linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 0/5] dyndbg add exclusive class support
@ 2022-03-11  4:47 Jim Cromie
  2022-03-11  4:47 ` [PATCH 1/5] dyndbg: fix static_branch manipulation Jim Cromie
                   ` (4 more replies)
  0 siblings, 5 replies; 11+ messages in thread
From: Jim Cromie @ 2022-03-11  4:47 UTC (permalink / raw)
  To: jbaron, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx, Jim Cromie

Hi Greg, Jason,

Please consider these for char/misc or linux-next/soon/mumble.

This patchset adds exclusive class support to dyndbg, allowing it to
directly represent drm's debug_category.

It is the dyndbg half of:
https://lore.kernel.org/lkml/20220217034829.64395-1-jim.cromie@gmail.com/

The DRM half of that patchset uses this support to reimplement
drm.debug on dyndbg, and uses its callsite patching to avoid the
runtime checking done by drm_debug_enabled().

Background:

In the past, various extensions to dyndbg/pr_debug have been proposed,
none seemed to gain any consensus.  This list is certainly incomplete.

https://lore.kernel.org/lkml/20200609104604.1594-1-stanimir.varbanov@linaro.org/

pr_levels was discussed around this patchset, in revs 3,4,5
This search helps:
	s:venus: s:dynamic f:stanimir.varbanov@linaro.org

IMO, pr_levels suffers from implied meaning between the levels: 2 > 1.
In contrast, DRM has logically disjoint categories, and is implemented
in an enum (despite its flag/mask values, a micro-optimization).

https://lore.kernel.org/lkml/20210813151734.1236324-1-jim.cromie@gmail.com/

That patchset used pr_debug in DRM, and prepended "drm:kms:" etc to
the format strings so each category was selectable; "format drm:kms:"
in the query.  This worked, but it made the format config-dependent,
and was hard to explain without undue "artifact".


So this patchset adds .class_id field (4-bits) to dynamic-debug
callsites, and 'class N' query/command support to select upon it.
Existing callsites and queries get .class_id=15, so 0-14 are available
for use by the client (DRM wants 0-10).

The DRM patchset then:
. renumbers drm_debug_category to fit into the 4-bit .class_id
  the new enumerations *are* the bit-positions in drm/parameters/debug.
. adapts the category-macro layer to use _CLS macros, mapping categories.
. adds macro layer under the category-macro layer
  which wraps drm_*dbg inside a dyndbg Factory macro
. uses DEFINE_DYNAMIC_DEBUG_CLASSBITS to tie to __drm_debug
  callbacks ref the var, so drm_debug_enabled(cat) just works.

Jim Cromie (5):

dyndbg: fix static_branch manipulation
  fixes a latent bug, before a 2nd "enable" flag exposes it.

dyndbg: add class_id field and query support
  will allow (with that drm patchset):
  #> # turn on DRM_ATOMIC in amdgpu
  #> echo module amdgpu class 4 +p > /proc/dynamic_debug/control
  #> # turn on DRM_CORE in drm
  #> echo module drm class 0 +p > /proc/dynamic_debug/control

dyndbg: add DEFINE_DYNAMIC_DEBUG_CLASSBITS macro
  adds macro & callbacks to support drm.debug bitmap
  #> echo 4 > /sys/module/drm/parameters/debug

dyndbg: drop EXPORTed dynamic_debug_exec_queries
  unused yet, obsoleted by 2,3

dyndbg: show both old and new in change-info
  minor debug improvement


 .../admin-guide/dynamic-debug-howto.rst       |   7 +
 include/linux/dynamic_debug.h                 | 111 ++++++++++---
 lib/dynamic_debug.c                           | 150 ++++++++++++++----
 3 files changed, 213 insertions(+), 55 deletions(-)

-- 
2.35.1


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH 1/5] dyndbg: fix static_branch manipulation
  2022-03-11  4:47 [PATCH 0/5] dyndbg add exclusive class support Jim Cromie
@ 2022-03-11  4:47 ` Jim Cromie
  2022-03-11 18:03   ` Jason Baron
  2022-03-11  4:47 ` [PATCH 2/5] dyndbg: add class_id field and query support Jim Cromie
                   ` (3 subsequent siblings)
  4 siblings, 1 reply; 11+ messages in thread
From: Jim Cromie @ 2022-03-11  4:47 UTC (permalink / raw)
  To: jbaron, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx, Jim Cromie,
	vincent.whitchurch

In https://lore.kernel.org/lkml/20211209150910.GA23668@axis.com/

Vincent's patch commented on, and worked around, a bug toggling
static_branch's, when a 2nd PRINTK-ish flag was added.  The bug
results in a premature static_branch_disable when the 1st of 2 flags
was disabled.

The cited commit computed newflags, but then in the JUMP_LABEL block,
failed to use that result, instead using just one of the terms in it.
Using newflags instead made the code work properly.

This is Vincents test-case, reduced.  It needs the 2nd flag to work
properly, but it's explanatory here.

pt_test() {
    echo 5 > /sys/module/dynamic_debug/verbose

    site="module tcp" # just one callsite
    echo " $site =_ " > /proc/dynamic_debug/control # clear it

    # A B ~A ~B
    for flg in +T +p "-T #broke here" -p; do
	echo " $site $flg " > /proc/dynamic_debug/control
    done;

    # A B ~B ~A
    for flg in +T +p "-p #broke here" -T; do
	echo " $site $flg " > /proc/dynamic_debug/control
    done
}
pt_test

Fixes: 84da83a6ffc0 dyndbg: combine flags & mask into a struct, simplify with it
CC: vincent.whitchurch@axis.com
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>

--
.drop @stable, no exposed bug.
---
 lib/dynamic_debug.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index dd7f56af9aed..a56c1286ffa4 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -211,10 +211,11 @@ static int ddebug_change(const struct ddebug_query *query,
 				continue;
 #ifdef CONFIG_JUMP_LABEL
 			if (dp->flags & _DPRINTK_FLAGS_PRINT) {
-				if (!(modifiers->flags & _DPRINTK_FLAGS_PRINT))
+				if (!(newflags & _DPRINTK_FLAGS_PRINT))
 					static_branch_disable(&dp->key.dd_key_true);
-			} else if (modifiers->flags & _DPRINTK_FLAGS_PRINT)
+			} else if (newflags & _DPRINTK_FLAGS_PRINT) {
 				static_branch_enable(&dp->key.dd_key_true);
+			}
 #endif
 			dp->flags = newflags;
 			v4pr_info("changed %s:%d [%s]%s =%s\n",
-- 
2.35.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH 2/5] dyndbg: add class_id field and query support
  2022-03-11  4:47 [PATCH 0/5] dyndbg add exclusive class support Jim Cromie
  2022-03-11  4:47 ` [PATCH 1/5] dyndbg: fix static_branch manipulation Jim Cromie
@ 2022-03-11  4:47 ` Jim Cromie
  2022-03-11 19:06   ` Jason Baron
  2022-03-11  4:47 ` [PATCH 3/5] dyndbg: add DEFINE_DYNAMIC_DEBUG_CLASSBITS macro Jim Cromie
                   ` (2 subsequent siblings)
  4 siblings, 1 reply; 11+ messages in thread
From: Jim Cromie @ 2022-03-11  4:47 UTC (permalink / raw)
  To: jbaron, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx, Jim Cromie

DRM defines/uses 10 enum drm_debug_category's to create exclusive
classes of debug messages.  To support this directly in dynamic-debug,
add the following:

- struct _ddebug.class_id:4 - 4 bits is enough
- define _DPRINTK_SITE_UNCLASSED 15 - see below

and the query support:
- struct _ddebug_query.class_id
- ddebug_parse_query	- looks for "class" keyword, then calls..
- parse_class		- accepts uint: 0-15, sets query.class_id and marker
- vpr_info_dq		- displays new field
- ddebug_proc_show	- append column with "cls:%d" if !UNCLASSED

With the patch, one can enable current/unclassed callsites by:

  #> echo drm class 15 +p > /proc/dynamic_debug/control

parse_class() accepts 0 .. _DPRINTK_SITE_UNCLASSED, which allows the
>control interface to explicitly manipulate unclassed callsites.

After parsing keywords, ddebug_parse_query() sets .class_id=15, iff it
wasnt explicitly set.  This allows future classed/categorized
callsites to be untouched by legacy (class unaware) queries.

DEFINE_DYNAMIC_DEBUG_METADATA gets _CLS(cls,) suffix and arg, and
initializes the new .class_id=cls field.  The old name gets the default.

Then, these _CLS(cls,...) modifications are repeated up through the
stack of *dynamic_func_call* macros that use the METADATA initializer,
so as to actually supply the category into it.

NOTES:

_DPRINTK_SITE_UNCLASSED: this symbol is used to initialize all
existing/unclassed pr-debug callsites.  Normally, the default would be
zero, but DRM_UT_CORE "uses" that value, in the sense that 0 is
exposed as a bit position in drm.debug.  Using 15 allows identity
mapping from category to class, avoiding fiddly offsets.

Default .class_id = 15 means that ``echo +p > control`` no longer
toggles ALL the callsites, only the unclassed ones.  This was only
useful for static-branch toggle load testing anyway.

RFC:

The new _CLS macro flavor gets a warning from DRM/dri-devel's CI,
but not from checkpatch (on this subject).

a8f6c71f283e dyndbg: add class_id field and query support
-:141: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'id' - possible side-effects?
+#define __dynamic_func_call_cls(id, cls, fmt, func, ...) do {	\
+	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);	\
+	if (DYNAMIC_DEBUG_BRANCH(id))				\
+		func(&id, ##__VA_ARGS__);			\
 } while (0)

I couldn't fix it with a ``typeof(id) _id = id`` construct.  I haven't
seen the warning myself, on the _CLS extended macro, nor the original.

CC: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
 .../admin-guide/dynamic-debug-howto.rst       |  7 +++
 include/linux/dynamic_debug.h                 | 54 ++++++++++++++-----
 lib/dynamic_debug.c                           | 48 ++++++++++++++---
 3 files changed, 88 insertions(+), 21 deletions(-)

diff --git a/Documentation/admin-guide/dynamic-debug-howto.rst b/Documentation/admin-guide/dynamic-debug-howto.rst
index a89cfa083155..8ef8d7dcd140 100644
--- a/Documentation/admin-guide/dynamic-debug-howto.rst
+++ b/Documentation/admin-guide/dynamic-debug-howto.rst
@@ -35,6 +35,7 @@ Dynamic debug has even more useful features:
    - line number (including ranges of line numbers)
    - module name
    - format string
+   - class number:0-15
 
  * Provides a debugfs control file: ``<debugfs>/dynamic_debug/control``
    which can be read to display the complete list of known debug
@@ -143,6 +144,7 @@ against.  Possible keywords are:::
 		 'module' string |
 		 'format' string |
 		 'line' line-range
+		 'class' integer:[0-15]
 
   line-range ::= lineno |
 		 '-'lineno |
@@ -217,6 +219,11 @@ line
 	line -1605          // the 1605 lines from line 1 to line 1605
 	line 1600-          // all lines from line 1600 to the end of the file
 
+class
+    This expects a single integer in range: 0-15.
+    15 is used/reserved for existing/unclassed callsites,
+    and is defaulted in unless specified to >control
+
 The flags specification comprises a change operation followed
 by one or more flag characters.  The change operation is one
 of the characters::
diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h
index dce631e678dd..d4b48f3cc6e8 100644
--- a/include/linux/dynamic_debug.h
+++ b/include/linux/dynamic_debug.h
@@ -6,6 +6,8 @@
 #include <linux/jump_label.h>
 #endif
 
+#include <linux/build_bug.h>
+
 /*
  * An instance of this structure is created in a special
  * ELF section at every dynamic debug callsite.  At runtime,
@@ -21,6 +23,9 @@ struct _ddebug {
 	const char *filename;
 	const char *format;
 	unsigned int lineno:18;
+#define CLS_BITS 4
+	unsigned int class_id:CLS_BITS;
+#define _DPRINTK_SITE_UNCLASSED		((1 << CLS_BITS) - 1)
 	/*
 	 * The flags field controls the behaviour at the callsite.
 	 * The bits here are changed dynamically when the user
@@ -87,7 +92,7 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
 			 const struct ib_device *ibdev,
 			 const char *fmt, ...);
 
-#define DEFINE_DYNAMIC_DEBUG_METADATA(name, fmt)		\
+#define DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, cls, fmt)	\
 	static struct _ddebug  __aligned(8)			\
 	__section("__dyndbg") name = {				\
 		.modname = KBUILD_MODNAME,			\
@@ -96,8 +101,14 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
 		.format = (fmt),				\
 		.lineno = __LINE__,				\
 		.flags = _DPRINTK_FLAGS_DEFAULT,		\
+		.class_id = cls,				\
 		_DPRINTK_KEY_INIT				\
-	}
+	};							\
+	BUILD_BUG_ON_MSG(cls > _DPRINTK_SITE_UNCLASSED,		\
+			 "classid value overflow")
+
+#define DEFINE_DYNAMIC_DEBUG_METADATA(name, fmt)		\
+	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, _DPRINTK_SITE_UNCLASSED, fmt)
 
 #ifdef CONFIG_JUMP_LABEL
 
@@ -128,18 +139,26 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
 
 #endif /* CONFIG_JUMP_LABEL */
 
-#define __dynamic_func_call(id, fmt, func, ...) do {	\
-	DEFINE_DYNAMIC_DEBUG_METADATA(id, fmt);		\
-	if (DYNAMIC_DEBUG_BRANCH(id))			\
-		func(&id, ##__VA_ARGS__);		\
+#define __dynamic_func_call_cls(id, cls, fmt, func, ...) do {	\
+	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);	\
+	if (DYNAMIC_DEBUG_BRANCH(id))				\
+		func(&id, ##__VA_ARGS__);			\
 } while (0)
 
-#define __dynamic_func_call_no_desc(id, fmt, func, ...) do {	\
-	DEFINE_DYNAMIC_DEBUG_METADATA(id, fmt);			\
-	if (DYNAMIC_DEBUG_BRANCH(id))				\
-		func(__VA_ARGS__);				\
+#define __dynamic_func_call_no_desc_cls(id, cls, fmt, func, ...) do {	\
+	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);		\
+	if (DYNAMIC_DEBUG_BRANCH(id))					\
+		func(__VA_ARGS__);					\
 } while (0)
 
+#define __dynamic_func_call(id, fmt, func, ...)				\
+	__dynamic_func_call_cls(id, _DPRINTK_SITE_UNCLASSED,		\
+				fmt, func, ##__VA_ARGS__)
+
+#define __dynamic_func_call_no_desc(id, fmt, func, ...)			\
+	__dynamic_func_call_no_desc_cls(id, _DPRINTK_SITE_UNCLASSED,	\
+					fmt, func, ##__VA_ARGS__)
+
 /*
  * "Factory macro" for generating a call to func, guarded by a
  * DYNAMIC_DEBUG_BRANCH. The dynamic debug descriptor will be
@@ -148,15 +167,24 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
  * the varargs. Note that fmt is repeated in invocations of this
  * macro.
  */
+#define _dynamic_func_call_cls(cls, fmt, func, ...)			\
+	__dynamic_func_call_cls(__UNIQUE_ID(ddebug), cls, fmt, func, ##__VA_ARGS__)
 #define _dynamic_func_call(fmt, func, ...)				\
-	__dynamic_func_call(__UNIQUE_ID(ddebug), fmt, func, ##__VA_ARGS__)
+	_dynamic_func_call_cls(_DPRINTK_SITE_UNCLASSED, fmt, func, ##__VA_ARGS__)
+
 /*
  * A variant that does the same, except that the descriptor is not
  * passed as the first argument to the function; it is only called
  * with precisely the macro's varargs.
  */
-#define _dynamic_func_call_no_desc(fmt, func, ...)	\
-	__dynamic_func_call_no_desc(__UNIQUE_ID(ddebug), fmt, func, ##__VA_ARGS__)
+#define _dynamic_func_call_no_desc_cls(fmt, cat, func, ...)		\
+	__dynamic_func_call_no_desc_cls(__UNIQUE_ID(ddebug), cat,	\
+					fmt, func, ##__VA_ARGS__)
+
+#define _dynamic_func_call_no_desc(fmt, func, ...)			\
+	__dynamic_func_call_no_desc_cls(__UNIQUE_ID(ddebug),		\
+					_DPRINTK_SITE_UNCLASSED,	\
+					fmt, func, ##__VA_ARGS__)
 
 #define dynamic_pr_debug(fmt, ...)				\
 	_dynamic_func_call(fmt,	__dynamic_pr_debug,		\
diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index a56c1286ffa4..ee2129becacc 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -55,6 +55,8 @@ struct ddebug_query {
 	const char *function;
 	const char *format;
 	unsigned int first_lineno, last_lineno;
+	unsigned int class_id;
+	unsigned int class_marked:1;
 };
 
 struct ddebug_iter {
@@ -134,13 +136,13 @@ static void vpr_info_dq(const struct ddebug_query *query, const char *msg)
 			fmtlen--;
 	}
 
-	v3pr_info("%s: func=\"%s\" file=\"%s\" module=\"%s\" format=\"%.*s\" lineno=%u-%u\n",
-		 msg,
-		 query->function ?: "",
-		 query->filename ?: "",
-		 query->module ?: "",
-		 fmtlen, query->format ?: "",
-		 query->first_lineno, query->last_lineno);
+	v3pr_info("%s: func=\"%s\" file=\"%s\" module=\"%s\" format=\"%.*s\" lineno=%u-%u class=%u\n",
+		  msg,
+		  query->function ?: "",
+		  query->filename ?: "",
+		  query->module ?: "",
+		  fmtlen, query->format ?: "",
+		  query->first_lineno, query->last_lineno, query->class_id);
 }
 
 /*
@@ -170,6 +172,10 @@ static int ddebug_change(const struct ddebug_query *query,
 		for (i = 0; i < dt->num_ddebugs; i++) {
 			struct _ddebug *dp = &dt->ddebugs[i];
 
+			/* match against the class_id, either given or default */
+			if (query->class_id != dp->class_id)
+				continue;
+
 			/* match against the source filename */
 			if (query->filename &&
 			    !match_wildcard(query->filename, dp->filename) &&
@@ -308,6 +314,21 @@ static inline int parse_lineno(const char *str, unsigned int *val)
 	return 0;
 }
 
+static inline int parse_class(struct ddebug_query *query, const char *str)
+{
+	int rc;
+	unsigned int val;
+
+	rc = kstrtouint(str, 10, &val);
+	if (rc < 0 || val > _DPRINTK_SITE_UNCLASSED) {
+		pr_err("expecting class:[0-%d], not %s\n", _DPRINTK_SITE_UNCLASSED, str);
+		return -EINVAL;
+	}
+	query->class_id = val;
+	query->class_marked = 1;
+	return 0;
+}
+
 static int parse_linerange(struct ddebug_query *query, const char *first)
 {
 	char *last = strchr(first, '-');
@@ -421,6 +442,9 @@ static int ddebug_parse_query(char *words[], int nwords,
 		} else if (!strcmp(keyword, "line")) {
 			if (parse_linerange(query, arg))
 				return -EINVAL;
+		} else if (!strcmp(keyword, "class")) {
+			if (parse_class(query, arg))
+				return -EINVAL;
 		} else {
 			pr_err("unknown keyword \"%s\"\n", keyword);
 			return -EINVAL;
@@ -428,6 +452,10 @@ static int ddebug_parse_query(char *words[], int nwords,
 		if (rc)
 			return rc;
 	}
+	/* post-validate the query, set default */
+	if (!query->class_marked)
+		query->class_id = _DPRINTK_SITE_UNCLASSED;
+
 	vpr_info_dq(query, "parsed");
 	return 0;
 }
@@ -900,7 +928,11 @@ static int ddebug_proc_show(struct seq_file *m, void *p)
 		   iter->table->mod_name, dp->function,
 		   ddebug_describe_flags(dp->flags, &flags));
 	seq_escape(m, dp->format, "\t\r\n\"");
-	seq_puts(m, "\"\n");
+	seq_puts(m, "\"");
+
+	if (dp->class_id != _DPRINTK_SITE_UNCLASSED)
+		seq_printf(m, " cls:%u", dp->class_id);
+	seq_puts(m, "\n");
 
 	return 0;
 }
-- 
2.35.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH 3/5] dyndbg: add DEFINE_DYNAMIC_DEBUG_CLASSBITS macro
  2022-03-11  4:47 [PATCH 0/5] dyndbg add exclusive class support Jim Cromie
  2022-03-11  4:47 ` [PATCH 1/5] dyndbg: fix static_branch manipulation Jim Cromie
  2022-03-11  4:47 ` [PATCH 2/5] dyndbg: add class_id field and query support Jim Cromie
@ 2022-03-11  4:47 ` Jim Cromie
  2022-03-11  4:47 ` [PATCH 4/5] dyndbg: drop EXPORTed dynamic_debug_exec_queries Jim Cromie
  2022-03-11  4:47 ` [PATCH 5/5] dyndbg: show both old and new in change-info Jim Cromie
  4 siblings, 0 replies; 11+ messages in thread
From: Jim Cromie @ 2022-03-11  4:47 UTC (permalink / raw)
  To: jbaron, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx, Jim Cromie

DEFINE_DYNAMIC_DEBUG_CLASSBITS(fsname, var, bitmap_desc, classes..)
allows users to create a drm.debug style (bitmap) sysfs interface, to
control sets of pr_debug's according to their .class_id's

This wraps existing "class" keyword and behavior:

   bash-5.1# echo <<END  > /proc/dynamic_debug/control
   > module drm class 0 +p
   > module drm class 2 +p
   > END

With the macro used (in a client), this is basically equivalent:

   # but this also clears other flags, if theyre set
   echo 0x05 > /sys/module/drm/parameters/debug

To use:

DEFINE_DYNAMIC_DEBUG_CLASSBITS(debug, __drm_debug, "pmfl",
	"drm.debug - bits => categories:",
	/* vector of uint:4 symbols, ala enum drm_debug_category, 15 is EOL */
	DRM_UT_CORE,
	DRM_UT_DRIVER,
	DRM_UT_KMS ... );

The 3rd arg is a string with any of the dyndbg.flags [pmflt_]+
Full exposure of the flags here lets the module author:

- fully customize/take-over the decorations of enabled sites.
  generally leaving decorations to user is preferred.

- aim the debug-stream:
  now printk, later tracefs.
  using both together means more work (p or T, in practice)
  iface doesn't care about new flags added later

- declare 2 separate sysfs-knobs, one each for p, T, if desired.

- decorations are per callsite,
  shared across sysfs-knobs for any controlled classes

To support the macro, the patch adds:

 - int param_set_dyndbg_classbits()
 - int param_get_dyndbg_classbits()
 - struct kernel_param_ops param_ops_dyndbg_classbits

Following the model of kernel/params.c STANDARD_PARAM_DEFS, these are
non-static and exported.

get/set use an augmented kernel_param; the arg refs a new struct
dyndbg_bitmap_param containing:

A- the vector of classes (drm.debug "categories") being controlled

This in-line vector of constants (uint [0-14]) specifies a sequence of
controlling bits (by position, starting at 0) with the values naming
the class_id's mapped to that bit.

A value of _DPRINTK_SITE_UNCLASSED terminates the vector processing by
param_set_dyndbg_classbits(), and is appended by the macro to insure a
defined termination after max 15 classes are applied.

Technically, the vector is a flex-array, but its size is practically
limited to max 15 in length (repeats are pointless).

B- a pointer to the user module's ulong holding the bits/state.

By accessing client's bit-state, we coordinate with existing code
that still uses drm_debug_enabled(), so they work unchanged.
The change to ulong allows use of BIT() etc.

NOTES:

_DPRINTK_SITE_UNCLASSED = 15, ie 2**CLS_BITS-1, deserves special
mention; it already marks all existing pr-debug callsites, only the
new drm.debug callsites are initialized to other (DRM_UT_*) values.

_DPRINTK_SITE_UNCLASSED is used in param_set_dyndbg_classbits() to
limit the range of bits that are processed to what fits in uint:4.
It also terminates the class-id list passed into the macro, so dont
use it halfway through your list of classes-to-control.

param_set_dyndbg_classbits() compares new vs old bits, and only
updates each class on changes.  This maximally preserves the
underlying state, which may have been customized at some point via
`echo $cmd >control`.  So if users really want to know that all
prdbgs are set precisely, they must pre-clear then set.

Identity mapping in (A) is encouraged.  Your vector should exclude 15,
if used, it terminates the list prematurely; any following bit
mappings will be ignored (it is the default/non category).

The whole (A) vector/list passed to the macro is:

- not strictly needed: 0-N bits are scanned, only N is needed in the
  macro interface to do this, not the whole list.

- 0-N list allows juggling the bit->class map
  Identity map is preferred.
  15 terminates list if used. (macro impl does this)

That said, (A) is self-documenting; the explicit list is greppable,
'DRM_UT_*' provides lots of clues.  Further, it could be upgraded,
something like:

  _pick_sym_(DRM_UT_CORE, "mumble something useful about CORE debug")

_pick_sym_(a,b) a		// gives us what we need here
_pick_help_(a,b) #a " : " b	// mod-info fodder

Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
 include/linux/dynamic_debug.h | 50 +++++++++++++++++++++++
 lib/dynamic_debug.c           | 77 +++++++++++++++++++++++++++++++++++
 2 files changed, 127 insertions(+)

diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h
index d4b48f3cc6e8..e83c4e36ad29 100644
--- a/include/linux/dynamic_debug.h
+++ b/include/linux/dynamic_debug.h
@@ -209,6 +209,10 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
 				   KERN_DEBUG, prefix_str, prefix_type,	\
 				   rowsize, groupsize, buf, len, ascii)
 
+struct kernel_param;
+int param_set_dyndbg_classbits(const char *instr, const struct kernel_param *kp);
+int param_get_dyndbg_classbits(char *buffer, const struct kernel_param *kp);
+
 #else /* !CONFIG_DYNAMIC_DEBUG_CORE */
 
 #include <linux/string.h>
@@ -255,6 +259,52 @@ static inline int dynamic_debug_exec_queries(const char *query, const char *modn
 	return 0;
 }
 
+struct kernel_param;
+static inline int param_set_dyndbg_classbits(const char *instr, const struct kernel_param *kp)
+{ return 0; }
+static inline int param_get_dyndbg_classbits(char *buffer, const struct kernel_param *kp)
+{ return 0; }
+
 #endif /* !CONFIG_DYNAMIC_DEBUG_CORE */
 
+#define FLAGS_LEN 8
+struct dyndbg_classbits_param {
+	unsigned long *bits;		/* ref to shared state */
+	const char flags[FLAGS_LEN];	/* toggle these flags on bit-changes */
+	const int classes[];		/* indexed by bitpos */
+};
+
+#if defined(CONFIG_DYNAMIC_DEBUG) || defined(CONFIG_DYNAMIC_DEBUG_CORE)
+/**
+ * DEFINE_DYNAMIC_DEBUG_CLASSBITS() - bitmap control of classed pr_debugs
+ * @sysname: sysfs-node name
+ * @_var:    C-identifier holding bit-vector (Bits 0-14 are usable)
+ * @_flgs:   string with dyndbg flags: 'p' and/or 'T', and maybe "fmlt" also.
+ * @desc:    string summarizing the controls provided
+ * @classes: vector of callsite.class_id's (uint:4, 15 is reserved)
+ *
+ * This macro implements a DRM.debug API style bitmap, mapping bits
+ * 0-14 to classes of prdbg's, as initialized in their .class_id fields.
+ * @_flgs chooses the debug recipient; p - syslog, T - tracefs, and
+ * can include log decorations; m - module, f - function, l - line_num
+ */
+#define DEFINE_DYNAMIC_DEBUG_CLASSBITS(fsname, _var, _flgs, desc, ...)	\
+	MODULE_PARM_DESC(fsname, desc);					\
+	static struct dyndbg_classbits_param ddcats_##_var = {		\
+		.bits = &(_var),					\
+		.flags = _flgs,						\
+		.classes = { __VA_ARGS__, _DPRINTK_SITE_UNCLASSED }	\
+	};								\
+	module_param_cb(fsname, &param_ops_dyndbg_classbits,		\
+			&ddcats_##_var, 0644)
+
+extern const struct kernel_param_ops param_ops_dyndbg_classbits;
+
+#else /* no dyndbg configured, throw error on macro use */
+
+#define DEFINE_DYNAMIC_DEBUG_CLASSBITS(fsname, var, bitmap_desc, ...)	\
+	BUILD_BUG_ON_MSG(1, "CONFIG_DYNAMIC_DEBUG* needed to use this macro: " #fsname)
+
+#endif
+
 #endif
diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index ee2129becacc..704361af5b23 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -611,6 +611,83 @@ int dynamic_debug_exec_queries(const char *query, const char *modname)
 }
 EXPORT_SYMBOL_GPL(dynamic_debug_exec_queries);
 
+#ifdef CONFIG_MODULES
+#define KP_MOD_NAME kp->mod->name
+#else
+#define KP_MOD_NAME NULL /* wildcard */
+#endif
+#define FMT_QUERY_SIZE 128 /* typically need <40 */
+/**
+ * param_set_dyndbg_classbits - bits => categories >control setter
+ * @instr: string echo>d to sysfs
+ * @kp:    kp->arg has state: bits, map
+ *
+ * Enable/disable prdbgs by their "category", as specified in the
+ * DEFINE_DYNAMIC_DEBUG_BITGRPS.classbits argument.
+ *
+ * Returns: 0 or <0 if error.
+ */
+int param_set_dyndbg_classbits(const char *instr, const struct kernel_param *kp)
+{
+	unsigned long inbits;
+	int rc, i, matches = 0, totct = 0;
+	char query[FMT_QUERY_SIZE];
+	const struct dyndbg_classbits_param *dcp = kp->arg;
+
+	if (!dcp) {
+		pr_err("set_dyndbg_classbits: no bits=>queries map\n");
+		return -EINVAL;
+	}
+	rc = kstrtoul(instr, 0, &inbits);
+	if (rc) {
+		pr_err("set_dyndbg_classbits: expecting unsigned int\n");
+		return rc;
+	}
+	vpr_info("set_dyndbg_classbits: new 0x%lx old 0x%lx\n", inbits, *dcp->bits);
+
+	for (i = 0; (i < _DPRINTK_SITE_UNCLASSED &&
+		     dcp->classes[i] < _DPRINTK_SITE_UNCLASSED); i++) {
+
+		if (test_bit(i, &inbits) == test_bit(i, dcp->bits))
+			continue;
+		snprintf(query, FMT_QUERY_SIZE, "class %d %cT",
+			 dcp->classes[i], test_bit(i, &inbits) ? '+' : '-');
+
+		matches = ddebug_exec_queries(query, KP_MOD_NAME);
+
+		v2pr_info("bit-%d: %d matches on class:%u\n", i,
+			  matches, dcp->classes[i]);
+		totct += matches;
+	}
+	*dcp->bits = inbits;
+	vpr_info("total matches: %d\n", totct);
+	return 0;
+}
+EXPORT_SYMBOL(param_set_dyndbg_classbits);
+
+/**
+ * param_get_dyndbg_classbits - classbits reader
+ * @buffer: string description of controlled bits -> classes
+ * @kp:     kp->arg has state: bits, map
+ *
+ * Reads last written bits, underlying prdbg state may have changed since.
+ * Returns: #chars written or <0 on error
+ */
+int param_get_dyndbg_classbits(char *buffer, const struct kernel_param *kp)
+{
+	const struct dyndbg_classbits_param *p = kp->arg;
+	unsigned long val = *p->bits;
+
+	return scnprintf(buffer, PAGE_SIZE, "0x%lx\n", val);
+}
+EXPORT_SYMBOL(param_get_dyndbg_classbits);
+
+const struct kernel_param_ops param_ops_dyndbg_classbits = {
+	.set = param_set_dyndbg_classbits,
+	.get = param_get_dyndbg_classbits,
+};
+EXPORT_SYMBOL(param_ops_dyndbg_classbits);
+
 #define PREFIX_SIZE 64
 
 static int remaining(int wrote)
-- 
2.35.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH 4/5] dyndbg: drop EXPORTed dynamic_debug_exec_queries
  2022-03-11  4:47 [PATCH 0/5] dyndbg add exclusive class support Jim Cromie
                   ` (2 preceding siblings ...)
  2022-03-11  4:47 ` [PATCH 3/5] dyndbg: add DEFINE_DYNAMIC_DEBUG_CLASSBITS macro Jim Cromie
@ 2022-03-11  4:47 ` Jim Cromie
  2022-03-11  4:47 ` [PATCH 5/5] dyndbg: show both old and new in change-info Jim Cromie
  4 siblings, 0 replies; 11+ messages in thread
From: Jim Cromie @ 2022-03-11  4:47 UTC (permalink / raw)
  To: jbaron, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx, Jim Cromie

This exported fn is effectively obsoleted by Commit:HEAD~2, so remove it.

Its intent was to allow drm.debug to use the exported function to
implement its drm.debug bitmap api using dynamic_debug.  Instead,
HEAD~2 implements the bitmap inside dyndbg, using the internal fn that
the export wraps.

Since there are no other expected users, and any prospects would
likely reuse the bitmap or a straightforward extension of it, we can
drop the exported function until its really needed.

This also drops the CONFIG_DYNAMIC_DEBUG=N stub-func, and its
pr_warn(), which I avoided in 2012, then added in 2020 :-/

Fixes: a2d375eda771 ("dyndbg: refine export, rename to dynamic_debug_exec_queries()")
Fixes: 4c0d77828d4f ("dyndbg: export ddebug_exec_queries")
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
 include/linux/dynamic_debug.h |  9 ---------
 lib/dynamic_debug.c           | 29 -----------------------------
 2 files changed, 38 deletions(-)

diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h
index e83c4e36ad29..664bb83778d2 100644
--- a/include/linux/dynamic_debug.h
+++ b/include/linux/dynamic_debug.h
@@ -60,9 +60,6 @@ struct _ddebug {
 
 #if defined(CONFIG_DYNAMIC_DEBUG_CORE)
 
-/* exported for module authors to exercise >control */
-int dynamic_debug_exec_queries(const char *query, const char *modname);
-
 int ddebug_add_module(struct _ddebug *tab, unsigned int n,
 				const char *modname);
 extern int ddebug_remove_module(const char *mod_name);
@@ -253,12 +250,6 @@ static inline int ddebug_dyndbg_module_param_cb(char *param, char *val,
 				rowsize, groupsize, buf, len, ascii);	\
 	} while (0)
 
-static inline int dynamic_debug_exec_queries(const char *query, const char *modname)
-{
-	pr_warn("kernel not built with CONFIG_DYNAMIC_DEBUG_CORE\n");
-	return 0;
-}
-
 struct kernel_param;
 static inline int param_set_dyndbg_classbits(const char *instr, const struct kernel_param *kp)
 { return 0; }
diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index 704361af5b23..b15a9c715e5b 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -582,35 +582,6 @@ static int ddebug_exec_queries(char *query, const char *modname)
 	return nfound;
 }
 
-/**
- * dynamic_debug_exec_queries - select and change dynamic-debug prints
- * @query: query-string described in admin-guide/dynamic-debug-howto
- * @modname: string containing module name, usually &module.mod_name
- *
- * This uses the >/proc/dynamic_debug/control reader, allowing module
- * authors to modify their dynamic-debug callsites. The modname is
- * canonically struct module.mod_name, but can also be null or a
- * module-wildcard, for example: "drm*".
- */
-int dynamic_debug_exec_queries(const char *query, const char *modname)
-{
-	int rc;
-	char *qry; /* writable copy of query */
-
-	if (!query) {
-		pr_err("non-null query/command string expected\n");
-		return -EINVAL;
-	}
-	qry = kstrndup(query, PAGE_SIZE, GFP_KERNEL);
-	if (!qry)
-		return -ENOMEM;
-
-	rc = ddebug_exec_queries(qry, modname);
-	kfree(qry);
-	return rc;
-}
-EXPORT_SYMBOL_GPL(dynamic_debug_exec_queries);
-
 #ifdef CONFIG_MODULES
 #define KP_MOD_NAME kp->mod->name
 #else
-- 
2.35.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH 5/5] dyndbg: show both old and new in change-info
  2022-03-11  4:47 [PATCH 0/5] dyndbg add exclusive class support Jim Cromie
                   ` (3 preceding siblings ...)
  2022-03-11  4:47 ` [PATCH 4/5] dyndbg: drop EXPORTed dynamic_debug_exec_queries Jim Cromie
@ 2022-03-11  4:47 ` Jim Cromie
  4 siblings, 0 replies; 11+ messages in thread
From: Jim Cromie @ 2022-03-11  4:47 UTC (permalink / raw)
  To: jbaron, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx, Jim Cromie

print old -> new flag values in the info("change") message.

Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
 lib/dynamic_debug.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index b15a9c715e5b..cceac8ebbacd 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -158,7 +158,7 @@ static int ddebug_change(const struct ddebug_query *query,
 	struct ddebug_table *dt;
 	unsigned int newflags;
 	unsigned int nfound = 0;
-	struct flagsbuf fbuf;
+	struct flagsbuf fbuf, nbuf;
 
 	/* search for matching ddebugs */
 	mutex_lock(&ddebug_lock);
@@ -223,11 +223,12 @@ static int ddebug_change(const struct ddebug_query *query,
 				static_branch_enable(&dp->key.dd_key_true);
 			}
 #endif
+			v4pr_info("changed %s:%d [%s]%s %s -> %s\n",
+				  trim_prefix(dp->filename), dp->lineno,
+				  dt->mod_name, dp->function,
+				  ddebug_describe_flags(dp->flags, &fbuf),
+				  ddebug_describe_flags(newflags, &nbuf));
 			dp->flags = newflags;
-			v4pr_info("changed %s:%d [%s]%s =%s\n",
-				 trim_prefix(dp->filename), dp->lineno,
-				 dt->mod_name, dp->function,
-				 ddebug_describe_flags(dp->flags, &fbuf));
 		}
 	}
 	mutex_unlock(&ddebug_lock);
-- 
2.35.1


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* Re: [PATCH 1/5] dyndbg: fix static_branch manipulation
  2022-03-11  4:47 ` [PATCH 1/5] dyndbg: fix static_branch manipulation Jim Cromie
@ 2022-03-11 18:03   ` Jason Baron
  0 siblings, 0 replies; 11+ messages in thread
From: Jason Baron @ 2022-03-11 18:03 UTC (permalink / raw)
  To: Jim Cromie, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx, vincent.whitchurch



On 3/10/22 23:47, Jim Cromie wrote:
> In https://urldefense.com/v3/__https://lore.kernel.org/lkml/20211209150910.GA23668@axis.com/__;!!GjvTz_vk!HGKKoni4RVdEBgv_V0zPSNSX428bpf02zkCy2WbeQkBdVtp1QJqGX-lJYlRDGg$ 
> 
> Vincent's patch commented on, and worked around, a bug toggling
> static_branch's, when a 2nd PRINTK-ish flag was added.  The bug
> results in a premature static_branch_disable when the 1st of 2 flags
> was disabled.
> 
> The cited commit computed newflags, but then in the JUMP_LABEL block,
> failed to use that result, instead using just one of the terms in it.
> Using newflags instead made the code work properly.
> 
> This is Vincents test-case, reduced.  It needs the 2nd flag to work
> properly, but it's explanatory here.
> 
> pt_test() {
>     echo 5 > /sys/module/dynamic_debug/verbose
> 
>     site="module tcp" # just one callsite
>     echo " $site =_ " > /proc/dynamic_debug/control # clear it
> 
>     # A B ~A ~B
>     for flg in +T +p "-T #broke here" -p; do
> 	echo " $site $flg " > /proc/dynamic_debug/control
>     done;
> 
>     # A B ~B ~A
>     for flg in +T +p "-p #broke here" -T; do
> 	echo " $site $flg " > /proc/dynamic_debug/control
>     done
> }
> pt_test
> 
> Fixes: 84da83a6ffc0 dyndbg: combine flags & mask into a struct, simplify with it
> CC: vincent.whitchurch@axis.com
> Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
> 
> --
> .drop @stable, no exposed bug.
> ---
>  lib/dynamic_debug.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
> index dd7f56af9aed..a56c1286ffa4 100644
> --- a/lib/dynamic_debug.c
> +++ b/lib/dynamic_debug.c
> @@ -211,10 +211,11 @@ static int ddebug_change(const struct ddebug_query *query,
>  				continue;
>  #ifdef CONFIG_JUMP_LABEL
>  			if (dp->flags & _DPRINTK_FLAGS_PRINT) {
> -				if (!(modifiers->flags & _DPRINTK_FLAGS_PRINT))
> +				if (!(newflags & _DPRINTK_FLAGS_PRINT))
>  					static_branch_disable(&dp->key.dd_key_true);
> -			} else if (modifiers->flags & _DPRINTK_FLAGS_PRINT)
> +			} else if (newflags & _DPRINTK_FLAGS_PRINT) {
>  				static_branch_enable(&dp->key.dd_key_true);
> +			}
>  #endif
>  			dp->flags = newflags;
>  			v4pr_info("changed %s:%d [%s]%s =%s\n",



Hi Jim,

If iiuc this is currently a bug but could be if we add a second 'print' bit
such as for printing to the tracing logs. That said I agree that using 'newflags'
here makes the code more straightforward/readable. So this one is fine with
me.

Acked-by: Jason Baron <jbaron@akamai.com>

Thanks,

-Jason

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 2/5] dyndbg: add class_id field and query support
  2022-03-11  4:47 ` [PATCH 2/5] dyndbg: add class_id field and query support Jim Cromie
@ 2022-03-11 19:06   ` Jason Baron
  2022-03-12  1:06     ` jim.cromie
  0 siblings, 1 reply; 11+ messages in thread
From: Jason Baron @ 2022-03-11 19:06 UTC (permalink / raw)
  To: Jim Cromie, gregkh, linux-kernel
  Cc: daniel.vetter, seanpaul, robdclark, linux, joe, dri-devel,
	amd-gfx, intel-gvt-dev, intel-gfx



On 3/10/22 23:47, Jim Cromie wrote:
> DRM defines/uses 10 enum drm_debug_category's to create exclusive
> classes of debug messages.  To support this directly in dynamic-debug,
> add the following:
> 
> - struct _ddebug.class_id:4 - 4 bits is enough
> - define _DPRINTK_SITE_UNCLASSED 15 - see below
> 
> and the query support:
> - struct _ddebug_query.class_id
> - ddebug_parse_query	- looks for "class" keyword, then calls..
> - parse_class		- accepts uint: 0-15, sets query.class_id and marker
> - vpr_info_dq		- displays new field
> - ddebug_proc_show	- append column with "cls:%d" if !UNCLASSED
> 
> With the patch, one can enable current/unclassed callsites by:
> 
>   #> echo drm class 15 +p > /proc/dynamic_debug/control
> 

To me, this is hard to read, what the heck is '15'? I have to go look it
up in the control file and it's not descriptive. I think that using
classes/categories makes sense but I'm wondering if it can be a bit more
user friendly? Perhaps, we can pass an array of strings that is indexed
by the class id to each pr_debug() site that wants to use class. So
something like:

enum levels {
	LOW,
	MEDIUM,
	HIGH
};

static const char * const level_to_strings[] = {
        [LOW] = "low",
        [MEDIUM] = "medium",
        [HIGH] = "high",
};

And then you'd have a wrapper macros in your driver:

#define module_foo_pr_debug_class(level, fmt, args...)
	pr_debug_class(level, level_to_strings, fmt, args);

Such that call sites look like:

module_foo_pr_debug_class(LOW, fmt, args...);

Such that you're not always passing the strings array around. Now, this
does mean another pointer for struct _ddebug and most wouldn't have it.
Maybe we could just add another linker section for these so as to save
space.

> parse_class() accepts 0 .. _DPRINTK_SITE_UNCLASSED, which allows the
>> control interface to explicitly manipulate unclassed callsites.
> 
> After parsing keywords, ddebug_parse_query() sets .class_id=15, iff it
> wasnt explicitly set.  This allows future classed/categorized
> callsites to be untouched by legacy (class unaware) queries.
> 
> DEFINE_DYNAMIC_DEBUG_METADATA gets _CLS(cls,) suffix and arg, and
> initializes the new .class_id=cls field.  The old name gets the default.
> 
> Then, these _CLS(cls,...) modifications are repeated up through the
> stack of *dynamic_func_call* macros that use the METADATA initializer,
> so as to actually supply the category into it.
> 
> NOTES:
> 
> _DPRINTK_SITE_UNCLASSED: this symbol is used to initialize all
> existing/unclassed pr-debug callsites.  Normally, the default would be
> zero, but DRM_UT_CORE "uses" that value, in the sense that 0 is
> exposed as a bit position in drm.debug.  Using 15 allows identity
> mapping from category to class, avoiding fiddly offsets.
> 
> Default .class_id = 15 means that ``echo +p > control`` no longer
> toggles ALL the callsites, only the unclassed ones.  This was only
> useful for static-branch toggle load testing anyway.
> 

I think that # echo +p > control should continue to work as is, why
should the introduction of classes change that ?

> RFC:
> 
> The new _CLS macro flavor gets a warning from DRM/dri-devel's CI,
> but not from checkpatch (on this subject).
> 
> a8f6c71f283e dyndbg: add class_id field and query support
> -:141: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'id' - possible side-effects?
> +#define __dynamic_func_call_cls(id, cls, fmt, func, ...) do {	\
> +	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);	\
> +	if (DYNAMIC_DEBUG_BRANCH(id))				\
> +		func(&id, ##__VA_ARGS__);			\
>  } while (0)
> 
> I couldn't fix it with a ``typeof(id) _id = id`` construct.  I haven't
> seen the warning myself, on the _CLS extended macro, nor the original.
> 
> CC: Rasmus Villemoes <linux@rasmusvillemoes.dk>
> Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
> ---
>  .../admin-guide/dynamic-debug-howto.rst       |  7 +++
>  include/linux/dynamic_debug.h                 | 54 ++++++++++++++-----
>  lib/dynamic_debug.c                           | 48 ++++++++++++++---
>  3 files changed, 88 insertions(+), 21 deletions(-)
> 
> diff --git a/Documentation/admin-guide/dynamic-debug-howto.rst b/Documentation/admin-guide/dynamic-debug-howto.rst
> index a89cfa083155..8ef8d7dcd140 100644
> --- a/Documentation/admin-guide/dynamic-debug-howto.rst
> +++ b/Documentation/admin-guide/dynamic-debug-howto.rst
> @@ -35,6 +35,7 @@ Dynamic debug has even more useful features:
>     - line number (including ranges of line numbers)
>     - module name
>     - format string
> +   - class number:0-15
>  
>   * Provides a debugfs control file: ``<debugfs>/dynamic_debug/control``
>     which can be read to display the complete list of known debug
> @@ -143,6 +144,7 @@ against.  Possible keywords are:::
>  		 'module' string |
>  		 'format' string |
>  		 'line' line-range
> +		 'class' integer:[0-15]
>  
>    line-range ::= lineno |
>  		 '-'lineno |
> @@ -217,6 +219,11 @@ line
>  	line -1605          // the 1605 lines from line 1 to line 1605
>  	line 1600-          // all lines from line 1600 to the end of the file
>  
> +class
> +    This expects a single integer in range: 0-15.
> +    15 is used/reserved for existing/unclassed callsites,
> +    and is defaulted in unless specified to >control
> +
>  The flags specification comprises a change operation followed
>  by one or more flag characters.  The change operation is one
>  of the characters::
> diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h
> index dce631e678dd..d4b48f3cc6e8 100644
> --- a/include/linux/dynamic_debug.h
> +++ b/include/linux/dynamic_debug.h
> @@ -6,6 +6,8 @@
>  #include <linux/jump_label.h>
>  #endif
>  
> +#include <linux/build_bug.h>
> +
>  /*
>   * An instance of this structure is created in a special
>   * ELF section at every dynamic debug callsite.  At runtime,
> @@ -21,6 +23,9 @@ struct _ddebug {
>  	const char *filename;
>  	const char *format;
>  	unsigned int lineno:18;
> +#define CLS_BITS 4
> +	unsigned int class_id:CLS_BITS;
> +#define _DPRINTK_SITE_UNCLASSED		((1 << CLS_BITS) - 1)
>  	/*
>  	 * The flags field controls the behaviour at the callsite.
>  	 * The bits here are changed dynamically when the user
> @@ -87,7 +92,7 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
>  			 const struct ib_device *ibdev,
>  			 const char *fmt, ...);
>  
> -#define DEFINE_DYNAMIC_DEBUG_METADATA(name, fmt)		\
> +#define DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, cls, fmt)	\
>  	static struct _ddebug  __aligned(8)			\
>  	__section("__dyndbg") name = {				\
>  		.modname = KBUILD_MODNAME,			\
> @@ -96,8 +101,14 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
>  		.format = (fmt),				\
>  		.lineno = __LINE__,				\
>  		.flags = _DPRINTK_FLAGS_DEFAULT,		\
> +		.class_id = cls,				\
>  		_DPRINTK_KEY_INIT				\
> -	}
> +	};							\
> +	BUILD_BUG_ON_MSG(cls > _DPRINTK_SITE_UNCLASSED,		\
> +			 "classid value overflow")
> +
> +#define DEFINE_DYNAMIC_DEBUG_METADATA(name, fmt)		\
> +	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, _DPRINTK_SITE_UNCLASSED, fmt)
>  
>  #ifdef CONFIG_JUMP_LABEL
>  
> @@ -128,18 +139,26 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
>  
>  #endif /* CONFIG_JUMP_LABEL */
>  
> -#define __dynamic_func_call(id, fmt, func, ...) do {	\
> -	DEFINE_DYNAMIC_DEBUG_METADATA(id, fmt);		\
> -	if (DYNAMIC_DEBUG_BRANCH(id))			\
> -		func(&id, ##__VA_ARGS__);		\
> +#define __dynamic_func_call_cls(id, cls, fmt, func, ...) do {	\
> +	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);	\
> +	if (DYNAMIC_DEBUG_BRANCH(id))				\
> +		func(&id, ##__VA_ARGS__);			\
>  } while (0)
>  
> -#define __dynamic_func_call_no_desc(id, fmt, func, ...) do {	\
> -	DEFINE_DYNAMIC_DEBUG_METADATA(id, fmt);			\
> -	if (DYNAMIC_DEBUG_BRANCH(id))				\
> -		func(__VA_ARGS__);				\
> +#define __dynamic_func_call_no_desc_cls(id, cls, fmt, func, ...) do {	\
> +	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);		\
> +	if (DYNAMIC_DEBUG_BRANCH(id))					\
> +		func(__VA_ARGS__);					\
>  } while (0)
>  
> +#define __dynamic_func_call(id, fmt, func, ...)				\
> +	__dynamic_func_call_cls(id, _DPRINTK_SITE_UNCLASSED,		\
> +				fmt, func, ##__VA_ARGS__)
> +
> +#define __dynamic_func_call_no_desc(id, fmt, func, ...)			\
> +	__dynamic_func_call_no_desc_cls(id, _DPRINTK_SITE_UNCLASSED,	\
> +					fmt, func, ##__VA_ARGS__)
> +
>  /*
>   * "Factory macro" for generating a call to func, guarded by a
>   * DYNAMIC_DEBUG_BRANCH. The dynamic debug descriptor will be
> @@ -148,15 +167,24 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
>   * the varargs. Note that fmt is repeated in invocations of this
>   * macro.
>   */
> +#define _dynamic_func_call_cls(cls, fmt, func, ...)			\
> +	__dynamic_func_call_cls(__UNIQUE_ID(ddebug), cls, fmt, func, ##__VA_ARGS__)
>  #define _dynamic_func_call(fmt, func, ...)				\
> -	__dynamic_func_call(__UNIQUE_ID(ddebug), fmt, func, ##__VA_ARGS__)
> +	_dynamic_func_call_cls(_DPRINTK_SITE_UNCLASSED, fmt, func, ##__VA_ARGS__)
> +
>  /*
>   * A variant that does the same, except that the descriptor is not
>   * passed as the first argument to the function; it is only called
>   * with precisely the macro's varargs.
>   */
> -#define _dynamic_func_call_no_desc(fmt, func, ...)	\
> -	__dynamic_func_call_no_desc(__UNIQUE_ID(ddebug), fmt, func, ##__VA_ARGS__)
> +#define _dynamic_func_call_no_desc_cls(fmt, cat, func, ...)		\
> +	__dynamic_func_call_no_desc_cls(__UNIQUE_ID(ddebug), cat,	\
> +					fmt, func, ##__VA_ARGS__)
> +
> +#define _dynamic_func_call_no_desc(fmt, func, ...)			\
> +	__dynamic_func_call_no_desc_cls(__UNIQUE_ID(ddebug),		\
> +					_DPRINTK_SITE_UNCLASSED,	\
> +					fmt, func, ##__VA_ARGS__)
>  
>  #define dynamic_pr_debug(fmt, ...)				\
>  	_dynamic_func_call(fmt,	__dynamic_pr_debug,		\
> diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
> index a56c1286ffa4..ee2129becacc 100644
> --- a/lib/dynamic_debug.c
> +++ b/lib/dynamic_debug.c
> @@ -55,6 +55,8 @@ struct ddebug_query {
>  	const char *function;
>  	const char *format;
>  	unsigned int first_lineno, last_lineno;
> +	unsigned int class_id;
> +	unsigned int class_marked:1;
>  };
>  
>  struct ddebug_iter {
> @@ -134,13 +136,13 @@ static void vpr_info_dq(const struct ddebug_query *query, const char *msg)
>  			fmtlen--;
>  	}
>  
> -	v3pr_info("%s: func=\"%s\" file=\"%s\" module=\"%s\" format=\"%.*s\" lineno=%u-%u\n",
> -		 msg,
> -		 query->function ?: "",
> -		 query->filename ?: "",
> -		 query->module ?: "",
> -		 fmtlen, query->format ?: "",
> -		 query->first_lineno, query->last_lineno);
> +	v3pr_info("%s: func=\"%s\" file=\"%s\" module=\"%s\" format=\"%.*s\" lineno=%u-%u class=%u\n",
> +		  msg,
> +		  query->function ?: "",
> +		  query->filename ?: "",
> +		  query->module ?: "",
> +		  fmtlen, query->format ?: "",
> +		  query->first_lineno, query->last_lineno, query->class_id);
>  }
>  
>  /*
> @@ -170,6 +172,10 @@ static int ddebug_change(const struct ddebug_query *query,
>  		for (i = 0; i < dt->num_ddebugs; i++) {
>  			struct _ddebug *dp = &dt->ddebugs[i];
>  
> +			/* match against the class_id, either given or default */
> +			if (query->class_id != dp->class_id)
> +				continue;
> +

Could that instead be:

if (query->class_id != _DPRINTK_SITE_UNCLASSED && query->class_id != dp->class_id)
	continue;

That way by default we fall through and enable callsites for queries that do not have
the 'class' keyboard specified. I think I would also disallow setting
_DPRINTK_SITE_UNCLASSED explicitly since it would now mean 'class *', and what if
we expand class_id size to greater than 15 at some point?

>  			/* match against the source filename */
>  			if (query->filename &&
>  			    !match_wildcard(query->filename, dp->filename) &&
> @@ -308,6 +314,21 @@ static inline int parse_lineno(const char *str, unsigned int *val)
>  	return 0;
>  }
>  
> +static inline int parse_class(struct ddebug_query *query, const char *str)
> +{
> +	int rc;
> +	unsigned int val;
> +
> +	rc = kstrtouint(str, 10, &val);
> +	if (rc < 0 || val > _DPRINTK_SITE_UNCLASSED) {
> +		pr_err("expecting class:[0-%d], not %s\n", _DPRINTK_SITE_UNCLASSED, str);
> +		return -EINVAL;
> +	}
> +	query->class_id = val;
> +	query->class_marked = 1;
> +	return 0;
> +}
> +
>  static int parse_linerange(struct ddebug_query *query, const char *first)
>  {
>  	char *last = strchr(first, '-');
> @@ -421,6 +442,9 @@ static int ddebug_parse_query(char *words[], int nwords,
>  		} else if (!strcmp(keyword, "line")) {
>  			if (parse_linerange(query, arg))
>  				return -EINVAL;
> +		} else if (!strcmp(keyword, "class")) {
> +			if (parse_class(query, arg))
> +				return -EINVAL;
>  		} else {
>  			pr_err("unknown keyword \"%s\"\n", keyword);
>  			return -EINVAL;
> @@ -428,6 +452,10 @@ static int ddebug_parse_query(char *words[], int nwords,
>  		if (rc)
>  			return rc;
>  	}
> +	/* post-validate the query, set default */
> +	if (!query->class_marked)
> +		query->class_id = _DPRINTK_SITE_UNCLASSED;
> +
>  	vpr_info_dq(query, "parsed");
>  	return 0;
>  }
> @@ -900,7 +928,11 @@ static int ddebug_proc_show(struct seq_file *m, void *p)
>  		   iter->table->mod_name, dp->function,
>  		   ddebug_describe_flags(dp->flags, &flags));
>  	seq_escape(m, dp->format, "\t\r\n\"");
> -	seq_puts(m, "\"\n");
> +	seq_puts(m, "\"");
> +
> +	if (dp->class_id != _DPRINTK_SITE_UNCLASSED)
> +		seq_printf(m, " cls:%u", dp->class_id);
> +	seq_puts(m, "\n");
>  
>  	return 0;
>  }

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 2/5] dyndbg: add class_id field and query support
  2022-03-11 19:06   ` Jason Baron
@ 2022-03-12  1:06     ` jim.cromie
  2022-03-14 21:29       ` Jason Baron
  0 siblings, 1 reply; 11+ messages in thread
From: jim.cromie @ 2022-03-12  1:06 UTC (permalink / raw)
  To: Jason Baron
  Cc: Greg KH, LKML, Daniel Vetter, Sean Paul, robdclark,
	Rasmus Villemoes, Joe Perches, dri-devel, amd-gfx mailing list,
	intel-gvt-dev, Intel Graphics Development

On Fri, Mar 11, 2022 at 12:06 PM Jason Baron <jbaron@akamai.com> wrote:
>
>
>
> On 3/10/22 23:47, Jim Cromie wrote:
> > DRM defines/uses 10 enum drm_debug_category's to create exclusive
> > classes of debug messages.  To support this directly in dynamic-debug,
> > add the following:
> >
> > - struct _ddebug.class_id:4 - 4 bits is enough
> > - define _DPRINTK_SITE_UNCLASSED 15 - see below
> >
> > and the query support:
> > - struct _ddebug_query.class_id
> > - ddebug_parse_query  - looks for "class" keyword, then calls..
> > - parse_class         - accepts uint: 0-15, sets query.class_id and marker
> > - vpr_info_dq         - displays new field
> > - ddebug_proc_show    - append column with "cls:%d" if !UNCLASSED
> >
> > With the patch, one can enable current/unclassed callsites by:
> >
> >   #> echo drm class 15 +p > /proc/dynamic_debug/control
> >
>
> To me, this is hard to read, what the heck is '15'? I have to go look it
> up in the control file and it's not descriptive. I think that using
> classes/categories makes sense but I'm wondering if it can be a bit more
> user friendly? Perhaps, we can pass an array of strings that is indexed
> by the class id to each pr_debug() site that wants to use class. So
> something like:
>

Im not at all averse to nice names, but as something added on.

1st, the interface to make friendlier is DRM's

echo 0x04 > /sys/module/drm/parameters/debug   # which category ?

parm:           debug:Enable debug output, where each bit enables a
debug category.
Bit 0 (0x01)  will enable CORE messages (drm core code)
Bit 1 (0x02)  will enable DRIVER messages (drm controller code)
Bit 2 (0x04)  will enable KMS messages (modesetting code)

echo DRM_UT_DRIVER,DRM_UT_KMS > /sys/module/drm/parameters/debug   #
now its pretty clear

If that works, its cuz DEFINE_DYNAMIC_DEBUG_CLASSBITS()
plucks class symbols out of its __VA_ARGS__, and #stringifes them.
So that macro could then build the 1-per-module static constant string array
and (only) the callbacks would be able to use it.

from there, it shouldnt be hard to ref that from the module's ddebug_table,
so parse_query could validate class args against the module given (or
fail if none given)

Speaking strictly, theres a gap:
   echo module * class DRM_UT_KMS +p > control
is nonsense for * other than drm + drivers,
but its fair/safe to just disallow wildcards on modname for this purpose.

it does however imply that module foo must exist for FOO_CAT_1 to be usable.
thats not currently the case:
bash-5.1# echo module foo +p > /proc/dynamic_debug/control
[   15.403749] dyndbg: read 14 bytes from userspace
[   15.405413] dyndbg: query 0: "module foo +p" mod:*
[   15.406486] dyndbg: split into words: "module" "foo" "+p"
[   15.407070] dyndbg: op='+'
[   15.407388] dyndbg: flags=0x1
[   15.407809] dyndbg: *flagsp=0x1 *maskp=0xffffffff
[   15.408300] dyndbg: parsed: func="" file="" module="foo" format=""
lineno=0-0 class=15
[   15.409151] dyndbg: no matches for query
[   15.409591] dyndbg: no-match: func="" file="" module="foo"
format="" lineno=0-0 class=15
[   15.410524] dyndbg: processed 1 queries, with 0 matches, 0 errs
bash-5.1#

ISTM we can keep that "0 errs" response for that case, but still reject this:

   echo module foo class FOO_NOT_HERE +p > /proc/dynamic_debug/control


> enum levels {
>         LOW,
>         MEDIUM,
>         HIGH
> };

I want to steer clear of "level" anything,
since 2>1 implies non independence of the categories



>
> static const char * const level_to_strings[] = {
>         [LOW] = "low",
>         [MEDIUM] = "medium",
>         [HIGH] = "high",
> };
>
> And then you'd have a wrapper macros in your driver:
>
> #define module_foo_pr_debug_class(level, fmt, args...)
>         pr_debug_class(level, level_to_strings, fmt, args);
>
> Such that call sites look like:
>
> module_foo_pr_debug_class(LOW, fmt, args...);
>

That macro, minus the "module_foo_" prefix,
could go into dynamic_debug.h

I didnt do that, for 2 reasons:

DRM didnt need it - it had an enum var,
and a set of macros to encapsulate the categories.

-  the "prototype" looks like this might be ok:
    define LOW "low"
    pr_debug_class(LOW, "mumble about something %p %p\n", foo, bar)
ok thats a stretch, but...

Basically, I didnt want to deal with creating a new interface. KIS

> Such that you're not always passing the strings array around. Now, this
> does mean another pointer for struct _ddebug and most wouldn't have it.
> Maybe we could just add another linker section for these so as to save
> space.

From a space consideration, adding DRM users means adding
thousands of callsites ~2k for i915, ~4k for amdgpu
Also increasing per/callsite memory seems counterproductive.

see DEFINE_DYNAMIC_DEBUG_CLASSBITS for the way forward.

> > parse_class() accepts 0 .. _DPRINTK_SITE_UNCLASSED, which allows the
> >> control interface to explicitly manipulate unclassed callsites.
> >
> > After parsing keywords, ddebug_parse_query() sets .class_id=15, iff it
> > wasnt explicitly set.  This allows future classed/categorized
> > callsites to be untouched by legacy (class unaware) queries.
> >
> > DEFINE_DYNAMIC_DEBUG_METADATA gets _CLS(cls,) suffix and arg, and
> > initializes the new .class_id=cls field.  The old name gets the default.
> >
> > Then, these _CLS(cls,...) modifications are repeated up through the
> > stack of *dynamic_func_call* macros that use the METADATA initializer,
> > so as to actually supply the category into it.
> >
> > NOTES:
> >
> > _DPRINTK_SITE_UNCLASSED: this symbol is used to initialize all
> > existing/unclassed pr-debug callsites.  Normally, the default would be
> > zero, but DRM_UT_CORE "uses" that value, in the sense that 0 is
> > exposed as a bit position in drm.debug.  Using 15 allows identity
> > mapping from category to class, avoiding fiddly offsets.
> >
> > Default .class_id = 15 means that ``echo +p > control`` no longer
> > toggles ALL the callsites, only the unclassed ones.  This was only
> > useful for static-branch toggle load testing anyway.
> >
>
> I think that # echo +p > control should continue to work as is, why
> should the introduction of classes change that ?

1st, its a good way to flood your syslog, forex if theres a serial driver built.
And this might get dramatically worse if DRM_VBLANK got lumped in.

but basically, it still affects the same universe of callsites
as before DRM popped/s into existence (which could double the callsites)

ISTM   ``echo "+p;-p" >control``
is primarily useful for work-load generation.
whether 1 cycle is 4k or 8k toggles is unimportant.



>
> > RFC:
> >
> > The new _CLS macro flavor gets a warning from DRM/dri-devel's CI,
> > but not from checkpatch (on this subject).
> >
> > a8f6c71f283e dyndbg: add class_id field and query support
> > -:141: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'id' - possible side-effects?
> > +#define __dynamic_func_call_cls(id, cls, fmt, func, ...) do {        \
> > +     DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);        \
> > +     if (DYNAMIC_DEBUG_BRANCH(id))                           \
> > +             func(&id, ##__VA_ARGS__);                       \
> >  } while (0)
> >
> > I couldn't fix it with a ``typeof(id) _id = id`` construct.  I haven't
> > seen the warning myself, on the _CLS extended macro, nor the original.
> >
> > CC: Rasmus Villemoes <linux@rasmusvillemoes.dk>
> > Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
> > ---
> >  .../admin-guide/dynamic-debug-howto.rst       |  7 +++
> >  include/linux/dynamic_debug.h                 | 54 ++++++++++++++-----
> >  lib/dynamic_debug.c                           | 48 ++++++++++++++---
> >  3 files changed, 88 insertions(+), 21 deletions(-)
> >
> > diff --git a/Documentation/admin-guide/dynamic-debug-howto.rst b/Documentation/admin-guide/dynamic-debug-howto.rst
> > index a89cfa083155..8ef8d7dcd140 100644
> > --- a/Documentation/admin-guide/dynamic-debug-howto.rst
> > +++ b/Documentation/admin-guide/dynamic-debug-howto.rst
> > @@ -35,6 +35,7 @@ Dynamic debug has even more useful features:
> >     - line number (including ranges of line numbers)
> >     - module name
> >     - format string
> > +   - class number:0-15
> >
> >   * Provides a debugfs control file: ``<debugfs>/dynamic_debug/control``
> >     which can be read to display the complete list of known debug
> > @@ -143,6 +144,7 @@ against.  Possible keywords are:::
> >                'module' string |
> >                'format' string |
> >                'line' line-range
> > +              'class' integer:[0-15]
> >
> >    line-range ::= lineno |
> >                '-'lineno |
> > @@ -217,6 +219,11 @@ line
> >       line -1605          // the 1605 lines from line 1 to line 1605
> >       line 1600-          // all lines from line 1600 to the end of the file
> >
> > +class
> > +    This expects a single integer in range: 0-15.
> > +    15 is used/reserved for existing/unclassed callsites,
> > +    and is defaulted in unless specified to >control
> > +
> >  The flags specification comprises a change operation followed
> >  by one or more flag characters.  The change operation is one
> >  of the characters::
> > diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h
> > index dce631e678dd..d4b48f3cc6e8 100644
> > --- a/include/linux/dynamic_debug.h
> > +++ b/include/linux/dynamic_debug.h
> > @@ -6,6 +6,8 @@
> >  #include <linux/jump_label.h>
> >  #endif
> >
> > +#include <linux/build_bug.h>
> > +
> >  /*
> >   * An instance of this structure is created in a special
> >   * ELF section at every dynamic debug callsite.  At runtime,
> > @@ -21,6 +23,9 @@ struct _ddebug {
> >       const char *filename;
> >       const char *format;
> >       unsigned int lineno:18;
> > +#define CLS_BITS 4
> > +     unsigned int class_id:CLS_BITS;
> > +#define _DPRINTK_SITE_UNCLASSED              ((1 << CLS_BITS) - 1)
> >       /*
> >        * The flags field controls the behaviour at the callsite.
> >        * The bits here are changed dynamically when the user
> > @@ -87,7 +92,7 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
> >                        const struct ib_device *ibdev,
> >                        const char *fmt, ...);
> >
> > -#define DEFINE_DYNAMIC_DEBUG_METADATA(name, fmt)             \
> > +#define DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, cls, fmt)    \
> >       static struct _ddebug  __aligned(8)                     \
> >       __section("__dyndbg") name = {                          \
> >               .modname = KBUILD_MODNAME,                      \
> > @@ -96,8 +101,14 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
> >               .format = (fmt),                                \
> >               .lineno = __LINE__,                             \
> >               .flags = _DPRINTK_FLAGS_DEFAULT,                \
> > +             .class_id = cls,                                \
> >               _DPRINTK_KEY_INIT                               \
> > -     }
> > +     };                                                      \
> > +     BUILD_BUG_ON_MSG(cls > _DPRINTK_SITE_UNCLASSED,         \
> > +                      "classid value overflow")
> > +
> > +#define DEFINE_DYNAMIC_DEBUG_METADATA(name, fmt)             \
> > +     DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, _DPRINTK_SITE_UNCLASSED, fmt)
> >
> >  #ifdef CONFIG_JUMP_LABEL
> >
> > @@ -128,18 +139,26 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
> >
> >  #endif /* CONFIG_JUMP_LABEL */
> >
> > -#define __dynamic_func_call(id, fmt, func, ...) do { \
> > -     DEFINE_DYNAMIC_DEBUG_METADATA(id, fmt);         \
> > -     if (DYNAMIC_DEBUG_BRANCH(id))                   \
> > -             func(&id, ##__VA_ARGS__);               \
> > +#define __dynamic_func_call_cls(id, cls, fmt, func, ...) do {        \
> > +     DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);        \
> > +     if (DYNAMIC_DEBUG_BRANCH(id))                           \
> > +             func(&id, ##__VA_ARGS__);                       \
> >  } while (0)
> >
> > -#define __dynamic_func_call_no_desc(id, fmt, func, ...) do { \
> > -     DEFINE_DYNAMIC_DEBUG_METADATA(id, fmt);                 \
> > -     if (DYNAMIC_DEBUG_BRANCH(id))                           \
> > -             func(__VA_ARGS__);                              \
> > +#define __dynamic_func_call_no_desc_cls(id, cls, fmt, func, ...) do {        \
> > +     DEFINE_DYNAMIC_DEBUG_METADATA_CLS(id, cls, fmt);                \
> > +     if (DYNAMIC_DEBUG_BRANCH(id))                                   \
> > +             func(__VA_ARGS__);                                      \
> >  } while (0)
> >
> > +#define __dynamic_func_call(id, fmt, func, ...)                              \
> > +     __dynamic_func_call_cls(id, _DPRINTK_SITE_UNCLASSED,            \
> > +                             fmt, func, ##__VA_ARGS__)
> > +
> > +#define __dynamic_func_call_no_desc(id, fmt, func, ...)                      \
> > +     __dynamic_func_call_no_desc_cls(id, _DPRINTK_SITE_UNCLASSED,    \
> > +                                     fmt, func, ##__VA_ARGS__)
> > +
> >  /*
> >   * "Factory macro" for generating a call to func, guarded by a
> >   * DYNAMIC_DEBUG_BRANCH. The dynamic debug descriptor will be
> > @@ -148,15 +167,24 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
> >   * the varargs. Note that fmt is repeated in invocations of this
> >   * macro.
> >   */
> > +#define _dynamic_func_call_cls(cls, fmt, func, ...)                  \
> > +     __dynamic_func_call_cls(__UNIQUE_ID(ddebug), cls, fmt, func, ##__VA_ARGS__)
> >  #define _dynamic_func_call(fmt, func, ...)                           \
> > -     __dynamic_func_call(__UNIQUE_ID(ddebug), fmt, func, ##__VA_ARGS__)
> > +     _dynamic_func_call_cls(_DPRINTK_SITE_UNCLASSED, fmt, func, ##__VA_ARGS__)
> > +
> >  /*
> >   * A variant that does the same, except that the descriptor is not
> >   * passed as the first argument to the function; it is only called
> >   * with precisely the macro's varargs.
> >   */
> > -#define _dynamic_func_call_no_desc(fmt, func, ...)   \
> > -     __dynamic_func_call_no_desc(__UNIQUE_ID(ddebug), fmt, func, ##__VA_ARGS__)
> > +#define _dynamic_func_call_no_desc_cls(fmt, cat, func, ...)          \
> > +     __dynamic_func_call_no_desc_cls(__UNIQUE_ID(ddebug), cat,       \
> > +                                     fmt, func, ##__VA_ARGS__)
> > +
> > +#define _dynamic_func_call_no_desc(fmt, func, ...)                   \
> > +     __dynamic_func_call_no_desc_cls(__UNIQUE_ID(ddebug),            \
> > +                                     _DPRINTK_SITE_UNCLASSED,        \
> > +                                     fmt, func, ##__VA_ARGS__)
> >
> >  #define dynamic_pr_debug(fmt, ...)                           \
> >       _dynamic_func_call(fmt, __dynamic_pr_debug,             \
> > diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
> > index a56c1286ffa4..ee2129becacc 100644
> > --- a/lib/dynamic_debug.c
> > +++ b/lib/dynamic_debug.c
> > @@ -55,6 +55,8 @@ struct ddebug_query {
> >       const char *function;
> >       const char *format;
> >       unsigned int first_lineno, last_lineno;
> > +     unsigned int class_id;
> > +     unsigned int class_marked:1;
> >  };
> >
> >

> >  /*
> > @@ -170,6 +172,10 @@ static int ddebug_change(const struct ddebug_query *query,
> >               for (i = 0; i < dt->num_ddebugs; i++) {
> >                       struct _ddebug *dp = &dt->ddebugs[i];
> >
> > +                     /* match against the class_id, either given or default */
> > +                     if (query->class_id != dp->class_id)
> > +                             continue;
> > +
>
> Could that instead be:
>
> if (query->class_id != _DPRINTK_SITE_UNCLASSED && query->class_id != dp->class_id)
>         continue;
>

I think that is the effect currently.

query construction adds in the default if the term wasnt specified.

> That way by default we fall through and enable callsites for queries that do not have
> the 'class' keyboard specified.

thats what I was avoiding - I suppose its a matter of perspective

all existing callsites are unclassified. - ie 15
to preserve the universe affected by old queries,
adding 15 by default comports with this.
then when class !=15 callsites are added,
they are immune from effects by existing queries.

classed queries are the alternate universe.
and need "class N" in >control to get there.
Id expect DRM would like the isolation,
at least against inadvertent changes.

that said,   ``echo +p >control`` can only affect 1 universe at a time.
ISTM this is fine,
the only thing lost is 1/2 the workload for
   while true do echo "+p ; -p" >control; done

> I think I would also disallow setting
> _DPRINTK_SITE_UNCLASSED explicitly since it would now mean 'class *', and what if
> we expand class_id size to greater than 15 at some point?

it doesnt mean class *, it means class OTHER than 0-14

earlier version of parse_class disallowed 15 explicitly,
I changed it to simplify testing.

re-adding that constraint could be argued,
but if class 15 is the default when unspecified,
its hard to explain why "class 15" could draw an errror.

we'd just increase bit-field size, class_id:5
and bump 15 to 31.
is there an appropriate *_MAX macro to use,
to add the right implications / connotations ?
theres something vaguely overflowy / rollover-y here.
What choice words would convey it succinctly ?

>
> >                       /* match against the source filename */
> >                       if (query->filename &&
> >                           !match_wildcard(query->filename, dp->filename) &&
> > @@ -308,6 +314,21 @@ static inline int parse_lineno(const char *str, unsigned int *val)
> >       return 0;
> >  }
> >
> > +static inline int parse_class(struct ddebug_query *query, const char *str)
> > +{
> > +     int rc;
> > +     unsigned int val;
> > +
> > +     rc = kstrtouint(str, 10, &val);
> > +     if (rc < 0 || val > _DPRINTK_SITE_UNCLASSED) {
> > +             pr_err("expecting class:[0-%d], not %s\n", _DPRINTK_SITE_UNCLASSED, str);
> > +             return -EINVAL;
> > +     }
> > +     query->class_id = val;
> > +     query->class_marked = 1;
> > +     return 0;
> > +}
> > +
> >  static int parse_linerange(struct ddebug_query *query, const char *first)
> >  {
> >       char *last = strchr(first, '-');
> > @@ -421,6 +442,9 @@ static int ddebug_parse_query(char *words[], int nwords,
> >               } else if (!strcmp(keyword, "line")) {
> >                       if (parse_linerange(query, arg))
> >                               return -EINVAL;
> > +             } else if (!strcmp(keyword, "class")) {
> > +                     if (parse_class(query, arg))
> > +                             return -EINVAL;
> >               } else {
> >                       pr_err("unknown keyword \"%s\"\n", keyword);
> >                       return -EINVAL;
> > @@ -428,6 +452,10 @@ static int ddebug_parse_query(char *words[], int nwords,
> >               if (rc)
> >                       return rc;
> >       }
> > +     /* post-validate the query, set default */
> > +     if (!query->class_marked)
> > +             query->class_id = _DPRINTK_SITE_UNCLASSED;
> > +
> >       vpr_info_dq(query, "parsed");
> >       return 0;
> >  }
> > @@ -900,7 +928,11 @@ static int ddebug_proc_show(struct seq_file *m, void *p)
> >                  iter->table->mod_name, dp->function,
> >                  ddebug_describe_flags(dp->flags, &flags));
> >       seq_escape(m, dp->format, "\t\r\n\"");
> > -     seq_puts(m, "\"\n");
> > +     seq_puts(m, "\"");
> > +
> > +     if (dp->class_id != _DPRINTK_SITE_UNCLASSED)
> > +             seq_printf(m, " cls:%u", dp->class_id);
> > +     seq_puts(m, "\n");
> >
> >       return 0;
> >  }

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 2/5] dyndbg: add class_id field and query support
  2022-03-12  1:06     ` jim.cromie
@ 2022-03-14 21:29       ` Jason Baron
  2022-03-28 19:07         ` jim.cromie
  0 siblings, 1 reply; 11+ messages in thread
From: Jason Baron @ 2022-03-14 21:29 UTC (permalink / raw)
  To: jim.cromie
  Cc: Greg KH, LKML, Daniel Vetter, Sean Paul, robdclark,
	Rasmus Villemoes, Joe Perches, dri-devel, amd-gfx mailing list,
	intel-gvt-dev, Intel Graphics Development



On 3/11/22 20:06, jim.cromie@gmail.com wrote:
> On Fri, Mar 11, 2022 at 12:06 PM Jason Baron <jbaron@akamai.com> wrote:
>>
>>
>>
>> On 3/10/22 23:47, Jim Cromie wrote:
>>> DRM defines/uses 10 enum drm_debug_category's to create exclusive
>>> classes of debug messages.  To support this directly in dynamic-debug,
>>> add the following:
>>>
>>> - struct _ddebug.class_id:4 - 4 bits is enough
>>> - define _DPRINTK_SITE_UNCLASSED 15 - see below
>>>
>>> and the query support:
>>> - struct _ddebug_query.class_id
>>> - ddebug_parse_query  - looks for "class" keyword, then calls..
>>> - parse_class         - accepts uint: 0-15, sets query.class_id and marker
>>> - vpr_info_dq         - displays new field
>>> - ddebug_proc_show    - append column with "cls:%d" if !UNCLASSED
>>>
>>> With the patch, one can enable current/unclassed callsites by:
>>>
>>>   #> echo drm class 15 +p > /proc/dynamic_debug/control
>>>
>>
>> To me, this is hard to read, what the heck is '15'? I have to go look it
>> up in the control file and it's not descriptive. I think that using
>> classes/categories makes sense but I'm wondering if it can be a bit more
>> user friendly? Perhaps, we can pass an array of strings that is indexed
>> by the class id to each pr_debug() site that wants to use class. So
>> something like:
>>
> 
> Im not at all averse to nice names, but as something added on.
> 
> 1st, the interface to make friendlier is DRM's
> 
> echo 0x04 > /sys/module/drm/parameters/debug   # which category ?
> 
> parm:           debug:Enable debug output, where each bit enables a
> debug category.
> Bit 0 (0x01)  will enable CORE messages (drm core code)
> Bit 1 (0x02)  will enable DRIVER messages (drm controller code)
> Bit 2 (0x04)  will enable KMS messages (modesetting code)
> 
> echo DRM_UT_DRIVER,DRM_UT_KMS > /sys/module/drm/parameters/debug   #
> now its pretty clear
> 
> If that works, its cuz DEFINE_DYNAMIC_DEBUG_CLASSBITS()
> plucks class symbols out of its __VA_ARGS__, and #stringifes them.
> So that macro could then build the 1-per-module static constant string array
> and (only) the callbacks would be able to use it.
> 
> from there, it shouldnt be hard to ref that from the module's ddebug_table,
> so parse_query could validate class args against the module given (or
> fail if none given)
> 
> Speaking strictly, theres a gap:
>    echo module * class DRM_UT_KMS +p > control
> is nonsense for * other than drm + drivers,
> but its fair/safe to just disallow wildcards on modname for this purpose.
> 
> it does however imply that module foo must exist for FOO_CAT_1 to be usable.
> thats not currently the case:
> bash-5.1# echo module foo +p > /proc/dynamic_debug/control
> [   15.403749] dyndbg: read 14 bytes from userspace
> [   15.405413] dyndbg: query 0: "module foo +p" mod:*
> [   15.406486] dyndbg: split into words: "module" "foo" "+p"
> [   15.407070] dyndbg: op='+'
> [   15.407388] dyndbg: flags=0x1
> [   15.407809] dyndbg: *flagsp=0x1 *maskp=0xffffffff
> [   15.408300] dyndbg: parsed: func="" file="" module="foo" format=""
> lineno=0-0 class=15
> [   15.409151] dyndbg: no matches for query
> [   15.409591] dyndbg: no-match: func="" file="" module="foo"
> format="" lineno=0-0 class=15
> [   15.410524] dyndbg: processed 1 queries, with 0 matches, 0 errs
> bash-5.1#
> 
> ISTM we can keep that "0 errs" response for that case, but still reject this:
> 
>    echo module foo class FOO_NOT_HERE +p > /proc/dynamic_debug/control
> 
> 

Ok, yeah so I guess I'm thinking about the 'class' more as global namespace,
so that for example, it could span modules, or be specific to certain modules.
I'm also thinking of it as a 'string' which is maybe hierarchical, so that it's
clear what sub-system, or sub-sub-system it belongs to. So for DRM for example,
it could be a string like "DRM:CORE". The index num I think is still helpful for
implementation so we don't have to store a pointer size, but I don't think it's
really exposed (except perhaps in module params b/c drm is doing that already?).


>> enum levels {
>>         LOW,
>>         MEDIUM,
>>         HIGH
>> };
> 
> I want to steer clear of "level" anything,
> since 2>1 implies non independence of the categories
> 
> 

Agreed, that was a bad example on my part.

I've put together a rough patch on top of your series, to make my thinking
hopefully clear. I also think that the module param callback thing could be
implemented in this 'global' space with the 0-14 low numbers, by adding
some sort of offset into the table. IE you would add the low number +
the offset to get the 'string' to add to the ddebug query. I commented it
out in my patch below b/c I didn't implement that part.

Thoughts?


diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h
index 664bb83778d2..b0bc1b536d54 100644
--- a/include/linux/dynamic_debug.h
+++ b/include/linux/dynamic_debug.h
@@ -8,6 +8,14 @@

 #include <linux/build_bug.h>

+enum ddebug_categories {
+        FOO_BAR = 0,
+        DRM_A = 1,
+        DRM_B = 2,
+        DRM_C = 3,
+        //_DPRINTK_SITE_DEFAULT 63 is max
+};
+
 /*
  * An instance of this structure is created in a special
  * ELF section at every dynamic debug callsite.  At runtime,
@@ -23,9 +31,9 @@ struct _ddebug {
 	const char *filename;
 	const char *format;
 	unsigned int lineno:18;
-#define CLS_BITS 4
+#define CLS_BITS 6
 	unsigned int class_id:CLS_BITS;
-#define _DPRINTK_SITE_UNCLASSED		((1 << CLS_BITS) - 1)
+#define _DPRINTK_SITE_DEFAULT		((1 << CLS_BITS) - 1)
 	/*
 	 * The flags field controls the behaviour at the callsite.
 	 * The bits here are changed dynamically when the user
@@ -101,11 +109,11 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
 		.class_id = cls,				\
 		_DPRINTK_KEY_INIT				\
 	};							\
-	BUILD_BUG_ON_MSG(cls > _DPRINTK_SITE_UNCLASSED,		\
+	BUILD_BUG_ON_MSG(cls > _DPRINTK_SITE_DEFAULT,		\
 			 "classid value overflow")

 #define DEFINE_DYNAMIC_DEBUG_METADATA(name, fmt)		\
-	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, _DPRINTK_SITE_UNCLASSED, fmt)
+	DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, _DPRINTK_SITE_DEFAULT, fmt)

 #ifdef CONFIG_JUMP_LABEL

@@ -149,11 +157,11 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
 } while (0)

 #define __dynamic_func_call(id, fmt, func, ...)				\
-	__dynamic_func_call_cls(id, _DPRINTK_SITE_UNCLASSED,		\
+	__dynamic_func_call_cls(id, _DPRINTK_SITE_DEFAULT,		\
 				fmt, func, ##__VA_ARGS__)

 #define __dynamic_func_call_no_desc(id, fmt, func, ...)			\
-	__dynamic_func_call_no_desc_cls(id, _DPRINTK_SITE_UNCLASSED,	\
+	__dynamic_func_call_no_desc_cls(id, _DPRINTK_SITE_DEFAULT,	\
 					fmt, func, ##__VA_ARGS__)

 /*
@@ -167,7 +175,7 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,
 #define _dynamic_func_call_cls(cls, fmt, func, ...)			\
 	__dynamic_func_call_cls(__UNIQUE_ID(ddebug), cls, fmt, func, ##__VA_ARGS__)
 #define _dynamic_func_call(fmt, func, ...)				\
-	_dynamic_func_call_cls(_DPRINTK_SITE_UNCLASSED, fmt, func, ##__VA_ARGS__)
+	_dynamic_func_call_cls(_DPRINTK_SITE_DEFAULT, fmt, func, ##__VA_ARGS__)

 /*
  * A variant that does the same, except that the descriptor is not
@@ -180,7 +188,7 @@ void __dynamic_ibdev_dbg(struct _ddebug *descriptor,

 #define _dynamic_func_call_no_desc(fmt, func, ...)			\
 	__dynamic_func_call_no_desc_cls(__UNIQUE_ID(ddebug),		\
-					_DPRINTK_SITE_UNCLASSED,	\
+					_DPRINTK_SITE_DEFAULT,	\
 					fmt, func, ##__VA_ARGS__)

 #define dynamic_pr_debug(fmt, ...)				\
@@ -284,7 +292,7 @@ struct dyndbg_classbits_param {
 	static struct dyndbg_classbits_param ddcats_##_var = {		\
 		.bits = &(_var),					\
 		.flags = _flgs,						\
-		.classes = { __VA_ARGS__, _DPRINTK_SITE_UNCLASSED }	\
+		.classes = { __VA_ARGS__, _DPRINTK_SITE_DEFAULT }	\
 	};								\
 	module_param_cb(fsname, &param_ops_dyndbg_classbits,		\
 			&ddcats_##_var, 0644)
diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index 11c8d0771cd2..0f390638e46c 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -39,6 +39,14 @@

 #include <rdma/ib_verbs.h>

+static const char * const cat_to_strings[] = {
+        [FOO_BAR] = "foo:bar",
+        [DRM_A] = "drm:A",
+        [DRM_B] = "drm:B",
+        [DRM_C] = "drm:C",
+	[_DPRINTK_SITE_DEFAULT] = "default",
+};
+
 extern struct _ddebug __start___dyndbg[];
 extern struct _ddebug __stop___dyndbg[];

@@ -55,8 +63,7 @@ struct ddebug_query {
 	const char *function;
 	const char *format;
 	unsigned int first_lineno, last_lineno;
-	unsigned int class_id;
-	unsigned int class_marked:1;
+	const char *class_string;
 };

 struct ddebug_iter {
@@ -136,13 +143,13 @@ static void vpr_info_dq(const struct ddebug_query *query, const char *msg)
 			fmtlen--;
 	}

-	v3pr_info("%s: func=\"%s\" file=\"%s\" module=\"%s\" format=\"%.*s\" lineno=%u-%u class=%u\n",
+	v3pr_info("%s: func=\"%s\" file=\"%s\" module=\"%s\" format=\"%.*s\" lineno=%u-%u class=%s\n",
 		 msg,
 		 query->function ?: "",
 		 query->filename ?: "",
 		 query->module ?: "",
 		 fmtlen, query->format ?: "",
-		 query->first_lineno, query->last_lineno, query->class_id);
+		 query->first_lineno, query->last_lineno, query->class_string);
 }

 /*
@@ -158,7 +165,7 @@ static int ddebug_change(const struct ddebug_query *query,
 	struct ddebug_table *dt;
 	unsigned int newflags;
 	unsigned int nfound = 0;
-	struct flagsbuf fbuf;
+	struct flagsbuf fbuf, nbuf;

 	/* search for matching ddebugs */
 	mutex_lock(&ddebug_lock);
@@ -172,8 +179,8 @@ static int ddebug_change(const struct ddebug_query *query,
 		for (i = 0; i < dt->num_ddebugs; i++) {
 			struct _ddebug *dp = &dt->ddebugs[i];

-			/* match against the class_id, either given or default */
-			if (query->class_id != dp->class_id)
+			/* if class is not set fall through, ot check if it matches */
+			if (query->class_string && strcmp(query->class_string, cat_to_strings[dp->class_id]))
 				continue;

 			/* match against the source filename */
@@ -223,11 +230,12 @@ static int ddebug_change(const struct ddebug_query *query,
 				static_branch_enable(&dp->key.dd_key_true);
 			}
 #endif
+			v4pr_info("changed %s:%d [%s]%s %s -> %s\n",
+				  trim_prefix(dp->filename), dp->lineno,
+				  dt->mod_name, dp->function,
+				  ddebug_describe_flags(dp->flags, &fbuf),
+				  ddebug_describe_flags(newflags, &nbuf));
 			dp->flags = newflags;
-			v4pr_info("changed %s:%d [%s]%s =%s\n",
-				 trim_prefix(dp->filename), dp->lineno,
-				 dt->mod_name, dp->function,
-				 ddebug_describe_flags(dp->flags, &fbuf));
 		}
 	}
 	mutex_unlock(&ddebug_lock);
@@ -314,21 +322,6 @@ static inline int parse_lineno(const char *str, unsigned int *val)
 	return 0;
 }

-static inline int parse_class(struct ddebug_query *query, const char *str)
-{
-	int rc;
-	unsigned int val;
-
-	rc = kstrtouint(str, 10, &val);
-	if (rc < 0 || val > _DPRINTK_SITE_UNCLASSED) {
-		pr_err("expecting class:[0-%d], not %s\n", _DPRINTK_SITE_UNCLASSED, str);
-		return -EINVAL;
-	}
-	query->class_id = val;
-	query->class_marked = 1;
-	return 0;
-}
-
 static int parse_linerange(struct ddebug_query *query, const char *first)
 {
 	char *last = strchr(first, '-');
@@ -443,8 +436,7 @@ static int ddebug_parse_query(char *words[], int nwords,
 			if (parse_linerange(query, arg))
 				return -EINVAL;
 		} else if (!strcmp(keyword, "class")) {
-			if (parse_class(query, arg))
-				return -EINVAL;
+			rc = check_set(&query->class_string, arg, "class");
 		} else {
 			pr_err("unknown keyword \"%s\"\n", keyword);
 			return -EINVAL;
@@ -452,9 +444,6 @@ static int ddebug_parse_query(char *words[], int nwords,
 		if (rc)
 			return rc;
 	}
-	/* post-validate the query, set default */
-	if (!query->class_marked)
-		query->class_id = _DPRINTK_SITE_UNCLASSED;

 	vpr_info_dq(query, "parsed");
 	return 0;
@@ -616,6 +605,8 @@ int param_set_dyndbg_classbits(const char *instr, const struct kernel_param *kp)
 	}
 	vpr_info("set_dyndbg_classbits: new 0x%lx old 0x%lx\n", inbits, *dcp->bits);

+
+/*
 	for (i = 0; (i < _DPRINTK_SITE_UNCLASSED &&
 		     dcp->classes[i] < _DPRINTK_SITE_UNCLASSED); i++) {

@@ -630,6 +621,7 @@ int param_set_dyndbg_classbits(const char *instr, const struct kernel_param *kp)
 			  matches, dcp->classes[i]);
 		totct += matches;
 	}
+*/
 	*dcp->bits = inbits;
 	vpr_info("total matches: %d\n", totct);
 	return 0;
@@ -978,8 +970,7 @@ static int ddebug_proc_show(struct seq_file *m, void *p)
 	seq_escape(m, dp->format, "\t\r\n\"");
 	seq_puts(m, "\"");

-	if (dp->class_id != _DPRINTK_SITE_UNCLASSED)
-		seq_printf(m, " cls:%u", dp->class_id);
+	seq_printf(m, " cat:%s", cat_to_strings[dp->class_id]);
 	seq_puts(m, "\n");

 	return 0;

^ permalink raw reply related	[flat|nested] 11+ messages in thread

* Re: [PATCH 2/5] dyndbg: add class_id field and query support
  2022-03-14 21:29       ` Jason Baron
@ 2022-03-28 19:07         ` jim.cromie
  0 siblings, 0 replies; 11+ messages in thread
From: jim.cromie @ 2022-03-28 19:07 UTC (permalink / raw)
  To: Jason Baron
  Cc: Greg KH, LKML, Daniel Vetter, Sean Paul, robdclark,
	Rasmus Villemoes, Joe Perches, dri-devel, amd-gfx mailing list,
	intel-gvt-dev, Intel Graphics Development

On Mon, Mar 14, 2022 at 3:30 PM Jason Baron <jbaron@akamai.com> wrote:
>
>
>
> On 3/11/22 20:06, jim.cromie@gmail.com wrote:
> > On Fri, Mar 11, 2022 at 12:06 PM Jason Baron <jbaron@akamai.com> wrote:
> >>
> >>
> >>
> >> On 3/10/22 23:47, Jim Cromie wrote:
> >>>
> >>> With the patch, one can enable current/unclassed callsites by:
> >>>
> >>>   #> echo drm class 15 +p > /proc/dynamic_debug/control
> >>>
> >>
> >> To me, this is hard to read, what the heck is '15'? I have to go look it
> >> up in the control file and it's not descriptive. I think that using
> >> classes/categories makes sense but I'm wondering if it can be a bit more
> >> user friendly? Perhaps, we can pass an array of strings that is indexed
> >> by the class id to each pr_debug() site that wants to use class. So
> >> something like:

hi Jason,
Im now in basically full agreement with you

1.   .class_id  is a "global" space, in that all callsites have one.
2.    0-15 is an exceedingly small range for a global space

Fix that by
3. make it private (by removing "class N" parsing)
   now nobody can do
   echo module * class N +p >control

4. add private/per-module "string" -> class_id map
    each module will have to declare the class-strings they use/accept
    opt-in - want coordinated / shared names for DRM_UT_KMS etc.

5. validate input using the known class_string -> class_id

then, this is knowably right or wrong, depending on DRM_FOO:
     echo module drm class DRM_FOO +p > control

it also means that
     echo class DRM_UT_KMS +p >control
is both wellformed and minimal;
any module that has DRM_UT_KMS defined will know which class_id *they*
have mapped it to.
theres no global "DRM_UT_KMS" to be abused.

So Ive been working towards that..
Heres my current biggest roadblock

DEFINE_DYNAMIC_DEBUG_CLASSBITS
creates the class-strings[] array declaratively, at compile-time.
This array is attached to the kernel-param.args,
so it can be used by the callbacks (to construct the command)

But its not obviously available from outside the sysfs knob
that its attached to, as is needed to apply command >control generally.

If I can attach the class-strings[]  to struct ddebug_table,
then ddebug_change() can use it to validate >control input.

So, how to attach ?
its got to work for both builtin & loadable modules.
(which precludes something in struct module ?)

I looked for a kernel_param_register callback, came up empty.
Trying to add it feels invasive / imposing.


> >
> > If that works, its cuz DEFINE_DYNAMIC_DEBUG_CLASSBITS()
> > plucks class symbols out of its __VA_ARGS__, and #stringifes them.
> > So that macro could then build the 1-per-module static constant string array
> > and (only) the callbacks would be able to use it.
> >

So Ive been tinkering hard on this macro, since its pretty central to
the interface defn.
heres some choices:

this is what Ive been working towards.
using enum symbols directly like this associates them by grep,
in contrast, class-strings are mealy-mouthed, milquetoast.

DEFINE_DYNAMIC_DEBUG_CLASSBITS(debug, __drm_debug, "p",
        "enable drm.debug categories - 1 bit per category",
        DRM_UT_CORE,
        DRM_UT_DRIVER,
        DRM_UT_KMS,
        DRM_UT_PRIME,
        DRM_UT_ATOMIC,
        DRM_UT_VBL,
        DRM_UT_STATE,
        DRM_UT_LEASE,
        DRM_UT_DP,
        DRM_UT_DRMRES);

 I found a slick MAP ( ) macro to do this:

#define DEFINE_DYNAMIC_DEBUG_CLASSBITS(fsname, _var, _flgs, desc, ...) \
  MODULE_PARM_DESC(fsname, desc); \
  static struct dyndbg_classbits_param ddcats_##_var = { \
    .bits = &(_var), \
    .flags = _flgs, \
    .classes = { __VA_ARGS__, _DPRINTK_CLASS_DFLT }, \
    .class_names = { mgMAP(__stringify, sCOMMA, \
                                                __VA_ARGS__,
_DPRINTK_CLASS_DFLT) } \
}; \
module_param_cb(fsname, &param_ops_dyndbg_classbits, \
&ddcats_##_var, 0644)

 __VA_ARGS__   is used 2x
.class_names is available for validating command >control

As much as I like the above, the MAP macro has a longer, more risky
path to the kernel

so a more modest alternative: module user defines class-strings in interface,
but they *must* align manually with the enum values they correspond to;
the order determines the bit-pos in the sysfs node,
since the interface no longer provides the enum values themselves.

DEFINE_DYNAMIC_DEBUG_CLASS_STRINGS(debug, __drm_debug, "p",
        "enable drm.debug categories - 1 bit per category",
        "DRM_UT_CORE",
        "DRM_UT_DRIVER",
        "DRM_UT_KMS",

different name allows CLASSBITs or similar in future, if MAP works out.
class-strings are completely defined by users, drm can drop UT naming

TLDR: FWIW

iSTM that  the same macro will support the coordinated use of class-strings
across multiple modules.

drm_print.c - natural home for exposed sysfs node

amdgpu/, drm_helper/ i915/  nouveau/  will all need a DEFINE_DD added,
so that ddebug_change() can allow those .class_ids to be controlled.
sysfs perm inits can disable their nodes, since theyre coordinated.

> >
>
> Ok, yeah so I guess I'm thinking about the 'class' more as global namespace,
> so that for example, it could span modules, or be specific to certain modules.
> I'm also thinking of it as a 'string' which is maybe hierarchical, so that it's
> clear what sub-system, or sub-sub-system it belongs to. So for DRM for example,
> it could be a string like "DRM:CORE". The index num I think is still helpful for
> implementation so we don't have to store a pointer size, but I don't think it's
> really exposed (except perhaps in module params b/c drm is doing that already?).
>

So what Ive got here is as described above,
I just need a few bright ideas,
then I can bring it together.
got a spare tuit?

Jim

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2022-03-28 19:08 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-03-11  4:47 [PATCH 0/5] dyndbg add exclusive class support Jim Cromie
2022-03-11  4:47 ` [PATCH 1/5] dyndbg: fix static_branch manipulation Jim Cromie
2022-03-11 18:03   ` Jason Baron
2022-03-11  4:47 ` [PATCH 2/5] dyndbg: add class_id field and query support Jim Cromie
2022-03-11 19:06   ` Jason Baron
2022-03-12  1:06     ` jim.cromie
2022-03-14 21:29       ` Jason Baron
2022-03-28 19:07         ` jim.cromie
2022-03-11  4:47 ` [PATCH 3/5] dyndbg: add DEFINE_DYNAMIC_DEBUG_CLASSBITS macro Jim Cromie
2022-03-11  4:47 ` [PATCH 4/5] dyndbg: drop EXPORTed dynamic_debug_exec_queries Jim Cromie
2022-03-11  4:47 ` [PATCH 5/5] dyndbg: show both old and new in change-info Jim Cromie

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).