Skip to contentSkip to conversation

nouveau: fix the start/end range for migration

PATCH Superseded Ralph Campbell 2 participants 5 messages Aug 27, 2020 – Aug 31, 2020

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 nouveau: fix the start/end range for migration

The user level OpenCL code shouldn't have to align start and end addresses to a page boundary. That is better handled in the nouveau driver. The npages field is also redundant since it can be computed from the start and end addresses.

---

This is for Ben Skegg's nouveau tree.

I have been working with Karol Herbst on the OpenCL mesa changes for nouveau which will be merged upstream soon.
With or without those changes, the user visible effect of this patch only extends the range by one page (round up vs. round down to page boundary).

 drivers/gpu/drm/nouveau/nouveau_svm.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)
drivers/gpu/drm/nouveau/nouveau_svm.c +6 -11
index 2df1c0460559..888aa0908c5a 100644
@@ -105,11 +105,14 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct drm_nouveau_svm_bind *args = data;
 	unsigned target, cmd, priority;
-	unsigned long addr, end, size;
+	unsigned long addr, end;
 	struct mm_struct *mm;
 
 	args->va_start &= PAGE_MASK;
-	args->va_end &= PAGE_MASK;
+	args->va_end = ALIGN(args->va_end, PAGE_SIZE);
+	/* If no end address is given, assume a single page. */
+	if (args->va_end == 0)
+		args->va_end = args->va_start + PAGE_SIZE;
 
 	/* Sanity check arguments */
 	if (args->reserved0 || args->reserved1)
@@ -118,8 +121,6 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
 		return -EINVAL;
 	if (args->va_start >= args->va_end)
 		return -EINVAL;
-	if (!args->npages)
-		return -EINVAL;
 
 	cmd = args->header >> NOUVEAU_SVM_BIND_COMMAND_SHIFT;
 	cmd &= NOUVEAU_SVM_BIND_COMMAND_MASK;
@@ -151,12 +152,6 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
 	if (args->stride)
 		return -EINVAL;
 
-	size = ((unsigned long)args->npages) << PAGE_SHIFT;
-	if ((args->va_start + size) <= args->va_start)
-		return -EINVAL;
-	if ((args->va_start + size) > args->va_end)
-		return -EINVAL;
-
 	/*
 	 * Ok we are ask to do something sane, for now we only support migrate
 	 * commands but we will add things like memory policy (what to do on
@@ -171,7 +166,7 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
 		return -EINVAL;
 	}
 
-	for (addr = args->va_start, end = args->va_start + size; addr < end;) {
+	for (addr = args->va_start, end = args->va_end; addr < end;) {
 		struct vm_area_struct *vma;
 		unsigned long next;
 
Signature · 1 line
2.20.1
#1 lore raw
Quoted 25 lines · Ralph Campbell

The user level OpenCL code shouldn't have to align start and end addresses to a page boundary. That is better handled in the nouveau driver. The npages field is also redundant since it can be computed from the start and end addresses.

This is for Ben Skegg's nouveau tree.

I have been working with Karol Herbst on the OpenCL mesa changes for nouveau which will be merged upstream soon.
With or without those changes, the user visible effect of this patch only extends the range by one page (round up vs. round down to page boundary).

 drivers/gpu/drm/nouveau/nouveau_svm.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)
drivers/gpu/drm/nouveau/nouveau_svm.c +1 -0
index 2df1c0460559..888aa0908c5a 100644
+++ b/drivers/gpu/drm/nouveau/nouveau_svm.c
@@ -105,11 +105,14 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct drm_nouveau_svm_bind *args = data;
 	unsigned target, cmd, priority;

drivers/gpu/drm/nouveau/nouveau_svm.c nouveau_svmm_bind(struct drm_device *dev, void *data, · line 115

Quoted 10 lines
drivers/gpu/drm/nouveau/nouveau_svm.c +5 -2
-	unsigned long addr, end, size;
+	unsigned long addr, end;
 	struct mm_struct *mm;
 
 	args->va_start &= PAGE_MASK;
-	args->va_end &= PAGE_MASK;
+	args->va_end = ALIGN(args->va_end, PAGE_SIZE);
+	/* If no end address is given, assume a single page. */
+	if (args->va_end == 0)
+		args->va_end = args->va_start + PAGE_SIZE;

That is really weird, how is it useful for the kernel to map a region of unknown size and alignment to the GPU?

Jason

#2 lore raw
Quoted 40 lines · Jason Gunthorpe
Quoted 25 lines · Ralph Campbell

The user level OpenCL code shouldn't have to align start and end addresses to a page boundary. That is better handled in the nouveau driver. The npages field is also redundant since it can be computed from the start and end addresses.

This is for Ben Skegg's nouveau tree.

I have been working with Karol Herbst on the OpenCL mesa changes for nouveau which will be merged upstream soon.
With or without those changes, the user visible effect of this patch only extends the range by one page (round up vs. round down to page boundary).

  drivers/gpu/drm/nouveau/nouveau_svm.c | 17 ++++++-----------
  1 file changed, 6 insertions(+), 11 deletions(-)
drivers/gpu/drm/nouveau/nouveau_svm.c +1 -0
index 2df1c0460559..888aa0908c5a 100644
+++ b/drivers/gpu/drm/nouveau/nouveau_svm.c
@@ -105,11 +105,14 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
  	struct nouveau_cli *cli = nouveau_cli(file_priv);
  	struct drm_nouveau_svm_bind *args = data;
  	unsigned target, cmd, priority;

drivers/gpu/drm/nouveau/nouveau_svm.c nouveau_svmm_bind(struct drm_device *dev, void *data, · line 115

Quoted 10 lines
drivers/gpu/drm/nouveau/nouveau_svm.c +5 -2
-	unsigned long addr, end, size;
+	unsigned long addr, end;
  	struct mm_struct *mm;
 
  	args->va_start &= PAGE_MASK;
-	args->va_end &= PAGE_MASK;
+	args->va_end = ALIGN(args->va_end, PAGE_SIZE);
+	/* If no end address is given, assume a single page. */
+	if (args->va_end == 0)
+		args->va_end = args->va_start + PAGE_SIZE;

That is really weird, how is it useful for the kernel to map a region of unknown size and alignment to the GPU?

Quoted 1 line

Jason

I agree it is somewhat weird. The OpenCL 2.2 specification says that clEnqueueSVMMigrateMem() takes an array of pointers and sizes (in bytes) but the size is optional. There is no alignment required.
This "works" because the pointers have to be allocated with clSVMAlloc() and presumably, the implementation for clSVMAlloc()
keeps track of the length allocated and can fill that in if size is zero. However, requiring all allocations to be made with clSVMAlloc() defeats the goal of being able to use regular malloc() and mmap() allocations for OpenCL implementations that support fine-grained system allocations. (See https://github.com/KhronosGroup/OpenCL-Docs/issues/392)

So if the size isn't specified, the most logical choices are do nothing and return OK, return an error, or assume that at least one byte is being migrated and try migrate it.

#3 lore raw
Quoted 54 lines · Ralph Campbell
Quoted 39 lines · Jason Gunthorpe
Quoted 24 lines · Ralph Campbell

The user level OpenCL code shouldn't have to align start and end addresses to a page boundary. That is better handled in the nouveau driver. The npages field is also redundant since it can be computed from the start and end addresses.

This is for Ben Skegg's nouveau tree.

I have been working with Karol Herbst on the OpenCL mesa changes for nouveau which will be merged upstream soon.
With or without those changes, the user visible effect of this patch only extends the range by one page (round up vs. round down to page boundary).

  drivers/gpu/drm/nouveau/nouveau_svm.c | 17 ++++++-----------
  1 file changed, 6 insertions(+), 11 deletions(-)
drivers/gpu/drm/nouveau/nouveau_svm.c +1 -0
index 2df1c0460559..888aa0908c5a 100644
+++ b/drivers/gpu/drm/nouveau/nouveau_svm.c
@@ -105,11 +105,14 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
  	struct nouveau_cli *cli = nouveau_cli(file_priv);
  	struct drm_nouveau_svm_bind *args = data;

drivers/gpu/drm/nouveau/nouveau_svm.c nouveau_svmm_bind(struct drm_device *dev, void *data, · line 114

Quoted 10 lines
drivers/gpu/drm/nouveau/nouveau_svm.c +5 -2
  	unsigned target, cmd, priority;
-	unsigned long addr, end, size;
+	unsigned long addr, end;
  	struct mm_struct *mm;
  	args->va_start &= PAGE_MASK;
-	args->va_end &= PAGE_MASK;
+	args->va_end = ALIGN(args->va_end, PAGE_SIZE);
+	/* If no end address is given, assume a single page. */
+	if (args->va_end == 0)
+		args->va_end = args->va_start + PAGE_SIZE;

That is really weird, how is it useful for the kernel to map a region of unknown size and alignment to the GPU?

Quoted 1 line

Jason

I agree it is somewhat weird. The OpenCL 2.2 specification says that clEnqueueSVMMigrateMem() takes an array of pointers and sizes (in bytes) but the size is optional. There is no alignment required.
This "works" because the pointers have to be allocated with clSVMAlloc() and presumably, the implementation for clSVMAlloc()
keeps track of the length allocated and can fill that in if size is zero. However, requiring all allocations to be made with clSVMAlloc() defeats the goal of being able to use regular malloc() and mmap() allocations for OpenCL implementations that support fine-grained system allocations. (See https://github.com/KhronosGroup/OpenCL-Docs/issues/392)

Quoted 3 lines

So if the size isn't specified, the most logical choices are do nothing and return OK, return an error, or assume that at least one byte is being migrated and try migrate it.

So if the app migrates the wrong memory then nothing bad happens, it just might not get the performance from migration? Seems find but really weird.

Jason

#4 lore raw
Quoted 59 lines · Jason Gunthorpe
Quoted 54 lines · Ralph Campbell
Quoted 39 lines · Jason Gunthorpe
Quoted 24 lines · Ralph Campbell

The user level OpenCL code shouldn't have to align start and end addresses to a page boundary. That is better handled in the nouveau driver. The npages field is also redundant since it can be computed from the start and end addresses.

This is for Ben Skegg's nouveau tree.

I have been working with Karol Herbst on the OpenCL mesa changes for nouveau which will be merged upstream soon.
With or without those changes, the user visible effect of this patch only extends the range by one page (round up vs. round down to page boundary).

   drivers/gpu/drm/nouveau/nouveau_svm.c | 17 ++++++-----------
   1 file changed, 6 insertions(+), 11 deletions(-)
drivers/gpu/drm/nouveau/nouveau_svm.c +1 -0
index 2df1c0460559..888aa0908c5a 100644
+++ b/drivers/gpu/drm/nouveau/nouveau_svm.c
@@ -105,11 +105,14 @@ nouveau_svmm_bind(struct drm_device *dev, void *data,
   	struct nouveau_cli *cli = nouveau_cli(file_priv);
   	struct drm_nouveau_svm_bind *args = data;

drivers/gpu/drm/nouveau/nouveau_svm.c nouveau_svmm_bind(struct drm_device *dev, void *data, · line 114

Quoted 10 lines
drivers/gpu/drm/nouveau/nouveau_svm.c +5 -2
   	unsigned target, cmd, priority;
-	unsigned long addr, end, size;
+	unsigned long addr, end;
   	struct mm_struct *mm;
   	args->va_start &= PAGE_MASK;
-	args->va_end &= PAGE_MASK;
+	args->va_end = ALIGN(args->va_end, PAGE_SIZE);
+	/* If no end address is given, assume a single page. */
+	if (args->va_end == 0)
+		args->va_end = args->va_start + PAGE_SIZE;

That is really weird, how is it useful for the kernel to map a region of unknown size and alignment to the GPU?

Quoted 1 line

Jason

I agree it is somewhat weird. The OpenCL 2.2 specification says that clEnqueueSVMMigrateMem() takes an array of pointers and sizes (in bytes) but the size is optional. There is no alignment required.
This "works" because the pointers have to be allocated with clSVMAlloc() and presumably, the implementation for clSVMAlloc()
keeps track of the length allocated and can fill that in if size is zero. However, requiring all allocations to be made with clSVMAlloc() defeats the goal of being able to use regular malloc() and mmap() allocations for OpenCL implementations that support fine-grained system allocations. (See https://github.com/KhronosGroup/OpenCL-Docs/issues/392)

Quoted 3 lines

So if the size isn't specified, the most logical choices are do nothing and return OK, return an error, or assume that at least one byte is being migrated and try migrate it.

Quoted 6 lines

So if the app migrates the wrong memory then nothing bad happens, it just might not get the performance from migration? Seems find but really weird.

Jason

Given the principal of least surprise, I'm thinking the better choice is to return an error from the driver and leave any other actions to the user level library. I'll post a v2.