There are memory leaks reported by kmemleak:
...
unreferenced object 0xffff00213c141000 (size 1024):
comm "systemd-udevd", pid 2123, jiffies 4294909467 (age 6062.160s)
hex dump (first 32 bytes):
04 00 00 00 02 00 00 00 18 10 14 3c 21 00 ff ff ...........<!...
00 00 00 00 00 00 00 00 03 00 00 00 10 00 00 00 ................
backtrace:
[<000000004b7c9001>] __kmem_cache_alloc_node+0x2f8/0x348
[<00000000b0fc7ceb>] __kmalloc+0x58/0x108
[<0000000064ff4695>] acpi_os_allocate+0x2c/0x68
[<000000007d57d116>] acpi_ut_initialize_buffer+0x54/0xe0
[<0000000024583908>] acpi_evaluate_object+0x388/0x438
[<0000000017b2e72b>] acpi_evaluate_object_typed+0xe8/0x240
[<000000005df0eac2>] coresight_get_platform_data+0x1b4/0x988 [coresight]
...
The ACPI buffer memory (buf.pointer) should be freed. But the buffer
is also used after returning from acpi_get_dsd_graph().
Move the temporary variables buf to acpi_coresight_parse_graph(),
and free it before the function return to prevent memory leak.
Fixes: 76ffa5ab5b79 ("coresight: Support for ACPI bindings")
Signed-off-by: Junhao He <hejunhao3@huawei.com>
---
.../hwtracing/coresight/coresight-platform.c | 40 ++++++++++++-------
1 file changed, 26 insertions(+), 14 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index 7d7b641c0a71..9d550f5697fa 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -492,19 +492,18 @@ static inline bool acpi_validate_dsd_graph(const union acpi_object *graph)
/* acpi_get_dsd_graph - Find the _DSD Graph property for the given device. */
static const union acpi_object *
-acpi_get_dsd_graph(struct acpi_device *adev)
+acpi_get_dsd_graph(struct acpi_device *adev, struct acpi_buffer *buf)
{
int i;
- struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER };
acpi_status status;
const union acpi_object *dsd;
status = acpi_evaluate_object_typed(adev->handle, "_DSD", NULL,
- &buf, ACPI_TYPE_PACKAGE);
+ buf, ACPI_TYPE_PACKAGE);
if (ACPI_FAILURE(status))
return NULL;
- dsd = buf.pointer;
+ dsd = buf->pointer;
/*
* _DSD property consists tuples { Prop_UUID, Package() }
@@ -555,12 +554,12 @@ acpi_validate_coresight_graph(const union acpi_object *cs_graph)
* returns NULL.
*/
static const union acpi_object *
-acpi_get_coresight_graph(struct acpi_device *adev)
+acpi_get_coresight_graph(struct acpi_device *adev, struct acpi_buffer *buf)
{
const union acpi_object *graph_list, *graph;
int i, nr_graphs;
- graph_list = acpi_get_dsd_graph(adev);
+ graph_list = acpi_get_dsd_graph(adev, buf);
if (!graph_list)
return graph_list;
@@ -661,22 +660,24 @@ static int acpi_coresight_parse_graph(struct device *dev,
struct acpi_device *adev,
struct coresight_platform_data *pdata)
{
+ int ret = 0;
int i, nlinks;
const union acpi_object *graph;
struct coresight_connection conn, zero_conn = {};
struct coresight_connection *new_conn;
+ struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER, NULL };
- graph = acpi_get_coresight_graph(adev);
+ graph = acpi_get_coresight_graph(adev, &buf);
/*
* There are no graph connections, which is fine for some components.
* e.g., ETE
*/
if (!graph)
- return 0;
+ goto free;
nlinks = graph->package.elements[2].integer.value;
if (!nlinks)
- return 0;
+ goto free;
for (i = 0; i < nlinks; i++) {
const union acpi_object *link = &graph->package.elements[3 + i];
@@ -684,17 +685,28 @@ static int acpi_coresight_parse_graph(struct device *dev,
conn = zero_conn;
dir = acpi_coresight_parse_link(adev, link, &conn);
- if (dir < 0)
- return dir;
+ if (dir < 0) {
+ ret = dir;
+ goto free;
+ }
if (dir == ACPI_CORESIGHT_LINK_MASTER) {
new_conn = coresight_add_out_conn(dev, pdata, &conn);
- if (IS_ERR(new_conn))
- return PTR_ERR(new_conn);
+ if (IS_ERR(new_conn)) {
+ ret = PTR_ERR(new_conn);
+ goto free;
+ }
}
}
- return 0;
+free:
+ /*
+ * When ACPI fails to alloc a buffer, it will free the buffer
+ * created via ACPI_ALLOCATE_BUFFER and set to NULL.
+ * ACPI_FREE can handle NULL pointers, so free it directly.
+ */
+ ACPI_FREE(buf.pointer);
+ return ret;
}
/*
--
2.33.0
On 17/08/2023 09:59, Junhao He wrote:
> There are memory leaks reported by kmemleak:
> ...
> unreferenced object 0xffff00213c141000 (size 1024):
> comm "systemd-udevd", pid 2123, jiffies 4294909467 (age 6062.160s)
> hex dump (first 32 bytes):
> 04 00 00 00 02 00 00 00 18 10 14 3c 21 00 ff ff ...........<!...
> 00 00 00 00 00 00 00 00 03 00 00 00 10 00 00 00 ................
> backtrace:
> [<000000004b7c9001>] __kmem_cache_alloc_node+0x2f8/0x348
> [<00000000b0fc7ceb>] __kmalloc+0x58/0x108
> [<0000000064ff4695>] acpi_os_allocate+0x2c/0x68
> [<000000007d57d116>] acpi_ut_initialize_buffer+0x54/0xe0
> [<0000000024583908>] acpi_evaluate_object+0x388/0x438
> [<0000000017b2e72b>] acpi_evaluate_object_typed+0xe8/0x240
> [<000000005df0eac2>] coresight_get_platform_data+0x1b4/0x988 [coresight]
> ...
>
> The ACPI buffer memory (buf.pointer) should be freed. But the buffer
> is also used after returning from acpi_get_dsd_graph().
> Move the temporary variables buf to acpi_coresight_parse_graph(),
> and free it before the function return to prevent memory leak.
>
> Fixes: 76ffa5ab5b79 ("coresight: Support for ACPI bindings")
> Signed-off-by: Junhao He <hejunhao3@huawei.com>
I confirmed that the error gone. Thanks for the fix.
Reviewed-by: James Clark <james.clark@arm.com>
> ---
> .../hwtracing/coresight/coresight-platform.c | 40 ++++++++++++-------
> 1 file changed, 26 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index 7d7b641c0a71..9d550f5697fa 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -492,19 +492,18 @@ static inline bool acpi_validate_dsd_graph(const union acpi_object *graph)
>
> /* acpi_get_dsd_graph - Find the _DSD Graph property for the given device. */
> static const union acpi_object *
> -acpi_get_dsd_graph(struct acpi_device *adev)
> +acpi_get_dsd_graph(struct acpi_device *adev, struct acpi_buffer *buf)
> {
> int i;
> - struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER };
> acpi_status status;
> const union acpi_object *dsd;
>
> status = acpi_evaluate_object_typed(adev->handle, "_DSD", NULL,
> - &buf, ACPI_TYPE_PACKAGE);
> + buf, ACPI_TYPE_PACKAGE);
> if (ACPI_FAILURE(status))
> return NULL;
>
> - dsd = buf.pointer;
> + dsd = buf->pointer;
>
> /*
> * _DSD property consists tuples { Prop_UUID, Package() }
> @@ -555,12 +554,12 @@ acpi_validate_coresight_graph(const union acpi_object *cs_graph)
> * returns NULL.
> */
> static const union acpi_object *
> -acpi_get_coresight_graph(struct acpi_device *adev)
> +acpi_get_coresight_graph(struct acpi_device *adev, struct acpi_buffer *buf)
> {
> const union acpi_object *graph_list, *graph;
> int i, nr_graphs;
>
> - graph_list = acpi_get_dsd_graph(adev);
> + graph_list = acpi_get_dsd_graph(adev, buf);
> if (!graph_list)
> return graph_list;
>
> @@ -661,22 +660,24 @@ static int acpi_coresight_parse_graph(struct device *dev,
> struct acpi_device *adev,
> struct coresight_platform_data *pdata)
> {
> + int ret = 0;
> int i, nlinks;
> const union acpi_object *graph;
> struct coresight_connection conn, zero_conn = {};
> struct coresight_connection *new_conn;
> + struct acpi_buffer buf = { ACPI_ALLOCATE_BUFFER, NULL };
>
> - graph = acpi_get_coresight_graph(adev);
> + graph = acpi_get_coresight_graph(adev, &buf);
> /*
> * There are no graph connections, which is fine for some components.
> * e.g., ETE
> */
> if (!graph)
> - return 0;
> + goto free;
>
> nlinks = graph->package.elements[2].integer.value;
> if (!nlinks)
> - return 0;
> + goto free;
>
> for (i = 0; i < nlinks; i++) {
> const union acpi_object *link = &graph->package.elements[3 + i];
> @@ -684,17 +685,28 @@ static int acpi_coresight_parse_graph(struct device *dev,
>
> conn = zero_conn;
> dir = acpi_coresight_parse_link(adev, link, &conn);
> - if (dir < 0)
> - return dir;
> + if (dir < 0) {
> + ret = dir;
> + goto free;
> + }
>
> if (dir == ACPI_CORESIGHT_LINK_MASTER) {
> new_conn = coresight_add_out_conn(dev, pdata, &conn);
> - if (IS_ERR(new_conn))
> - return PTR_ERR(new_conn);
> + if (IS_ERR(new_conn)) {
> + ret = PTR_ERR(new_conn);
> + goto free;
> + }
> }
> }
>
> - return 0;
> +free:
> + /*
> + * When ACPI fails to alloc a buffer, it will free the buffer
> + * created via ACPI_ALLOCATE_BUFFER and set to NULL.
> + * ACPI_FREE can handle NULL pointers, so free it directly.
> + */
> + ACPI_FREE(buf.pointer);
> + return ret;
> }
>
> /*
On 17/08/2023 15:03, James Clark wrote:
>
>
> On 17/08/2023 09:59, Junhao He wrote:
>> There are memory leaks reported by kmemleak:
>> ...
>> unreferenced object 0xffff00213c141000 (size 1024):
>> comm "systemd-udevd", pid 2123, jiffies 4294909467 (age 6062.160s)
>> hex dump (first 32 bytes):
>> 04 00 00 00 02 00 00 00 18 10 14 3c 21 00 ff ff ...........<!...
>> 00 00 00 00 00 00 00 00 03 00 00 00 10 00 00 00 ................
>> backtrace:
>> [<000000004b7c9001>] __kmem_cache_alloc_node+0x2f8/0x348
>> [<00000000b0fc7ceb>] __kmalloc+0x58/0x108
>> [<0000000064ff4695>] acpi_os_allocate+0x2c/0x68
>> [<000000007d57d116>] acpi_ut_initialize_buffer+0x54/0xe0
>> [<0000000024583908>] acpi_evaluate_object+0x388/0x438
>> [<0000000017b2e72b>] acpi_evaluate_object_typed+0xe8/0x240
>> [<000000005df0eac2>] coresight_get_platform_data+0x1b4/0x988 [coresight]
>> ...
>>
>> The ACPI buffer memory (buf.pointer) should be freed. But the buffer
>> is also used after returning from acpi_get_dsd_graph().
>> Move the temporary variables buf to acpi_coresight_parse_graph(),
>> and free it before the function return to prevent memory leak.
>>
>> Fixes: 76ffa5ab5b79 ("coresight: Support for ACPI bindings")
>> Signed-off-by: Junhao He <hejunhao3@huawei.com>
>
> I confirmed that the error gone. Thanks for the fix.
>
> Reviewed-by: James Clark <james.clark@arm.com>
Queued:
https://git.kernel.org/coresight/c/1a9e02673e25
Thanks!
Suzuki
© 2016 - 2025 Red Hat, Inc.