mirror of
git://nv-tegra.nvidia.com/linux-nvgpu.git
synced 2025-12-23 18:16:01 +03:00
17451138cf
Fix a race condition in gk20a_get_channel_from_file() that returns a channel pointer from an fd: take a reference to the channel before putting the file ref back. Now the caller is responsible of releasing the channel reference eventually. Also document why dbg_session_channel_data has to hold a ref to the channel file instead of just the channel: that might deadlock if the fds were closed in "wrong" order. Change-Id: I8e91b809f5f7b1cb0c1487bd955ad6d643727a53 Signed-off-by: Konsta Holtta <kholtta@nvidia.com> Reviewed-on: https://git-master.nvidia.com/r/1549290 Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com> Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
151 lines
4.7 KiB
C
151 lines
4.7 KiB
C
/*
|
|
* Tegra GK20A GPU Debugger Driver
|
|
*
|
|
* Copyright (c) 2013-2017, NVIDIA CORPORATION. All rights reserved.
|
|
*
|
|
* This program is free software; you can redistribute it and/or modify it
|
|
* under the terms and conditions of the GNU General Public License,
|
|
* version 2, as published by the Free Software Foundation.
|
|
*
|
|
* This program is distributed in the hope it will be useful, but WITHOUT
|
|
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
|
|
* more details.
|
|
*
|
|
* You should have received a copy of the GNU General Public License
|
|
* along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
*/
|
|
#ifndef DBG_GPU_GK20A_H
|
|
#define DBG_GPU_GK20A_H
|
|
#include <linux/poll.h>
|
|
|
|
/* module debug driver interface */
|
|
int gk20a_dbg_gpu_dev_release(struct inode *inode, struct file *filp);
|
|
int gk20a_dbg_gpu_dev_open(struct inode *inode, struct file *filp);
|
|
long gk20a_dbg_gpu_dev_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
|
|
unsigned int gk20a_dbg_gpu_dev_poll(struct file *filep, poll_table *wait);
|
|
|
|
/* used by profiler driver interface */
|
|
int gk20a_prof_gpu_dev_open(struct inode *inode, struct file *filp);
|
|
|
|
/* used by the interrupt handler to post events */
|
|
void gk20a_dbg_gpu_post_events(struct channel_gk20a *fault_ch);
|
|
|
|
struct channel_gk20a *
|
|
nvgpu_dbg_gpu_get_session_channel(struct dbg_session_gk20a *dbg_s);
|
|
|
|
struct dbg_gpu_session_events {
|
|
wait_queue_head_t wait_queue;
|
|
bool events_enabled;
|
|
int num_pending_events;
|
|
};
|
|
|
|
struct dbg_session_gk20a {
|
|
/* dbg session id used for trace/prints */
|
|
int id;
|
|
|
|
/* profiler session, if any */
|
|
bool is_profiler;
|
|
|
|
/* has a valid profiler reservation */
|
|
bool has_profiler_reservation;
|
|
|
|
/* power enabled or disabled */
|
|
bool is_pg_disabled;
|
|
|
|
/* timeouts enabled or disabled */
|
|
bool is_timeout_disabled;
|
|
|
|
/*
|
|
* There can be different versions of the whitelists
|
|
* between both global and per-context sets; as well
|
|
* as between debugger and profiler interfaces.
|
|
*/
|
|
struct regops_whitelist *global;
|
|
struct regops_whitelist *per_context;
|
|
|
|
/* gpu module vagaries */
|
|
struct device *dev;
|
|
struct gk20a *g;
|
|
|
|
/* list of bound channels, if any */
|
|
struct nvgpu_list_node ch_list;
|
|
struct nvgpu_mutex ch_list_lock;
|
|
|
|
/* event support */
|
|
struct dbg_gpu_session_events dbg_events;
|
|
|
|
bool broadcast_stop_trigger;
|
|
|
|
struct nvgpu_mutex ioctl_lock;
|
|
};
|
|
|
|
struct dbg_session_data {
|
|
struct dbg_session_gk20a *dbg_s;
|
|
struct nvgpu_list_node dbg_s_entry;
|
|
};
|
|
|
|
static inline struct dbg_session_data *
|
|
dbg_session_data_from_dbg_s_entry(struct nvgpu_list_node *node)
|
|
{
|
|
return (struct dbg_session_data *)
|
|
((uintptr_t)node - offsetof(struct dbg_session_data, dbg_s_entry));
|
|
};
|
|
|
|
struct dbg_session_channel_data {
|
|
/*
|
|
* We have to keep a ref to the _file_, not the channel, because
|
|
* close(channel_fd) is synchronous and would deadlock if we had an
|
|
* open debug session fd holding a channel ref at that time. Holding a
|
|
* ref to the file makes close(channel_fd) just drop a kernel ref to
|
|
* the file; the channel will close when the last file ref is dropped.
|
|
*/
|
|
struct file *ch_f;
|
|
int channel_fd;
|
|
int chid;
|
|
struct nvgpu_list_node ch_entry;
|
|
struct dbg_session_data *session_data;
|
|
};
|
|
|
|
static inline struct dbg_session_channel_data *
|
|
dbg_session_channel_data_from_ch_entry(struct nvgpu_list_node *node)
|
|
{
|
|
return (struct dbg_session_channel_data *)
|
|
((uintptr_t)node - offsetof(struct dbg_session_channel_data, ch_entry));
|
|
};
|
|
|
|
struct dbg_profiler_object_data {
|
|
int session_id;
|
|
u32 prof_handle;
|
|
struct channel_gk20a *ch;
|
|
bool has_reservation;
|
|
struct nvgpu_list_node prof_obj_entry;
|
|
};
|
|
|
|
static inline struct dbg_profiler_object_data *
|
|
dbg_profiler_object_data_from_prof_obj_entry(struct nvgpu_list_node *node)
|
|
{
|
|
return (struct dbg_profiler_object_data *)
|
|
((uintptr_t)node - offsetof(struct dbg_profiler_object_data, prof_obj_entry));
|
|
};
|
|
|
|
int dbg_unbind_single_channel_gk20a(struct dbg_session_gk20a *dbg_s,
|
|
struct dbg_session_channel_data *ch_data);
|
|
|
|
bool gk20a_dbg_gpu_broadcast_stop_trigger(struct channel_gk20a *ch);
|
|
int gk20a_dbg_gpu_clear_broadcast_stop_trigger(struct channel_gk20a *ch);
|
|
|
|
int dbg_set_powergate(struct dbg_session_gk20a *dbg_s, u32 powermode);
|
|
bool nvgpu_check_and_set_global_reservation(
|
|
struct dbg_session_gk20a *dbg_s,
|
|
struct dbg_profiler_object_data *prof_obj);
|
|
bool nvgpu_check_and_set_context_reservation(
|
|
struct dbg_session_gk20a *dbg_s,
|
|
struct dbg_profiler_object_data *prof_obj);
|
|
void nvgpu_release_profiler_reservation(struct dbg_session_gk20a *dbg_s,
|
|
struct dbg_profiler_object_data *prof_obj);
|
|
int gk20a_perfbuf_enable_locked(struct gk20a *g, u64 offset, u32 size);
|
|
int gk20a_perfbuf_disable_locked(struct gk20a *g);
|
|
|
|
#endif /* DBG_GPU_GK20A_H */
|