Messages in this thread | | | From | "Rafael J. Wysocki" <> | Date | Tue, 25 Oct 2022 17:57:20 +0200 | Subject | Re: [PATCH] PM: hibernate: Fix potential memory leak in hibernate_preallocate_memory() and prepare_highmem_image() |
| |
On Wed, Sep 7, 2022 at 8:43 AM Jianglei Nie <niejianglei2021@163.com> wrote: > > hibernate_preallocate_memory() and prepare_highmem_image() allocates > memory chunk with memory_bm_create(). When the function gets some error > after memory_bm_create(), relavent memory should be released with > memory_bm_free().
swsusp_free() called at the end is expected to free all of these pages. Doesn't it do that?
> Fix it by calling memory_bm_free() at the right time. > > Signed-off-by: Jianglei Nie <niejianglei2021@163.com> > --- > kernel/power/snapshot.c | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c > index 2a406753af90..e5ec204ebe22 100644 > --- a/kernel/power/snapshot.c > +++ b/kernel/power/snapshot.c > @@ -1752,6 +1752,7 @@ int hibernate_preallocate_memory(void) > > error = memory_bm_create(©_bm, GFP_IMAGE, PG_ANY); > if (error) { > + memory_bm_free(&orig_bm, PG_UNSAFE_CLEAR); > pr_err("Cannot allocate copy bitmap\n"); > goto err_out; > } > @@ -2335,8 +2336,10 @@ static int prepare_highmem_image(struct memory_bitmap *bm, > if (memory_bm_create(bm, GFP_ATOMIC, PG_SAFE)) > return -ENOMEM; > > - if (get_highmem_buffer(PG_SAFE)) > + if (get_highmem_buffer(PG_SAFE)) { > + memory_bm_free(bm, PG_UNSAFE_CLEAR); > return -ENOMEM; > + } > > to_alloc = count_free_highmem_pages(); > if (to_alloc > *nr_highmem_p) > -- > 2.25.1 >
| |