tegrakernel/kernel/nvidia/drivers/misc/eventlib/eventlib_init.h

148 lines
5.2 KiB
C
Raw Permalink Normal View History

2022-02-16 09:13:02 -06:00
/*
* Copyright (c) 2016-2017, NVIDIA CORPORATION. All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
* DEALINGS IN THE SOFTWARE.
*/
#ifndef EVENTLIB_INIT_H
#define EVENTLIB_INIT_H
/* All pointers targeted to shared memory are annotated with 'shmptr' */
#define shmptr /* empty, pure annotation */
#include "eventlib.h"
#include "utility.h"
#include "eventlib_tbuf.h"
#include "eventlib_flt.h"
#define ALIGN_64_DOWN(x) ((x) & ~((typeof(x))(sizeof(uint64_t) - 1)))
#define ALIGN_64_UP(x) ALIGN_64_DOWN((x) + ((uint32_t)sizeof(uint64_t) - 1))
#define ALIGN_64_CHECK(x) ((x) == ALIGN_64_DOWN(x))
/* Eventlib uses two shared memory blocks, one is read only for readers,
* and other is read-write for readers.
*
* These blocks can be physically accessed by different eventlib builds
* simultaneuosly. Thus special mesures have to be taken to ensure
* compatibility.
*
* - Blocks are separated into subblocks corresponding to subsystems. At this
* time, subsystems are "tracebuf" and "filtering". One more could be -
* alive management. Room for subsystems costs 8 bytes per subsystem per
* shmem block - thus should not be an issue to reserve space for 8 of them.
* - Each subblock holds some sort of version/compatibility information, that
* can be used in subsystem-specific way to find out if reader's
* implementation of this subsystem can understand writer's implementation
* of this subsystem. Details are subsystem-specific.
* - Thus reader can understand some of writer's subsystems, but maybe not
* all. Understanding 'tracebuf' subsystem is required.Not understanding
* other subsystem is possible but then reader will have only limited
* information.
*/
#define EVENTLIB_SUBSYS_TRACEBUF 0
#define EVENTLIB_SUBSYS_FILTERING 1
#define EVENTLIB_SUBSYS_MAX 2
#define EVENTLIB_TBUFS_MAX 6
#define EVENTLIB_MAGIC_W2R 0x52574c45 /* 'ELWR' in little endian */
#define EVENTLIB_MAGIC_R2W 0x57524c45 /* 'ELRW' in little endian */
struct eventlib_shared {
/* Different magic is used for two shared memory blocks */
uint32_t magic;
/* Number of trace buffers
* Below value is updated from writer context
*/
uint32_t num_buffers;
/* Locatons of subsystem's information. If subsystem is not presented
* in particular block, corresponding offset/size are kept zeroed.
*/
struct {
uint32_t offset;
uint32_t size;
} subsys[EVENTLIB_SUBSYS_MAX];
} __attribute__((__packed__));
struct eventlib_init {
/* writer, init time: offsets of currently unallocated space */
uint32_t w2r_offset, r2w_offset;
/* Local copy of subsystem information is used to protect against
* unexpected asynchronous changes (not expected under normal use).
*/
struct eventlib_shared w2r_copy;
struct eventlib_shared r2w_copy;
struct eventlib_tbuf_ctx tbuf[EVENTLIB_TBUFS_MAX];
struct eventlib_flt_ctx flt;
};
static inline void *subsys_shm_area(shmptr struct eventlib_shared *sh,
struct eventlib_shared *copy, int subsys)
{
if (sh == NULL)
return NULL;
return ((void *)((uintptr_t)sh + copy->subsys[subsys].offset));
}
#define subsys_w2r_area(ctx, subsys) \
subsys_shm_area(ctx->w2r_shm, &ctx->priv->w2r_copy, \
EVENTLIB_SUBSYS_ ## subsys)
#define subsys_r2w_area(ctx, subsys) \
subsys_shm_area(ctx->r2w_shm, &ctx->priv->r2w_copy, \
EVENTLIB_SUBSYS_ ## subsys)
static inline uint32_t subsys_shm_size(struct eventlib_shared *copy, int subsys)
{
return copy->subsys[subsys].size;
}
#define subsys_w2r_size(ctx, subsys) \
subsys_shm_size(&ctx->priv->w2r_copy, EVENTLIB_SUBSYS_ ## subsys)
#define subsys_r2w_size(ctx, subsys) \
subsys_shm_size(&ctx->priv->r2w_copy, EVENTLIB_SUBSYS_ ## subsys)
static inline int subsys_shm_carve(struct eventlib_shared *copy,
uint32_t *shm_size, uint32_t *shm_offset, int subsys, uint32_t size)
{
if (size > *shm_size - *shm_offset)
return -ENOSPC;
copy->subsys[subsys].offset = *shm_offset;
copy->subsys[subsys].size = size;
*shm_offset += ALIGN_64_UP(size);
return 0;
}
#define subsys_w2r_carve(ctx, subsys, size) \
subsys_shm_carve(&ctx->priv->w2r_copy, &ctx->w2r_shm_size, \
&ctx->priv->w2r_offset, EVENTLIB_SUBSYS_ ## subsys, size)
#define subsys_r2w_carve(ctx, subsys, size) \
subsys_shm_carve(&ctx->priv->r2w_copy, &ctx->r2w_shm_size, \
&ctx->priv->r2w_offset, EVENTLIB_SUBSYS_ ## subsys, size)
#endif