On Mon, July 18, 2011 9:55 am, Fabian Oboril wrote: > Sorry, I forgot to add that. The error occurs here (FindInHash function > in mem_dep_unit_impl.cc called from WakeDependents function in > inst_queue_impl): > > 946 > <http://repo.gem5.org/gem5/file/9b48b795bfc5/src/cpu/o3/inst_queue_impl.hh#l946> > InstructionQueue<Impl>::wakeDependents(DynInstPtr &completed_inst) > [...] > > 967 > <http://repo.gem5.org/gem5/file/9b48b795bfc5/src/cpu/o3/inst_queue_impl.hh#l967> > if (completed_inst->isMemRef()) { > > 968 > <http://repo.gem5.org/gem5/file/9b48b795bfc5/src/cpu/o3/inst_queue_impl.hh#l968> > > memDepUnit[completed_inst->threadNumber].wakeDependents(completed_inst); > > 969 > <http://repo.gem5.org/gem5/file/9b48b795bfc5/src/cpu/o3/inst_queue_impl.hh#l969> > completeMemInst(completed_inst); > > > which calls > > 424 > <http://repo.gem5.org/gem5/file/9b48b795bfc5/src/cpu/o3/mem_dep_unit_impl.hh#l424> > MemDepUnit<MemDepPred, Impl>::wakeDependents(DynInstPtr &inst) > [...] > > 431 > <http://repo.gem5.org/gem5/file/9b48b795bfc5/src/cpu/o3/mem_dep_unit_impl.hh#l431> > MemDepEntryPtr inst_entry = findInHash(inst); > > > which calls at some point, > > assert(hash_it != memDepHash.end()); > > which leads to the segmentation fault. > > Best wishes > Fabian > > Am 18.07.2011 16:32, schrieb Nilay: >> On Mon, July 18, 2011 6:48 am, Fabian Oboril wrote: >>> Hi, >>> >>> when I use the actual version (the most recent unstable from today), >>> exactly the same error occurs! Can I upload somewhere my configuration >>> files? >>> >>> Best wishes >>> Fabian >>> >>> >>> Am 14.07.2011 16:19, schrieb Nilay: >>>> On Thu, July 14, 2011 8:02 am, Fabian Oboril wrote: >>>>> Hi guys, >>>>> >>>>> I have a problem with M5 and the FS mode with ALPHA. The version I am >>>>> using is based on a release in February which I modified for my own >>>>> purpose, so may be the problem I face is due to my additions. >>>>> However, >>>>> I >>>>> don't know this, I basically have no idea where the problem comes >>>>> from, >>>>> right now. >>>>> >>>>> If I decrease the size of the L2-Cache (detailed mode with L1& >>>>> L2-Cache) from 2 MByte to 512 KByte I get an segmentation fault when >>>>> booting in FS mode. The last actions thereby are: >>>>> >>>>> init started: BusyBox v1.1.0 (2007.03.04-01:07+0000) multi-call >>>>> binary >>>>> mounting filesystems... >>>>> >>>>> Has anyone seen such a behavior before or has an idea what my be the >>>>> root cause for the segmentation fault? With 2 MByte L2-Cache >>>>> everything >>>>> is working fine... >>>>> >>>> I do not expect booting of the operating system can be affected by the >>>> size of the cache. To figure out where the problem is, try booting >>>> with >>>> unmodified gem5 and 512kb of cache. >>>> >> Use gdb and track where the segmentation fault occurs. Post the function >> call stack trace, that should help in figuring out the problem. >>
Are you saying that the assert failed? Assertion failure is different from a segmentation fault. Again, would it possible for you to provide a call stack trace from gdb? -- Nilay _______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
