Skip to contentSkip to conversation

i2c: i801: Use PCI bus rescan mutex to protect P2SB access

PATCH Applied Heiner Kallweit 4 participants 4 messages Sep 5, 2021 – Sep 29, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 i2c: i801: Use PCI bus rescan mutex to protect P2SB access

As pointed out by Andy in [0] using a local mutex here isn't strictly wrong but not sufficient. We should hold the PCI rescan lock for P2SB operations.

[0] https://www.spinics.net/lists/linux-i2c/msg52717.html

Fixes 1a987c69ce2c ("i2c: i801: make p2sb_spinlock a mutex")
---
 drivers/i2c/busses/i2c-i801.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/i2c/busses/i2c-i801.c +2 -3
index 89ae78ef1..1f929e6c3 100644
@@ -1493,7 +1493,6 @@ static struct platform_device *
 i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 		 struct resource *tco_res)
 {
-	static DEFINE_MUTEX(p2sb_mutex);
 	struct resource *res;
 	unsigned int devfn;
 	u64 base64_addr;
@@ -1506,7 +1505,7 @@ i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 	 * enumerated by the PCI subsystem, so we need to unhide/hide it
 	 * to lookup the P2SB BAR.
 	 */
-	mutex_lock(&p2sb_mutex);
+	pci_lock_rescan_remove();
 
 	devfn = PCI_DEVFN(PCI_SLOT(pci_dev->devfn), 1);
 
@@ -1524,7 +1523,7 @@ i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 	/* Hide the P2SB device, if it was hidden before */
 	if (hidden)
 		pci_bus_write_config_byte(pci_dev->bus, devfn, 0xe1, hidden);
-	mutex_unlock(&p2sb_mutex);
+	pci_unlock_rescan_remove();
 
 	res = &tco_res[1];
 	if (pci_dev->device == PCI_DEVICE_ID_INTEL_DNV_SMBUS)
Signature · 1 line
2.33.0
Reviewed-by on v1 1/1
#1 lore raw
Quoted 3 lines · Heiner Kallweit

As pointed out by Andy in [0] using a local mutex here isn't strictly wrong but not sufficient. We should hold the PCI rescan lock for P2SB operations.

Thanks!

Quoted 42 lines

[0] https://www.spinics.net/lists/linux-i2c/msg52717.html

Fixes 1a987c69ce2c ("i2c: i801: make p2sb_spinlock a mutex")
---
 drivers/i2c/busses/i2c-i801.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/i2c/busses/i2c-i801.c +2 -3
index 89ae78ef1..1f929e6c3 100644
@@ -1493,7 +1493,6 @@ static struct platform_device *
 i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 		 struct resource *tco_res)
 {
-	static DEFINE_MUTEX(p2sb_mutex);
 	struct resource *res;
 	unsigned int devfn;
 	u64 base64_addr;
@@ -1506,7 +1505,7 @@ i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 	 * enumerated by the PCI subsystem, so we need to unhide/hide it
 	 * to lookup the P2SB BAR.
 	 */
-	mutex_lock(&p2sb_mutex);
+	pci_lock_rescan_remove();
 
 	devfn = PCI_DEVFN(PCI_SLOT(pci_dev->devfn), 1);
 
@@ -1524,7 +1523,7 @@ i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 	/* Hide the P2SB device, if it was hidden before */
 	if (hidden)
 		pci_bus_write_config_byte(pci_dev->bus, devfn, 0xe1, hidden);
-	mutex_unlock(&p2sb_mutex);
+	pci_unlock_rescan_remove();
 
 	res = &tco_res[1];
 	if (pci_dev->device == PCI_DEVICE_ID_INTEL_DNV_SMBUS)
Signature · 1 line
2.33.0
Signature · 2 lines
With Best Regards,
Andy Shevchenko
Reviewed-by on v1 1/1
#2 lore raw
In reply to Heiner Kallweit

Hi Heiner,

Quoted 35 lines · Heiner Kallweit

As pointed out by Andy in [0] using a local mutex here isn't strictly wrong but not sufficient. We should hold the PCI rescan lock for P2SB operations.

[0] https://www.spinics.net/lists/linux-i2c/msg52717.html

Fixes 1a987c69ce2c ("i2c: i801: make p2sb_spinlock a mutex")
---
 drivers/i2c/busses/i2c-i801.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/i2c/busses/i2c-i801.c +1 -2
index 89ae78ef1..1f929e6c3 100644
@@ -1493,7 +1493,6 @@ static struct platform_device *
 i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 		 struct resource *tco_res)
 {
-	static DEFINE_MUTEX(p2sb_mutex);
 	struct resource *res;
 	unsigned int devfn;
 	u64 base64_addr;
@@ -1506,7 +1505,7 @@ i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 	 * enumerated by the PCI subsystem, so we need to unhide/hide it
 	 * to lookup the P2SB BAR.
 	 */
-	mutex_lock(&p2sb_mutex);
+	pci_lock_rescan_remove();
 
 	devfn = PCI_DEVFN(PCI_SLOT(pci_dev->devfn), 1);
 

drivers/i2c/busses/i2c-i801.c i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev, · line 1529

Quoted 8 lines
drivers/i2c/busses/i2c-i801.c +1 -1
@@ -1524,7 +1523,7 @@ i801_add_tco_spt(struct i801_priv *priv, struct pci_dev *pci_dev,
 	/* Hide the P2SB device, if it was hidden before */
 	if (hidden)
 		pci_bus_write_config_byte(pci_dev->bus, devfn, 0xe1, hidden);
-	mutex_unlock(&p2sb_mutex);
+	pci_unlock_rescan_remove();
 
 	res = &tco_res[1];
 	if (pci_dev->device == PCI_DEVICE_ID_INTEL_DNV_SMBUS)

Looks good to me.

Signature · 2 lines
Jean Delvare
SUSE L3 Support
Applied
#3 lore raw
In reply to Heiner Kallweit
Quoted 9 lines · Heiner Kallweit

As pointed out by Andy in [0] using a local mutex here isn't strictly wrong but not sufficient. We should hold the PCI rescan lock for P2SB operations.

[0] https://www.spinics.net/lists/linux-i2c/msg52717.html

Fixes 1a987c69ce2c ("i2c: i801: make p2sb_spinlock a mutex")

Applied to for-next, thanks!