Calling swiotlb functions directly is nowadays considered harmful. See
https://lore.kernel.org/intel-gfx/[email protected]/
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.
Cc: Christoph Hellwig <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Cc: Thomas Hellstrom <[email protected]>
Cc: Matthew Auld <[email protected]>
v2: - restore UINT_MAX clamp in i915_sg_segment_size()
- drop PAGE_SIZE check as it will always be >= PAGE_SIZE
v3: - actually clamp to UINT_MAX in i915_sg_segment_size()
v4: - round down max segment size to PAGE_SIZE
v5: - fix checkpatch whitespace issue
Reviewed-by: Christoph Hellwig <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Signed-off-by: Robert Beckett <[email protected]>
---
drivers/gpu/drm/i915/gem/i915_gem_internal.c | 19 ++++---------------
drivers/gpu/drm/i915/gem/i915_gem_shmem.c | 2 +-
drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 4 ++--
drivers/gpu/drm/i915/gem/i915_gem_userptr.c | 2 +-
drivers/gpu/drm/i915/i915_scatterlist.h | 16 ++++------------
5 files changed, 12 insertions(+), 31 deletions(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_internal.c b/drivers/gpu/drm/i915/gem/i915_gem_internal.c
index c698f95af15f..24f37658f1bb 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_internal.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_internal.c
@@ -6,7 +6,6 @@
#include <linux/scatterlist.h>
#include <linux/slab.h>
-#include <linux/swiotlb.h>
#include "i915_drv.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;
unsigned int sg_page_sizes;
unsigned int npages;
- int max_order;
+ int max_order = MAX_ORDER;
+ unsigned int max_segment;
gfp_t gfp;
- max_order = MAX_ORDER;
-#ifdef CONFIG_SWIOTLB
- 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
+ max_segment = i915_sg_segment_size(i915->drm.dev) >> PAGE_SHIFT;
+ max_order = min(max_order, ilog2(max_segment));
gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;
if (IS_I965GM(i915) || IS_I965G(i915)) {
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
index 4eed3dd90ba8..34b9c76cd8e6 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
@@ -194,7 +194,7 @@ static int shmem_get_pages(struct drm_i915_gem_object *obj)
struct intel_memory_region *mem = obj->mm.region;
struct address_space *mapping = obj->base.filp->f_mapping;
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 sgt_iter sgt_iter;
struct page *page;
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
index 5a5cf332d8a5..7a828c9c0f6d 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
@@ -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 intel_memory_region *mr = i915->mm.regions[INTEL_MEMORY_SYSTEM];
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;
struct file *filp = i915_tt->filp;
struct sgt_iter sgt_iter;
@@ -568,7 +568,7 @@ static struct i915_refct_sgt *i915_ttm_tt_get_st(struct ttm_tt *ttm)
ret = sg_alloc_table_from_pages_segment(st,
ttm->pages, ttm->num_pages,
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) {
st->sgl = NULL;
return ERR_PTR(ret);
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
index 094f06b4ce33..dfc35905dba2 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
@@ -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)
{
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;
unsigned int sg_page_sizes;
struct page **pvec;
diff --git a/drivers/gpu/drm/i915/i915_scatterlist.h b/drivers/gpu/drm/i915/i915_scatterlist.h
index 9ddb3e743a3e..3dd40012e901 100644
--- a/drivers/gpu/drm/i915/i915_scatterlist.h
+++ b/drivers/gpu/drm/i915/i915_scatterlist.h
@@ -9,7 +9,7 @@
#include <linux/pfn.h>
#include <linux/scatterlist.h>
-#include <linux/swiotlb.h>
+#include <linux/dma-mapping.h>
#include "i915_gem.h"
@@ -127,19 +127,11 @@ static inline unsigned int i915_sg_dma_sizes(struct scatterlist *sg)
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;
-
- size = rounddown(size, PAGE_SIZE);
- /* swiotlb_max_segment_size can return 1 byte when it means one page. */
- if (size < PAGE_SIZE)
- size = PAGE_SIZE;
-
- return size;
+ return round_down(max, PAGE_SIZE);
}
bool i915_sg_trim(struct sg_table *orig_st);
--
2.25.1
Hi, [back from vacation]
On Tue, 2022-07-26 at 16:39 +0100, Robert Beckett wrote:
> Calling swiotlb functions directly is nowadays considered harmful.
> See
> https://lore.kernel.org/intel-gfx/[email protected]/
>
> 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.
>
> Cc: Christoph Hellwig <[email protected]>
> Cc: Tvrtko Ursulin <[email protected]>
> Cc: Thomas Hellstrom <[email protected]>
> Cc: Matthew Auld <[email protected]>
>
> v2: - restore UINT_MAX clamp in i915_sg_segment_size()
> - drop PAGE_SIZE check as it will always be >= PAGE_SIZE
> v3: - actually clamp to UINT_MAX in i915_sg_segment_size()
> v4: - round down max segment size to PAGE_SIZE
> v5: - fix checkpatch whitespace issue
>
> Reviewed-by: Christoph Hellwig <[email protected]>
> Reviewed-by: Tvrtko Ursulin <[email protected]>
> Signed-off-by: Robert Beckett <[email protected]>
Hmm,
This whole thing looks a bit strange to me since with SWIOTLB actually
used for i915, the driver should malfunction anyway as it doesn't do
any dma_sync_sg_for_cpu() or dma_sync_sg_for_device(), and the driver
assumes all coherent dma. Is that SWIOTLB=force kernel option still
available?
Also, correct me if I'm wrong, but the original driver segment size
appears to mean "the largest contiguous area that can be handled either
by the device or the dma mapping layer" rather than the total space
available for dma mappings? Not completely sure what
dma_max_mapping_size() is returning, though?
/Thomas
On 08/08/2022 16:48, Hellstrom, Thomas wrote:
> Hi, [back from vacation]
>
> On Tue, 2022-07-26 at 16:39 +0100, Robert Beckett wrote:
>> Calling swiotlb functions directly is nowadays considered harmful.
>> See
>> https://lore.kernel.org/intel-gfx/[email protected]/
>>
>> 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.
>>
>> Cc: Christoph Hellwig <[email protected]>
>> Cc: Tvrtko Ursulin <[email protected]>
>> Cc: Thomas Hellstrom <[email protected]>
>> Cc: Matthew Auld <[email protected]>
>>
>> v2: - restore UINT_MAX clamp in i915_sg_segment_size()
>> - drop PAGE_SIZE check as it will always be >= PAGE_SIZE
>> v3: - actually clamp to UINT_MAX in i915_sg_segment_size()
>> v4: - round down max segment size to PAGE_SIZE
>> v5: - fix checkpatch whitespace issue
>>
>> Reviewed-by: Christoph Hellwig <[email protected]>
>> Reviewed-by: Tvrtko Ursulin <[email protected]>
>> Signed-off-by: Robert Beckett <[email protected]>
>
> Hmm,
>
> This whole thing looks a bit strange to me since with SWIOTLB actually
> used for i915, the driver should malfunction anyway as it doesn't do
> any dma_sync_sg_for_cpu() or dma_sync_sg_for_device(), and the driver
> assumes all coherent dma. Is that SWIOTLB=force kernel option still
> available?
Don't know about these - but pretty sure in the past we had i915 break
if we did not respect swiotlb_max_segment.
Digging through git history at least running as Xen dom0 looks to have
been impacted, but commits such as abb0deacb5a6 ("drm/i915: Fallback to
single PAGE_SIZE segments for DMA remapping") are older and suggest
problem was generic. 1625e7e549c5 ("drm/i915: make compact dma scatter
lists creation work with SWIOTLB backend.") as well. So it looks it did
work behind swiotlb despite those missing calls you highlighted.
> Also, correct me if I'm wrong, but the original driver segment size
> appears to mean "the largest contiguous area that can be handled either
> by the device or the dma mapping layer" rather than the total space
> available for dma mappings? Not completely sure what
> dma_max_mapping_size() is returning, though?
AFAIU looks to be compatible on paper at least.:
dma_max_mapping_size -> "Returns the maximum size of a mapping for the
device."
So an individual mapping.
But then in case of swiotlb is implemented in swiotlb_max_mapping_size,
and not the same code as swiotlb_max_segment. I agree, ideally if
someone could clarify they are returning the same thing or there is a
miss somewhere.
Regards,
Tvrtko
On Tue, Aug 09, 2022 at 12:36:50PM +0100, Tvrtko Ursulin wrote:
>
> Digging through git history at least running as Xen dom0 looks to have been
> impacted, but commits such as abb0deacb5a6 ("drm/i915: Fallback to single
> PAGE_SIZE segments for DMA remapping") are older and suggest problem was
> generic. 1625e7e549c5 ("drm/i915: make compact dma scatter lists creation
> work with SWIOTLB backend.") as well. So it looks it did work behind
> swiotlb despite those missing calls you highlighted.
Hmm. xen-swiotlb bounce buffers as soon as any single mapping that
straddles a Xen page size boundary. Could the magic value there
somehow made all mappings small enough to just avoid bounce buffering
for Xen by more or less accident?
On Mon, Aug 08, 2022 at 03:48:02PM +0000, Hellstrom, Thomas wrote:
> This whole thing looks a bit strange to me since with SWIOTLB actually
> used for i915, the driver should malfunction anyway as it doesn't do
> any dma_sync_sg_for_cpu() or dma_sync_sg_for_device(),
Yeah, I can't actually see any syncs in i915.
> and the driver
> assumes all coherent dma. Is that SWIOTLB=force kernel option still
> available?
Yes, although it is (and always has been) lower case swiotlb for the
option.
> Also, correct me if I'm wrong, but the original driver segment size
> appears to mean "the largest contiguous area that can be handled either
> by the device or the dma mapping layer" rather than the total space
> available for dma mappings? Not completely sure what
> dma_max_mapping_size() is returning, though?
dma_max_mapping_size is sort of both. It is is the largest contigous
size, but we really should not go above that to avoid starvation.