From: justinmattock@yahoo.com (Justin Mattock) Date: Mon, 19 Dec 2011 20:43:27 -0800 (PST) Subject: [refpolicy] semodule segfaults with ubuntu In-Reply-To: <1320764942.66095.YahooMailNeo@web114310.mail.gq1.yahoo.com> References: <1320681595.98236.YahooMailNeo@web114319.mail.gq1.yahoo.com> <201111081022.29910.russell@coker.com.au> <1320764942.66095.YahooMailNeo@web114310.mail.gq1.yahoo.com> Message-ID: <1324356207.34656.YahooMailNeo@web114308.mail.gq1.yahoo.com> To: refpolicy@oss.tresys.com List-Id: refpolicy.oss.tresys.com ----- Original Message ----- From: Justin Mattock To: "russell at coker.com.au" Cc: SE-Linux Sent: Tuesday, November 8, 2011 7:09 AM Subject: Re: [refpolicy] semodule segfaults with ubuntu ----- Original Message ----- From: Russell Coker To: refpolicy at oss.tresys.com; Justin Mattock Cc: SE-Linux Sent: Monday, November 7, 2011 3:22 PM Subject: Re: [refpolicy] semodule segfaults with ubuntu On Tue, 8 Nov 2011, Justin Mattock wrote: > this is really annoying.. > I am getting a segfault with using git version of refpolicy, but using > ubuntu's policy I dont get this. why? Build it with debugging symbols and run it under gdb. yeah I was using valgrind to see whats really happening. Also this isn't a refpolicy issue, no matter what's wrong with the policy semodule shouldn't SEGV.? There is no excuse for a SEGV, any program which gives one (*) is buggy and needs to be fixed.? Please remove the refpolicy list from follow-ups. I was going to rebuild eglibc, but dont have the time or energy if things crash or burn. (*) Except of course for the unusual cases where hardware errors cause particular programs to SEGV in which case the hardware needs to be fixed/replaced. I hope there is no hardware issues..(but could occur since its an old machine) alright, will play things by ear then.. Thanks for the info! Justin P. Mattock hmm... this is odd I just finished with creating a systemd patch(RFC) will send it out after loading, but I cant seem to load due to a segfault like the above. system setup is simple: qemu with fedora 16 and Mainline policy. I am hitting this with my patch, as well as stock _unpatched_ policy.. I will try a bisect on the policy to see if I can find the bad commit, as well as libc to see what/where things are going wrong. in the meantime I will send out my systemd patch in the matter if anybody wants to have a go at it. Justin P. Mattock -- My Main Blog? ? ? ?? http://etbe.coker.com.au/ My Documents Blog? ? http://doc.coker.com.au/