musl is moving to a default of 64-bit time_t on all architectures,
glibc will follow later. This breaks reading timestamps through cmsg
data with the HCI_TIME_STAMP socket option.
Change both copies of hcidump to work on all architectures. This also
fixes x32, which has never worked, and carefully avoids breaking sparc64,
which is another special case.
I have only compiled this on one architecture, please at least test
it to ensure there are no regressions. The toolchain binaries from
http://musl.cc/ should allow testing with a 64-bit time_t, but it may
be hard to build all the dependencies first.
libpcap has the same bug and needs a similar fix to work on future
32-bit Linux systems. Everything else apparently uses the generic
SO_TIMESTAMP timestamps, which work correctly when using new enough
kernels with a time64 libc.
Signed-off-by: Arnd Bergmann <[email protected]>
---
monitor/hcidump.c | 32 +++++++++++++++++++++++++++++++-
tools/hcidump.c | 33 +++++++++++++++++++++++++++++++--
2 files changed, 62 insertions(+), 3 deletions(-)
diff --git a/monitor/hcidump.c b/monitor/hcidump.c
index 8b6f846d3..6d2330287 100644
--- a/monitor/hcidump.c
+++ b/monitor/hcidump.c
@@ -107,6 +107,36 @@ static int open_hci_dev(uint16_t index)
return fd;
}
+static struct timeval hci_tstamp_read(void *data)
+{
+ struct timeval tv;
+
+ /*
+ * On 64-bit architectures, the data matches the timeval
+ * format. Note that on sparc64 this is different from
+ * all others.
+ */
+ if (sizeof(long) == 8) {
+ memcpy(&tv, data, sizeof(tv));
+ }
+
+ /*
+ * On 32-bit architectures, the timeval definition may
+ * use 32-bit or 64-bit members depending on the C
+ * library and architecture.
+ * The cmsg data however always contains a pair of
+ * 32-bit values. Interpret as unsigned to make it work
+ * past y2038.
+ */
+ if (sizeof(long) == 4) {
+ unsigned int *stamp = data;
+ tv.tv_sec = stamp[0];
+ tv.tv_usec = stamp[1];
+ }
+
+ return tv;
+}
+
static void device_callback(int fd, uint32_t events, void *user_data)
{
struct hcidump_data *data = user_data;
@@ -150,7 +180,7 @@ static void device_callback(int fd, uint32_t events, void *user_data)
memcpy(&dir, CMSG_DATA(cmsg), sizeof(dir));
break;
case HCI_CMSG_TSTAMP:
- memcpy(&ctv, CMSG_DATA(cmsg), sizeof(ctv));
+ ctv = hci_tstamp_read(CMSG_DATA(cmsg));
tv = &ctv;
break;
}
diff --git a/tools/hcidump.c b/tools/hcidump.c
index 33d429b6c..be14d0930 100644
--- a/tools/hcidump.c
+++ b/tools/hcidump.c
@@ -136,6 +136,36 @@ static inline int write_n(int fd, char *buf, int len)
return t;
}
+static struct timeval hci_tstamp_read(void *data)
+{
+ struct timeval tv;
+
+ /*
+ * On 64-bit architectures, the data matches the timeval
+ * format. Note that on sparc64 this is different from
+ * all others.
+ */
+ if (sizeof(long) == 8) {
+ memcpy(&tv, data, sizeof(tv));
+ }
+
+ /*
+ * On 32-bit architectures, the timeval definition may
+ * use 32-bit or 64-bit members depending on the C
+ * library and architecture.
+ * The cmsg data however always contains a pair of
+ * 32-bit values. Interpret as unsigned to make it work
+ * past y2038.
+ */
+ if (sizeof(long) == 4) {
+ unsigned int *stamp = data;
+ tv.tv_sec = stamp[0];
+ tv.tv_usec = stamp[1];
+ }
+
+ return tv;
+}
+
static int process_frames(int dev, int sock, int fd, unsigned long flags)
{
struct cmsghdr *cmsg;
@@ -230,8 +260,7 @@ static int process_frames(int dev, int sock, int fd, unsigned long flags)
frm.in = (uint8_t) dir;
break;
case HCI_CMSG_TSTAMP:
- memcpy(&frm.ts, CMSG_DATA(cmsg),
- sizeof(struct timeval));
+ frm.ts = hci_tstamp_read(CMSG_DATA(cmsg));
break;
}
cmsg = CMSG_NXTHDR(&msg, cmsg);
--
2.20.0
On Fri, Jan 10, 2020 at 09:49:03PM +0100, Arnd Bergmann wrote:
> musl is moving to a default of 64-bit time_t on all architectures,
> glibc will follow later. This breaks reading timestamps through cmsg
> data with the HCI_TIME_STAMP socket option.
>
> Change both copies of hcidump to work on all architectures. This also
> fixes x32, which has never worked, and carefully avoids breaking sparc64,
> which is another special case.
Won't it be broken on rv32 though? Based on my (albeit perhaps
incomplete) reading of the thread, I think use of HCI_TIME_STAMP
should just be dropped entirely in favor of using SO_TIMESTAMPNS -- my
understanding was that it works with bluetooth sockets too.
Rich
On Fri, Jan 10, 2020 at 10:05 PM Rich Felker <[email protected]> wrote:
>
> On Fri, Jan 10, 2020 at 09:49:03PM +0100, Arnd Bergmann wrote:
> > musl is moving to a default of 64-bit time_t on all architectures,
> > glibc will follow later. This breaks reading timestamps through cmsg
> > data with the HCI_TIME_STAMP socket option.
> >
> > Change both copies of hcidump to work on all architectures. This also
> > fixes x32, which has never worked, and carefully avoids breaking sparc64,
> > which is another special case.
>
> Won't it be broken on rv32 though? Based on my (albeit perhaps
> incomplete) reading of the thread, I think use of HCI_TIME_STAMP
> should just be dropped entirely in favor of using SO_TIMESTAMPNS -- my
> understanding was that it works with bluetooth sockets too.
All 32-bit architectures use old_timeval32 timestamps in the kernel
here, even rv32 and x32. As a rule, we keep the types bug-for-bug
compatible between architectures and fix them all at the same time.
Changing hcidump to SO_TIMESTAMPNS would work as well, but
that is a much bigger change and I don't know how to test that.
Arnd
On Jan 10, 2020, at 12:49 PM, Arnd Bergmann <[email protected]> wrote:
> diff --git a/monitor/hcidump.c b/monitor/hcidump.c
> index 8b6f846d3..6d2330287 100644
> --- a/monitor/hcidump.c
> +++ b/monitor/hcidump.c
> @@ -107,6 +107,36 @@ static int open_hci_dev(uint16_t index)
> return fd;
> }
>
> +static struct timeval hci_tstamp_read(void *data)
> +{
> + struct timeval tv;
> +
> + /*
> + * On 64-bit architectures, the data matches the timeval
> + * format. Note that on sparc64 this is different from
> + * all others.
> + */
> + if (sizeof(long) == 8) {
> + memcpy(&tv, data, sizeof(tv));
> + }
> +
> + /*
> + * On 32-bit architectures, the timeval definition may
> + * use 32-bit or 64-bit members depending on the C
> + * library and architecture.
> + * The cmsg data however always contains a pair of
> + * 32-bit values. Interpret as unsigned to make it work
> + * past y2038.
> + */
> + if (sizeof(long) == 4) {
> + unsigned int *stamp = data;
> + tv.tv_sec = stamp[0];
> + tv.tv_usec = stamp[1];
> + }
> +
> + return tv;
> +}
Should it be something more like
if (sizeof(long) == 8) {
/*
* On 64-bit architectures, the data matches the timeval
* format. Note that on sparc64 this is different from
* all others.
*/
memcpy(&tv, data, sizeof(tv));
} else if (sizeof(long) == 4) {
/*
* On 32-bit architectures, the timeval definition may
* use 32-bit or 64-bit members depending on the C
* library and architecture.
* The cmsg data however always contains a pair of
* 32-bit values. Interpret as unsigned to make it work
* past y2038.
*/
unsigned int *stamp = data;
tv.tv_sec = stamp[0];
tv.tv_usec = stamp[1];
} else {
abort(); /* or some other "sorry, we're not ready for 128-bit or weird architectures yet" failure */
}
return tv;
> static void device_callback(int fd, uint32_t events, void *user_data)
> {
> struct hcidump_data *data = user_data;
> @@ -150,7 +180,7 @@ static void device_callback(int fd, uint32_t events, void *user_data)
> memcpy(&dir, CMSG_DATA(cmsg), sizeof(dir));
> break;
> case HCI_CMSG_TSTAMP:
> - memcpy(&ctv, CMSG_DATA(cmsg), sizeof(ctv));
> + ctv = hci_tstamp_read(CMSG_DATA(cmsg));
> tv = &ctv;
> break;
> }
And libpcap's Linux BT code should do the same thing, changing its memcpy() call?
If you want, you can submit a pull request, or I can make the change.
On Jan 10, 2020, at 1:19 PM, Arnd Bergmann <[email protected]> wrote:
> On Fri, Jan 10, 2020 at 10:05 PM Rich Felker <[email protected]> wrote:
>>
>> On Fri, Jan 10, 2020 at 09:49:03PM +0100, Arnd Bergmann wrote:
>>> musl is moving to a default of 64-bit time_t on all architectures,
>>> glibc will follow later. This breaks reading timestamps through cmsg
>>> data with the HCI_TIME_STAMP socket option.
>>>
>>> Change both copies of hcidump to work on all architectures. This also
>>> fixes x32, which has never worked, and carefully avoids breaking sparc64,
>>> which is another special case.
>>
>> Won't it be broken on rv32 though? Based on my (albeit perhaps
>> incomplete) reading of the thread, I think use of HCI_TIME_STAMP
>> should just be dropped entirely in favor of using SO_TIMESTAMPNS -- my
>> understanding was that it works with bluetooth sockets too.
>
> All 32-bit architectures use old_timeval32 timestamps in the kernel
> here, even rv32 and x32. As a rule, we keep the types bug-for-bug
> compatible between architectures and fix them all at the same time.
>
> Changing hcidump to SO_TIMESTAMPNS would work as well, but
> that is a much bigger change and I don't know how to test that.
If so, maybe I'll just do that for libpcap. Libpcap *does* have an API to request capturing with nanoseconds in tv_usec (and I plan to give it pcapng-flavored APIs to deliver higher-resolution time stamps, as well as metadata such as "incoming" vs. "outgoing", as well).
was On Sat, Jan 11, 2020 at 12:32 AM Guy Harris <[email protected]> wrote:
>
> On Jan 10, 2020, at 1:19 PM, Arnd Bergmann <[email protected]> wrote:
>
> > On Fri, Jan 10, 2020 at 10:05 PM Rich Felker <[email protected]> wrote:
> >>
> >> On Fri, Jan 10, 2020 at 09:49:03PM +0100, Arnd Bergmann wrote:
> >>> musl is moving to a default of 64-bit time_t on all architectures,
> >>> glibc will follow later. This breaks reading timestamps through cmsg
> >>> data with the HCI_TIME_STAMP socket option.
> >>>
> >>> Change both copies of hcidump to work on all architectures. This also
> >>> fixes x32, which has never worked, and carefully avoids breaking sparc64,
> >>> which is another special case.
> >>
> >> Won't it be broken on rv32 though? Based on my (albeit perhaps
> >> incomplete) reading of the thread, I think use of HCI_TIME_STAMP
> >> should just be dropped entirely in favor of using SO_TIMESTAMPNS -- my
> >> understanding was that it works with bluetooth sockets too.
> >
> > All 32-bit architectures use old_timeval32 timestamps in the kernel
> > here, even rv32 and x32. As a rule, we keep the types bug-for-bug
> > compatible between architectures and fix them all at the same time.
> >
> > Changing hcidump to SO_TIMESTAMPNS would work as well, but
> > that is a much bigger change and I don't know how to test that.
>
> If so, maybe I'll just do that for libpcap. Libpcap *does* have an API to request
>capturing with nanoseconds in tv_usec (and I plan to give it pcapng-flavored APIs
> to deliver higher-resolution time stamps, as well as metadata such as "incoming"
> vs. "outgoing", as well).
Sounds good to me, just make sure that you no longer reference
HCI_TIME_STAMP and it should be fine with both old and new libc
versions.
SO_TIMESTAMPNS was first added to the kernel in linux-2.6.22, which
is probably older than anything you care about, but I'm not sure what you
need to do to use it on bluetooth. It appears to have first been added along
with the monitor channel support in linux-3.4.
On a semi-related note, I see that there is a y2038 compatibility in the
libpcap definition of struct pcap_timeval, which uses a *signed* tv_sec
field. I assume the sizes of the fields cannot be changed, but it would
be good to investigate if you can just make it unsigned instead, changing
the supported time range from 1902...2038 to 1970...2106.
It probably makes sense to do this together with supporting nanosecond
timestamps. You could also consider changing the format to use a
64-bit nanosecond value starting at either 1970 or 1902 to give an
even larger range.
Arnd