From: Ø. H. <oyv...@zy...> - 2008-10-24 12:27:51
|
On Fri, Oct 24, 2008 at 12:24 PM, Spen <sp...@sp...> wrote: >> > Also try the attached patch, which ignores read errors from >> the target >> > entirely, even if the address+length is nonsensical. >> >> Committed. >> >> gdb_report_data_abort now ignores all errors upon reading >> target memory. >> > > Not keen on this approach. Me neither. > Should we not try and trace the cause rather than accepting it? I've already got bug reports out there to GDB.... not sure what more we can do... I tried understanding these issues in GDB, but they are very non-trivial. > > I did have similar problems with an older gdb turned out after further > debugging the problem was caused by ld > > are there still issues when using gdb directly, rather than through gdbmi. > > Cheers > Spen > > -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 XScale Cortex JTAG debugger and flash programmer |