Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com Message-Id: <13179A03C694D211B84700105A1E6C8B2FCFCF@cfx-server.cfds.harwell.aeat.co.uk> From: "Mortimer, Andy" To: "'cygwin AT sources DOT redhat DOT com'" Subject: RE: sig_send: wait for sig_complete event failed Date: Wed, 20 Sep 2000 11:23:55 +0100 Mime-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Hi, Chris Faylor writes: > On Tue, Sep 19, 2000 at 06:10:35PM +0100, Mortimer, Andy wrote: > >Hi Chris/list, > > > >Chris Faylor writes: > >>On Fri, Sep 15, 2000 at 12:01:52PM +0100, Mortimer, Andy wrote: > >>>1184486202 [sig] bash 28849 sig_send: wait for sig_complete > >>> event failed, sig 11, rc 258, Win32 error 0 > >> [...] > >As noted in another message, the latest snapshots didn't work for me, > >but I've applied just the patch referred to here and it has indeed > >stopped the hang. Thanks, Chris! Now, I'm just left with the problem > >of bash segfaulting inside Cygwin's signal handling thread ... ;-) > > Are you saying that you took the changes from the snapshot, > applied them > to 1.1.4 sources, and rebuilt a DLL? Yes. > Unfortunately, I can't decipher a snapshot if you're using a > home grown > DLL unless you're willing to tell me what all of the 6100* addresses > resolve to symbolically. I know. :-( > For that you can use 'addr2line' or gdb. I'd tried to use gdb to get this information, but I really couldn't persuade it to play with me; I hadn't come across addr2line before. (I even went to the lengths of getting ld to produce a map file, but I didn't get very far with that either :-/ ) The addresses in the previous dump I sent resolve with addr2line to: 6100922E stack_info::init(unsigned long) exceptions.cc:226 6100922E stack_info::init(unsigned long) exceptions.cc:226 6100A2A3 interrupt_on_return(unsigned long, int, sigaction &, void *) exceptions.cc:663 6100A806 call_handler(int, sigaction &, void *) exceptions.cc:778 6100AF00 sig_handle(int) exceptions.cc:972 6103EE0E wait_sig(void *) sigproc.cc:1269 6100331A thread_stub(void *) debug.cc:91 77F04EE8 ?? ??:0 I wasn't sure which DLL to look in for the last entry. With the changes I made to exceptions.cc, the lines referenced above are: 226 sf.AddrPC.Offset = ((DWORD *) ebp)[1]; 663 for (i = 0; i < 32 && thestack++ ; i++) 778 else if (!interrupt_on_return (ebp, sig, siga, handler)) 972 rc = call_handler (sig, thissig, handler); which shouldn't be out by more than a couple of lines from 1.1.4. > Btw, the argv[0] problem should be fixed in the next snapshot. Cool, perhaps I'll try that one out if I get a chance. Thanks! (I'll also try the replacement DLL you sent by private mail if I get a chance, but things are still running here, and it may be a while before I have an opportunity with nothing happening so I can unload the previous DLL! ;-) Thanks for your help, Andy -- Andy Mortimer, CFX-5 Architecture and Infrastructure Team andy DOT mortimer AT aeat DOT com [An automatic footer will be attached to this message. By posting to a public mailing list, I implicitly give permission to read/disclose/copy /delete/reply to/whatever the message.] *********************************************************************** This transmission contains information which may be confidential and which may also be privileged. It is intended for the named addressee only. Unless you are the named addressee, or authorised to receive it on behalf of the addressee you may not copy or use it, or disclose it to anyone else. If you have received this transmission in error please contact the sender. Thank you for your cooperation. *********************************************************************** For more information about AEA Technology please visit our website at http://www.aeat.co.uk AEA Technology plc registered office 329 Harwell, Didcot, Oxfordshire OX11 0QJ. Registered in England and Wales, number 3095862. -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com