unsigned cond_exe_offs;
u64 cond_exe_gpu_addr;
volatile u32 *cond_exe_cpu_addr;
- int vmid;
};
/*
unsigned vm_id, uint64_t pd_addr,
uint32_t gds_base, uint32_t gds_size,
uint32_t gws_base, uint32_t gws_size,
- uint32_t oa_base, uint32_t oa_size,
- bool vmid_switch);
+ uint32_t oa_base, uint32_t oa_size);
void amdgpu_vm_reset_id(struct amdgpu_device *adev, unsigned vm_id);
uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr);
int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
/* MM block clocks */
int (*set_uvd_clocks)(struct amdgpu_device *adev, u32 vclk, u32 dclk);
int (*set_vce_clocks)(struct amdgpu_device *adev, u32 evclk, u32 ecclk);
+ /* query virtual capabilities */
+ u32 (*get_virtual_caps)(struct amdgpu_device *adev);
};
/*
/* GPU virtualization */
+#define AMDGPU_VIRT_CAPS_SRIOV_EN (1 << 0)
+#define AMDGPU_VIRT_CAPS_IS_VF (1 << 1)
struct amdgpu_virtualization {
bool supports_sr_iov;
+ bool is_virtual;
+ u32 caps;
};
/*
#define amdgpu_asic_get_xclk(adev) (adev)->asic_funcs->get_xclk((adev))
#define amdgpu_asic_set_uvd_clocks(adev, v, d) (adev)->asic_funcs->set_uvd_clocks((adev), (v), (d))
#define amdgpu_asic_set_vce_clocks(adev, ev, ec) (adev)->asic_funcs->set_vce_clocks((adev), (ev), (ec))
+#define amdgpu_asic_get_virtual_caps(adev) ((adev)->asic_funcs->get_virtual_caps((adev)))
#define amdgpu_asic_get_gpu_clock_counter(adev) (adev)->asic_funcs->get_gpu_clock_counter((adev))
#define amdgpu_asic_read_disabled_bios(adev) (adev)->asic_funcs->read_disabled_bios((adev))
#define amdgpu_asic_read_bios_from_rom(adev, b, l) (adev)->asic_funcs->read_bios_from_rom((adev), (b), (l))
return 0;
}
+static bool amdgpu_device_is_virtual(void)
+{
+#ifdef CONFIG_X86
+ return boot_cpu_has(X86_FEATURE_HYPERVISOR);
+#else
+ return false;
+#endif
+}
+
/**
* amdgpu_device_init - initialize the driver
*
adev->virtualization.supports_sr_iov =
amdgpu_atombios_has_gpu_virtualization_table(adev);
+ /* Check if we are executing in a virtualized environment */
+ adev->virtualization.is_virtual = amdgpu_device_is_virtual();
+ adev->virtualization.caps = amdgpu_asic_get_virtual_caps(adev);
+
/* Post card if necessary */
- if (!amdgpu_card_posted(adev)) {
+ if (!amdgpu_card_posted(adev) ||
+ (adev->virtualization.is_virtual &&
+ !adev->virtualization.caps & AMDGPU_VIRT_CAPS_SRIOV_EN)) {
if (!adev->bios) {
dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n");
return -EINVAL;
bool skip_preamble, need_ctx_switch;
unsigned patch_offset = ~0;
struct amdgpu_vm *vm;
- int vmid = 0, old_vmid = ring->vmid;
struct fence *hwf;
uint64_t ctx;
if (job) {
vm = job->vm;
ctx = job->ctx;
- vmid = job->vm_id;
} else {
vm = NULL;
ctx = 0;
- vmid = 0;
}
if (!ring->ready) {
r = amdgpu_vm_flush(ring, job->vm_id, job->vm_pd_addr,
job->gds_base, job->gds_size,
job->gws_base, job->gws_size,
- job->oa_base, job->oa_size,
- (ring->current_ctx == ctx) && (old_vmid != vmid));
+ job->oa_base, job->oa_size);
if (r) {
amdgpu_ring_undo(ring);
return r;
need_ctx_switch = ring->current_ctx != ctx;
for (i = 0; i < num_ibs; ++i) {
ib = &ibs[i];
+
/* drop preamble IBs if we don't have a context switch */
if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) && skip_preamble)
continue;
amdgpu_ring_emit_ib(ring, ib, job ? job->vm_id : 0,
need_ctx_switch);
need_ctx_switch = false;
- ring->vmid = vmid;
}
if (ring->funcs->emit_hdp_invalidate)
dev_err(adev->dev, "failed to emit fence (%d)\n", r);
if (job && job->vm_id)
amdgpu_vm_reset_id(adev, job->vm_id);
- ring->vmid = old_vmid;
amdgpu_ring_undo(ring);
return r;
}
unsigned vm_id, uint64_t pd_addr,
uint32_t gds_base, uint32_t gds_size,
uint32_t gws_base, uint32_t gws_size,
- uint32_t oa_base, uint32_t oa_size,
- bool vmid_switch)
+ uint32_t oa_base, uint32_t oa_size)
{
struct amdgpu_device *adev = ring->adev;
struct amdgpu_vm_id *id = &adev->vm_manager.ids[vm_id];
int r;
if (ring->funcs->emit_pipeline_sync && (
- pd_addr != AMDGPU_VM_NO_FLUSH || gds_switch_needed || vmid_switch))
+ pd_addr != AMDGPU_VM_NO_FLUSH || gds_switch_needed ||
+ ring->type == AMDGPU_RING_TYPE_COMPUTE))
amdgpu_ring_emit_pipeline_sync(ring);
if (ring->funcs->emit_vm_flush &&
return true;
}
+static u32 cik_get_virtual_caps(struct amdgpu_device *adev)
+{
+ /* CIK does not support SR-IOV */
+ return 0;
+}
+
static const struct amdgpu_allowed_register_entry cik_allowed_read_registers[] = {
{mmGRBM_STATUS, false},
{mmGB_ADDR_CONFIG, false},
.get_xclk = &cik_get_xclk,
.set_uvd_clocks = &cik_set_uvd_clocks,
.set_vce_clocks = &cik_set_vce_clocks,
+ .get_virtual_caps = &cik_get_virtual_caps,
/* these should be moved to their own ip modules */
.get_gpu_clock_counter = &gfx_v7_0_get_gpu_clock_counter,
.wait_for_mc_idle = &gmc_v7_0_mc_wait_for_idle,
case 2:
for (i = 0; i < adev->gfx.num_compute_rings; i++) {
ring = &adev->gfx.compute_ring[i];
- if ((ring->me == me_id) & (ring->pipe == pipe_id))
+ if ((ring->me == me_id) && (ring->pipe == pipe_id))
amdgpu_fence_process(ring);
}
break;
return true;
}
+static u32 vi_get_virtual_caps(struct amdgpu_device *adev)
+{
+ u32 caps = 0;
+ u32 reg = RREG32(mmBIF_IOV_FUNC_IDENTIFIER);
+
+ if (REG_GET_FIELD(reg, BIF_IOV_FUNC_IDENTIFIER, IOV_ENABLE))
+ caps |= AMDGPU_VIRT_CAPS_SRIOV_EN;
+
+ if (REG_GET_FIELD(reg, BIF_IOV_FUNC_IDENTIFIER, FUNC_IDENTIFIER))
+ caps |= AMDGPU_VIRT_CAPS_IS_VF;
+
+ return caps;
+}
+
static const struct amdgpu_allowed_register_entry tonga_allowed_read_registers[] = {
{mmGB_MACROTILE_MODE7, true},
};
.get_xclk = &vi_get_xclk,
.set_uvd_clocks = &vi_set_uvd_clocks,
.set_vce_clocks = &vi_set_vce_clocks,
+ .get_virtual_caps = &vi_get_virtual_caps,
/* these should be moved to their own ip modules */
.get_gpu_clock_counter = &gfx_v8_0_get_gpu_clock_counter,
.wait_for_mc_idle = &gmc_v8_0_mc_wait_for_idle,
uint8_t phases;
uint8_t cks_enable;
uint8_t cks_voffset;
+ uint32_t sclk_offset;
};
typedef struct phm_ppt_v1_clock_voltage_dependency_record phm_ppt_v1_clock_voltage_dependency_record;
vddci = phm_find_closest_vddci(&(data->vddci_voltage_table),
(dep_table->entries[i].vddc -
(uint16_t)data->vddc_vddci_delta));
- *voltage |= (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT;
+ *voltage |= (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT;
}
if (POLARIS10_VOLTAGE_CONTROL_NONE == data->mvdd_control)
ATOM_Tonga_State *state_entry = (ATOM_Tonga_State *)state;
ATOM_Tonga_POWERPLAYTABLE *powerplay_table =
(ATOM_Tonga_POWERPLAYTABLE *)pp_table;
- ATOM_Tonga_SCLK_Dependency_Table *sclk_dep_table =
- (ATOM_Tonga_SCLK_Dependency_Table *)
+ PPTable_Generic_SubTable_Header *sclk_dep_table =
+ (PPTable_Generic_SubTable_Header *)
(((unsigned long)powerplay_table) +
le16_to_cpu(powerplay_table->usSclkDependencyTableOffset));
+
ATOM_Tonga_MCLK_Dependency_Table *mclk_dep_table =
(ATOM_Tonga_MCLK_Dependency_Table *)
(((unsigned long)powerplay_table) +
/* Performance levels are arranged from low to high. */
performance_level->memory_clock = mclk_dep_table->entries
[state_entry->ucMemoryClockIndexLow].ulMclk;
- performance_level->engine_clock = sclk_dep_table->entries
+ if (sclk_dep_table->ucRevId == 0)
+ performance_level->engine_clock = ((ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table)->entries
+ [state_entry->ucEngineClockIndexLow].ulSclk;
+ else if (sclk_dep_table->ucRevId == 1)
+ performance_level->engine_clock = ((ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table)->entries
[state_entry->ucEngineClockIndexLow].ulSclk;
performance_level->pcie_gen = get_pcie_gen_support(data->pcie_gen_cap,
state_entry->ucPCIEGenLow);
[polaris10_power_state->performance_level_count++]);
performance_level->memory_clock = mclk_dep_table->entries
[state_entry->ucMemoryClockIndexHigh].ulMclk;
- performance_level->engine_clock = sclk_dep_table->entries
+
+ if (sclk_dep_table->ucRevId == 0)
+ performance_level->engine_clock = ((ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table)->entries
[state_entry->ucEngineClockIndexHigh].ulSclk;
+ else if (sclk_dep_table->ucRevId == 1)
+ performance_level->engine_clock = ((ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table)->entries
+ [state_entry->ucEngineClockIndexHigh].ulSclk;
+
performance_level->pcie_gen = get_pcie_gen_support(data->pcie_gen_cap,
state_entry->ucPCIEGenHigh);
performance_level->pcie_lane = get_pcie_lane_support(data->pcie_lane_cap,
switch (state->classification.ui_label) {
case PP_StateUILabel_Performance:
data->use_pcie_performance_levels = true;
-
for (i = 0; i < ps->performance_level_count; i++) {
if (data->pcie_gen_performance.max <
ps->performance_levels[i].pcie_gen)
ps->performance_levels[i].pcie_lane)
data->pcie_lane_performance.max =
ps->performance_levels[i].pcie_lane;
-
if (data->pcie_lane_performance.min >
ps->performance_levels[i].pcie_lane)
data->pcie_lane_performance.min =
{
struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend);
uint32_t mm_boot_level_offset, mm_boot_level_value;
- struct phm_ppt_v1_information *table_info =
- (struct phm_ppt_v1_information *)(hwmgr->pptable);
if (!bgate) {
- data->smc_state_table.SamuBootLevel =
- (uint8_t) (table_info->mm_dep_table->count - 1);
+ data->smc_state_table.SamuBootLevel = 0;
mm_boot_level_offset = data->dpm_table_start +
offsetof(SMU74_Discrete_DpmTable, SamuBootLevel);
mm_boot_level_offset /= 4;
ATOM_Tonga_SCLK_Dependency_Record entries[1]; /* Dynamically allocate entries. */
} ATOM_Tonga_SCLK_Dependency_Table;
+typedef struct _ATOM_Polaris_SCLK_Dependency_Record {
+ UCHAR ucVddInd; /* Base voltage */
+ USHORT usVddcOffset; /* Offset relative to base voltage */
+ ULONG ulSclk;
+ USHORT usEdcCurrent;
+ UCHAR ucReliabilityTemperature;
+ UCHAR ucCKSVOffsetandDisable; /* Bits 0~6: Voltage offset for CKS, Bit 7: Disable/enable for the SCLK level. */
+ ULONG ulSclkOffset;
+} ATOM_Polaris_SCLK_Dependency_Record;
+
+typedef struct _ATOM_Polaris_SCLK_Dependency_Table {
+ UCHAR ucRevId;
+ UCHAR ucNumEntries; /* Number of entries. */
+ ATOM_Polaris_SCLK_Dependency_Record entries[1]; /* Dynamically allocate entries. */
+} ATOM_Polaris_SCLK_Dependency_Table;
+
typedef struct _ATOM_Tonga_PCIE_Record {
UCHAR ucPCIEGenSpeed;
UCHAR usPCIELaneWidth;
static int get_sclk_voltage_dependency_table(
struct pp_hwmgr *hwmgr,
phm_ppt_v1_clock_voltage_dependency_table **pp_tonga_sclk_dep_table,
- const ATOM_Tonga_SCLK_Dependency_Table * sclk_dep_table
+ const PPTable_Generic_SubTable_Header *sclk_dep_table
)
{
uint32_t table_size, i;
phm_ppt_v1_clock_voltage_dependency_table *sclk_table;
- PP_ASSERT_WITH_CODE((0 != sclk_dep_table->ucNumEntries),
- "Invalid PowerPlay Table!", return -1);
+ if (sclk_dep_table->ucRevId < 1) {
+ const ATOM_Tonga_SCLK_Dependency_Table *tonga_table =
+ (ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table;
- table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record)
- * sclk_dep_table->ucNumEntries;
+ PP_ASSERT_WITH_CODE((0 != tonga_table->ucNumEntries),
+ "Invalid PowerPlay Table!", return -1);
- sclk_table = (phm_ppt_v1_clock_voltage_dependency_table *)
- kzalloc(table_size, GFP_KERNEL);
+ table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record)
+ * tonga_table->ucNumEntries;
- if (NULL == sclk_table)
- return -ENOMEM;
+ sclk_table = (phm_ppt_v1_clock_voltage_dependency_table *)
+ kzalloc(table_size, GFP_KERNEL);
- memset(sclk_table, 0x00, table_size);
-
- sclk_table->count = (uint32_t)sclk_dep_table->ucNumEntries;
-
- for (i = 0; i < sclk_dep_table->ucNumEntries; i++) {
- sclk_table->entries[i].vddInd =
- sclk_dep_table->entries[i].ucVddInd;
- sclk_table->entries[i].vdd_offset =
- sclk_dep_table->entries[i].usVddcOffset;
- sclk_table->entries[i].clk =
- sclk_dep_table->entries[i].ulSclk;
- sclk_table->entries[i].cks_enable =
- (((sclk_dep_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0;
- sclk_table->entries[i].cks_voffset =
- (sclk_dep_table->entries[i].ucCKSVOffsetandDisable & 0x7F);
- }
+ if (NULL == sclk_table)
+ return -ENOMEM;
+
+ memset(sclk_table, 0x00, table_size);
+
+ sclk_table->count = (uint32_t)tonga_table->ucNumEntries;
+
+ for (i = 0; i < tonga_table->ucNumEntries; i++) {
+ sclk_table->entries[i].vddInd =
+ tonga_table->entries[i].ucVddInd;
+ sclk_table->entries[i].vdd_offset =
+ tonga_table->entries[i].usVddcOffset;
+ sclk_table->entries[i].clk =
+ tonga_table->entries[i].ulSclk;
+ sclk_table->entries[i].cks_enable =
+ (((tonga_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0;
+ sclk_table->entries[i].cks_voffset =
+ (tonga_table->entries[i].ucCKSVOffsetandDisable & 0x7F);
+ }
+ } else {
+ const ATOM_Polaris_SCLK_Dependency_Table *polaris_table =
+ (ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table;
+ PP_ASSERT_WITH_CODE((0 != polaris_table->ucNumEntries),
+ "Invalid PowerPlay Table!", return -1);
+
+ table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record)
+ * polaris_table->ucNumEntries;
+
+ sclk_table = (phm_ppt_v1_clock_voltage_dependency_table *)
+ kzalloc(table_size, GFP_KERNEL);
+
+ if (NULL == sclk_table)
+ return -ENOMEM;
+
+ memset(sclk_table, 0x00, table_size);
+
+ sclk_table->count = (uint32_t)polaris_table->ucNumEntries;
+
+ for (i = 0; i < polaris_table->ucNumEntries; i++) {
+ sclk_table->entries[i].vddInd =
+ polaris_table->entries[i].ucVddInd;
+ sclk_table->entries[i].vdd_offset =
+ polaris_table->entries[i].usVddcOffset;
+ sclk_table->entries[i].clk =
+ polaris_table->entries[i].ulSclk;
+ sclk_table->entries[i].cks_enable =
+ (((polaris_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0;
+ sclk_table->entries[i].cks_voffset =
+ (polaris_table->entries[i].ucCKSVOffsetandDisable & 0x7F);
+ sclk_table->entries[i].sclk_offset = polaris_table->entries[i].ulSclkOffset;
+ }
+ }
*pp_tonga_sclk_dep_table = sclk_table;
return 0;
const ATOM_Tonga_MCLK_Dependency_Table *mclk_dep_table =
(const ATOM_Tonga_MCLK_Dependency_Table *)(((unsigned long) powerplay_table) +
le16_to_cpu(powerplay_table->usMclkDependencyTableOffset));
- const ATOM_Tonga_SCLK_Dependency_Table *sclk_dep_table =
- (const ATOM_Tonga_SCLK_Dependency_Table *)(((unsigned long) powerplay_table) +
+ const PPTable_Generic_SubTable_Header *sclk_dep_table =
+ (const PPTable_Generic_SubTable_Header *)(((unsigned long) powerplay_table) +
le16_to_cpu(powerplay_table->usSclkDependencyTableOffset));
const ATOM_Tonga_Hard_Limit_Table *pHardLimits =
(const ATOM_Tonga_Hard_Limit_Table *)(((unsigned long) powerplay_table) +
if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev) || ASIC_IS_DCE8(rdev))
radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
/* use frac fb div on RS780/RS880 */
- if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
+ if (((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
+ && !radeon_crtc->ss_enabled)
radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
if (ASIC_IS_DCE32(rdev) && mode->clock > 165000)
radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
if (radeon_crtc->ss.refdiv) {
radeon_crtc->pll_flags |= RADEON_PLL_USE_REF_DIV;
radeon_crtc->pll_reference_div = radeon_crtc->ss.refdiv;
- if (ASIC_IS_AVIVO(rdev))
+ if (rdev->family >= CHIP_RV770)
radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
}
}
/*
* GPU helpers function.
*/
+
+/**
+ * radeon_device_is_virtual - check if we are running is a virtual environment
+ *
+ * Check if the asic has been passed through to a VM (all asics).
+ * Used at driver startup.
+ * Returns true if virtual or false if not.
+ */
+static bool radeon_device_is_virtual(void)
+{
+#ifdef CONFIG_X86
+ return boot_cpu_has(X86_FEATURE_HYPERVISOR);
+#else
+ return false;
+#endif
+}
+
/**
* radeon_card_posted - check if the hw has already been initialized
*
{
uint32_t reg;
+ /* for pass through, always force asic_init */
+ if (radeon_device_is_virtual())
+ return false;
+
/* required for EFI mode on macbook2,1 which uses an r5xx asic */
if (efi_enabled(EFI_BOOT) &&
(rdev->pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE) &&
radeon_agp_suspend(rdev);
pci_save_state(dev->pdev);
- if (freeze && rdev->family >= CHIP_R600) {
+ if (freeze && rdev->family >= CHIP_CEDAR) {
rdev->asic->asic_reset(rdev, true);
pci_restore_state(dev->pdev);
} else if (suspend) {
projects / cascardo / linux.git / commitdiff