[PATCH 1/1] regmap: debugfs: runtime resume a device when reading registers

[PATCH 1/1] regmap: debugfs: runtime resume a device when reading registers

[Alexander Stein]

If the registers shall be read from device, runtime resume it in order
to enable a possiblly attached power domain before accessing the device.

Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
---
 drivers/base/regmap/regmap-debugfs.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/base/regmap/regmap-debugfs.c b/drivers/base/regmap/regmap-debugfs.c
index fb84cda92a753..063ede23173af 100644
--- a/drivers/base/regmap/regmap-debugfs.c
+++ b/drivers/base/regmap/regmap-debugfs.c
@@ -12,6 +12,7 @@
 #include <linux/uaccess.h>
 #include <linux/device.h>
 #include <linux/list.h>
+#include <linux/pm_runtime.h>
 
 #include "internal.h"
 
@@ -233,6 +234,10 @@ static ssize_t regmap_read_debugfs(struct regmap *map, unsigned int from,
 	if (!buf)
 		return -ENOMEM;
 
+	ret = pm_runtime_resume_and_get(map->dev);
+	if (ret)
+		goto out_free;
+
 	regmap_calc_tot_len(map, buf, count);
 
 	/* Work out which register we're starting at */
@@ -277,6 +282,8 @@ static ssize_t regmap_read_debugfs(struct regmap *map, unsigned int from,
 	*ppos += buf_pos;
 
 out:
+	pm_runtime_put(map->dev);
+out_free:
 	kfree(buf);
 	return ret;
 }
-- 
2.34.1

Re: [PATCH 1/1] regmap: debugfs: runtime resume a device when reading registers

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 402 bytes --]

On Wed, Jan 10, 2024 at 10:53:58AM +0100, Alexander Stein wrote:

> If the registers shall be read from device, runtime resume it in order
> to enable a possiblly attached power domain before accessing the device.

The idea is that the debugfs interface isn't supposed to be disruptive
to the thing being debugged.  It would be better to detect if there will
be problems and report the status as busy.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 1/1] regmap: debugfs: runtime resume a device when reading registers

[Alexander Stein]

Am Mittwoch, 10. Januar 2024, 13:23:17 CET schrieb Mark Brown:
> On Wed, Jan 10, 2024 at 10:53:58AM +0100, Alexander Stein wrote:
> > If the registers shall be read from device, runtime resume it in order
> > to enable a possiblly attached power domain before accessing the device.
> 
> The idea is that the debugfs interface isn't supposed to be disruptive
> to the thing being debugged.  It would be better to detect if there will
> be problems and report the status as busy.

In my case the device is actually unused, runtime suspended, thus disabled 
power domain. That's totally different to busy. In this case dumping the 
registers is non-disruptive, unless you account enabling/disabling the power 
domain as well. Any attached clock is already enabled/disabled, but power 
domains are not.

Best regards,
Alexander
-- 
TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany
Amtsgericht München, HRB 105018
Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider
http://www.tq-group.com/

Re: [PATCH 1/1] regmap: debugfs: runtime resume a device when reading registers

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 837 bytes --]

On Wed, Jan 10, 2024 at 01:46:31PM +0100, Alexander Stein wrote:
> Am Mittwoch, 10. Januar 2024, 13:23:17 CET schrieb Mark Brown:

> > The idea is that the debugfs interface isn't supposed to be disruptive
> > to the thing being debugged.  It would be better to detect if there will
> > be problems and report the status as busy.

> In my case the device is actually unused, runtime suspended, thus disabled 
> power domain. That's totally different to busy. In this case dumping the 
> registers is non-disruptive, unless you account enabling/disabling the power 
> domain as well. Any attached clock is already enabled/disabled, but power 
> domains are not.

At a regmap level we map that onto -EBUSY for devices with a cache
(using a cache for your specific device ought to DTRT, it's not a
general solution though).

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]