From: Ben Widawsky <benjamin.widawsky@intel.com>
To: Intel GFX <intel-gfx@lists.freedesktop.org>
Cc: Ben Widawsky <ben@bwidawsk.net>,
Ben Widawsky <benjamin.widawsky@intel.com>
Subject: [PATCH 22/56] drm/i915: Clean up pagetable DMA map & unmap
Date: Fri, 9 May 2014 20:59:17 -0700 [thread overview]
Message-ID: <1399694391-3935-23-git-send-email-benjamin.widawsky@intel.com> (raw)
In-Reply-To: <1399694391-3935-1-git-send-email-benjamin.widawsky@intel.com>
Map and unmap are common operations across all generations for
pagetables. With a simple helper, we can get a nice net code reduction
as well as simplified complexity.
There is some room for optimization here, for instance with the multiple
page mapping, that can be done in one pci_map operation. In that case
however, the max value we'll ever see there is 512, and so I believe the
simpler code makes this a worthwhile trade-off. Also, the range mapping
functions are place holders to help transition the code. Eventually,
mapping will only occur during a page allocation which will always be a
discrete operation.
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
---
drivers/gpu/drm/i915/i915_gem_gtt.c | 147 +++++++++++++++++++++---------------
1 file changed, 85 insertions(+), 62 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index e396b89..e8d4dfa 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -211,6 +211,76 @@ static gen6_gtt_pte_t iris_pte_encode(dma_addr_t addr,
return pte;
}
+#define dma_unmap_pt_single(pt, dev) do { \
+ pci_unmap_page((dev)->pdev, (pt)->daddr, 4096, PCI_DMA_BIDIRECTIONAL); \
+} while (0);
+
+
+static void dma_unmap_pt_range(struct i915_pagedir *pd,
+ unsigned pde, size_t n,
+ struct drm_device *dev)
+{
+ if (WARN_ON(pde + n > I915_PDES_PER_PD))
+ n = I915_PDES_PER_PD - pde;
+
+ n += pde;
+
+ for (; pde < n; pde++)
+ dma_unmap_pt_single(pd->page_tables[pde], dev);
+}
+
+/**
+ * dma_map_pt_single() - Create a dma mapping for a page table
+ * @pt: Page table to get a DMA map for
+ * @dev: drm device
+ *
+ * Page table allocations are unified across all gens. They always require a
+ * single 4k allocation, as well as a DMA mapping.
+ *
+ * Return: 0 if success.
+ */
+static int dma_map_pt_single(struct i915_pagetab *pt, struct drm_device *dev)
+{
+ struct page *page;
+ dma_addr_t pt_addr;
+ int ret;
+
+ page = pt->page;
+ pt_addr = pci_map_page(dev->pdev, page, 0, 4096,
+ PCI_DMA_BIDIRECTIONAL);
+
+ ret = pci_dma_mapping_error(dev->pdev, pt_addr);
+ if (ret)
+ return ret;
+
+ pt->daddr = pt_addr;
+
+ return 0;
+}
+
+static int dma_map_pt_range(struct i915_pagedir *pd,
+ unsigned pde, size_t n,
+ struct drm_device *dev)
+{
+ const int first = pde;
+
+ if (WARN_ON(pde + n > I915_PDES_PER_PD))
+ n = I915_PDES_PER_PD - pde;
+
+ n += pde;
+
+ for (; pde < n; pde++) {
+ int ret;
+ ret = dma_map_pt_single(pd->page_tables[pde], dev);
+ if (ret) {
+ dma_unmap_pt_range(pd, first, pde, dev);
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
static void free_pt_single(struct i915_pagetab *pt)
{
if (WARN_ON(!pt->page))
@@ -219,7 +289,7 @@ static void free_pt_single(struct i915_pagetab *pt)
kfree(pt);
}
-static struct i915_pagetab *alloc_pt_single(void)
+static struct i915_pagetab *alloc_pt_single(struct drm_device *dev)
{
struct i915_pagetab *pt;
@@ -242,6 +312,7 @@ static struct i915_pagetab *alloc_pt_single(void)
* available to point to the allocated page tables.
* @pde: First page directory entry for which we are allocating.
* @count: Number of pages to allocate.
+ * @dev DRM device used for DMA mapping.
*
* Allocates multiple page table pages and sets the appropriate entries in the
* page table structure within the page directory. Function cleans up after
@@ -249,7 +320,8 @@ static struct i915_pagetab *alloc_pt_single(void)
*
* Return: 0 if allocation succeeded.
*/
-static int alloc_pt_range(struct i915_pagedir *pd, uint16_t pde, size_t count)
+static int alloc_pt_range(struct i915_pagedir *pd, uint16_t pde, size_t count,
+ struct drm_device *dev)
{
int i, ret;
@@ -259,7 +331,7 @@ static int alloc_pt_range(struct i915_pagedir *pd, uint16_t pde, size_t count)
BUG_ON(pde + count > I915_PDES_PER_PD);
for (i = pde; i < pde + count; i++) {
- struct i915_pagetab *pt = alloc_pt_single();
+ struct i915_pagetab *pt = alloc_pt_single(dev);
if (IS_ERR(pt)) {
ret = PTR_ERR(pt);
goto err_out;
@@ -515,7 +587,7 @@ static int gen8_ppgtt_allocate_page_tables(struct i915_hw_ppgtt *ppgtt)
for (i = 0; i < ppgtt->num_pd_pages; i++) {
ret = alloc_pt_range(ppgtt->pdp.pagedir[i],
- 0, I915_PDES_PER_PD);
+ 0, I915_PDES_PER_PD, ppgtt->base.dev);
if (ret)
goto unwind_out;
}
@@ -594,27 +666,6 @@ static int gen8_ppgtt_setup_page_directories(struct i915_hw_ppgtt *ppgtt,
return 0;
}
-static int gen8_ppgtt_setup_page_tables(struct i915_hw_ppgtt *ppgtt,
- const int pdpe,
- const int pde)
-{
- dma_addr_t pt_addr;
- struct i915_pagedir *pd = ppgtt->pdp.pagedir[pdpe];
- struct i915_pagetab *pt = pd->page_tables[pde];
- struct page *p = pt->page;
- int ret;
-
- pt_addr = pci_map_page(ppgtt->base.dev->pdev,
- p, 0, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
- ret = pci_dma_mapping_error(ppgtt->base.dev->pdev, pt_addr);
- if (ret)
- return ret;
-
- pt->daddr = pt_addr;
-
- return 0;
-}
-
/**
* GEN8 legacy ppgtt programming is accomplished through a max 4 PDP registers
* with a net effect resembling a 2-level page table in normal x86 terms. Each
@@ -643,12 +694,15 @@ static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt, uint64_t size)
* 2. Create DMA mappings for the page directories and page tables.
*/
for (i = 0; i < max_pdp; i++) {
+ struct i915_pagedir *pd;
ret = gen8_ppgtt_setup_page_directories(ppgtt, i);
if (ret)
goto bail;
+ pd = ppgtt->pdp.pagedir[i];
+
for (j = 0; j < I915_PDES_PER_PD; j++) {
- ret = gen8_ppgtt_setup_page_tables(ppgtt, i, j);
+ ret = dma_map_pt_single(pd->page_tables[j], ppgtt->base.dev);
if (ret)
goto bail;
}
@@ -1066,16 +1120,6 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
kunmap_atomic(pt_vaddr);
}
-static void gen6_ppgtt_dma_unmap_pages(struct i915_hw_ppgtt *ppgtt)
-{
- int i;
-
- for (i = 0; i < ppgtt->num_pd_entries; i++)
- pci_unmap_page(ppgtt->base.dev->pdev,
- ppgtt->pd.page_tables[i]->daddr,
- 4096, PCI_DMA_BIDIRECTIONAL);
-}
-
static void gen6_ppgtt_free(struct i915_hw_ppgtt *ppgtt)
{
int i;
@@ -1095,7 +1139,7 @@ static void gen6_ppgtt_cleanup(struct i915_address_space *vm)
drm_mm_takedown(&ppgtt->base.mm);
drm_mm_remove_node(&ppgtt->node);
- gen6_ppgtt_dma_unmap_pages(ppgtt);
+ dma_unmap_pt_range(&ppgtt->pd, 0, ppgtt->num_pd_entries, vm->dev);
gen6_ppgtt_free(ppgtt);
}
@@ -1146,7 +1190,8 @@ static int gen6_ppgtt_alloc(struct i915_hw_ppgtt *ppgtt)
if (ret)
return ret;
- ret = alloc_pt_range(&ppgtt->pd, 0, ppgtt->num_pd_entries);
+ ret = alloc_pt_range(&ppgtt->pd, 0, ppgtt->num_pd_entries,
+ ppgtt->base.dev);
if (ret) {
drm_mm_remove_node(&ppgtt->node);
return ret;
@@ -1155,29 +1200,6 @@ static int gen6_ppgtt_alloc(struct i915_hw_ppgtt *ppgtt)
return 0;
}
-static int gen6_ppgtt_setup_page_tables(struct i915_hw_ppgtt *ppgtt)
-{
- struct drm_device *dev = ppgtt->base.dev;
- int i;
-
- for (i = 0; i < ppgtt->num_pd_entries; i++) {
- struct page *page;
- dma_addr_t pt_addr;
-
- page = ppgtt->pd.page_tables[i]->page;
- pt_addr = pci_map_page(dev->pdev, page, 0, 4096,
- PCI_DMA_BIDIRECTIONAL);
-
- if (pci_dma_mapping_error(dev->pdev, pt_addr)) {
- gen6_ppgtt_dma_unmap_pages(ppgtt);
- return -EIO;
- }
-
- ppgtt->pd.page_tables[i]->daddr = pt_addr;
- }
-
- return 0;
-}
static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
{
@@ -1202,7 +1224,8 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
if (ret)
return ret;
- ret = gen6_ppgtt_setup_page_tables(ppgtt);
+ ret = dma_map_pt_range(&ppgtt->pd, 0, ppgtt->num_pd_entries,
+ ppgtt->base.dev);
if (ret) {
gen6_ppgtt_free(ppgtt);
return ret;
--
1.9.2
next prev parent reply other threads:[~2014-05-10 4:01 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-05-10 3:58 [PATCH 00/56] [RFCish] Dynamic page table alloc, 64b, and GPU/CPU mirror Ben Widawsky
2014-05-10 3:58 ` [PATCH 01/56] drm/i915: Fix flush before context switch comment Ben Widawsky
2014-05-10 3:58 ` [PATCH 02/56] Revert "drm/i915: Drop I915_PARAM_HAS_FULL_PPGTT again" Ben Widawsky
2014-05-10 3:58 ` [PATCH 03/56] drm/i915: Prevent signals from interrupting close() Ben Widawsky
2014-05-10 3:58 ` [PATCH 04/56] drm/i915: Wrap VMA binding Ben Widawsky
2014-05-10 3:59 ` [PATCH 05/56] drm/i915: Make pin global flags explicit Ben Widawsky
2014-05-10 3:59 ` [PATCH 06/56] drm/i915: Split out aliasing binds Ben Widawsky
2014-05-10 3:59 ` [PATCH 07/56] drm/i915: fix gtt_total_entries() Ben Widawsky
2014-05-10 3:59 ` [PATCH 08/56] drm/i915: Rename to GEN8_LEGACY_PDPES Ben Widawsky
2014-05-10 3:59 ` [PATCH 09/56] drm/i915: Split out verbose PPGTT dumping Ben Widawsky
2014-05-10 3:59 ` [PATCH 10/56] drm/i915: s/pd/pdpe, s/pt/pde Ben Widawsky
2014-05-10 3:59 ` [PATCH 11/56] drm/i915: rename map/unmap to dma_map/unmap Ben Widawsky
2014-05-10 3:59 ` [PATCH 12/56] drm/i915: Setup less PPGTT on failed pagedir Ben Widawsky
2014-05-10 3:59 ` [PATCH 13/56] drm/i915: clean up PPGTT init error path Ben Widawsky
2014-05-10 3:59 ` [PATCH 14/56] drm/i915: Un-hardcode number of page directories Ben Widawsky
2014-05-10 3:59 ` [PATCH 15/56] drm/i915: Make gen6_write_pdes gen6_map_page_tables Ben Widawsky
2014-05-10 3:59 ` [PATCH 16/56] drm/i915: Range clearing is PPGTT agnostic Ben Widawsky
2014-05-10 3:59 ` [PATCH 17/56] drm/i915: Page table helpers, and define renames Ben Widawsky
2014-05-10 3:59 ` [PATCH 18/56] drm/i915: construct page table abstractions Ben Widawsky
2014-05-10 3:59 ` [PATCH 19/56] drm/i915: Complete page table structures Ben Widawsky
2014-05-10 3:59 ` [PATCH 20/56] drm/i915: Create page table allocators Ben Widawsky
2014-05-10 3:59 ` [PATCH 21/56] drm/i915: Generalize GEN6 mapping Ben Widawsky
2014-05-10 3:59 ` Ben Widawsky [this message]
2014-05-10 3:59 ` [PATCH 23/56] drm/i915: Always dma map page table allocations Ben Widawsky
2014-05-10 3:59 ` [PATCH 24/56] drm/i915: Consolidate dma mappings Ben Widawsky
2014-05-10 3:59 ` [PATCH 25/56] drm/i915: Always dma map page directory allocations Ben Widawsky
2014-05-10 3:59 ` [PATCH 26/56] drm/i915: Track GEN6 page table usage Ben Widawsky
2014-05-10 3:59 ` [PATCH 27/56] drm/i915: Extract context switch skip logic Ben Widawsky
2014-05-10 3:59 ` [PATCH 28/56] drm/i915: Force pd restore when PDEs change, gen6-7 Ben Widawsky
2014-05-10 3:59 ` [PATCH 29/56] drm/i915: Finish gen6/7 dynamic page table allocation Ben Widawsky
2014-05-10 3:59 ` [PATCH 30/56] drm/i915/bdw: Use dynamic allocation idioms on free Ben Widawsky
2014-05-10 3:59 ` [PATCH 31/56] drm/i915/bdw: pagedirs rework allocation Ben Widawsky
2014-05-10 3:59 ` [PATCH 32/56] drm/i915/bdw: pagetable allocation rework Ben Widawsky
2014-05-10 3:59 ` [PATCH 33/56] drm/i915/bdw: Make the pdp switch a bit less hacky Ben Widawsky
2014-05-10 3:59 ` [PATCH 34/56] drm/i915: num_pd_pages/num_pd_entries isn't useful Ben Widawsky
2014-05-10 3:59 ` [PATCH 35/56] drm/i915: Extract PPGTT param from pagedir alloc Ben Widawsky
2014-05-10 3:59 ` [PATCH 36/56] drm/i915/bdw: Split out mappings Ben Widawsky
2014-05-10 3:59 ` [PATCH 37/56] drm/i915/bdw: begin bitmap tracking Ben Widawsky
2014-05-10 3:59 ` [PATCH 38/56] drm/i915/bdw: Dynamic page table allocations Ben Widawsky
2014-05-10 3:59 ` [PATCH 39/56] drm/i915/bdw: Scratch unused pages Ben Widawsky
2014-05-10 3:59 ` [PATCH 40/56] drm/i915/bdw: Add ppgtt info for dynamic pages Ben Widawsky
2014-05-10 3:59 ` [PATCH 41/56] drm/i915/bdw: Optimize PDP loads Ben Widawsky
2014-05-10 3:59 ` [PATCH 42/56] TESTME: Either drop the last patch or fix it Ben Widawsky
2014-05-10 3:59 ` [PATCH 43/56] drm/i915/bdw: Add dynamic page trace events Ben Widawsky
2014-05-10 3:59 ` [PATCH 44/56] drm/i915/bdw: Make pdp allocation more dynamic Ben Widawsky
2014-05-10 3:59 ` [PATCH 45/56] drm/i915/bdw: Abstract PDP usage Ben Widawsky
2014-05-10 3:59 ` [PATCH 46/56] drm/i915/bdw: implement alloc/teardown for 4lvl Ben Widawsky
2014-05-10 3:59 ` [PATCH 47/56] drm/i915/bdw: 4 level pages tables Ben Widawsky
2014-05-10 3:59 ` [PATCH 48/56] drm/i915: Restructure map vs. insert entries Ben Widawsky
2014-05-10 3:59 ` [PATCH 49/56] drm/i915/bdw: make aliasing PPGTT dynamic Ben Widawsky
2014-05-10 3:59 ` [PATCH 50/56] drm/i915: Expand error state's address width to 64b Ben Widawsky
2014-05-10 3:59 ` [PATCH 51/56] drm/i915/bdw: Flip the 48b switch Ben Widawsky
2014-05-10 3:59 ` [PATCH 52/56] TESTME: GFX_TLB_INVALIDATE_EXPLICIT Ben Widawsky
2014-05-10 3:59 ` [PATCH 53/56] TESTME: Always force invalidate Ben Widawsky
2014-05-10 3:59 ` [PATCH 54/56] drm/i915: Introduce mapping of user pages into video memory (userptr) ioctl Ben Widawsky
2014-05-10 3:59 ` [PATCH 55/56] drm/i915: Track userptr VMAs Ben Widawsky
2014-05-10 3:59 ` [PATCH 56/56] drm/i915/userptr: Mirror GPU addr at ioctl (HACK/POC) Ben Widawsky
2014-05-11 17:33 ` [PATCH 00/56] [RFCish] Dynamic page table alloc, 64b, and GPU/CPU mirror Daniel Vetter
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1399694391-3935-23-git-send-email-benjamin.widawsky@intel.com \
--to=benjamin.widawsky@intel.com \
--cc=ben@bwidawsk.net \
--cc=intel-gfx@lists.freedesktop.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox