[PATCH v7 7/7] ptp: ptp_vmclock: return TAI not UTC
From: Takahiro Itazuri <hidden>
Date: 2026-01-30 17:38:00
Also in:
linux-devicetree, lkml
Subsystem:
networking drivers, ptp hardware clock support, ptp vmclock support, the rest · Maintainers:
Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Richard Cochran, David Woodhouse, Linus Torvalds
From: David Woodhouse <redacted> To output UTC would involve complex calculations about whether the time elapsed since the reference time has crossed the end of the month when a leap second takes effect. I've prototyped that, but it made me sad. Much better to report TAI, which is what PHCs should do anyway. And much much simpler. Signed-off-by: David Woodhouse <redacted> Signed-off-by: Babis Chalios <redacted> Tested-by: Takahiro Itazuri <redacted> --- drivers/ptp/ptp_vmclock.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/ptp/ptp_vmclock.c b/drivers/ptp/ptp_vmclock.c
index b6e9d4c96..c7c75e19f 100644
--- a/drivers/ptp/ptp_vmclock.c
+++ b/drivers/ptp/ptp_vmclock.c@@ -83,13 +83,13 @@ static uint64_t mul_u64_u64_shr_add_u64(uint64_t *res_hi, uint64_t delta, static bool tai_adjust(struct vmclock_abi *clk, uint64_t *sec) { - if (likely(clk->time_type == VMCLOCK_TIME_UTC)) + if (clk->time_type == VMCLOCK_TIME_TAI) return true; - if (clk->time_type == VMCLOCK_TIME_TAI && + if (clk->time_type == VMCLOCK_TIME_UTC && (le64_to_cpu(clk->flags) & VMCLOCK_FLAG_TAI_OFFSET_VALID)) { if (sec) - *sec += (int16_t)le16_to_cpu(clk->tai_offset_sec); + *sec -= (int16_t)le16_to_cpu(clk->tai_offset_sec); return true; } return false;
@@ -350,9 +350,9 @@ static struct ptp_clock *vmclock_ptp_register(struct device *dev, return NULL; } - /* Only UTC, or TAI with offset */ + /* Accept TAI directly, or UTC with valid offset for conversion to TAI */ if (!tai_adjust(st->clk, NULL)) { - dev_info(dev, "vmclock does not provide unambiguous UTC\n"); + dev_info(dev, "vmclock does not provide unambiguous time\n"); return NULL; }
--
2.50.1