lkml.org 
[lkml]   [2018]   [Oct]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH 3/3] staging: gasket: page_table: add mapping flags
Date
From: Nick Ewalt <nicholasewalt@google.com>

This allows for more precise dma_direction in the dma_map_page requests.
Also leaves room for adding more flags later.

Signed-off-by: Nick Ewalt <nicholasewalt@google.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
---
drivers/staging/gasket/gasket.h | 33 ++++++++
drivers/staging/gasket/gasket_ioctl.c | 62 +++++++++++----
drivers/staging/gasket/gasket_page_table.c | 87 ++++++++++++++--------
drivers/staging/gasket/gasket_page_table.h | 4 +-
4 files changed, 141 insertions(+), 45 deletions(-)

diff --git a/drivers/staging/gasket/gasket.h b/drivers/staging/gasket/gasket.h
index a0f065c517a52..93e7af1551975 100644
--- a/drivers/staging/gasket/gasket.h
+++ b/drivers/staging/gasket/gasket.h
@@ -37,6 +37,31 @@ struct gasket_page_table_ioctl {
u64 device_address;
};

+/*
+ * Structure for ioctl mapping buffers with flags when using the Gasket
+ * page_table module.
+ */
+struct gasket_page_table_ioctl_flags {
+ struct gasket_page_table_ioctl base;
+ /*
+ * Flags indicating status and attribute requests from the host.
+ * NOTE: STATUS bit does not need to be set in this request.
+ * Set RESERVED bits to 0 to ensure backwards compatibility.
+ *
+ * Bitfields:
+ * [0] - STATUS: indicates if this entry/slot is free
+ * 0 = PTE_FREE
+ * 1 = PTE_INUSE
+ * [2:1] - DMA_DIRECTION: dma_data_direction requested by host
+ * 00 = DMA_BIDIRECTIONAL
+ * 01 = DMA_TO_DEVICE
+ * 10 = DMA_FROM_DEVICE
+ * 11 = DMA_NONE
+ * [31:3] - RESERVED
+ */
+ u32 flags;
+};
+
/*
* Common structure for ioctls mapping and unmapping buffers when using the
* Gasket page_table module.
@@ -119,4 +144,12 @@ struct gasket_coherent_alloc_config_ioctl {
#define GASKET_IOCTL_CONFIG_COHERENT_ALLOCATOR \
_IOWR(GASKET_IOCTL_BASE, 11, struct gasket_coherent_alloc_config_ioctl)

+/*
+ * Tells the kernel to map size bytes at host_address to device_address in
+ * page_table_index page table. Passes flags to indicate additional attribute
+ * requests for the mapped memory.
+ */
+#define GASKET_IOCTL_MAP_BUFFER_FLAGS \
+ _IOW(GASKET_IOCTL_BASE, 12, struct gasket_page_table_ioctl_flags)
+
#endif /* __GASKET_H__ */
diff --git a/drivers/staging/gasket/gasket_ioctl.c b/drivers/staging/gasket/gasket_ioctl.c
index 0ca48e688818f..e80f38509f47c 100644
--- a/drivers/staging/gasket/gasket_ioctl.c
+++ b/drivers/staging/gasket/gasket_ioctl.c
@@ -20,6 +20,7 @@
#define trace_gasket_ioctl_integer_data(x)
#define trace_gasket_ioctl_eventfd_data(x, ...)
#define trace_gasket_ioctl_page_table_data(x, ...)
+#define trace_gasket_ioctl_page_table_flags_data(x, ...)
#define trace_gasket_ioctl_config_coherent_allocator(x, ...)
#endif

@@ -130,29 +131,59 @@ static int gasket_partition_page_table(
}

/* Map a userspace buffer to a device virtual address. */
+static int gasket_map_buffers_common(
+ struct gasket_dev *gasket_dev,
+ struct gasket_page_table_ioctl_flags *pibuf)
+{
+ if (pibuf->base.page_table_index >= gasket_dev->num_page_tables)
+ return -EFAULT;
+
+ if (gasket_page_table_are_addrs_bad(gasket_dev->page_table[pibuf->base.page_table_index],
+ pibuf->base.host_address,
+ pibuf->base.device_address,
+ pibuf->base.size))
+ return -EINVAL;
+
+ return gasket_page_table_map(gasket_dev->page_table[pibuf->base.page_table_index],
+ pibuf->base.host_address,
+ pibuf->base.device_address,
+ pibuf->base.size / PAGE_SIZE,
+ pibuf->flags);
+}
+
static int gasket_map_buffers(struct gasket_dev *gasket_dev,
struct gasket_page_table_ioctl __user *argp)
{
- struct gasket_page_table_ioctl ibuf;
+ struct gasket_page_table_ioctl_flags ibuf;

- if (copy_from_user(&ibuf, argp, sizeof(struct gasket_page_table_ioctl)))
+ if (copy_from_user(&ibuf.base, argp, sizeof(struct gasket_page_table_ioctl)))
return -EFAULT;

- trace_gasket_ioctl_page_table_data(ibuf.page_table_index, ibuf.size,
- ibuf.host_address,
- ibuf.device_address);
+ ibuf.flags = 0;

- if (ibuf.page_table_index >= gasket_dev->num_page_tables)
+ trace_gasket_ioctl_page_table_data(ibuf.base.page_table_index,
+ ibuf.base.size,
+ ibuf.base.host_address,
+ ibuf.base.device_address);
+
+ return gasket_map_buffers_common(gasket_dev, &ibuf);
+}
+
+static int gasket_map_buffers_flags(struct gasket_dev *gasket_dev,
+ struct gasket_page_table_ioctl_flags __user *argp)
+{
+ struct gasket_page_table_ioctl_flags ibuf;
+
+ if (copy_from_user(&ibuf, argp, sizeof(struct gasket_page_table_ioctl_flags)))
return -EFAULT;

- if (gasket_page_table_are_addrs_bad(gasket_dev->page_table[ibuf.page_table_index],
- ibuf.host_address,
- ibuf.device_address, ibuf.size))
- return -EINVAL;
+ trace_gasket_ioctl_page_table_flags_data(ibuf.base.page_table_index,
+ ibuf.base.size,
+ ibuf.base.host_address,
+ ibuf.base.device_address,
+ ibuf.flags);

- return gasket_page_table_map(gasket_dev->page_table[ibuf.page_table_index],
- ibuf.host_address, ibuf.device_address,
- ibuf.size / PAGE_SIZE);
+ return gasket_map_buffers_common(gasket_dev, &ibuf);
}

/* Unmap a userspace buffer from a device virtual address. */
@@ -252,6 +283,7 @@ static bool gasket_ioctl_check_permissions(struct file *filp, uint cmd)
return alive && write;

case GASKET_IOCTL_MAP_BUFFER:
+ case GASKET_IOCTL_MAP_BUFFER_FLAGS:
case GASKET_IOCTL_UNMAP_BUFFER:
return alive && write;

@@ -336,6 +368,9 @@ long gasket_handle_ioctl(struct file *filp, uint cmd, void __user *argp)
case GASKET_IOCTL_MAP_BUFFER:
retval = gasket_map_buffers(gasket_dev, argp);
break;
+ case GASKET_IOCTL_MAP_BUFFER_FLAGS:
+ retval = gasket_map_buffers_flags(gasket_dev, argp);
+ break;
case GASKET_IOCTL_CONFIG_COHERENT_ALLOCATOR:
retval = gasket_config_coherent_allocator(gasket_dev, argp);
break;
@@ -381,6 +416,7 @@ long gasket_is_supported_ioctl(uint cmd)
case GASKET_IOCTL_PAGE_TABLE_SIZE:
case GASKET_IOCTL_SIMPLE_PAGE_TABLE_SIZE:
case GASKET_IOCTL_MAP_BUFFER:
+ case GASKET_IOCTL_MAP_BUFFER_FLAGS:
case GASKET_IOCTL_UNMAP_BUFFER:
case GASKET_IOCTL_CLEAR_INTERRUPT_COUNTS:
case GASKET_IOCTL_CONFIG_COHERENT_ALLOCATOR:
diff --git a/drivers/staging/gasket/gasket_page_table.c b/drivers/staging/gasket/gasket_page_table.c
index b7d460cf15fbc..06e188f5b905c 100644
--- a/drivers/staging/gasket/gasket_page_table.c
+++ b/drivers/staging/gasket/gasket_page_table.c
@@ -87,6 +87,19 @@
*/
#define GASKET_EXTENDED_LVL1_SHIFT 12

+/*
+ * Utilities for accessing flags bitfields.
+ */
+#define MASK(field) (((1u << field##_WIDTH) - 1) << field##_SHIFT)
+#define GET(field, flags) (((flags) & MASK(field)) >> field##_SHIFT)
+#define SET(field, flags, val) (((flags) & ~MASK(field)) | ((val) << field##_SHIFT))
+
+#define FLAGS_STATUS_SHIFT 0
+#define FLAGS_STATUS_WIDTH 1
+
+#define FLAGS_DMA_DIRECTION_SHIFT 1
+#define FLAGS_DMA_DIRECTION_WIDTH 2
+
/* Type declarations */
/* Valid states for a struct gasket_page_table_entry. */
enum pte_status {
@@ -108,8 +121,12 @@ enum pte_status {
* to the actual page mapped and described by this structure.
*/
struct gasket_page_table_entry {
- /* The status of this entry/slot: free or in use. */
- enum pte_status status;
+ /*
+ * Internal structure matches gasket_page_table_ioctl_flags.flags.
+ * NOTE: All fields should have a default value of 0. This ensures that
+ * the kernel will be backwards compatible with old drivers.
+ */
+ u32 flags;

/*
* Index for alignment into host vaddrs.
@@ -305,7 +322,7 @@ static bool gasket_is_pte_range_free(struct gasket_page_table_entry *ptes,
int i;

for (i = 0; i < num_entries; i++) {
- if (ptes[i].status != PTE_FREE)
+ if (GET(FLAGS_STATUS, ptes[i].flags) != PTE_FREE)
return false;
}

@@ -321,7 +338,7 @@ static void gasket_free_extended_subtable(struct gasket_page_table *pg_tbl,
u64 __iomem *slot)
{
/* Release the page table from the driver */
- pte->status = PTE_FREE;
+ pte->flags = SET(FLAGS_STATUS, pte->flags, PTE_FREE);

/* Release the page table from the device */
writeq(0, slot);
@@ -355,7 +372,7 @@ gasket_page_table_garbage_collect_nolock(struct gasket_page_table *pg_tbl)
slot = pg_tbl->base_slot + pg_tbl->num_simple_entries;
pte < pg_tbl->entries + pg_tbl->config.total_entries;
pte++, slot++) {
- if (pte->status == PTE_INUSE) {
+ if (GET(FLAGS_STATUS, pte->flags) == PTE_INUSE) {
if (gasket_is_pte_range_free(pte->sublevel,
GASKET_PAGES_PER_SUBTABLE))
gasket_free_extended_subtable(pg_tbl, pte,
@@ -404,7 +421,7 @@ int gasket_page_table_partition(struct gasket_page_table *pg_tbl,
start = min(pg_tbl->num_simple_entries, num_simple_entries);

for (i = start; i < pg_tbl->config.total_entries; i++) {
- if (pg_tbl->entries[i].status != PTE_FREE) {
+ if (GET(FLAGS_STATUS, pg_tbl->entries[i].flags) != PTE_FREE) {
dev_err(pg_tbl->device, "entry %d is not free\n", i);
mutex_unlock(&pg_tbl->mutex);
return -EBUSY;
@@ -465,7 +482,8 @@ static bool gasket_release_page(struct page *page)
static int gasket_perform_mapping(struct gasket_page_table *pg_tbl,
struct gasket_page_table_entry *ptes,
u64 __iomem *slots, ulong host_addr,
- uint num_pages, int is_simple_mapping)
+ uint num_pages, u32 flags,
+ int is_simple_mapping)
{
int ret;
ulong offset;
@@ -474,6 +492,12 @@ static int gasket_perform_mapping(struct gasket_page_table *pg_tbl,
ulong page_addr;
int i;

+ if (GET(FLAGS_DMA_DIRECTION, flags) == DMA_NONE) {
+ dev_err(pg_tbl->device, "invalid DMA direction flags=0x%lx\n",
+ (unsigned long)flags);
+ return -EINVAL;
+ }
+
for (i = 0; i < num_pages; i++) {
page_addr = host_addr + i * PAGE_SIZE;
offset = page_addr & (PAGE_SIZE - 1);
@@ -502,9 +526,8 @@ static int gasket_perform_mapping(struct gasket_page_table *pg_tbl,
ptes[i].offset = offset;

/* Map the page into DMA space. */
- ptes[i].dma_addr =
- dma_map_page(pg_tbl->device, page, 0, PAGE_SIZE,
- DMA_BIDIRECTIONAL);
+ ptes[i].dma_addr = dma_map_page(pg_tbl->device, page, 0, PAGE_SIZE,
+ GET(FLAGS_DMA_DIRECTION, flags));

if (dma_mapping_error(pg_tbl->device,
ptes[i].dma_addr)) {
@@ -531,7 +554,9 @@ static int gasket_perform_mapping(struct gasket_page_table *pg_tbl,
(void *)&((u64 __force *)slots)[i],
sizeof(u64), DMA_TO_DEVICE);
}
- ptes[i].status = PTE_INUSE;
+
+ /* Set PTE flags equal to flags param with STATUS=PTE_INUSE. */
+ ptes[i].flags = SET(FLAGS_STATUS, flags, PTE_INUSE);
}
return 0;
}
@@ -600,7 +625,8 @@ static void gasket_perform_unmapping(struct gasket_page_table *pg_tbl,
*/
for (i = 0; i < num_pages; i++) {
/* release the address from the device, */
- if (is_simple_mapping || ptes[i].status == PTE_INUSE) {
+ if (is_simple_mapping ||
+ GET(FLAGS_STATUS, ptes[i].flags) == PTE_INUSE) {
writeq(0, &slots[i]);
} else {
((u64 __force *)slots)[i] = 0;
@@ -609,15 +635,14 @@ static void gasket_perform_unmapping(struct gasket_page_table *pg_tbl,
}

/* release the address from the driver, */
- if (ptes[i].status == PTE_INUSE) {
+ if (GET(FLAGS_STATUS, ptes[i].flags) == PTE_INUSE) {
if (ptes[i].page && ptes[i].dma_addr) {
- dma_unmap_page(pg_tbl->device, ptes[i].dma_addr,
- PAGE_SIZE, DMA_BIDIRECTIONAL);
+ dma_unmap_page(pg_tbl->device, ptes[i].dma_addr, PAGE_SIZE,
+ GET(FLAGS_DMA_DIRECTION, ptes[i].flags));
+ if (gasket_release_page(ptes[i].page))
+ --pg_tbl->num_active_pages;
}
- if (gasket_release_page(ptes[i].page))
- --pg_tbl->num_active_pages;
}
-
/* and clear the PTE. */
memset(&ptes[i], 0, sizeof(struct gasket_page_table_entry));
}
@@ -656,7 +681,7 @@ static void gasket_unmap_extended_pages(struct gasket_page_table *pg_tbl,
/* TODO: Add check to ensure pte remains valid? */
len = min(remain, GASKET_PAGES_PER_SUBTABLE - slot_idx);

- if (pte->status == PTE_INUSE) {
+ if (GET(FLAGS_STATUS, pte->flags) == PTE_INUSE) {
slot_base = (u64 __iomem *)(page_address(pte->page) +
pte->offset);
gasket_perform_unmapping(pg_tbl,
@@ -818,7 +843,7 @@ static void gasket_page_table_unmap_nolock(struct gasket_page_table *pg_tbl,
*/
static int gasket_map_simple_pages(struct gasket_page_table *pg_tbl,
ulong host_addr, ulong dev_addr,
- uint num_pages)
+ uint num_pages, u32 flags)
{
int ret;
uint slot_idx = gasket_simple_page_idx(pg_tbl, dev_addr);
@@ -833,7 +858,7 @@ static int gasket_map_simple_pages(struct gasket_page_table *pg_tbl,

ret = gasket_perform_mapping(pg_tbl, pg_tbl->entries + slot_idx,
pg_tbl->base_slot + slot_idx, host_addr,
- num_pages, 1);
+ num_pages, flags, 1);

if (ret) {
gasket_page_table_unmap_nolock(pg_tbl, dev_addr, num_pages);
@@ -888,7 +913,7 @@ static int gasket_alloc_extended_subtable(struct gasket_page_table *pg_tbl,
dma_addr = (pte->dma_addr + pte->offset) | GASKET_VALID_SLOT_FLAG;
writeq(dma_addr, slot);

- pte->status = PTE_INUSE;
+ pte->flags = SET(FLAGS_STATUS, pte->flags, PTE_INUSE);

return 0;
}
@@ -926,7 +951,7 @@ static int gasket_alloc_extended_entries(struct gasket_page_table *pg_tbl,
len = min(remain,
GASKET_PAGES_PER_SUBTABLE - subtable_slot_idx);

- if (pte->status == PTE_FREE) {
+ if (GET(FLAGS_STATUS, pte->flags) == PTE_FREE) {
ret = gasket_alloc_extended_subtable(pg_tbl, pte, slot);
if (ret) {
dev_err(pg_tbl->device,
@@ -954,7 +979,7 @@ static int gasket_alloc_extended_entries(struct gasket_page_table *pg_tbl,
*/
static int gasket_map_extended_pages(struct gasket_page_table *pg_tbl,
ulong host_addr, ulong dev_addr,
- uint num_pages)
+ uint num_pages, u32 flags)
{
int ret;
ulong dev_addr_end;
@@ -988,7 +1013,7 @@ static int gasket_map_extended_pages(struct gasket_page_table *pg_tbl,
(u64 __iomem *)(page_address(pte->page) + pte->offset);
ret = gasket_perform_mapping(pg_tbl, pte->sublevel + slot_idx,
slot_base + slot_idx, host_addr,
- len, 0);
+ len, flags, 0);
if (ret) {
gasket_page_table_unmap_nolock(pg_tbl, dev_addr,
num_pages);
@@ -1013,7 +1038,7 @@ static int gasket_map_extended_pages(struct gasket_page_table *pg_tbl,
* The page table mutex is held for the entire operation.
*/
int gasket_page_table_map(struct gasket_page_table *pg_tbl, ulong host_addr,
- ulong dev_addr, uint num_pages)
+ ulong dev_addr, uint num_pages, u32 flags)
{
int ret;

@@ -1024,10 +1049,10 @@ int gasket_page_table_map(struct gasket_page_table *pg_tbl, ulong host_addr,

if (gasket_addr_is_simple(pg_tbl, dev_addr)) {
ret = gasket_map_simple_pages(pg_tbl, host_addr, dev_addr,
- num_pages);
+ num_pages, flags);
} else {
ret = gasket_map_extended_pages(pg_tbl, host_addr, dev_addr,
- num_pages);
+ num_pages, flags);
}

mutex_unlock(&pg_tbl->mutex);
@@ -1102,7 +1127,7 @@ int gasket_page_table_lookup_page(
goto fail;

pte = pg_tbl->entries + page_num;
- if (pte->status != PTE_INUSE)
+ if (GET(FLAGS_STATUS, pte->flags) != PTE_INUSE)
goto fail;
} else {
/* Find the level 0 entry, */
@@ -1111,13 +1136,13 @@ int gasket_page_table_lookup_page(
goto fail;

pte = pg_tbl->entries + pg_tbl->num_simple_entries + page_num;
- if (pte->status != PTE_INUSE)
+ if (GET(FLAGS_STATUS, pte->flags) != PTE_INUSE)
goto fail;

/* and its contained level 1 entry. */
page_num = gasket_extended_lvl1_page_idx(pg_tbl, dev_addr);
pte = pte->sublevel + page_num;
- if (pte->status != PTE_INUSE)
+ if (GET(FLAGS_STATUS, pte->flags) != PTE_INUSE)
goto fail;
}

diff --git a/drivers/staging/gasket/gasket_page_table.h b/drivers/staging/gasket/gasket_page_table.h
index 7b01b73ea3e70..d4df6a198e9e0 100644
--- a/drivers/staging/gasket/gasket_page_table.h
+++ b/drivers/staging/gasket/gasket_page_table.h
@@ -85,6 +85,8 @@ int gasket_page_table_partition(struct gasket_page_table *page_table,
* @host_addr: Starting host virtual memory address of the pages.
* @dev_addr: Starting device address of the pages.
* @num_pages: Number of [4kB] pages to map.
+ * @flags: Specifies attributes to apply to the pages.
+ * Internal structure matches gasket_page_table_ioctl_flags.flags.
*
* Description: Maps the "num_pages" pages of host memory pointed to by
* host_addr to the address "dev_addr" in device memory.
@@ -95,7 +97,7 @@ int gasket_page_table_partition(struct gasket_page_table *page_table,
* If there is an error, no pages are mapped.
*/
int gasket_page_table_map(struct gasket_page_table *page_table, ulong host_addr,
- ulong dev_addr, uint num_pages);
+ ulong dev_addr, uint num_pages, u32 flags);

/*
* Un-map host pages from device memory.
--
2.19.1.331.ge82ca0e54c-goog
\
 
 \ /
  Last update: 2018-10-16 14:04    [W:0.048 / U:0.224 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site