nv-tegra-mirror/linux-nvgpu
1
0
Fork 0
mirror of git://nv-tegra.nvidia.com/linux-nvgpu.git synced 2025-12-22 09:12:24 +03:00
Code Issues Packages Projects Releases Wiki Activity

gpu: nvgpu: Expose more config info for SM

Browse Source 
This patch exposes the following fields to userspace for
each SM, indexed by its global SM index:
- Logical GPC index
- Virtual GPC index
- Local Logical TPC index
- Global Logical TPC index
- Migratable TPC index
- Local SM index

Bug 4115114

Change-Id: Ie596c8f617d46d7b6f1563f7afb210b59f250f3c
Signed-off-by: Martin Radev <mradev@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2909784
Reviewed-by: Rajesh Devaraj <rdevaraj@nvidia.com>
Reviewed-by: Ankur Kishore <ankkishore@nvidia.com>
Reviewed-by: Vijayakumar Subbu <vsubbu@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
This commit is contained in:
Martin Radev
2023-05-24 13:11:48 +03:00
committed by mobile promotions
parent d6359b5adc
commit 78bef566dd
2 changed files with 60 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
drivers/gpu/nvgpu/os/linux/ioctl_ctrl.c
View File

@@ -1151,31 +1151,65 @@ static int gk20a_ctrl_vsm_mapping(struct gk20a *g,
struct nvgpu_gr_config *gr_config, struct nvgpu_gpu_vsms_mapping *args)
{
int err = 0;
u32 gpc_max_count;
u32 no_of_sm = nvgpu_gr_config_get_no_of_sm(gr_config);
size_t write_size = no_of_sm *
sizeof(struct nvgpu_gpu_vsms_mapping_entry);
struct nvgpu_gpu_vsms_mapping_entry *vsms_buf;
u32 i;
u32 gpc_logical_index;
u32 gpc_virtual_index;
u32 tpc_local_logical_index;
u32 tpc_global_logical_index;
u32 tpc_migratable_index;
u32 sm_local_id;
gpc_max_count = nvgpu_gr_config_get_max_gpc_count(gr_config);
vsms_buf = nvgpu_kzalloc(g, write_size);
if (vsms_buf == NULL)
return -ENOMEM;
for (i = 0; i < no_of_sm; i++) {
struct nvgpu_sm_info *sm_info =
nvgpu_gr_config_get_sm_info(gr_config, i);
struct nvgpu_sm_info *sm_info = nvgpu_gr_config_get_sm_info(gr_config, i);
vsms_buf[i].gpc_index =
nvgpu_gr_config_get_sm_info_gpc_index(sm_info);
gpc_logical_index = nvgpu_gr_config_get_sm_info_gpc_index(sm_info);
gpc_virtual_index = nvgpu_gr_config_get_sm_info_virtual_gpc_index(sm_info);
if (g->ops.gr.init.get_nonpes_aware_tpc)
vsms_buf[i].tpc_index =
tpc_local_logical_index =
g->ops.gr.init.get_nonpes_aware_tpc(g,
nvgpu_gr_config_get_sm_info_gpc_index(sm_info),
nvgpu_gr_config_get_sm_info_tpc_index(sm_info),
gpc_logical_index,
nvgpu_gr_config_get_sm_info_tpc_index(sm_info),
gr_config);
else
vsms_buf[i].tpc_index =
nvgpu_gr_config_get_sm_info_tpc_index(sm_info);
tpc_local_logical_index = nvgpu_gr_config_get_sm_info_tpc_index(sm_info);
tpc_global_logical_index = nvgpu_gr_config_get_sm_info_global_tpc_index(sm_info);
sm_local_id = nvgpu_gr_config_get_sm_info_sm_index(sm_info);
if (vsms_buf[i].gpc_virtual_index < gpc_max_count)
tpc_migratable_index = tpc_local_logical_index;
else {
/*
* The TPC migratable index is 0 for any Virtual GPC index >= the maximal number of GPCs.
*/
tpc_migratable_index = 0U;
}
if (gpc_logical_index > U8_MAX ||
gpc_virtual_index > U8_MAX ||
tpc_local_logical_index > U8_MAX ||
tpc_global_logical_index > U8_MAX ||
tpc_migratable_index > U8_MAX ||
sm_local_id > U8_MAX) {
nvgpu_kfree(g, vsms_buf);
return -EOVERFLOW;
}
vsms_buf[i].gpc_logical_index = (u8)gpc_logical_index;
vsms_buf[i].gpc_virtual_index = (u8)gpc_virtual_index;
vsms_buf[i].tpc_local_logical_index = (u8)tpc_local_logical_index;
vsms_buf[i].tpc_global_logical_index = (u8)tpc_global_logical_index;
vsms_buf[i].sm_local_id = (u8)sm_local_id;
vsms_buf[i].tpc_migratable_index = (u8)tpc_migratable_index;
}
err = copy_to_user((void __user *)(uintptr_t)