From mboxrd@z Thu Jan 1 00:00:00 1970 References: <857546.30206.qm@domain.hid> <490B32EF.8050503@domain.hid> <327543.30247.qm@domain.hid> <490B37D6.8050809@domain.hid> <352894.28773.qm@domain.hid> <490B3E58.3010206@domain.hid> Date: Fri, 31 Oct 2008 18:41:19 +0000 (GMT) From: Gabriele Moabiti MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="0-2118816037-1225478479=:91697" Message-ID: <125728.91697.qm@domain.hid> Subject: Re: [Xenomai-help] floating point (int 16) exception List-Id: Help regarding installation and common use of Xenomai List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Gilles Chanteperdrix Cc: xenomai@xenomai.org --0-2118816037-1225478479=:91697 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable =0A>You mean you want to handle a division by zero without ending the=0A>pr= ogram, and without even switching to secondary mode? Then in this=0A>case, = I am afraid you are on your own.=0A=0Ammm... it's better if i explain the s= cenario to make sense.=0A=0AI am developing a driver for a CNC using xenoma= i.=0ANow this cnc run on dos/win (and linux) and the real-time is based on = nmi (with=0Aa pci card generate the nmi @ 0.125 us - 2ms attached to the pe= ripherals and servodrives).=0ANow we want to switch to linux only changing = the pci cart with an eth card and rtnet to=0Acommunicate.=0A=0AI have alrea= dy done the ring 0 driver on linux (on windows is very similar) in a the *o= rrible* way (I know!), =0Areprogramming the IDT and using cli and sti and u= sing using xenomai only for small task=0Abut we want to use rtnet so I have= to better integrate into the xenomai context and=0Athe only way to do it p= roper is to shift cnc to user rt domain.=0A(NOTE: it's a long story but cnc= , written in asm x86 (500k) need to take the first 16 MB and this=0Ais done= with dirty tricks on gdt on ring 0, absolutely incompatible with linux/xen= omai but=0Ain ring 3 with some mapping this problems disappears of course)= =0A=0AThe cnc need to take different decisions on fpu exceptions:=0A1) the = intel defaults actions on masked fpu exceptions aren't right for cnc=0A2) a= n exception sometimes can occur but this must not stop cnc!=0A (the cnc = register the problem continue if it can be resumed)=0ASo custom exceptions = handling is *really* needed.=0A=0AHowever I think a robust hard real time a= pplication may need to remain in rt even if an=0Afpu exception is raised an= d a custom exception handling is not so strange.=0A=0A Gabriele=0A=0A=0A= Unisciti alla community di Io fotografo e video, il nuovo corso di fo= tografia di Gazzetta dello sport:=0Ahttp://www.flickr.com/groups/iofotograf= oevideo --0-2118816037-1225478479=:91697 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable

<= br>>You mean you want to handle a division by zero without ending the>program, and without even switching to secondary mode? Then in this>case, I am afraid you are on your own.

mmm... it's better if i = explain the scenario to make sense.

I am developing a driver for a C= NC using xenomai.
Now this cnc run on dos/win (and linux) and the real-t= ime is based on nmi (with
a pci card generate the nmi @ 0.125 us - 2ms a= ttached to the peripherals and servodrives).
Now we want to switch to li= nux only changing the pci cart with an eth card and rtnet to
communicate= .

I have already done the ring 0 driver on linux (on windows is very similar) in a the *orrible* way (I know!),
reprogramming the IDT and u= sing cli and sti and using using xenomai only for small task
but we want= to use rtnet so I have to better integrate into the xenomai context andthe only way to do it proper is to shift cnc to user rt domain.
(NOTE: = it's a long story but cnc, written in asm x86 (500k) need to take the first= 16 MB and this
is done with dirty tricks on gdt on ring 0, absolutely i= ncompatible with linux/xenomai but
in ring 3 with some mapping this prob= lems disappears of course)

The cnc need to take different decisions = on fpu exceptions:
=0A1) the intel defaults actions on masked fpu except= ions aren't right for cnc
=0A2) an exception sometimes can occur but thi= s must not stop cnc!
=0A (the cnc register the problem= continue if it can be resumed)
=0ASo custom exceptions handling is *rea= lly* needed.
=0A
However I think a robust hard real time application = may need to remain in rt even if an
fpu exception is raised and a custom= exception handling is not so strange.

Gabriele

=0A=0A=0A=0A

Scopri la community di Io fotografo e video
=0AIl nuovo corso di Gazzetta dello sport per= diventare veri fotografi! --0-2118816037-1225478479=:91697--