1. These days I am studying the exception handler with
kdb. I add some breakpoints into the kernel code,
press "go" and result in the machine reboot. The
detailed steps are listed as below:
  1. enter KDB mode.
  2. id divide_error
     The ASM codes on my screen are:
        0xc010926c divide_errorpush $0x0
        0xc010926e divide_error+0x2push $0xc0109890
        ...
        0xc01092a4 error_code+0x30and %esp,%ebx
        0xc01092a6 error_code+0x32call *%edi
        ...
  3. bp 0xc01092a6 /* the call instruction */
  4. go

I run this test on my p4 and p3 machines with kernel
2.4.2 and 2.4.14 with corresponding KDB patch. The
results are same: reboot.

Is it a bug?

2. I test bph command as below:
   bph testkdbbph datar /*testkdbbph is a variant name
in one of my own module*/
   then press go;

   I call some function in my own module to modify the
testkdbbph value and expect the breakpoint invoked.
But no response, why?

Thanks.

jiangang


Reply via email to