of: unittest: Remove refcount underflow tests
PATCH 1/1 of: unittest: Remove refcount underflow tests
With CONFIG_OF_UNITTEST=y and CONFIG_BUG_ON_DATA_CORRUPTION=y, the machine no longer boots after commit 5b5224220747 ("refcount: Use CONFIG_BUG_ON_DATA_CORRUPTION for UAF-related errors").
of_unittest_lifecycle() deliberately calls of_node_put() after the node's reference count has reached zero. The test expects a warning and continues to check that a subsequent put does not produce another warning. With BUG_ON_DATA_CORRUPTION enabled, refcount underflow now triggers BUG() instead. This panics the kernel during boot:
kernel BUG at lib/refcount.c:37! Oops: invalid opcode: 0000 [#1] SMP NOPTI CPU: 1 UID: 0 PID: 1 Comm: swapper/0 Tainted: G N 7.3.0+ #14 PREEMPT(full) RIP: 0010:refcount_warn_saturate+0x89/0x90 Call Trace: of_unittest_lifecycle+0x158/0x310 of_unittest+0x105/0x130 do_one_initcall+0x127/0x350 do_initcall_level+0x7c/0xe0 do_initcalls+0x59/0xa0 kernel_init_freeable+0x118/0x1a0 kernel_init+0x1f/0x140 Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
Remove the refcount underflow tests, which exercise the refcount implementation rather than OF-specific behavior.
--- drivers/of/unittest.c | 43 ------------------------------------------- 1 file changed, 43 deletions(-)
drivers/of/unittest.c +0 -43
@@ -3678,9 +3678,6 @@ static void __init of_unittest_lifecycle(void)* - Decrementing refcount to zero via of_node_put() should cause the * attempt to free the node memory by of_node_release() to fail * because the node is not a dynamic node. -* -* - Decrementing refcount past zero should result in additional -* errors reported. */ np =of_find_node_by_path (refcount_path);
@@ -3714,46 +3711,6 @@ static void __init of_unittest_lifecycle(void)EXPECT_END (KERN_INFO ,"OF: ERROR: of_node_release() detected bad of_node_put() on /testcase-data/refcount-node" ); - -/* -* expect stack trace for subsequent of_node_put(): -* __refcount_sub_and_test() calls: -* refcount_warn_saturate(r, REFCOUNT_SUB_UAF) -* -* Not capturing entire WARN_ONCE() trace with EXPECT_*(), just -* the first three lines, and the last line. -*/ -EXPECT_BEGIN (KERN_INFO ,"------------[ cut here ]------------" ); -EXPECT_BEGIN (KERN_INFO ,"WARNING: <<all>>" ); -EXPECT_BEGIN (KERN_INFO ,"refcount_t: underflow; use-after-free." ); -EXPECT_BEGIN (KERN_INFO ,"---[ end trace <<int>> ]---" ); - -/* refcount is now zero, this should fail */ -unittest (1 ,"/testcase-data/refcount-node is zero" ); -of_node_put (np); - -EXPECT_END (KERN_INFO ,"---[ end trace <<int>> ]---" ); -EXPECT_END (KERN_INFO ,"refcount_t: underflow; use-after-free." ); -EXPECT_END (KERN_INFO ,"WARNING: <<all>>" ); -EXPECT_END (KERN_INFO ,"------------[ cut here ]------------" ); - -/* -* Q. do we expect to get yet another warning? -* A. no, the WARNING is from WARN_ONCE() -*/ -EXPECT_NOT_BEGIN (KERN_INFO ,"------------[ cut here ]------------" ); -EXPECT_NOT_BEGIN (KERN_INFO ,"WARNING: <<all>>" ); -EXPECT_NOT_BEGIN (KERN_INFO ,"refcount_t: underflow; use-after-free." ); -EXPECT_NOT_BEGIN (KERN_INFO ,"---[ end trace <<int>> ]---" ); - -unittest (1 ,"/testcase-data/refcount-node is zero, second time" ); -of_node_put (np); - -EXPECT_NOT_END (KERN_INFO ,"---[ end trace <<int>> ]---" ); -EXPECT_NOT_END (KERN_INFO ,"refcount_t: underflow; use-after-free." ); -EXPECT_NOT_END (KERN_INFO ,"WARNING: <<all>>" ); -EXPECT_NOT_END (KERN_INFO ,"------------[ cut here ]------------" ); -/* * refcount of zero will trigger stack traces from any further * attempt to of_node_get() node "refcount-node". One example of
base-commit: a287f97d4256b5f5527f5b5a54f8ada55e1482c7
Signature · 1 line
2.56.0
Quoted 24 lines · Danilo Krummrich
With
CONFIG_OF_UNITTEST=yandCONFIG_BUG_ON_DATA_CORRUPTION=y, the machine no longer boots after commit5b5224220747("refcount: UseCONFIG_BUG_ON_DATA_CORRUPTIONfor UAF-related errors").
of_unittest_lifecycle()deliberately callsof_node_put()after the node's reference count has reached zero. The test expects a warning and continues to check that a subsequent put does not produce another warning. WithBUG_ON_DATA_CORRUPTIONenabled, refcount underflow now triggersBUG()instead. This panics the kernel during boot:kernel BUG at lib/refcount.c:37! Oops: invalid opcode: 0000 [#1] SMP NOPTI CPU: 1 UID: 0 PID: 1 Comm: swapper/0 Tainted: G N 7.3.0+ #14 PREEMPT(full) RIP: 0010:refcount_warn_saturate+0x89/0x90 Call Trace: of_unittest_lifecycle+0x158/0x310 of_unittest+0x105/0x130 do_one_initcall+0x127/0x350 do_initcall_level+0x7c/0xe0 do_initcalls+0x59/0xa0 kernel_init_freeable+0x118/0x1a0 kernel_init+0x1f/0x140 Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
Quoted 8 lines
Remove the refcount underflow tests, which exercise the refcount implementation rather than OF-specific behavior.
Fixes 5b5224220747 ("refcount: Use CONFIG_BUG_ON_DATA_CORRUPTION for UAF-related errors")Signed-off-by Danilo Krummrich <[email protected]>--- drivers/of/unittest.c | 43 ------------------------------------------- 1 file changed, 43 deletions(-)
Applied, thanks.
Rob