M. Koehrer wrote: > Hi! > > Here are the latest test results: > When I use the standard linux driver e100 instead of eepro100 I never see the > error situation. > I had a test running with more than 17.000.000 cycles and I never saw the > issue.
That's good to know. > I had a closer look into the e100 driver and found out that they do more > things in the ISR > as masking interrupts, reading the status register more than once... > This is slightly different than the eepro100 driver. > Perhaps cloning this behaviour of the e100 could help... > Did you identified the part that may make a difference here? The question is if we should port that stuff over, which may be straightforward when it is about a few lines, or if we should better switch to e100 as basis for RTnet. I also refrained from this step as e100 appeared more complex to trim for RT. But the latter step may help to track future changes better, also when considering that eepro100 will vanish from the kernel on the mid-term. ATM there is only one minor issue pending /wrt e100 and one of our boxes, but the maintainer @Intel promised to work out a fix. Jan PS: Your test case likely revealed a nasty IRQ latency problem in the I-pipe patch. Funny. Philippe and I are on it.
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
