You say this only happens when a process raises a signal. Maybe the signal trampoline is getting messed up. Do you see the same behaviour with and without a signal handler installed for the raised signal? Regards, Brad ----- Original Message ----- From: "Tim Wu" <chtimwu@libra.seed.net.tw> To: "Bradley D. LaRonde" <brad@ltc.com> Cc: "linux-mips" <linux-mips@oss.sgi.com> Sent: Saturday, April 13, 2002 4:02 AM Subject: Re: LEXRA MIPS > I insert a show_registers() into the begining of do_ri(). > It seems that CPU runs into data section. > > $0 : 00000000 1000fc00 00000014 00000000 00000001 00000000 00000000 00000000 > $8 : 0000fc00 7fff7a78 00000000 00000002 00000000 00000000 00000000 00000002 > $16: 00000004 7fff7e10 7fff7e18 00000000 100604b0 00000600 7fff7e28 00000010 > $24: 810d6080 00000000 1005e870 7fff7a58 7fff7db0 7fff7a68 > Hi : 00000000 > Lo : 00000000 > epc : 7fff7aa8 Not tainted > Status: 0000fc0c > Cause : 10000028 > Process ping (pid: 24, stackpage=81ff2000) > Stack: 386d4381 31323100 00000000 00000000 24021000 00000000 00000000 0000fc0c > 2ab39a1c 00000000 00000000 00000000 2ab9f550 00000000 00000200 00000000 > 00000000 00000000 00000003 00000000 7fff7d08 00000000 000000c0 00000000 > 00000000 00000000 0000fc00 00000000 81ff0210 00000000 00000000 00000000 > 00000003 00000000 00000000 00000000 00000000 00000000 00000000 00000000 > 00000002 ... > Call Trace: > Code: 00000000 00000003 00000000 <7fff7d08> 00000000 000000c0 00000000 > 00000000 00000000 > > > > Bradley D. LaRonde wrote: > > > ----- Original Message ----- > > From: "Tim Wu" <chtimwu@libra.seed.net.tw> > > To: "linux-mips" <linux-mips@oss.sgi.com> > > Sent: Friday, April 12, 2002 3:29 AM > > Subject: LEXRA MIPS > > > > > > > >>I traced the kernel source and found SIGILL is sent by the exception > >>handler, do_ri(). > >> > > > > Can you tell what/where the reserved (illegal) instruction is that causes > > the trap? > > > > Regards, > > Brad > > > > > > > > >
