remote debugging: "call" just leads to "Cannot insert breakpoint 0"

remote debugging: "call" just leads to "Cannot insert breakpoint 0"

[Steffen Dettmer]

Hi,

using gdb-arm-elf (non-linux) running on x86 linux I think remote
functions can be invoked (such as "(gdb) print strlen(str)"), but
I just get "Cannot insert breakpoint 0".

Why is gdb trying to break at 0?

Any ideas what could cause this?

Do I do something wrong?
Do I need to setup anything special?

This issue was reported for a wide range of gdb versions (6.4,
6.8, 7.1/7.2), actually never worked for us, but I think it
should work, shouldn't it?

The trace shows that when you type the "call" command, GDB sets
up the registers, sets a breakpoint (to be reached at the end of
the function execution, I believe), and resumes the execution.
The problem is that the breakpoint is always set at address 0, on
4 bytes:

"Sending packet: $Z0,0,4#46...Ack"
-> It makes no sense to set such a breakpoint at address 0, and
BTW this breakpoint is never reached. It fails immediately
because it is impossible to set this breakpoint.

p &toto
p &toto
$1 = (int ()) 0xc0000248 <toto>
set debug remote 1
set debug remote 1
call toto()
call toto()
Sending packet: $Pe=00000000#72...Ack
Packet received: OK
Packet P (set-register) is supported
Sending packet: $Pd=980702c0#be...Ack
Packet received: OK
Sending packet: $Pf=480200c0#b4...Ack
Packet received: OK
Sending packet: $Z0,0,4#46...Ack
Packet received: E00
Sending packet: $Pd=a00702c0#de...Ack
Packet received: OK
Sending packet: $Pe=100200c0#a8...Ack
Packet received: OK
Sending packet: $Pf=520200c0#af...Ack
Packet received: OK
Warning:
Cannot insert breakpoint 0.
Error accessing memory address 0x0: Unknown error 4294967295.


Any hints appreciated.

oki,

Steffen

Re: remote debugging: "call" just leads to "Cannot insert breakpoint 0"

[Pedro Alves]

On Monday 20 September 2010 15:41:45, Steffen Dettmer wrote:
> Hi,
> 
> using gdb-arm-elf (non-linux) running on x86 linux I think remote
> functions can be invoked (such as "(gdb) print strlen(str)"), but
> I just get "Cannot insert breakpoint 0".
> 
> Why is gdb trying to break at 0?
> 
> Any ideas what could cause this?

GDB defaults to place the call dummy breakpoint at the entry point.
Your binary probably has entry point == 0.

> 
> Do I do something wrong?
> Do I need to setup anything special?
> 
> This issue was reported for a wide range of gdb versions (6.4,
> 6.8, 7.1/7.2), actually never worked for us, but I think it
> should work, shouldn't it?
> 
> The trace shows that when you type the "call" command, GDB sets
> up the registers, sets a breakpoint (to be reached at the end of
> the function execution, I believe), and resumes the execution.
> The problem is that the breakpoint is always set at address 0, on
> 4 bytes:
> 
> "Sending packet: $Z0,0,4#46...Ack"
> -> It makes no sense to set such a breakpoint at address 0, and
> BTW this breakpoint is never reached. It fails immediately
> because it is impossible to set this breakpoint.
> 
> p &toto
> p &toto
> $1 = (int ()) 0xc0000248 <toto>
> set debug remote 1
> set debug remote 1
> call toto()
> call toto()
> Sending packet: $Pe=00000000#72...Ack
> Packet received: OK
> Packet P (set-register) is supported
> Sending packet: $Pd=980702c0#be...Ack
> Packet received: OK
> Sending packet: $Pf=480200c0#b4...Ack
> Packet received: OK
> Sending packet: $Z0,0,4#46...Ack
> Packet received: E00
> Sending packet: $Pd=a00702c0#de...Ack
> Packet received: OK
> Sending packet: $Pe=100200c0#a8...Ack
> Packet received: OK
> Sending packet: $Pf=520200c0#af...Ack
> Packet received: OK
> Warning:
> Cannot insert breakpoint 0.
> Error accessing memory address 0x0: Unknown error 4294967295.
> 
> 
> Any hints appreciated.
> 
> oki,
> 
> Steffen
> 

-- 
Pedro Alves