What is -fdata-sections in gcc?

[_maibi]

Hello, my old question: https://devzone.nordicsemi.com/f/nordic-q-a/48087/nrf51822-s130-hardfault-error

I found something.

After removing -fdata-sections in makefile, Hardfault error in my old question never occur. 

What is role of data-sections flag?

I want to know purpose and role of this flag.

From https://gcc.gnu.org/onlinedocs/gcc-2.95.2/gcc_2.html#SEC10

-ffunction-sections-fdata-sectionsPlace each function or data item into its own section in the output file if the target supports arbitrary sections. The name of the function or the name of the data item determines the section's name in the output file.

Use these options on systems where the linker can perform optimizations to improve locality of reference in the instruction space. HPPA processors running HP-UX and Sparc processors running Solaris 2 have linkers with such optimizations. Other systems using the ELF object format as well as AIX may have these optimizations in the future.

Only use these options when there are significant benefits from doing so. When you specify these options, the assembler and linker will create larger object and executable files and will also be slower. You will not be able to use gprof on all systems if you specify this option and you may have problems with debugging if you specify both this option and `-g'.

[awneil]

It's just a way of collecting data together in memory

https://gcc.gnu.org/onlinedocs/gcc-6.1.0/gccint/Sections.html

[_maibi]

Is this flag related on memory alignment?

if so, Is it possible that Hardfault I suffered is related on m0 memory unalignment?

if it make sense, how I can debug where this error occur?

[_maibi]

hi! thanks for response, I checked that this was not a flag problem.

same stack trace occur... I just modified my code a little and compiled it. 

what should I do next step for debugging? It's really hard to me.

[Vidar Berg]

Yes, verify that the function pointer stays the same by using the if check you had earlier (m_conn_params_config.evt_handler). The event number is invalid which may be a sign of a buffer overrun.

[_maibi]

as your mention, I concentrate debugging overriding memory, finally I found the bug!! thank your advice!

[awneil]

_maibi said:

finally I found the bug!

For the benefit of other readers who may come along with similar problems, please describe what was the bug.

And how did you find it?

[_maibi]

This is what happens when I try to access the wrong handler address.

my actual handler address is 0x200010 (example). 

and compare value of variable containing handler address in another module.

it is contaminated partially.

(gdb) print actual_handler

0x200010

(gdb) print m_evt_handler ( in another module)

0x2000ff

so I try to debugging upper m_evt_handler address. 

(gdb) x/256b (m_evt_handler address-0x80)

0x20004a58 <m_read_buf>:        0xff    0xff    0xff    0xff    0xff   0xff    0xff   0xff

.... 

0x20004aXX <m_read_buf>:        0xff    0xff    0xff    0xff    0xff   0xff    0xff   0xff

0x20004aXX <m_read_buf>:        0xff    0xff    0xff    0xff    0xff   0xff    0xff   0xff

0x20004aXX <m_read_buf>:        0xff    0xff    0xff    0xff    0xff   0xff    0xff   0xff

0x20004aXX <m_evt_handler>:        0xff    0x00    0x20    ....

and I check the usage of m_read_buf in my code. and I found overrun bug.