[PATCH v4 2/2] drm/amdkfd: force raven as "dgpu" path (v4)
Felix Kuehling
felix.kuehling at amd.com
Fri Aug 21 04:20:38 UTC 2020
Am 2020-08-20 um 11:53 p.m. schrieb Huang Rui:
> On Fri, Aug 21, 2020 at 10:41:17AM +0800, Kuehling, Felix wrote:
>> Am 2020-08-20 um 4:40 a.m. schrieb Huang Rui:
>>> We still have a few iommu issues which need to address, so force raven
>>> as "dgpu" path for the moment.
>>>
>>> This is to add the fallback path to bypass IOMMU if IOMMU v2 is disabled
>>> or ACPI CRAT table not correct.
>>>
>>> v2: Use ignore_crat parameter to decide whether it will go with IOMMUv2.
>>> v3: Align with existed thunk, don't change the way of raven, only renoir
>>> will use "dgpu" path by default.
>>> v4: don't update global ignore_crat in the driver, and revise fallback
>>> function if CRAT is broken.
>>>
>>> Signed-off-by: Huang Rui <ray.huang at amd.com>
>>> ---
>>> drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 5 ++++-
>>> drivers/gpu/drm/amd/amdkfd/kfd_crat.c | 23 +++++++++++++++++++++--
>>> drivers/gpu/drm/amd/amdkfd/kfd_device.c | 5 ++++-
>>> drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 10 ++++++++--
>>> drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 20 ++++++++++++++++++++
>>> 5 files changed, 57 insertions(+), 6 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>>> index a9a4319c24ae..189f9d7e190d 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>>> @@ -684,11 +684,14 @@ MODULE_PARM_DESC(debug_largebar,
>>> * Ignore CRAT table during KFD initialization. By default, KFD uses the ACPI CRAT
>>> * table to get information about AMD APUs. This option can serve as a workaround on
>>> * systems with a broken CRAT table.
>>> + *
>>> + * Default is auto (according to asic type, iommu_v2, and crat table, to decide
>>> + * whehter use CRAT)
>>> */
>>> int ignore_crat;
>>> module_param(ignore_crat, int, 0444);
>>> MODULE_PARM_DESC(ignore_crat,
>>> - "Ignore CRAT table during KFD initialization (0 = use CRAT (default), 1 = ignore CRAT)");
>>> + "Ignore CRAT table during KFD initialization (0 = auto (default), 1 = ignore CRAT)");
>>>
>>> /**
>>> * DOC: halt_if_hws_hang (int)
>>> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
>>> index 59557e3e206a..a17cfc290072 100644
>>> --- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
>>> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
>>> @@ -22,6 +22,7 @@
>>>
>>> #include <linux/pci.h>
>>> #include <linux/acpi.h>
>>> +#include <asm/processor.h>
>>> #include "kfd_crat.h"
>>> #include "kfd_priv.h"
>>> #include "kfd_topology.h"
>>> @@ -740,6 +741,25 @@ static int kfd_fill_gpu_cache_info(struct kfd_dev *kdev,
>>> return 0;
>>> }
>>>
>>> +
>>> +#ifdef CONFIG_ACPI
>>> +
>>> +bool kfd_ignore_crat(void)
>>> +{
>>> + bool ret;
>>> +
>>> + if (ignore_crat)
>>> + return true;
>>> +
>>> +#ifndef KFD_SUPPORT_IOMMU_V2
>>> + ret = true;
>>> +#else
>>> + ret = false;
>>> +#endif
>>> +
>>> + return ret;
>>> +}
>>> +
>>> /*
>>> * kfd_create_crat_image_acpi - Allocates memory for CRAT image and
>>> * copies CRAT from ACPI (if available).
>>> @@ -751,7 +771,6 @@ static int kfd_fill_gpu_cache_info(struct kfd_dev *kdev,
>>> *
>>> * Return 0 if successful else return error code
>>> */
>>> -#ifdef CONFIG_ACPI
>>> int kfd_create_crat_image_acpi(void **crat_image, size_t *size)
>>> {
>>> struct acpi_table_header *crat_table;
>>> @@ -775,7 +794,7 @@ int kfd_create_crat_image_acpi(void **crat_image, size_t *size)
>>> return -EINVAL;
>>> }
>>>
>>> - if (ignore_crat) {
>>> + if (kfd_ignore_crat()) {
>>> pr_info("CRAT table disabled by module option\n");
>>> return -ENODATA;
>>> }
>>> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
>>> index 2c030c2b5b8d..fdf64d361be3 100644
>>> --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
>>> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
>>> @@ -112,6 +112,7 @@ static const struct kfd_device_info carrizo_device_info = {
>>> .num_xgmi_sdma_engines = 0,
>>> .num_sdma_queues_per_engine = 2,
>>> };
>>> +#endif
>>>
>>> static const struct kfd_device_info raven_device_info = {
>>> .asic_family = CHIP_RAVEN,
>>> @@ -130,7 +131,6 @@ static const struct kfd_device_info raven_device_info = {
>>> .num_xgmi_sdma_engines = 0,
>>> .num_sdma_queues_per_engine = 2,
>>> };
>>> -#endif
>>>
>>> static const struct kfd_device_info hawaii_device_info = {
>>> .asic_family = CHIP_HAWAII,
>>> @@ -688,6 +688,9 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
>>> goto gws_error;
>>> }
>>>
>>> + /* If CRAT is broken, won't set iommu enabled */
>>> + kfd_double_confirm_iommu_support(kfd);
>>> +
>>> if (kfd_iommu_device_init(kfd)) {
>>> dev_err(kfd_device, "Error initializing iommuv2\n");
>>> goto device_iommu_error;
>>> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
>>> index 82f955750e75..5b70fbe429f1 100644
>>> --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
>>> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
>>> @@ -308,12 +308,14 @@ struct kfd_dev {
>>>
>>> /* xGMI */
>>> uint64_t hive_id;
>>> -
>>> /* UUID */
>>> uint64_t unique_id;
>>>
>>> bool pci_atomic_requested;
>>>
>>> + /* Use IOMMU v2 flag */
>>> + bool use_iommu_v2;
>>> +
>>> /* SRAM ECC flag */
>>> atomic_t sram_ecc_flag;
>>>
>>> @@ -1009,6 +1011,7 @@ struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev);
>>> struct kfd_dev *kfd_device_by_kgd(const struct kgd_dev *kgd);
>>> int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev);
>>> int kfd_numa_node_to_apic_id(int numa_node_id);
>>> +void kfd_double_confirm_iommu_support(struct kfd_dev *gpu);
>>>
>>> /* Interrupts */
>>> int kfd_interrupt_init(struct kfd_dev *dev);
>>> @@ -1232,9 +1235,12 @@ static inline int kfd_devcgroup_check_permission(struct kfd_dev *kfd)
>>> #endif
>>> }
>>>
>>> +bool kfd_ignore_crat(void);
>>> +
>>> static inline bool kfd_device_use_iommu_v2(const struct kfd_dev *dev)
>>> {
>>> - return dev && dev->device_info->needs_iommu_device;
>>> + return !kfd_ignore_crat() && dev && dev->use_iommu_v2 &&
>>> + dev->device_info->needs_iommu_device;
>> I think this could now be simplified:
>>
>> return dev && dev->use_iommu_v2;
>>
>> So maybe you don't need this function any more.
> In Renoir, if ACPI CRAT from SBIOS is good, we may still use
> dev->device_info->needs_iommu_device to confirm whether we should go dGPU.
That should be reflected correctly in the dev->use_iommu_v2 flag.
dev->use_iommu_v2 should not be set on any GPU that doesn't have
dev->device_info->needs_iommu_device. See below.
>
>>
>>> }
>>>
>>> /* Debugfs */
>>> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
>>> index 4b29815e9205..8907b5317103 100644
>>> --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
>>> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
>>> @@ -1579,6 +1579,26 @@ int kfd_numa_node_to_apic_id(int numa_node_id)
>>> return kfd_cpumask_to_apic_id(cpumask_of_node(numa_node_id));
>>> }
>>>
>>> +void kfd_double_confirm_iommu_support(struct kfd_dev *gpu)
>>> +{
>>> + struct kfd_topology_device *dev;
>>> +
>>> + unsigned temp = 0;
>>> +
>>> + down_read(&topology_lock);
>>> +
>>> + /* The cpu_cores_count and simd_count aren't zero at the same time in
>>> + * APU node.
>>> + */
>>> + list_for_each_entry(dev, &topology_device_list, list)
>>> + temp |= dev->node_props.cpu_cores_count *
>>> + dev->node_props.simd_count;
>> You shouldn't look at all GPUs, only at the GPU currently being
>> initialized. Otherwise all your dGPUs in an APU system will also have
>> use_iommu_v2 == true, which would be confusing.
>>
>> I'd do this in kfd_assign_gpu, because at that point you have access to
>> the kfd_topology_device and the kfd_dev at the same time without having
>> to add another loop.
>>
> Actually, I follow your comment to do it like this, however, we have to set
> the use_iommu_v2 before kfd_iommu_device_init(). kfd_assign_gpu in kfd_topology_add_device()
> is a little late.
I see, I missed that. But we still should make sure we set
gpu->use_iommu_v2 only on devices that have
gpu->device_info->needs_iommu_device, and only if this GPU can be
assigned to an APU node in the topology. So then
kfd_double_confirm_iommu_support would look like this:
+ if (!gpu->device_info->needs_iommu_device)
+ return;
+ down_read(&topology_lock);
+ /* Only use IOMMUv2 if there is an APU topology node with no GPU assigned yet.
+ * This GPU will be assigned to it.
+ */
+ list_for_each_entry(dev, &topology_device_list, list)
+ if (dev->node_props.cpu_cores_count &&
+ dev->node_props.simd_count &&
+ !dev->gpu) {
+ gpu->use_iommu_v2 = true;
+ break;
+ }
+ up_read(&topology_lock);
Regards,
Felix
>
> Thanks,
> Ray
>
>> ...
>> list_for_each_entry(dev, &topology_device_list, list) {
>> /* Discrete GPUs need their own topology device list
>> * entries. Don't assign them to CPU/APU nodes.
>> */
>> if (!gpu->device_info->needs_iommu_device &&
>> dev->node_props.cpu_cores_count)
>> continue;
>>
>> if (!dev->gpu && (dev->node_props.simd_count > 0)) {
>> + if (dev->node_props.cpu_cores_count)
>> + dev->use_iommu_v2 = true;
>> ...
>>
>> Regards,
>> Felix
>>
>>
>>> +
>>> + up_read(&topology_lock);
>>> +
>>> + gpu->use_iommu_v2 = temp ? true : false;
>>> +}
>>> +
>>> #if defined(CONFIG_DEBUG_FS)
>>>
>>> int kfd_debugfs_hqds_by_device(struct seq_file *m, void *data)
More information about the amd-gfx
mailing list