[PATCH v2 00/11] Remove use of i2c_match_id in PMBUS

[PATCH v2 00/11] Remove use of i2c_match_id in PMBUS

[Andrew Davis]

Hello all,

Much like we already have done for HWMON top level drivers,
we are removing the i2c_match_id() function from all hwmon/pmbus/
drivers now. Using i2c_get_match_data() can simplify code and has
some other benefits described in the patches.

I don't have most of this hardware, so testing is very welcome :)

Thanks,
Andrew

Changes for v2:
 - Moved enum value [2/11] to prevent 0 from matching valid device
 - Trimmed all commit messages to only patch relevant info

Andrew Davis (11):
  hwmon: (pmbus/bel-pfe) Remove use of i2c_match_id()
  hwmon: (pmbus/ibm-cffps) Remove use of i2c_match_id()
  hwmon: (pmbus/isl68137) Remove use of i2c_match_id()
  hwmon: (pmbus/max20730) Remove use of i2c_match_id()
  hwmon: (pmbus/max34440) Remove use of i2c_match_id()
  hwmon: (pmbus) Remove use of i2c_match_id()
  hwmon: (pmbus/q54sj108a2) Remove use of i2c_match_id()
  hwmon: (pmbus/tps53679) Remove use of i2c_match_id()
  hwmon: (pmbus/fsp-3y) Remove use of i2c_match_id()
  hwmon: (pmbus/ltc2978) Remove use of i2c_match_id()
  hwmon: (pmbus/max16601) Remove use of i2c_match_id()

 drivers/hwmon/pmbus/bel-pfe.c    |  5 +----
 drivers/hwmon/pmbus/fsp-3y.c     | 13 ++++++-------
 drivers/hwmon/pmbus/ibm-cffps.c  | 16 +++-------------
 drivers/hwmon/pmbus/isl68137.c   |  4 +---
 drivers/hwmon/pmbus/ltc2978.c    |  2 +-
 drivers/hwmon/pmbus/max16601.c   |  2 +-
 drivers/hwmon/pmbus/max20730.c   |  5 +----
 drivers/hwmon/pmbus/max34440.c   |  4 +---
 drivers/hwmon/pmbus/pmbus.c      |  4 +---
 drivers/hwmon/pmbus/q54sj108a2.c |  5 +----
 drivers/hwmon/pmbus/tps53679.c   |  5 +----
 11 files changed, 18 insertions(+), 47 deletions(-)

-- 
2.39.2

[PATCH v2 01/11] hwmon: (pmbus/bel-pfe) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has another benefit:
 * It doesn't need the i2c_device_id passed in so we do not need
   to have that forward declared, allowing us to remove that.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/bel-pfe.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/hwmon/pmbus/bel-pfe.c b/drivers/hwmon/pmbus/bel-pfe.c
index ddf9d9a2958c4..6499556f735b4 100644
--- a/drivers/hwmon/pmbus/bel-pfe.c
+++ b/drivers/hwmon/pmbus/bel-pfe.c
@@ -88,13 +88,10 @@ static struct pmbus_driver_info pfe_driver_info[] = {
 	},
 };
 
-static const struct i2c_device_id pfe_device_id[];
-
 static int pfe_pmbus_probe(struct i2c_client *client)
 {
-	int model;
+	int model = (uintptr_t)i2c_get_match_data(client);
 
-	model = (int)i2c_match_id(pfe_device_id, client)->driver_data;
 	client->dev.platform_data = &pfe_plat_data;
 
 	/*
-- 
2.39.2

[PATCH v2 02/11] hwmon: (pmbus/ibm-cffps) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has another benefit:
 * It also checks for device match data, which allows for OF based
   probing. That means we do not have to manually check those first
   and can remove that check.

As i2c_get_match_data() return NULL/0 on failure which also matches
the enum for "cffps1", switch around the enum order so cffps_unknown
is index 0 and existing behavior is preserved.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/ibm-cffps.c | 16 +++-------------
 1 file changed, 3 insertions(+), 13 deletions(-)

diff --git a/drivers/hwmon/pmbus/ibm-cffps.c b/drivers/hwmon/pmbus/ibm-cffps.c
index d05ef7a968a96..6c7256d997f46 100644
--- a/drivers/hwmon/pmbus/ibm-cffps.c
+++ b/drivers/hwmon/pmbus/ibm-cffps.c
@@ -58,7 +58,7 @@ enum {
 	CFFPS_DEBUGFS_NUM_ENTRIES
 };
 
-enum versions { cffps1, cffps2, cffps_unknown };
+enum versions { cffps_unknown, cffps1, cffps2 };
 
 struct ibm_cffps {
 	enum versions version;
@@ -482,19 +482,9 @@ MODULE_DEVICE_TABLE(i2c, ibm_cffps_id);
 static int ibm_cffps_probe(struct i2c_client *client)
 {
 	int i, rc;
-	enum versions vs = cffps_unknown;
+	enum versions vs = (uintptr_t)i2c_get_match_data(client);
 	struct dentry *debugfs;
 	struct ibm_cffps *psu;
-	const void *md = of_device_get_match_data(&client->dev);
-	const struct i2c_device_id *id;
-
-	if (md) {
-		vs = (uintptr_t)md;
-	} else {
-		id = i2c_match_id(ibm_cffps_id, client);
-		if (id)
-			vs = (enum versions)id->driver_data;
-	}
 
 	if (vs == cffps_unknown) {
 		u16 ccin_revision = 0;
@@ -534,7 +524,7 @@ static int ibm_cffps_probe(struct i2c_client *client)
 		}
 
 		/* Set the client name to include the version number. */
-		snprintf(client->name, I2C_NAME_SIZE, "cffps%d", vs + 1);
+		snprintf(client->name, I2C_NAME_SIZE, "cffps%d", vs);
 	}
 
 	client->dev.platform_data = &ibm_cffps_pdata;
-- 
2.39.2

[PATCH v2 03/11] hwmon: (pmbus/isl68137) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has a couple other benefits:
 * It doesn't need the i2c_device_id passed in so we do not need
   to have that forward declared, allowing us to remove that.
 * It also checks for device match data, which allows for OF and
   ACPI based probing.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/isl68137.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/hwmon/pmbus/isl68137.c b/drivers/hwmon/pmbus/isl68137.c
index 97b61836f53a4..78cff97128541 100644
--- a/drivers/hwmon/pmbus/isl68137.c
+++ b/drivers/hwmon/pmbus/isl68137.c
@@ -90,8 +90,6 @@ struct isl68137_data {
 
 #define to_isl68137_data(x)	container_of(x, struct isl68137_data, info)
 
-static const struct i2c_device_id raa_dmpvr_id[];
-
 static ssize_t isl68137_avs_enable_show_page(struct i2c_client *client,
 					     int page,
 					     char *buf)
@@ -375,7 +373,7 @@ static int isl68137_probe(struct i2c_client *client)
 	memcpy(&data->info, &raa_dmpvr_info, sizeof(data->info));
 	info = &data->info;
 
-	switch (i2c_match_id(raa_dmpvr_id, client)->driver_data) {
+	switch ((uintptr_t)i2c_get_match_data(client)) {
 	case raa_dmpvr1_2rail:
 		info->pages = 2;
 		info->R[PSC_VOLTAGE_IN] = 3;
-- 
2.39.2

[PATCH v2 04/11] hwmon: (pmbus/max20730) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has another benefit:
 * It also checks for device match data. That means we do not have
   to manually check that first.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/max20730.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/hwmon/pmbus/max20730.c b/drivers/hwmon/pmbus/max20730.c
index 95869d198ecf0..fe03164788dfa 100644
--- a/drivers/hwmon/pmbus/max20730.c
+++ b/drivers/hwmon/pmbus/max20730.c
@@ -715,10 +715,7 @@ static int max20730_probe(struct i2c_client *client)
 		return -ENODEV;
 	}
 
-	if (client->dev.of_node)
-		chip_id = (uintptr_t)of_device_get_match_data(dev);
-	else
-		chip_id = i2c_match_id(max20730_id, client)->driver_data;
+	chip_id = (uintptr_t)i2c_get_match_data(client);
 
 	data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
 	if (!data)
-- 
2.39.2

[PATCH v2 05/11] hwmon: (pmbus/max34440) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has another benefit:
 * It doesn't need the i2c_device_id passed in so we do not need
   to have that forward declared, allowing us to remove that.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/max34440.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/hwmon/pmbus/max34440.c b/drivers/hwmon/pmbus/max34440.c
index 8ea4e68d4e9de..cc96bb22f8f5a 100644
--- a/drivers/hwmon/pmbus/max34440.c
+++ b/drivers/hwmon/pmbus/max34440.c
@@ -71,8 +71,6 @@ struct max34440_data {
 
 #define to_max34440_data(x)  container_of(x, struct max34440_data, info)
 
-static const struct i2c_device_id max34440_id[];
-
 static int max34440_read_word_data(struct i2c_client *client, int page,
 				   int phase, int reg)
 {
@@ -628,7 +626,7 @@ static int max34440_probe(struct i2c_client *client)
 			    GFP_KERNEL);
 	if (!data)
 		return -ENOMEM;
-	data->id = i2c_match_id(max34440_id, client)->driver_data;
+	data->id = (uintptr_t)i2c_get_match_data(client);
 	data->info = max34440_info[data->id];
 	data->iout_oc_fault_limit = MAX34440_IOUT_OC_FAULT_LIMIT;
 	data->iout_oc_warn_limit = MAX34440_IOUT_OC_WARN_LIMIT;
-- 
2.39.2

[PATCH v2 06/11] hwmon: (pmbus) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has another benefit:
 * It doesn't need the i2c_device_id passed in so we do not need
   to have that forward declared, allowing us to remove that.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/pmbus.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/hwmon/pmbus/pmbus.c b/drivers/hwmon/pmbus/pmbus.c
index 920cd5408141a..d1844c7a51eef 100644
--- a/drivers/hwmon/pmbus/pmbus.c
+++ b/drivers/hwmon/pmbus/pmbus.c
@@ -20,8 +20,6 @@ struct pmbus_device_info {
 	u32 flags;
 };
 
-static const struct i2c_device_id pmbus_id[];
-
 /*
  * Find sensor groups and status registers on each page.
  */
@@ -174,7 +172,7 @@ static int pmbus_probe(struct i2c_client *client)
 	if (!info)
 		return -ENOMEM;
 
-	device_info = (struct pmbus_device_info *)i2c_match_id(pmbus_id, client)->driver_data;
+	device_info = (struct pmbus_device_info *)i2c_get_match_data(client);
 	if (device_info->flags) {
 		pdata = devm_kzalloc(dev, sizeof(struct pmbus_platform_data),
 				     GFP_KERNEL);
-- 
2.39.2

[PATCH v2 07/11] hwmon: (pmbus/q54sj108a2) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has another benefit:
 * It also checks for device match data, which means we do not have
   to manually check that first.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/q54sj108a2.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/hwmon/pmbus/q54sj108a2.c b/drivers/hwmon/pmbus/q54sj108a2.c
index fc030ca34480c..88c25290ffce1 100644
--- a/drivers/hwmon/pmbus/q54sj108a2.c
+++ b/drivers/hwmon/pmbus/q54sj108a2.c
@@ -291,10 +291,7 @@ static int q54sj108a2_probe(struct i2c_client *client)
 				     I2C_FUNC_SMBUS_BLOCK_DATA))
 		return -ENODEV;
 
-	if (client->dev.of_node)
-		chip_id = (enum chips)(unsigned long)of_device_get_match_data(dev);
-	else
-		chip_id = i2c_match_id(q54sj108a2_id, client)->driver_data;
+	chip_id = (enum chips)(uintptr_t)i2c_get_match_data(client);
 
 	ret = i2c_smbus_read_block_data(client, PMBUS_MFR_ID, buf);
 	if (ret < 0) {
-- 
2.39.2

[PATCH v2 08/11] hwmon: (pmbus/tps53679) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().

This helper has another benefit:
 * It also checks for device match data, which means we do not have
   to manually check that first.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/tps53679.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/hwmon/pmbus/tps53679.c b/drivers/hwmon/pmbus/tps53679.c
index ca2bfa25eb04c..df2726659a4ed 100644
--- a/drivers/hwmon/pmbus/tps53679.c
+++ b/drivers/hwmon/pmbus/tps53679.c
@@ -253,10 +253,7 @@ static int tps53679_probe(struct i2c_client *client)
 	struct pmbus_driver_info *info;
 	enum chips chip_id;
 
-	if (dev->of_node)
-		chip_id = (uintptr_t)of_device_get_match_data(dev);
-	else
-		chip_id = i2c_match_id(tps53679_id, client)->driver_data;
+	chip_id = (uintptr_t)i2c_get_match_data(client);
 
 	info = devm_kmemdup(dev, &tps53679_info, sizeof(*info), GFP_KERNEL);
 	if (!info)
-- 
2.39.2

[PATCH v2 09/11] hwmon: (pmbus/fsp-3y) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This can be done instead with
i2c_client_get_device_id() which doesn't need the i2c_device_id
passed in so we do not need to have that forward declared, allowing
us to move the i2c_device_id table down to its more natural spot
with the other module info.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/fsp-3y.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/hwmon/pmbus/fsp-3y.c b/drivers/hwmon/pmbus/fsp-3y.c
index a4dc09e2ef75a..cad4d23300036 100644
--- a/drivers/hwmon/pmbus/fsp-3y.c
+++ b/drivers/hwmon/pmbus/fsp-3y.c
@@ -222,12 +222,6 @@ static int fsp3y_detect(struct i2c_client *client)
 	return -ENODEV;
 }
 
-static const struct i2c_device_id fsp3y_id[] = {
-	{"ym2151e", ym2151e},
-	{"yh5151e", yh5151e},
-	{ }
-};
-
 static int fsp3y_probe(struct i2c_client *client)
 {
 	struct fsp3y_data *data;
@@ -242,7 +236,7 @@ static int fsp3y_probe(struct i2c_client *client)
 	if (data->chip < 0)
 		return data->chip;
 
-	id = i2c_match_id(fsp3y_id, client);
+	id = i2c_client_get_device_id(client);
 	if (data->chip != id->driver_data)
 		dev_warn(&client->dev, "Device mismatch: Configured %s (%d), detected %d\n",
 			 id->name, (int)id->driver_data, data->chip);
@@ -276,6 +270,11 @@ static int fsp3y_probe(struct i2c_client *client)
 	return pmbus_do_probe(client, &data->info);
 }
 
+static const struct i2c_device_id fsp3y_id[] = {
+	{"ym2151e", ym2151e},
+	{"yh5151e", yh5151e},
+	{ }
+};
 MODULE_DEVICE_TABLE(i2c, fsp3y_id);
 
 static struct i2c_driver fsp3y_driver = {
-- 
2.39.2

[PATCH v2 10/11] hwmon: (pmbus/ltc2978) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This can instead be done with
i2c_client_get_device_id(). For this driver functionality should
not change. Switch over to remove the last couple users of the
i2c_match_id() function from kernel.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/ltc2978.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/hwmon/pmbus/ltc2978.c b/drivers/hwmon/pmbus/ltc2978.c
index 8f5be520a15db..d69a5e675e80e 100644
--- a/drivers/hwmon/pmbus/ltc2978.c
+++ b/drivers/hwmon/pmbus/ltc2978.c
@@ -733,7 +733,7 @@ static int ltc2978_probe(struct i2c_client *client)
 		return chip_id;
 
 	data->id = chip_id;
-	id = i2c_match_id(ltc2978_id, client);
+	id = i2c_client_get_device_id(client);
 	if (data->id != id->driver_data)
 		dev_warn(&client->dev,
 			 "Device mismatch: Configured %s (%d), detected %d\n",
-- 
2.39.2

[PATCH v2 11/11] hwmon: (pmbus/max16601) Remove use of i2c_match_id()

[Andrew Davis]

The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This can instead be done with
i2c_client_get_device_id(). For this driver functionality should
not change. Switch over to remove the last couple users of the
i2c_match_id() function from kernel.

Signed-off-by: Andrew Davis <afd@ti.com>
---
 drivers/hwmon/pmbus/max16601.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/hwmon/pmbus/max16601.c b/drivers/hwmon/pmbus/max16601.c
index d696e506aafba..36dc13424d929 100644
--- a/drivers/hwmon/pmbus/max16601.c
+++ b/drivers/hwmon/pmbus/max16601.c
@@ -318,7 +318,7 @@ static int max16601_probe(struct i2c_client *client)
 	if (chip_id < 0)
 		return chip_id;
 
-	id = i2c_match_id(max16601_id, client);
+	id = i2c_client_get_device_id(client);
 	if (chip_id != id->driver_data)
 		dev_warn(&client->dev,
 			 "Device mismatch: Configured %s (%d), detected %d\n",
-- 
2.39.2

Re: [PATCH v2 10/11] hwmon: (pmbus/ltc2978) Remove use of i2c_match_id()

[Guenter Roeck]

On Fri, Mar 06, 2026 at 11:16:51AM -0600, Andrew Davis wrote:
> The function i2c_match_id() is used to fetch the matching ID from
> the i2c_device_id table. This can instead be done with
> i2c_client_get_device_id(). For this driver functionality should
> not change. Switch over to remove the last couple users of the
> i2c_match_id() function from kernel.
> 
> Signed-off-by: Andrew Davis <afd@ti.com>
> ---
>  drivers/hwmon/pmbus/ltc2978.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/hwmon/pmbus/ltc2978.c b/drivers/hwmon/pmbus/ltc2978.c
> index 8f5be520a15db..d69a5e675e80e 100644
> --- a/drivers/hwmon/pmbus/ltc2978.c
> +++ b/drivers/hwmon/pmbus/ltc2978.c
> @@ -733,7 +733,7 @@ static int ltc2978_probe(struct i2c_client *client)
>  		return chip_id;
>  
>  	data->id = chip_id;
> -	id = i2c_match_id(ltc2978_id, client);
> +	id = i2c_client_get_device_id(client);

AI feedback:

  Is `id` guaranteed to be non-NULL here?

  If the device is instantiated via ACPI `PRP0001` or using a fallback DT
  compatible string where the first compatible string is not in the
  `ltc2978_id` table, `i2c_client_get_device_id()` will return `NULL`.
  This leads to a NULL pointer dereference when accessing `id->driver_data`.

  While this vulnerability existed in the old code with `i2c_match_id()`,
  adding a NULL check here might be a good idea while the code is being
  refactored.

I never know if this is real. Any idea ?

Thanks,
Guenter

>  	if (data->id != id->driver_data)
>  		dev_warn(&client->dev,
>  			 "Device mismatch: Configured %s (%d), detected %d\n",
> -- 
> 2.39.2
> 

Re: [PATCH v2 00/11] Remove use of i2c_match_id in PMBUS

[Guenter Roeck]

On 3/6/26 09:16, Andrew Davis wrote:
> Hello all,
> 
> Much like we already have done for HWMON top level drivers,
> we are removing the i2c_match_id() function from all hwmon/pmbus/
> drivers now. Using i2c_get_match_data() can simplify code and has
> some other benefits described in the patches.
> 
> I don't have most of this hardware, so testing is very welcome :)
> 
It looks safe to me, so I'll apply the series. There is a potential
pre-existing problem in the ltc2978 driver, but that is really unrelated
to this series, and it is better to fix that separately anyway.

Thanks,
Guenter

> Thanks,
> Andrew
> 
> Changes for v2:
>   - Moved enum value [2/11] to prevent 0 from matching valid device
>   - Trimmed all commit messages to only patch relevant info
> 
> Andrew Davis (11):
>    hwmon: (pmbus/bel-pfe) Remove use of i2c_match_id()
>    hwmon: (pmbus/ibm-cffps) Remove use of i2c_match_id()
>    hwmon: (pmbus/isl68137) Remove use of i2c_match_id()
>    hwmon: (pmbus/max20730) Remove use of i2c_match_id()
>    hwmon: (pmbus/max34440) Remove use of i2c_match_id()
>    hwmon: (pmbus) Remove use of i2c_match_id()
>    hwmon: (pmbus/q54sj108a2) Remove use of i2c_match_id()
>    hwmon: (pmbus/tps53679) Remove use of i2c_match_id()
>    hwmon: (pmbus/fsp-3y) Remove use of i2c_match_id()
>    hwmon: (pmbus/ltc2978) Remove use of i2c_match_id()
>    hwmon: (pmbus/max16601) Remove use of i2c_match_id()
> 
>   drivers/hwmon/pmbus/bel-pfe.c    |  5 +----
>   drivers/hwmon/pmbus/fsp-3y.c     | 13 ++++++-------
>   drivers/hwmon/pmbus/ibm-cffps.c  | 16 +++-------------
>   drivers/hwmon/pmbus/isl68137.c   |  4 +---
>   drivers/hwmon/pmbus/ltc2978.c    |  2 +-
>   drivers/hwmon/pmbus/max16601.c   |  2 +-
>   drivers/hwmon/pmbus/max20730.c   |  5 +----
>   drivers/hwmon/pmbus/max34440.c   |  4 +---
>   drivers/hwmon/pmbus/pmbus.c      |  4 +---
>   drivers/hwmon/pmbus/q54sj108a2.c |  5 +----
>   drivers/hwmon/pmbus/tps53679.c   |  5 +----
>   11 files changed, 18 insertions(+), 47 deletions(-)
> 

Re: [PATCH v2 10/11] hwmon: (pmbus/ltc2978) Remove use of i2c_match_id()

[Andrew Davis]

On 3/6/26 12:10 PM, Guenter Roeck wrote:
> On Fri, Mar 06, 2026 at 11:16:51AM -0600, Andrew Davis wrote:
>> The function i2c_match_id() is used to fetch the matching ID from
>> the i2c_device_id table. This can instead be done with
>> i2c_client_get_device_id(). For this driver functionality should
>> not change. Switch over to remove the last couple users of the
>> i2c_match_id() function from kernel.
>>
>> Signed-off-by: Andrew Davis <afd@ti.com>
>> ---
>>   drivers/hwmon/pmbus/ltc2978.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/hwmon/pmbus/ltc2978.c b/drivers/hwmon/pmbus/ltc2978.c
>> index 8f5be520a15db..d69a5e675e80e 100644
>> --- a/drivers/hwmon/pmbus/ltc2978.c
>> +++ b/drivers/hwmon/pmbus/ltc2978.c
>> @@ -733,7 +733,7 @@ static int ltc2978_probe(struct i2c_client *client)
>>   		return chip_id;
>>   
>>   	data->id = chip_id;
>> -	id = i2c_match_id(ltc2978_id, client);
>> +	id = i2c_client_get_device_id(client);
> 
> AI feedback:
> 
>    Is `id` guaranteed to be non-NULL here?
> 
>    If the device is instantiated via ACPI `PRP0001` or using a fallback DT
>    compatible string where the first compatible string is not in the
>    `ltc2978_id` table, `i2c_client_get_device_id()` will return `NULL`.
>    This leads to a NULL pointer dereference when accessing `id->driver_data`.
> 
>    While this vulnerability existed in the old code with `i2c_match_id()`,
>    adding a NULL check here might be a good idea while the code is being
>    refactored.
> 
> I never know if this is real. Any idea ?
> 

The AI is right on both parts, the second being the important one that
this was preexisting. i2c_match_id() should never return NULL in
practice as we must have matched on something to have gotten probe()'d
in the first place. And for the same reason i2c_client_get_device_id()
shouldn't ever return NULL either. So no change here.

But I see how this would confuse an AI as both functions have a
"return NULL;" statement in them. This problem of returning the matched
ID in-band with returning 0 for errors (which could also be a valid ID)
is one of the motivating reasons for me removing i2c_match_id(). Even
though its replacement has the same issue, once we get everyone over
to the same single API then switching that one API over to something
safer becomes possible.

Andrew

> Thanks,
> Guenter
> 
>>   	if (data->id != id->driver_data)
>>   		dev_warn(&client->dev,
>>   			 "Device mismatch: Configured %s (%d), detected %d\n",
>> -- 
>> 2.39.2
>>