Setting register to force ordering to prevent read/write or write/read
hazards for un-cached modes.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c| 22 +--
drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c| 8 +++
.../include/asic_reg/gc
Setting register to force ordering to prevent read/write or write/read
hazards for un-cached modes.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c| 8
.../gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_offset.h| 2 ++
2 files changed, 10
ref count functionality was missing on fuse/dax.c.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
---
fs/dax.c| 4 ++--
fs/ext4/inode.c | 5 +
fs/fuse/dax.c | 4 +---
fs/xfs/xfs_file.c | 4 +---
include/linux/dax.h | 10 ++
5 files changed, 15
1-2 Rebased Ralph Campbell's ZONE_DEVICE page refcounting patches.
Patches 4-5 are for context to show how we are looking up the SPM
memory and registering it with devmap.
Patches 3,6-8 are the changes we are trying to upstream or rework to
make them acceptable upstream.
Patches 9-13 add ZON
.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 17 -
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
index 21b86c35a4f2
claim it as MEMORY_DEVICE_GENERIC using
devm_memremap_pages.
Signed-off-by: Alex Sierra
---
kernel/resource.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/kernel/resource.c b/kernel/resource.c
index 627e61b0c124..2d4abd395c73 100644
--- a/kernel/resource.c
+++ b/kernel/resource.c
@@ -783,6
Two helpers added. One checks if zone device page is generic
type. The other if page is either private or generic type.
Signed-off-by: Alex Sierra
---
include/linux/mm.h | 8
1 file changed, 8 insertions(+)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index c0fcb47d7641
supports two different SP ranges
of at least 256MB size. This could be specified in the
kernel parameter variable efi_fake_mem. Ex. Two SP ranges
of 1GB starting at 0x1 & 0x14000 physical address.
efi_fake_mem=1G@0x1:0x4,1G@0x14000:0x4
Signed-off-by: Alex Si
Device generic type case added for migrate_vma_pages and
migrate_vma_check_page helpers.
Both, generic and private device types have the same
conditions to decide to migrate pages from/to device
memory.
Signed-off-by: Alex Sierra
---
mm/migrate.c | 20 +---
1 file changed, 9
Add two more parameters to set spm_addr_dev0 & spm_addr_dev1
addresses. These two parameters configure the start SP
addresses for each device in test_hmm driver.
Consequently, this configures zone device type as generic.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/test_hmm.sh
ed to zero.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
---
arch/powerpc/kvm/book3s_hv_uvmem.c | 2 +-
drivers/gpu/drm/nouveau/nouveau_dmem.c | 2 +-
fs/dax.c | 4 +-
include/linux/dax.h| 2 +-
include/linux/memre
Add MEMORY_DEVICE_GENERIC case to free_zone_device_page callback.
Device generic type memory case is now able to free its pages properly.
Signed-off-by: Alex Sierra
---
mm/memremap.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/mm/memremap.c b/mm/memremap.c
index
migrate_ping_pong test case added to test explicit migration
from device to sys memory for both private and generic
zone types.
Helpers to migrate from device to sys memory and vicerversa
were also added.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/hmm-tests.c | 142
new ioctl cmd added to query zone device type. This will be
used once the test_hmm adds zone device generic type.
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 15 ++-
lib/test_hmm_uapi.h | 7 +++
2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/lib
Generic device type memory on VRAM to RAM migration,
has similar access as System RAM from the CPU. This flag sets
the source from the sender. Which in Generic type case,
should be set as SYSTEM.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd
ap.type = MEMORY_DEVICE_PRIVATE at
dmirror_allocate_chunk that was forcing to configure pagemap.type
to MEMORY_DEVICE_PRIVATE
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 46 -
lib/test_hmm_uapi.h | 1 +
2 files changed, 34 insertions(+), 13 deletions(-)
diff --gi
g directories:
$sudo mkdir /mnt/ram0
$sudo mkdir /mnt/ram1
Everytime you boot, you need to create ext4 file system for the emulated
persistent memory partitions.
$sudo mkfs.ext4 /dev/pmem0
$sudo mkfs.ext4 /dev/pmem1
To run the tests:
$sudo ./check -g quick
Alex Sierra (11):
kernel: resource: lo
ref count functionality was missing on fuse/dax.c.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
---
fs/dax.c| 4 ++--
fs/ext4/inode.c | 5 +
fs/fuse/dax.c | 4 +---
fs/xfs/xfs_file.c | 4 +---
include/linux/dax.h | 10 ++
5 files changed, 15
claim it as MEMORY_DEVICE_GENERIC using
devm_memremap_pages.
Signed-off-by: Alex Sierra
---
kernel/resource.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/kernel/resource.c b/kernel/resource.c
index ca9f5198a01f..227fc9fab573 100644
--- a/kernel/resource.c
+++ b/kernel/resource.c
@@ -772,6
.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 17 -
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
b/drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
index dab290a4d19d
ed to zero.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
---
arch/powerpc/kvm/book3s_hv_uvmem.c | 2 +-
drivers/gpu/drm/nouveau/nouveau_dmem.c | 2 +-
fs/dax.c | 4 +-
include/linux/dax.h| 2 +-
include/linux/memre
Generic device type memory on VRAM to RAM migration,
has similar access as System RAM from the CPU. This flag sets
the source from the sender. Which in Generic type case,
should be set as SYSTEM.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd
Two helpers added. One checks if zone device page is generic
type. The other if page is either private or generic type.
Signed-off-by: Alex Sierra
---
include/linux/mm.h | 8
1 file changed, 8 insertions(+)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index d48a1f0889d1
ap.type = MEMORY_DEVICE_PRIVATE at
dmirror_allocate_chunk that was forcing to configure pagemap.type
to MEMORY_DEVICE_PRIVATE
v6:
Check for null pointers for resource and memremap references
at dmirror_allocate_chunk
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 56 ---
Add MEMORY_DEVICE_GENERIC case to free_zone_device_page callback.
Device generic type memory case is now able to free its pages properly.
Signed-off-by: Alex Sierra
---
mm/memremap.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/mm/memremap.c b/mm/memremap.c
index
Device generic type case added for migrate_vma_pages and
migrate_vma_check_page helpers.
Both, generic and private device types have the same
conditions to decide to migrate pages from/to device
memory.
Signed-off-by: Alex Sierra
---
mm/migrate.c | 18 +++---
1 file changed, 11
Add two more parameters to set spm_addr_dev0 & spm_addr_dev1
addresses. These two parameters configure the start SP
addresses for each device in test_hmm driver.
Consequently, this configures zone device type as generic.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/test_hmm.sh
new ioctl cmd added to query zone device type. This will be
used once the test_hmm adds zone device generic type.
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 15 ++-
lib/test_hmm_uapi.h | 7 +++
2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/lib
supports two different SP ranges
of at least 256MB size. This could be specified in the
kernel parameter variable efi_fake_mem. Ex. Two SP ranges
of 1GB starting at 0x1 & 0x14000 physical address.
efi_fake_mem=1G@0x1:0x4,1G@0x14000:0x4
Signed-off-by: Alex Si
migrate_ping_pong test case added to test explicit migration
from device to sys memory for both private and generic
zone types.
Helpers to migrate from device to sys memory and vicerversa
were also added.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/hmm-tests.c | 142
Public device type memory on VRAM to RAM migration,
has similar access as System RAM from the CPU. This flag sets
the source from the sender. Which in Public type case,
should be set as IOMEM.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c
Ref counter from device pages is init to zero during memmap init zone.
The first time a new device page is allocated to migrate data into it,
its ref counter needs to be initialized to one.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 2 +-
1 file changed, 1
supports two different SP ranges of at least
256MB size. This could be specified in the kernel parameter variable
efi_fake_mem. Ex. Two SP ranges of 1GB starting at 0x1 &
0x14000 physical address. Ex.
efi_fake_mem=1G@0x1:0x4,1G@0x14000:0x4
Signed-off-by: Alex Si
In order to configure device public in test_hmm, two module parameters
should be passed, which correspond to the SP start address of each
device (2) spm_addr_dev0 & spm_addr_dev1. If no parameters are passed,
private device type is configured.
Signed-off-by: Alex Sierra
---
v5:
Remove de
In this case, this is used to migrate pages from device memory, back to
system memory. This particular device memory type should be accessible
by the CPU, through IOMEM access. Typically, zone device public type
memory falls into this category.
Signed-off-by: Alex Sierra
---
include/linux
Add device public type case to migrate_vma_insert_page,
migrate_vma_pages and migrate_vma_check_page helpers.
Signed-off-by: Alex Sierra
---
mm/migrate.c | 19 ---
1 file changed, 12 insertions(+), 7 deletions(-)
diff --git a/mm/migrate.c b/mm/migrate.c
index d4ae2da99607
From: Ralph Campbell
There are several places where ZONE_DEVICE struct pages assume a reference
count == 1 means the page is idle and free. Instead of open coding this,
add a helper function to hide this detail.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph
migrate_ping_pong test case added to test explicit migration
from device to sys memory for both private and public
zone types.
Helpers to migrate from device to sys memory and vicerversa
were also added.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/hmm-tests.c | 142
.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
v7:
Remove lookup_resource call, so export symbol for this function
is not longer required. Patch dropped "kernel: resource:
lookup_resource as exported symbol"
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 29 +++--
count doesn't need to
be treated specially for ZONE_DEVICE.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
---
v2:
AS: merged this patch in linux 5.11 version
v5:
AS: add condition at try_grab_page to check for the zone device type, while
page ref counter is checked less/equal to
-off-by: Alex Sierra
---
include/linux/memremap.h | 8
include/linux/mm.h | 8
mm/memcontrol.c | 6 +++---
mm/memory-failure.c | 6 +-
mm/memremap.c| 1 +
5 files changed, 25 insertions(+), 4 deletions(-)
diff --git a/include/linux/memremap.h b
new ioctl cmd added to query zone device type. This will be
used once the test_hmm adds zone device public type.
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 15 ++-
lib/test_hmm_uapi.h | 7 +++
2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/lib
previous
reviews of this patch series (see
https://patchwork.freedesktop.org/series/90706/).
Finally, we extended hmm_test to cover migration of MEMORY_DEVICE_PUBLIC.
This work is based on HMM and our SVM memory manager, which has landed
in Linux 5.14 recently.
Alex Sierra (12):
mm: add iomem
Add MEMORY_DEVICE_PUBLIC case to free_zone_device_page callback.
Device public type memory case is now able to free its pages properly.
Signed-off-by: Alex Sierra
---
mm/memremap.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/mm/memremap.c b/mm/memremap.c
index
Add two more parameters to set spm_addr_dev0 & spm_addr_dev1
addresses. These two parameters configure the start SP
addresses for each device in test_hmm driver.
Consequently, this configures zone device type as public.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/test_hmm.sh
During svm restore pages interrupt handler, kfd_process ref count was
never dropped when xnack was disabled. Therefore, the object was never
released.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a
From: Ralph Campbell
There are several places where ZONE_DEVICE struct pages assume a reference
count == 1 means the page is idle and free. Instead of open coding this,
add a helper function to hide this detail.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph
Minor changes:
Swap patch order 03 and 04.
Addings
Add VM_BUG_ON_PAGE(page_ref_count(page), page) to patch 'drm/amdkfd:
ref count init for device pages', to make sure page hasn't been used
Alex Sierra (10):
mm: add zone device public type memory support
mm: add device public vma
-off-by: Alex Sierra
---
include/linux/memremap.h | 8
include/linux/mm.h | 8
mm/memcontrol.c | 6 +++---
mm/memory-failure.c | 6 +-
mm/memremap.c| 2 ++
mm/migrate.c | 19 ---
6 files changed, 38 insertions
This case is used to migrate pages from device memory, back to system
memory. Device public type memory is cache coherent from device and CPU
point of view.
Signed-off-by: Alex Sierra
---
v2:
condition added when migrations from device public pages.
---
include/linux/migrate.h | 1 +
mm
Public device type memory on VRAM to RAM migration, has similar access
as System RAM from the CPU. This flag sets the source from the sender.
Which in Public type case, should be set as
MIGRATE_VMA_SELECT_DEVICE_PUBLIC.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm
Ref counter from device pages is init to zero during memmap init zone.
The first time a new device page is allocated to migrate data into it,
its ref counter needs to be initialized to one.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 3 ++-
1 file changed, 2
count doesn't need to
be treated specially for ZONE_DEVICE.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph Hellwig
---
v2:
AS: merged this patch in linux 5.11 version
v5:
AS: add condition at try_grab_page to check for the zone device type, while
pag
.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
v7:
Remove lookup_resource call, so export symbol for this function
is not longer required. Patch dropped "kernel: resource:
lookup_resource as exported symbol"
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 32 +++--
In order to configure device public in test_hmm, two module parameters
should be passed, which correspond to the SP start address of each
device (2) spm_addr_dev0 & spm_addr_dev1. If no parameters are passed,
private device type is configured.
Signed-off-by: Alex Sierra
---
v5:
Remove de
supports two different SP ranges of at least
256MB size. This could be specified in the kernel parameter variable
efi_fake_mem. Ex. Two SP ranges of 1GB starting at 0x1 &
0x14000 physical address. Ex.
efi_fake_mem=1G@0x1:0x4,1G@0x14000:0x4
Signed-off-by: Alex Si
Add two more parameters to set spm_addr_dev0 & spm_addr_dev1
addresses. These two parameters configure the start SP
addresses for each device in test_hmm driver.
Consequently, this configures zone device type as public.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/test_hmm.sh
migrate_ping_pong test case added to test explicit migration
from device to sys memory for both private and public
zone types.
Helpers to migrate from device to sys memory and vicerversa
were also added.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/hmm-tests.c | 142
new ioctl cmd added to query zone device type. This will be
used once the test_hmm adds zone device public type.
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 15 ++-
lib/test_hmm_uapi.h | 7 +++
2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/lib
svm_range_list svms declaration removed to avoid werror when
CONFIG_HSA_AMD_SVM is not enabled.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
b
x27;t need to hold the BO reservation anymore during validation
and mapping. Now the physical addresses are taken from hmm_range_fault.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 5 -
1 file changed, 5 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm
x27;t need to hold the BO reservation anymore during validation
and mapping. Now the physical addresses are taken from hmm_range_fault.
We also take migrate_mutex to prevent range migration while
validate_and_map update GPU page table.
Signed-off-by: Alex Sierra
Signed-off-by: Felix Kue
builds on HMM and our SVM memory manager already merged in 5.14.
We would like to complete review and merge this migration patchset for
5.16.
Alex Sierra (10):
mm: add zone device coherent type memory support
mm: add device coherent vma selection for memory migration
drm/amdkfd: ref count init
From: Ralph Campbell
There are several places where ZONE_DEVICE struct pages assume a reference
count == 1 means the page is idle and free. Instead of open coding this,
add a helper function to hide this detail.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph
-off-by: Alex Sierra
---
include/linux/memremap.h | 8
include/linux/mm.h | 8
mm/memcontrol.c | 6 +++---
mm/memory-failure.c | 6 +-
mm/memremap.c| 2 ++
mm/migrate.c | 19 ---
6 files changed, 38 insertions
count doesn't need to
be treated specially for ZONE_DEVICE.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph Hellwig
---
v2:
AS: merged this patch in linux 5.11 version
v5:
AS: add condition at try_grab_page to check for the zone device type, while
pag
This case is used to migrate pages from device memory, back to system
memory. Device coherent type memory is cache coherent from device and CPU
point of view.
Signed-off-by: Alex Sierra
---
v2:
condition added when migrations from device coherent pages.
---
include/linux/migrate.h | 1 +
mm
Ref counter from device pages is init to zero during memmap init zone.
The first time a new device page is allocated to migrate data into it,
its ref counter needs to be initialized to one.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 3 ++-
1 file changed, 2
.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
v7:
Remove lookup_resource call, so export symbol for this function
is not longer required. Patch dropped "kernel: resource:
lookup_resource as exported symbol"
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 32 +++--
Coherent device type memory on VRAM to RAM migration, has similar access
as System RAM from the CPU. This flag sets the source from the sender.
Which in Coherent type case, should be set as
MIGRATE_VMA_SELECT_DEVICE_COHERENT.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers
new ioctl cmd added to query zone device type. This will be
used once the test_hmm adds zone device coherent type.
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 15 ++-
lib/test_hmm_uapi.h | 7 +++
2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/lib
In order to configure device coherent in test_hmm, two module parameters
should be passed, which correspond to the SP start address of each
device (2) spm_addr_dev0 & spm_addr_dev1. If no parameters are passed,
private device type is configured.
Signed-off-by: Alex Sierra
---
lib/test_h
Add two more parameters to set spm_addr_dev0 & spm_addr_dev1
addresses. These two parameters configure the start SP
addresses for each device in test_hmm driver.
Consequently, this configures zone device type as coherent.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/test_hm
migrate_ping_pong test case added to test explicit migration
from device to sys memory for both private and coherent zone
types.
Helpers to migrate from device to sys memory and vicerversa
were also added.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/hmm-tests.c | 137
supports two different SP ranges of at least
256MB size. This could be specified in the kernel parameter variable
efi_fake_mem. Ex. Two SP ranges of 1GB starting at 0x1 &
0x14000 physical address. Ex.
efi_fake_mem=1G@0x1:0x4,1G@0x14000:0x4
Signed-off-by: Alex Si
From: Ralph Campbell
There are several places where ZONE_DEVICE struct pages assume a reference
count == 1 means the page is idle and free. Instead of open coding this,
add a helper function to hide this detail.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph
This patch cleans up ZONE_DEVICE page refcounting. Quoting Matthew
Wilcox, "it removes a ton of cruft from every call to put_page()"
This work was originally done by Ralph Campbell and submitted as RFC.
As of today, it has been ack by Theodore Ts'o / Darrick J. Wong, and
reviewed by Christoph Hellw
count doesn't need to
be treated specially for ZONE_DEVICE.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph Hellwig
---
v2:
AS: merged this patch in linux 5.11 version
v5:
AS: add condition at try_grab_page to check for the zone device type, while
pag
ction. If equal, the migration will be
ignored.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 4
drivers/gpu/drm/amd/amdkfd/kfd_priv.h| 1 +
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 3 +++
3 files changed, 8 insertions(+)
diff --git a/drivers/gpu/drm/amd/
mber is checked against "current" reference at
svm_migrate_to_ram callback function. If equal, the migration will be
ignored.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 4
drivers/gpu/drm/amd/amdkfd/kfd_priv.h| 1 +
drivers/gpu/drm/amd/amdkfd/kfd_
space to map memory at this low
region.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
index 2e86692def19
.
Signed-off-by: Alex Sierra
---
include/linux/memremap.h | 8
include/linux/mm.h | 9 +
mm/memcontrol.c | 6 +++---
mm/memory-failure.c | 6 +-
mm/memremap.c| 5 -
mm/migrate.c | 21 +
6 files changed, 42
.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
v7:
Remove lookup_resource call, so export symbol for this function
is not longer required. Patch dropped "kernel: resource:
lookup_resource as exported symbol"
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 29 +++--
new ioctl cmd added to query zone device type. This will be
used once the test_hmm adds zone device coherent type.
Signed-off-by: Alex Sierra
---
lib/test_hmm.c | 15 ++-
lib/test_hmm_uapi.h | 7 +++
2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/lib
Coherent device type memory on VRAM to RAM migration, has similar access
as System RAM from the CPU. This flag sets the source from the sender.
Which in Coherent type case, should be set as
MIGRATE_VMA_SELECT_DEVICE_COHERENT.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
drivers
Add two more parameters to set spm_addr_dev0 & spm_addr_dev1
addresses. These two parameters configure the start SP
addresses for each device in test_hmm driver.
Consequently, this configures zone device type as coherent.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/test_hm
This case is used to migrate pages from device memory, back to system
memory. Device coherent type memory is cache coherent from device and CPU
point of view.
Signed-off-by: Alex Sierra
---
v2:
condition added when migrations from device coherent pages.
---
include/linux/migrate.h | 1 +
mm
In order to configure device coherent in test_hmm, two module parameters
should be passed, which correspond to the SP start address of each
device (2) spm_addr_dev0 & spm_addr_dev1. If no parameters are passed,
private device type is configured.
Signed-off-by: Alex Sierra
---
lib/test_h
supports two different SP ranges of at least
256MB size. This could be specified in the kernel parameter variable
efi_fake_mem. Ex. Two SP ranges of 1GB starting at 0x1 &
0x14000 physical address. Ex.
efi_fake_mem=1G@0x1:0x4,1G@0x14000:0x4
Signed-off-by: Alex Si
and our SVM memory manager already merged in
5.15.
Alex Sierra (9):
mm: add zone device coherent type memory support
mm: add device coherent vma selection for memory migration
drm/amdkfd: add SPM support for SVM
drm/amdkfd: coherent type as sys mem on migration to ram
lib: test_hmm add
migrate_ping_pong test case added to test explicit migration
from device to sys memory for both private and coherent zone
types.
Helpers to migrate from device to sys memory and vicerversa
were also added.
Signed-off-by: Alex Sierra
---
tools/testing/selftests/vm/hmm-tests.c | 156
This patch includes Ralph Campbell’s ZONE_DEVICE refcount cleanup and
additionally the changes necessary to avoid breaking the separately
submitted MEMORY_DEVICE_COHERENT page migration code.
Ralph’s original description:
ZONE_DEVICE struct pages have an extra reference count that complicates
the c
count doesn't need to
be treated specially for ZONE_DEVICE.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph Hellwig
---
arch/powerpc/kvm/book3s_hv_uvmem.c | 2 +-
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 3 +-
drivers/gpu/drm/nouveau/nouveau_d
From: Ralph Campbell
There are several places where ZONE_DEVICE struct pages assume a reference
count == 1 means the page is idle and free. Instead of open coding this,
add a helper function to hide this detail.
Signed-off-by: Ralph Campbell
Signed-off-by: Alex Sierra
Reviewed-by: Christoph
This case is used to migrate pages from device memory, back to system
memory. Device coherent type memory is cache coherent from device and CPU
point of view.
Signed-off-by: Alex Sierra
---
v2:
condition added when migrations from device coherent pages.
---
include/linux/migrate.h | 1 +
mm
new hmm_gup_test has been added to the hmm-test to test this
functionality. It makes use of the gup_test module to long term pin
user pages that have been migrate to device memory first.
- Other patch corrections made by Felix, Alistair and Christoph.
Alex Sierra (11):
mm: add zone device cohe
Avoid long term pinning for Coherent device type pages. This could
interfere with their own device memory manager.
If caller tries to get user device coherent pages with PIN_LONGTERM flag
set, those pages will be migrated back to system memory.
Signed-off-by: Alex Sierra
---
mm/gup.c | 32
.
Signed-off-by: Alex Sierra
Reviewed-by: Felix Kuehling
---
v7:
Remove lookup_resource call, so export symbol for this function
is not longer required. Patch dropped "kernel: resource:
lookup_resource as exported symbol"
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 29 +++--
In order to configure device coherent in test_hmm, two module parameters
should be passed, which correspond to the SP start address of each
device (2) spm_addr_dev0 & spm_addr_dev1. If no parameters are passed,
private device type is configured.
Signed-off-by: Alex Sierra
---
lib/test_h
.
Signed-off-by: Alex Sierra
---
include/linux/memremap.h | 8
include/linux/mm.h | 16
mm/memcontrol.c | 6 +++---
mm/memory-failure.c | 8 ++--
mm/memremap.c| 5 -
mm/migrate.c | 21 +
6 files changed
1 - 100 of 491 matches
Mail list logo
