Skip to contentSkip to conversation

regmap: Synchronize cache for the page selector

PATCH In review Andy Shevchenko 4 participants 22 messages Jan 16, 2025 – Feb 3, 2025

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 regmap: Synchronize cache for the page selector

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

    // Real registers
    18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
    ...
    // Virtual (per port)
    40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
    50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
    60: 01 ff 00 00 ff ff 00 00 00 00 00 00
    70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
    80: 03 ff 00 00 00 00 00 00 00 00 00 ff
    90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

    // Real registers
    18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
    ...
    // Virtual (per port)
    40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
    50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
    60: 02 ff 00 00 ff ff 00 00 00 00 00 00
    70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
    80: 04 ff 00 00 00 00 00 00 00 00 00 ff
    90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

---
v3: rebased on the top of latest vanilla

 drivers/base/regmap/regmap.c | 28 ++++++++++++++++++++++------
 1 file changed, 22 insertions(+), 6 deletions(-)
drivers/base/regmap/regmap.c +22 -6
index f2843f814675..197c79b66828 100644
@@ -1557,24 +1557,40 @@ static int _regmap_select_page(struct regmap *map, unsigned int *reg,
 			return -EINVAL;
 	}
 
-	/* It is possible to have selector register inside data window.
-	   In that case, selector register is located on every page and
-	   it needs no page switching, when accessed alone. */
+	/*
+	 * It is possible to have selector register inside data window.
+	 * In that case, selector register is located on every page and it
+	 * needs no page switching, when accessed alone.
+	 *
+	 * Nevertheless we should synchronize the cache values for it.
+	 */
 	if (val_num > 1 ||
 	    range->window_start + win_offset != range->selector_reg) {
+		unsigned int page_off = win_page * range->window_len;
+		unsigned int sel_offset = range->selector_reg - range->window_start;
+		unsigned int sel_register = range->range_min + page_off + sel_offset;
+		unsigned int val = win_page << range->selector_shift;
+		unsigned int mask = range->selector_mask;
+
 		/* Use separate work_buf during page switching */
 		orig_work_buf = map->work_buf;
 		map->work_buf = map->selector_work_buf;
 
-		ret = _regmap_update_bits(map, range->selector_reg,
-					  range->selector_mask,
-					  win_page << range->selector_shift,
+		ret = _regmap_update_bits(map, range->selector_reg, mask, val,
 					  &page_chg, false);
 
 		map->work_buf = orig_work_buf;
 
 		if (ret != 0)
 			return ret;
+
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);
 	}
 
 	*reg = range->window_start + win_offset;
Signature · 1 line
2.43.0.rc1.1336.g36b5255a03ac
maintainer
#1 lore raw
Quoted 8 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

[...]

Applied to

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap.git for-next

Thanks!

[1/1] regmap: Synchronize cache for the page selector

      commit: 1fd60ed1700cf25d7ee233580c08fda755f9a61b

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

#2 lore raw
In reply to Andy Shevchenko

Dear All,

Quoted 31 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

     // Real registers
     18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
     ...
     // Virtual (per port)
     40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
     50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
     60: 01 ff 00 00 ff ff 00 00 00 00 00 00
     70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
     80: 03 ff 00 00 00 00 00 00 00 00 00 ff
     90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

     // Real registers
     18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
     ...
     // Virtual (per port)
     40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
     50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
     60: 02 ff 00 00 ff ff 00 00 00 00 00 00
     70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
     80: 04 ff 00 00 00 00 00 00 00 00 00 ff
     90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Quoted 2 lines
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver. With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for 
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with 
error -2

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Quoted 46 lines

---
v3: rebased on the top of latest vanilla

  drivers/base/regmap/regmap.c | 28 ++++++++++++++++++++++------
  1 file changed, 22 insertions(+), 6 deletions(-)
drivers/base/regmap/regmap.c +15 -6
index f2843f814675..197c79b66828 100644
@@ -1557,24 +1557,40 @@ static int _regmap_select_page(struct regmap *map, unsigned int *reg,
  			return -EINVAL;
  	}
 
-	/* It is possible to have selector register inside data window.
-	   In that case, selector register is located on every page and
-	   it needs no page switching, when accessed alone. */
+	/*
+	 * It is possible to have selector register inside data window.
+	 * In that case, selector register is located on every page and it
+	 * needs no page switching, when accessed alone.
+	 *
+	 * Nevertheless we should synchronize the cache values for it.
+	 */
  	if (val_num > 1 ||
  	    range->window_start + win_offset != range->selector_reg) {
+		unsigned int page_off = win_page * range->window_len;
+		unsigned int sel_offset = range->selector_reg - range->window_start;
+		unsigned int sel_register = range->range_min + page_off + sel_offset;
+		unsigned int val = win_page << range->selector_shift;
+		unsigned int mask = range->selector_mask;
+
  		/* Use separate work_buf during page switching */
  		orig_work_buf = map->work_buf;
  		map->work_buf = map->selector_work_buf;
 
-		ret = _regmap_update_bits(map, range->selector_reg,
-					  range->selector_mask,
-					  win_page << range->selector_shift,
+		ret = _regmap_update_bits(map, range->selector_reg, mask, val,
  					  &page_chg, false);
 
  		map->work_buf = orig_work_buf;
 
  		if (ret != 0)
  			return ret;
+

drivers/base/regmap/regmap.c static int _regmap_select_page(struct regmap *map, unsigned int *reg, · line 1596

Quoted 10 lines
drivers/base/regmap/regmap.c +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);
  	}
 
  	*reg = range->window_start + win_offset;

Best regards

Signature · 2 lines
Marek Szyprowski, PhD
Samsung R&D Institute Poland
#3 lore raw
Quoted 32 lines · Marek Szyprowski
Quoted 31 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

     // Real registers
     18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
     ...
     // Virtual (per port)
     40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
     50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
     60: 01 ff 00 00 ff ff 00 00 00 00 00 00
     70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
     80: 03 ff 00 00 00 00 00 00 00 00 00 ff
     90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

     // Real registers
     18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
     ...
     // Virtual (per port)
     40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
     50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
     60: 02 ff 00 00 ff ff 00 00 00 00 00 00
     70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
     80: 04 ff 00 00 00 00 00 00 00 00 00 ff
     90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Quoted 6 lines
Quoted 2 lines
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver.

Is there any datasheet link to the HW in question?

(FWIW, I have tested this with the CY8C9540 GPIO I²C expander on Intel Galileo
Gen 1 board.)

Quoted 11 lines

With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for 
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with 
error -2
Quoted 7 lines

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Yes, see below. And thank you for your report!

...

Quoted 7 lines

drivers/base/regmap/regmap.c static int _regmap_select_page(struct regmap *map, unsigned int *reg, · line 1593

Quoted 7 lines
drivers/base/regmap/regmap.c +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);

Can you add a test printk() here to show

page_chg
range->selector_reg, range->window_start, range->window_len sel_register, mask, val

?

And would commenting these three lines make it work again?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#4 lore raw
Quoted 78 lines · Andy Shevchenko
Quoted 32 lines · Marek Szyprowski
Quoted 31 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

     // Real registers
     18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
     ...
     // Virtual (per port)
     40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
     50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
     60: 01 ff 00 00 ff ff 00 00 00 00 00 00
     70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
     80: 03 ff 00 00 00 00 00 00 00 00 00 ff
     90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

     // Real registers
     18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
     ...
     // Virtual (per port)
     40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
     50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
     60: 02 ff 00 00 ff ff 00 00 00 00 00 00
     70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
     80: 04 ff 00 00 00 00 00 00 00 00 00 ff
     90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Quoted 6 lines
Quoted 2 lines
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver.

Is there any datasheet link to the HW in question?

(FWIW, I have tested this with the CY8C9540 GPIO I²C expander on Intel Galileo
Gen 1 board.)

Quoted 11 lines

With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for 
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with 
error -2
Quoted 7 lines

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Yes, see below. And thank you for your report!

...

Quoted 7 lines

Review

Quoted 7 lines
(quoted diff) +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);

Can you add a test printk() here to show

Quoted 7 lines

page_chg
range->selector_reg, range->window_start, range->window_len sel_register, mask, val

?

And would commenting these three lines make it work again?

Also try to apply this patch (while having the above lines not commented):

From 0fe5fe51d8b86305a4ca1ae44ede34a24fe2f9d7 Mon Sep 17 00:00:00 2001
From: Andy Shevchenko <[email protected]>
Date: Fri, 17 Jan 2025 16:29:19 +0200
Subject: [PATCH 1/1] TBD:

---
 drivers/gpu/drm/bridge/lontium-lt9611uxc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/bridge/lontium-lt9611uxc.c +1 -1
index f4c3ff1fdc69..35a1dd568bbb 100644
@@ -69,7 +69,7 @@ struct lt9611uxc {
 static const struct regmap_range_cfg lt9611uxc_ranges[] = {
 	{
 		.name = "register_range",
-		.range_min =  0,
+		.range_min = 0x0100,
 		.range_max = 0xd0ff,
 		.selector_reg = LT9611_PAGE_CONTROL,
 		.selector_mask = 0xff,
Signature · 5 lines
2.43.0.rc1.1336.g36b5255a03ac

-- 
With Best Regards,
Andy Shevchenko
maintainer
#5 lore raw

drivers/gpu/drm/bridge/lontium-lt9611uxc.c struct lt9611uxc { · line 75

Quoted 11 lines · Andy Shevchenko

On Fri, Jan 17, 2025 at 04:09:58PM +0200, Andy Shevchenko wrote:

drivers/gpu/drm/bridge/lontium-lt9611uxc.c +1 -1
@@ -69,7 +69,7 @@ struct lt9611uxc {
 static const struct regmap_range_cfg lt9611uxc_ranges[] = {
 	{
 		.name = "register_range",
-		.range_min =  0,
+		.range_min = 0x0100,
 		.range_max = 0xd0ff,
 		.selector_reg = LT9611_PAGE_CONTROL,
 		.selector_mask = 0xff,

There's a bunch of other (generally broken) drivers with their ranges overlapping their windows, we probably have to drop this for now.

#21 lore raw
In reply to Andy Shevchenko
Quoted 35 lines · Andy Shevchenko
Quoted 34 lines · Marek Szyprowski
Quoted 33 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

      // Real registers
      18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
      ...
      // Virtual (per port)
      40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
      50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
      60: 01 ff 00 00 ff ff 00 00 00 00 00 00
      70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
      80: 03 ff 00 00 00 00 00 00 00 00 00 ff
      90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

      // Real registers
      18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
      ...
      // Virtual (per port)
      40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
      50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
      60: 02 ff 00 00 ff ff 00 00 00 00 00 00
      70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
      80: 04 ff 00 00 00 00 00 00 00 00 00 ff
      90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")
Quoted 7 lines
Quoted 3 lines

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver.

Is there any datasheet link to the HW in question?

(FWIW, I have tested this with the CY8C9540 GPIO I²C expander on Intel Galileo
Gen 1 board.)

I'm not aware of it, but I'm not related to lt9611uxc driver development at all...

Quoted 31 lines
Quoted 11 lines

With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with
error -2
Quoted 7 lines

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Yes, see below. And thank you for your report!

...

Quoted 7 lines

drivers/base/regmap/regmap.c static int _regmap_select_page(struct regmap *map, unsigned int *reg, · line 1593

Quoted 7 lines
drivers/base/regmap/regmap.c +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);

Can you add a test printk() here to show

Quoted 7 lines

page_chg
range->selector_reg, range->window_start, range->window_len sel_register, mask, val

?

And would commenting these three lines make it work again?

I've added the following debug message before that check:

printk("page_chg %d range->selector_reg %x range->window_start %x range->window_len %x sel_register %x mask %x val %x\n", page_chg, range->selector_reg, range->window_start, range->window_len, sel_register, mask, val);

Here is the result:

root@target:~# modprobe lontium_lt9611uxc
[   29.892962] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   29.907419] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 81ff mask ff val 81
[   29.926712] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 81ff mask ff val 81
[   29.958301] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 81ff mask ff val 81
[   29.979951] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   29.987002] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   30.052370] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   30.066808] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register b0ff mask ff val b0
[   30.079859] lt9611uxc 5-002b: LT9611 version: 0x00
[   30.085990] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   30.151735] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   30.158830] lt9611uxc 5-002b: Direct firmware load for 
lt9611uxc_fw.bin failed with error -2
[   30.168653] lt9611uxc 5-002b: probe with driver lt9611uxc failed with 
error -2

after disabling the above mentioned 3 lines:

root@target:~# modprobe lontium_lt9611uxc
[   44.584893] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   44.597589] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 81ff mask ff val 81
[   44.609936] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 81ff mask ff val 81
[   44.622277] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 81ff mask ff val 81
[   44.634579] lt9611uxc 5-002b: LT9611 revision: 0x17.04.93
[   44.641444] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   44.710694] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   44.724579] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register b0ff mask ff val b0
[   44.736853] lt9611uxc 5-002b: LT9611 version: 0x40
[   44.743253] page_chg 1 range->selector_reg ff range->window_start 0 
range->window_len 100 sel_register 80ff mask ff val 80
[   44.816433] msm_dpu ae01000.display-controller: bound ae94000.dsi 
(ops dsi_ops [msm])
[   44.843944] msm_dpu ae01000.display-controller: bound 
ae90000.displayport-controller (ops msm_dp_display_comp_ops [msm])
[   44.862886] adreno 3d00000.gpu: supply vdd not found, using dummy 
regulator
[   44.870974] adreno 3d00000.gpu: supply vddcx not found, using dummy 
regulator
[   44.897116] platform 3d6a000.gmu: Adding to iommu group 27
[   44.938871] msm_dpu ae01000.display-controller: bound 3d00000.gpu 
(ops a3xx_ops [msm])
[   44.952435] [drm:dpu_kms_hw_init:1156] dpu hardware revision:0x60000000
[   44.976496] [drm] Initialized msm 1.12.0 for 
ae01000.display-controller on minor 0
...

Best regards

Signature · 2 lines
Marek Szyprowski, PhD
Samsung R&D Institute Poland
#6 lore raw
In reply to Andy Shevchenko
Quoted 75 lines · Andy Shevchenko
Quoted 74 lines · Andy Shevchenko
Quoted 32 lines · Marek Szyprowski
Quoted 31 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

      // Real registers
      18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
      ...
      // Virtual (per port)
      40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
      50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
      60: 01 ff 00 00 ff ff 00 00 00 00 00 00
      70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
      80: 03 ff 00 00 00 00 00 00 00 00 00 ff
      90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

      // Real registers
      18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
      ...
      // Virtual (per port)
      40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
      50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
      60: 02 ff 00 00 ff ff 00 00 00 00 00 00
      70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
      80: 04 ff 00 00 00 00 00 00 00 00 00 ff
      90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Quoted 5 lines
Quoted 2 lines
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver.

Is there any datasheet link to the HW in question?

(FWIW, I have tested this with the CY8C9540 GPIO I²C expander on Intel Galileo
Gen 1 board.)

Quoted 11 lines

With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with
error -2
Quoted 7 lines

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Yes, see below. And thank you for your report!

...

Quoted 7 lines

Review

Quoted 7 lines
(quoted diff) +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);

Can you add a test printk() here to show

Quoted 8 lines
Quoted 7 lines

page_chg
range->selector_reg, range->window_start, range->window_len sel_register, mask, val

?

And would commenting these three lines make it work again?

Also try to apply this patch (while having the above lines not commented):

This patch applied alone doesn't change anything. Probe still fails.

However, with the mentioned 3 lines in the regmap code commented AND this patch applied, lt9611uxc driver probe also fails. Does it mean that there is really a bug in the driver?

Quoted 15 lines

From 0fe5fe51d8b86305a4ca1ae44ede34a24fe2f9d7 Mon Sep 17 00:00:00 2001
From: Andy Shevchenko <[email protected]>
Date: Fri, 17 Jan 2025 16:29:19 +0200
Subject: [PATCH 1/1] TBD:

---
  drivers/gpu/drm/bridge/lontium-lt9611uxc.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/bridge/lontium-lt9611uxc.c +0 -0
index f4c3ff1fdc69..35a1dd568bbb 100644

drivers/gpu/drm/bridge/lontium-lt9611uxc.c struct lt9611uxc { · line 75

Quoted 8 lines
drivers/gpu/drm/bridge/lontium-lt9611uxc.c +1 -1
@@ -69,7 +69,7 @@ struct lt9611uxc {
  static const struct regmap_range_cfg lt9611uxc_ranges[] = {
  	{
  		.name = "register_range",
-		.range_min =  0,
+		.range_min = 0x0100,
  		.range_max = 0xd0ff,
  		.selector_reg = LT9611_PAGE_CONTROL,
  		.selector_mask = 0xff,

Best regards

Signature · 2 lines
Marek Szyprowski, PhD
Samsung R&D Institute Poland
#7 lore raw
Quoted 64 lines · Marek Szyprowski
Quoted 63 lines · Andy Shevchenko
Quoted 62 lines · Andy Shevchenko
Quoted 32 lines · Marek Szyprowski
Quoted 31 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

      // Real registers
      18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
      ...
      // Virtual (per port)
      40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
      50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
      60: 01 ff 00 00 ff ff 00 00 00 00 00 00
      70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
      80: 03 ff 00 00 00 00 00 00 00 00 00 ff
      90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

      // Real registers
      18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
      ...
      // Virtual (per port)
      40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
      50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
      60: 02 ff 00 00 ff ff 00 00 00 00 00 00
      70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
      80: 04 ff 00 00 00 00 00 00 00 00 00 ff
      90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Quoted 5 lines
Quoted 2 lines
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver.

Is there any datasheet link to the HW in question?

(FWIW, I have tested this with the CY8C9540 GPIO I²C expander on Intel Galileo
Gen 1 board.)

Quoted 11 lines

With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with
error -2
Quoted 7 lines

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Yes, see below. And thank you for your report!

...

Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 7 lines

Review

Quoted 7 lines
(quoted diff) +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);

Can you add a test printk() here to show

page_chg
range->selector_reg, range->window_start, range->window_len sel_register, mask, val

?

Quoted 7 lines
Quoted 2 lines
Quoted 1 line

And would commenting these three lines make it work again?

Also try to apply this patch (while having the above lines not commented):

This patch applied alone doesn't change anything. Probe still fails.

However, with the mentioned 3 lines in the regmap code commented AND this patch applied, lt9611uxc driver probe also fails.

Does it fail in the same way?

Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

Quoted 11 lines
Quoted 11 lines
drivers/gpu/drm/bridge/lontium-lt9611uxc.c +1 -1
@@ -69,7 +69,7 @@ struct lt9611uxc {
  static const struct regmap_range_cfg lt9611uxc_ranges[] = {
  	{
  		.name = "register_range",
-		.range_min =  0,
+		.range_min = 0x0100,
  		.range_max = 0xd0ff,
  		.selector_reg = LT9611_PAGE_CONTROL,
  		.selector_mask = 0xff,
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#8 lore raw
Quoted 82 lines · Andy Shevchenko
Quoted 64 lines · Marek Szyprowski
Quoted 63 lines · Andy Shevchenko
Quoted 62 lines · Andy Shevchenko
Quoted 32 lines · Marek Szyprowski
Quoted 31 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

       // Real registers
       18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
       ...
       // Virtual (per port)
       40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
       50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
       60: 01 ff 00 00 ff ff 00 00 00 00 00 00
       70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
       80: 03 ff 00 00 00 00 00 00 00 00 00 ff
       90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

       // Real registers
       18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
       ...
       // Virtual (per port)
       40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
       50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
       60: 02 ff 00 00 ff ff 00 00 00 00 00 00
       70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
       80: 04 ff 00 00 00 00 00 00 00 00 00 ff
       90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Quoted 5 lines
Quoted 2 lines
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver.

Is there any datasheet link to the HW in question?

(FWIW, I have tested this with the CY8C9540 GPIO I²C expander on Intel Galileo
Gen 1 board.)

Quoted 11 lines

With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with
error -2
Quoted 7 lines

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Yes, see below. And thank you for your report!

...

Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 7 lines

Review

Quoted 7 lines
(quoted diff) +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);

Can you add a test printk() here to show

page_chg
range->selector_reg, range->window_start, range->window_len sel_register, mask, val

?

Quoted 7 lines
Quoted 6 lines
Quoted 2 lines
Quoted 1 line

And would commenting these three lines make it work again?

Also try to apply this patch (while having the above lines not commented):

This patch applied alone doesn't change anything. Probe still fails.

However, with the mentioned 3 lines in the regmap code commented AND this patch applied, lt9611uxc driver probe also fails.

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Quoted 9 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

Quoted 11 lines
Quoted 11 lines

drivers/gpu/drm/bridge/lontium-lt9611uxc.c struct lt9611uxc { · line 75

Quoted 11 lines
drivers/gpu/drm/bridge/lontium-lt9611uxc.c +1 -1
@@ -69,7 +69,7 @@ struct lt9611uxc {
   static const struct regmap_range_cfg lt9611uxc_ranges[] = {
   	{
   		.name = "register_range",
-		.range_min =  0,
+		.range_min = 0x0100,
   		.range_max = 0xd0ff,
   		.selector_reg = LT9611_PAGE_CONTROL,
   		.selector_mask = 0xff,

Best regards

Signature · 2 lines
Marek Szyprowski, PhD
Samsung R&D Institute Poland
#9 lore raw
In reply to Andy Shevchenko
Quoted 82 lines · Andy Shevchenko
Quoted 64 lines · Marek Szyprowski
Quoted 63 lines · Andy Shevchenko
Quoted 62 lines · Andy Shevchenko
Quoted 32 lines · Marek Szyprowski
Quoted 31 lines · Andy Shevchenko

If the selector register is represented in each page, its value in accordance to the debugfs is stale because it gets synchronized only after the real page switch happens. Synchronize cache for the page selector.

Before (offset followed by hexdump, the first byte is selector):

       // Real registers
       18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
       ...
       // Virtual (per port)
       40: 05 ff 00 00 e0 e0 00 00 00 00 00 1f
       50: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
       60: 01 ff 00 00 ff ff 00 00 00 00 00 00
       70: 02 ff 00 00 cf f3 00 00 00 00 00 0c
       80: 03 ff 00 00 00 00 00 00 00 00 00 ff
       90: 04 ff 00 00 ff 0f 00 00 f0 00 00 00

After:

       // Real registers
       18: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
       ...
       // Virtual (per port)
       40: 00 ff 00 00 e0 e0 00 00 00 00 00 1f
       50: 01 ff 00 00 e0 e0 00 00 00 00 00 1f
       60: 02 ff 00 00 ff ff 00 00 00 00 00 00
       70: 03 ff 00 00 cf f3 00 00 00 00 00 0c
       80: 04 ff 00 00 00 00 00 00 00 00 00 ff
       90: 05 ff 00 00 ff 0f 00 00 f0 00 00 00
Quoted 5 lines
Quoted 2 lines
Fixes 6863ca622759 ("regmap: Add support for register indirect addressing.")

This patch landed in linux-next some time ago as commit 1fd60ed1700c ("regmap: Synchronize cache for the page selector"). Today I've noticed that it causes a regression for Lontium LT9611UXC HDMI bridge driver.

Is there any datasheet link to the HW in question?

(FWIW, I have tested this with the CY8C9540 GPIO I²C expander on Intel Galileo
Gen 1 board.)

Quoted 11 lines

With today's linux-next I got the following messages on QCom RB5 board:

# dmesg | grep  lt9611uxc
[   13.737346] lt9611uxc 5-002b: LT9611 revision: 0x00.00.00
[   13.804190] lt9611uxc 5-002b: LT9611 version: 0x00
[   13.870564] lt9611uxc 5-002b: FW version 0, enforcing firmware update
[   13.877437] lt9611uxc 5-002b: Direct firmware load for
lt9611uxc_fw.bin failed with error -2
[   13.887517] lt9611uxc 5-002b: probe with driver lt9611uxc failed with
error -2
Quoted 7 lines

after reverting the $subject patch, the driver probes fine on that board.

I'm not sure if this is really a bug caused by this change or simply the driver already was aligned to old regmap behavior. Dmitry, could you check the regamp usage and review the changes introduced by this patch? Let me know if there is anything to check on the real hardware to help resolving this issue.

Yes, see below. And thank you for your report!

...

Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 7 lines

Review

Quoted 7 lines
(quoted diff) +7 -0
+		/*
+		 * If selector register has been just updated, update the respective
+		 * virtual copy as well.
+		 */
+		if (page_chg &&
+		    in_range(range->selector_reg, range->window_start, range->window_len))
+			_regmap_update_bits(map, sel_register, mask, val, NULL, false);

Can you add a test printk() here to show

page_chg
range->selector_reg, range->window_start, range->window_len sel_register, mask, val

?

Quoted 7 lines
Quoted 6 lines
Quoted 2 lines
Quoted 1 line

And would commenting these three lines make it work again?

Also try to apply this patch (while having the above lines not commented):

This patch applied alone doesn't change anything. Probe still fails.

However, with the mentioned 3 lines in the regmap code commented AND this patch applied, lt9611uxc driver probe also fails.

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Quoted 9 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

Quoted 11 lines
Quoted 11 lines

drivers/gpu/drm/bridge/lontium-lt9611uxc.c struct lt9611uxc { · line 75

Quoted 11 lines
drivers/gpu/drm/bridge/lontium-lt9611uxc.c +1 -1
@@ -69,7 +69,7 @@ struct lt9611uxc {
   static const struct regmap_range_cfg lt9611uxc_ranges[] = {
   	{
   		.name = "register_range",
-		.range_min =  0,
+		.range_min = 0x0100,
   		.range_max = 0xd0ff,
   		.selector_reg = LT9611_PAGE_CONTROL,
   		.selector_mask = 0xff,

Best regards

Signature · 2 lines
Marek Szyprowski, PhD
Samsung R&D Institute Poland
#10 lore raw
Quoted 7 lines · Marek Szyprowski
Quoted 3 lines · Andy Shevchenko

On Fri, Jan 17, 2025 at 05:05:42PM +0100, Marek Szyprowski wrote:

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Hmm... This is very interesting! It means that the page selector is a bit magical there. Dmitry, can you chime in and perhaps shed some light on this?

Quoted 7 lines
Quoted 7 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

Quoted 5 lines
Quoted 2 lines

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

It would be helpful! Traces from the failed and non-failed cases till the firmware revision and chip ID reading would be enough to start with.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#11 lore raw

Hi Andy,

Quoted 17 lines · Andy Shevchenko
Quoted 6 lines · Marek Szyprowski
Quoted 3 lines · Andy Shevchenko

On Fri, Jan 17, 2025 at 05:05:42PM +0100, Marek Szyprowski wrote:

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Hmm... This is very interesting! It means that the page selector is a bit magical there. Dmitry, can you chime in and perhaps shed some light on this?

Quoted 7 lines
Quoted 7 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

It would be helpful! Traces from the failed and non-failed cases till the firmware revision and chip ID reading would be enough to start with.

I'm sorry for the delay, I was a bit busy with other stuff.

Here are logs (all values are in hex):

next-20250128 (probe broken):

root@target:~# dmesg | grep regmap
[   14.817604] regmap_write reg 80ee <- 1
[   14.823036] regmap_read reg 8100 -> 0
[   14.827631] regmap_read reg 8101 -> 0
[   14.832130] regmap_read reg 8102 -> 0
[   14.841477] regmap_write reg 80ee <- 0
[   14.901796] regmap_write reg 80ee <- 1
[   14.909895] regmap_read reg b021 -> 0
[   14.927409] regmap_write reg 80ee <- 0

next-20250128 + 1fd60ed1700c reverted (probe okay):

root@target:~# dmesg | grep regmap
[   13.565920] regmap_write reg 80ee <- 1
[   13.567509] regmap_read reg 8100 -> 17
[   13.568219] regmap_read reg 8101 -> 4
[   13.568909] regmap_read reg 8102 -> 93
[   13.568927] regmap_write reg 80ee <- 0
[   13.625844] regmap_write reg 80ee <- 1
[   13.630963] regmap_read reg b021 -> 40
[   13.645522] regmap_write reg 80ee <- 0
[   13.880192] regmap_write reg 80ee <- 1
[   13.884921] regmap_read reg b023 -> 0
[   13.892159] regmap_write reg 80ee <- 0
[   13.954508] regmap_write reg 80ee <- 1
[   13.959441] regmap_read reg b023 -> 0
[   13.963269] regmap_write reg 80ee <- 0
[   14.029818] regmap_write reg 80ee <- 1
[   14.034644] regmap_read reg b023 -> 0
[   14.038491] regmap_write reg 80ee <- 0
[   16.278387] regmap_write reg 80ee <- 1
[   16.283902] regmap_read reg b023 -> 0
[   16.287944] regmap_write reg 80ee <- 0
[   16.347570] regmap_write reg 80ee <- 1
[   16.355894] regmap_read reg b023 -> 0
[   16.359748] regmap_write reg 80ee <- 0
[   16.452472] regmap_write reg 80ee <- 1
[   16.454387] regmap_write reg d00d <- 4
[   16.456228] regmap_write reg d00e <- 65
[   16.462096] regmap_write reg d00f <- 4
[   16.462971] regmap_write reg d010 <- 38
[   16.463741] regmap_write reg d011 <- 8
[   16.464741] regmap_write reg d012 <- 98
[   16.465250] regmap_write reg d013 <- 7
[   16.465764] regmap_write reg d014 <- 80
[   16.466481] regmap_write reg d015 <- 5
[   16.467156] regmap_update_bits reg d016 <- 0/f
[   16.467817] regmap_write reg d017 <- 2c
[   16.468325] regmap_update_bits reg d018 <- 0/f
[   16.469101] regmap_write reg d019 <- 4
[   16.469749] regmap_update_bits reg d01a <- 0/f
[   16.470509] regmap_write reg d01b <- 58
[   16.471041] regmap_write reg 80ee <- 0

Best regards

Signature · 2 lines
Marek Szyprowski, PhD
Samsung R&D Institute Poland
#12 lore raw
Quoted 20 lines · Marek Szyprowski
Quoted 19 lines · Andy Shevchenko
Quoted 6 lines · Marek Szyprowski
Quoted 3 lines · Andy Shevchenko

On Fri, Jan 17, 2025 at 05:05:42PM +0100, Marek Szyprowski wrote:

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Hmm... This is very interesting! It means that the page selector is a bit magical there. Dmitry, can you chime in and perhaps shed some light on this?

Quoted 9 lines
Quoted 9 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

It would be helpful! Traces from the failed and non-failed cases till the firmware revision and chip ID reading would be enough to start with.

I'm sorry for the delay, I was a bit busy with other stuff.

No problem and thanks for sharing.

Quoted 11 lines

Here are logs (all values are in hex):

next-20250128 (probe broken):

root@target:~# dmesg | grep regmap
[   14.817604] regmap_write reg 80ee <- 1
[   14.823036] regmap_read reg 8100 -> 0
[   14.827631] regmap_read reg 8101 -> 0
[   14.832130] regmap_read reg 8102 -> 0
Quoted 6 lines

next-20250128 + 1fd60ed1700c reverted (probe okay):

root@target:~# dmesg | grep regmap
[   13.565920] regmap_write reg 80ee <- 1
[   13.567509] regmap_read reg 8100 -> 17
[   13.568219] regmap_read reg 8101 -> 4
[   13.568909] regmap_read reg 8102 -> 93

Something is missing here. Like we have an identical start and an immediate failure. If you did it via printk() approach, it's probably wrong as my patch uses internal regmap function. Most likely you need to enable trace events for regmap and collect those for let's say 2 seconds:

	echo 1 > ...trace events...
	modprobe ...
	sleep 2
	echo 0 > ...trace events...

and dump the buffer to a file. It might have though more than needed as some other devices might also use regmap at the same time. I don't remember if the trace events for regmap have a device instance name field which can be used as a filter.

Alternatively you may also try to add a printk() into regmap core, but I don't think it's more practical than trace events.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#13 lore raw
Quoted 59 lines · Andy Shevchenko
Quoted 20 lines · Marek Szyprowski
Quoted 19 lines · Andy Shevchenko
Quoted 6 lines · Marek Szyprowski
Quoted 3 lines · Andy Shevchenko

On Fri, Jan 17, 2025 at 05:05:42PM +0100, Marek Szyprowski wrote:

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Hmm... This is very interesting! It means that the page selector is a bit magical there. Dmitry, can you chime in and perhaps shed some light on this?

Quoted 9 lines
Quoted 9 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

It would be helpful! Traces from the failed and non-failed cases till the firmware revision and chip ID reading would be enough to start with.

I'm sorry for the delay, I was a bit busy with other stuff.

No problem and thanks for sharing.

Quoted 11 lines

Here are logs (all values are in hex):

next-20250128 (probe broken):

root@target:~# dmesg | grep regmap
[   14.817604] regmap_write reg 80ee <- 1
[   14.823036] regmap_read reg 8100 -> 0
[   14.827631] regmap_read reg 8101 -> 0
[   14.832130] regmap_read reg 8102 -> 0
Quoted 6 lines

next-20250128 + 1fd60ed1700c reverted (probe okay):

root@target:~# dmesg | grep regmap
[   13.565920] regmap_write reg 80ee <- 1
[   13.567509] regmap_read reg 8100 -> 17
[   13.568219] regmap_read reg 8101 -> 4
[   13.568909] regmap_read reg 8102 -> 93

Something is missing here. Like we have an identical start and an immediate failure. If you did it via printk() approach, it's probably wrong as my patch uses internal regmap function. Most likely you need to enable trace events for regmap and collect those for let's say 2 seconds:

	echo 1 > ...trace events...
	modprobe ...
	sleep 2
	echo 0 > ...trace events...
Quoted 7 lines

and dump the buffer to a file. It might have though more than needed as some other devices might also use regmap at the same time. I don't remember if the trace events for regmap have a device instance name field which can be used as a filter.

Alternatively you may also try to add a printk() into regmap core, but I don't think it's more practical than trace events.

Meanwhile, can you test this patch (on top of non-working case)?

drivers/base/regmap/regmap.c +12 -1
index 2314744201b4..f799a7a80231 100644
@@ -1553,8 +1553,19 @@ static int _regmap_select_page(struct regmap *map, unsigned int *reg,
 		 * virtual copy as well.
 		 */
 		if (page_chg &&
-		    in_range(range->selector_reg, range->window_start, range->window_len))
+		    in_range(range->selector_reg, range->window_start, range->window_len)) {
+			bool bypass, cache_only;
+
+			bypass = map->cache_bypass;
+			cache_only = map->cache_only;
+			map->cache_bypass = false;
+			map->cache_only = true;
+
 			_regmap_update_bits(map, sel_register, mask, val, NULL, false);
+
+			map->cache_bypass = bypass;
+			map->cache_only = cache_only;
+		}
 	}
 
 	*reg = range->window_start + win_offset;

If I understood the case, the affected driver doesn't use case and we actually write to the selector register twice which most likely messes up the things. But this is only a theory (since we don't have the traces yet).

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#14 lore raw
Quoted 29 lines · Andy Shevchenko
Quoted 1 line

[...]

Meanwhile, can you test this patch (on top of non-working case)?

drivers/base/regmap/regmap.c +12 -1
index 2314744201b4..f799a7a80231 100644
@@ -1553,8 +1553,19 @@ static int _regmap_select_page(struct regmap *map, unsigned int *reg,
  		 * virtual copy as well.
  		 */
  		if (page_chg &&
-		    in_range(range->selector_reg, range->window_start, range->window_len))
+		    in_range(range->selector_reg, range->window_start, range->window_len)) {
+			bool bypass, cache_only;
+
+			bypass = map->cache_bypass;
+			cache_only = map->cache_only;
+			map->cache_bypass = false;
+			map->cache_only = true;
+
  			_regmap_update_bits(map, sel_register, mask, val, NULL, false);
+
+			map->cache_bypass = bypass;
+			map->cache_only = cache_only;
+		}
  	}
 
  	*reg = range->window_start + win_offset;
Quoted 3 lines

If I understood the case, the affected driver doesn't use case and we actually write to the selector register twice which most likely messes up the things. But this is only a theory (since we don't have the traces yet).

Bingo! This patch (on top of current linux-next) fixed the probe issue. Feel free to add:

(although I'm not sure if this is a fix for the generic case or just this driver)

Best regards

Signature · 2 lines
Marek Szyprowski, PhD
Samsung R&D Institute Poland
#16 lore raw
In reply to Marek Szyprowski
Quoted 8 lines · Marek Szyprowski
Quoted 6 lines · Andy Shevchenko
Quoted 1 line

[...]

Meanwhile, can you test this patch (on top of non-working case)?

If I understood the case, the affected driver doesn't use case and we actually write to the selector register twice which most likely messes up the things. But this is only a theory (since we don't have the traces yet).

Quoted 5 lines

Bingo! This patch (on top of current linux-next) fixed the probe issue. Feel free to add:

Thank you! I will test it on my HW as well to confirm that there no regressions and will submit a formal fix.

Quoted 2 lines

(although I'm not sure if this is a fix for the generic case or just this driver)

This is for all cases where cache is not in use.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#17 lore raw
In reply to Andy Shevchenko
Quoted 96 lines · Andy Shevchenko
Quoted 59 lines · Andy Shevchenko
Quoted 20 lines · Marek Szyprowski
Quoted 19 lines · Andy Shevchenko
Quoted 6 lines · Marek Szyprowski
Quoted 3 lines · Andy Shevchenko

On Fri, Jan 17, 2025 at 05:05:42PM +0100, Marek Szyprowski wrote:

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Hmm... This is very interesting! It means that the page selector is a bit magical there. Dmitry, can you chime in and perhaps shed some light on this?

Quoted 9 lines
Quoted 9 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

It would be helpful! Traces from the failed and non-failed cases till the firmware revision and chip ID reading would be enough to start with.

I'm sorry for the delay, I was a bit busy with other stuff.

No problem and thanks for sharing.

Quoted 11 lines

Here are logs (all values are in hex):

next-20250128 (probe broken):

root@target:~# dmesg | grep regmap
[   14.817604] regmap_write reg 80ee <- 1
[   14.823036] regmap_read reg 8100 -> 0
[   14.827631] regmap_read reg 8101 -> 0
[   14.832130] regmap_read reg 8102 -> 0
Quoted 6 lines

next-20250128 + 1fd60ed1700c reverted (probe okay):

root@target:~# dmesg | grep regmap
[   13.565920] regmap_write reg 80ee <- 1
[   13.567509] regmap_read reg 8100 -> 17
[   13.568219] regmap_read reg 8101 -> 4
[   13.568909] regmap_read reg 8102 -> 93

Something is missing here. Like we have an identical start and an immediate failure. If you did it via printk() approach, it's probably wrong as my patch uses internal regmap function. Most likely you need to enable trace events for regmap and collect those for let's say 2 seconds:

	echo 1 > ...trace events...
	modprobe ...
	sleep 2
	echo 0 > ...trace events...
Quoted 7 lines

and dump the buffer to a file. It might have though more than needed as some other devices might also use regmap at the same time. I don't remember if the trace events for regmap have a device instance name field which can be used as a filter.

Alternatively you may also try to add a printk() into regmap core, but I don't think it's more practical than trace events.

Meanwhile, can you test this patch (on top of non-working case)?

drivers/base/regmap/regmap.c +12 -1
index 2314744201b4..f799a7a80231 100644
@@ -1553,8 +1553,19 @@ static int _regmap_select_page(struct regmap *map, unsigned int *reg,
 		 * virtual copy as well.
 		 */
 		if (page_chg &&
-		    in_range(range->selector_reg, range->window_start, range->window_len))
+		    in_range(range->selector_reg, range->window_start, range->window_len)) {
+			bool bypass, cache_only;
+
+			bypass = map->cache_bypass;
+			cache_only = map->cache_only;
+			map->cache_bypass = false;
+			map->cache_only = true;
+
 			_regmap_update_bits(map, sel_register, mask, val, NULL, false);
+
+			map->cache_bypass = bypass;
+			map->cache_only = cache_only;
+		}
 	}
 
 	*reg = range->window_start + win_offset;
Quoted 2 lines

If I understood the case, the affected driver doesn't use case and we actually write to the selector register twice which most likely messes up the things.

Unfortunately I can not comment regarding the LT9611UXC itself, the datasheet that I have here is pretty cryptic, incomplete and partially written in Mandarin.

This patch though fixes an issue for me too, So:

Quoted 7 lines

But this is only a theory (since we don't have the traces yet).

Signature · 2 lines
With Best Regards,
Andy Shevchenko
Signature · 2 lines
With best wishes
Dmitry
#18 lore raw
Quoted 104 lines · Dmitry Baryshkov
Quoted 96 lines · Andy Shevchenko
Quoted 59 lines · Andy Shevchenko
Quoted 20 lines · Marek Szyprowski
Quoted 19 lines · Andy Shevchenko
Quoted 6 lines · Marek Szyprowski
Quoted 3 lines · Andy Shevchenko

On Fri, Jan 17, 2025 at 05:05:42PM +0100, Marek Szyprowski wrote:

Does it fail in the same way?

Yes, the hw revision is reported as zero in this case: LT9611 revision:
0x00.00.00

Hmm... This is very interesting! It means that the page selector is a bit magical there. Dmitry, can you chime in and perhaps shed some light on this?

Quoted 9 lines
Quoted 9 lines
Quoted 1 line

Does it mean that there is really a bug in the driver?

Without looking at the datasheet it's hard to say. At least what I found so far is one page of the I²C traffic dump on Windows as an example how to use their evaluation board and software, but it doesn't unveil the bigger picture. At least what I think is going on here is that the programming is not so easy as just paging. Something is more complicated there.

But at least (and as Mark said) the most of the regmap based drivers got the ranges wrong (so, at least there is one bug in the driver).

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

I can do more experiments if this helps. Do you need a dump of all regmap accesses or i2c traffic from this driver?

It would be helpful! Traces from the failed and non-failed cases till the firmware revision and chip ID reading would be enough to start with.

I'm sorry for the delay, I was a bit busy with other stuff.

No problem and thanks for sharing.

Quoted 11 lines

Here are logs (all values are in hex):

next-20250128 (probe broken):

root@target:~# dmesg | grep regmap
[   14.817604] regmap_write reg 80ee <- 1
[   14.823036] regmap_read reg 8100 -> 0
[   14.827631] regmap_read reg 8101 -> 0
[   14.832130] regmap_read reg 8102 -> 0
Quoted 6 lines

next-20250128 + 1fd60ed1700c reverted (probe okay):

root@target:~# dmesg | grep regmap
[   13.565920] regmap_write reg 80ee <- 1
[   13.567509] regmap_read reg 8100 -> 17
[   13.568219] regmap_read reg 8101 -> 4
[   13.568909] regmap_read reg 8102 -> 93

Something is missing here. Like we have an identical start and an immediate failure. If you did it via printk() approach, it's probably wrong as my patch uses internal regmap function. Most likely you need to enable trace events for regmap and collect those for let's say 2 seconds:

	echo 1 > ...trace events...
	modprobe ...
	sleep 2
	echo 0 > ...trace events...
Quoted 7 lines

and dump the buffer to a file. It might have though more than needed as some other devices might also use regmap at the same time. I don't remember if the trace events for regmap have a device instance name field which can be used as a filter.

Alternatively you may also try to add a printk() into regmap core, but I don't think it's more practical than trace events.

Meanwhile, can you test this patch (on top of non-working case)?

drivers/base/regmap/regmap.c +12 -1
index 2314744201b4..f799a7a80231 100644
@@ -1553,8 +1553,19 @@ static int _regmap_select_page(struct regmap *map, unsigned int *reg,
 		 * virtual copy as well.
 		 */
 		if (page_chg &&
-		    in_range(range->selector_reg, range->window_start, range->window_len))
+		    in_range(range->selector_reg, range->window_start, range->window_len)) {
+			bool bypass, cache_only;
+
+			bypass = map->cache_bypass;
+			cache_only = map->cache_only;
+			map->cache_bypass = false;
+			map->cache_only = true;
+
 			_regmap_update_bits(map, sel_register, mask, val, NULL, false);
+
+			map->cache_bypass = bypass;
+			map->cache_only = cache_only;
+		}
 	}
 
 	*reg = range->window_start + win_offset;
Quoted 2 lines

If I understood the case, the affected driver doesn't use case and we actually write to the selector register twice which most likely messes up the things.

Unfortunately I can not comment regarding the LT9611UXC itself, the datasheet that I have here is pretty cryptic, incomplete and partially written in Mandarin.

Quoted 3 lines

This patch though fixes an issue for me too, So:

Thank you, guys, for reporting an testing, but it seems the simple problem to solve requires a lot of changes to be done without regressions (this fix on fix makes a regression to those who have cache enabled), which means that for now I propose to revert it (or drop) if possible, Mark, what is your preference?

Quoted 1 line
Quoted 1 line

But this is only a theory (since we don't have the traces yet).

Signature · 2 lines
With Best Regards,
Andy Shevchenko
maintainer
#19 lore raw
Quoted 5 lines · Andy Shevchenko

Thank you, guys, for reporting an testing, but it seems the simple problem to solve requires a lot of changes to be done without regressions (this fix on fix makes a regression to those who have cache enabled), which means that for now I propose to revert it (or drop) if possible, Mark, what is your preference?

I dropped it a while ago, it's not in mainline.

#20 lore raw
Quoted 9 lines · Mark Brown
Quoted 5 lines · Andy Shevchenko

Thank you, guys, for reporting an testing, but it seems the simple problem to solve requires a lot of changes to be done without regressions (this fix on fix makes a regression to those who have cache enabled), which means that for now I propose to revert it (or drop) if possible, Mark, what is your preference?

I dropped it a while ago, it's not in mainline.

Okay, thanks!

Signature · 2 lines
With Best Regards,
Andy Shevchenko