From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Weinberger Subject: Re: Mixing cobalt and alchemy? Date: Wed, 06 Mar 2019 21:20:17 +0100 Message-ID: <1763748.uoOIHctIBT@blindfold> In-Reply-To: References: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" List-Id: Discussions about the Xenomai project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jan Kiszka , xenomai@xenomai.org Am Mittwoch, 6. M=E4rz 2019, 18:25:56 CET schrieb Jan Kiszka: > On 06.03.19 10:59, Richard Weinberger via Xenomai wrote: > > Hi! > >=20 > > I have a Xenomai 2.6 application which makes heavy usage of > > rt_dev_ioctl() and shadow threads > > with rt_task_shadow(). > >=20 > > If I understand the new Xenomai 3 API correctly, rt_dev_ioctl() for > > userspace is gone and the cobalt wrappers should be used such that > > plain ioctl() calls will be routed through libcobalt. > > What I don't fully understand is how to mix these with the alchemy > > skin such that the application can use rt_task_shadow() and friends > > because the cobalt wrapper also wraps pthread_create(). > > The application simply does not fully fit in either cobalt nor alchemy = skins. > >=20 > > Is it legal to modify lib/cobalt/cobalt.wrappers to wrap only read/clos= e/ioctl? > > Or can I directly issue cobalt syscalls from the application? > >=20 >=20 > You can do __RT(ioctl(...)), that's also what the transition kit (trank) = does to=20 > provide rt_dev_ioctl. Good to know! Then I'll use alchemy and the __RT() wrappers. Thanks, //richard