Messages in this thread | | | Date | Mon, 04 Aug 2008 16:11:19 +0200 | From | Stefan Richter <> | Subject | Re: [PATCH 2.6.27-rc1 2/25] mdb: Merkey's Kernel Debugger 2.6.27-rc1 |
| |
jmerkey@wolfmountaingroup.com wrote: > Netware style debugger for Linux written by Jeffrey Vernon Merkey > > --- a/arch/x86/kernel/traps_32.c 2008-08-03 16:09:54.000000000 -0600 > +++ b/arch/x86/kernel/traps_32.c 2008-08-03 17:33:06.000000000 -0600 ... > @@ -865,6 +896,16 @@ > } > #endif > > +#if defined(CONFIG_MDB) && !defined(CONFIG_KPROBES) > +fastcall void do_int3(struct pt_regs * regs, long error_code) > +{ > +#ifdef CONFIG_MDB > + if (mdb(BREAKPOINT_EXCEPTION, error_code, regs)) > + return; > +#endif > + do_trap(3, SIGTRAP, "int3", 1, regs, error_code, NULL); > +} > +#endif /* CONFIG_KDB && !CONFIG_KPROBES */
The inner #ifdef is always true.
> /* > * Our handling of the processor debug registers is non-trivial. > * We do not clear them on entry and exit from the kernel. Therefore > @@ -895,6 +936,10 @@ > trace_hardirqs_fixup(); > > get_debugreg(condition, 6); > +#ifdef CONFIG_MDB > + if (mdb(DEBUGGER_EXCEPTION, error_code, regs)) > + return; > +#endif > > /* > * The processor cleared BTF, so don't mark that we need it set.
I wonder if it would be beneficial to - include an mdb header file in source files like this one, - in the mdb header, define mdb() as a static inline function which does nothing but return false if CONFIG_MDB is not defined.
Would get rid of the visual clutter introduced by #ifdef/#endif and catch potential syntax errors when people later modify this code. I assume that a compiler is used which eliminates code properly when mdb() is nothing but constant false... -- Stefan Richter -=====-==--- =--- --=-- http://arcgraph.de/sr/
| |