[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <bc73b7ca-d61d-f805-f4ce-e72f7f6cce6e@quicinc.com>
Date: Tue, 15 Aug 2023 00:42:35 +0530
From: Dikshita Agarwal <quic_dikshita@...cinc.com>
To: Krzysztof Kozlowski <krzk@...nel.org>,
Vikash Garodia <quic_vgarodia@...cinc.com>,
<stanimir.k.varbanov@...il.com>, <agross@...nel.org>,
<andersson@...nel.org>, <konrad.dybcio@...aro.org>,
<mchehab@...nel.org>, <hans.verkuil@...co.com>,
<linux-kernel@...r.kernel.org>, <linux-media@...r.kernel.org>,
<linux-arm-msm@...r.kernel.org>
Subject: Re: [PATCH 24/33] iris: vidc: add debug files
On 8/1/2023 3:01 AM, Krzysztof Kozlowski wrote:
> On 28/07/2023 15:23, Vikash Garodia wrote:
>> this implements the debugging framework.
>
> Your commit msgs are not helping to understand why do you need it and
> what is this doing. Based on this commit description I would ask you to
> drop most of this code as it looks useless. Extend the commit msg to
> provide proper justification and list of features each unit provides.
>
> Please do not use "This commit/patch", but imperative mood. See longer
> explanation here:
> https://elixir.bootlin.com/linux/v5.17.1/source/Documentation/process/submitting-patches.rst#L95
>
>>
>> Signed-off-by: Dikshita Agarwal <quic_dikshita@...cinc.com>
>> Signed-off-by: Vikash Garodia <quic_vgarodia@...cinc.com>
>> ---
>> .../platform/qcom/iris/vidc/inc/msm_vidc_debug.h | 186 +++++++
>> .../platform/qcom/iris/vidc/src/msm_vidc_debug.c | 581 +++++++++++++++++++++
>> 2 files changed, 767 insertions(+)
>> create mode 100644 drivers/media/platform/qcom/iris/vidc/inc/msm_vidc_debug.h
>> create mode 100644 drivers/media/platform/qcom/iris/vidc/src/msm_vidc_debug.c
>>
>> diff --git a/drivers/media/platform/qcom/iris/vidc/inc/msm_vidc_debug.h b/drivers/media/platform/qcom/iris/vidc/inc/msm_vidc_debug.h
>> new file mode 100644
>> index 0000000..ffced01
>> --- /dev/null
>> +++ b/drivers/media/platform/qcom/iris/vidc/inc/msm_vidc_debug.h
>> @@ -0,0 +1,186 @@
>> +/* SPDX-License-Identifier: GPL-2.0-only */
>> +/*
>> + * Copyright (c) 2020-2021, The Linux Foundation. All rights reserved.
>> + * Copyright (c) 2023 Qualcomm Innovation Center, Inc. All rights reserved.
>> + */
>> +
>> +#ifndef __MSM_VIDC_DEBUG__
>> +#define __MSM_VIDC_DEBUG__
>> +
>> +#include <linux/debugfs.h>
>> +#include <linux/delay.h>
>> +#include <linux/errno.h>
>> +#include <linux/module.h>
>> +#include <linux/moduleparam.h>
>> +#include <linux/types.h>
>> +
>> +struct msm_vidc_core;
>> +struct msm_vidc_inst;
>> +
>> +#ifndef VIDC_DBG_LABEL
>> +#define VIDC_DBG_LABEL "msm_vidc"
>> +#endif
>
> Drop these three. Don't re-invent Linux kernel API.
>
>> +
>> +/* Allow only 6 prints/sec */
>> +#define VIDC_DBG_SESSION_RATELIMIT_INTERVAL (1 * HZ)
>> +#define VIDC_DBG_SESSION_RATELIMIT_BURST 6
>> +
>> +#define VIDC_DBG_TAG_INST VIDC_DBG_LABEL ": %4s: %s: "
>> +#define VIDC_DBG_TAG_CORE VIDC_DBG_LABEL ": %4s: %08x: %s: "
>> +#define FW_DBG_TAG VIDC_DBG_LABEL ": %6s: "
>> +#define DEFAULT_SID ((u32)-1)
>> +
>> +#ifndef MSM_VIDC_EMPTY_BRACE
>> +#define MSM_VIDC_EMPTY_BRACE {},
>
> That's the funniest code I saw since some time.
>
>> +#endif
>> +
>> +extern unsigned int msm_vidc_debug;
>
> Nope.
>
>> +extern unsigned int msm_fw_debug;
>
> Nope.
>
>> +extern bool msm_vidc_fw_dump;
>
> Nope.
>
>> +
>> +/* do not modify the log message as it is used in test scripts */
>> +#define FMT_STRING_SET_CTRL \
>> + "%s: state %s, name %s, id 0x%x value %d\n"
>> +#define FMT_STRING_STATE_CHANGE \
>> + "%s: state changed to %s from %s\n"
>> +#define FMT_STRING_MSG_SFR \
>> + "SFR Message from FW: %s\n"
>> +#define FMT_STRING_FAULT_HANDLER \
>> + "%s: faulting address: %lx\n"
>> +#define FMT_STRING_SET_CAP \
>> + "set cap: name: %24s, cap value: %#10x, hfi: %#10llx\n"
>> +
>> +/* To enable messages OR these values and
>> + * echo the result to debugfs file.
>> + *
>> + * To enable all messages set msm_vidc_debug = 0x101F
>> + */
>> +
>> +enum vidc_msg_prio_drv {
>> + VIDC_ERR = 0x00000001,
>> + VIDC_HIGH = 0x00000002,
>> + VIDC_LOW = 0x00000004,
>> + VIDC_PERF = 0x00000008,
>> + VIDC_PKT = 0x00000010,
>> + VIDC_BUS = 0x00000020,
>> + VIDC_STAT = 0x00000040,
>> + VIDC_ENCODER = 0x00000100,
>> + VIDC_DECODER = 0x00000200,
>> + VIDC_PRINTK = 0x10000000,
>> + VIDC_FTRACE = 0x20000000,
>> +};
>> +
>> +enum vidc_msg_prio_fw {
>> + FW_LOW = 0x00000001,
>> + FW_MED = 0x00000002,
>> + FW_HIGH = 0x00000004,
>> + FW_ERROR = 0x00000008,
>> + FW_FATAL = 0x00000010,
>> + FW_PERF = 0x00000020,
>> + FW_CACHE_LOW = 0x00000100,
>> + FW_CACHE_MED = 0x00000200,
>> + FW_CACHE_HIGH = 0x00000400,
>> + FW_CACHE_ERROR = 0x00000800,
>> + FW_CACHE_FATAL = 0x00001000,
>> + FW_CACHE_PERF = 0x00002000,
>> + FW_PRINTK = 0x10000000,
>> + FW_FTRACE = 0x20000000,
>> +};
>> +
>> +#define DRV_LOG (VIDC_ERR | VIDC_PRINTK)
>> +#define DRV_LOGSHIFT (0)
>> +#define DRV_LOGMASK (0x0FFFFFFF)
>> +
>> +#define FW_LOG (FW_ERROR | FW_FATAL | FW_PRINTK)
>> +#define FW_LOGSHIFT (0)
>> +#define FW_LOGMASK (0x0FFFFFFF)
>> +
>> +#define dprintk_inst(__level, __level_str, inst, __fmt, ...) \
>> + do { \
>> + if (inst && (msm_vidc_debug & (__level))) { \
>> + pr_info(VIDC_DBG_TAG_INST __fmt, \
>> + __level_str, \
>> + inst->debug_str, \
>> + ##__VA_ARGS__); \
>> + } \
>> + } while (0)
>> +
>> +#define i_vpr_e(inst, __fmt, ...) dprintk_inst(VIDC_ERR, "err ", inst, __fmt, ##__VA_ARGS__)
>> +#define i_vpr_i(inst, __fmt, ...) dprintk_inst(VIDC_HIGH, "high", inst, __fmt, ##__VA_ARGS__)
>> +#define i_vpr_h(inst, __fmt, ...) dprintk_inst(VIDC_HIGH, "high", inst, __fmt, ##__VA_ARGS__)
>> +#define i_vpr_l(inst, __fmt, ...) dprintk_inst(VIDC_LOW, "low ", inst, __fmt, ##__VA_ARGS__)
>> +#define i_vpr_p(inst, __fmt, ...) dprintk_inst(VIDC_PERF, "perf", inst, __fmt, ##__VA_ARGS__)
>> +#define i_vpr_t(inst, __fmt, ...) dprintk_inst(VIDC_PKT, "pkt ", inst, __fmt, ##__VA_ARGS__)
>> +#define i_vpr_b(inst, __fmt, ...) dprintk_inst(VIDC_BUS, "bus ", inst, __fmt, ##__VA_ARGS__)
>
> NAK for entire interface. Please use standard debugging functions, not
> pr_info for everything.
>
> dev_dbg, dev_info, dev_warn, dev_err. Only these.
>
>
>> +#define i_vpr_s(inst, __fmt, ...) dprintk_inst(VIDC_STAT, "stat", inst, __fmt, ##__VA_ARGS__)
>> +
>> +#define i_vpr_hp(inst, __fmt, ...) \
>> + dprintk_inst(VIDC_HIGH | VIDC_PERF, "high", inst, __fmt, ##__VA_ARGS__)
>> +#define i_vpr_hs(inst, __fmt, ...) \
>> + dprintk_inst(VIDC_HIGH | VIDC_STAT, "stat", inst, __fmt, ##__VA_ARGS__)
>> +> +#define dprintk_core(__level, __level_str, __fmt, ...) \
>
> NAK
>
>> + do { \
>> + if (msm_vidc_debug & (__level)) { \
>> + pr_info(VIDC_DBG_TAG_CORE __fmt, \
>> + __level_str, \
>> + DEFAULT_SID, \
>> + "codec", \
>> + ##__VA_ARGS__); \
>> + } \
>> + } while (0)
>> +
>> +#define d_vpr_e(__fmt, ...) dprintk_core(VIDC_ERR, "err ", __fmt, ##__VA_ARGS__)
>> +#define d_vpr_h(__fmt, ...) dprintk_core(VIDC_HIGH, "high", __fmt, ##__VA_ARGS__)
>> +#define d_vpr_l(__fmt, ...) dprintk_core(VIDC_LOW, "low ", __fmt, ##__VA_ARGS__)
>> +#define d_vpr_p(__fmt, ...) dprintk_core(VIDC_PERF, "perf", __fmt, ##__VA_ARGS__)
>> +#define d_vpr_t(__fmt, ...) dprintk_core(VIDC_PKT, "pkt ", __fmt, ##__VA_ARGS__)
>> +#define d_vpr_b(__fmt, ...) dprintk_core(VIDC_BUS, "bus ", __fmt, ##__VA_ARGS__)
>> +#define d_vpr_s(__fmt, ...) dprintk_core(VIDC_STAT, "stat", __fmt, ##__VA_ARGS__)
>> +#define d_vpr_hs(__fmt, ...) \
>> + dprintk_core(VIDC_HIGH | VIDC_STAT, "high", __fmt, ##__VA_ARGS__)
>> +
>> +#define dprintk_ratelimit(__level, __level_str, __fmt, ...) \
>> + do { \
>> + if (msm_vidc_check_ratelimit()) { \
>> + dprintk_core(__level, __level_str, __fmt, ##__VA_ARGS__); \
>> + } \
>> + } while (0)
>> +
>> +#define dprintk_firmware(__level, __fmt, ...) \
>> + do { \
>> + if ((msm_fw_debug & (__level)) & FW_PRINTK) { \
>> + pr_info(FW_DBG_TAG __fmt, \
>> + "fw", \
>> + ##__VA_ARGS__); \
>> + } \
>> + } while (0)
>> +
>> +enum msm_vidc_debugfs_event {
>> + MSM_VIDC_DEBUGFS_EVENT_ETB,
>> + MSM_VIDC_DEBUGFS_EVENT_EBD,
>> + MSM_VIDC_DEBUGFS_EVENT_FTB,
>> + MSM_VIDC_DEBUGFS_EVENT_FBD,
>> +};
>> +
>> +enum msm_vidc_bug_on_error {
>> + MSM_VIDC_BUG_ON_FATAL = BIT(0),
>> + MSM_VIDC_BUG_ON_NOC = BIT(1),
>> + MSM_VIDC_BUG_ON_WD_TIMEOUT = BIT(2),
>> +};
>> +
>> +struct dentry *msm_vidc_debugfs_init_drv(void);
>> +struct dentry *msm_vidc_debugfs_init_core(struct msm_vidc_core *core);
>> +struct dentry *msm_vidc_debugfs_init_inst(struct msm_vidc_inst *inst,
>> + struct dentry *parent);
>> +void msm_vidc_debugfs_deinit_inst(struct msm_vidc_inst *inst);
>> +void msm_vidc_debugfs_update(struct msm_vidc_inst *inst,
>> + enum msm_vidc_debugfs_event e);
>> +int msm_vidc_check_ratelimit(void);
>> +
>> +static inline bool is_stats_enabled(void)
>> +{
>> + return !!(msm_vidc_debug & VIDC_STAT);
>
> ...
>
>> +struct dentry *msm_vidc_debugfs_init_core(struct msm_vidc_core *core)
>> +{
>> + struct dentry *dir = NULL;
>> + char debugfs_name[MAX_DEBUGFS_NAME];
>> + struct dentry *parent;
>> +
>> + if (!core->debugfs_parent) {
>> + d_vpr_e("%s: invalid params\n", __func__);
>> + goto failed_create_dir;
>> + }
>> + parent = core->debugfs_parent;
>> +
>> + snprintf(debugfs_name, MAX_DEBUGFS_NAME, "core");
>> + dir = debugfs_create_dir(debugfs_name, parent);
>> + if (IS_ERR_OR_NULL(dir)) {
>> + dir = NULL;
>> + d_vpr_e("Failed to create debugfs for msm_vidc\n");
>> + goto failed_create_dir;
>> + }
>> + if (!debugfs_create_file("info", 0444, dir, core, &core_info_fops)) {
>> + d_vpr_e("debugfs_create_file: fail\n");
>> + goto failed_create_dir;
>> + }
>> +
>> + if (!debugfs_create_file("stats_delay_ms", 0644, dir, core, &stats_delay_fops)) {
>> + d_vpr_e("debugfs_create_file: fail\n");
>
>
> What is this entire debugfs supposed to provide?
will remove this whole file in next version as part of custom debug
wrappers removal.
Thanks,
Dikshita
>
>
>
> Best regards,
> Krzysztof
>
Powered by blists - more mailing lists