Message ID | 20200810222657.1841322-1-jcrouse@codeaurora.org (mailing list archive) |
---|---|
Headers | show |
Series | iommu/arm-smmu: Add Adreno SMMU specific implementation | expand |
On Mon, Aug 10, 2020 at 04:26:44PM -0600, Jordan Crouse wrote: > This series adds an Adreno SMMU implementation to arm-smmu to allow GPU hardware > pagetable switching. > > The Adreno GPU has built in capabilities to switch the TTBR0 pagetable during > runtime to allow each individual instance or application to have its own > pagetable. In order to take advantage of the HW capabilities there are certain > requirements needed of the SMMU hardware. "capabilities" is a polite way of putting it ;) Anyway, modulo two design comments, I think this is about as nice as we're going to get this. Thanks for persevering, and sorry that you have to deal with such dreadful hardware. Hopefully the next version will be the one, although I'd like Robin to take a quick look as well if he gets a chance. Will
From: Rob Clark <robdclark@chromium.org>
NOTE: Since Jordan was out today, and I wanted to keep things moving on
this, I took the liberty of respinning his series (originally
titled "iommu/arm-smmu: Add Adreno SMMU specific implementation")
with updates based on Will's review comments, and some fixes and
extra bits that I found in stress testing the series. Original
commit msg and updated version history below.
In general I like the private interface between adreno-smmu and
the GPU driver. It should make for a more straightforward way
to extend things to optimize TLB invalidation in the future, for
example, rather than shoe-horning everything thru domain attrs.
And it lets us describe the get_ttbr1_cfg/set_ttrb0_cfg dance
more clearly. Although it is going to make landing this via
iommu vs drm tree a bit more difficult. Maybe there are some
arm-smmu parts of this series that could be pulled out to make
it not conflicty to land the private interface and adreno-smmu
bits via the drm tree? (But I'm jumping a bit ahead here. Just
wanted to point out that issue.)
The complete series can be found at:
https://gitlab.freedesktop.org/drm/msm/-/commits/msm-next-pgtables
This series adds an Adreno SMMU implementation to arm-smmu to allow GPU hardware
pagetable switching.
The Adreno GPU has built in capabilities to switch the TTBR0 pagetable during
runtime to allow each individual instance or application to have its own
pagetable. In order to take advantage of the HW capabilities there are certain
requirements needed of the SMMU hardware.
This series adds support for an Adreno specific arm-smmu implementation. The new
implementation 1) ensures that the GPU domain is always assigned context bank 0,
2) enables split pagetable support (TTBR1) so that the instance specific
pagetable can be swapped while the global memory remains in place and 3) shares
the current pagetable configuration with the GPU driver to allow it to create
its own io-pgtable instances.
The series then adds the drm/msm code to enable these features. For targets that
support it allocate new pagetables using the io-pgtable configuration shared by
the arm-smmu driver and swap them in during runtime.
This version of the series merges the previous patchset(s) [1] and [2]
with the following improvements:
v13: (Respin by Rob)
- Switch to a private interface between adreno-smmu and GPU driver,
dropping the custom domain attr (Will Deacon)
- Rework the SCTLR.HUPCF patch to add new fields in smmu_domain->cfg
rather than adding new impl hook (Will Deacon)
- Drop for_each_cfg_sme() in favor of plain for() loop (Will Deacon)
- Fix context refcnt'ing issue which was causing problems with GPU
crash recover stress testing.
- Spiff up $debugfs/gem to show process information associated with
VMAs
v12:
- Nitpick cleanups in gpu/drm/msm/msm_iommu.c (Rob Clark)
- Reorg in gpu/drm/msm/msm_gpu.c (Rob Clark)
- Use the default asid for the context bank so that iommu_tlb_flush_all works
- Flush the UCHE after a page switch
- Add the SCTLR.HUPCF patch at the end of the series
v11:
- Add implementation specific get_attr/set_attr functions (per Rob Clark)
- Fix context bank allocation (per Bjorn Andersson)
v10:
- arm-smmu: add implementation hook to allocate context banks
- arm-smmu: Match the GPU domain by stream ID instead of compatible string
- arm-smmu: Make DOMAIN_ATTR_PGTABLE_CFG bi-directional. The leaf driver
queries the configuration to create a pagetable and then sends the newly
created configuration back to the smmu-driver to enable TTBR0
- drm/msm: Add context reference counting for submissions
- drm/msm: Use dummy functions to skip TLB operations on per-instance
pagetables
[1] https://lists.linuxfoundation.org/pipermail/iommu/2020-June/045653.html
[2] https://lists.linuxfoundation.org/pipermail/iommu/2020-June/045659.html
Jordan Crouse (12):
iommu/arm-smmu: Pass io-pgtable config to implementation specific
function
iommu/arm-smmu: Add support for split pagetables
iommu/arm-smmu: Prepare for the adreno-smmu implementation
iommu/arm-smmu-qcom: Add implementation for the adreno GPU SMMU
dt-bindings: arm-smmu: Add compatible string for Adreno GPU SMMU
drm/msm: Add a context pointer to the submitqueue
drm/msm: Drop context arg to gpu->submit()
drm/msm: Set the global virtual address range from the IOMMU domain
drm/msm: Add support to create a local pagetable
drm/msm: Add support for private address space instances
drm/msm/a6xx: Add support for per-instance pagetables
arm: dts: qcom: sm845: Set the compatible string for the GPU SMMU
Rob Clark (7):
drm/msm: remove dangling submitqueue references
iommu: add private interface for adreno-smmu
drm/msm/gpu: add dev_to_gpu() helper
drm/msm: set adreno_smmu as gpu's drvdata
iommu/arm-smmu: constify some helpers
iommu/arm-smmu: add a way for implementations to influence SCTLR
drm/msm: show process names in gem_describe
.../devicetree/bindings/iommu/arm,smmu.yaml | 4 +
arch/arm64/boot/dts/qcom/sdm845.dtsi | 2 +-
drivers/gpu/drm/msm/adreno/a5xx_gpu.c | 12 +-
drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 75 ++++++-
drivers/gpu/drm/msm/adreno/a6xx_gpu.h | 1 +
drivers/gpu/drm/msm/adreno/adreno_device.c | 12 +-
drivers/gpu/drm/msm/adreno/adreno_gpu.c | 18 +-
drivers/gpu/drm/msm/adreno/adreno_gpu.h | 3 +-
drivers/gpu/drm/msm/msm_drv.c | 16 +-
drivers/gpu/drm/msm/msm_drv.h | 25 +++
drivers/gpu/drm/msm/msm_gem.c | 25 ++-
drivers/gpu/drm/msm/msm_gem.h | 6 +
drivers/gpu/drm/msm/msm_gem_submit.c | 8 +-
drivers/gpu/drm/msm/msm_gem_vma.c | 10 +
drivers/gpu/drm/msm/msm_gpu.c | 41 +++-
drivers/gpu/drm/msm/msm_gpu.h | 21 +-
drivers/gpu/drm/msm/msm_gpummu.c | 2 +-
drivers/gpu/drm/msm/msm_iommu.c | 206 +++++++++++++++++-
drivers/gpu/drm/msm/msm_mmu.h | 16 +-
drivers/gpu/drm/msm/msm_ringbuffer.h | 1 +
drivers/gpu/drm/msm/msm_submitqueue.c | 7 +-
drivers/iommu/arm/arm-smmu/arm-smmu-impl.c | 6 +-
drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c | 155 ++++++++++++-
drivers/iommu/arm/arm-smmu/arm-smmu.c | 102 ++++-----
drivers/iommu/arm/arm-smmu/arm-smmu.h | 87 +++++++-
include/linux/adreno-smmu-priv.h | 36 +++
26 files changed, 765 insertions(+), 132 deletions(-)
create mode 100644 include/linux/adreno-smmu-priv.h
On Thu, Aug 13, 2020 at 07:40:55PM -0700, Rob Clark wrote: > From: Rob Clark <robdclark@chromium.org> > > NOTE: Since Jordan was out today, and I wanted to keep things moving on > this, I took the liberty of respinning his series (originally > titled "iommu/arm-smmu: Add Adreno SMMU specific implementation") > with updates based on Will's review comments, and some fixes and > extra bits that I found in stress testing the series. Original > commit msg and updated version history below. > > In general I like the private interface between adreno-smmu and > the GPU driver. It should make for a more straightforward way > to extend things to optimize TLB invalidation in the future, for > example, rather than shoe-horning everything thru domain attrs. > And it lets us describe the get_ttbr1_cfg/set_ttrb0_cfg dance > more clearly. Although it is going to make landing this via > iommu vs drm tree a bit more difficult. Maybe there are some > arm-smmu parts of this series that could be pulled out to make > it not conflicty to land the private interface and adreno-smmu > bits via the drm tree? (But I'm jumping a bit ahead here. Just > wanted to point out that issue.) > > The complete series can be found at: > https://gitlab.freedesktop.org/drm/msm/-/commits/msm-next-pgtables > > This series adds an Adreno SMMU implementation to arm-smmu to allow GPU hardware > pagetable switching. > > The Adreno GPU has built in capabilities to switch the TTBR0 pagetable during > runtime to allow each individual instance or application to have its own > pagetable. In order to take advantage of the HW capabilities there are certain > requirements needed of the SMMU hardware. > > This series adds support for an Adreno specific arm-smmu implementation. The new > implementation 1) ensures that the GPU domain is always assigned context bank 0, > 2) enables split pagetable support (TTBR1) so that the instance specific > pagetable can be swapped while the global memory remains in place and 3) shares > the current pagetable configuration with the GPU driver to allow it to create > its own io-pgtable instances. > > The series then adds the drm/msm code to enable these features. For targets that > support it allocate new pagetables using the io-pgtable configuration shared by > the arm-smmu driver and swap them in during runtime. > > This version of the series merges the previous patchset(s) [1] and [2] > with the following improvements: > > v13: (Respin by Rob) > - Switch to a private interface between adreno-smmu and GPU driver, > dropping the custom domain attr (Will Deacon) > - Rework the SCTLR.HUPCF patch to add new fields in smmu_domain->cfg > rather than adding new impl hook (Will Deacon) > - Drop for_each_cfg_sme() in favor of plain for() loop (Will Deacon) > - Fix context refcnt'ing issue which was causing problems with GPU > crash recover stress testing. > - Spiff up $debugfs/gem to show process information associated with > VMAs I'll add the tags to Rob's code but in general I ack all these changes. I also like the private interface - it gives us the most flexibility without either changing the IOMMU API or giving up entirely and making an internal SMMU implementation inside drm/msm. <snip> Jordan