LTP-20050804
– Applied fix from Greg Edwards for 64bit execution.
– self_exec magic required to run child functions on uClinux
– Applied patch from Mike Frysinger:trying to do a build on uClibc will abort in lib/tlibio.c because we dont
provide aio.h find attached a patch which updates the check to include UCLIBC alongside
UCLINUX
– remove call to create.sh script that checks for obscure c++ rpms
– remove all references to and creation of non-std /usr/local/bin/perl5
– fix ballista.cpp to not core dump with std c++ lib
– fix to add librt to MakefileTarget for running aio_suspend test:missing clock_gettime on linux
– remove printf.h and stdio.h from testcases/commands/ade/ld/rd1.c.
– Change to fix the addition of 2 minutes without going over 60 for the seconds
– patch to fix up the install target in disktest to match the install targets of everything else
– patch to fix writetest Makefile to not always rebuild the writetest binary
regardless of whether you ran `make` or `make install` or whatever
– newer toolchains complain about redefining ‘log’ since it’s a math function
provided by the libc find attached a simple patch to rename the ‘log’ variable in
testcases/kernel/ipc/ipc_stress/message_queue_test _04.c to ‘logit’
– when running make in silent mode (make -s) the verbose mode of AR ‘gets in the
way’ attached patch drops the -v and adds -c so that ar wont display the ‘ar:
creating blah.a’ message either
– the current mallocstress.c emits a warning about newsize being used
uninitialized because gcc doesnt detect the abort(0) path
find attached a simple patch to prevent the warning from being issued
– Running nptl01 can fail if the test lasts longer than 300 seconds, patch to lower interations to 100000.
– Fixed clone04 to return correct failure code.”
Link: ltp.sf.net
