drm/amdgpu: select ARCH_HAS_HMM and ZONE_DEVICE option
Subsystems (2)
PATCH 1/1 drm/amdgpu: select ARCH_HAS_HMM and ZONE_DEVICE option
Those options are needed to support HMM
--- drivers/gpu/drm/amd/amdgpu/Kconfig | 2 ++ 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/amdgpu/Kconfig +2 -0
@@ -26,7 +26,9 @@ config DRM_AMDGPU_CIK config DRM_AMDGPU_USERPTR bool "Always enable userptr write support" depends on DRM_AMDGPU + select ARCH_HAS_HMM select HMM_MIRROR + select ZONE_DEVICE help This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it isn't already selected to enabled full userptr support.
Signature · 6 lines
2.17.1 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): [email protected]
[+Jerome]
Why to we need ZONE_DEVICE. I didn't think this was needed for mirroring CPU page tables to device page tables.
ARCH_HAS_HMM depends on (X86_64 || PPC64). Do we have some alternative for ARM support?
Also, the name ARCH_HAS_HMM looks like it's meant to be selected by the CPU architecture rather than any driver. Jerome, do you have any advice?
Thanks,
Felix
Quoted 13 lines · Philip
Those options are needed to support HMM
Change-Id Ieb7bb3bcec07245d79a02793e6728228decc400aSigned-off-by Philip Yang <[email protected]>--- drivers/gpu/drm/amd/amdgpu/Kconfig | 2 ++ 1 file changed, 2 insertions(+)drivers/gpu/drm/amd/amdgpu/Kconfig +0
-0
drivers/gpu/drm/amd/amdgpu/Kconfig config DRM_AMDGPU_CIK · line 34
Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/Kconfig +2
-0@@ -26,7 +26,9 @@ config DRM_AMDGPU_CIK config DRM_AMDGPU_USERPTR bool "Always enable userptr write support" depends on DRM_AMDGPU + select ARCH_HAS_HMM select HMM_MIRROR + select ZONE_DEVICE help This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it isn't already selected to enabled full userptr support.
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): 1dfb54a4-67bf-2e96-40f4-7bfe24284232-5C7GfCeVMHo@public.gmane.org
Quoted 10 lines · Felix
[+Jerome]
Why to we need
ZONE_DEVICE. I didn't think this was needed for mirroring CPU page tables to device page tables.
ARCH_HAS_HMMdepends on (X86_64|| PPC64). Do we have some alternative for ARM support?Also, the name
ARCH_HAS_HMMlooks like it's meant to be selected by the CPU architecture rather than any driver. Jerome, do you have any advice?
This patch is wrong you need to depend on ARCH_HAS_HMM and select HMM_MIRROR you do not need to select ZONE_DEVICE
So it should look like:
config DRM_AMDGPU_USERPTR
bool "Always enable userptr write support" depends on DRM_AMDGPU
depends on ARCH_HAS_HMM
select HMM_MIRROR
help
This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it isn't already selected to enabled full userptr support.
I have not got around to work on amdgpu on that respect yet but it is on my todo list unless someone else beat me to it :)
Cheers,
Jérôme
Quoted 13 lines
Thanks,
FelixQuoted 8 lines · Philip
Those options are needed to support HMM
Change-Id Ieb7bb3bcec07245d79a02793e6728228decc400aSigned-off-by Philip Yang <[email protected]>--- drivers/gpu/drm/amd/amdgpu/Kconfig | 2 ++ 1 file changed, 2 insertions(+)
Quoted 14 lines
Quoted 5 lines
drivers/gpu/drm/amd/amdgpu/Kconfig +0
-0drivers/gpu/drm/amd/amdgpu/Kconfig config DRM_AMDGPU_CIK · line 34
Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/Kconfig +2
-0@@ -26,7 +26,9 @@ config DRM_AMDGPU_CIK config DRM_AMDGPU_USERPTR bool "Always enable userptr write support" depends on DRM_AMDGPU + select ARCH_HAS_HMM select HMM_MIRROR + select ZONE_DEVICE help This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it isn't already selected to enabled full userptr support.
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 6 lines · Jerome Glisse
Quoted 5 lines · Felix
[+Jerome]
Why to we need
ZONE_DEVICE. I didn't think this was needed for mirroring CPU page tables to device page tables.
Quoted 6 lines
Quoted 5 lines
ARCH_HAS_HMMdepends on (X86_64|| PPC64). Do we have some alternative for ARM support?Also, the name
ARCH_HAS_HMMlooks like it's meant to be selected by the CPU architecture rather than any driver. Jerome, do you have any advice?This patch is wrong you need to depend on
ARCH_HAS_HMMand
Who selects ARCH_HAS_HMM? Currently I don't see this selected anywhere. So any config option that depends on it will be invisible in menuconfig. Do we need ARCH_HAS_HMM somewhere in the arch/x86/Kconfig and arch/powerpc/Kconfig?
Also, ARCH_HAS_HMM does not currently support ARM. Does that mean we can't have ARM support in AMDGPU if we start using HMM?
Finally, ARCH_HAS_HMM has a bunch of dependencies. If they are not met, I guess it can't be enabled. Should those be "select"s instead?
config ARCH_HAS_HMM
bool
default y
depends on (X86_64 || PPC64)
depends on ZONE_DEVICE
depends on MMU && 64BIT
depends on MEMORY_HOTPLUG
depends on MEMORY_HOTREMOVE
depends on SPARSEMEM_VMEMMAP
Regards,
Felix
Quoted 31 lines
select
HMM_MIRRORyou do not need to selectZONE_DEVICESo it should look like:
config
DRM_AMDGPU_USERPTR
bool "Always enable userptr write support" depends onDRM_AMDGPU
depends onARCH_HAS_HMM
selectHMM_MIRROR
help
This option selectsCONFIG_HMMandCONFIG_HMM_MIRRORif it isn't already selected to enabled full userptr support.I have not got around to work on amdgpu on that respect yet but it is on my todo list unless someone else beat me to it :)
Cheers,
JérômeQuoted 12 lines
Thanks,
FelixQuoted 8 lines · Philip
Those options are needed to support HMM
Change-Id Ieb7bb3bcec07245d79a02793e6728228decc400aSigned-off-by Philip Yang <[email protected]>--- drivers/gpu/drm/amd/amdgpu/Kconfig | 2 ++ 1 file changed, 2 insertions(+)
Quoted 14 lines
Quoted 6 lines
Quoted 6 lines
drivers/gpu/drm/amd/amdgpu/Kconfig +0
-0@@ -26,7 +26,9 @@ config DRM_AMDGPU_CIK config DRM_AMDGPU_USERPTRQuoted 8 lines
drivers/gpu/drm/amd/amdgpu/Kconfig config DRM_AMDGPU_CIK · line 34
Quoted 8 lines
drivers/gpu/drm/amd/amdgpu/Kconfig +2
-0bool "Always enable userptr write support" depends on DRM_AMDGPU + select ARCH_HAS_HMM select HMM_MIRROR + select ZONE_DEVICE help This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it isn't already selected to enabled full userptr support.
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): d52f74c1-a9c2-d265-993d-dae3d367bf70-5C7GfCeVMHo@public.gmane.org
Quoted 14 lines · Felix
Quoted 6 lines · Jerome Glisse
Quoted 5 lines · Felix
[+Jerome]
Why to we need
ZONE_DEVICE. I didn't think this was needed for mirroring CPU page tables to device page tables.Quoted 6 lines
Quoted 5 lines
ARCH_HAS_HMMdepends on (X86_64|| PPC64). Do we have some alternative for ARM support?Also, the name
ARCH_HAS_HMMlooks like it's meant to be selected by the CPU architecture rather than any driver. Jerome, do you have any advice?This patch is wrong you need to depend on
ARCH_HAS_HMMand
Quoted 7 lines
Who selects
ARCH_HAS_HMM? Currently I don't see this selected anywhere. So any config option that depends on it will be invisible in menuconfig. Do we needARCH_HAS_HMMsomewhere in thearch/x86/Kconfigandarch/powerpc/Kconfig?Also,
ARCH_HAS_HMMdoes not currently support ARM. Does that mean we can't have ARM support in AMDGPU if we start using HMM?
ARCH_HAS_HMM is defined by architecture that support HMM. So par x86 and PPC. It should not be hard to add it to ARM (i can not remember if ARM has DAX yet or not, if ARM does not have DAX then you need to add that first).
Quoted 3 lines
Finally,
ARCH_HAS_HMMhas a bunch of dependencies. If they are not met, I guess it can't be enabled. Should those be "select"s instead?
No they should not be selected, people configuring their system need to have the freedom of doing so. All those option are selected in all the big distribution.
Quoted 10 lines
config
ARCH_HAS_HMM
bool
default y
depends on (X86_64|| PPC64)
depends onZONE_DEVICE
depends on MMU && 64BIT
depends onMEMORY_HOTPLUG
depends onMEMORY_HOTREMOVE
depends onSPARSEMEM_VMEMMAP
Cheers,
Jérôme
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 19 lines · Jerome Glisse
Quoted 18 lines · Felix
Quoted 6 lines · Jerome Glisse
Quoted 5 lines · Felix
[+Jerome]
Why to we need
ZONE_DEVICE. I didn't think this was needed for mirroring CPU page tables to device page tables.Quoted 6 lines
Quoted 5 lines
ARCH_HAS_HMMdepends on (X86_64|| PPC64). Do we have some alternative for ARM support?Also, the name
ARCH_HAS_HMMlooks like it's meant to be selected by the CPU architecture rather than any driver. Jerome, do you have any advice?This patch is wrong you need to depend on
ARCH_HAS_HMMandWho selects
ARCH_HAS_HMM? Currently I don't see this selected anywhere. So any config option that depends on it will be invisible in menuconfig. Do we needARCH_HAS_HMMsomewhere in thearch/x86/Kconfigandarch/powerpc/Kconfig?
Quoted 6 lines
Quoted 2 lines
Also,
ARCH_HAS_HMMdoes not currently support ARM. Does that mean we can't have ARM support in AMDGPU if we start using HMM?
ARCH_HAS_HMMis defined by architecture that support HMM. So par x86 and PPC. It should not be hard to add it to ARM (i can not remember if ARM has DAX yet or not, if ARM does not have DAX then you need to add that first).
Not having ARM support is a bummer. I just enabled KFD on ARM a few weeks ago. Now depending on HMM makes KFD unusable on ARM. [+Mark FYI] I hope this is only a temporary setback.
Quoted 5 lines
Quoted 2 lines
Finally,
ARCH_HAS_HMMhas a bunch of dependencies. If they are not met, I guess it can't be enabled. Should those be "select"s instead?No they should not be selected, people configuring their system need to have the freedom of doing so. All those option are selected in all the big distribution.
As far as I can tell, the arch/x86/Kconfig doesn't select ARCH_HAS_HMM. Its default is "y", so it should be enabled on anything that meets the dependencies. But ZONE_DEVICE was not enabled by default. I think that's what broke our kernel configs.
We'll fix our own kernel configs to enable ZONE_DEVICE and ARCH_HAS_HMM to get our internal builds to work again.
I suspect other users with their own kernel configs will stumble over this and wonder why KFD and userptr support are disabled in their builds.
Regards,
Felix
Quoted 11 lines
Quoted 11 lines
config
ARCH_HAS_HMM
bool
default y
depends on (X86_64|| PPC64)
depends onZONE_DEVICE
depends on MMU && 64BIT
depends onMEMORY_HOTPLUG
depends onMEMORY_HOTREMOVE
depends onSPARSEMEM_VMEMMAP
Quoted 2 lines
Cheers,
Jérôme
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): 3b6bcf86-c871-b23d-1da8-fe5b4d4ea081-5C7GfCeVMHo@public.gmane.org
Quoted 22 lines · Felix
Quoted 21 lines · Jerome Glisse
Quoted 20 lines · Felix
Quoted 6 lines · Jerome Glisse
Quoted 5 lines · Felix
[+Jerome]
Why to we need
ZONE_DEVICE. I didn't think this was needed for mirroring CPU page tables to device page tables.Quoted 6 lines
Quoted 5 lines
ARCH_HAS_HMMdepends on (X86_64|| PPC64). Do we have some alternative for ARM support?Also, the name
ARCH_HAS_HMMlooks like it's meant to be selected by the CPU architecture rather than any driver. Jerome, do you have any advice?This patch is wrong you need to depend on
ARCH_HAS_HMMandWho selects
ARCH_HAS_HMM? Currently I don't see this selected anywhere. So any config option that depends on it will be invisible in menuconfig. Do we needARCH_HAS_HMMsomewhere in thearch/x86/Kconfigandarch/powerpc/Kconfig?Also,
ARCH_HAS_HMMdoes not currently support ARM. Does that mean we can't have ARM support in AMDGPU if we start using HMM?
Quoted 8 lines
Quoted 4 lines
ARCH_HAS_HMMis defined by architecture that support HMM. So par x86 and PPC. It should not be hard to add it to ARM (i can not remember if ARM has DAX yet or not, if ARM does not have DAX then you need to add that first).Not having ARM support is a bummer. I just enabled KFD on ARM a few weeks ago. Now depending on HMM makes KFD unusable on ARM. [+Mark FYI] I hope this is only a temporary setback.
It should not be hard to add in fact all it might need is a Kconfig patch. I have no easy access to ARM with PCIE so i have not tackle this yet.
Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
Finally,
ARCH_HAS_HMMhas a bunch of dependencies. If they are not met, I guess it can't be enabled. Should those be "select"s instead?No they should not be selected, people configuring their system need to have the freedom of doing so. All those option are selected in all the big distribution.
Quoted 7 lines
As far as I can tell, the
arch/x86/Kconfigdoesn't selectARCH_HAS_HMM. Its default is "y", so it should be enabled on anything that meets the dependencies. ButZONE_DEVICEwas not enabled by default. I think that's what broke our kernel configs.We'll fix our own kernel configs to enable
ZONE_DEVICEandARCH_HAS_HMMto get our internal builds to work again.
You seem to be doing weird thing with your kconfig ...
Quoted 3 lines
I suspect other users with their own kernel configs will stumble over this and wonder why KFD and userptr support are disabled in their builds.
Patch to improve kconfig are welcome but they should not force select thing. Configuration is there to give user freedom to select fewature they want to give up.
Maybe following would help:ARCH_HAS_HMM
- bool
- default y
+ def_bool y
Cheers,
Jérôme
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): [email protected]
Thanks Jerome for the the correct HMM config option, only select HMM_MIRROR is not good enough because CONFIG_HMM option maybe missing, add depends on ARCH_HAS_HMM will solve the issue.
I will submit new patch to fix the compilation error if HMM_MIRROR config is missing and the HMM config dependency issue.
Philip
Quoted 67 lines · Jerome Glisse
Quoted 22 lines · Felix
Quoted 21 lines · Jerome Glisse
Quoted 20 lines · Felix
Quoted 6 lines · Jerome Glisse
Quoted 5 lines · Felix
[+Jerome]
Why to we need
ZONE_DEVICE. I didn't think this was needed for mirroring CPU page tables to device page tables.Quoted 6 lines
Quoted 5 lines
ARCH_HAS_HMMdepends on (X86_64|| PPC64). Do we have some alternative for ARM support?Also, the name
ARCH_HAS_HMMlooks like it's meant to be selected by the CPU architecture rather than any driver. Jerome, do you have any advice?This patch is wrong you need to depend on
ARCH_HAS_HMMandWho selects
ARCH_HAS_HMM? Currently I don't see this selected anywhere. So any config option that depends on it will be invisible in menuconfig. Do we needARCH_HAS_HMMsomewhere in thearch/x86/Kconfigandarch/powerpc/Kconfig?Also,
ARCH_HAS_HMMdoes not currently support ARM. Does that mean we can't have ARM support in AMDGPU if we start using HMM?Quoted 8 lines
Quoted 4 lines
ARCH_HAS_HMMis defined by architecture that support HMM. So par x86 and PPC. It should not be hard to add it to ARM (i can not remember if ARM has DAX yet or not, if ARM does not have DAX then you need to add that first).Not having ARM support is a bummer. I just enabled KFD on ARM a few weeks ago. Now depending on HMM makes KFD unusable on ARM. [+Mark FYI] I hope this is only a temporary setback.
It should not be hard to add in fact all it might need is a
Kconfigpatch. I have no easy access to ARM with PCIE so i have not tackle this yet.Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
Finally,
ARCH_HAS_HMMhas a bunch of dependencies. If they are not met, I guess it can't be enabled. Should those be "select"s instead?No they should not be selected, people configuring their system need to have the freedom of doing so. All those option are selected in all the big distribution.
Quoted 7 lines
As far as I can tell, the
arch/x86/Kconfigdoesn't selectARCH_HAS_HMM. Its default is "y", so it should be enabled on anything that meets the dependencies. ButZONE_DEVICEwas not enabled by default. I think that's what broke our kernel configs.We'll fix our own kernel configs to enable
ZONE_DEVICEandARCH_HAS_HMMto get our internal builds to work again.You seem to be doing weird thing with your kconfig ...
Quoted 3 lines
I suspect other users with their own kernel configs will stumble over this and wonder why KFD and userptr support are disabled in their builds.
Patch to improve kconfig are welcome but they should not force select thing. Configuration is there to give user freedom to select fewature they want to give up.
Maybe following would help:
ARCH_HAS_HMM(quoted diff) +1
-2- bool - default y + def_bool y
Quoted 3 lines
Cheers,
Jérôme
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx