` Robin Murphy
2016-02-17 14:34 ` Robin Murphy
2016-02-10 15:48 ` Will Deacon
2016-02-10 15:48 ` Will Deacon
[not found] ` <20160210154815.GO1052-5wv7dgnIgG8@public.gmane.org>
2016-02-11 16:13 ` [PATCH 4/3] iommu/io-pgtable: Rationalise quirk handling Robin Murphy
2016-02-11 16:13 ` Robin Murphy
[not found] ` <ea1f3d3feec0dc9c18d43393678370b50b1275f5.1455206840.git.robin.murphy-5wv7dgnIgG8@public.gmane.org>
2016-02-11 20:39 ` Laurent Pinchart
2016-02-11 20:39 ` Laurent Pinchart
2016-02-12 12:08 ` Will Deacon
2016-02-12 12:08 ` Will Deacon
[not found] ` <20160212120857.GN25087-5wv7dgnIgG8@public.gmane.org>
2016-02-12 12:32 ` Laurent Pinchart
2016-02-12 12:32 ` Laurent Pinchart
2016-02-12 15:23 ` Robin Murphy
2016-02-12 15:23 ` Robin Murphy
[not found] ` <56BDF8D8.4010200-5wv7dgnIgG8@public.gmane.org>
2016-02-12 15:55 ` Will Deacon
2016-02-12 15:55 ` Will Deacon
2016-03-01 12:01 ` [PATCH v3 1/3] iommu/io-pgtable: Add ARMv7 short descriptor support Geert Uytterhoeven
2016-03-01 12:01 ` Geert Uytterhoeven
2016-03-01 13:13 ` Laurent Pinchart
2016-03-01 13:13 ` Laurent Pinchart
2016-03-01 13:51 ` Robin Murphy
2016-03-01 13:51 ` Robin Murphy
2016-01-26 17:13 ` [PATCH v3 2/3] iommu/io-pgtable: Add helper functions for TLB ops Robin Murphy
2016-01-26 17:13 ` Robin Murphy
2016-01-26 17:13 ` [PATCH v3 3/3] iommu/io-pgtable: Avoid redundant TLB syncs Robin Murphy
2016-01-26 17:13 ` Robin Murphy
[not found] ` <81a5e784494a04713e152659c95b34e92d944f32.1453723096.git.robin.murphy-5wv7dgnIgG8@public.gmane.org>
2016-02-10 14:21 ` Will Deacon
2016-02-10 14:21 ` Will Deacon
2016-02-12 17:09 ` [PATCH v2] iommu/io-pgtable: Rationalise quirk handling Robin Murphy
2016-02-12 17:09 ` Robin Murphy
[not found] ` <62860b753f304beef7b4b6f1e13d7094252f2f37.1455294748.git.robin.murphy-5wv7dgnIgG8@public.gmane.org>
2016-02-12 17:24 ` Laurent Pinchart
2016-02-12 17:24 ` Laurent Pinchart
2016-02-12 17:25 ` Robin Murphy
2016-02-12 17:25 ` Robin Murphy
2016-02-12 17:27 ` Will Deacon
2016-02-12 17:27 ` Will Deacon
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.