Commit f7c826ad authored by Alexey Skidanov's avatar Alexey Skidanov Committed by Oded Gabbay

drm/amdkfd: Add number of watch points to topology

This patch adds the number of watch points to the node capabilities in the
topology module
Signed-off-by: default avatarAlexey Skidanov <Alexey.Skidanov@amd.com>
Signed-off-by: default avatarOded Gabbay <oded.gabbay@amd.com>
parent c93546a5
...@@ -32,6 +32,7 @@ ...@@ -32,6 +32,7 @@
static const struct kfd_device_info kaveri_device_info = { static const struct kfd_device_info kaveri_device_info = {
.max_pasid_bits = 16, .max_pasid_bits = 16,
.ih_ring_entry_size = 4 * sizeof(uint32_t), .ih_ring_entry_size = 4 * sizeof(uint32_t),
.num_of_watch_points = 4,
.mqd_size_aligned = MQD_SIZE_ALIGNED .mqd_size_aligned = MQD_SIZE_ALIGNED
}; };
......
...@@ -107,6 +107,7 @@ enum cache_policy { ...@@ -107,6 +107,7 @@ enum cache_policy {
struct kfd_device_info { struct kfd_device_info {
unsigned int max_pasid_bits; unsigned int max_pasid_bits;
size_t ih_ring_entry_size; size_t ih_ring_entry_size;
uint8_t num_of_watch_points;
uint16_t mqd_size_aligned; uint16_t mqd_size_aligned;
}; };
......
...@@ -27,6 +27,7 @@ ...@@ -27,6 +27,7 @@
#include <linux/acpi.h> #include <linux/acpi.h>
#include <linux/hash.h> #include <linux/hash.h>
#include <linux/cpufreq.h> #include <linux/cpufreq.h>
#include <linux/log2.h>
#include "kfd_priv.h" #include "kfd_priv.h"
#include "kfd_crat.h" #include "kfd_crat.h"
...@@ -630,10 +631,10 @@ static struct kobj_type cache_type = { ...@@ -630,10 +631,10 @@ static struct kobj_type cache_type = {
static ssize_t node_show(struct kobject *kobj, struct attribute *attr, static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
char *buffer) char *buffer)
{ {
ssize_t ret;
struct kfd_topology_device *dev; struct kfd_topology_device *dev;
char public_name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE]; char public_name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE];
uint32_t i; uint32_t i;
uint32_t log_max_watch_addr;
/* Making sure that the buffer is an empty string */ /* Making sure that the buffer is an empty string */
buffer[0] = 0; buffer[0] = 0;
...@@ -641,8 +642,10 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, ...@@ -641,8 +642,10 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
if (strcmp(attr->name, "gpu_id") == 0) { if (strcmp(attr->name, "gpu_id") == 0) {
dev = container_of(attr, struct kfd_topology_device, dev = container_of(attr, struct kfd_topology_device,
attr_gpuid); attr_gpuid);
ret = sysfs_show_32bit_val(buffer, dev->gpu_id); return sysfs_show_32bit_val(buffer, dev->gpu_id);
} else if (strcmp(attr->name, "name") == 0) { }
if (strcmp(attr->name, "name") == 0) {
dev = container_of(attr, struct kfd_topology_device, dev = container_of(attr, struct kfd_topology_device,
attr_name); attr_name);
for (i = 0; i < KFD_TOPOLOGY_PUBLIC_NAME_SIZE; i++) { for (i = 0; i < KFD_TOPOLOGY_PUBLIC_NAME_SIZE; i++) {
...@@ -652,80 +655,90 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, ...@@ -652,80 +655,90 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
break; break;
} }
public_name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE-1] = 0x0; public_name[KFD_TOPOLOGY_PUBLIC_NAME_SIZE-1] = 0x0;
ret = sysfs_show_str_val(buffer, public_name); return sysfs_show_str_val(buffer, public_name);
} else { }
dev = container_of(attr, struct kfd_topology_device,
attr_props);
sysfs_show_32bit_prop(buffer, "cpu_cores_count",
dev->node_props.cpu_cores_count);
sysfs_show_32bit_prop(buffer, "simd_count",
dev->node_props.simd_count);
if (dev->mem_bank_count < dev->node_props.mem_banks_count) {
pr_warn("kfd: mem_banks_count truncated from %d to %d\n",
dev->node_props.mem_banks_count,
dev->mem_bank_count);
sysfs_show_32bit_prop(buffer, "mem_banks_count",
dev->mem_bank_count);
} else {
sysfs_show_32bit_prop(buffer, "mem_banks_count",
dev->node_props.mem_banks_count);
}
sysfs_show_32bit_prop(buffer, "caches_count", dev = container_of(attr, struct kfd_topology_device,
dev->node_props.caches_count); attr_props);
sysfs_show_32bit_prop(buffer, "io_links_count", sysfs_show_32bit_prop(buffer, "cpu_cores_count",
dev->node_props.io_links_count); dev->node_props.cpu_cores_count);
sysfs_show_32bit_prop(buffer, "cpu_core_id_base", sysfs_show_32bit_prop(buffer, "simd_count",
dev->node_props.cpu_core_id_base); dev->node_props.simd_count);
sysfs_show_32bit_prop(buffer, "simd_id_base",
dev->node_props.simd_id_base); if (dev->mem_bank_count < dev->node_props.mem_banks_count) {
sysfs_show_32bit_prop(buffer, "capability", pr_warn("kfd: mem_banks_count truncated from %d to %d\n",
dev->node_props.capability); dev->node_props.mem_banks_count,
sysfs_show_32bit_prop(buffer, "max_waves_per_simd", dev->mem_bank_count);
dev->node_props.max_waves_per_simd); sysfs_show_32bit_prop(buffer, "mem_banks_count",
sysfs_show_32bit_prop(buffer, "lds_size_in_kb", dev->mem_bank_count);
dev->node_props.lds_size_in_kb); } else {
sysfs_show_32bit_prop(buffer, "gds_size_in_kb", sysfs_show_32bit_prop(buffer, "mem_banks_count",
dev->node_props.gds_size_in_kb); dev->node_props.mem_banks_count);
sysfs_show_32bit_prop(buffer, "wave_front_size", }
dev->node_props.wave_front_size);
sysfs_show_32bit_prop(buffer, "array_count",
dev->node_props.array_count);
sysfs_show_32bit_prop(buffer, "simd_arrays_per_engine",
dev->node_props.simd_arrays_per_engine);
sysfs_show_32bit_prop(buffer, "cu_per_simd_array",
dev->node_props.cu_per_simd_array);
sysfs_show_32bit_prop(buffer, "simd_per_cu",
dev->node_props.simd_per_cu);
sysfs_show_32bit_prop(buffer, "max_slots_scratch_cu",
dev->node_props.max_slots_scratch_cu);
sysfs_show_32bit_prop(buffer, "vendor_id",
dev->node_props.vendor_id);
sysfs_show_32bit_prop(buffer, "device_id",
dev->node_props.device_id);
sysfs_show_32bit_prop(buffer, "location_id",
dev->node_props.location_id);
if (dev->gpu) {
sysfs_show_32bit_prop(buffer, "max_engine_clk_fcompute",
kfd2kgd->get_max_engine_clock_in_mhz(
dev->gpu->kgd));
sysfs_show_64bit_prop(buffer, "local_mem_size",
kfd2kgd->get_vmem_size(dev->gpu->kgd));
sysfs_show_32bit_prop(buffer, "fw_version",
kfd2kgd->get_fw_version(
dev->gpu->kgd,
KGD_ENGINE_MEC1));
sysfs_show_32bit_prop(buffer, "caches_count",
dev->node_props.caches_count);
sysfs_show_32bit_prop(buffer, "io_links_count",
dev->node_props.io_links_count);
sysfs_show_32bit_prop(buffer, "cpu_core_id_base",
dev->node_props.cpu_core_id_base);
sysfs_show_32bit_prop(buffer, "simd_id_base",
dev->node_props.simd_id_base);
sysfs_show_32bit_prop(buffer, "capability",
dev->node_props.capability);
sysfs_show_32bit_prop(buffer, "max_waves_per_simd",
dev->node_props.max_waves_per_simd);
sysfs_show_32bit_prop(buffer, "lds_size_in_kb",
dev->node_props.lds_size_in_kb);
sysfs_show_32bit_prop(buffer, "gds_size_in_kb",
dev->node_props.gds_size_in_kb);
sysfs_show_32bit_prop(buffer, "wave_front_size",
dev->node_props.wave_front_size);
sysfs_show_32bit_prop(buffer, "array_count",
dev->node_props.array_count);
sysfs_show_32bit_prop(buffer, "simd_arrays_per_engine",
dev->node_props.simd_arrays_per_engine);
sysfs_show_32bit_prop(buffer, "cu_per_simd_array",
dev->node_props.cu_per_simd_array);
sysfs_show_32bit_prop(buffer, "simd_per_cu",
dev->node_props.simd_per_cu);
sysfs_show_32bit_prop(buffer, "max_slots_scratch_cu",
dev->node_props.max_slots_scratch_cu);
sysfs_show_32bit_prop(buffer, "vendor_id",
dev->node_props.vendor_id);
sysfs_show_32bit_prop(buffer, "device_id",
dev->node_props.device_id);
sysfs_show_32bit_prop(buffer, "location_id",
dev->node_props.location_id);
if (dev->gpu) {
log_max_watch_addr =
__ilog2_u32(dev->gpu->device_info->num_of_watch_points);
if (log_max_watch_addr) {
dev->node_props.capability |=
HSA_CAP_WATCH_POINTS_SUPPORTED;
dev->node_props.capability |=
((log_max_watch_addr <<
HSA_CAP_WATCH_POINTS_TOTALBITS_SHIFT) &
HSA_CAP_WATCH_POINTS_TOTALBITS_MASK);
} }
ret = sysfs_show_32bit_prop(buffer, "max_engine_clk_ccompute", sysfs_show_32bit_prop(buffer, "max_engine_clk_fcompute",
cpufreq_quick_get_max(0)/1000); kfd2kgd->get_max_engine_clock_in_mhz(
dev->gpu->kgd));
sysfs_show_64bit_prop(buffer, "local_mem_size",
kfd2kgd->get_vmem_size(dev->gpu->kgd));
sysfs_show_32bit_prop(buffer, "fw_version",
kfd2kgd->get_fw_version(
dev->gpu->kgd,
KGD_ENGINE_MEC1));
} }
return ret; return sysfs_show_32bit_prop(buffer, "max_engine_clk_ccompute",
cpufreq_quick_get_max(0)/1000);
} }
static const struct sysfs_ops node_ops = { static const struct sysfs_ops node_ops = {
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment