Hi Andreas, yes, this works perfectly, thank you. The modification immediately eliminated my task delays. However, this modification requires some additional modifications in the driver, like described in http://kerneltrap.org/mailarchive/linux-kernel/2008/1/8/548252 But that's easy now ! Claus On Wednesday 28 January 2009 09:09:37 am Lampersperger Andreas wrote: > Hello, > > try to use .unlocked_ioctl instead of .ioctl for your drivers > file_operations struct. > > Andreas > > > -----Original Message----- > > From: linux-rt-users-owner@xxxxxxxxxxxxxxx [mailto:linux-rt-users- > > owner@xxxxxxxxxxxxxxx] On Behalf Of Claus Gindhart > > Sent: Wednesday, January 28, 2009 8:34 AM > > To: linux-rt-users@xxxxxxxxxxxxxxx; Gerd Koenig; Peter Widmann > > Cc: klaus weschmann; Jochen Beier > > Subject: Realtime capability of ioctl() vs. read() and write() > > > > I am running a Linux 2.6.24.7-rt21 Kernel on our HW. > > > > This HW contains a number of IO-Ports, and the drivers for these IO- > > Ports are > > designed in a way, that various ioctl() function provide the interface > > to > > these Ports. > > > > Now, i recognized, that processes, which are using such ioctl()'s, > > absolutely > > do not run in realtime (regardless on the priority). If i cat a huge > > file to > > a /dev/ttyX device, the process jitters for many milliseconds. > > > > This does NOT happen, if i modify the driver and the user space process > > to > > access IO's via read() and write() instead of ioctl()'s. > > > > My question is: Is it a known and intended behaviour, that ioctl()'s > > ruin > > realtime behaviour, and that IO's should only be accessed via read() > > and > > write() ? > > I didnt find any documentation, which would advice me not to use > > ioctl() in > > realtime applications. > > > > -- > > Mit freundlichen Gruessen / Best regards > > > > Claus Gindhart > > SW R&D > > Kontron Modular Computers > > phone :++49 (0)8341-803-374 > > mailto:claus.gindhart@xxxxxxxxxxxxxxxxxxx > > http://www.kontron.com > > -- > > To unsubscribe from this list: send the line "unsubscribe linux-rt- > > users" in > > the body of a message to majordomo@xxxxxxxxxxxxxxx > > More majordomo info at http://vger.kernel.org/majordomo-info.html -- Mit freundlichen Gruessen / Best regards Claus Gindhart SW R&D Kontron Modular Computers phone :++49 (0)8341-803-374 mailto:claus.gindhart@xxxxxxxxxxxxxxxxxxx http://www.kontron.com -- To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
