2021-05-28 07:41:11

by Takashi Iwai

[permalink] [raw]
Subject: [PATCH v3 1/2] Input: elants_i2c - Fix NULL dereference at probing

The recent change in elants_i2c driver to support more chips
introduced a regression leading to Oops at probing. The driver reads
id->driver_data, but the id may be NULL depending on the device type
the driver gets bound.

Replace the driver data extraction with the device_get_match_data()
helper, and define the driver data in OF table, too.

Fixes: 9517b95bdc46 ("Input: elants_i2c - add support for eKTF3624")
BugLink: https://bugzilla.suse.com/show_bug.cgi?id=1186454
Cc: <[email protected]>
Signed-off-by: Takashi Iwai <[email protected]>
---
v1->v2: Use device_get_match_data()
v2->v3: Fix pointer cast

drivers/input/touchscreen/elants_i2c.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/input/touchscreen/elants_i2c.c b/drivers/input/touchscreen/elants_i2c.c
index 17540bdb1eaf..0f9e3ec99aae 100644
--- a/drivers/input/touchscreen/elants_i2c.c
+++ b/drivers/input/touchscreen/elants_i2c.c
@@ -1396,7 +1396,7 @@ static int elants_i2c_probe(struct i2c_client *client,
init_completion(&ts->cmd_done);

ts->client = client;
- ts->chip_id = (enum elants_chip_id)id->driver_data;
+ ts->chip_id = (enum elants_chip_id)(uintptr_t)device_get_match_data(&client->dev);
i2c_set_clientdata(client, ts);

ts->vcc33 = devm_regulator_get(&client->dev, "vcc33");
@@ -1636,8 +1636,8 @@ MODULE_DEVICE_TABLE(acpi, elants_acpi_id);

#ifdef CONFIG_OF
static const struct of_device_id elants_of_match[] = {
- { .compatible = "elan,ekth3500" },
- { .compatible = "elan,ektf3624" },
+ { .compatible = "elan,ekth3500", .data = (void *)EKTH3500 },
+ { .compatible = "elan,ektf3624", .data = (void *)EKTF3624 },
{ /* sentinel */ }
};
MODULE_DEVICE_TABLE(of, elants_of_match);
--
2.26.2


2021-05-28 11:14:58

by Takashi Iwai

[permalink] [raw]
Subject: [PATCH v3 2/2] Input: elants_i2c - Switch to probe_new

Now that we get rid of the usage of id argument at probe again, let's
switch to the new i2c probe method; this will avoid for people
misusing the possibly unassigned id pointer again.

Signed-off-by: Takashi Iwai <[email protected]>
---
drivers/input/touchscreen/elants_i2c.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/input/touchscreen/elants_i2c.c b/drivers/input/touchscreen/elants_i2c.c
index 0f9e3ec99aae..68f542bb809f 100644
--- a/drivers/input/touchscreen/elants_i2c.c
+++ b/drivers/input/touchscreen/elants_i2c.c
@@ -1369,8 +1369,7 @@ static bool elants_acpi_is_hid_device(struct device *dev)
}
#endif

-static int elants_i2c_probe(struct i2c_client *client,
- const struct i2c_device_id *id)
+static int elants_i2c_probe(struct i2c_client *client)
{
union i2c_smbus_data dummy;
struct elants_data *ts;
@@ -1644,7 +1643,7 @@ MODULE_DEVICE_TABLE(of, elants_of_match);
#endif

static struct i2c_driver elants_i2c_driver = {
- .probe = elants_i2c_probe,
+ .probe_new = elants_i2c_probe,
.id_table = elants_i2c_id,
.driver = {
.name = DEVICE_NAME,
--
2.26.2

2021-06-02 04:24:21

by Dmitry Torokhov

[permalink] [raw]
Subject: Re: [PATCH v3 1/2] Input: elants_i2c - Fix NULL dereference at probing

On Fri, May 28, 2021 at 09:10:23AM +0200, Takashi Iwai wrote:
> The recent change in elants_i2c driver to support more chips
> introduced a regression leading to Oops at probing. The driver reads
> id->driver_data, but the id may be NULL depending on the device type
> the driver gets bound.
>
> Replace the driver data extraction with the device_get_match_data()
> helper, and define the driver data in OF table, too.
>
> Fixes: 9517b95bdc46 ("Input: elants_i2c - add support for eKTF3624")
> BugLink: https://bugzilla.suse.com/show_bug.cgi?id=1186454
> Cc: <[email protected]>
> Signed-off-by: Takashi Iwai <[email protected]>

Applied, thank you.

--
Dmitry

2021-06-02 04:24:31

by Dmitry Torokhov

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] Input: elants_i2c - Switch to probe_new

On Fri, May 28, 2021 at 09:10:24AM +0200, Takashi Iwai wrote:
> Now that we get rid of the usage of id argument at probe again, let's
> switch to the new i2c probe method; this will avoid for people
> misusing the possibly unassigned id pointer again.
>
> Signed-off-by: Takashi Iwai <[email protected]>

Applied, thank you.

--
Dmitry