On x86 kernel is not allow using FPU or floating point.
You can utilise floating point emulation.
Here you can find some example
http://www.linuxsmiths.com/blog/?p=253
Ciao
Fede
On 09/26/10 00:41, Jon Elson wrote:
> I'm trying to correct a bug in the hal_ppmc.c driver. There is a place
> where a
A subset of standard C math functions are available when you #include
"rtapi_math.h". This includes fabs(), which is currently used by other
realtime components such as stepgen.
Jeff
--
Start uncovering the many advantag
Jeff Epler wrote:
> A subset of standard C math functions are available when you #include
> "rtapi_math.h". This includes fabs(), which is currently used by other
> realtime components such as stepgen.
>
>
OK, I grep'ed hal/drivers, but didn't think to grep hal/components to
see if they
used a
Federico Manzan wrote:
> On x86 kernel is not allow using FPU or floating point.
> You can utilise floating point emulation.
> Here you can find some example
> http://www.linuxsmiths.com/blog/?p=253
>
I'm sorry if this appears rude, but if you know so little about EMC, you
should not
be comment
From: Pavel Shramov
When call returns INERP_EXECUTE_FINISH value check if call depth
was changed and if so issue execute(0) commands until it returns
to pre-call value.
Uses 0xff to mark command as NULL since it's not possible to set
char[256] to NULL value.
Signed-off-by: Pavel Shramov
---
s
I've played a bit with EMC and found that it's impossible to call function with
some
M66 or probing instructions in MDI mode. To be correct it's possible but stops
at first
instruction returing INTERP_EXECUTE_FINISH. So I'm sending for review patch set
fixing
this problem. It is split into 3 pat
From: Pavel Shramov
Returns call_level value from interpreter
Signed-off-by: Pavel Shramov
---
src/emc/nml_intf/emc.hh |1 +
src/emc/task/emctask.cc | 11 +++
2 files changed, 12 insertions(+), 0 deletions(-)
diff --git a/src/emc/nml_intf/emc.hh b/src/emc/nml_intf/emc.hh
index 4
From: Pavel Shramov
When instruction iterpreted with INTERP_EXECUTE_FINISH return value
is found in O-call set mdi_interrupt flag so subsequent calls to
Interp::execute(0) will continue execution. Also don't reset on
such errors.
Also added call_level function to get current call stack depth.
S
Thanks for your submission--the patches applied without trouble.
Unfortunately, I ran into some problems when testing it.
I modified smartprobe.ngc to be a gcode subroutine, used the
servo_sim.ini sample configuration (but modified to use the AXIS GUI)
and then called smartprobe from mdi:
o ca
On Sun, Sep 26, 2010 at 03:58:24PM -0500, Jeff Epler wrote:
> I found that trying to interrupt the gcode, for instance by hitting ESC
> to abort, or F3 to switch to manual mode, the GUI would become
> unresponsive and the terminal would still show lines like
> LOG(G1 X0.30 Y-0.66 Z0.907
Split mdi flag to wait and next for waiting and feeding respectively.
Abort on all conditions where emcTaskAbort is called.
Signed-off-by: Pavel Shramov
---
src/emc/task/emctaskmain.cc | 37 -
1 files changed, 28 insertions(+), 9 deletions(-)
diff --git a/s
11 matches
Mail list logo