medium memory model debug issues

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

medium memory model debug issues

Dockter, Michael
I have a project that recently pushed our application in to the 20 bit address range. After trying in vain to build on mspgcc 4.6.3 , I installed 4.7.0 and was able to build successfully using -mmemory-model=medium' '-mcpu=430xv2 switches. I am prototyping off a MSP-EXP430F5529LP eval board.

Problem now is getting a stable debug environment. My original setup is Eclipse with CDT and first running mspdebug v0.22 with 'tilib'  and 'gdb' as the arguments for the driver and proxy. I would then start my debug configuration with msp430-gdb and target remote :2000 with the configuration check boxes to load the image and symbols.

msp430-gdb attempts a load but there is an error, from mspdebug;

Client connected from 127.0.0.1:53929
Clearing all breakpoints...
Reading    2 bytes from 0x4400
Writing 8176 bytes to 0x4400
Writing 8176 bytes to 0x63f0
Writing 8176 bytes to 0x83e0
Writing 8176 bytes to 0xa3d0
Writing 8176 bytes to 0xc3c0
Writing 2518 bytes to 0xe3b0
Writing 3062 bytes to 0xed88
Writing   72 bytes to 0xf97e
Writing  128 bytes to 0xff80
Writing 8176 bytes to 0x0000
tilib: MSP430_Memory: Could not perform access, BSL memory segments are protected (error = 62)
Reading    2 bytes from 0x43c4
Reading    2 bytes from 0x43c6
Reading    2 bytes from 0x43c8
Reading    2 bytes from 0x443e
Reading    2 bytes from 0x443e
Reading    2 bytes from 0x4440
Reading    2 bytes from 0x4446
Reading    2 bytes from 0x443e
Reading    2 bytes from 0x443e
Reading    2 bytes from 0x4440
Reading    2 bytes from 0x444c

And from msp430-gdb
~"Loading section .far.text, size 0x3cce lma 0x0\n"

I tried to set the opt enable_bsl_access true in mspdebug with no affect. I also bricked a couple of eval boards trying to turn on the enable_locked_flash_ access option. It also appears that msp430-gdb is loading the lower 64K but then choking when it gets to .fat.text as it doesn't know the proper address or it has overflowed the index in msp430-gdb.

I can however use mspdebug to load the file


(mspdebug) load c:\btstack\eZ430-RF2560\example\spp_counter.elf
load c:\btstack\eZ430-RF2560\example\spp_counter.elf
Writing 4096 bytes at 10000 [section: .far.text]...
Writing 4096 bytes at 11000 [section: .far.text]...
Writing 4096 bytes at 12000 [section: .far.text]...
Writing 3278 bytes at 13000 [section: .far.text]...
Writing 4096 bytes at 4400 [section: .text]...
Writing 4096 bytes at 5400 [section: .text]...
Writing 4096 bytes at 6400 [section: .text]...
Writing 4096 bytes at 7400 [section: .text]...
Writing 4096 bytes at 8400 [section: .text]...
Writing 4096 bytes at 9400 [section: .text]...
Writing 4096 bytes at a400 [section: .text]...
Writing 4096 bytes at b400 [section: .text]...
Writing 4096 bytes at c400 [section: .text]...
Writing 4096 bytes at d400 [section: .text]...
Writing 2438 bytes at e400 [section: .text]...
Writing 3062 bytes at ed88 [section: .rodata]...
Writing   72 bytes at f97e [section: .data]...
Writing  128 bytes at ff80 [section: .vectors]...
Done, 62226 bytes total
(mspdebug)

This however leads to other issues even after loading the symbol file in msp430-gdb it seems almost impossible to get the application to hit the breakpoints unless I manually enter them in msp430-gdb.

Is there any trick to setting this up cleanly or something else I should be doing? I would like to go back to just using mspdebug as proxy only and then loading the file in gdb.

I also get this in gdb when connecting to the proxy, not sure if it's relevant..
target remote :2000
target remote :2000
&"target remote :2000\n"
=thread-group-started,id="i1",pid="42000"
=thread-created,id="1",group-id="i1"
~"_reset_vector__ () at ../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S:118\n"
_reset_vector__ () at ../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S:118
&"118\t../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S: No such file or directory.\n"
118    ../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S: No such file or directory.
*stopped,frame={addr="0x00004400",func="_reset_vector__",args=[],file="../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S",line="118"},thread-id="1",stopped-threads="all"
748^done

TIA


Michael Dockter


------------------------------------------------------------------------------
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos.
Get unparalleled scalability from the best Selenium testing platform available
Simple to use. Nothing to install. Get started now for free."
http://p.sf.net/sfu/SauceLabs
_______________________________________________
Mspgcc-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mspgcc-users
Reply | Threaded
Open this post in threaded view
|

Re: medium memory model debug issues

Eric Decker
the gdb  that is part of 4.7 doesn't work

changes for 20 bit gdb for the msp430 haven't been done (for the 4.7 toolchain)

the new tool chain that redhat/ti is working on maybe more functional but I haven't checked it out yet



>From cPhone


> On May 12, 2014, at 7:49 AM, "Dockter, Michael" <[hidden email]> wrote:
>
> I have a project that recently pushed our application in to the 20 bit address range. After trying in vain to build on mspgcc 4.6.3 , I installed 4.7.0 and was able to build successfully using -mmemory-model=medium' '-mcpu=430xv2 switches. I am prototyping off a MSP-EXP430F5529LP eval board.
>
> Problem now is getting a stable debug environment. My original setup is Eclipse with CDT and first running mspdebug v0.22 with 'tilib'  and 'gdb' as the arguments for the driver and proxy. I would then start my debug configuration with msp430-gdb and target remote :2000 with the configuration check boxes to load the image and symbols.
>
> msp430-gdb attempts a load but there is an error, from mspdebug;
>
> Client connected from 127.0.0.1:53929
> Clearing all breakpoints...
> Reading    2 bytes from 0x4400
> Writing 8176 bytes to 0x4400
> Writing 8176 bytes to 0x63f0
> Writing 8176 bytes to 0x83e0
> Writing 8176 bytes to 0xa3d0
> Writing 8176 bytes to 0xc3c0
> Writing 2518 bytes to 0xe3b0
> Writing 3062 bytes to 0xed88
> Writing   72 bytes to 0xf97e
> Writing  128 bytes to 0xff80
> Writing 8176 bytes to 0x0000
> tilib: MSP430_Memory: Could not perform access, BSL memory segments are protected (error = 62)
> Reading    2 bytes from 0x43c4
> Reading    2 bytes from 0x43c6
> Reading    2 bytes from 0x43c8
> Reading    2 bytes from 0x443e
> Reading    2 bytes from 0x443e
> Reading    2 bytes from 0x4440
> Reading    2 bytes from 0x4446
> Reading    2 bytes from 0x443e
> Reading    2 bytes from 0x443e
> Reading    2 bytes from 0x4440
> Reading    2 bytes from 0x444c
>
> And from msp430-gdb
> ~"Loading section .far.text, size 0x3cce lma 0x0\n"
>
> I tried to set the opt enable_bsl_access true in mspdebug with no affect. I also bricked a couple of eval boards trying to turn on the enable_locked_flash_ access option. It also appears that msp430-gdb is loading the lower 64K but then choking when it gets to .fat.text as it doesn't know the proper address or it has overflowed the index in msp430-gdb.
>
> I can however use mspdebug to load the file
>
>
> (mspdebug) load c:\btstack\eZ430-RF2560\example\spp_counter.elf
> load c:\btstack\eZ430-RF2560\example\spp_counter.elf
> Writing 4096 bytes at 10000 [section: .far.text]...
> Writing 4096 bytes at 11000 [section: .far.text]...
> Writing 4096 bytes at 12000 [section: .far.text]...
> Writing 3278 bytes at 13000 [section: .far.text]...
> Writing 4096 bytes at 4400 [section: .text]...
> Writing 4096 bytes at 5400 [section: .text]...
> Writing 4096 bytes at 6400 [section: .text]...
> Writing 4096 bytes at 7400 [section: .text]...
> Writing 4096 bytes at 8400 [section: .text]...
> Writing 4096 bytes at 9400 [section: .text]...
> Writing 4096 bytes at a400 [section: .text]...
> Writing 4096 bytes at b400 [section: .text]...
> Writing 4096 bytes at c400 [section: .text]...
> Writing 4096 bytes at d400 [section: .text]...
> Writing 2438 bytes at e400 [section: .text]...
> Writing 3062 bytes at ed88 [section: .rodata]...
> Writing   72 bytes at f97e [section: .data]...
> Writing  128 bytes at ff80 [section: .vectors]...
> Done, 62226 bytes total
> (mspdebug)
>
> This however leads to other issues even after loading the symbol file in msp430-gdb it seems almost impossible to get the application to hit the breakpoints unless I manually enter them in msp430-gdb.
>
> Is there any trick to setting this up cleanly or something else I should be doing? I would like to go back to just using mspdebug as proxy only and then loading the file in gdb.
>
> I also get this in gdb when connecting to the proxy, not sure if it's relevant..
> target remote :2000
> target remote :2000
> &"target remote :2000\n"
> =thread-group-started,id="i1",pid="42000"
> =thread-created,id="1",group-id="i1"
> ~"_reset_vector__ () at ../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S:118\n"
> _reset_vector__ () at ../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S:118
> &"118\t../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S: No such file or directory.\n"
> 118    ../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S: No such file or directory.
> *stopped,frame={addr="0x00004400",func="_reset_vector__",args=[],file="../../../../../../../../gcc-4.7.0/libgcc/config/msp430/crt0.S",line="118"},thread-id="1",stopped-threads="all"
> 748^done
>
> TIA
>
>
> Michael Dockter
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Mspgcc-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/mspgcc-users

------------------------------------------------------------------------------
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos.
Get unparalleled scalability from the best Selenium testing platform available
Simple to use. Nothing to install. Get started now for free."
http://p.sf.net/sfu/SauceLabs
_______________________________________________
Mspgcc-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mspgcc-users