Re: [PATCH RESEND 2/2] mm/damon/core-test: Fix memory leak in damon_new_ctx()
From: SeongJae Park
Date: Mon Sep 18 2023 - 01:51:36 EST
Hi Jinjie,
On Mon, 18 Sep 2023 13:10:44 +0800 Jinjie Ruan <ruanjinjie@xxxxxxxxxx> wrote:
> The damon_ctx which is allocated by kzalloc() in damon_new_ctx() in
> damon_test_ops_registration() and damon_test_set_attrs() are not freed and
> it causes below memory leak. So use damon_destroy_ctx() to free it.
>
> unreferenced object 0xffff2b49c6968800 (size 512):
> comm "kunit_try_catch", pid 350, jiffies 4294895294 (age 557.028s)
> hex dump (first 32 bytes):
> 88 13 00 00 00 00 00 00 a0 86 01 00 00 00 00 00 ................
> 00 87 93 03 00 00 00 00 0a 00 00 00 00 00 00 00 ................
> backtrace:
> [<0000000088e71769>] slab_post_alloc_hook+0xb8/0x368
> [<0000000073acab3b>] __kmem_cache_alloc_node+0x174/0x290
> [<00000000b5f89cef>] kmalloc_trace+0x40/0x164
> [<00000000eb19e83f>] damon_new_ctx+0x28/0xb4
> [<00000000daf6227b>] damon_test_ops_registration+0x34/0x328
> [<00000000559c4801>] kunit_try_run_case+0x50/0xac
> [<000000003932ed49>] kunit_generic_run_threadfn_adapter+0x20/0x2c
> [<000000003c3e9211>] kthread+0x124/0x130
> [<0000000028f85bdd>] ret_from_fork+0x10/0x20
> unreferenced object 0xffff2b49c1a9cc00 (size 512):
> comm "kunit_try_catch", pid 356, jiffies 4294895306 (age 557.000s)
> hex dump (first 32 bytes):
> 88 13 00 00 00 00 00 00 a0 86 01 00 00 00 00 00 ................
> 00 00 00 00 00 00 00 00 0a 00 00 00 00 00 00 00 ................
> backtrace:
> [<0000000088e71769>] slab_post_alloc_hook+0xb8/0x368
> [<0000000073acab3b>] __kmem_cache_alloc_node+0x174/0x290
> [<00000000b5f89cef>] kmalloc_trace+0x40/0x164
> [<00000000eb19e83f>] damon_new_ctx+0x28/0xb4
> [<00000000058495c4>] damon_test_set_attrs+0x30/0x1a8
> [<00000000559c4801>] kunit_try_run_case+0x50/0xac
> [<000000003932ed49>] kunit_generic_run_threadfn_adapter+0x20/0x2c
> [<000000003c3e9211>] kthread+0x124/0x130
> [<0000000028f85bdd>] ret_from_fork+0x10/0x20
Same to the comment on the first patch of this patchset, simply letting others
know the name of the tool you used for making this cool output could be
helpful. Could you please add it?
Also, I prefer adding four spaces indent to quoted command outputs on commit
messages. Could you please add that indentation?
Also, I cannot 'git am' this patch on top of mm-unstable[1]. Could you please
double check and rebase on latest mm-unstable if I din't miss something?
[1] https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git/log/?h=mm-unstable
>
> Fixes: d1836a3b2a9a ("mm/damon/core-test: initialise context before test in damon_test_set_attrs()")
> Fixes: 4f540f5ab4f2 ("mm/damon/core-test: add a kunit test case for ops registration")
> Signed-off-by: Jinjie Ruan <ruanjinjie@xxxxxxxxxx>
> ---
> mm/damon/core-test.h | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/mm/damon/core-test.h b/mm/damon/core-test.h
> index 255f8c925c00..ce86ed30fb47 100644
> --- a/mm/damon/core-test.h
> +++ b/mm/damon/core-test.h
> @@ -266,6 +266,8 @@ static void damon_test_ops_registration(struct kunit *test)
>
> /* Check double-registration failure again */
> KUNIT_EXPECT_EQ(test, damon_register_ops(&ops), -EINVAL);
> +
> + damon_destroy_ctx(c);
> }
>
> static void damon_test_set_regions(struct kunit *test)
> @@ -342,6 +344,8 @@ static void damon_test_set_attrs(struct kunit *test)
> invalid_attrs = valid_attrs;
> invalid_attrs.aggr_interval = 4999;
> KUNIT_EXPECT_EQ(test, damon_set_attrs(c, &invalid_attrs), -EINVAL);
> +
> + damon_destroy_ctx(c);
> }
>
> static void damos_test_new_filter(struct kunit *test)
Other than above comments, changes look good to me.
> --
> 2.34.1
>
Thanks,
SJ