Hi all! I have 2.2.18+RTL3.0 but also with 2.14rtl23 i had the same problem. My hardware is a VME board P2/233 with Tundra pci/vme bridge (model is a VMIC 7591). I'm acquiring data at 4KHz, using as a trigger a VME IRQ (declared as RTirq) in a TundraVME/PCI modified for RTLinux driver. But the problem is this: sometimes, in particualr cpu load conditions (i.e. X11 screen saver (!!!) or doing bigg disk access) I loose some irqs (max 3 or 4 a time). Is this normal? Maybe the PIC is programmed badly(masqueraded for long time) by a (disk) driver? Any ideas? Just to report some measurements: the IRQ response time (from IRQ to the first VME access in my routine that re-enable also the IRQ) is 5micro seconds(measured with the oscilloscope)in cpu free conditions. Loading the CPU with a user process (i.e. XFRACTINT) this value goes to 10 micro seconds (average, it jtter from 5 to 20 micro). Are these values good? If someone is interested on it, I have more detailed measurements (email me if interested) Thanks in advance and excuse me for the complexity (and my bad english) Raffaele -- [rtl] --- To unsubscribe: echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED] -- For more information on Real-Time Linux see: http://www.rtlinux.org/rtlinux/
