timekeeping: Fix timex status validation for auxiliary clocks
Miroslav Lichvar
reviewerPATCH 1/1 timekeeping: Fix timex status validation for auxiliary clocks
The timekeeping_validate_timex() function validated the timex status of an auxiliary system clock even when the status is not to be changed, which caused unexpected errors for applications that make read-only clock_adjtime() calls, or set some other timex fields, but without clearing the status field.
Do the AUX-specific status validation only when the modes field contains ADJ_STATUS, i.e. the application is actually trying to change the status. This makes the AUX-specific clock_adjtime() behavior consistent with CLOCK_REALTIME.
--- kernel/time/timekeeping.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
kernel/time/timekeeping.c +4 -2
@@ -2653,7 +2653,8 @@ static int timekeeping_validate_timex(const struct __kernel_timex *txc, bool auxif (aux_clock) {/* Auxiliary clocks are similar to TAI and do not have leap seconds */ -if (txc->status & (STA_INS |STA_DEL )) +if (txc->modes &ADJ_STATUS && + txc->status & (STA_INS |STA_DEL ))return -EINVAL ;/* No TAI offset setting */
@@ -2661,7 +2662,8 @@ static int timekeeping_validate_timex(const struct __kernel_timex *txc, bool auxreturn -EINVAL ;/* No PPS support either */ -if (txc->status & (STA_PPSFREQ |STA_PPSTIME )) +if (txc->modes &ADJ_STATUS && + txc->status & (STA_PPSFREQ |STA_PPSTIME ))return -EINVAL ; }
Signature · 1 line
2.53.0
The following commit has been merged into the timers/urgent branch of tip:
The following commit has been merged into the timers/urgent branch of tip:
Commit-ID: e48a869957a70cc39b4090cd27c36a86f8db9b92 Gitweb: https://git.kernel.org/tip/e48a869957a70cc39b4090cd27c36a86f8db9b92 Author: Miroslav Lichvar <[email protected]> AuthorDate: Wed, 25 Feb 2026 09:51:35 +01:00 Committer: Thomas Gleixner <[email protected]> CommitterDate: Wed, 04 Mar 2026 20:05:37 +01:00
timekeeping: Fix timex status validation for auxiliary clocks
The timekeeping_validate_timex() function validates the timex status of an auxiliary system clock even when the status is not to be changed, which causes unexpected errors for applications that make read-only clock_adjtime() calls, or set some other timex fields, but without clearing the status field.
Do the AUX-specific status validation only when the modes field contains ADJ_STATUS, i.e. the application is actually trying to change the status. This makes the AUX-specific clock_adjtime() behavior consistent with CLOCK_REALTIME.
--- kernel/time/timekeeping.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
kernel/time/timekeeping.c +4 -2
@@ -2653,7 +2653,8 @@ static int timekeeping_validate_timex(const struct __kernel_timex *txc, bool auxif (aux_clock) {/* Auxiliary clocks are similar to TAI and do not have leap seconds */ -if (txc->status & (STA_INS |STA_DEL )) +if (txc->modes &ADJ_STATUS && + txc->status & (STA_INS |STA_DEL ))return -EINVAL ;/* No TAI offset setting */
@@ -2661,7 +2662,8 @@ static int timekeeping_validate_timex(const struct __kernel_timex *txc, bool auxreturn -EINVAL ;/* No PPS support either */ -if (txc->status & (STA_PPSFREQ |STA_PPSTIME )) +if (txc->modes &ADJ_STATUS && + txc->status & (STA_PPSFREQ |STA_PPSTIME ))return -EINVAL ; }