[gpsd-commit-watch] [SCM] GPSD branch, master, updated. release-3.16-4-g7815c41

[Eric S. Raymond]

This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GPSD".

The branch, master has been updated
       via  7815c4184ff6950267f1713dff44f5cba4a903f6 (commit)
       via  4da3d327a8e283228ebd96e1131bb646813ba71a (commit)
      from  63e3a8a3123a1625ae857025ada400e1bf686345 (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit 7815c4184ff6950267f1713dff44f5cba4a903f6
Author: Eric S. Raymond <address@hidden>
Date:   Sat Jan 16 12:06:58 2016 -0500

    In regress-driver, a shelll syntax cleanup.

commit 4da3d327a8e283228ebd96e1131bb646813ba71a
Author: Eric S. Raymond <address@hidden>
Date:   Sat Jan 16 12:04:35 2016 -0500

    Prevent a subtle bug in the SiRF driver that sometimes caused NTP problems.
    
    Simon Hradecky <address@hidden> writes:
    
    I just verified a bug in the driver_sirf.c (version 3.15, I just discovered
    3.16 was released a few days ago, didn't yet check this version), routine
    sirf_time_offset:
    
    Line 637:
    
    double retval = NAN;
    
    This sets the default offset to NAN, which later results in the real time in
    ntp shared memory etc. to be set to 0 with a valid time stamp, further
    causing the NTP daemon to suddenly jump to year 1970, if the SIRF last 
status
    is not known to the driver.
    
    I believe, this default offset would be correct to read:
    
    double retval = 0;
    
    The problem with the real time set to 0 as result of the NAN time offset in
    the SIRF driver (while the clock time was correct) in ntp shared memory
    segment occurs frequently (but not always) with GlobalSat SIRF modules, in
    our case EB5531RE. This can happen at any time of operation, it is possible
    that the gpsd starts out entirely correct with proper time stamps in all
    fields and suddenly forces the NTP to 0 seconds since 1970 because the 
offset
    becomes NAN.

-----------------------------------------------------------------------

Summary of changes:
 driver_sirf.c  |    2 +-
 regress-driver |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)


hooks/post-receive
-- 
GPSD