[PATCH] bnxt: fix bnxt_hwrm_fw_set_time for y2038

[PATCH] bnxt: fix bnxt_hwrm_fw_set_time for y2038

[Arnd Bergmann]

On 32-bit architectures, rtc_time_to_tm() returns incorrect results
in 2038 or later, and do_gettimeofday() is broken for the same reason.

This changes the code to use ktime_get_real_seconds() and time64_to_tm()
instead, both of them are 2038-safe, and we can also get rid of the
CONFIG_RTC_LIB dependency that way.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 96416f5d97f3..5c963a8c2b0e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -4915,16 +4915,14 @@ static int bnxt_hwrm_ver_get(struct bnxt *bp)
 
 int bnxt_hwrm_fw_set_time(struct bnxt *bp)
 {
-#if IS_ENABLED(CONFIG_RTC_LIB)
 	struct hwrm_fw_set_time_input req = {0};
-	struct rtc_time tm;
-	struct timeval tv;
+	struct tm tm;
+	time64_t now = ktime_get_real_seconds();
 
 	if (bp->hwrm_spec_code < 0x10400)
 		return -EOPNOTSUPP;
 
-	do_gettimeofday(&tv);
-	rtc_time_to_tm(tv.tv_sec, &tm);
+	time64_to_tm(now, 0, &tm);
 	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_FW_SET_TIME, -1, -1);
 	req.year = cpu_to_le16(1900 + tm.tm_year);
 	req.month = 1 + tm.tm_mon;
@@ -4933,9 +4931,6 @@ int bnxt_hwrm_fw_set_time(struct bnxt *bp)
 	req.minute = tm.tm_min;
 	req.second = tm.tm_sec;
 	return hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
-#else
-	return -EOPNOTSUPP;
-#endif
 }
 
 static int bnxt_hwrm_port_qstats(struct bnxt *bp)
-- 
2.9.0

Re: [PATCH] bnxt: fix bnxt_hwrm_fw_set_time for y2038

[David Miller]

From: Arnd Bergmann <arnd@arndb.de>
Date: Mon,  6 Nov 2017 15:04:39 +0100

> On 32-bit architectures, rtc_time_to_tm() returns incorrect results
> in 2038 or later, and do_gettimeofday() is broken for the same reason.
> 
> This changes the code to use ktime_get_real_seconds() and time64_to_tm()
> instead, both of them are 2038-safe, and we can also get rid of the
> CONFIG_RTC_LIB dependency that way.
> 
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Applied, thanks Arnd.

The RTC_LIB dependency should have been done via KConfig anyways.