Trying to debug LPC1768 project (with FreeRTOS) on hardware with J-Link 
GDB Server I encountered problem: debugger receives SIGTRAP signal on 
instruction SVC 0 (in function vPortStartFirstTask). As if breakpoint 
catches. I can't understand why. I'm trying resume but it triggers 
again. On single step it hangs and j-link shows a lot of these events 
(target halted, started CPU, target halted, etc.). Moreover, I tried 
standard example CORTEX_LPC1768_GCC_RedSuite from FreeRTOS. The same. 
But IAR projects work fine. I'm using Yagarto toolchain. (Tried both old 
arm-elf and latest arm-eabi versions). gdb commands I specified in 
Eclipse debug configuration:
monitor endian little
monitor flash device = LPC1768
monitor flash download = 1
monitor flash breakpoints = 1
monitor speed 1000
monitor reg r13 = (0x00000000)
monitor reg pc = (0x00000004)
Also I checked Reset and Delay and Halt options. Set breakpoint at: main 
and Resume also set and works fine (Program stopped at main, then I 
resume and it stopped on SVC 0)
I worked with ARM7TDMI earlier and didn't experienced such problems.

I figured out what was wrong. Target didn't resetted. "Reset and Delay" 
and "Halt" options have no effect actually. It's a bug/feature in "C/C++ 
GDB Hardware Debugging" (version 7.0.0.201009241320) feature of Eclipse 
CDT. You have to specify "monitor reset 0" command manually.
It's a pity because I could observe it easy as all being watched 
closelly in gdb server messages window and didn't noticed that 
"Resetting target" message doesn't appears anywhere :)