Documentation/ABI/testing/sysfs-pps | 27 +++++++++++++++++++++++ Documentation/driver-api/pps.rst | 19 ++++++++++++++++ drivers/pps/kapi.c | 24 +++++++++++++++++--- drivers/pps/sysfs.c | 34 +++++++++++++++++++++++++++++ include/linux/pps_kernel.h | 2 ++ 5 files changed, 103 insertions(+), 3 deletions(-)
Some applications like Android needs elapsed realtime timestamping
to PPS pulse for its clock management. Add sysfs node for this.
Signed-off-by: Alexander Komrakov <alexander.komrakov@broadcom.com>
---
Documentation/ABI/testing/sysfs-pps | 27 +++++++++++++++++++++++
Documentation/driver-api/pps.rst | 19 ++++++++++++++++
drivers/pps/kapi.c | 24 +++++++++++++++++---
drivers/pps/sysfs.c | 34 +++++++++++++++++++++++++++++
include/linux/pps_kernel.h | 2 ++
5 files changed, 103 insertions(+), 3 deletions(-)
diff --git a/Documentation/ABI/testing/sysfs-pps b/Documentation/ABI/testing/sysfs-pps
index 25028c7bc37d..031ec89e1ed6 100644
--- a/Documentation/ABI/testing/sysfs-pps
+++ b/Documentation/ABI/testing/sysfs-pps
@@ -1,3 +1,30 @@
+What: /sys/class/pps/pps0/assert_elapsed
+Date: October 2021
+Contact: Alexander Komrakov <alexander.komrakov@broadcom.com>
+Description:
+ The /sys/class/pps/ppsX/assert_elapsed file reports the
+ elapsed real-time assert events and the elapsed
+ real-time assert sequence number of the X-th source
+ in the form:
+
+ <secs>.<nsec>#<sequence>
+
+ If the source has no elapsed real-time assert events
+ the content of this file is empty.
+
+What: /sys/class/pps/ppsX/clear_elapsed
+Date: October 2021
+Contact: Alexander Komrakov <alexander.komrakov@broadcom.com>
+Description:
+ The /sys/class/pps/ppsX/clear_elapsed file reports the elapsed
+ real-time clear events and the elapsed real-time clear
+ sequence number of the X-th source in the form:
+
+ <secs>.<nsec>#<sequence>
+
+ If the source has no elapsed real-time clear events the
+ content of this file is empty.
+
What: /sys/class/pps/
Date: February 2008
Contact: Rodolfo Giometti <giometti@linux.it>
diff --git a/Documentation/driver-api/pps.rst b/Documentation/driver-api/pps.rst
index 2d6b99766ee8..bd0e63f15f95 100644
--- a/Documentation/driver-api/pps.rst
+++ b/Documentation/driver-api/pps.rst
@@ -167,6 +167,25 @@ sequence number. Other files are:
* path: reports the PPS source's device path, that is the device the
PPS source is connected to (if it exists).
+Real-time assert event::
+
+ Calculate the monotonic clock from the timespec clock to generate PPS elapsed
+ real-time event value and store the result into /sys/class/pps/pps0/assert_elapsed.
+ Because we have requirements to make sure the delta between standard time,
+ say the GPS Time, and elapsedRealtime < 1 millisecond,
+ regular linux clock timestamp is not enough for our use case.
+ The pin PPS will generate elapsedRealtime event at 1 sec boundary
+ which is an exact value of the monotonic clock from the kernel PPS driver
+ /sys/class/pps/pps0/assert_elapsed.
+ Whenever AP receives this pulse, kernel's pps driver timestamp this
+ elapsedRealtime event and let this time available via sysfs node
+ /sys/class/pps/pps0/assert_elapsed or the IOCTL call.
+
+ The device's Unix epoch time system clock is not monotonic; it can jump forwards
+ or backwards unpredictably and may be changed at any time by the user,
+ so this time should not be used to order or compare locations.
+ Prefer for that purpose, as the elapsed realtime clock is guaranteed to be monotonic.
+
Testing the PPS support
-----------------------
diff --git a/drivers/pps/kapi.c b/drivers/pps/kapi.c
index d9d566f70ed1..1f3ddccf9642 100644
--- a/drivers/pps/kapi.c
+++ b/drivers/pps/kapi.c
@@ -162,11 +162,21 @@ void pps_event(struct pps_device *pps, struct pps_event_time *ts, int event,
unsigned long flags;
int captured = 0;
struct pps_ktime ts_real = { .sec = 0, .nsec = 0, .flags = 0 };
+ struct pps_ktime ts_real_elapsed = { .sec = 0, .nsec = 0, .flags = 0 };
+ struct timespec64 ts64 = { .tv_sec = 0, .tv_nsec = 0 };
/* check event type */
BUG_ON((event & (PPS_CAPTUREASSERT | PPS_CAPTURECLEAR)) == 0);
- dev_dbg(pps->dev, "PPS event at %lld.%09ld\n",
+ /* Calculate the monotonic clock from the timespec clock and stores the result in pps_ktime format
+ ktime_get_boottime_ts64() : because elapsed realtime includes time spent in sleep */
+ ktime_get_boottime_ts64(&ts64);
+ timespec_to_pps_ktime(&ts_real_elapsed,ts64);
+
+ dev_dbg(pps->dev, "PPS event (monotonic) at %lld.%09d\n",
+ (s64)ts_real_elapsed.sec, ts_real_elapsed.nsec);
+
+ dev_dbg(pps->dev, "PPS event (timestamp) at %lld.%09ld\n",
(s64)ts->ts_real.tv_sec, ts->ts_real.tv_nsec);
timespec_to_pps_ktime(&ts_real, ts->ts_real);
@@ -181,11 +191,15 @@ void pps_event(struct pps_device *pps, struct pps_event_time *ts, int event,
pps->current_mode = pps->params.mode;
if (event & pps->params.mode & PPS_CAPTUREASSERT) {
/* We have to add an offset? */
- if (pps->params.mode & PPS_OFFSETASSERT)
+ if (pps->params.mode & PPS_OFFSETASSERT) {
+ pps_add_offset(&ts_real_elapsed,
+ &pps->params.assert_off_tu);
pps_add_offset(&ts_real,
&pps->params.assert_off_tu);
+ }
/* Save the time stamp */
+ pps->assert_elapsed_tu = ts_real_elapsed;
pps->assert_tu = ts_real;
pps->assert_sequence++;
dev_dbg(pps->dev, "capture assert seq #%u\n",
@@ -195,11 +209,15 @@ void pps_event(struct pps_device *pps, struct pps_event_time *ts, int event,
}
if (event & pps->params.mode & PPS_CAPTURECLEAR) {
/* We have to add an offset? */
- if (pps->params.mode & PPS_OFFSETCLEAR)
+ if (pps->params.mode & PPS_OFFSETCLEAR) {
+ pps_add_offset(&ts_real_elapsed,
+ &pps->params.clear_off_tu);
pps_add_offset(&ts_real,
&pps->params.clear_off_tu);
+ }
/* Save the time stamp */
+ pps->clear_elapsed_tu = ts_real_elapsed;
pps->clear_tu = ts_real;
pps->clear_sequence++;
dev_dbg(pps->dev, "capture clear seq #%u\n",
diff --git a/drivers/pps/sysfs.c b/drivers/pps/sysfs.c
index 134bc33f6ad0..983f16b86a62 100644
--- a/drivers/pps/sysfs.c
+++ b/drivers/pps/sysfs.c
@@ -10,6 +10,7 @@
#include <linux/module.h>
#include <linux/string.h>
#include <linux/pps_kernel.h>
+#include <linux/sysfs.h>
/*
* Attribute functions
@@ -29,6 +30,21 @@ static ssize_t assert_show(struct device *dev, struct device_attribute *attr,
}
static DEVICE_ATTR_RO(assert);
+static ssize_t assert_elapsed_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct pps_device *pps = dev_get_drvdata(dev);
+
+ if (!(pps->info.mode & PPS_CAPTUREASSERT))
+ return 0;
+
+ return sysfs_emit(buf, "%lld.%09d#%d\n",
+ (long long) pps->assert_elapsed_tu.sec,
+ pps->assert_elapsed_tu.nsec,
+ pps->assert_sequence);
+}
+static DEVICE_ATTR_RO(assert_elapsed);
+
static ssize_t clear_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
@@ -43,6 +59,22 @@ static ssize_t clear_show(struct device *dev, struct device_attribute *attr,
}
static DEVICE_ATTR_RO(clear);
+static ssize_t clear_elapsed_show(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct pps_device *pps = dev_get_drvdata(dev);
+
+ if (!(pps->info.mode & PPS_CAPTURECLEAR))
+ return 0;
+
+ return sysfs_emit(buf, "%lld.%09d#%d\n",
+ (long long) pps->clear_elapsed_tu.sec,
+ pps->clear_elapsed_tu.nsec,
+ pps->clear_sequence);
+}
+static DEVICE_ATTR_RO(clear_elapsed);
+
static ssize_t mode_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
@@ -82,6 +114,8 @@ static DEVICE_ATTR_RO(path);
static struct attribute *pps_attrs[] = {
&dev_attr_assert.attr,
&dev_attr_clear.attr,
+ &dev_attr_assert_elapsed.attr,
+ &dev_attr_clear_elapsed.attr,
&dev_attr_mode.attr,
&dev_attr_echo.attr,
&dev_attr_name.attr,
diff --git a/include/linux/pps_kernel.h b/include/linux/pps_kernel.h
index 78c8ac4951b5..807642da12f7 100644
--- a/include/linux/pps_kernel.h
+++ b/include/linux/pps_kernel.h
@@ -47,6 +47,8 @@ struct pps_device {
__u32 assert_sequence; /* PPS assert event seq # */
__u32 clear_sequence; /* PPS clear event seq # */
+ struct pps_ktime assert_elapsed_tu; /* PPS elapsed rt assert seq # */
+ struct pps_ktime clear_elapsed_tu; /* PPS elapsed rt clear event seq */
struct pps_ktime assert_tu;
struct pps_ktime clear_tu;
int current_mode; /* PPS mode at event time */
--
2.25.1
On Thu, Mar 30, 2023 at 12:20:13AM -0700, Alexander Komrakov wrote: > Some applications like Android needs elapsed realtime timestamping > to PPS pulse for its clock management. Add sysfs node for this. Where is the Android code that needs this? Have a pointer to it anywhere? > > Signed-off-by: Alexander Komrakov <alexander.komrakov@broadcom.com> > --- > Documentation/ABI/testing/sysfs-pps | 27 +++++++++++++++++++++++ > Documentation/driver-api/pps.rst | 19 ++++++++++++++++ > drivers/pps/kapi.c | 24 +++++++++++++++++--- > drivers/pps/sysfs.c | 34 +++++++++++++++++++++++++++++ > include/linux/pps_kernel.h | 2 ++ > 5 files changed, 103 insertions(+), 3 deletions(-) My bot will remind you that you dropped the version information off :( > diff --git a/Documentation/ABI/testing/sysfs-pps b/Documentation/ABI/testing/sysfs-pps > index 25028c7bc37d..031ec89e1ed6 100644 > --- a/Documentation/ABI/testing/sysfs-pps > +++ b/Documentation/ABI/testing/sysfs-pps > @@ -1,3 +1,30 @@ > +What: /sys/class/pps/pps0/assert_elapsed That should be "ppsX" right? > +Date: October 2021 > +Contact: Alexander Komrakov <alexander.komrakov@broadcom.com> > +Description: > + The /sys/class/pps/ppsX/assert_elapsed file reports the > + elapsed real-time assert events and the elapsed > + real-time assert sequence number of the X-th source > + in the form: > + > + <secs>.<nsec>#<sequence> > + > + If the source has no elapsed real-time assert events > + the content of this file is empty. > + > +What: /sys/class/pps/ppsX/clear_elapsed > +Date: October 2021 > +Contact: Alexander Komrakov <alexander.komrakov@broadcom.com> > +Description: > + The /sys/class/pps/ppsX/clear_elapsed file reports the elapsed > + real-time clear events and the elapsed real-time clear > + sequence number of the X-th source in the form: > + > + <secs>.<nsec>#<sequence> > + > + If the source has no elapsed real-time clear events the > + content of this file is empty. > + > What: /sys/class/pps/ Why are your new entries above the main pps entry? Shouldn't they be below it? thanks, greg k-h
On Thu, Mar 30, 2023 at 12:20:13AM -0700, Alexander Komrakov wrote: > Some applications like Android needs elapsed realtime timestamping > to PPS pulse for its clock management. Add sysfs node for this. > > Signed-off-by: Alexander Komrakov <alexander.komrakov@broadcom.com> > --- > Documentation/ABI/testing/sysfs-pps | 27 +++++++++++++++++++++++ > Documentation/driver-api/pps.rst | 19 ++++++++++++++++ > drivers/pps/kapi.c | 24 +++++++++++++++++--- > drivers/pps/sysfs.c | 34 +++++++++++++++++++++++++++++ > include/linux/pps_kernel.h | 2 ++ > 5 files changed, 103 insertions(+), 3 deletions(-) > Hi, This is the friendly patch-bot of Greg Kroah-Hartman. You have sent him a patch that has triggered this response. He used to manually respond to these common problems, but in order to save his sanity (he kept writing the same thing over and over, yet to different people), I was created. Hopefully you will not take offence and will fix the problem in your patch and resubmit it so that it can be accepted into the Linux kernel tree. You are receiving this message because of the following common error(s) as indicated below: - This looks like a new version of a previously submitted patch, but you did not list below the --- line any changes from the previous version. Please read the section entitled "The canonical patch format" in the kernel file, Documentation/process/submitting-patches.rst for what needs to be done here to properly describe this. If you wish to discuss this problem further, or you have questions about how to resolve this issue, please feel free to respond to this email and Greg will reply once he has dug out from the pending patches received from other developers. thanks, greg k-h's patch email bot
© 2016 - 2026 Red Hat, Inc.