diff options
| -rw-r--r-- | arch/powerpc/platforms/powernv/npu-dma.c | 64 | ||||
| -rw-r--r-- | arch/powerpc/platforms/powernv/pci-ioda.c | 112 | ||||
| -rw-r--r-- | arch/powerpc/platforms/powernv/pci.h | 6 |
3 files changed, 176 insertions, 6 deletions
diff --git a/arch/powerpc/platforms/powernv/npu-dma.c b/arch/powerpc/platforms/powernv/npu-dma.c index cb2d1da..0459e10 100644 --- a/arch/powerpc/platforms/powernv/npu-dma.c +++ b/arch/powerpc/platforms/powernv/npu-dma.c @@ -12,6 +12,7 @@ #include <linux/export.h> #include <linux/pci.h> #include <linux/memblock.h> +#include <linux/iommu.h> #include <asm/iommu.h> #include <asm/pnv-pci.h> @@ -154,7 +155,7 @@ static struct pnv_ioda_pe *get_gpu_pci_dev_and_pe(struct pnv_ioda_pe *npe, return pe; } -static long pnv_npu_set_window(struct pnv_ioda_pe *npe, +long pnv_npu_set_window(struct pnv_ioda_pe *npe, int num, struct iommu_table *tbl) { struct pnv_phb *phb = npe->phb; @@ -182,13 +183,13 @@ static long pnv_npu_set_window(struct pnv_ioda_pe *npe, pnv_pci_ioda2_tce_invalidate_entire(phb, false); /* Add the table to the list so its TCE cache will get invalidated */ - pnv_pci_link_table_and_group(phb->hose->node, 0, + pnv_pci_link_table_and_group(phb->hose->node, num, tbl, &npe->table_group); return 0; } -static long pnv_npu_unset_window(struct pnv_ioda_pe *npe) +long pnv_npu_unset_window(struct pnv_ioda_pe *npe, int num) { struct pnv_phb *phb = npe->phb; int64_t rc; @@ -205,7 +206,7 @@ static long pnv_npu_unset_window(struct pnv_ioda_pe *npe) } pnv_pci_ioda2_tce_invalidate_entire(phb, false); - pnv_pci_unlink_table_and_group(npe->table_group.tables[0], + pnv_pci_unlink_table_and_group(npe->table_group.tables[num], &npe->table_group); return 0; @@ -231,7 +232,7 @@ static void pnv_npu_dma_set_32(struct pnv_ioda_pe *npe) if (!gpe) return; - rc = pnv_npu_set_window(npe, gpe->table_group.tables[0]); + rc = pnv_npu_set_window(npe, 0, gpe->table_group.tables[0]); /* * We don't initialise npu_pe->tce32_table as we always use @@ -255,7 +256,7 @@ static int pnv_npu_dma_set_bypass(struct pnv_ioda_pe *npe) if (phb->type != PNV_PHB_NPU || !npe->pdev) return -EINVAL; - rc = pnv_npu_unset_window(npe); + rc = pnv_npu_unset_window(npe, 0); if (rc != OPAL_SUCCESS) return rc; @@ -307,3 +308,54 @@ void pnv_npu_try_dma_set_bypass(struct pci_dev *gpdev, bool bypass) } } } + +/* Switch ownership from platform code to external user (e.g. VFIO) */ +void pnv_npu_take_ownership(struct pnv_ioda_pe *npe) +{ + struct pnv_phb *phb = npe->phb; + int64_t rc; + + /* + * Note: NPU has just a single TVE in the hardware which means that + * while used by the kernel, it can have either 32bit window or + * DMA bypass but never both. So we deconfigure 32bit window only + * if it was enabled at the moment of ownership change. + */ + if (npe->table_group.tables[0]) { + pnv_npu_unset_window(npe, 0); + return; + } + + /* Disable bypass */ + rc = opal_pci_map_pe_dma_window_real(phb->opal_id, + npe->pe_number, npe->pe_number, + 0 /* bypass base */, 0); + if (rc) { + pe_err(npe, "Failed to disable bypass, err %lld\n", rc); + return; + } + pnv_pci_ioda2_tce_invalidate_entire(npe->phb, false); +} + +struct pnv_ioda_pe *pnv_pci_npu_setup_iommu(struct pnv_ioda_pe *npe) +{ + struct pnv_phb *phb = npe->phb; + struct pci_bus *pbus = phb->hose->bus; + struct pci_dev *npdev, *gpdev = NULL, *gptmp; + struct pnv_ioda_pe *gpe = get_gpu_pci_dev_and_pe(npe, &gpdev); + + if (!gpe || !gpdev) + return NULL; + + list_for_each_entry(npdev, &pbus->devices, bus_list) { + gptmp = pnv_pci_get_gpu_dev(npdev); + + if (gptmp != gpdev) + continue; + + pe_info(gpe, "Attached NPU %s\n", dev_name(&npdev->dev)); + iommu_group_add_device(gpe->table_group.group, &npdev->dev); + } + + return gpe; +} diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c index ef31218..4cfe704 100644 --- a/arch/powerpc/platforms/powernv/pci-ioda.c +++ b/arch/powerpc/platforms/powernv/pci-ioda.c @@ -2389,6 +2389,116 @@ static struct iommu_table_group_ops pnv_pci_ioda2_ops = { .take_ownership = pnv_ioda2_take_ownership, .release_ownership = pnv_ioda2_release_ownership, }; + +static int gpe_table_group_to_npe_cb(struct device *dev, void *opaque) +{ + struct pci_controller *hose; + struct pnv_phb *phb; + struct pnv_ioda_pe **ptmppe = opaque; + struct pci_dev *pdev = container_of(dev, struct pci_dev, dev); + struct pci_dn *pdn = pci_get_pdn(pdev); + + if (!pdn || pdn->pe_number == IODA_INVALID_PE) + return 0; + + hose = pci_bus_to_host(pdev->bus); + phb = hose->private_data; + if (phb->type != PNV_PHB_NPU) + return 0; + + *ptmppe = &phb->ioda.pe_array[pdn->pe_number]; + + return 1; +} + +/* + * This returns PE of associated NPU. + * This assumes that NPU is in the same IOMMU group with GPU and there is + * no other PEs. + */ +static struct pnv_ioda_pe *gpe_table_group_to_npe( + struct iommu_table_group *table_group) +{ + struct pnv_ioda_pe *npe = NULL; + int ret = iommu_group_for_each_dev(table_group->group, &npe, + gpe_table_group_to_npe_cb); + + BUG_ON(!ret || !npe); + + return npe; +} + +static long pnv_pci_ioda2_npu_set_window(struct iommu_table_group *table_group, + int num, struct iommu_table *tbl) +{ + long ret = pnv_pci_ioda2_set_window(table_group, num, tbl); + + if (ret) + return ret; + + ret = pnv_npu_set_window(gpe_table_group_to_npe(table_group), num, tbl); + if (ret) + pnv_pci_ioda2_unset_window(table_group, num); + + return ret; +} + +static long pnv_pci_ioda2_npu_unset_window( + struct iommu_table_group *table_group, + int num) +{ + long ret = pnv_pci_ioda2_unset_window(table_group, num); + + if (ret) + return ret; + + return pnv_npu_unset_window(gpe_table_group_to_npe(table_group), num); +} + +static void pnv_ioda2_npu_take_ownership(struct iommu_table_group *table_group) +{ + /* + * Detach NPU first as pnv_ioda2_take_ownership() will destroy + * the iommu_table if 32bit DMA is enabled. + */ + pnv_npu_take_ownership(gpe_table_group_to_npe(table_group)); + pnv_ioda2_take_ownership(table_group); +} + +static struct iommu_table_group_ops pnv_pci_ioda2_npu_ops = { + .get_table_size = pnv_pci_ioda2_get_table_size, + .create_table = pnv_pci_ioda2_create_table, + .set_window = pnv_pci_ioda2_npu_set_window, + .unset_window = pnv_pci_ioda2_npu_unset_window, + .take_ownership = pnv_ioda2_npu_take_ownership, + .release_ownership = pnv_ioda2_release_ownership, +}; + +static void pnv_pci_ioda_setup_iommu_api(void) +{ + struct pci_controller *hose, *tmp; + struct pnv_phb *phb; + struct pnv_ioda_pe *pe, *gpe; + + /* + * Now we have all PHBs discovered, time to add NPU devices to + * the corresponding IOMMU groups. + */ + list_for_each_entry_safe(hose, tmp, &hose_list, list_node) { + phb = hose->private_data; + + if (phb->type != PNV_PHB_NPU) + continue; + + list_for_each_entry(pe, &phb->ioda.pe_list, list) { + gpe = pnv_pci_npu_setup_iommu(pe); + if (gpe) + gpe->table_group.ops = &pnv_pci_ioda2_npu_ops; + } + } +} +#else /* !CONFIG_IOMMU_API */ +static void pnv_pci_ioda_setup_iommu_api(void) { }; #endif static void pnv_pci_ioda_setup_opal_tce_kill(struct pnv_phb *phb) @@ -3115,6 +3225,8 @@ static void pnv_pci_ioda_setup_DMA(void) phb = hose->private_data; phb->initialized = 1; } + + pnv_pci_ioda_setup_iommu_api(); } static void pnv_pci_ioda_create_dbgfs(void) diff --git a/arch/powerpc/platforms/powernv/pci.h b/arch/powerpc/platforms/powernv/pci.h index 2b35606..7dee25e 100644 --- a/arch/powerpc/platforms/powernv/pci.h +++ b/arch/powerpc/platforms/powernv/pci.h @@ -225,5 +225,11 @@ extern void pe_level_printk(const struct pnv_ioda_pe *pe, const char *level, /* Nvlink functions */ extern void pnv_npu_try_dma_set_bypass(struct pci_dev *gpdev, bool bypass); extern void pnv_pci_ioda2_tce_invalidate_entire(struct pnv_phb *phb, bool rm); +extern struct pnv_ioda_pe *pnv_pci_npu_setup_iommu(struct pnv_ioda_pe *npe); +extern long pnv_npu_set_window(struct pnv_ioda_pe *npe, int num, + struct iommu_table *tbl); +extern long pnv_npu_unset_window(struct pnv_ioda_pe *npe, int num); +extern void pnv_npu_take_ownership(struct pnv_ioda_pe *npe); +extern void pnv_npu_release_ownership(struct pnv_ioda_pe *npe); #endif /* __POWERNV_PCI_H */ |
