Philippe, Thanks for working on this! I will test it on my app in the next few days.
With regard to your note on userspace apps, I will add that I have had substantial "weird" issues when trying to build Xenomai pthread- skin user apps on the Blackfin in the past. I think I got some simple programs to work but was never able to build anything real without getting segfaults or other very bizarre error messages. Sorry I can't provide more details but its been a few months since I've tried and I can't remember the exact problems encountered. Terry On Apr 10, 2007, at 6:54 AM, Philippe Gerum wrote: > On Sun, 2007-04-08 at 10:26 -0600, Terry Laurenzo wrote: >>> >>> If you could help me to test and validate the upgraded I-pipe patch >>> over >>> Xenomai on your side, I could provide such patch in a few days. >> >> Absolutely. I need to upgrade and re-test my product with the >> current Xenomai anyway. I would be happy to run the Xenomai/IPipe >> tests on my boards and work through any issues. >> > > The following patch applies against the latest uClinux release for the > Blackfin architecture: > http://download.gna.org/adeos/patches/v2.6/blackfin/adeos-ipipe- > bf53x-R07R1RC3-1.6-01.patch > > This is the result of a combined work from PhilW and myself, since I > have upgraded Phil's February patch with the latest -noarch updates > from > my Adeos git tree. Both Xenomai v2.4-devel (i.e. SVN trunk/) and our > current stable tree (v2.3.x/) run over this patch on the bf537 I > have at > hand, so things look good, at least wrt the latency test in user- > space. > > The cited SVN branches from the Xenomai project can be retrieved here: > https://gna.org/svn/?group=xenomai > > Note: While rebuilding some user-space apps, I stumbled upon what > seems > to be a linker bug, so you are likely to get it too when compiling > Xenomai's user-space libs and executables (building cyclictest fails, > "make -k" likely required). Basically, bfin-linux-uclibc-ld (07r1) > raises a segfault when reading an option file passed using the @file > directive. This file contains a list of --wrap options, and some > wrapped > symbols in particular seem to cause this issue. I'll try to produce a > stripped down testcase later. > > -- > Philippe. > > _______________________________________________ Uclinux-dist-devel mailing list Uclinux-dist-devel@blackfin.uclinux.org http://blackfin.uclinux.org/mailman/listinfo/uclinux-dist-devel
