gpu: nvgpu: add separate unit for debugger

Rename gk20a/dbg_gpu_gk20a.c to common/debugger.c and make it a
separate common unit
Also rename gk20a/dbg_gpu_gk20a.h to include/nvgpu/debugger.h

We had two different HALs for debugger - gops.debugger and
gops.dbg_session_ops
Combine them into one single HAL gops.debugger and remove
gops.dbg_session_ops

Rename all exported APIs from debugger.h to be in the form of
nvgpu_*()

Jira NVGPU-1013

Change-Id: I136dc7786e3b2065921eb03b99f16049212f3cd2
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1920075
Reviewed-by: Sachin Jadhav <sachinj@nvidia.com>
Tested-by: Sachin Jadhav <sachinj@nvidia.com>
This commit is contained in:
Deepak Nibade
2018-10-05 16:18:01 +05:30
committed by Sachin Jadhav
parent 625fa68d2a
commit 1b2a0833e0
25 changed files with 63 additions and 94 deletions

View File

@@ -22,8 +22,6 @@
#include <nvgpu/ctxsw_trace.h>
#include <nvgpu/gk20a.h>
#include "gk20a/dbg_gpu_gk20a.h"
#include "ioctl_channel.h"
#include "ioctl_ctrl.h"
#include "ioctl_as.h"

View File

@@ -38,7 +38,6 @@
#include <nvgpu/channel.h>
#include <nvgpu/channel_sync.h>
#include "gk20a/dbg_gpu_gk20a.h"
#include "gk20a/fence_gk20a.h"
#include "platform_gk20a.h"

View File

@@ -29,6 +29,7 @@
#include <nvgpu/vm.h>
#include <nvgpu/atomic.h>
#include <nvgpu/cond.h>
#include <nvgpu/debugger.h>
#include <nvgpu/utils.h>
#include <nvgpu/gk20a.h>
#include <nvgpu/channel.h>
@@ -38,7 +39,6 @@
#include "gk20a/gr_gk20a.h"
#include "gk20a/regops_gk20a.h"
#include "gk20a/dbg_gpu_gk20a.h"
#include "os_linux.h"
#include "platform_gk20a.h"
#include "ioctl_dbg.h"
@@ -214,7 +214,7 @@ int gk20a_dbg_gpu_dev_release(struct inode *inode, struct file *filp)
dbg_profiler_object_data, prof_obj_entry) {
if (prof_obj->session_id == dbg_s->id) {
if (prof_obj->has_reservation)
g->ops.dbg_session_ops.
g->ops.debugger.
release_profiler_reservation(dbg_s, prof_obj);
nvgpu_list_del(&prof_obj->prof_obj_entry);
nvgpu_kfree(g, prof_obj);
@@ -469,7 +469,7 @@ static int dbg_unbind_single_channel_gk20a(struct dbg_session_gk20a *dbg_s,
if ((prof_obj->session_id == dbg_s->id) &&
(prof_obj->ch->chid == chid)) {
if (prof_obj->has_reservation) {
g->ops.dbg_session_ops.
g->ops.debugger.
release_profiler_reservation(dbg_s, prof_obj);
}
nvgpu_list_del(&prof_obj->prof_obj_entry);
@@ -1189,7 +1189,7 @@ static int nvgpu_ioctl_free_profiler_object(
break;
}
if (prof_obj->has_reservation)
g->ops.dbg_session_ops.
g->ops.debugger.
release_profiler_reservation(dbg_s, prof_obj);
nvgpu_list_del(&prof_obj->prof_obj_entry);
nvgpu_kfree(g, prof_obj);
@@ -1645,7 +1645,7 @@ static int nvgpu_profiler_reserve_release(struct dbg_session_gk20a *dbg_s,
}
if (prof_obj->has_reservation)
g->ops.dbg_session_ops.release_profiler_reservation(dbg_s, prof_obj);
g->ops.debugger.release_profiler_reservation(dbg_s, prof_obj);
else {
nvgpu_err(g, "No reservation found");
err = -EINVAL;
@@ -1691,7 +1691,7 @@ static int nvgpu_profiler_reserve_acquire(struct dbg_session_gk20a *dbg_s,
/* Global reservations are only allowed if there are no other
* global or per-context reservations currently held
*/
if (!g->ops.dbg_session_ops.check_and_set_global_reservation(
if (!g->ops.debugger.check_and_set_global_reservation(
dbg_s, my_prof_obj)) {
nvgpu_err(g,
"global reserve: have existing reservation");
@@ -1721,7 +1721,7 @@ static int nvgpu_profiler_reserve_acquire(struct dbg_session_gk20a *dbg_s,
}
}
if (!g->ops.dbg_session_ops.check_and_set_context_reservation(
if (!g->ops.debugger.check_and_set_context_reservation(
dbg_s, my_prof_obj)) {
/* Another guest OS has the global reservation */
nvgpu_err(g,
@@ -1745,7 +1745,7 @@ static int nvgpu_profiler_reserve_acquire(struct dbg_session_gk20a *dbg_s,
}
}
if (!g->ops.dbg_session_ops.check_and_set_context_reservation(
if (!g->ops.debugger.check_and_set_context_reservation(
dbg_s, my_prof_obj)) {
/* Another guest OS has the global reservation */
nvgpu_err(g,