From mboxrd@z Thu Jan 1 00:00:00 1970 From: Grant Edwards To: danish iftikhar Cc: fnasser@cygnus.com, gdb@sourceware.cygnus.com Subject: Re: Jeeni & ARM720T with GDB Date: Wed, 22 Nov 2000 06:36:00 -0000 Message-id: <20001122083920.C15146@visi.com> References: X-SW-Source: 2000-11/msg00228.html On Wed, Nov 22, 2000 at 10:35:04AM -0000, danish iftikhar wrote: > After load Pc shows the address : 0x8060 > > The start uo code is : > (gdb) x/30i $pc > 0x8060 : mrs r7, cpsr > 0x8064 : and r7, r7, #31 ; 0x1f > 0x8068 : cmp r7, #19 ; 0x13 > 0x806c : beq 0x80a0 That looks right. > i am not able to do step or cont but using stepi i went through this > startup function . It seems to loop around between 0x815c & 0x8164 and is > not going ahead of that . > the code at this address is as follows : > > 0x8154 : cmp r1, r2 > - 0x8158 : beq 0x8168 > - 0x815c : str r0, [r1], #4 > - 0x8160 : cmp r1, r2 > - 0x8164 : bne 0x815c > - 0x8168 : bl 0x1f194 > - 0x816c : bl 0x1ee10 > > - 0x8170 : bl 0x1f358 It looks to me like the debugger and emulator are working fine. The loop you're looking at is zeroing out the bss section. R2 should contain the address of the end of bss, and R1 should start at the beginning of bss and increment through the whole section. This loop will probably execute tens of thousands of times, so you can get pretty bored doing a stepi... Do the addresses in R1 and R2 correspond to your bss section? Can you set a breakpoint immediately following the loop and do a cont? -- Grant Edwards grante@visi.com >From jimb@zwingli.cygnus.com Wed Nov 22 07:21:00 2000 From: Jim Blandy To: David Taylor Cc: gdb@sourceware.cygnus.com, gdb-patches@sourceware.cygnus.com Subject: Re: pathmap (again) Date: Wed, 22 Nov 2000 07:21:00 -0000 Message-id: References: <200011212037.PAA22391@texas.cygnus.com> X-SW-Source: 2000-11/msg00229.html Content-length: 86 > Comments? It looks reasonable. I'd say, give it a try and see what annoys people. >From taylor@cygnus.com Wed Nov 22 08:54:00 2000 From: David Taylor To: gdb@sourceware.cygnus.com Cc: gdb-patches@sourceware.cygnus.com Subject: pathmap semantics issues Date: Wed, 22 Nov 2000 08:54:00 -0000 Message-id: <200011221653.LAA22540@texas.cygnus.com> X-SW-Source: 2000-11/msg00230.html Content-length: 402 Okay, here's some questions concerning how pathmap should behave: . should it translate all paths? Or just some paths? If some, which? Source paths? Others? That is, should it be consulted when looking for sources?, objects?, executables?, and shared objects? . should it be searched instead of the existing path? Or in addition? If in addition, which should be searched first? Do people care? >From cgf@redhat.com Wed Nov 22 08:58:00 2000 From: Christopher Faylor To: gdb@sources.redhat.com, gdb-patches@sources.redhat.com Subject: Re: pathmap semantics issues Date: Wed, 22 Nov 2000 08:58:00 -0000 Message-id: <20001122115802.B7346@redhat.com> References: <200011221653.LAA22540@texas.cygnus.com> X-SW-Source: 2000-11/msg00231.html Content-length: 769 On Wed, Nov 22, 2000 at 11:53:38AM -0500, David Taylor wrote: >Okay, here's some questions concerning how pathmap should behave: > >. should it translate all paths? Or just some paths? If some, which? >Source paths? Others? That is, should it be consulted when looking >for sources?, objects?, executables?, and shared objects? > >. should it be searched instead of the existing path? Or in addition? >If in addition, which should be searched first? > >Do people care? I would think that it should be "all paths". If you've gone to the effort of issuing the pathmap command then it must be for the reason that you don't have the specific path on your system and need to translate it. So, using the pathmap translation for everything would make sense, IMO. cgf >From msnyder@redhat.com Wed Nov 22 09:17:00 2000 From: Michael Snyder To: David Taylor Cc: gdb@sourceware.cygnus.com, gdb-patches@sourceware.cygnus.com Subject: Re: pathmap semantics issues Date: Wed, 22 Nov 2000 09:17:00 -0000 Message-id: <3A1BFFAF.1E9A@redhat.com> References: <200011221653.LAA22540@texas.cygnus.com> X-SW-Source: 2000-11/msg00232.html Content-length: 550 David Taylor wrote: > > Okay, here's some questions concerning how pathmap should behave: > > . should it translate all paths? Or just some paths? If some, which? > Source paths? Others? That is, should it be consulted when looking > for sources?, objects?, executables?, and shared objects? > > . should it be searched instead of the existing path? Or in addition? > If in addition, which should be searched first? I hope it won't interfere with the path for searching for shared libraries. I've just spent some effort sanifying that... >From ac131313@cygnus.com Wed Nov 22 20:55:00 2000 From: Andrew Cagney To: GDB Discussion Subject: gdb@gnu.org Discussion Date: Wed, 22 Nov 2000 20:55:00 -0000 Message-id: <3A1CA166.A24792EF@cygnus.com> X-SW-Source: 2000-11/msg00233.html Content-length: 148 People on this mailing list should be aware of the discussion: http://sources.redhat.com/ml/overseers/2000-q4/threads.html#00214 sorry, Andrew >From d_iftikhar@hotmail.com Wed Nov 22 23:15:00 2000 From: "danish iftikhar" To: grante@visi.com Cc: fnasser@cygnus.com, gdb@sourceware.cygnus.com Subject: Re: Jeeni & ARM720T with GDB Date: Wed, 22 Nov 2000 23:15:00 -0000 Message-id: X-SW-Source: 2000-11/msg00234.html Content-length: 3399 Hi i went through the code and checked the addresses of bss as well as other sections and they are perfectly o.k. ya , grant , it's zeroing the bss address only and i was able to come out of that loop . where it stucks is in : take a look at code below : > > 0x8154 : cmp r1, r2 > > - 0x8158 : beq 0x8168 > > - 0x815c : str r0, [r1], #4 > > - 0x8160 : cmp r1, r2 > > - 0x8164 : bne 0x815c > > - 0x8168 : bl 0x1f194 > > > - 0x816c : bl 0x1ee10 > > > > - 0x8170 : bl 0x1f358 things seem to work perfectly fine for h/w initialisation . i wonder whether some special options have to be configured in ecos build to make it work. i have a basic doubt ..how the diagnostic output is handled . i don't think that jeeni will be able to send it back to console .. or will connecting through the serial port work . is the diag printf 's are the one causing it to hang. thanks danish. >From: Grant Edwards >To: danish iftikhar >CC: fnasser@cygnus.com, gdb@sourceware.cygnus.com >Subject: Re: Jeeni & ARM720T with GDB >Date: Wed, 22 Nov 2000 08:39:21 -0600 > >On Wed, Nov 22, 2000 at 10:35:04AM -0000, danish iftikhar wrote: > > > After load Pc shows the address : 0x8060 > > > > The start uo code is : > > (gdb) x/30i $pc > > 0x8060 : mrs r7, cpsr > > 0x8064 : and r7, r7, #31 ; 0x1f > > 0x8068 : cmp r7, #19 ; 0x13 > > 0x806c : beq 0x80a0 > >That looks right. > > > i am not able to do step or cont but using stepi i went through this > > startup function . It seems to loop around between 0x815c & 0x8164 and >is > > not going ahead of that . > > the code at this address is as follows : > > > > 0x8154 : cmp r1, r2 > > - 0x8158 : beq 0x8168 > > - 0x815c : str r0, [r1], #4 > > - 0x8160 : cmp r1, r2 > > - 0x8164 : bne 0x815c > > - 0x8168 : bl 0x1f194 > > > - 0x816c : bl 0x1ee10 > > > > - 0x8170 : bl 0x1f358 > >It looks to me like the debugger and emulator are working fine. >The loop you're looking at is zeroing out the bss section. R2 >should contain the address of the end of bss, and R1 should >start at the beginning of bss and increment through the whole >section. This loop will probably execute tens of thousands of >times, so you can get pretty bored doing a stepi... > >Do the addresses in R1 and R2 correspond to your bss section? > >Can you set a breakpoint immediately following the loop and do >a cont? > >-- >Grant Edwards >grante@visi.com _____________________________________________________________________________________ Get more from the Web. FREE MSN Explorer download : http://explorer.msn.com