On Monday 23 November 2015 23:52:58 Amitoj Kaur Chawla wrote:
> 32 bit systems using 'struct timeval' will break in the year 2038, so
> we modify the code appropriately.
>
> This patch replaces the use of struct timeval and do_gettimeofday()
> with ktime_get_real_seconds() which returns a 64 bit value which is
> safer than struct timeval.
>
> Signed-off-by: Amitoj Kaur Chawla <[email protected]>
The comment doesn't match what the patch does:
> diff --git a/drivers/scsi/bfa/bfa_svc.c b/drivers/scsi/bfa/bfa_svc.c
> index 6521896..0240554 100644
> --- a/drivers/scsi/bfa/bfa_svc.c
> +++ b/drivers/scsi/bfa/bfa_svc.c
> @@ -3424,13 +3424,10 @@ __bfa_cb_fcport_stats_clr(void *cbarg, bfa_boolean_t
> complete)
> struct list_head *qe, *qen;
>
> if (complete) {
> - struct timeval tv;
> -
> /*
> * re-initialize time stamp for stats reset
> */
> - do_gettimeofday(&tv);
> - fcport->stats_reset_time = tv.tv_sec;
> + fcport->stats_reset_time = ktime_get_seconds();
> list_for_each_safe(qe, qen, &fcport->statsclr_pending_q) {
> bfa_q_deq(&fcport->statsclr_pending_q, &qe);
> cb = (struct bfa_cb_pending_q_s *)qe;
You use ktime_get_seconds() rather than ktime_get_real_seconds(), and you don't
store the result in a 64-bit variable, so the range extension doesn't work.
This is ok if all uses of the variable are changed to monotonic time, but
in this case they are not:
From looking at the uses of stats_reset_time, it seems that you need to change
__bfa_cb_fcport_stats_get at the same time.
Arnd
_______________________________________________
Y2038 mailing list
[email protected]
https://lists.linaro.org/mailman/listinfo/y2038
