From: Jim Cromie <jim.cromie@gmail.com>
To: linux-kernel@vger.kernel.org, jbaron@akamai.com,
gregkh@linuxfoundation.org, ukaszb@chromium.org
Cc: intel-gfx-trybot@lists.freedesktop.org,
dri-devel@lists.freedesktop.org, amd-gfx@lists.freedesktop.org,
intel-gvt-dev@lists.freedesktop.org,
intel-gfx@lists.freedesktop.org, daniel.vetter@ffwll.ch,
tvrtko.ursulin@linux.intel.com, jani.nikula@intel.com,
ville.syrjala@linux.intel.com, Jim Cromie <jim.cromie@gmail.com>
Subject: [PATCH 38/63] dyndbg: drop "protection" of class'd pr_debugs from legacy queries
Date: Fri, 24 Jan 2025 23:45:52 -0700 [thread overview]
Message-ID: <20250125064619.8305-39-jim.cromie@gmail.com> (raw)
In-Reply-To: <20250125064619.8305-1-jim.cromie@gmail.com>
Current classmap code protects class'd pr_debugs from unintended
changes by "legacy" unclassed queries:
# this doesn't disable all of DRM_UT_* categories
echo "-p" > /proc/dynamic_debug/control
# name the class to change it - protective but tedious
echo "class DRM_UT_CORE +p" > /proc/dynamic_debug/control
# or do it the subsystem way
echo 1 > /sys/module/drm/parameters/debug
This "name the class to change it" behavior gave a modicum of
protection to classmap users (ie DRM) so their debug settings aren't
trivially and unintentionally altered underneath them. Afterall,
__drm_debug is authoritative w/o dyndbg under it.
But this made the class keyword special in some sense; the other
keywords follow the rule: no "keyword value" given means no skipping
on that criterion.
Jason Baron didn't like this special case when I 1st proposed it;
I argued 2 points:
- "protection gives stable-debug, improving utility"
- "class _DFLT" assumed in query is not that special
I thought I'd convinced him back then, (and the patchset got merged),
but he noted it again when he reviewed this series. So this commit
undoes the special case, indirectly.
The difference in behavior comes down to a choice on how to handle
class-mismatches; ddebug_client_module_protects_classes() names this
choice (and defines it false). On class-mismatches of this kind, it
is "called" before skipping the legacy cmd on a class'd site.
So this makes it a policy-choice, and reverts to original policy.
Later, if any user/module wants to protect its classes, we could add a
flag to ddebug_table, a means to set it from CLASSMAP_DEFINE, and
check it when applying a classless query/cmd. Further, runtime mods
to the protection are possible via an exported fn, or a grammar
addition to >control.
NOTES
In ddebug_change(), the 2-part class code is mostly unchanged:
1. query->class_str is validated once per module, and maps good ones
to valid_class, now renamed to slctd_class to convey non-boolean-ness.
2. in the inner per-site loop, site mismatches are now 2 sub-cases:
a. a class_str given explicitly, and site doesn't match it.
skip/continue.
b. no class_str given, _CLASS_DFLT is inferred
skip if ddebug_client_module_protects_classes()
else fall thru to site-change
CC: jbaron@akamai.com
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
lib/dynamic_debug.c | 34 +++++++++++++++++++++++++---------
1 file changed, 25 insertions(+), 9 deletions(-)
diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index 68b5a77c2d79..9a278c7ca365 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -193,6 +193,17 @@ static int ddebug_find_valid_class(struct ddebug_table const *dt, const char *cl
return -ENOENT;
}
+/*
+ * classmaps-v1 protected classes from changes by legacy commands
+ * (those selecting _DPRINTK_CLASS_DFLT by omission), v2 undoes that
+ * special treatment. State so explicitly. Later we could give
+ * modules the choice to protect their classes or to keep v2 behavior.
+ */
+static inline bool ddebug_client_module_protects_classes(const struct ddebug_table *dt)
+{
+ return false;
+}
+
/*
* Search the tables for _ddebug's which match the given `query' and
* apply the `flags' and `mask' to them. Returns number of matching
@@ -206,7 +217,7 @@ static int ddebug_change(const struct ddebug_query *query, struct flag_settings
unsigned int newflags;
unsigned int nfound = 0;
struct flagsbuf fbuf, nbuf;
- int valid_class;
+ int slctd_class;
/* search for matching ddebugs */
mutex_lock(&ddebug_lock);
@@ -218,21 +229,26 @@ static int ddebug_change(const struct ddebug_query *query, struct flag_settings
continue;
if (query->class_string) {
- valid_class = ddebug_find_valid_class(dt, query->class_string);
- if (valid_class < 0)
+ slctd_class = ddebug_find_valid_class(dt, query->class_string);
+ if (slctd_class < 0)
+ /* skip/reject classes unknown by module */
continue;
} else {
- /* constrain query, do not touch class'd callsites */
- valid_class = _DPRINTK_CLASS_DFLT;
+ slctd_class = _DPRINTK_CLASS_DFLT;
}
for (i = 0; i < dt->info.descs.len; i++) {
struct _ddebug *dp = &dt->info.descs.start[i];
- /* match site against query-class */
- if (dp->class_id != valid_class)
- continue;
-
+ if (dp->class_id != slctd_class) {
+ if (query->class_string)
+ /* site.class != given class */
+ continue;
+ /* legacy query, class'd site */
+ else if (ddebug_client_module_protects_classes(dt))
+ continue;
+ /* allow change on class'd pr_debug */
+ }
/* match against the source filename */
if (query->filename &&
!match_wildcard(query->filename, dp->filename) &&
--
2.48.1
next prev parent reply other threads:[~2025-01-25 6:47 UTC|newest]
Thread overview: 103+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-01-25 6:45 [PATCH 00/63] Fix CONFIG_DRM_USE_DYNAMIC_DEBUG=y Jim Cromie
2025-01-25 6:45 ` [PATCH 01/63] docs/dyndbg: update examples \012 to \n Jim Cromie
2025-02-25 14:30 ` Louis Chauvet
2025-03-16 13:50 ` jim.cromie
2025-01-25 6:45 ` [PATCH 02/63] test-dyndbg: fixup CLASSMAP usage error Jim Cromie
2025-02-25 11:22 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 03/63] dyndbg: reword "class unknown," to "class:_UNKNOWN_" Jim Cromie
2025-02-25 11:22 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 04/63] dyndbg: make ddebug_class_param union members same size Jim Cromie
2025-02-25 11:24 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 05/63] dyndbg: replace classmap list with a vector Jim Cromie
2025-02-25 14:08 ` Louis Chauvet
2025-03-16 14:04 ` jim.cromie
2025-01-25 6:45 ` [PATCH 06/63] dyndbg: ddebug_apply_class_bitmap - add module arg, select on it Jim Cromie
2025-02-25 14:09 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 07/63] dyndbg: split param_set_dyndbg_classes to _module & wrapper fns Jim Cromie
2025-02-25 14:09 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 08/63] dyndbg: drop NUM_TYPE_ARRAY Jim Cromie
2025-02-25 14:09 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 09/63] dyndbg: reduce verbose/debug clutter Jim Cromie
2025-02-25 14:10 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 10/63] dyndbg: silence debugs with no-change updates Jim Cromie
2025-02-25 14:10 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 11/63] dyndbg: tighten ddebug_class_name() 1st arg type Jim Cromie
2025-02-25 14:11 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 12/63] dyndbg: tighten fn-sig of ddebug_apply_class_bitmap Jim Cromie
2025-02-25 14:12 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 13/63] dyndbg: reduce verbose=3 messages in ddebug_add_module Jim Cromie
2025-02-25 14:12 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 14/63] dyndbg-API: remove DD_CLASS_TYPE_(DISJOINT|LEVEL)_NAMES and code Jim Cromie
2025-01-25 6:45 ` [PATCH 15/63] checkpatch: add an exception to the do-while wrapper advice Jim Cromie
2025-02-25 14:12 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 16/63] dyndbg-API: replace DECLARE_DYNDBG_CLASSMAP Jim Cromie
2025-02-25 14:16 ` Louis Chauvet
2025-03-16 19:46 ` jim.cromie
2025-03-24 15:07 ` Louis Chauvet
2025-03-25 16:03 ` jim.cromie
2025-01-25 6:45 ` [PATCH 17/63] dyndbg: check DYNDBG_CLASSMAP_DEFINE args at compile-time Jim Cromie
2025-02-25 14:17 ` Louis Chauvet
2025-03-16 20:46 ` jim.cromie
2025-01-25 6:45 ` [PATCH 18/63] dyndbg: add/use for_subvec() to reduce boilerplate Jim Cromie
2025-02-25 14:18 ` Louis Chauvet
2025-03-16 20:50 ` jim.cromie
2025-01-25 6:45 ` [PATCH 19/63] dyndbg: make proper substructs in _ddebug_info Jim Cromie
2025-02-25 14:19 ` Louis Chauvet
2025-03-16 20:58 ` jim.cromie
2025-01-25 6:45 ` [PATCH 20/63] dyndbg: drop premature optimization in ddebug_add_module Jim Cromie
2025-02-25 14:26 ` Louis Chauvet
2025-01-25 6:45 ` [PATCH 21/63] dyndbg: allow ddebug_add_module to fail Jim Cromie
2025-02-25 14:26 ` Louis Chauvet
2025-03-16 21:04 ` jim.cromie
2025-01-25 6:45 ` [PATCH 22/63] dyndbg: rework ddebug_attach_*module_classes() Jim Cromie
2025-01-25 6:45 ` [PATCH 23/63] dyndbg: fail modprobe on ddebug_class_range_overlap() Jim Cromie
2025-02-25 14:27 ` Louis Chauvet
2025-03-16 21:07 ` jim.cromie
2025-01-25 6:45 ` [PATCH 24/63] dyndbg: hoist the range-overlap checks Jim Cromie
2025-01-25 6:45 ` [PATCH 25/63] ddebug: cleanup-range-overlap fails Jim Cromie
2025-01-25 6:45 ` [PATCH 26/63] dyndbg-test: change do_prints testpoint to accept a loopct Jim Cromie
2025-01-25 6:45 ` [PATCH 27/63] selftests-dyndbg: add tools/testing/selftests/dynamic_debug/* Jim Cromie
2025-01-25 6:45 ` [PATCH 28/63] dyndbg-API: promote DYNDBG_CLASSMAP_PARAM to API Jim Cromie
2025-02-25 14:29 ` Louis Chauvet
2025-03-16 21:14 ` jim.cromie
2025-03-24 22:58 ` jim.cromie
2025-01-25 6:45 ` [PATCH 29/63] dyndbg-doc: add classmap info to howto Jim Cromie
2025-01-25 6:45 ` [PATCH 30/63] dyndbg: treat comma as a token separator Jim Cromie
2025-01-25 6:45 ` [PATCH 31/63] selftests-dyndbg: add comma_terminator_tests Jim Cromie
2025-01-25 6:45 ` [PATCH 32/63] dyndbg: split multi-query strings with % Jim Cromie
2025-01-25 6:45 ` [PATCH 33/63] selftests-dyndbg: test_percent_splitting Jim Cromie
2025-01-25 6:45 ` [PATCH 34/63] docs/dyndbg: explain new delimiters: comma, percent Jim Cromie
2025-01-25 6:45 ` [PATCH 35/63] selftests-dyndbg: add test_mod_submod Jim Cromie
2025-01-25 6:45 ` [PATCH 36/63] docs/dyndbg: explain flags parse 1st Jim Cromie
2025-01-25 6:45 ` [PATCH 37/63] dyndbg: change __dynamic_func_call_cls* macros into expressions Jim Cromie
2025-01-25 6:45 ` Jim Cromie [this message]
2025-01-25 6:45 ` [PATCH 39/63] drm: use correct ccflags-y spelling Jim Cromie
2025-01-25 6:45 ` [PATCH 40/63] checkpatch: dont warn about unused macro arg on empty body Jim Cromie
2025-01-25 6:45 ` [PATCH 41/63] drm-dyndbg: adapt drm core to use dyndbg classmaps-v2 Jim Cromie
2025-01-25 6:45 ` [PATCH 42/63] drm-dyndbg: adapt DRM to invoke DYNDBG_CLASSMAP_PARAM Jim Cromie
2025-01-25 6:45 ` [PATCH 43/63] drm-print: fix config-dependent unused variable Jim Cromie
2025-01-25 6:45 ` [PATCH 44/63] drm-dyndbg: DRM_CLASSMAP_USE in amdgpu driver Jim Cromie
2025-01-25 6:45 ` [PATCH 45/63] drm-dyndbg: DRM_CLASSMAP_USE in i915 driver Jim Cromie
2025-01-25 6:46 ` [PATCH 46/63] drm-dyndbg: DRM_CLASSMAP_USE in drm_crtc_helper Jim Cromie
2025-01-25 6:46 ` [PATCH 47/63] drm-dyndbg: DRM_CLASSMAP_USE in drm_dp_helper Jim Cromie
2025-01-25 6:46 ` [PATCH 48/63] drm-dyndbg: DRM_CLASSMAP_USE in nouveau Jim Cromie
2025-01-25 6:46 ` [PATCH 49/63] drm-dyndbg: add DRM_CLASSMAP_USE to Xe driver Jim Cromie
2025-01-25 6:46 ` [PATCH 50/63] drm-dyndbg: add DRM_CLASSMAP_USE to virtio_gpu Jim Cromie
2025-01-25 6:46 ` [PATCH 51/63] drm-dyndbg: add DRM_CLASSMAP_USE to simpledrm Jim Cromie
2025-01-25 6:46 ` [PATCH 52/63] drm-dyndbg: add DRM_CLASSMAP_USE to bochs Jim Cromie
2025-01-25 6:46 ` [PATCH 53/63] drm-dyndbg: add DRM_CLASSMAP_USE to etnaviv Jim Cromie
2025-01-25 6:46 ` [PATCH 54/63] drm-dyndbg: add DRM_CLASSMAP_USE to gma500 driver Jim Cromie
2025-01-25 6:46 ` [PATCH 55/63] drm-dyndbg: add DRM_CLASSMAP_USE to radeon Jim Cromie
2025-01-25 6:46 ` [PATCH 56/63] drm-dyndbg: add DRM_CLASSMAP_USE to vmwgfx driver Jim Cromie
2025-01-25 6:46 ` [PATCH 57/63] drm-dyndbg: add DRM_CLASSMAP_USE to vkms driver Jim Cromie
2025-01-25 6:46 ` [PATCH 58/63] drm-dyndbg: add DRM_CLASSMAP_USE to udl driver Jim Cromie
2025-01-25 6:46 ` [PATCH 59/63] drm-dyndbg: add DRM_CLASSMAP_USE to mgag200 driver Jim Cromie
2025-01-25 6:46 ` [PATCH 60/63] drm-dyndbg: add DRM_CLASSMAP_USE to the gud driver Jim Cromie
2025-01-25 6:46 ` [PATCH 61/63] drm-dyndbg: add DRM_CLASSMAP_USE to the qxl driver Jim Cromie
2025-01-25 6:46 ` [PATCH 62/63] drm-dyndbg: add DRM_CLASSMAP_USE to the drm_gem_shmem_helper driver Jim Cromie
2025-01-25 6:46 ` [PATCH 63/63] drm: restore CONFIG_DRM_USE_DYNAMIC_DEBUG un-BROKEN Jim Cromie
2025-01-25 6:46 ` [PATCH 64/64] This series fixes dyndbg (dynamic debug) classmap support for DRM Jim Cromie
2025-02-20 8:31 ` [PATCH 00/63] Fix CONFIG_DRM_USE_DYNAMIC_DEBUG=y Greg KH
2025-02-20 9:45 ` Simona Vetter
2025-02-28 16:24 ` Louis Chauvet
2025-03-12 16:26 ` jim.cromie
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250125064619.8305-39-jim.cromie@gmail.com \
--to=jim.cromie@gmail.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=daniel.vetter@ffwll.ch \
--cc=dri-devel@lists.freedesktop.org \
--cc=gregkh@linuxfoundation.org \
--cc=intel-gfx-trybot@lists.freedesktop.org \
--cc=intel-gfx@lists.freedesktop.org \
--cc=intel-gvt-dev@lists.freedesktop.org \
--cc=jani.nikula@intel.com \
--cc=jbaron@akamai.com \
--cc=linux-kernel@vger.kernel.org \
--cc=tvrtko.ursulin@linux.intel.com \
--cc=ukaszb@chromium.org \
--cc=ville.syrjala@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).