[PATCH v3 0/4] hwmon (it87): Correct handling for configuration mode

[PATCH v3 0/4] hwmon (it87): Correct handling for configuration mode

[Frank Crawford]

On various Gigabyte boards, incorrectly entering configuration mode
causes the second Super-IO chip to generate LPC bus access errors.
This was preivously fixed by ensuring that the second chip receives the
code to enter configuration mode before the first chip.

On discussion with people who have access to the specification documents
it was noted that this is wrong, and you should not enter or leave
configuration mode for the second chip, as it is enable during
initialisation and should not be changed.

In particular, this was found to be the case on the Gigabyte X670E Aorus
Master board, where it was reporting a totally wrong chip ID (0x8883)
using the previous method.  This was corrected by not entering
configuration mode, and this has been found to still work with older boards.

---

Changes since v1:
 * renamed the feature to FEAT_NOCONF and macro to has_noconf.
 * reworked patch descriptions to be more explicit about the actions and
   remove any reference to the BIOS.

Changes since v2:
 * rename variable from opened to enabled it87_find().
 * corrected final call to superio_exit() in it87_find().
 * minor update to patch documentation for it87_find() change.

Frank Crawford (4):
  Rename FEAT_CONF_NOEXIT to FEAT_NOCONF as more descriptive of requirement
  Do not enter configuration mode for some chiptypes
  Test for chipset before entering configuration mode
  Remove tests nolonger required

 drivers/hwmon/it87.c | 127 +++++++++++++++++++++----------------------
 1 file changed, 61 insertions(+), 66 deletions(-)

-- 
2.44.0

[PATCH v3 1/4] hwmon (it87): Rename FEAT_CONF_NOEXIT to FEAT_NOCONF as more descriptive of requirement

[Frank Crawford]

Rename previous definitions to match the new information that they are
preinitialised as enabled and should not receive codes to enter or exit
configuration mode.

Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>
---

v2:
 * renamed the feature to FEAT_NOCONF and macro to has_noconf following
   review.

v3:
 * No change.

---
 drivers/hwmon/it87.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/drivers/hwmon/it87.c b/drivers/hwmon/it87.c
index fbe86cec6055..6eeba3a01e3c 100644
--- a/drivers/hwmon/it87.c
+++ b/drivers/hwmon/it87.c
@@ -320,7 +320,7 @@ struct it87_devices {
  * second SIO address. Never exit configuration mode on these
  * chips to avoid the problem.
  */
-#define FEAT_CONF_NOEXIT	BIT(19)	/* Chip should not exit conf mode */
+#define FEAT_NOCONF		BIT(19)	/* Chip conf mode enabled on startup */
 #define FEAT_FOUR_FANS		BIT(20)	/* Supports four fans */
 #define FEAT_FOUR_PWM		BIT(21)	/* Supports four fan controls */
 #define FEAT_FOUR_TEMP		BIT(22)
@@ -452,7 +452,7 @@ static const struct it87_devices it87_devices[] = {
 		.model = "IT8790E",
 		.features = FEAT_NEWER_AUTOPWM | FEAT_12MV_ADC | FEAT_16BIT_FANS
 		  | FEAT_TEMP_OFFSET | FEAT_TEMP_PECI | FEAT_IN7_INTERNAL
-		  | FEAT_PWM_FREQ2 | FEAT_FANCTL_ONOFF | FEAT_CONF_NOEXIT,
+		  | FEAT_PWM_FREQ2 | FEAT_FANCTL_ONOFF | FEAT_NOCONF,
 		.peci_mask = 0x07,
 	},
 	[it8792] = {
@@ -461,7 +461,7 @@ static const struct it87_devices it87_devices[] = {
 		.features = FEAT_NEWER_AUTOPWM | FEAT_16BIT_FANS
 		  | FEAT_TEMP_OFFSET | FEAT_TEMP_OLD_PECI | FEAT_TEMP_PECI
 		  | FEAT_10_9MV_ADC | FEAT_IN7_INTERNAL | FEAT_FANCTL_ONOFF
-		  | FEAT_CONF_NOEXIT,
+		  | FEAT_NOCONF,
 		.peci_mask = 0x07,
 		.old_peci_mask = 0x02,	/* Actually reports PCH */
 	},
@@ -507,7 +507,7 @@ static const struct it87_devices it87_devices[] = {
 		.features = FEAT_NEWER_AUTOPWM | FEAT_16BIT_FANS
 		  | FEAT_TEMP_OFFSET | FEAT_TEMP_OLD_PECI | FEAT_TEMP_PECI
 		  | FEAT_10_9MV_ADC | FEAT_IN7_INTERNAL | FEAT_FANCTL_ONOFF
-		  | FEAT_CONF_NOEXIT,
+		  | FEAT_NOCONF,
 		.peci_mask = 0x07,
 		.old_peci_mask = 0x02,	/* Actually reports PCH */
 	},
@@ -544,7 +544,7 @@ static const struct it87_devices it87_devices[] = {
 #define has_four_temp(data)	((data)->features & FEAT_FOUR_TEMP)
 #define has_six_temp(data)	((data)->features & FEAT_SIX_TEMP)
 #define has_vin3_5v(data)	((data)->features & FEAT_VIN3_5V)
-#define has_conf_noexit(data)	((data)->features & FEAT_CONF_NOEXIT)
+#define has_noconf(data)	((data)->features & FEAT_NOCONF)
 #define has_scaling(data)	((data)->features & (FEAT_12MV_ADC | \
 						     FEAT_10_9MV_ADC))
 #define has_fanctl_onoff(data)	((data)->features & FEAT_FANCTL_ONOFF)
@@ -748,7 +748,7 @@ static int smbus_disable(struct it87_data *data)
 		superio_select(data->sioaddr, PME);
 		superio_outb(data->sioaddr, IT87_SPECIAL_CFG_REG,
 			     data->ec_special_config & ~data->smbus_bitmap);
-		superio_exit(data->sioaddr, has_conf_noexit(data));
+		superio_exit(data->sioaddr, has_noconf(data));
 	}
 	return 0;
 }
@@ -765,7 +765,7 @@ static int smbus_enable(struct it87_data *data)
 		superio_select(data->sioaddr, PME);
 		superio_outb(data->sioaddr, IT87_SPECIAL_CFG_REG,
 			     data->ec_special_config);
-		superio_exit(data->sioaddr, has_conf_noexit(data));
+		superio_exit(data->sioaddr, has_noconf(data));
 	}
 	return 0;
 }
@@ -3143,7 +3143,7 @@ static int __init it87_find(int sioaddr, unsigned short *address,
 	}
 
 exit:
-	superio_exit(sioaddr, config ? has_conf_noexit(config) : false);
+	superio_exit(sioaddr, config ? has_noconf(config) : false);
 	return err;
 }
 
@@ -3540,7 +3540,7 @@ static void it87_resume_sio(struct platform_device *pdev)
 			     reg2c);
 	}
 
-	superio_exit(data->sioaddr, has_conf_noexit(data));
+	superio_exit(data->sioaddr, has_noconf(data));
 }
 
 static int it87_resume(struct device *dev)
-- 
2.44.0

[PATCH v3 2/4] hwmon (it87): Do not enter configuration mode for some chiptypes

[Frank Crawford]

Update the configuration mode entry code to allow conditional entry, and
apply to all calls.

Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>
---

v2:
 * change macro name from has_conf_biosopen to has_noconf.

v3:
 * no change.

---
 drivers/hwmon/it87.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/hwmon/it87.c b/drivers/hwmon/it87.c
index 6eeba3a01e3c..396c2d3afbf7 100644
--- a/drivers/hwmon/it87.c
+++ b/drivers/hwmon/it87.c
@@ -117,7 +117,7 @@ static inline void superio_select(int ioreg, int ldn)
 	outb(ldn, ioreg + 1);
 }
 
-static inline int superio_enter(int ioreg)
+static inline int superio_enter(int ioreg, bool noentry)
 {
 	/*
 	 * Try to reserve ioreg and ioreg + 1 for exclusive access.
@@ -125,7 +125,8 @@ static inline int superio_enter(int ioreg)
 	if (!request_muxed_region(ioreg, 2, DRVNAME))
 		return -EBUSY;
 
-	__superio_enter(ioreg);
+	if (!noentry)
+		__superio_enter(ioreg);
 	return 0;
 }
 
@@ -742,7 +743,7 @@ static int smbus_disable(struct it87_data *data)
 	int err;
 
 	if (data->smbus_bitmap) {
-		err = superio_enter(data->sioaddr);
+		err = superio_enter(data->sioaddr, has_noconf(data));
 		if (err)
 			return err;
 		superio_select(data->sioaddr, PME);
@@ -758,7 +759,7 @@ static int smbus_enable(struct it87_data *data)
 	int err;
 
 	if (data->smbus_bitmap) {
-		err = superio_enter(data->sioaddr);
+		err = superio_enter(data->sioaddr, has_noconf(data));
 		if (err)
 			return err;
 
@@ -2674,7 +2675,7 @@ static int __init it87_find(int sioaddr, unsigned short *address,
 	u16 chip_type;
 	const struct it87_devices *config = NULL;
 
-	err = superio_enter(sioaddr);
+	err = superio_enter(sioaddr, false);
 	if (err)
 		return err;
 
@@ -3520,7 +3521,7 @@ static void it87_resume_sio(struct platform_device *pdev)
 	if (!data->need_in7_reroute)
 		return;
 
-	err = superio_enter(data->sioaddr);
+	err = superio_enter(data->sioaddr, has_noconf(data));
 	if (err) {
 		dev_warn(&pdev->dev,
 			 "Unable to enter Super I/O to reroute in7 (%d)",
-- 
2.44.0

[PATCH v3 3/4] hwmon (it87): Test for chipset before entering configuration mode

[Frank Crawford]

Major part of the change for the new method to avoid chipset issues.

The actual update does the following:

1) Lock the memory, but does not perform a SIO entry (previously it
would have performed an SIO entry).

2) Attempt to read the chipID.  This should be safe no matter which
chip we have.

3) If step (2) fails, then perform SIO entry and retry chipID read.  For
older chips and on failure it acts similarly to prior to this patch.

4) Set the sio_data->type, similar to previously.

5) If we have not performed an SIO entry, and this is not a chip type
with the NOCONF feature, then it will perform an SIO entry at this
point.

6) Proceed with setup as prior to this patch.

7) Any following access to the SIO registers will invoke the SIO entry
and SIO exit steps unless it is a chip with the NOCONF feature set.
This was set up in the previous patches in this patchset.

8) Update to the exit based on if it had performed a SIO entry or not.

Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>
---

v2:
 * rename macro from has_conf_biosopen to has_noconf.
 * improve patch documentation on new test of chipID.

v3:
 * rename variable from opened to enabled.
 * corrected final call to superio_exit().
 * minor update to patch documentation.

---
 drivers/hwmon/it87.c | 52 +++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 47 insertions(+), 5 deletions(-)

diff --git a/drivers/hwmon/it87.c b/drivers/hwmon/it87.c
index 396c2d3afbf7..6a77f2f6e1e1 100644
--- a/drivers/hwmon/it87.c
+++ b/drivers/hwmon/it87.c
@@ -2667,6 +2667,27 @@ static const struct attribute_group it87_group_auto_pwm = {
 	.is_visible = it87_auto_pwm_is_visible,
 };
 
+/*
+ * Original explanation:
+ * On various Gigabyte AM4 boards (AB350, AX370), the second Super-IO chip
+ * (IT8792E) needs to be in configuration mode before accessing the first
+ * due to a bug in IT8792E which otherwise results in LPC bus access errors.
+ * This needs to be done before accessing the first Super-IO chip since
+ * the second chip may have been accessed prior to loading this driver.
+ *
+ * The problem is also reported to affect IT8795E, which is used on X299 boards
+ * and has the same chip ID as IT8792E (0x8733). It also appears to affect
+ * systems with IT8790E, which is used on some Z97X-Gaming boards as well as
+ * Z87X-OC.
+ *
+ * From other information supplied:
+ * ChipIDs 0x8733, 0x8695 (early ID for IT87952E) and 0x8790 are intialised
+ * and left in configuration mode, and entering and/or exiting configuration
+ * mode is what causes the crash.
+ *
+ * The recommendation is to look up the chipID before doing any mode swap
+ * and then act accordingly.
+ */
 /* SuperIO detection - will change isa_address if a chip is found */
 static int __init it87_find(int sioaddr, unsigned short *address,
 			    struct it87_sio_data *sio_data, int chip_cnt)
@@ -2674,16 +2695,25 @@ static int __init it87_find(int sioaddr, unsigned short *address,
 	int err;
 	u16 chip_type;
 	const struct it87_devices *config = NULL;
+	bool enabled = false;
 
-	err = superio_enter(sioaddr, false);
+	/* First step, lock memory but don't enter configuration mode */
+	err = superio_enter(sioaddr, true);
 	if (err)
 		return err;
 
 	err = -ENODEV;
 	chip_type = superio_inw(sioaddr, DEVID);
-	/* check first for a valid chip before forcing chip id */
-	if (chip_type == 0xffff)
-		goto exit;
+	/* Check for a valid chip before forcing chip id */
+	if (chip_type == 0xffff) {
+		/* Enter configuration mode */
+		__superio_enter(sioaddr);
+		enabled = true;
+		/* and then try again */
+		chip_type = superio_inw(sioaddr, DEVID);
+		if (chip_type == 0xffff)
+			goto exit;
+	}
 
 	if (force_id_cnt == 1) {
 		/* If only one value given use for all chips */
@@ -2767,6 +2797,18 @@ static int __init it87_find(int sioaddr, unsigned short *address,
 
 	config = &it87_devices[sio_data->type];
 
+	/*
+	 * If previously we didn't enter configuration mode and it isn't a
+	 * chip we know is initialised in configuration mode, then enter
+	 * configuration mode.
+	 *
+	 * I don't know if any such chips can exist but be defensive.
+	 */
+	if (!enabled && !has_noconf(config)) {
+		__superio_enter(sioaddr);
+		enabled = true;
+	}
+
 	superio_select(sioaddr, PME);
 	if (!(superio_inb(sioaddr, IT87_ACT_REG) & 0x01)) {
 		pr_info("Device (chip %s ioreg 0x%x) not activated, skipping\n",
@@ -3144,7 +3186,7 @@ static int __init it87_find(int sioaddr, unsigned short *address,
 	}
 
 exit:
-	superio_exit(sioaddr, config ? has_noconf(config) : false);
+	superio_exit(sioaddr, !enabled);
 	return err;
 }
 
-- 
2.44.0

[PATCH v3 4/4] hwmon (it87): Remove tests nolonger required

[Frank Crawford]

Remove DMI tests for boards that are known to have issues with entering
configuration mode, as now we are testing the chips directly and no
longer need to rely on matching the board.

Leave the DMI table in place, for the nVIDIA board, and any future
expansions.

Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>
---

v2:
 * no change.

v3:
 * no change.

---
 drivers/hwmon/it87.c | 48 --------------------------------------------
 1 file changed, 48 deletions(-)

diff --git a/drivers/hwmon/it87.c b/drivers/hwmon/it87.c
index 6a77f2f6e1e1..b850eb3e5907 100644
--- a/drivers/hwmon/it87.c
+++ b/drivers/hwmon/it87.c
@@ -3683,27 +3683,6 @@ static int it87_dmi_cb(const struct dmi_system_id *dmi_entry)
 	return 1;
 }
 
-/*
- * On various Gigabyte AM4 boards (AB350, AX370), the second Super-IO chip
- * (IT8792E) needs to be in configuration mode before accessing the first
- * due to a bug in IT8792E which otherwise results in LPC bus access errors.
- * This needs to be done before accessing the first Super-IO chip since
- * the second chip may have been accessed prior to loading this driver.
- *
- * The problem is also reported to affect IT8795E, which is used on X299 boards
- * and has the same chip ID as IT8792E (0x8733). It also appears to affect
- * systems with IT8790E, which is used on some Z97X-Gaming boards as well as
- * Z87X-OC.
- * DMI entries for those systems will be added as they become available and
- * as the problem is confirmed to affect those boards.
- */
-static int it87_sio_force(const struct dmi_system_id *dmi_entry)
-{
-	__superio_enter(REG_4E);
-
-	return it87_dmi_cb(dmi_entry);
-};
-
 /*
  * On the Shuttle SN68PT, FAN_CTL2 is apparently not
  * connected to a fan, but to something else. One user
@@ -3726,34 +3705,7 @@ static struct it87_dmi_data nvidia_fn68pt = {
 		.driver_data = data, \
 	}
 
-#define IT87_DMI_MATCH_GBT(name, cb, data) \
-	IT87_DMI_MATCH_VND("Gigabyte Technology Co., Ltd.", name, cb, data)
-
 static const struct dmi_system_id it87_dmi_table[] __initconst = {
-	IT87_DMI_MATCH_GBT("AB350", it87_sio_force, NULL),
-		/* ? + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("AX370", it87_sio_force, NULL),
-		/* ? + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("Z97X-Gaming G1", it87_sio_force, NULL),
-		/* ? + IT8790E */
-	IT87_DMI_MATCH_GBT("TRX40 AORUS XTREME", it87_sio_force, NULL),
-		/* IT8688E + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("Z390 AORUS ULTRA-CF", it87_sio_force, NULL),
-		/* IT8688E + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("B550 AORUS PRO AC", it87_sio_force, NULL),
-		/* IT8688E + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("X570 AORUS MASTER", it87_sio_force, NULL),
-		/* IT8688E + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("X570 AORUS PRO", it87_sio_force, NULL),
-		/* IT8688E + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("X570 AORUS PRO WIFI", it87_sio_force, NULL),
-		/* IT8688E + IT8792E/IT8795E */
-	IT87_DMI_MATCH_GBT("X570S AERO G", it87_sio_force, NULL),
-		/* IT8689E + IT87952E */
-	IT87_DMI_MATCH_GBT("Z690 AORUS PRO DDR4", it87_sio_force, NULL),
-		/* IT8689E + IT87952E */
-	IT87_DMI_MATCH_GBT("Z690 AORUS PRO", it87_sio_force, NULL),
-		/* IT8689E + IT87952E */
 	IT87_DMI_MATCH_VND("nVIDIA", "FN68PT", it87_dmi_cb, &nvidia_fn68pt),
 	{ }
 
-- 
2.44.0

Re: [PATCH v3 1/4] hwmon (it87): Rename FEAT_CONF_NOEXIT to FEAT_NOCONF as more descriptive of requirement

[Guenter Roeck]

On Sun, Apr 28, 2024 at 04:06:33PM +1000, Frank Crawford wrote:
> Rename previous definitions to match the new information that they are
> preinitialised as enabled and should not receive codes to enter or exit
> configuration mode.
> 
> Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>

Applied.

Guenter

Re: [PATCH v3 2/4] hwmon (it87): Do not enter configuration mode for some chiptypes

[Guenter Roeck]

On Sun, Apr 28, 2024 at 04:06:34PM +1000, Frank Crawford wrote:
> Update the configuration mode entry code to allow conditional entry, and
> apply to all calls.
> 
> Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>

Applied.

Guenter

Re: [PATCH v3 3/4] hwmon (it87): Test for chipset before entering configuration mode

[Guenter Roeck]

On Sun, Apr 28, 2024 at 04:06:35PM +1000, Frank Crawford wrote:
> Major part of the change for the new method to avoid chipset issues.
> 
> The actual update does the following:
> 
> 1) Lock the memory, but does not perform a SIO entry (previously it
> would have performed an SIO entry).
> 
> 2) Attempt to read the chipID.  This should be safe no matter which
> chip we have.
> 
> 3) If step (2) fails, then perform SIO entry and retry chipID read.  For
> older chips and on failure it acts similarly to prior to this patch.
> 
> 4) Set the sio_data->type, similar to previously.
> 
> 5) If we have not performed an SIO entry, and this is not a chip type
> with the NOCONF feature, then it will perform an SIO entry at this
> point.
> 
> 6) Proceed with setup as prior to this patch.
> 
> 7) Any following access to the SIO registers will invoke the SIO entry
> and SIO exit steps unless it is a chip with the NOCONF feature set.
> This was set up in the previous patches in this patchset.
> 
> 8) Update to the exit based on if it had performed a SIO entry or not.
> 
> Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>

Ah, checkpatch is useful.

> + * ChipIDs 0x8733, 0x8695 (early ID for IT87952E) and 0x8790 are intialised
                                                                    ^^^^^^^^^^

No need to resend, but please run checkpatch on future patches.

Thanks,
Guenter

Re: [PATCH v3 4/4] hwmon (it87): Remove tests nolonger required

[Guenter Roeck]

On Sun, Apr 28, 2024 at 04:06:36PM +1000, Frank Crawford wrote:
> Remove DMI tests for boards that are known to have issues with entering
> configuration mode, as now we are testing the chips directly and no
> longer need to rely on matching the board.
> 
> Leave the DMI table in place, for the nVIDIA board, and any future
> expansions.
> 
> Signed-off-by: Frank Crawford <frank@crawford.emu.id.au>

Applied.

Thanks,
Guenter