struct linux_rtc_time promised, struct tm provided, is this ok?

[matthewnourse]

In sys-utils/hwclock-rtc.c :
# define RTC_RD_TIME _IOR('p', 0x09, struct linux_rtc_time)

...but in do_rtc_read_ioctl(int rtc_fd, struct tm *tm) :
rc = ioctl(rtc_fd, RTC_RD_TIME, tm);

On my system the fields in linux_rtc_time appear in the same order as struct tm so all is well I guess...but...
...is this guaranteed to be the case everywhere & into the future?
...and is it ok that tm.tm_gmtoff and tm.tm_zone are uninitialized?

If not & there's a potential problem...would you like a pull request that passes a linux_rtc_time to ioctl and then copies the results to the tm (along the lines of what's already done for sparc)?

[karelzak]

just for the record, it should be fixed by commit a423fe1