Re: [PATCH v3 1/4] iommu/arm-smmu: introduction of ACTLR for custom prefetcher settings

[Bibek Kumar Patro]

On 11/27/2023 9:37 PM, Dmitry Baryshkov wrote:
> On Mon, 27 Nov 2023 at 16:54, Bibek Kumar Patro
> <quic_bibekkum@xxxxxxxxxxx> wrote:
> > Currently in Qualcomm  SoCs the default prefetch is set to 1 which allows
> > the TLB to fetch just the next page table. MMU-500 features ACTLR
> > register which is implementation defined and is used for Qualcomm SoCs
> > to have a prefetch setting of 1/3/7/15 enabling TLB to prefetch
> > the next set of page tables accordingly allowing for faster translations.
> >
> > ACTLR value is unique for each SMR (Stream matching register) and stored
> > in a pre-populated table. This value is set to the register during
> > context bank initialisation.
> >
> > Signed-off-by: Bibek Kumar Patro <quic_bibekkum@xxxxxxxxxxx>
> >
> > ---
> >   drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c | 56 +++++++++++++++++++++-
> >   drivers/iommu/arm/arm-smmu/arm-smmu-qcom.h |  6 ++-
> >   drivers/iommu/arm/arm-smmu/arm-smmu.c      |  5 +-
> >   drivers/iommu/arm/arm-smmu/arm-smmu.h      |  5 ++
> >   4 files changed, 68 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c
> > index 7f52ac67495f..4a38cae29be2 100644
> > --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c
> > +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c
> > @@ -14,6 +14,12 @@
> >
> >   #define QCOM_DUMMY_VAL -1
> >
> > +struct actlr_config {
> > +       u16 sid;
> > +       u16 mask;
> > +       u32 actlr;
> > +};
> > +
> >   static struct qcom_smmu *to_qcom_smmu(struct arm_smmu_device *smmu)
> >   {
> >          return container_of(smmu, struct qcom_smmu, smmu);
> > @@ -205,10 +211,40 @@ static bool qcom_adreno_can_do_ttbr1(struct arm_smmu_device *smmu)
> >          return true;
> >   }
> >
> > +static void arm_smmu_set_actlr(struct device *dev, struct arm_smmu_device *smmu, int cbndx,
> > +               const struct actlr_config *actlrcfg, size_t actlrcfg_size)
> > +{
> > +       struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev);
> > +       struct arm_smmu_master_cfg *cfg = dev_iommu_priv_get(dev);
> > +       struct arm_smmu_smr *smr;
> > +       int i;
> > +       int j;
> > +       u16 id;
> > +       u16 mask;
> > +       int idx;
> > +
> > +       for (i = 0; i < actlrcfg_size; ++i) {
> > +               id = (actlrcfg + i)->sid;
> > +               mask = (actlrcfg + i)->mask;
> > +
> > +               for_each_cfg_sme(cfg, fwspec, j, idx) {
> > +                       smr = &smmu->smrs[idx];
> > +                       if (smr_is_subset(*smr, id, mask))
> > +                               arm_smmu_cb_write(smmu, cbndx, ARM_SMMU_CB_ACTLR,
> > +                                               (actlrcfg + i)->actlr);
> > +               }
> > +       }
> > +}
> > +
> >   static int qcom_adreno_smmu_init_context(struct arm_smmu_domain *smmu_domain,
> >                  struct io_pgtable_cfg *pgtbl_cfg, struct device *dev)
> >   {
> >          struct adreno_smmu_priv *priv;
> > +       struct arm_smmu_device *smmu = smmu_domain->smmu;
> > +       struct qcom_smmu *qsmmu = to_qcom_smmu(smmu);
> > +       const struct actlr_config *actlrcfg;
> > +       size_t actlrcfg_size;
> > +       int cbndx = smmu_domain->cfg.cbndx;
> >
> >          smmu_domain->cfg.flush_walk_prefer_tlbiasid = true;
> >
> > @@ -238,6 +274,12 @@ static int qcom_adreno_smmu_init_context(struct arm_smmu_domain *smmu_domain,
> >          priv->set_stall = qcom_adreno_smmu_set_stall;
> >          priv->resume_translation = qcom_adreno_smmu_resume_translation;
> >
> > +       if (qsmmu->data->actlrcfg_gfx) {
> > +               actlrcfg = qsmmu->data->actlrcfg_gfx;
> > +               actlrcfg_size = qsmmu->data->actlrcfg_gfx_size;
> > +               arm_smmu_set_actlr(dev, smmu, cbndx, actlrcfg, actlrcfg_size);
> > +       }
> > +
> >          return 0;
> >   }
> >
> > @@ -263,6 +305,18 @@ static const struct of_device_id qcom_smmu_client_of_match[] __maybe_unused = {
> >   static int qcom_smmu_init_context(struct arm_smmu_domain *smmu_domain,
> >                  struct io_pgtable_cfg *pgtbl_cfg, struct device *dev)
> >   {
> > +       struct arm_smmu_device *smmu = smmu_domain->smmu;
> > +       struct qcom_smmu *qsmmu = to_qcom_smmu(smmu);
> > +       const struct actlr_config *actlrcfg;
> > +       size_t actlrcfg_size;
> > +       int cbndx = smmu_domain->cfg.cbndx;
> > +
> > +       if (qsmmu->data->actlrcfg) {
> > +               actlrcfg = qsmmu->data->actlrcfg;
> > +               actlrcfg_size = qsmmu->data->actlrcfg_size;
> > +               arm_smmu_set_actlr(dev, smmu, cbndx, actlrcfg, actlrcfg_size);
> > +       }
> > +
> >          smmu_domain->cfg.flush_walk_prefer_tlbiasid = true;
> >
> >          return 0;
> > @@ -464,7 +518,7 @@ static struct arm_smmu_device *qcom_smmu_create(struct arm_smmu_device *smmu,
> >                  return ERR_PTR(-ENOMEM);
> >
> >          qsmmu->smmu.impl = impl;
> > -       qsmmu->cfg = data->cfg;
> > +       qsmmu->data = data;
>
> This should go to a separate commit. It is not related to ACTLR support

qsmmu->data has the actlrcfg/actlrcfg_gfx as well hence clubbed this 
change here as rightly suggested by Robin[1] on v2 revision.
Initially planned[2] for separate patch, but later clubbing it with 
actlr patch looked like a cleaner approach.
Would it be okay to keep it here? Or separate patch would be better?

[1]:https://lore.kernel.org/all/c75d107a-44cb-4df3-b583-13719df1f8be@xxxxxxx/
[2]:https://lore.kernel.org/all/9b406a7c-57b8-4b5f-8fbc-714560cce8cf@xxxxxxxxxxx/
> >          return &qsmmu->smmu;
> >   }
> > diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.h b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.h
> > index 593910567b88..138fc57f7b0d 100644
> > --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.h
> > +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.h
> > @@ -8,7 +8,7 @@
> >
> >   struct qcom_smmu {
> >          struct arm_smmu_device smmu;
> > -       const struct qcom_smmu_config *cfg;
> > +       const struct qcom_smmu_match_data *data;
> >          bool bypass_quirk;
> >          u8 bypass_cbndx;
> >          u32 stall_enabled;
> > @@ -25,6 +25,10 @@ struct qcom_smmu_config {
> >   };
> >
> >   struct qcom_smmu_match_data {
> > +       const struct actlr_config *actlrcfg;
> > +       size_t actlrcfg_size;
> > +       const struct actlr_config *actlrcfg_gfx;
> > +       size_t actlrcfg_gfx_size;
> >          const struct qcom_smmu_config *cfg;
> >          const struct arm_smmu_impl *impl;
> >          const struct arm_smmu_impl *adreno_impl;
> > diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu.c b/drivers/iommu/arm/arm-smmu/arm-smmu.c
> > index d6d1a2a55cc0..8e4faf015286 100644
> > --- a/drivers/iommu/arm/arm-smmu/arm-smmu.c
> > +++ b/drivers/iommu/arm/arm-smmu/arm-smmu.c
> > @@ -990,9 +990,10 @@ static int arm_smmu_find_sme(struct arm_smmu_device *smmu, u16 id, u16 mask)
> >                   * expect simply identical entries for this case, but there's
> >                   * no harm in accommodating the generalisation.
> >                   */
> > -               if ((mask & smrs[i].mask) == mask &&
> > -                   !((id ^ smrs[i].id) & ~smrs[i].mask))
> > +
> > +               if (smr_is_subset(smrs[i], id, mask))
> >                          return i;
> > +
> >                  /*
> >                   * If the new entry has any other overlap with an existing one,
> >                   * though, then there always exists at least one stream ID
> > diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu.h b/drivers/iommu/arm/arm-smmu/arm-smmu.h
> > index 703fd5817ec1..b1638bbc41d4 100644
> > --- a/drivers/iommu/arm/arm-smmu/arm-smmu.h
> > +++ b/drivers/iommu/arm/arm-smmu/arm-smmu.h
> > @@ -501,6 +501,11 @@ static inline void arm_smmu_writeq(struct arm_smmu_device *smmu, int page,
> >                  writeq_relaxed(val, arm_smmu_page(smmu, page) + offset);
> >   }
> >
> > +static inline bool smr_is_subset(struct arm_smmu_smr smrs, u16 id, u16 mask)
>
> A pointer to the struct, please
>
> > +{
> > +       return (mask & smrs.mask) == mask && !((id ^ smrs.id) & ~smrs.mask);
> > +}
> > +
> >   #define ARM_SMMU_GR0           0
> >   #define ARM_SMMU_GR1           1
> >   #define ARM_SMMU_CB(s, n)      ((s)->numpage + (n))
> > --
> > 2.17.1