korg/linux
0
0
Fork 0
mirror of https://mirrors.bfsu.edu.cn/git/linux.git synced 2024-11-14 15:54:15 +08:00
Code Issues Packages Projects Releases Wiki Activity

drm/i915: stop abusing swiotlb_max_segment

Browse Source 
swiotlb_max_segment used to return either the maximum size that swiotlb
could bounce, or for Xen PV PAGE_SIZE even if swiotlb could bounce buffer
larger mappings.  This made i915 on Xen PV work as it bypasses the
coherency aspect of the DMA API and can't cope with bounce buffering
and this avoided bounce buffering for the Xen/PV case.

So instead of adding this hack back, check for Xen/PV directly in i915
for the Xen case and otherwise use the proper DMA API helper to query
the maximum mapping size.

Replace swiotlb_max_segment() calls with dma_max_mapping_size().
In i915_gem_object_get_pages_internal() no longer consider max_segment
only if CONFIG_SWIOTLB is enabled. There can be other (iommu related)
causes of specific max segment sizes.

Fixes: a2daa27c0c ("swiotlb: simplify swiotlb_max_segment")
Reported-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Signed-off-by: Robert Beckett <bob.beckett@collabora.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
[hch: added the Xen hack, rewrote the changelog]
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020110308.1582518-1-hch@lst.de
(cherry picked from commit 78a07fe777)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
This commit is contained in:
Robert Beckett 2022-10-20 13:03:08 +02:00 committed by Tvrtko Ursulin
parent d7164a5048
commit d3f6bacfca
5 changed files with 28 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
drivers/gpu/drm/i915/gem/i915_gem_internal.c
View File

@ -6,7 +6,6 @@
#include <linux/scatterlist.h> #include <linux/scatterlist.h>
#include <linux/slab.h> #include <linux/slab.h>
#include <linux/swiotlb.h>
#include "i915_drv.h" #include "i915_drv.h"
#include "i915_gem.h" #include "i915_gem.h"
@ -38,22 +37,12 @@ static int i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
struct scatterlist *sg; struct scatterlist *sg;
unsigned int sg_page_sizes; unsigned int sg_page_sizes;
unsigned int npages; unsigned int npages;
int max_order; int max_order = MAX_ORDER;
unsigned int max_segment;
gfp_t gfp; gfp_t gfp;
max_order = MAX_ORDER; max_segment = i915_sg_segment_size(i915->drm.dev) >> PAGE_SHIFT;
#ifdef CONFIG_SWIOTLB max_order = min(max_order, get_order(max_segment));
if (is_swiotlb_active(obj->base.dev->dev)) {
unsigned int max_segment;
max_segment = swiotlb_max_segment();
if (max_segment) {
max_segment = max_t(unsigned int, max_segment,
PAGE_SIZE) >> PAGE_SHIFT;
max_order = min(max_order, ilog2(max_segment));
}
}
#endif
gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE; gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;
if (IS_I965GM(i915) || IS_I965G(i915)) { if (IS_I965GM(i915) || IS_I965G(i915)) {

2
drivers/gpu/drm/i915/gem/i915_gem_shmem.c
View File

@ -194,7 +194,7 @@ static int shmem_get_pages(struct drm_i915_gem_object *obj)
struct intel_memory_region *mem = obj->mm.region; struct intel_memory_region *mem = obj->mm.region;
struct address_space *mapping = obj->base.filp->f_mapping; struct address_space *mapping = obj->base.filp->f_mapping;
const unsigned long page_count = obj->base.size / PAGE_SIZE; const unsigned long page_count = obj->base.size / PAGE_SIZE;
unsigned int max_segment = i915_sg_segment_size(); unsigned int max_segment = i915_sg_segment_size(i915->drm.dev);
struct sg_table *st; struct sg_table *st;
struct sgt_iter sgt_iter; struct sgt_iter sgt_iter;
struct page *page; struct page *page;

4
drivers/gpu/drm/i915/gem/i915_gem_ttm.c
View File

@ -189,7 +189,7 @@ static int i915_ttm_tt_shmem_populate(struct ttm_device *bdev,
struct drm_i915_private *i915 = container_of(bdev, typeof(*i915), bdev); struct drm_i915_private *i915 = container_of(bdev, typeof(*i915), bdev);
struct intel_memory_region *mr = i915->mm.regions[INTEL_MEMORY_SYSTEM]; struct intel_memory_region *mr = i915->mm.regions[INTEL_MEMORY_SYSTEM];
struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm); struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
const unsigned int max_segment = i915_sg_segment_size(); const unsigned int max_segment = i915_sg_segment_size(i915->drm.dev);
const size_t size = (size_t)ttm->num_pages << PAGE_SHIFT; const size_t size = (size_t)ttm->num_pages << PAGE_SHIFT;
struct file *filp = i915_tt->filp; struct file *filp = i915_tt->filp;
struct sgt_iter sgt_iter; struct sgt_iter sgt_iter;
@ -538,7 +538,7 @@ static struct i915_refct_sgt *i915_ttm_tt_get_st(struct ttm_tt *ttm)
ret = sg_alloc_table_from_pages_segment(st, ret = sg_alloc_table_from_pages_segment(st,
ttm->pages, ttm->num_pages, ttm->pages, ttm->num_pages,
0, (unsigned long)ttm->num_pages << PAGE_SHIFT, 0, (unsigned long)ttm->num_pages << PAGE_SHIFT,
i915_sg_segment_size(), GFP_KERNEL); i915_sg_segment_size(i915_tt->dev), GFP_KERNEL);
if (ret) { if (ret) {
st->sgl = NULL; st->sgl = NULL;
return ERR_PTR(ret); return ERR_PTR(ret);

2
drivers/gpu/drm/i915/gem/i915_gem_userptr.c
View File

@ -129,7 +129,7 @@ static void i915_gem_object_userptr_drop_ref(struct drm_i915_gem_object *obj)
static int i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj) static int i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
{ {
const unsigned long num_pages = obj->base.size >> PAGE_SHIFT; const unsigned long num_pages = obj->base.size >> PAGE_SHIFT;
unsigned int max_segment = i915_sg_segment_size(); unsigned int max_segment = i915_sg_segment_size(obj->base.dev->dev);
struct sg_table *st; struct sg_table *st;
unsigned int sg_page_sizes; unsigned int sg_page_sizes;
struct page **pvec; struct page **pvec;

32
drivers/gpu/drm/i915/i915_scatterlist.h
View File

@ -9,7 +9,8 @@
#include <linux/pfn.h> #include <linux/pfn.h>
#include <linux/scatterlist.h> #include <linux/scatterlist.h>
#include <linux/swiotlb.h> #include <linux/dma-mapping.h>
#include <xen/xen.h>
#include "i915_gem.h" #include "i915_gem.h"
@ -127,19 +128,26 @@ static inline unsigned int i915_sg_dma_sizes(struct scatterlist *sg)
return page_sizes; return page_sizes;
} }
static inline unsigned int i915_sg_segment_size(void) static inline unsigned int i915_sg_segment_size(struct device *dev)
{ {
unsigned int size = swiotlb_max_segment(); size_t max = min_t(size_t, UINT_MAX, dma_max_mapping_size(dev));
if (size == 0) /*
size = UINT_MAX; * For Xen PV guests pages aren't contiguous in DMA (machine) address
* space. The DMA API takes care of that both in dma_alloc_* (by
size = rounddown(size, PAGE_SIZE); * calling into the hypervisor to make the pages contiguous) and in
/* swiotlb_max_segment_size can return 1 byte when it means one page. */ * dma_map_* (by bounce buffering). But i915 abuses ignores the
if (size < PAGE_SIZE) * coherency aspects of the DMA API and thus can't cope with bounce
size = PAGE_SIZE; * buffering actually happening, so add a hack here to force small
* allocations and mappings when running in PV mode on Xen.
return size; *
* Note this will still break if bounce buffering is required for other
* reasons, like confidential computing hypervisors or PCIe root ports
* with addressing limitations.
*/
if (xen_pv_domain())
max = PAGE_SIZE;
return round_down(max, PAGE_SIZE);
} }
bool i915_sg_trim(struct sg_table *orig_st); bool i915_sg_trim(struct sg_table *orig_st);