This is probably due to the very recent release of the newest official version of Xenomai, which includes all the necessary patches. I haven't updated my scripts to reflect this yet.
The ipipe patch is now in the xenomai repository, and doesn't need to be generated from the ipipe repository any more. Feel free to fix this up and send me a pull request :), or I can probably get around to fixing it in the next day or so. On the plus side, not having to generate the ipipe patch will save a *LOT* of drive space, plus less download time for first-time builds. On 10/8/2013 2:14 PM, John Syne wrote: > Hi Charles, > > I downloaded your linux-dev git and checkout the 3.8.13-bone27-xenomai > branch. Everything seems to work fine until I get this error: > > dir: xenomai - ipipe > reference: v3.8.13, kernel version: 3.8.13 > ipipe-core-3.8.13-x86-3.patch > ipipe-core-3.8.13-arm-2.patch > ipipe-core-3.8.13-blackfin-2.patch > ipipe-core-3.8.13-powerpc-2.patch > fatal: can't open patch > '/home/<userid>/Download/GIT/linux-xenomai/ignore/ipipe/ipipe-core-3.8.13-a > rm-1.patch': No such file or directory > > I looked in the ignore/ipipe and I do not see this patch. > > Regards, > John > > > > > > > > On 9/30/13 8:12 PM, "Charles Steinkuehler" <[email protected]> > wrote: > >> AFAIK, preempt-rt on the 3.8 kernel is either brand new or still a work >> in process. The running kernels I'm aware of use a patched version of >> 3.2, but I haven't been following along real close. >> >> In my tests with preempt-rt (many months ago), I was getting pretty >> lousy worst-case latency numbers, so I switched to using Xenomai. Note >> that the latest Xenomai from git (soon to be officially released) >> supports the 3.8.13 kernel as used on the BeagleBone, and has a pre/post >> patch set to make it easy to build the kernel. I've even got kernel >> build scripts if you don't like doing everything from scratch: >> >> https://github.com/cdsteinkuehler/linux-dev/tree/3.8.13-bone27-xenomai >> >> Since I looked at preempt-rt, the OSADL folks have brought up a test >> system with an AM3359 (not a BeagleBone, but the same SoC), and they are >> seeing much better latency numbers than I was. Again, this is still a >> 3.2 kernel, not a 3.8. I'm not sure if the better numbers are due to a >> different kernel config or improved drivers. Preempt-rt relies on the >> Linux kernel code and drivers being written for high-performance SMP >> processing and generally playing nice, while xenomai just preempts the >> entire Linux kernel (including any mis-behaving drivers) so it works >> well with ...um... "sub-optimal" vendor-supplied drivers, which are all >> too common in the ARM space. :) >> >> Let us all know if you get a 3.8 kernel with preempt-rt running on a >> 'Bone (white or black). >> >> On 9/30/2013 8:44 PM, wittend wrote: >>> I am interested in tightening up the performance of the BBB as well. I >>> would like to hear about either your experience with this or that of >>> anyone >>> else who has tried it. >>> >>> Dave >>> >>> On Monday, September 30, 2013 2:32:07 PM UTC-5, jeshu wrote: >>>> >>>> Hello List, >>>> >>>> Previously I have implemented my applications with Real time kernel in >>>> my >>>> Beaglebone White using Robert Nelson's kernel. Now I am going to >>>> change the >>>> platform/board to BBB, I am unable to find any RT patched kernel. Can >>>> anybody post me the links where I can find the PT patched 3.8. kernel ? >>>> >>>> Thanks, >>>> Jeshwanth >>>> Bangalore >>>> >>> >> >> >> -- >> Charles Steinkuehler >> [email protected] >> > > -- Charles Steinkuehler [email protected] -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
