Skip to contentSkip to conversation

bpf_wq followup series

PATCH Superseded Benjamin Tissoires 3 participants 7 messages Apr 25, 2024 – Apr 30, 2024

Subsystems (3)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

cover bpf_wq followup series

Few patches that should have been there from day 1.

Anyway, they are coming now.

---
Benjamin Tissoires (3):
bpf: do not walk twice the map on free
bpf: do not walk twice the hash map on free selftests/bpf: drop an unused local variable

 kernel/bpf/arraymap.c                       | 15 ++++++++-------
 kernel/bpf/hashtab.c                        | 16 +++++-----------
 tools/testing/selftests/bpf/prog_tests/wq.c |  2 --
 3 files changed, 13 insertions(+), 20 deletions(-)

---
base-commit: 52578f7f53ff8fe3a8f6f3bc8b5956615c07a16e
change-id: 20240425-bpf-next-2114350587e3

Best regards,

Signature · 1 line
Benjamin Tissoires <[email protected]>
#1 lore raw

PATCH 1/3 bpf: do not walk twice the map on free

If someone stores both a timer and a workqueue in a map, on free we would walk it twice.
Add a check in array_map_free_timers_wq and free the timers and workqueues if they are present.

Fixes 246331e3f1ea ("bpf: allow struct bpf_wq to be embedded in arraymaps and hashmaps")
---
 kernel/bpf/arraymap.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)
kernel/bpf/arraymap.c +8 -7
index 580d07b15471..feabc0193852 100644
@@ -436,13 +436,14 @@ static void array_map_free_timers_wq(struct bpf_map *map)
 	/* We don't reset or free fields other than timer and workqueue
 	 * on uref dropping to zero.
 	 */
-	if (btf_record_has_field(map->record, BPF_TIMER))
-		for (i = 0; i < array->map.max_entries; i++)
-			bpf_obj_free_timer(map->record, array_map_elem_ptr(array, i));
-
-	if (btf_record_has_field(map->record, BPF_WORKQUEUE))
-		for (i = 0; i < array->map.max_entries; i++)
-			bpf_obj_free_workqueue(map->record, array_map_elem_ptr(array, i));
+	if (btf_record_has_field(map->record, BPF_TIMER | BPF_WORKQUEUE)) {
+		for (i = 0; i < array->map.max_entries; i++) {
+			if (btf_record_has_field(map->record, BPF_TIMER))
+				bpf_obj_free_timer(map->record, array_map_elem_ptr(array, i));
+			if (btf_record_has_field(map->record, BPF_WORKQUEUE))
+				bpf_obj_free_workqueue(map->record, array_map_elem_ptr(array, i));
+		}
+	}
 }
 
 /* Called when map->refcnt goes to zero, either from workqueue or from syscall */
Signature · 1 line
2.44.0
#2 lore raw

PATCH 2/3 bpf: do not walk twice the hash map on free

In reply to Benjamin Tissoires

If someone stores both a timer and a workqueue in a hash map, on free, we would walk it twice.
Add a check in htab_free_malloced_timers_or_wq and free the timers and workqueues if they are present.

Fixes 246331e3f1ea ("bpf: allow struct bpf_wq to be embedded in arraymaps and hashmaps")
---
 kernel/bpf/hashtab.c | 16 +++++-----------
 1 file changed, 5 insertions(+), 11 deletions(-)
kernel/bpf/hashtab.c +5 -11
index 0179183c543a..20162ae741e9 100644
@@ -1515,7 +1515,7 @@ static void delete_all_elements(struct bpf_htab *htab)
 	migrate_enable();
 }
 
-static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab, bool is_timer)
+static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab)
 {
 	int i;
 
@@ -1527,10 +1527,10 @@ static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab, bool is_timer
 
 		hlist_nulls_for_each_entry(l, n, head, hash_node) {
 			/* We only free timer on uref dropping to zero */
-			if (is_timer)
+			if (btf_record_has_field(htab->map.record, BPF_TIMER))
 				bpf_obj_free_timer(htab->map.record,
 						   l->key + round_up(htab->map.key_size, 8));
-			else
+			if (btf_record_has_field(htab->map.record, BPF_WORKQUEUE))
 				bpf_obj_free_workqueue(htab->map.record,
 						       l->key + round_up(htab->map.key_size, 8));
 		}
@@ -1544,18 +1544,12 @@ static void htab_map_free_timers_and_wq(struct bpf_map *map)
 	struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 
 	/* We only free timer and workqueue on uref dropping to zero */
-	if (btf_record_has_field(htab->map.record, BPF_TIMER)) {
+	if (btf_record_has_field(htab->map.record, BPF_TIMER | BPF_WORKQUEUE)) {
 		if (!htab_is_prealloc(htab))
-			htab_free_malloced_timers_or_wq(htab, true);
+			htab_free_malloced_timers_or_wq(htab);
 		else
 			htab_free_prealloced_timers(htab);
 	}
-	if (btf_record_has_field(htab->map.record, BPF_WORKQUEUE)) {
-		if (!htab_is_prealloc(htab))
-			htab_free_malloced_timers_or_wq(htab, false);
-		else
-			htab_free_prealloced_wq(htab);
-	}
 }
 
 /* Called when map->refcnt goes to zero, either from workqueue or from syscall */
Signature · 1 line
2.44.0
#3 lore raw

Re: [PATCH 2/3] bpf: do not walk twice the hash map on free

Quoted 45 lines · Benjamin Tissoires

If someone stores both a timer and a workqueue in a hash map, on free, we would walk it twice.
Add a check in htab_free_malloced_timers_or_wq and free the timers and workqueues if they are present.

Fixes 246331e3f1ea ("bpf: allow struct bpf_wq to be embedded in arraymaps and hashmaps")
---
 kernel/bpf/hashtab.c | 16 +++++-----------
 1 file changed, 5 insertions(+), 11 deletions(-)
kernel/bpf/hashtab.c +4 -4
index 0179183c543a..20162ae741e9 100644
@@ -1515,7 +1515,7 @@ static void delete_all_elements(struct bpf_htab *htab)
        migrate_enable();
 }
 
-static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab, bool is_timer)
+static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab)
 {
        int i;
 
@@ -1527,10 +1527,10 @@ static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab, bool is_timer
 
                hlist_nulls_for_each_entry(l, n, head, hash_node) {
                        /* We only free timer on uref dropping to zero */
-                       if (is_timer)
+                       if (btf_record_has_field(htab->map.record, BPF_TIMER))
                                bpf_obj_free_timer(htab->map.record,
                                                   l->key + round_up(htab->map.key_size, 8));
-                       else
+                       if (btf_record_has_field(htab->map.record, BPF_WORKQUEUE))
                                bpf_obj_free_workqueue(htab->map.record,
                                                       l->key + round_up(htab->map.key_size, 8));
                }
@@ -1544,18 +1544,12 @@ static void htab_map_free_timers_and_wq(struct bpf_map *map)
        struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 
        /* We only free timer and workqueue on uref dropping to zero */
-       if (btf_record_has_field(htab->map.record, BPF_TIMER)) {
+       if (btf_record_has_field(htab->map.record, BPF_TIMER | BPF_WORKQUEUE)) {
                if (!htab_is_prealloc(htab))

kernel/bpf/hashtab.c static void htab_map_free_timers_and_wq(struct bpf_map *map) · line 1557

Quoted 10 lines
kernel/bpf/hashtab.c +1 -6
-                       htab_free_malloced_timers_or_wq(htab, true);
+                       htab_free_malloced_timers_or_wq(htab);
                else
                        htab_free_prealloced_timers(htab);
        }
-       if (btf_record_has_field(htab->map.record, BPF_WORKQUEUE)) {
-               if (!htab_is_prealloc(htab))
-                       htab_free_malloced_timers_or_wq(htab, false);
-               else
-                       htab_free_prealloced_wq(htab);

This looks wrong.
htab_free_prealloced_wq() is now unused as compiler says:
../kernel/bpf/hashtab.c:243:13: warning: ‘htab_free_prealloced_wq’ defined but not used [-Wunused-function]
243 | static void htab_free_prealloced_wq(struct bpf_htab *htab)

      |             ^~~~~~~~~~~~~~~~~~~~~~~

and prealloced maps with wq leak wq-s.

pw-bot: cr

#4 lore raw

Re: [PATCH 2/3] bpf: do not walk twice the hash map on free

Quoted 62 lines · Alexei Starovoitov
Quoted 45 lines · Benjamin Tissoires

If someone stores both a timer and a workqueue in a hash map, on free, we would walk it twice.
Add a check in htab_free_malloced_timers_or_wq and free the timers and workqueues if they are present.

Fixes 246331e3f1ea ("bpf: allow struct bpf_wq to be embedded in arraymaps and hashmaps")
---
 kernel/bpf/hashtab.c | 16 +++++-----------
 1 file changed, 5 insertions(+), 11 deletions(-)
kernel/bpf/hashtab.c +4 -4
index 0179183c543a..20162ae741e9 100644
@@ -1515,7 +1515,7 @@ static void delete_all_elements(struct bpf_htab *htab)
        migrate_enable();
 }
 
-static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab, bool is_timer)
+static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab)
 {
        int i;
 
@@ -1527,10 +1527,10 @@ static void htab_free_malloced_timers_or_wq(struct bpf_htab *htab, bool is_timer
 
                hlist_nulls_for_each_entry(l, n, head, hash_node) {
                        /* We only free timer on uref dropping to zero */
-                       if (is_timer)
+                       if (btf_record_has_field(htab->map.record, BPF_TIMER))
                                bpf_obj_free_timer(htab->map.record,
                                                   l->key + round_up(htab->map.key_size, 8));
-                       else
+                       if (btf_record_has_field(htab->map.record, BPF_WORKQUEUE))
                                bpf_obj_free_workqueue(htab->map.record,
                                                       l->key + round_up(htab->map.key_size, 8));
                }
@@ -1544,18 +1544,12 @@ static void htab_map_free_timers_and_wq(struct bpf_map *map)
        struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 
        /* We only free timer and workqueue on uref dropping to zero */
-       if (btf_record_has_field(htab->map.record, BPF_TIMER)) {
+       if (btf_record_has_field(htab->map.record, BPF_TIMER | BPF_WORKQUEUE)) {
                if (!htab_is_prealloc(htab))

kernel/bpf/hashtab.c static void htab_map_free_timers_and_wq(struct bpf_map *map) · line 1557

Quoted 10 lines
kernel/bpf/hashtab.c +1 -6
-                       htab_free_malloced_timers_or_wq(htab, true);
+                       htab_free_malloced_timers_or_wq(htab);
                else
                        htab_free_prealloced_timers(htab);
        }
-       if (btf_record_has_field(htab->map.record, BPF_WORKQUEUE)) {
-               if (!htab_is_prealloc(htab))
-                       htab_free_malloced_timers_or_wq(htab, false);
-               else
-                       htab_free_prealloced_wq(htab);

This looks wrong.
htab_free_prealloced_wq() is now unused as compiler says:
../kernel/bpf/hashtab.c:243:13: warning: ‘htab_free_prealloced_wq’ defined but not used [-Wunused-function]
243 | static void htab_free_prealloced_wq(struct bpf_htab *htab)

Quoted 3 lines
      |             ^~~~~~~~~~~~~~~~~~~~~~~

and prealloced maps with wq leak wq-s.

oops, you are right. Sending a v2 right away (sorry for the delay).

Cheers,
Benjamin

#5 lore raw

Re: [PATCH 2/3] bpf: do not walk twice the hash map on free

In reply to Benjamin Tissoires
kernel test robot noticed the following build warnings:

Hi Benjamin,

kernel test robot noticed the following build warnings:

[auto build test WARNING on 52578f7f53ff8fe3a8f6f3bc8b5956615c07a16e]

url:    https://github.com/intel-lab-lkp/linux/commits/Benjamin-Tissoires/bpf-do-not-walk-twice-the-map-on-free/20240425-220322
base:   52578f7f53ff8fe3a8f6f3bc8b5956615c07a16e
patch link:    https://lore.kernel.org/r/20240425-bpf-next-v1-2-1d8330e6c643%40kernel.org
patch subject: [PATCH 2/3] bpf: do not walk twice the hash map on free
config: arc-randconfig-002-20240426 (https://download.01.org/0day-ci/archive/20240426/[email protected]/config)
compiler: arc-elf-gcc (GCC) 13.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240426/[email protected]/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/

All warnings (new ones prefixed by >>):

>> kernel/bpf/hashtab.c:243:13: warning: 'htab_free_prealloced_wq' defined but not used [-Wunused-function]

243 | static void htab_free_prealloced_wq(struct bpf_htab *htab)
| ^~~~~~~~~~~~~~~~~~~~~~~

vim +/htab_free_prealloced_wq +243 kernel/bpf/hashtab.c

68134668c17f31 Alexei Starovoitov 2021-07-14 242

246331e3f1eac9 Benjamin Tissoires 2024-04-20 @243  static void htab_free_prealloced_wq(struct bpf_htab *htab)
246331e3f1eac9 Benjamin Tissoires 2024-04-20  244  {
246331e3f1eac9 Benjamin Tissoires 2024-04-20  245  	u32 num_entries = htab->map.max_entries;
246331e3f1eac9 Benjamin Tissoires 2024-04-20  246  	int i;
246331e3f1eac9 Benjamin Tissoires 2024-04-20  247  
246331e3f1eac9 Benjamin Tissoires 2024-04-20  248  	if (!btf_record_has_field(htab->map.record, BPF_WORKQUEUE))
246331e3f1eac9 Benjamin Tissoires 2024-04-20  249  		return;
246331e3f1eac9 Benjamin Tissoires 2024-04-20  250  	if (htab_has_extra_elems(htab))
246331e3f1eac9 Benjamin Tissoires 2024-04-20  251  		num_entries += num_possible_cpus();
246331e3f1eac9 Benjamin Tissoires 2024-04-20  252  
246331e3f1eac9 Benjamin Tissoires 2024-04-20  253  	for (i = 0; i < num_entries; i++) {
246331e3f1eac9 Benjamin Tissoires 2024-04-20  254  		struct htab_elem *elem;
246331e3f1eac9 Benjamin Tissoires 2024-04-20  255  
246331e3f1eac9 Benjamin Tissoires 2024-04-20  256  		elem = get_htab_elem(htab, i);
246331e3f1eac9 Benjamin Tissoires 2024-04-20  257  		bpf_obj_free_workqueue(htab->map.record,
246331e3f1eac9 Benjamin Tissoires 2024-04-20  258  				       elem->key + round_up(htab->map.key_size, 8));
246331e3f1eac9 Benjamin Tissoires 2024-04-20  259  		cond_resched();
246331e3f1eac9 Benjamin Tissoires 2024-04-20  260  	}
246331e3f1eac9 Benjamin Tissoires 2024-04-20  261  }
246331e3f1eac9 Benjamin Tissoires 2024-04-20  262  
Signature · 2 lines
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
#6 lore raw

PATCH 3/3 selftests/bpf: drop an unused local variable

In reply to Benjamin Tissoires

Some copy/paste leftover, this is never used

Fixes e3d9eac99afd ("selftests/bpf: wq: add bpf_wq_init() checks")
---
 tools/testing/selftests/bpf/prog_tests/wq.c | 2 --
 1 file changed, 2 deletions(-)
tools/testing/selftests/bpf/prog_tests/wq.c +0 -2
index c4bacd3160e1..99e438fe12ac 100644
@@ -36,7 +36,5 @@ void serial_test_wq(void)
 
 void serial_test_failures_wq(void)
 {
-	LIBBPF_OPTS(bpf_test_run_opts, topts);
-
 	RUN_TESTS(wq_failures);
 }
Signature · 1 line
2.44.0