struct mon_bin_hdr allows for a 64-bit seconds timestamp. The code
currently uses 'struct timeval' to populate the timestamp in mon_bin_hdr,
which has a 32-bit seconds field and will overflow in year 2038 and beyond.
This patch replaces 'struct timeval' with 'struct timespec64' which is
y2038 safe.
Signed-off-by: Tina Ruchandani <[email protected]>
---
drivers/usb/mon/mon_bin.c | 14 +++++++-------
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/usb/mon/mon_bin.c b/drivers/usb/mon/mon_bin.c
index 9a62e89..bfc2ebd 100644
--- a/drivers/usb/mon/mon_bin.c
+++ b/drivers/usb/mon/mon_bin.c
@@ -18,6 +18,7 @@
#include <linux/mm.h>
#include <linux/scatterlist.h>
#include <linux/slab.h>
+#include <linux/time64.h>
#include <asm/uaccess.h>
@@ -483,7 +484,7 @@ static void mon_bin_event(struct mon_reader_bin *rp, struct urb *urb,
char ev_type, int status)
{
const struct usb_endpoint_descriptor *epd = &urb->ep->desc;
- struct timeval ts;
+ struct timespec64 ts;
unsigned long flags;
unsigned int urb_length;
unsigned int offset;
@@ -494,7 +495,7 @@ static void mon_bin_event(struct mon_reader_bin *rp, struct urb *urb,
struct mon_bin_hdr *ep;
char data_tag = 0;
- do_gettimeofday(&ts);
+ getnstimeofday64(&ts);
spin_lock_irqsave(&rp->b_lock, flags);
@@ -568,7 +569,7 @@ static void mon_bin_event(struct mon_reader_bin *rp, struct urb *urb,
ep->busnum = urb->dev->bus->busnum;
ep->id = (unsigned long) urb;
ep->ts_sec = ts.tv_sec;
- ep->ts_usec = ts.tv_usec;
+ ep->ts_usec = ts.tv_nsec / NSEC_PER_USEC;
ep->status = status;
ep->len_urb = urb_length;
ep->len_cap = length + lendesc;
@@ -629,12 +630,12 @@ static void mon_bin_complete(void *data, struct urb *urb, int status)
static void mon_bin_error(void *data, struct urb *urb, int error)
{
struct mon_reader_bin *rp = data;
- struct timeval ts;
+ struct timespec64 ts;
unsigned long flags;
unsigned int offset;
struct mon_bin_hdr *ep;
- do_gettimeofday(&ts);
+ getnstimeofday64(&ts);
spin_lock_irqsave(&rp->b_lock, flags);
@@ -656,9 +657,8 @@ static void mon_bin_error(void *data, struct urb *urb, int error)
ep->busnum = urb->dev->bus->busnum;
ep->id = (unsigned long) urb;
ep->ts_sec = ts.tv_sec;
- ep->ts_usec = ts.tv_usec;
+ ep->ts_usec = ts.tv_nsec / NSEC_PER_USEC;
ep->status = error;
-
ep->flag_setup = '-';
ep->flag_data = 'E';
--
2.2.0.rc0.207.ga3a616c
On Tue, 2015-05-05 at 11:50 +0530, Tina Ruchandani wrote:
> struct mon_bin_hdr allows for a 64-bit seconds timestamp. The code
> currently uses 'struct timeval' to populate the timestamp in mon_bin_hdr,
> which has a 32-bit seconds field and will overflow in year 2038 and beyond.
> This patch replaces 'struct timeval' with 'struct timespec64' which is
> y2038 safe.
Hi,
but the timestamp will also overflow. So what is the point?
Regards
Oliver
> but the timestamp will also overflow. So what is the point?
>
The 32-bit timestamp obtained using do_gettimeofday() will overflow in
year 2038. However, with 64-bit timestamps, we get practically
infinite time. mon_bin_hdr already has support for a 64-bit seconds
timestamp. This patch changes how it is populated.
Thanks,
Tina
On Mon, 4 May 2015, Tina Ruchandani wrote:
> > but the timestamp will also overflow. So what is the point?
> >
>
> The 32-bit timestamp obtained using do_gettimeofday() will overflow in
> year 2038. However, with 64-bit timestamps, we get practically
> infinite time. mon_bin_hdr already has support for a 64-bit seconds
> timestamp. This patch changes how it is populated.
Surely the real point was mentioned in the patch description:
> ... This patch is part of a larger
> attempt to remove instances of struct timeval and other 32-bit timekeeping
> (time_t, struct timespec) from the kernel.
The point is to get rid of 32-bit timekeeping routines. Overflow and
the exact nature of the timestamps in usbmon are irrelevant.
Alan Stern