From: Jim Cromie via B4 Relay <devnull+jim.cromie.gmail.com@kernel.org>
To: Jason Baron <jbaron@akamai.com>, Shuah Khan <shuah@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
Maxime Ripard <mripard@kernel.org>,
Thomas Zimmermann <tzimmermann@suse.de>,
David Airlie <airlied@gmail.com>,
Simona Vetter <simona@ffwll.ch>, Arnd Bergmann <arnd@arndb.de>,
Luis Chamberlain <mcgrof@kernel.org>,
Petr Pavlu <petr.pavlu@suse.com>,
Daniel Gomez <da.gomez@kernel.org>,
Sami Tolvanen <samitolvanen@google.com>,
Aaron Tomlin <atomlin@atomlin.com>,
Jonathan Corbet <corbet@net.net>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Nathan Chancellor <nathan@kernel.org>,
Nicolas Schier <nsc@kernel.org>,
Shuah Khan <skhan@linuxfoundation.org>,
Randy Dunlap <rdunlap@infradead.org>,
"Rafael J. Wysocki" <rafael@kernel.org>,
Pavel Machek <pavel@kernel.org>, Len Brown <lenb@kernel.org>
Cc: linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
dri-devel@lists.freedesktop.org, linux-arch@vger.kernel.org,
linux-modules@vger.kernel.org, linux-doc@vger.kernel.org,
linux-kbuild@vger.kernel.org, linux-pm@vger.kernel.org,
Jim Cromie <jim.cromie@gmail.com>
Subject: [PATCH v10 33/38] dyndbg: harden classmap and descriptor validation
Date: Wed, 16 Sep 2026 09:33:25 -0600 [thread overview]
Message-ID: <20260916-dd-cmap-part2-clean-v10-33-af4cf4767707@gmail.com> (raw)
In-Reply-To: <20260916-dd-cmap-part2-clean-v10-0-af4cf4767707@gmail.com>
From: Jim Cromie <jim.cromie@gmail.com>
Dynamic debug classmaps allow modules to _DEFINE and/or _USE multiple
classmaps, but this requires coordination amongst the classmaps.
Previously, class validation done by DYNAMIC_DEBUG_CLASSMAP_DEFINE at
compile-time, and ddebug_class_range_overlap() at modprobe-time, was
incomplete, and DYNAMIC_DEBUG_CLASSMAP_USE_ had no validation. This
could allow broken classmaps, making them harder to use well.
This commit improves classmap and descriptor validation:
- Add run-time overlap checks for _USEd classmaps in ddebug_add_module()
to prevent collisions between private maps and imported APIs.
- Scan module descriptors at load time to print a single warning per
missing class_id, rather than waiting for a user query to trip over it.
- Downgrade the global WARN_ONCE in ddebug_match_desc() to a
pr_warn_ratelimited, since orphaned class IDs are now tracked and
warned about early at module load.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
v10:
. fix pr_warn_ratelimited() line wrap in ddebug_match_desc().
---
lib/dynamic_debug.c | 68 ++++++++++++++++++++--
lib/test_dynamic_debug.c | 16 +++--
.../selftests/dynamic_debug/dyndbg_selftest.sh | 22 +++----
3 files changed, 84 insertions(+), 22 deletions(-)
diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
index ac54a0499369..eb28831a6cb2 100644
--- a/lib/dynamic_debug.c
+++ b/lib/dynamic_debug.c
@@ -317,7 +317,8 @@ static bool ddebug_match_desc(const struct ddebug_query *query,
/* site is class'd */
site_map = ddebug_find_map_by_class_id(di, dp->class_id);
if (!site_map) {
- WARN_ONCE(1, "unknown class_id %d, check %s's CLASSMAP definitions", dp->class_id, di->mod_name);
+ pr_warn_ratelimited("unknown class_id %d, check %s's CLASSMAP definitions\n",
+ dp->class_id, di->mod_name);
return false;
}
/* module(-param) decides protection */
@@ -1490,6 +1491,23 @@ static int ddebug_class_range_overlap(struct ddebug_class_map *cm, u64 *reserved
return 0;
}
+static int ddebug_class_user_overlap(struct ddebug_class_user *cli,
+ u64 *reserved_ids)
+{
+ struct ddebug_class_map *cm = cli->map;
+ int base = cm->base + cli->offset;
+ u64 range = (((1ULL << cm->length) - 1) << base);
+
+ if (range & *reserved_ids) {
+ pr_err("module %s: [%d..%d] (from %s) conflicts with %llx\n",
+ cli->mod_name, base, base + cm->length - 1,
+ cm->class_names[0], *reserved_ids);
+ return -EINVAL;
+ }
+ *reserved_ids |= range;
+ return 0;
+}
+
/*
* Allocate a new ddebug_table for the given module
* and add it to the global list.
@@ -1500,7 +1518,8 @@ static int ddebug_add_module(struct _ddebug_info *di)
struct ddebug_class_map *cm;
struct ddebug_class_user *cli;
u64 reserved_ids = 0;
- int i;
+ u64 bad_ids = 0;
+ int i, err = 0;
if (!di->descs.len)
return 0;
@@ -1531,10 +1550,47 @@ static int ddebug_add_module(struct _ddebug_info *di)
dd_set_module_subrange(i, cm, &dt->info, maps);
dd_set_module_subrange(i, cli, &dt->info, users);
- /* insure 2+ classmaps share the per-module 0..62 class_id space */
+ /* validate the per-module shared 0..62 class_id space */
for_subvec(i, cm, &dt->info, maps)
if (ddebug_class_range_overlap(cm, &reserved_ids))
- goto cleanup;
+ err = -EINVAL;
+
+ for_subvec(i, cli, &dt->info, users) {
+ cm = cli->map;
+ if (!cm) {
+ pr_err("module %s: classmap not found for user\n", di->mod_name);
+ err = -EINVAL;
+ continue;
+ }
+
+ if (cm->base + cm->length + cli->offset > _DPRINTK_CLASS_DFLT) {
+ pr_err("module %s: base:%d + classes.len:%d + cli.offset:%d must be <= %d\n",
+ di->mod_name, cm->base, cm->length,
+ cli->offset, _DPRINTK_CLASS_DFLT);
+ err = -EINVAL;
+ continue;
+ }
+
+ if (ddebug_class_user_overlap(cli, &reserved_ids))
+ err = -EINVAL;
+ }
+ if (err)
+ goto cleanup;
+
+ /* validate all class_ids against module's classmaps/users */
+ for (i = 0; i < dt->info.descs.len; i++) {
+ struct _ddebug *dp = &dt->info.descs.start[i];
+
+ if (dp->class_id == _DPRINTK_CLASS_DFLT)
+ continue;
+ if (bad_ids & (1ULL << dp->class_id))
+ continue;
+ if (!ddebug_find_map_by_class_id(&dt->info, dp->class_id)) {
+ pr_warn("module %s uses unknown class_id %d\n",
+ dt->info.mod_name, dp->class_id);
+ bad_ids |= (1ULL << dp->class_id);
+ }
+ }
mutex_lock(&ddebug_lock);
list_add_tail(&dt->link, &ddebug_tables);
@@ -1546,7 +1602,7 @@ static int ddebug_add_module(struct _ddebug_info *di)
dt->info.descs.len, dt->info.mod_name);
return 0;
cleanup:
- WARN_ONCE(1, "dyndbg multi-classmap conflict in %s\n", di->mod_name);
+ pr_err("dyndbg multi-classmap conflict in %s\n", di->mod_name);
kfree(dt);
return -EINVAL;
}
@@ -1633,7 +1689,7 @@ static int ddebug_module_notify(struct notifier_block *self, unsigned long val,
mod->dyndbg_info.mod_name = mod->name;
ret = ddebug_add_module(&mod->dyndbg_info);
if (ret)
- WARN(1, "Failed to allocate memory: dyndbg may not work properly.\n");
+ pr_err("dyndbg: failed to add module %s: %d\n", mod->name, ret);
break;
case MODULE_STATE_GOING:
ddebug_remove_module(mod->name);
diff --git a/lib/test_dynamic_debug.c b/lib/test_dynamic_debug.c
index 7c10a1ddc630..fe5b885398f3 100644
--- a/lib/test_dynamic_debug.c
+++ b/lib/test_dynamic_debug.c
@@ -161,14 +161,20 @@ DYNAMIC_DEBUG_CLASSMAP_DEFINE(fail_base_len, 0, 60,
#endif
#else /* TEST_DYNAMIC_DEBUG_SUBMOD */
-
/*
- * in submod/drm-drivers, use the classmaps defined in top/parent
- * module above.
+ * In submod (drm-drivers/helpers) use the classmaps defined in
+ * top/parent module above. We _USE_() with offset, to test the
+ * non-zero case.
*/
-
DYNAMIC_DEBUG_CLASSMAP_USE(map_disjoint_bits);
-DYNAMIC_DEBUG_CLASSMAP_USE_(map_level_num, 7);
+/*
+ * maybe force failure of runtime sanity test of classmap.length + offset < 63
+ */
+#if !defined(DD_RUNTIME_CLASS_CHECK)
+DYNAMIC_DEBUG_CLASSMAP_USE_(map_level_num, 8);
+#else
+DYNAMIC_DEBUG_CLASSMAP_USE_(map_level_num, 55);
+#endif
#if defined(DD_MACRO_ARGCHECK)
DYNAMIC_DEBUG_CLASSMAP_USE_(fail_offset_big, 100);
diff --git a/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh b/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh
index a91a0ffa92d4..ef5f41ca7f7d 100755
--- a/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh
+++ b/tools/testing/selftests/dynamic_debug/dyndbg_selftest.sh
@@ -583,12 +583,12 @@ function GOLDEN_RECORDS {
#K= c518a50ba30ba8099d0dc874a27ecf16 FT_multi_query.4
#K= 8f18ea82c09460434b6e9e4cd12543e8 FT_test_classes.1
#K= a15ec4843acd721fbdfddc0b512c8032 FT_test_classes.2
-#K= 20d4545f9753e677e72e3adf52527fd3 FT_test_classes.3
-#K= 934d8677872fe26bd636a6c3d6416aa2 FT_classmap_inheritance.1
-#K= cd1389958807063baa1ea4b06c61fa02 FT_classmap_inheritance.2
+#K= bc074e657d8d7f0bbc960e4dfa8de25d FT_test_classes.3
+#K= a60284143b51b5d13c4562f7f4a2e807 FT_classmap_inheritance.1
+#K= 1d43471dca765597d88a1be25d776c92 FT_classmap_inheritance.2
#K= 0708a283f0f1959135c797e36119e4af FT_classmap_inheritance.3
-#K= 05f6efb80299d24cde65174f64d97308 FT_classmap_inheritance.4
-#K= 7e92245008439ee79fe2460aeaa16a9b FT_classmap_inheritance.5
+#K= a0b9a8a9530706b85c4f8bf911607db0 FT_classmap_inheritance.4
+#K= f43e0aff8a4b38435b73d90ed8100d1b FT_classmap_inheritance.5
#K= 53d1b6875b65c79cfd759e209ae50b11 FT_modprobe_w_param.1
#K= 53d1b6875b65c79cfd759e209ae50b11 FT_modprobe_w_param.2
#K= 79d912e2aeb04dea70f9c7e701333f0a FT_modprobe_w_param.3
@@ -599,11 +599,11 @@ function GOLDEN_RECORDS {
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.8
#K= f9cee4512e5604603e0262fba4bea223 FT_modprobe_w_param.9
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.10
-#K= b6a62433165f6423b81417f782551ab9 FT_modprobe_w_param.11
+#K= 7c645692e3e3796d4d3b8604c49583f7 FT_modprobe_w_param.11
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.12
-#K= de89753b843449d11da14240c4da4cad FT_modprobe_w_param.13
+#K= e995f70c0509334a33aec40ada9e6e6c FT_modprobe_w_param.13
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.14
-#K= 1d966d0cae735457791c93a86e9e1650 FT_modprobe_w_param.15
+#K= 72f97ca3e8f4c3e7f34fbe91de62e223 FT_modprobe_w_param.15
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.16
#K= 1aacb7c196a8354c46c60c73d78c6bf2 FT_modprobe_w_param.17
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.18
@@ -617,11 +617,11 @@ function GOLDEN_RECORDS {
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.26
#K= 7fa4c84490c42c750986614c3539d56a FT_modprobe_w_param.27
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.28
-#K= 1eb866a813551061cd73178ba7833543 FT_modprobe_w_param.29
+#K= fc098897aa3e1472a536e7ae8069c84f FT_modprobe_w_param.29
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.30
-#K= a9e7424ed7b02696b5e12108971792dd FT_modprobe_w_param.31
+#K= 38ea63fed62a0308241325e57020b28e FT_modprobe_w_param.31
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.32
-#K= 5323d7746d983cdcfda4866255cd5123 FT_modprobe_w_param.33
+#K= 4b896adb5e7cfcc79f9ffa879d791f37 FT_modprobe_w_param.33
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.34
#K= 7f961a7d3facd89bdd4b9d8e7b5541e9 FT_modprobe_w_param.35
#K= 6a320774e2b535ceb89d6cdde6f5d0fb FT_modprobe_w_param.36
--
2.55.0
next prev parent reply other threads:[~2026-09-16 15:33 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-16 15:32 [PATCH v10 00/38] dyndbg: Fix classmaps API for subsystems, query extensions, selftests Jim Cromie via B4 Relay
2026-09-16 15:32 ` [PATCH v10 01/38] selftests/dyndbg: Add kselftest script to verify dynamic-debug Jim Cromie via B4 Relay
2026-09-16 15:41 ` sashiko-bot
2026-09-16 15:32 ` [PATCH v10 02/38] vmlinux.lds.h: refactor BOUNDED_SECTION_* macros into bounded_sections.lds.h Jim Cromie via B4 Relay
2026-09-16 15:32 ` [PATCH v10 03/38] vmlinux.lds.h: drop unused HEADERED_SECTION* macros Jim Cromie via B4 Relay
2026-09-16 15:32 ` [PATCH v10 04/38] vmlinux.lds.h: Fix ALIGN(8) omission causing NULL ptr on i386 Jim Cromie via B4 Relay
2026-09-16 15:32 ` [PATCH v10 05/38] vmlinux.lds.h: remove redundant ALIGN(8) directives Jim Cromie via B4 Relay
2026-09-16 15:32 ` [PATCH v10 06/38] dyndbg.lds.S: fix lost dyndbg sections in modules Jim Cromie via B4 Relay
2026-09-16 15:32 ` [PATCH v10 07/38] dyndbg: factor ddebug_match_desc out from ddebug_change Jim Cromie via B4 Relay
2026-09-16 15:47 ` sashiko-bot
2026-09-16 15:33 ` [PATCH v10 08/38] dyndbg: add stub macro for DECLARE_DYNDBG_CLASSMAP Jim Cromie via B4 Relay
2026-09-16 16:00 ` sashiko-bot
2026-09-16 22:42 ` jim.cromie
2026-09-16 15:33 ` [PATCH v10 09/38] dyndbg: reword "class unknown," to "class:_UNKNOWN_" Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 10/38] dyndbg-API: remove DD_CLASS_TYPE_(DISJOINT|LEVEL)_NAMES and code Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 11/38] dyndbg: drop NUM_TYPE_ARGS Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 12/38] dyndbg: bump num-tokens in a query-cmd from 9 to 15 Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 13/38] dyndbg: reduce verbose/debug clutter Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 14/38] dyndbg: Bind callsites and classmaps to DDEBUG_MODNAME Jim Cromie via B4 Relay
2026-09-16 16:05 ` sashiko-bot
2026-09-16 15:33 ` [PATCH v10 15/38] dyndbg: refactor param_set_dyndbg_classes and below Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 16/38] dyndbg: tighten fn-sig of ddebug_apply_class_bitmap Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 17/38] dyndbg: replace classmap list with an array-slice Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 18/38] dyndbg: macrofy a 2-index for-loop pattern Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 19/38] dyndbg: reduce class param storage to u32 Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 20/38] dyndbg,module: make proper substructs in _ddebug_info Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 21/38] dyndbg: move mod_name down from struct ddebug_table to _ddebug_info Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 22/38] dyndbg: hoist classmap-filter-by-modname up to ddebug_add_module Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 23/38] dyndbg-API: replace DECLARE_DYNDBG_CLASSMAP Jim Cromie via B4 Relay
2026-09-16 16:28 ` sashiko-bot
2026-09-20 20:26 ` jim.cromie
2026-09-16 15:33 ` [PATCH v10 24/38] selftests/dyndbg: Enable FT_classmap_inheritance Jim Cromie via B4 Relay
2026-09-16 16:24 ` sashiko-bot
2026-09-20 17:25 ` jim.cromie
2026-09-16 15:33 ` [PATCH v10 25/38] dyndbg: detect class_id reservation conflicts Jim Cromie via B4 Relay
2026-09-16 16:30 ` sashiko-bot
2026-09-16 15:33 ` [PATCH v10 26/38] dyndbg: check DYNAMIC_DEBUG_CLASSMAP_{DEFINE,USE_} args at compile-time Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 27/38] dyndbg-test: add do_bulk testpoint, rename do_prints to do_classes Jim Cromie via B4 Relay
2026-09-16 16:37 ` sashiko-bot
2026-09-16 15:33 ` [PATCH v10 28/38] dyndbg-API: promote DYNAMIC_DEBUG_CLASSMAP_PARAM to API Jim Cromie via B4 Relay
2026-09-16 16:40 ` sashiko-bot
2026-09-16 15:33 ` [PATCH v10 29/38] dyndbg: control-parser: treat comma as a token separator Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 30/38] selftests: enable comma-terminator tests Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 31/38] dyndbg: split multi-query strings with @ Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 32/38] dyndbg: resolve "protection" of class'd pr_debug Jim Cromie via B4 Relay
2026-09-16 16:56 ` sashiko-bot
2026-09-16 15:33 ` Jim Cromie via B4 Relay [this message]
2026-09-16 15:33 ` [PATCH v10 34/38] docs/dyndbg: add classmap info to howto Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 35/38] dyndbg: Ignore additional arguments from pr_fmt Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 36/38] dyndbg: add epilogue to dynamic_debug/control file Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 37/38] dyndbg: add +c flag to count advantage of classmaps for DRM Jim Cromie via B4 Relay
2026-09-16 15:33 ` [PATCH v10 38/38] dyndbg: add DEBUG-biased fallback stubs for _dynamic_func_call_cls Jim Cromie via B4 Relay
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=20260916-dd-cmap-part2-clean-v10-33-af4cf4767707@gmail.com \
--to=devnull+jim.cromie.gmail.com@kernel.org \
--cc=airlied@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=arnd@arndb.de \
--cc=atomlin@atomlin.com \
--cc=corbet@net.net \
--cc=da.gomez@kernel.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=gregkh@linuxfoundation.org \
--cc=jbaron@akamai.com \
--cc=jim.cromie@gmail.com \
--cc=lenb@kernel.org \
--cc=linux-arch@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kbuild@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-modules@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=maarten.lankhorst@linux.intel.com \
--cc=mcgrof@kernel.org \
--cc=mripard@kernel.org \
--cc=nathan@kernel.org \
--cc=nsc@kernel.org \
--cc=pavel@kernel.org \
--cc=petr.pavlu@suse.com \
--cc=rafael@kernel.org \
--cc=rdunlap@infradead.org \
--cc=samitolvanen@google.com \
--cc=shuah@kernel.org \
--cc=simona@ffwll.ch \
--cc=skhan@linuxfoundation.org \
--cc=tzimmermann@suse.de \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.