Skip to contentSkip to conversation

i2c: i801: Restore INTREN on unload

PATCH Applied Jean Delvare 3 participants 4 messages Nov 9, 2021 – Nov 23, 2021

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

PATCH 1/1 i2c: i801: Restore INTREN on unload

If driver interrupts are enabled, SMBHSTCNT_INTREN will be 1 after the first transaction, and will stay to that value forever. This means that interrupts will be generated for both host-initiated transactions and also SMBus Alert events even after the driver is unloaded. To be on the safe side, we should restore the initial state of this bit at suspend and reboot time, as we do for several other configuration bits already and for the same reason: the BIOS should be handed the device in the same configuration state in which we received it. Otherwise interrupts may be generated which nobody will process.

---
This probably doesn't change much on its own in practice, however it is mandatory to make this change before Jarkko's fix for the SMB_ALERT interrupt storm gets applied, otherwise the fix will be incomplete.

Jarkko, this is not exactly the patch you tested, I added restoration to the suspend path as well to be 100% safe.

 drivers/i2c/busses/i2c-i801.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/i2c/busses/i2c-i801.c +6 -1
@@ -258,6 +258,7 @@ struct i801_priv {
 	struct i2c_adapter adapter;
 	unsigned long smba;
 	unsigned char original_hstcfg;
+	unsigned char original_hstcnt;
 	unsigned char original_slvcmd;
 	struct pci_dev *pci_dev;
 	unsigned int features;
@@ -1799,7 +1800,8 @@ static int i801_probe(struct pci_dev *de
 		outb_p(inb_p(SMBAUXCTL(priv)) &
 		       ~(SMBAUXCTL_CRC | SMBAUXCTL_E32B), SMBAUXCTL(priv));
 
-	/* Remember original Host Notify setting */
+	/* Remember original Interrupt and Host Notify settings */
+	priv->original_hstcnt = inb_p(SMBHSTCNT(priv)) & ~SMBHSTCNT_KILL;
 	if (priv->features & FEATURE_HOST_NOTIFY)
 		priv->original_slvcmd = inb_p(SMBSLVCMD(priv));
 
@@ -1863,6 +1865,7 @@ static void i801_remove(struct pci_dev *
 {
 	struct i801_priv *priv = pci_get_drvdata(dev);
 
+	outb_p(priv->original_hstcnt, SMBHSTCNT(priv));
 	i801_disable_host_notify(priv);
 	i801_del_mux(priv);
 	i2c_del_adapter(&priv->adapter);
@@ -1886,6 +1889,7 @@ static void i801_shutdown(struct pci_dev
 	struct i801_priv *priv = pci_get_drvdata(dev);
 
 	/* Restore config registers to avoid hard hang on some systems */
+	outb_p(priv->original_hstcnt, SMBHSTCNT(priv));
 	i801_disable_host_notify(priv);
 	pci_write_config_byte(dev, SMBHSTCFG, priv->original_hstcfg);
 }
@@ -1895,6 +1899,7 @@ static int i801_suspend(struct device *d
 {
 	struct i801_priv *priv = dev_get_drvdata(dev);
 
+	outb_p(priv->original_hstcnt, SMBHSTCNT(priv));
 	pci_write_config_byte(priv->pci_dev, SMBHSTCFG, priv->original_hstcfg);
 	return 0;
 }
Signature · 2 lines
Jean Delvare
SUSE L3 Support
#1 lore raw
Quoted 13 lines · Jean Delvare

If driver interrupts are enabled, SMBHSTCNT_INTREN will be 1 after the first transaction, and will stay to that value forever. This means that interrupts will be generated for both host-initiated transactions and also SMBus Alert events even after the driver is unloaded. To be on the safe side, we should restore the initial state of this bit at suspend and reboot time, as we do for several other configuration bits already and for the same reason: the BIOS should be handed the device in the same configuration state in which we received it. Otherwise interrupts may be generated which nobody will process.

Quoted 8 lines

---
This probably doesn't change much on its own in practice, however it is mandatory to make this change before Jarkko's fix for the SMB_ALERT interrupt storm gets applied, otherwise the fix will be incomplete.

Jarkko, this is not exactly the patch you tested, I added restoration to the suspend path as well to be 100% safe.

Same Tested-by holds here.

We have one laptop with RMI4 SMBus connected touchpad and trackpad. I wanted to check it does the suspend path here cease them to not work but they work after your patch.

Now we restore INTREN on suspend but do not explicitly enable it after resume. I didn't fully get how RMI4 stack resumes but perhaps they will do some power on, reset, etc command transaction and that gets the INTREN enabled and allow host notify.

Jarkko

#2 lore raw
Quoted 2 lines · Jarkko Nikula

Now we restore INTREN on suspend but do not explicitly enable it after resume.

We also don't enable it when loading the driver, so this is consistent.

Quoted 3 lines

I didn't fully get how RMI4 stack resumes but perhaps they will do some power on, reset, etc command transaction and that gets the INTREN enabled and allow host notify.

I didn't check either, but it shouldn't matter. If INTREN has to be set for RMI4 to work, then it must have been set even before the i2c-i801 driver was loaded. In this case, my new code is a no-op (SMBHSTCNT will be set to the value it already had, that is, INTREN is enabled at all times).

Signature · 2 lines
Jean Delvare
SUSE L3 Support
Applied
#3 lore raw
In reply to Jean Delvare
Quoted 11 lines · Jean Delvare

If driver interrupts are enabled, SMBHSTCNT_INTREN will be 1 after the first transaction, and will stay to that value forever. This means that interrupts will be generated for both host-initiated transactions and also SMBus Alert events even after the driver is unloaded. To be on the safe side, we should restore the initial state of this bit at suspend and reboot time, as we do for several other configuration bits already and for the same reason: the BIOS should be handed the device in the same configuration state in which we received it. Otherwise interrupts may be generated which nobody will process.

Quoted 2 lines

Applied to for-current, thanks!