Skip to contentSkip to conversation

s390/stp: Drop CLOCK_SYNC_STP

PATCH In review Sven Schnelle 3 participants 5 messages Aug 14 – Sep 9 s390

lore
Share & tools
lore

Sven Schnelle

author reviewer
#0 lore raw

PATCH 1/1 s390/stp: Drop CLOCK_SYNC_STP

CLOCK_SYNC_STP is never set when the stp=1 kernel command line option is used, or stp is enabled by default. This cause get_phys_clock to return -EACCES. Fix this by testing stp_online and remove CLOCK_SYNC_STP.

---
 arch/s390/kernel/time.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)
arch/s390/kernel/time.c +3 -8
index 2b989bebd220..de80e2a316ba 100644
@@ -69,6 +69,7 @@ EXPORT_SYMBOL(ptff_function_mask);
 
 static unsigned long lpar_offset;
 static unsigned long initial_leap_seconds;
+static bool stp_online = true;
 
 /*
  * Get time offsets with PTFF
@@ -274,8 +275,7 @@ static DEFINE_MUTEX(stp_mutex);
 static unsigned long clock_sync_flags;
 
 #define CLOCK_SYNC_HAS_STP		0
-#define CLOCK_SYNC_STP			1
-#define CLOCK_SYNC_STPINFO_VALID	2
+#define CLOCK_SYNC_STPINFO_VALID	1
 
 /*
  * The get_clock function for the physical clock. It will get the current
@@ -299,7 +299,7 @@ int get_phys_clock(unsigned long *clock)
 		return 0;
 	if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
 		return -EOPNOTSUPP;
-	if (!test_bit(CLOCK_SYNC_STP, &clock_sync_flags))
+	if (!stp_online)
 		return -EACCES;
 	return -EAGAIN;
 }
@@ -397,7 +397,6 @@ struct clock_sync_data {
 /*
  * Server Time Protocol (STP) code.
  */
-static bool stp_online = true;
 static struct stp_sstpi stp_info;
 static void *stp_page;
 
@@ -795,10 +794,6 @@ static ssize_t online_store(struct device *dev,
 		return -EOPNOTSUPP;
 	mutex_lock(&stp_mutex);
 	stp_online = value;
-	if (stp_online)
-		set_bit(CLOCK_SYNC_STP, &clock_sync_flags);
-	else
-		clear_bit(CLOCK_SYNC_STP, &clock_sync_flags);
 	queue_work(time_sync_wq, &stp_work);
 	mutex_unlock(&stp_mutex);
 	return count;
Signature · 1 line
2.53.0
#1 lore raw
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider: - [High] Booting on systems with XRC-capable DASDs but no active STP network will hang indefinitely due to…

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [High] Booting on systems with XRC-capable DASDs but no active STP network will hang indefinitely due to an unhandled -EAGAIN loop.
- [Low] Lockless plain read/write of stp_online introduces a formal data race.
--

--- Patch [1]: [PATCH] s390/stp: Drop CLOCK_SYNC_STP --- commit 39bb08d28d4e120a87017a08eb301d774dd6baa2
Author: Sven Schnelle <[email protected]>

s390/stp: Drop CLOCK_SYNC_STP

This commit removes the CLOCK_SYNC_STP flag in favor of using the stp_online boolean directly. This change prevents get_phys_clock() from returning -EACCES when the kernel command line option stp=1 is used or STP is enabled by default.

Quoted 4 lines
arch/s390/kernel/time.c +0 -0
index 2b989bebd220a..de80e2a316ba7 100644

[ ... ]

arch/s390/kernel/time.c int get_phys_clock(unsigned long *clock) · line 302

Quoted 6 lines
arch/s390/kernel/time.c +1 -1
@@ -299,7 +299,7 @@ int get_phys_clock(unsigned long *clock)
 		return 0;
 	if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
 		return -EOPNOTSUPP;
-	if (!test_bit(CLOCK_SYNC_STP, &clock_sync_flags))
+	if (!stp_online)

[Severity: Low]
Does this introduce a formal data race under the LKMM?

get_phys_clock() now performs a lockless plain read of stp_online. Meanwhile, online_store() modifies it concurrently via a plain write:

arch/s390/kernel/time.c:online_store() {
    ...
    mutex_lock(&stp_mutex);
    stp_online = value;
    ...
}

Could this trigger KCSAN warnings when userspace toggles the sysfs online attribute while get_phys_clock() is executing?

arch/s390/kernel/time.c int get_phys_clock(unsigned long *clock) · line 305

Quoted 3 lines
arch/s390/kernel/time.c +0 -0
 		return -EACCES;
 	return -EAGAIN;
 }

[Severity: High]
Could this change cause a boot hang on systems with XRC-capable DASDs but no active STP network?

Because stp_online is initialized to true, if the machine supports the STP facility but is not connected to a usable STP time source, time synchronization won't occur and stp_online will remain true.

With stp_online remaining true, get_phys_clock() will continuously fall through and return -EAGAIN, rather than safely returning -EACCES as it did prior to this patch.

The DASD driver ignores -EACCES but propagates -EAGAIN:

drivers/s390/block/dasd_eckd.c:set_timestamp() {
    ...
	if ((rc && !private->rdc_data.facilities.XRC_supported) ||
	    rc == -EOPNOTSUPP || rc == -EACCES)
		return 0;
    ...
}

If -EAGAIN is propagated up, will the block layer endlessly retry the I/O, resulting in a complete hang during boot?

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
#2 lore raw

Am 14.08.26 um 09:34 schrieb [email protected]:

Quoted 69 lines

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [High] Booting on systems with XRC-capable DASDs but no active STP network will hang indefinitely due to an unhandled -EAGAIN loop.
- [Low] Lockless plain read/write of stp_online introduces a formal data race.
--

--- Patch [1]: [PATCH] s390/stp: Drop CLOCK_SYNC_STP --- commit 39bb08d28d4e120a87017a08eb301d774dd6baa2
Author: Sven Schnelle <[email protected]>

s390/stp: Drop CLOCK_SYNC_STP

This commit removes the CLOCK_SYNC_STP flag in favor of using the stp_online boolean directly. This change prevents get_phys_clock() from returning -EACCES when the kernel command line option stp=1 is used or STP is enabled by default.

Quoted 4 lines
arch/s390/kernel/time.c +0 -0
index 2b989bebd220a..de80e2a316ba7 100644

[ ... ]

arch/s390/kernel/time.c int get_phys_clock(unsigned long *clock) · line 302

Quoted 6 lines
arch/s390/kernel/time.c +1 -1
@@ -299,7 +299,7 @@ int get_phys_clock(unsigned long *clock)
 		return 0;
 	if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
 		return -EOPNOTSUPP;
-	if (!test_bit(CLOCK_SYNC_STP, &clock_sync_flags))
+	if (!stp_online)

[Severity: Low]
Does this introduce a formal data race under the LKMM?

get_phys_clock() now performs a lockless plain read of stp_online. Meanwhile, online_store() modifies it concurrently via a plain write:

arch/s390/kernel/time.c:online_store() {
    ...
    mutex_lock(&stp_mutex);
    stp_online = value;
    ...
}

Could this trigger KCSAN warnings when userspace toggles the sysfs online attribute while get_phys_clock() is executing?

arch/s390/kernel/time.c int get_phys_clock(unsigned long *clock) · line 305

Quoted 3 lines
arch/s390/kernel/time.c +0 -0
 		return -EACCES;
 	return -EAGAIN;
 }

[Severity: High]
Could this change cause a boot hang on systems with XRC-capable DASDs but no active STP network?

Because stp_online is initialized to true, if the machine supports the STP facility but is not connected to a usable STP time source, time synchronization won't occur and stp_online will remain true.

With stp_online remaining true, get_phys_clock() will continuously fall through and return -EAGAIN, rather than safely returning -EACCES as it did prior to this patch.

The DASD driver ignores -EACCES but propagates -EAGAIN:

drivers/s390/block/dasd_eckd.c:set_timestamp() {
    ...
	if ((rc && !private->rdc_data.facilities.XRC_supported) ||
	    rc == -EOPNOTSUPP || rc == -EACCES)
		return 0;
    ...
}
Quoted 3 lines

If -EAGAIN is propagated up, will the block layer endlessly retry the I/O, resulting in a complete hang during boot?

From what I see, Sashiko is right. This will break how the DASD driver actually checks if the clock is enabled and in sync.

The DASD driver relies on -EACCES to know that the clock exists but STP sync is not online/applicable which in turn will skip it, whereas -EAGAIN means the clock does exist but is not in sync yet and we should wait and retry, which in turn leads to endless loop of request never being built because we never get a valid timestamp.

So I guess we will still need a third bit to differentiate those states.

author reviewer
#3 lore raw
Quoted 30 lines · Stefan Haberland

Am 14.08.26 um 09:34 schrieb [email protected]:

Quoted 22 lines

[Severity: High]
Could this change cause a boot hang on systems with XRC-capable DASDs but no active STP network?

Because stp_online is initialized to true, if the machine supports the STP facility but is not connected to a usable STP time source, time synchronization won't occur and stp_online will remain true.

With stp_online remaining true, get_phys_clock() will continuously fall through and return -EAGAIN, rather than safely returning -EACCES as it did prior to this patch.

The DASD driver ignores -EACCES but propagates -EAGAIN:

drivers/s390/block/dasd_eckd.c:set_timestamp() {
    ...
	if ((rc && !private->rdc_data.facilities.XRC_supported) ||
	    rc == -EOPNOTSUPP || rc == -EACCES)
		return 0;
    ...
}
Quoted 3 lines

If -EAGAIN is propagated up, will the block layer endlessly retry the I/O, resulting in a complete hang during boot?

From what I see, Sashiko is right. This will break how the DASD driver actually checks if the clock is enabled and in sync.

Quoted 7 lines

The DASD driver relies on -EACCES to know that the clock exists but STP sync is not online/applicable which in turn will skip it, whereas -EAGAIN means the clock does exist but is not in sync yet and we should wait and retry, which in turn leads to endless loop of request never being built because we never get a valid timestamp.

So I guess we will still need a third bit to differentiate those states.

If I understood the code correctly, it would only block until STP is in sync, so this is expected behaviour - looking at z/VM documentation (especially the XRC_OPTional flag), DASD I/O should be blocked until time is in sync.

I think the correct way would be to add a patch on top that adds the XRC_OPTional parameter to either block DASD I/O when STP is in unsynchronized state (current behaviour) or just omits the timestamp when XRC_OPTional is set.

#4 lore raw

Am 09.09.26 um 14:29 schrieb Sven Schnelle:

Quoted 45 lines
Quoted 30 lines · Stefan Haberland

Am 14.08.26 um 09:34 schrieb [email protected]:

Quoted 22 lines

[Severity: High]
Could this change cause a boot hang on systems with XRC-capable DASDs but no active STP network?

Because stp_online is initialized to true, if the machine supports the STP facility but is not connected to a usable STP time source, time synchronization won't occur and stp_online will remain true.

With stp_online remaining true, get_phys_clock() will continuously fall through and return -EAGAIN, rather than safely returning -EACCES as it did prior to this patch.

The DASD driver ignores -EACCES but propagates -EAGAIN:

drivers/s390/block/dasd_eckd.c:set_timestamp() {
    ...
	if ((rc && !private->rdc_data.facilities.XRC_supported) ||
	    rc == -EOPNOTSUPP || rc == -EACCES)
		return 0;
    ...
}
Quoted 3 lines

If -EAGAIN is propagated up, will the block layer endlessly retry the I/O, resulting in a complete hang during boot?

From what I see, Sashiko is right. This will break how the DASD driver actually checks if the clock is enabled and in sync.

Quoted 7 lines

The DASD driver relies on -EACCES to know that the clock exists but STP sync is not online/applicable which in turn will skip it, whereas -EAGAIN means the clock does exist but is not in sync yet and we should wait and retry, which in turn leads to endless loop of request never being built because we never get a valid timestamp.

So I guess we will still need a third bit to differentiate those states.

If I understood the code correctly, it would only block until STP is in sync, so this is expected behaviour - looking at z/VM documentation (especially the XRC_OPTional flag), DASD I/O should be blocked until time is in sync.

Quoted 4 lines

I think the correct way would be to add a patch on top that adds the XRC_OPTional parameter to either block DASD I/O when STP is in unsynchronized state (current behaviour) or just omits the timestamp when XRC_OPTional is set.

Not sure XRC_OPTional fully covers the case here, though. The existing "don't count -EAGAIN against retries" logic assumes the clock will eventually converge, a real STP network mid-sync.

But the scenario here is different: no CTN/network attached at all, so there's nothing to converge the state only changes via admin action, not by waiting.

Wouldn't a mandatory-XRC device on a machine with no STP network at all still block forever either way, regardless of XRC_OPTional? That flag decides whether a given volume needs a timestamp, but not whether the system can ever produce one.

Could it make sense for get_phys_clock() to distinguish "converging" from "no usable source at all" itself independent of the XRC_OPTional patch?