| Message ID | 20230310140848.94485-4-zhangpeng.00@bytedance.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | Some fixes and cleanup for maple tree. | expand |
* Peng Zhang <zhangpeng.00@bytedance.com> [230310 09:09]: > There is a concurrency bug that may cause the wrong value to be loaded > when a CPU is modifying the maple tree. > > CPU1: > mtree_insert_range() > mas_insert() > mas_store_root() > ... > mas_root_expand() > ... > rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); > ma_set_meta(node, maple_leaf_64, 0, slot); <---IP > > CPU2: > mtree_load() > mtree_lookup_walk() > ma_data_end(); > > When CPU1 is about to execute the instruction pointed to by IP, > the ma_data_end() executed by CPU2 may return the wrong end position, > which will cause the value loaded by mtree_load() to be wrong. > > An example of triggering the bug: > > Add mdelay(100) between rcu_assign_pointer() and ma_set_meta() in > mas_root_expand(). > > static DEFINE_MTREE(tree); > int work(void *p) { > unsigned long val; > for (int i = 0 ; i< 30; ++i) { > val = (unsigned long)mtree_load(&tree, 8); > mdelay(5); > pr_info("%lu",val); > } > return 0; > } > > mt_init_flags(&tree, MT_FLAGS_USE_RCU); > mtree_insert(&tree, 0, (void*)12345, GFP_KERNEL); > run_thread(work) > mtree_insert(&tree, 1, (void*)56789, GFP_KERNEL); > > In RCU mode, mtree_load() should always return the value before or after > the data structure is modified, and in this example mtree_load(&tree, 8) > may return 56789 which is not expected, it should always return NULL. > Fix it by put ma_set_meta() before rcu_assign_pointer(). Are you able to write a test case for this issue? I understand it's a race so it may be difficult to catch. > > Signed-off-by: Peng Zhang <zhangpeng.00@bytedance.com> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> > --- > lib/maple_tree.c | 3 +-- > 1 file changed, 1 insertion(+), 2 deletions(-) > > diff --git a/lib/maple_tree.c b/lib/maple_tree.c > index 4d15202a0692..de43ff19da72 100644 > --- a/lib/maple_tree.c > +++ b/lib/maple_tree.c > @@ -3635,10 +3635,9 @@ static inline int mas_root_expand(struct ma_state *mas, void *entry) > slot++; > mas->depth = 1; > mas_set_height(mas); > - > + ma_set_meta(node, maple_leaf_64, 0, slot); > /* swap the new root into the tree */ > rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); > - ma_set_meta(node, maple_leaf_64, 0, slot); > return slot; > } > > -- > 2.20.1 >
在 2023/3/11 02:27, Liam R. Howlett 写道: > * Peng Zhang <zhangpeng.00@bytedance.com> [230310 09:09]: >> There is a concurrency bug that may cause the wrong value to be loaded >> when a CPU is modifying the maple tree. >> >> CPU1: >> mtree_insert_range() >> mas_insert() >> mas_store_root() >> ... >> mas_root_expand() >> ... >> rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); >> ma_set_meta(node, maple_leaf_64, 0, slot); <---IP >> >> CPU2: >> mtree_load() >> mtree_lookup_walk() >> ma_data_end(); >> >> When CPU1 is about to execute the instruction pointed to by IP, >> the ma_data_end() executed by CPU2 may return the wrong end position, >> which will cause the value loaded by mtree_load() to be wrong. >> >> An example of triggering the bug: >> >> Add mdelay(100) between rcu_assign_pointer() and ma_set_meta() in >> mas_root_expand(). >> >> static DEFINE_MTREE(tree); >> int work(void *p) { >> unsigned long val; >> for (int i = 0 ; i< 30; ++i) { >> val = (unsigned long)mtree_load(&tree, 8); >> mdelay(5); >> pr_info("%lu",val); >> } >> return 0; >> } >> >> mt_init_flags(&tree, MT_FLAGS_USE_RCU); >> mtree_insert(&tree, 0, (void*)12345, GFP_KERNEL); >> run_thread(work) >> mtree_insert(&tree, 1, (void*)56789, GFP_KERNEL); >> >> In RCU mode, mtree_load() should always return the value before or after >> the data structure is modified, and in this example mtree_load(&tree, 8) >> may return 56789 which is not expected, it should always return NULL. >> Fix it by put ma_set_meta() before rcu_assign_pointer(). > Are you able to write a test case for this issue? I understand it's a > race so it may be difficult to catch. Yes it's hard to catch. I'll try to think of a test case next week. It is difficult because of the need to expand the competitive area. > >> Signed-off-by: Peng Zhang <zhangpeng.00@bytedance.com> > Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> > >> --- >> lib/maple_tree.c | 3 +-- >> 1 file changed, 1 insertion(+), 2 deletions(-) >> >> diff --git a/lib/maple_tree.c b/lib/maple_tree.c >> index 4d15202a0692..de43ff19da72 100644 >> --- a/lib/maple_tree.c >> +++ b/lib/maple_tree.c >> @@ -3635,10 +3635,9 @@ static inline int mas_root_expand(struct ma_state *mas, void *entry) >> slot++; >> mas->depth = 1; >> mas_set_height(mas); >> - >> + ma_set_meta(node, maple_leaf_64, 0, slot); >> /* swap the new root into the tree */ >> rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); >> - ma_set_meta(node, maple_leaf_64, 0, slot); >> return slot; >> } >> >> -- >> 2.20.1 >>
* Peng Zhang <zhangpeng.00@bytedance.com> [230310 14:03]: > > 在 2023/3/11 02:27, Liam R. Howlett 写道: > > * Peng Zhang <zhangpeng.00@bytedance.com> [230310 09:09]: > > > There is a concurrency bug that may cause the wrong value to be loaded > > > when a CPU is modifying the maple tree. > > > > > > CPU1: > > > mtree_insert_range() > > > mas_insert() > > > mas_store_root() > > > ... > > > mas_root_expand() > > > ... > > > rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); > > > ma_set_meta(node, maple_leaf_64, 0, slot); <---IP > > > > > > CPU2: > > > mtree_load() > > > mtree_lookup_walk() > > > ma_data_end(); > > > > > > When CPU1 is about to execute the instruction pointed to by IP, > > > the ma_data_end() executed by CPU2 may return the wrong end position, > > > which will cause the value loaded by mtree_load() to be wrong. > > > > > > An example of triggering the bug: > > > > > > Add mdelay(100) between rcu_assign_pointer() and ma_set_meta() in > > > mas_root_expand(). > > > > > > static DEFINE_MTREE(tree); > > > int work(void *p) { > > > unsigned long val; > > > for (int i = 0 ; i< 30; ++i) { > > > val = (unsigned long)mtree_load(&tree, 8); > > > mdelay(5); > > > pr_info("%lu",val); > > > } > > > return 0; > > > } > > > > > > mt_init_flags(&tree, MT_FLAGS_USE_RCU); > > > mtree_insert(&tree, 0, (void*)12345, GFP_KERNEL); > > > run_thread(work) > > > mtree_insert(&tree, 1, (void*)56789, GFP_KERNEL); > > > > > > In RCU mode, mtree_load() should always return the value before or after > > > the data structure is modified, and in this example mtree_load(&tree, 8) > > > may return 56789 which is not expected, it should always return NULL. > > > Fix it by put ma_set_meta() before rcu_assign_pointer(). > > Are you able to write a test case for this issue? I understand it's a > > race so it may be difficult to catch. > Yes it's hard to catch. I'll try to think of a test case next week. > It is difficult because of the need to expand the competitive area. This should have a fixes tag as well. Fixes: 54a611b60590 ("Maple Tree: add new data structure") > > > > > Signed-off-by: Peng Zhang <zhangpeng.00@bytedance.com> > > Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com> > > > > > --- > > > lib/maple_tree.c | 3 +-- > > > 1 file changed, 1 insertion(+), 2 deletions(-) > > > > > > diff --git a/lib/maple_tree.c b/lib/maple_tree.c > > > index 4d15202a0692..de43ff19da72 100644 > > > --- a/lib/maple_tree.c > > > +++ b/lib/maple_tree.c > > > @@ -3635,10 +3635,9 @@ static inline int mas_root_expand(struct ma_state *mas, void *entry) > > > slot++; > > > mas->depth = 1; > > > mas_set_height(mas); > > > - > > > + ma_set_meta(node, maple_leaf_64, 0, slot); > > > /* swap the new root into the tree */ > > > rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); > > > - ma_set_meta(node, maple_leaf_64, 0, slot); > > > return slot; > > > } > > > -- > > > 2.20.1 > > >
diff --git a/lib/maple_tree.c b/lib/maple_tree.c index 4d15202a0692..de43ff19da72 100644 --- a/lib/maple_tree.c +++ b/lib/maple_tree.c @@ -3635,10 +3635,9 @@ static inline int mas_root_expand(struct ma_state *mas, void *entry) slot++; mas->depth = 1; mas_set_height(mas); - + ma_set_meta(node, maple_leaf_64, 0, slot); /* swap the new root into the tree */ rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); - ma_set_meta(node, maple_leaf_64, 0, slot); return slot; }
There is a concurrency bug that may cause the wrong value to be loaded when a CPU is modifying the maple tree. CPU1: mtree_insert_range() mas_insert() mas_store_root() ... mas_root_expand() ... rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); ma_set_meta(node, maple_leaf_64, 0, slot); <---IP CPU2: mtree_load() mtree_lookup_walk() ma_data_end(); When CPU1 is about to execute the instruction pointed to by IP, the ma_data_end() executed by CPU2 may return the wrong end position, which will cause the value loaded by mtree_load() to be wrong. An example of triggering the bug: Add mdelay(100) between rcu_assign_pointer() and ma_set_meta() in mas_root_expand(). static DEFINE_MTREE(tree); int work(void *p) { unsigned long val; for (int i = 0 ; i< 30; ++i) { val = (unsigned long)mtree_load(&tree, 8); mdelay(5); pr_info("%lu",val); } return 0; } mt_init_flags(&tree, MT_FLAGS_USE_RCU); mtree_insert(&tree, 0, (void*)12345, GFP_KERNEL); run_thread(work) mtree_insert(&tree, 1, (void*)56789, GFP_KERNEL); In RCU mode, mtree_load() should always return the value before or after the data structure is modified, and in this example mtree_load(&tree, 8) may return 56789 which is not expected, it should always return NULL. Fix it by put ma_set_meta() before rcu_assign_pointer(). Signed-off-by: Peng Zhang <zhangpeng.00@bytedance.com> --- lib/maple_tree.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)