From: Jeff Dike <jdike@ka...> - 2001-03-21 22:10:24
> I am trying to break on uml_net_probe(). Bringing up the kernel
> debugger and setting a breakpoint on uml_net_probe() doesn't seem to
> stop the kernel.
Strange. How about putting breakpoints further up the call chain from
uml_net_probe and see where your ability to set them changes?
Jeff Dike wrote:
> glonnon@... said:
> > I am trying to break on uml_net_probe(). Bringing up the kernel
> > debugger and setting a breakpoint on uml_net_probe() doesn't seem to
> > stop the kernel.
> Strange. How about putting breakpoints further up the call chain from
> uml_net_probe and see where your ability to set them changes?
I put breakpoints all the way up the uml_net_probe call stack... Didn't
hit a single one (uml_net_probe, network_probe, net_device_init,
net_dev_init, device_init, partition_setup).
Then, I put a breakpoint on main.c:do_initcalls(). That doesn't break
either, but the call to main.c:start_context_thread() does break, its
right before the call to main.c:do_initcalls()... hmmm...
So, if I put a break at start_context_thread(), continue to there, and
then put a breakpoint at uml_net_probe() and continue. It works...
Breakpoint 1 at 0x1000a7e4: file panic.c, line 54.
Breakpoint 2 at 0x100994ef: file user_util.c, line 186.
Breakpoint 3 at 0x10001553: file init/main.c, line 531.
Breakpoint 3, start_kernel () at init/main.c:531
(gdb) b start_context_thread
Breakpoint 4 at 0x100176eb: file context.c, line 151.
Breakpoint 4, start_context_thread () at context.c:151
151 kernel_thread(context_thread, NULL, CLONE_FS |
(gdb) b uml_net_probe
Breakpoint 5 at 0x10006c39: file eth_kern.c, line 69.
Breakpoint 5, uml_net_probe () at eth_kern.c:69
Greg Lonnon mailto:glonnon@...