of: unittest: kmleak detected memory leaks
cover of: unittest: kmleak detected memory leaks
From: Frank Rowand <[email protected]>
Original bug report:
https://bugzilla.kernel.org/show_bug.cgi?id=206203
https://lore.kernel.org/r/[email protected]
kmemleak detected many memory leaks originating if devicetree unittests.
Five separate causes were found. Four causes were bugs and one was expected.
Fix the four bugs and modify the test that led to the expected memory leak to no longer have a memory leak.
Frank Rowand (5):
of: unittest: kmemleak on changeset destroy
of: unittest: kmemleak in of_unittest_platform_populate()
of: unittest: kmemleak in of_unittest_overlay_high_level()
of: overlay: kmemleak in dup_and_fixup_symbol_prop()
of: unittest: kmemleak in duplicate property update
drivers/of/overlay.c | 2 ++ .../of/unittest-data/overlay_bad_add_dup_prop.dts | 23 ++++++++++++++---- drivers/of/unittest.c | 28 +++++++++++++++------- 3 files changed, 40 insertions(+), 13 deletions(-)
Signature · 1 line
Frank Rowand <[email protected]>
PATCH 1/5 of: unittest: kmemleak on changeset destroy
From: Frank Rowand <[email protected]>
kmemleak reports several memory leaks from devicetree unittest. This is the fix for problem 1 of 5.
of_unittest_changeset() reaches deeply into the dynamic devicetree functions. Several nodes were left with an elevated reference count and thus were not properly cleaned up. Fix the reference counts so that the memory will be freed.
--- drivers/of/unittest.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/of/unittest.c +4 -0
@@ -861,6 +861,10 @@ static void __init of_unittest_changeset(void)unittest (!of_changeset_revert (&chgset),"revert failed\n" );of_changeset_destroy (&chgset); + +of_node_put (n1); +of_node_put (n2); +of_node_put (n21);#endif }
Signature · 1 line
Frank Rowand <[email protected]>
PATCH 2/5 of: unittest: kmemleak in of_unittest_platform_populate()
In reply to [email protected]From: Frank Rowand <[email protected]>
kmemleak reports several memory leaks from devicetree unittest. This is the fix for problem 2 of 5.
of_unittest_platform_populate() left an elevated reference count for grandchild nodes (which are platform devices). Fix the platform device reference counts so that the memory will be freed.
--- drivers/of/unittest.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/of/unittest.c +5 -2
@@ -1247,10 +1247,13 @@ static void __init of_unittest_platform_populate(void)of_platform_populate (np, match,NULL , &test_bus->dev);for_each_child_of_node (np, child) { -for_each_child_of_node (child, grandchild) -unittest (of_find_device_by_node (grandchild), +for_each_child_of_node (child, grandchild) { + pdev =of_find_device_by_node (grandchild); +unittest (pdev,"Could not create device for node '%pOFn'\n" , grandchild); +of_dev_put (pdev); + } }of_platform_depopulate (&test_bus->dev);
Signature · 1 line
Frank Rowand <[email protected]>
PATCH 3/5 of: unittest: kmemleak in of_unittest_overlay_high_level()
In reply to [email protected]From: Frank Rowand <[email protected]>
kmemleak reports several memory leaks from devicetree unittest. This is the fix for problem 3 of 5.
of_unittest_overlay_high_level() failed to kfree the newly created property when the property named 'name' is skipped.
--- drivers/of/unittest.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/of/unittest.c +4 -1
@@ -3094,8 +3094,11 @@ static __init void of_unittest_overlay_high_level(void) goto err_unlock; }if (__of_add_property (of_symbols, new_prop)) { +kfree (new_prop->name); +kfree (new_prop->value); +kfree (new_prop);/* "name" auto-generated by unflatten */ -if (!strcmp (new_prop->name,"name" )) +if (!strcmp (prop->name,"name" ))continue ;unittest (0 ,"duplicate property '%s' in overlay_base node __symbols__" , prop->name);
Signature · 1 line
Frank Rowand <[email protected]>
PATCH 4/5 of: overlay: kmemleak in dup_and_fixup_symbol_prop()
In reply to [email protected]From: Frank Rowand <[email protected]>
kmemleak reports several memory leaks from devicetree unittest. This is the fix for problem 4 of 5.
target_path was not freed in the non-error path.
--- drivers/of/overlay.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/of/overlay.c +2 -0
@@ -261,6 +261,8 @@ static struct property *dup_and_fixup_symbol_prop(of_property_set_flag (new_prop,OF_DYNAMIC ); +kfree (target_path); +return new_prop; err_free_new_prop:
Signature · 1 line
Frank Rowand <[email protected]>
PATCH 5/5 of: unittest: kmemleak in duplicate property update
In reply to [email protected]From: Frank Rowand <[email protected]>
kmemleak reports several memory leaks from devicetree unittest. This is the fix for problem 5 of 5.
When overlay 'overlay_bad_add_dup_prop' is applied, the apply code properly detects that a memory leak will occur if the overlay is removed since the duplicate property is located in a base devicetree node and reports via printk():
OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_avail
OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_avail
The overlay is removed when the apply code detects multiple changesets modifying the same property. This is reported via printk():
OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/rpm_avail
As a result of this error, the overlay is removed resulting in the expected memory leak.
Add another device node level to the overlay so that the duplicate property is located in a node added by the overlay, thus no memory leak will occur when the overlay is removed.
Thus users of kmemleak will not have to debug this leak in the future.
--- .../of/unittest-data/overlay_bad_add_dup_prop.dts | 23 ++++++++++++++++++---- drivers/of/unittest.c | 12 +++++------ 2 files changed, 25 insertions(+), 10 deletions(-)
drivers/of/unittest-data/overlay_bad_add_dup_prop.dts +19 -4
@@ -3,22 +3,37 @@ /plugin/; /* - * &electric_1/motor-1 and &spin_ctrl_1 are the same node: - * /testcase-data-2/substation@100/motor-1 + * &electric_1/motor-1/electric and &spin_ctrl_1/electric are the same node: + * /testcase-data-2/substation@100/motor-1/electric * * Thus the property "rpm_avail" in each fragment will * result in an attempt to update the same property twice. * This will result in an error and the overlay apply * will fail. + * + * The previous version of this test did not include the extra + * level of node 'electric'. That resulted in the 'rpm_avail' + * property being located in the pre-existing node 'motor-1'. + * Modifying a property results in a WARNING that a memory leak + * will occur if the overlay is removed. Since the overlay apply + * fails, the memory leak does actually occur, and kmemleak will + * further report the memory leak if CONFIG_DEBUG_KMEMLEAK is + * enabled. Adding the overlay node 'electric' avoids the + * memory leak and thus people who use kmemleak will not + * have to debug this non-problem again. */ &electric_1 { motor-1 { - rpm_avail = < 100 >; + electric { + rpm_avail = < 100 >; + }; }; }; &spin_ctrl_1 { - rpm_avail = < 100 200 >; + electric { + rpm_avail = < 100 200 >; + }; };
drivers/of/unittest.c +6 -6
@@ -3181,21 +3181,21 @@ static __init void of_unittest_overlay_high_level(void) "OF: overlay: ERROR: multiple fragments add and/or delete node /testcase-data-2/substation@100/motor-1/controller");EXPECT_BEGIN (KERN_ERR , -"OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_avail" ); +"OF: overlay: ERROR: multiple fragments add and/or delete node /testcase-data-2/substation@100/motor-1/electric" );EXPECT_BEGIN (KERN_ERR , -"OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_avail" ); +"OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/electric/rpm_avail" );EXPECT_BEGIN (KERN_ERR , -"OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/rpm_avail" ); +"OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/electric/name" );unittest (overlay_data_apply ("overlay_bad_add_dup_prop" ,NULL ),"Adding overlay 'overlay_bad_add_dup_prop' failed\n" );EXPECT_END (KERN_ERR , -"OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/rpm_avail" ); +"OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/electric/name" );EXPECT_END (KERN_ERR , -"OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_avail" ); +"OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/electric/rpm_avail" );EXPECT_END (KERN_ERR , -"OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_avail" ); +"OF: overlay: ERROR: multiple fragments add and/or delete node /testcase-data-2/substation@100/motor-1/electric" );unittest (overlay_data_apply ("overlay_bad_phandle" ,NULL ),"Adding overlay 'overlay_bad_phandle' failed\n" );
Signature · 1 line
Frank Rowand <[email protected]>
Re: [PATCH 5/5] of: unittest: kmemleak in duplicate property update
Quoted 27 lines · On Thu, 16 Apr 2020 16:42:50 -0500, [email protected] wrote:
From: Frank Rowand <[email protected]>
kmemleak reports several memory leaks from devicetree unittest. This is the fix for problem 5 of 5.
When overlay '
overlay_bad_add_dup_prop' is applied, the apply code properly detects that a memory leak will occur if the overlay is removed since the duplicate property is located in a base devicetree node and reports viaprintk():OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_avail
OF: overlay: WARNING: memory leak will occur if overlay removed, property: /testcase-data-2/substation@100/motor-1/rpm_availThe overlay is removed when the apply code detects multiple changesets modifying the same property. This is reported via
printk():OF: overlay: ERROR: multiple fragments add, update, and/or delete property /testcase-data-2/substation@100/motor-1/rpm_avail
As a result of this error, the overlay is removed resulting in the expected memory leak.
Add another device node level to the overlay so that the duplicate property is located in a node added by the overlay, thus no memory leak will occur when the overlay is removed.
Thus users of kmemleak will not have to debug this leak in the future.
Quoted 8 lines
Fixes 2fe0e8769df9 ("of: overlay: check prevents multiple fragments touching same property")Reported-by Erhard F. <[email protected]>Signed-off-by Frank Rowand <[email protected]>--- .../of/unittest-data/overlay_bad_add_dup_prop.dts | 23 ++++++++++++++++++---- drivers/of/unittest.c | 12 +++++------ 2 files changed, 25 insertions(+), 10 deletions(-)
Applied, thanks.
Rob
Rob Herring
maintainerQuoted 16 lines · On Thu, Apr 16, 2020 at 04:42:45PM -0500, [email protected] wrote:
From: Frank Rowand <[email protected]>
Original bug report:
https://bugzilla.kernel.org/show_bug.cgi?id=206203
https://lore.kernel.org/r/[email protected]kmemleak detected many memory leaks originating if devicetree unittests.
Five separate causes were found. Four causes were bugs and one was expected.
Fix the four bugs and modify the test that led to the expected memory leak to no longer have a memory leak.
Quoted 6 lines
Frank Rowand (5):
of: unittest: kmemleak on changeset destroy
of: unittest: kmemleak inof_unittest_platform_populate()
of: unittest: kmemleak inof_unittest_overlay_high_level()
of: overlay: kmemleak indup_and_fixup_symbol_prop()
of: unittest: kmemleak in duplicate property update
Series applied. (To get into today's -next and I plan to send to Linus tomorrow.)
Rob