- 20 Jan, 2012 2 commits
-
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
- 19 Jan, 2012 2 commits
-
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
ADJ_FREQ_MAX is machine-dependent, but is used by servo.c too, so I declare it in pptp.h as extern and define it for each platform. Only gnu-linux is currently supported
-
- 18 Jan, 2012 4 commits
-
-
Aurelio Colosimo authored
All structs defined; pp_update_delay and pp_init_clock are the first implemented servo functions
-
Aurelio Colosimo authored
From now on, this macro will be used instead of ppi->rt_opts. Old src code should be fixed too.
-
Aurelio Colosimo authored
virtualization of Linux adjtimex system clock time adjustment
-
Aurelio Colosimo authored
-
- 17 Jan, 2012 4 commits
-
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
- 16 Jan, 2012 3 commits
-
-
Aurelio Colosimo authored
msg_issue_xxx functions declared, implemented and used See that pp_get_tstamp now returns a TimeInternal struct. Actually timestamp struct might be defined outside of ieee_1588_types.h file, since it is not defined in the spec. For the moment, I leave it there, but in the future it must be redefined for our purposes (and in standard_c_notation)
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
As a consequence, pp_send_packet is now correctly defined by accepting as input the desired channel type. This leads to an incoherence between send and receive, since in receive we are not interested in which socket received the packet, whereas in send we need to specify it.
-
- 10 Jan, 2012 12 commits
-
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
a general notice is also put in AUTHORS file
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
to_TimeInternal and from_TimeInternal now return int (0 if no errors, -1 in case of error normalize_TimeInternal is now declared as static (was not used externally of arith.c file)
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
-
Aurelio Colosimo authored
previously commented "st_com_add_foreign" header is now removed (not used externally of common-fun.c)
-
- 09 Jan, 2012 13 commits
-
-
Alessandro Rubini authored
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
This change is needed to avoid requiring libgcc in freestanding environments. The div64 function I copied from the kernel only works with unsigned numbers, but arth.h had them signed. I'm not sure at this point whether PTP uses negative nanoseconds in this conversion, so I added a diagnostic message. If it triggers, I can look for a solution, but I suspect it doesn't happen. Please note that a 64-bit arch may define the function as an inline that does the calculation with C operators, but currently arch-gnu-linux is used for both 32-bit and 64-bit hosts. Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
This is the 64-bit unsigned division and remainder from the Linux kernel, for 32-bit host systems. The new lib/ directory can be used by archs to pick files, and that's already exemplified by arch-gnu-linux and arch-bare-linux. Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
Maybe we should consider some generic lib, as most of this can be shared across freestanding archs (we must check whether other freestanding places are really so lib-less os they have these minimal functions for example from newlib) Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
This allows to uncomment the original DBGV in ptp code, until they are turned to serious diagnostic messages. Then these two functions should disappear, in my opinion. Most DBGV use no arguments, so the default function is smaller; and printf-like DBGV can be turned to DBGV_ARGS instead. Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
I removed the trailing "interval" in all of them but announce, that has both receipt and interval. This shortens some almost-80-cols lines. I'm not sure it's the right thing to do, but timer names are unreadable anyways. Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
the "ppi" object instance should consistently be the first argument of functions that receive it. Besides being more consistent, it's smaller and faster too (on processors with more than 4 registers, I mean). Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
-
Alessandro Rubini authored
Boolean is braindead. We decided to leave them in the mandated structures by now (until we discover they are not really mandated) but not use them ourselves. This "found" variable wanted to be int, so be it int. Signed-off-by: Alessandro Rubini <rubini@gnudd.com> Acked-by: Aurelio Colosimo <aurelio@aureliocolosimo.it>
-